Prehistoric Cultural Development and Interregional Interaction in the Tropical Montane Forests of Peru Volume 1 A Dissertation Presented to the Faculty of the Graduate School of Yale University in Candidacy for the Degree of Doctor of Philosophy by Warren Brooks Church Dissertation Director: Richard L. Burger December, 1996 © 1996 by Warren Brooks Church ALL RIGHTS RESERVED Dedicated to my mother and father, Sylvina W. and Thomas T. Church ABSTRACT PREHISTORIC CULTURAL DEVELOPMENT-AND INTERREGIONAL INTERACTION IN THE TROPICAL MONTANE FORESTS OF PERU Warren Brooks Church Yale University 1996 The forested eastern slopes of the Central Andes have been characterized alternately as the impenetrable eastern frontier of Andean Civilization, as an empty migratory corridor, a sparsely populated buffer zone separating Andean and Amazonian populations, and as a remote ecological zone servicing highland political economies. This dissertation presents a new cultural sequence excavated from the stratified site of Manachaqui Cave at the upper edge of Peru's northeastern Andean montane forest. Analyses of the archaeological materials including ceramic, lithic, faunal and botanical remains are featured. These data undermine popular population movement and colonization hypotheses, instead providing evidence for autochthonous montane forest cultural development dating from the Preceramic Period (prior to 2000 B.C.). Manachaqui Cave's location beside a pre-Hispanic road, coupled with ethnographic analogies and ethnohistorical and archaeological evidence indicate that the rockshelter functioned primarily as a wayside camp for persons engaged in long-distance travel and exchange between the Central Andean, Amazonian and Northern Andean regions. The interpretation of Manachaqui Cave as a wayside station allows a unique perspective on the development of interregional interaction before, during and after the emergence of Central Andean civilization. Evidence from Manachaqui Cave and from other localities on the eastern slopes suggests that, rather than a remote frontier, the montane forest was the locus of intense boundary interaction. The mediation of long-distance exchange by autonomous montane forest populations strategically situated along the Andean slopes was vital to the development and maintenance of political economies and exchange networks in adjacent highland and lowland regions throughout prehistory. ACKNOWLEDGEMENTS Many kind individuals and generous institutions provided crucial support getting to Manachaqui Cave, performing the fieldwork there, processing the archaeological materials in Trujillo (Peru), and finally writing the dissertation back in New Haven (USA) . I would first like to thank Dr. Thomas J. Lennon for offering me the opportunity to join the Rio Abiseo National Park Research Project in 1984, and for his ongoing support and encouragement. The 1990 excavations at Manachaqui were funded by a generous donation to the University of ColoradoBoulder's Rio Abiseo Project by Lee and Betsy Turner. Permission to undertake research in and around the Rio Abiseo National Park was granted by the Ministry of Agriculture's Oficina del Director General de Forestal y Fauna in Lima. The 1990 archaeological fieldwork was performed under the auspices of the University of ColoradoBoulder, and authorized by the Institute Nacional de Cultura (INC), Lima, and Resoluci6n Ministerial No. 1514-90-ED. I am grateful to Dr. Rogger Ravines, then of the INC, and to Rolando Paredes for providing valuable guidance. Thanks to Sr. Enrique Goicochea Hoyle and his family for making me feel at home in Lima. In Trujillo, Lie. Ana Maria Hoyle and Dr. Santiago Uceda of the Institute Regional de CulturaRegi6n La Libertad kindly facilitated the proposed fieldwork. iii iv I am grateful to the people of Pataz and Los Alisos for receiving us with such kindness. The 1990 field season benefitted from a team of enthusiastic and dedicated archaeologists from Trujillo that worked long days and weeks under difficult conditions. Edgardo Silva V. served as INC field supervisor, and Victor Pimentel S. assisted direction of the fieldwork. The archaeologists were Cesar A. Cornelio Lecca, Lucia Medina de la Cruz, Monica Panaifo Texiera, Victor Hugo Rios Cisneros and Laura Nacarino Flores. From Pataz, Tito Hurtado B. served as cook and managed the camp, while Macedonia Gonzales, Marcos Salirrosas, Jose Escobedo, Celestino Ramirez, Julio Armas Villalobos, Jorge Alguilar and Mariano Alguilar assisted with the day to day work. Carlos Sillones provided short-wave radio support from Trujillo. The subsequent laboratory work in Trujillo was funded by Yale University's Joseph Albers and Augusta Hazzard Funds, and National Science Foundation Dissertation Research Grant No. DBC-9200799. Laboratory processing of the archaeological materials was overseen with the assistance of Daisy Barreto Cedamanos, while Cesar Cornelio Lecca, Elena Goicochea Diaz and Violeta Chamorro Castillo participated. Lie. Segundo Vasquez S. generously gave me access to his preliminary description and illustrations of ceramics recovered from Manachaqui Cave in 1988. Edwin Blas Carranza drew most of the artifact illustrations with help by v Marisabel Paredes de Pimentel and Manuel Tam Chang. I am most grateful to my parents-in-law, the late Dr. Otto Cedr6n Alva and Sra. Estela Viuda de Cedr6n and their family for allowing me to utilize a large portion of their back yard for the laboratory work. The processing of Manachaqui Cave's faunal and botanical materials, and their analyses by Drs. Jonathan Kent and Deborah Pearsall was funded by the Wenner-Gren Foundation for Anthropological Research Grant No. 5425. Exportation of archaeological materials for analysis was authorized by Resoluci6n Suprema Nos. 007-93-ED, 033-93-ED and 035-93-ED. Ms. Patricia Moore provided invaluable and varied kinds of assistance from the States during this stage of the project. I should also point out that Dr. Kent was in Peru when the chronological analysis of Manachaqui Cave's stratigraphy was completed, and he did not have the opportunity to check the recalculations of phase totals featured in this dissertation. Back in New Haven, support for the writing of my dissertation was provided by a Yale University Mellon Dissertation Writing Fellowship. Many specialists graciously supplied useful suggestions and access to unpublished data at the interpretive stages of my work, especially Drs. Kenneth Young, Donald Rodbell, Ruth Shady S., Warren DeBoer, Theresa Topic, Karen Stothert, MaryElizabeth Reeve, Sabina Hyland, Michael Coe and Frank Hole. vi Thanks also to archaeologists Arthur Rostoker, Ismael Perez C. and Alfredo Melly for offering unpublished information regarding their work. I am certain that I neglect to mention many other individuals whose generosity made this a better thesis. My wife Elke assisted my work on occasion, and showed considerable patience during the dissertation process. The success of this project is in large part owed to the unwavering support and friendship offered by my mentor Dr. Richard Burger and by his wife Lucy Burger Salazar. That I was able to complete the project is due to the loving and unconditional support of my father and mother to whom I have dedicated this dissertation. of course, my responsibility. Any of its short-comings are, TABLE OF CONTENTS ACKNOWLEDGEMENTS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii LIST OF ILLUSTRATIONS ................................ xviii LIST OF PLATES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxix LIST OF TABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . XXX Chapter 1. THE PROBLEM OF MONTANE FOREST PREHISTORY ........... 1 Andean Environments and Culture Areas ............. 11 The Central Andes. . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 The Central Andean Tropical Montane Forests ............................ 17 The Eastern Montane and Premontane Forests ......................... 19 Interpretive Frameworks for Andean Migrations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 Unitary Origins and Population Movements ..... 26 Andean Political Economy, Verticality and Colonization. . . . . . . . . . . . . . . . . . . . . . . . . . . 29 The Problem of Montane Forest Prehistory: Migration or Interaction? ...................... 34 Concepts and Strategies for Evaluating Migration Hypotheses ..................... 39 Contexts for Inquiry: Manachaqui Cave and the Greater Eastern Montane Forest ......... 51 The Alternative Hypothesis: Local Development and interaction ................ 58 2. CIVILIZING MIGRATIONS, VERTICALITY COLONIZATION AND THE EASTERN MONTANE FORESTS .................. 64 Civilizing Migrations and the Eastern Montane Forest. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 Primordial Migrations and the Eastern Montane Forest. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 vii viii Civilizing Chachapoyas ....................... 74 Population Movements and "Cuelap Civilization" ......................... 75 Highlandizing the Montane Forests ....... 78 Neolithic Migrations and the Eastern Montane Forest. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 Quechuas, Maize and the Eastern Montane Forest. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92 Verticality, Colonization and the Eastern Montane Pores t . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 9 The Southeastern Montane Forest .............. 105 The Central Montane Forest ................... 106 Colonization as Huari Political Economy ..................... 108 Verticality Colonies in the Tarma Canyon .......................... 113 The Northeastern Montane Forest .............. 120 Alternative Viewpoints ............................ 125 3. THE PATAZ-ABISEO STUDY AREA: PRESENT AND PAST ...... 129 Life Zones and Land-use in the Pataz-Abiseo Area ............................... 13 6 The Dry Forest Zone ......................... 139 The Moist Montane Zone ...................... 139 The Tropical Alpine Zone .................... 142 The Tropical Montane Rain Forest Zone ....... 143 The Huallabamba and Huallaga Valley Premontane Forests ........................ 146 Ethnohistory of the Pataz-Abiseo Area ............. 148 Southern Chachapoyas ......................... 149 Chachapoyas Culture ..................... 150 ix Chachapoyas Under the Incas ............. 154 Spanish Conquest and Administrative Units in Southern Chachapoyas ......... l55 Demographic Collapse .................... 157 The Central Huallaga Valley .................. 159 Early Missionary Contacts ............... 160 Ethnohistoric Evidence for HighlandLowland Boundaries ......................... 168 Ethnohistoric Evidence for HighlandLowland Interaction ........................ 172 Archaeology of Southern Chachapoyas and the Central Huallaga ............................ 175 Archaeology of the Pataz-Abiseo Area ......... 180 Pataz District .......................... 181 The Abiseo Drainage and Central Huallaga Lowlands ..................... 185 Summary ........................................... 18 8 4. MANACHAQUI CAVE AND ITS ENVIRONMENTAL CONTEXT ...... 195 The Changing Manachaqui Valley Environment ........ 198 Geology and Pleistocene Glaciation ........... 199 Holocene Environmental Change ................ 201 Flora and Fauna .............................. 203 The Manachaqui Valley and Manachaqui Cave ......... 205 Manachaqui Cave .............................. 2 0 6 The Pre-Hispanic Road System ................. 209 Hypothetical Functions of Manachaqui Cave ......... 212 Modern Uses of Highland Rockshelters ......... 213 Ancient Uses of Highland Rockshelters ........ 215 Manachaqui Cave as a Wayside Station ......... 216 X Surnrnary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 2 5 5. EXCAVATIONS AT MANACHAQUI CAVE ..................... 228 The 1988 Excavations .............................. 23 0 The 1990 Excavations .............................. 234 Site Formation and Stratigraphy ................... 238 Site formation ............................... 239 Stratigraphy ................................. 245 Stratum 4 ............................... 247 Stratum 3 ............................... 247 Stratum 2 ............................... 248 Stratum 1 ............................... 249 Archaeological Features, Floors and Hearths ................................ 2 50 Sector A Features and Floors ............ 251 Sector B Features and Floors ............ 261 Changing Hearth Morphology and Function ........... 264 Radiocarbon Evidence for Chronology ............... 266 Sector A Radiocarbon Evidence ................ 266 Sector B Radiocarbon Evidence and Stratigraphic Correlations with Sector A ................................... 269 Materials Analysis ................................ 272 Ceramic Analysis ............................. 273 Lithic Analysis .............................. 277 Analysis of Organic Remains .................. 277 Botanical Analysis ...................... 278 Faunal Analysis ......................... 279 xi 6 . THE MANACHAQUI PHASE . . . • . . • • • . • . . • . • . . . . . . . . . . . . . . . 2 8 2 Manachaqui Phase Antecedents . . . . . . . . . . . . . . . . . . . . . . 282 Manachaqui Phase Ceramics . . . . . . . . . . . . . . . . . . . . . . . . . 285 Ceramic Paste A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285 Morphology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287 Decoration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 0 9 Ceramic Paste Group B . . . . . . . . . . . . . . . . . . . . . . . . 321 Paste B1 • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 322 Paste B2 • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 323 Paste B3 • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 324 Lithic Remains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 2 4 Rocks and Minerals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 326 Botanical Remains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 327 Faunal Remains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 327 Manachaqui Phase Chronology . . . . . . . . . . . . . . . . . . . . . . . 328 Manachaqui Phase Ceramic Relationships ............ 329 Paste A Relationships: The Central Andes . . . . . . . . . . . . . . . . . . . . . . . . . . 331 Paste A Relationships: The Amazonian Lowlands . . . . . . . . . . . . . . . . . . . . . 341 Paste A Relationships: The Northern Andes . . . . . . . . . . . . . . . . . . . . . . . . . 346 Paste B Origins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 352 7. The Sui tacocha Phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 55 Suitacocha Phase Ceramics . . . . . . . . . . . . . . . . . . . . . . . . . 355 Ceramic Paste Group A . . . . . . . . . . . . . . . . . . . . . . . . 357 Morphology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 359 Decoration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 7 2 xii Ceramic Paste Group B . . . . . . . . . . . . . . . . . . . . . . . . 3 8 4 Paste B4 • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 384 Paste B5 • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 3 85 Lithic Remains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 8 6 Rocks and Minerals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 87 Botanical Remains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 87 Faunal Remains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 88 Suitacocha Phase Chronology . . . . . . . . . . . . . . . . . . . . . . . 388 Paste A Relationships: The Central Andes .......... 390 Paste A Relationships: The Amazonian Lowlands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 398 Paste A Relationships: The Northern Andes ......... 404 Paste B Origins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 411 Paste B4 • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 412 Paste B5 • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 413 8. THE COLPAR PHASE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 414 Colpar Phase Ceramics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 415 Ceramic Paste Group A . . . . . . . . . . . . . . . . . . . . . . . . 417 Morphology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 417 Decoration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 421 Ceramic Paste Group B . . . . . . . . . . . . . . . . . . . . . . . . 423 Paste B6 • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 424 Paste B7 • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 424 Paste B8 • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 425 Paste B9 • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 425 Paste B10 • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 42 6 Ceramic Paste C1 • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 42 6 xiii Lithic Remains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 427 Rocks and Minerals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 428 Botanical Remains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 8 Faunal Remains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 429 Radiocarbon Evidence for Colpar Phase Chronology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 9 Paste A Relationships: The Central Andes .......... 430 Paste A Relationships: The Amazonian Lowlands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 443 Paste A Relationships: The Northern Andes ......... 445 Paste B Origins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 448 Paste B6 • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 449 Paste B7 • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 449 Paste B8 • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 450 Paste B9 • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 451 Paste B10 • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 451 Paste C1 Origins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 451 Ceramic Evidence for Colpar Phase Chronology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 452 9. THE EMPEDRADA PHASE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 454 Empedrada Phase Ceramics . . . . . . . . . . . . . . . . . . . . . . . . . . 455 Ceramic Paste Group A . . . . . . . . . . . . . . . . . . . . . . . . 457 Morphology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 457 Decoration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 464 Ceramic Paste Group B . . . . . . . . . . . . . . . . . . . . . . . . 465 Paste B11 • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 465 Paste B12 • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 466 Paste B13 • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 466 xiv Paste B14 • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 466 Paste B15 • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 467 Paste B16 • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 467 Paste B17 • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 468 Paste B18 • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 468 Ceramic Paste Group C . . . . . . . . . . . . . . . . . . . . . . . . 469 Paste C2 a-h . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 7 0 Lithic Remains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 7 4 Rocks and Minerals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 476 Botanical Remains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 77 Faunal Remains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 77 Radiocarbon Evidence for Ernpedrada Phase Chronology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 77 Paste A Relationships: The Central Andes .......... 478 Paste A Relationships: The Amazonian Lowlands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 484 Paste A Relationships: The Northern Andes ......... 486 Paste B Origins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 487 Paste B11 • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 487 Paste B12 • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 488 Paste B13 • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 488 Paste B14 • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 489 Paste B15 • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 490 Paste B16 • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 491 Paste B17 • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 491 Paste B18 • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 492 Paste C2 Origins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 494 XV Paste C2 • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 494 Paste C2 a-h . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 495 Lithic Evidence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 497 Ceramic Evidence for Empedrada Phase Chronology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 499 10. MANACHAQUI CAVE AND MONTANE FOREST MIGRATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 501 Evidence for Migrations from Historical Linguistics . . . . . . . . . . . . . . . . . . . . . . . . . . 502 Chachapoyas Languages . . . . . . . . . . . . . . . . . . . . . . . . 503 Chol6n and Hivito Languages . . . . . . . . . . . . . . . . . . 505 Evidence for Migrations from Physical Anthropology . . . . . . . . . . . . . . . . . . . . . . . . . . . 508 Evidence for Migrations from Archaeology: Manachaqui Cave and the Pataz-Abiseo Area ....... 509 Pre-Lavasen Occupations and the Lavas en Phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 513 The Manachaqui Phase Occupation . . . . . . . . . . . . . . 518 Evidence from Botanical and Faunal Remains . . . . . . . . . . . . . . . . . . . 519 Evidence from Artifacts . . . . . . . . . . . . . . . . . 521 Manachaqui Phase Evidence for Migrations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 527 The Suitacocha Phase Occupation . . . . . . . . . . . . . . 533 Evidence from Botanical and Faunal Remains . . . . . . . . . . . . . . . . . . . . . . . . 533 Evidence from Artifacts . . . . . . . . . . . . . . . . . 534 Suitacocha Phase Evidence for Migrations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 536 The Colpar Phase Occupation . . . . . . . . . . . . . . . . . . 541 Colpar Phase Evidence for Migrations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 542 xvi The Empedrada Phase Occupation . . . . . . . . . . . . . . . 543 Evidence from Botanical and Faunal Remains . . . . . . . . . . . . . . . . . . . . . . 544 Evidence from Artifacts . . . . . . . . . . . . . . . 545 Empedrada Phase Evidence for Migrations . . . . . . . . . . . . . . . . . . . . . . . . . . 546 Summary of Evidence for Migrations at Manachaqui Cave . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 549 Evidence for Colonization in the Eastern Montane Forest . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 552 The Southeastern Montane Forest . . . . . . . . . . . . . . 552 The Central Montane Forest . . . . . . . . . . . . . . . . . . . 553 The Northeastern Montane Forest . . . . . . . . . . . . . . 556 11. MONTANE FOREST CULTURAL DEVELOPMENT AND INTERREGIONAL INTERACTION . . . . . . . . . . . . . . . . . . . . . . . . 560 Pre-Lavasen and Lavasen Phase Evidence for Interaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 563 Manachaqui Phase Evidence for Interaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 565 Suitacocha Phase Evidence for Interaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 570 The Chavin Horizon Hiatus . . . . . . . . . . . . . . . . . . . . 575 Colpar Phase Evidence for Interaction ........ 577 Empedrada Phase Evidence for Interaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 580 Boundary Interaction and the Central Andean Tropical Montane Forests . . . . . . . . . . . . . . . . . . . . . . . . 583 Pre-Chavin Horizon Boundary Interaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 589 Post-Chavin Horizon Boundary Interaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 599 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 608 xvii REFERENCES CITED ........................................ 619 APPENDIX A: ILLUSTRATIONS AND ARTIFACT PROVENIENCE ........................................... 67 6 APPENDIX B: PLATES ...................................... 803 APPENDIX C: TABLES ...................................... 812 APPENDIX D: MACROCHRONOLOGY ............................. 826 APPENDIX E: VESSEL SHAPES AND RIM FORMS BY PHASE ........ 836 APPENDIX F: REPORT ON BOTANICAL REMAINS BY DEBORAH M. PEARSALL ................................ 8 6 8 LIST OF ILLUSTRATIONS Fig. 1. Map of the Central Andes, western Amazonia and the Northern Andes (after Institute Geografico Nacional 1984) . . . . . . . . . . . . . 677 Fig. 2. Map of the Central Andean Montane Forests in Peru (after Young and Leon 1993: Fig. 1). The highland montane forest is not shown ........ 678 Fig. 3. Map of some sites mentioned in the text ......... 679 Fig. 4. Map of the northeastern Peruvian Andes. Shaded area is above 3,000 m (after Institute Geografico Nacional 1984 and NASA 1976) ......... 680 Fig. 5. Map of the Pataz-Abiseo study area and macro-ecological zones described in the text (after Young 1993: Fig. 3 and Young et al. 1994: Fig. 1) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 681 Fig. 6. Profile of the Maraffon-Huallaga divide showing distribution of macro-ecological zones . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 682 Fig. 7. Map of the Manachaqui and Montecristo River valleys and the pre-Hispanic roads ........ 683 Fig. 8. Plan map of the Manachaqui Cave site complex (Site M-1) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 684 Fig. 9. Plan map with detail of Manachaqui Cave (Site M-1A) interior. The principal interior space is shaded . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 8 5 Fig. 10. Map of Manachaqui Cave excavation units 1988-1990 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 686 ·Fig. 11. West profile of Units 1, 2 and 3 . . . . . . . . . . . . . . . 688 Fig. 12. West profile of Units 4, 5, 6 and 35. Features R-2 and R-3 are shown . . . . . . . . . . . . . . . . . 689 Fig. 13. West profile of Units 38' 39, 7 and 8 .......... 690 Fig. 14. East profile of Units 3' 2 and 1 . . . . . . . . . . . . . . . 691 xviii xix Fig. 15. East profile of Units 35, 6, 5 and 4. Feature R-4 is shown ........................... 692 Fig. 16. East profile of Units 7, 8, 39 and 38 .......... 693 Fig. 17. East profile of Units 37, 32 and 28 ............ 694 Fig. 18. South profile of Units 11, 1, 13 and 12. Feature R-5 is shown ........................... 695 Fig. 19. North profile of Units 12 and 13. Feature R-5 is shown and Sector A floors are indicated in right margin ...................... 696 Fig. 20. South profile of Units 21, 20, 19 and 18. Feature R-3 and Floor Z's Hearths 1 and 2 are shown .................................... 697 Fig. 21. North profile of Units 29, 30, 35, 31 and 3 2 ......................................... 69 8 Fig. 22. Schematic diagram illustrating relationships between 1988 excavation levels and strata as seen in Unit 5 and Unit 6 east profiles ....................... 699 Fig. 23. Schematic diagram illustrating relationships between 1990 excavation levels and strata as seen in Unit 31 and Unit 32 north profiles ......................... 700 Fig. 24. Plan view of Feature R-5, Sector A............. 701 Fig. 25. Sector A floors in Units 14-17. a: Floor FF and associated rock-filled hearth; b: Floor EE and associated rock-filled hearth ..... 702 Fig. 26. Sector A floors in Units 14-17. a: Floor CC and associated rock-filled hearth; b: Floor BB and associated hearth ................. 703 Fig. 27. Feature R-7 in rear of Sector A, Units 9 and 10. a: north face of stone wall in Unit 10. b: east profile of Feature R-7. Cave wall protrusion is shaded ................. 704 Fig. 28. Plan view of Manachaqui Cave interior showing rock wall features R-6, R-7 and R-8 ............................................ 705 XX Fig. 29. Sector A floors in Units 14-17. a: Floor AA and associated hearth; b: Floor Z and associated Hearths 1 and 2 . . . . . . . . . . . . . . . . . . . . . 706 Fig. 30. Sector A Floor P and associated hearth in Units 14-17 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 707 Fig. 31. Sector A floors containing hearths with embedded stones. a: Floor S in Unit 17; b: Floor M? in Unit 16 . . . . . . . . . . . . . . . . . . . . . . . . . 708 Fig. 32. Plan views of Sector Brock wall features. a: Feature R-6 in Units 23 and 28 (Rockshelter exterior is shaded). b: Feature R-8 in Units 18 and 19 . . . . . . . . . . . . . . . . . 709 Fig. 33. Sector B Feature R-4 in Unit 31, Level 16. a: Plan view of hearth. b: Profile of hearth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 710 Fig. 34. Manachaqui Phase Paste A body profiles. a-h: globular, semi-carinated and carinated profiles. i-p: same, with applique medial ribs. q-t: rare "stepped" shoulder profiles with shoulder ribs ........... 718 Fig. 35. Manachaqui Phase Paste A, Shape A rim p7ofil~s. a~h: Ri~ 1. i: hypothetical R1m 1 Jar. J-p: R1m 2 . . . . . . . . . . . . . . . . . . . . . . . . . . 719 Fig. 36. Manachaqui Phase Paste A, Shape A rim profiles. a, b, g: Rim 2 decorated rims. c-f: Rim 2 vessels with large orifices. h: hypothetical Rim 2 jar . . . . . . . . . . . . . . . . . . . . . . 720 Fig. 37. Manachaqui Phase Paste A, Shape A rim profiles. a-c: Rim 3. d-f: Rim 4. g-i: Rim 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 21 .Fig. 38. Manachaqui Phase Paste A, Shape A rim profiles. a-e: Rim 6. f-i: Rim 7. j: hypothetical Rim 7 jar. k-m: Rim 8. n -p: Rim 9 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 2 2 Fig. 39. Manachaqui Phase Paste A and B rim profiles. a-e: Shape A Rim 10. f: Shape A Rim 11. g: Shape A Rim 12. h: Shape A Rim 13. i-p: Shape BRim la. q-r: Shape B Rim lb. y-aa: Shape BRim lc . . . . . . . . . . . . . . . . . . . 723 xxi Fig. 40. Manachaqui Phase Paste A, Shape B rim profiles. a: Rim la with handle. b: hypothetical Rim lb jar. c-p: Rim 2a ........... 724 Fig. 41. Manachaqui Phase Paste A, Shape B rim profiles. a, b: Rim 2a profile and top view. c-i: Rim 2b. j-p: Rim 3a. q-w: Rim 3b .......... 725 Fig. 42. Manachaqui Phase Paste A, Shape Brim profiles. a-h: Rim 4. i-p: Rim 5 ............... 726 Fig. 43. Manachaqui Phase Paste A, Shape Brim profiles. a, b: Rim 5 with applique decoration. c: hypothetical Rim 5 jar .......... 727 Fig. 44. Manachaqui Phase Paste A, Shape B rim profiles. a: Rim 5 with rare notched decoration. b-d: Rim 5 with incision and applique decoration ............................ 728 Fig. 45. Manachaqui Phase Paste A, Shape B rim profiles. a-g: Rim 6. h-1: Rim 7. m-o: Rim 8 .......................................... 7 2 9 Fig. 46. Manachaqui Phase Paste A, Shape B rim profiles. a: Rim 8 partially reconstructed j ar . b, c : Rim 9 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 3 0 Fig. 47. Manachaqui Phase Paste A, Shape B Rim 9 profiles ................................. 731 Fig. 48. Manachaqui Phase Paste A, Shape B rim profiles. a, b: Rim 9. c: Rim 9 partially reconstructed vessel. d, e: Rim 10 ............. 732 Fig. 49. Manachaqui Phase Paste A, Shape B rim profiles. a: Rim 10 partially reconstructed jar. b, c: Rim 10 profiles with top views ...... 733 .Fig. 50. Manachaqui Phase Paste A, Shape C rim profiles. a-k: Rim la. 1-p: Rim lb. q: hypothetical Rim la bowl ....................... 734 Fig. 51. Manachaqui Phase Paste A, Shape C rim profiles. a-i: Rim 2a. j-n: Rim 2b. o: hypothetical Rim 2a bowl ....................... 735 Fig. 52. Manachaqui Phase Paste A, Shape C rim profiles. a, b: Rim 3. c-e: Rim 4. f: Rim 5 . g : Rim 6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 3 6 xxii Fig. 53. Manachaqui Phase Paste A, ShapeD rim profiles. a-h: Rim 1. i-1: Rim 2. m: hypothetical Rim 3 bowl ........................ 737 Fig. 54. Manachaqui Phase Paste A, Shape D rim profiles. a-d: Rim 3. e-g: Rim 4. h-j: Rim 5 . k, 1 : Rim 6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 3 8 Fig. 55. Manachaqui Phase Paste A, Shape E rim profiles. a, b: Rim 1. c, d: Rim 2. e-h: Rim 3. i, j: Rim 4. k: Rim 5 ................... 739 Fig. 56. Manachaqui Phase Paste A, Shape E rim profiles. a: Rim 5. b: Rim 6. c: Rim 7 ......... 740 Fig. 57. Manachaqui Phase Paste A miscellaneous shapes. a: spoon fragment. b, c: fragments of unidentified artifacts ...................... 741 Fig. 58. Manachaqui Phase Paste A applique rib decorations. a-d: unembellished medial ribs. e-g: unembellished shoulder ribs. h-m: Notched A medial ribs ..................... 742 Fig. 59. Manachaqui Phase Paste A applique rib decorations. a, b: Notched A medial ribs. c: Notched B shoulder rib. d: Notched B medial rib. e, g: Incised A medial ribs. f: Incised A shoulder rib ...................... 743 Fig. 60. Manachaqui Phase Paste A applique rib and band decorations. a: Incised A medial rib. b-e: Incised B medial and shoulder flanges. f-k: notched bands .................... 744 Fig. 61. Manachaqui Phase Paste A applique band decorations. a, b: notched bands with Incised A ribs. c, d: bands with circular punctations. e, f: bands with ovoid incisions. g: applique serpent head ............ 745 Fig. 62. Manachaqui Phase Paste A assorted applique decorations. a: Incised A rib/band. b: lug with notched medial rib ana notched band. c: incised button on unembellished rib. d: incised button on Incised A medial rib. e, f: snake adornos on unembellished ribs .......................... 746 Fig. 63. Manachaqui Phase Paste A adornos. a: head of amphibean or fish. b: bird wing or fish fin. c: zoomorphic heads .................. 747 xxiii Fig. 64. Manachaqui Phase Paste A incised decoration. a-c: Incised Vessel 1 with zoned punctation. d, e: Incised Vessel 2 with zoned punctation. f: incised sherd ........ 748 Fig. 65. Manachaqui Phase Paste Bl vessels. a: Vessel 1, b: Vessel 2, c: Vessel 3 ............. 749 Fig. 66. Manachaqui Phase Paste Group B. a: Paste Bl' Vessel 4. b: Paste Bl Vessel 5. c: Paste B2 vessel ............................. 750 Fig. 67. Manachaqui and Suitacocha Phase artifacts. a: Manachaqui Phase Paste B3 vessel. b: Manachaqui Phase incised slate disk. c: Suitacocha Phase sherd with net impressions on interior surface ............................ 751 Fig. 68. Suitacocha Phase Paste A, Shape A and B rim profiles. a-c: Shape A Rim 14. d-f: Shape A rim 15. g, h: Shape A Rim 16. i: Shape A Rim 17. j-s: Shape BRim lla ........... 752 Fig. 69. Suitacocha Phase Paste A, Shape Brim profiles. a-c: Rim lla. d-k: Rim llb. 1-r: Rim llc ................................... 753 Fig. 70. Suitacocha Phase Paste A, Shape Brim profiles. a-h: Rim lld. i-1: Rim lle. m-s: Rim llf. t-w: Rim llg ..................... 754 Fig. 71. Suitacocha Phase Paste A, Shapes B and C rim profiles. a-f: Shape B Rim 12. g-1: Shape B Rim 13. m-u: Shape C Rim 7 ............. 755 Fig. 72. Suitacocha Phase Paste A, Shape C rim profiles. a: Rim 7. b-d: Rim 8 ................. 756 Fig. 73. Suitacocha Phase Paste A, Shape C and E rim profiles. a-c: Shape C Rim 9. d: Shape C Rim 10. e-h: Shape E Rim 8. i: Shape E Rim 9. j: Shape E Rim 10. k, 1: Shape E Rim 11 ................................. 757 Fig. 74. Suitacocha Phase Paste A partially reconstructed Shape F Rim 1 vessel ............. 758 Fig. 75. Suitacocha Phase Paste A, Shape F rim profiles. a-n: Rim la. o-v: Rim lb ............. 759 Fig. 76. Suitacocha Phase Paste A, Shape F rim profiles. a-k: Rim 2. 1-n: Rim 3 ............... 760 xxiv Fig. 77. Suitacocha Phase Paste A, Shape F Rim 3 profiles ....................................... 7 61 Fig. 78. Suitacocha Phase Paste A, Shape F rim profiles. a: Rim 3. b-d: Rim 4a. e, f: Rim 4b. g, h: Rim 5. i, j: Rim 6. k: Rim 7 .......................................... 7 6 2 Fig. 79. Suitacocha Phase Paste A, Shape F and X rim profiles. a-c: Shape F Rim 8. d-j: Shape X Rim 1. k-n: Shape X Rim 2. o-r: Shape X Rim 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 6 3 Fig. 80. Suitacocha Phase Paste A rim profiles and basal sherds. a: Shape X Rim 4. b: Shape X Rim 5. c: Shape X Rim 6. d: jar or bowl base. e-h: basal angles from carinated bowls .......................................... 764 Fig. 81. Suitacocha Phase Paste A miscellaneous shapes. a-c: fragments of mammiform vessel legs. d-h: handle fragments? i: unidentified fragment .......................... 7 65 Fig. 82. Suitacocha Phase Paste A sherds with notched and punctate rib or band decoration. a-d: notched. e: parallel notched. f, g: high relief notched. h, i: ovoid punctation. j: round punctation ................ 766 Fig. 83. Suitacocha Phase Paste A sherds with applique and incised decoration. a: round punctation. b, c: unembellished applique. d-f: flanges. g-i: incised lines. j: divergent arrays of parallel lines. k, 1: cross-hatching ................................. 767 Fig. 84. Suitacocha Phase Paste A decorated sherds with stamped circles. a: in row. b, h: with paint. c-f, i-1: with incision. j, k: with punctation ................................ 768 Fig. 85. Suitacocha Phase Paste A decorated sherds. a-f: stamped circles and applique with round punctations. g, h: punctation and incision ....................................... 7 69 Fig. 86. Suitacocha Phase Paste A decorated sherds with punctation and incision ................... 770 XXV Fig. 87. Suitacocha Phase Paste A sherds with punctation. a-h: with punctation and incision. i: punctation and applique with round punctation. j: punctation, incision, notched applique and red paint . . . . . . . . . . . . . . . . . 771 Fig. 88. Suitacocha Phase Paste A assorted decorated sherds. a, b: punctation, incision and notched applique. c, d: combing. e: brushing. f: rouletting. g: incised boss. h: fine-line scratched ........... 772 Fig. 89. Suitacocha Phase Paste A assorted sherds. a, b: mat and/or fabric impressed. c: zoomorphic adorno depicting parrot head. d: zoomorphic adorno {opposum head?}. e: zoomorphic adorno (probably a bat's head and face} . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 773 Fig. 90. Suitacocha Phase Paste A adorno and Group B sherds. a: anthropomorphic head adorno. b: Paste Be. c, d: Paste B4 jar rims. e-g: Paste B4 sherds with punctation and incision. h: Paste B5 sherd from bowl ...... 774 Fig. 91. Colpar Phase Paste A, Shape B and C rim profiles. a-e: Shape BRim 11h. f-h: Shape BRim 11i. i, j: Shape BRim 11j. k-o: Shape C Rim 11. p-s: Shape C Rim 12. t: Shape C Rim 13. u: Shape C Rim 14. v: Shape C Rim 15 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 775 Fig. 92. Colpar Phase Paste A, Shape E rim profiles and decorated sherds. a: Rim 3, b, c: Rim 9. d-j: Rim 12. k, 1: Rim 13. m: Rim 14. n-p: notched applique ............... 776 Fig. 93. Colpar Phase Paste Group B. a, b: Paste B6 rim and basal angle profiles. c-e: Paste ~ rim profiles. f: Paste ~ rim with iridescent red paint. g: Paste B8 rim with white-on-red decoration ............... 777 Fig. 94. Colpar Phase Paste Groups B and C. a, b: Paste B8 rim profiles. c-e: Paste B9 rims with red paint. f: Paste B10 rim with orange-red paint. g: Paste C1 rim with red paint . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 8 Fig. 95. Ernpedrada Phase Paste A Shape E rim profiles. a-c: Rim 3. d-k: Rim 9. 1-s: Rim 15. t-w: Rim 16 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 779 xxvi Fig. 96: Empedrada Phase Paste A Shape E rim profiles. a-d: Rim 17. e-h: Rim 18. i-k: Rim 19. 1-n: Rim 20. o: Rim 21. p, q: Rim 22. r: Rim 23. s: Rim 24 ................... 780 Fig. 97. Ernpedrada Phase Paste A Shape GRim profiles. a-m: Rim 1. . . . . . . . . . . . . . . . . . . . . . . . . . . 781 Fig. 98. Ernpedrada profiles. Rim 4. o: s, t: Rim Phase Paste A Shape G rim a-h: Rim 2. i-k: Rim 3. 1-n: Rim 5. p, q: Rim 6. r: Rim 7. 8. u: Rim 9. v: Rim 11 ............... 782 Fig. 99. Ernpedrada Phase Paste A Shape H rim profiles. a-d: Rim 1. e: Rim 2. f: Rim 3. g : Rim 4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 8 3 Fig. 100. Ernpedrada Phase artifacts. a, b: Paste A spoon fragments. c: unidentified artifact. d-e: notched applique. f: Paste Bu jar rim profile. g: Paste Bu rim profile ......... 784 Fig. 101: Empedrada Phase Paste Group B. a-c: Paste Btl bowl with red paint. d-f: Paste B14 sherds from bowls with red and brown paint ......................................... 7 85 Fig. 102. Empedrada Phase Paste Group B. a, b: Paste B15 bowl with dark red paint, pale orange slip on exterior. c: Paste B16 : bowl with iridescent red paint. d, e: Paste B17 sherds with orange and white paint on brown slip. f: Paste B18 bowl with orange slip and negative-resist decoration ........... 786 Fig. 103. Empedrada Phase Paste Group C2 • a-p: unpainted. q-w: traces of red paint. x: Paste C2a with red paint ...................... 787 Fig. 104. Empedrada Phase Paste Group C2 • a-e: Paste C2a with red paint ...................... 788 Fig. 105. Empedrada Phase Paste Group C2 • a-e: Paste C2 a with red paint ...................... 789 Fig. 106. Empedrada Phase Paste Group C2 • a-e: Paste C2a with red paint. f-n: Paste C2a with orange slip .............................. 790 Fig. 107. Empedrada Phase Paste C2 • a-e: Paste C2b with black paint .............................. 791 xxvii Fig. 108. Empedrada Phase Paste C2 • a-d: Paste C2b with black paint. e-g: Paste C2c with red and black paint ........................... 792 Fig. 109. Empedrada Phase Paste C2 • a-c: Paste C2c with red and black paint. d-f: Paste C7d with tan paint. g, h: Paste C1 e with black paint on orange slip .................... 793 Fig. 110. Empedrada Phase Pastes C2 • a: Paste C2 f with brown slip (bottom), negative-resist smudging (top) and white painted circles on natural tan paste (middle) . b: Paste C2g with red paint on light brown paste. c: Paste C2h unpainted jar rim. d: notched ground slate disk from Unit 10 Level 2 ........ 794 Fig. 111. Pre-Lavasen Phase animal size distribution (NISP=Number of Individual Specimens) .................................... 7 9 5 Fig. 112. Lavasen Phase animal size distribution (NISP=Number of Individual Specimens) ......... 795 Fig. 113. Manachaqui Phase animal size distribution (NISP=Number of Individual Specimens) ......... 796 Fig. 114. Suitacocha Phase animal size distribution (NISP=Number of Individual Specimens) ......... 796 Fig. 115. Colpar Phase animal size distribution (NISP=Number of Individual Specimens) ......... 797 Fig. 116. Empedrada Phase animal size distribution (NISP=Number of Individual Specimens) ......... 797 Fig. 117. Guaman Perna's rendering of a camelid with a cloth or net alforja containing jars for transport (from Guaman Poma 1936:524) ......... 798 Fig. 118. Map showing relative degrees of stylistic similarity to the Manachaqui Phase Paste A style, and the probable sources of Paste Group B sherds. Darkest shading indicates closest similarity to Manachaqui Paste A............................ 799 xxviii Fig. 119. Map showing relative degrees of stylistic similarity to the Suitacocha Phase Paste A style, and the probable sources of Paste Group B sherds. Darkest shading indicates closest similarity to Suitacocha Paste A ............................ 800 Fig. 120. Map showing relative degrees of stylistic similarity to the Colpar Phase Paste A style, and the probable sources of Paste Group B and C sherds. Darkest shading indicates closest similarity to Colpar Paste A............................. 801 Fig. 121. Map showing relative degrees of stylistic similarity to the Empedrada Phase Paste A style, and the probable sources of Paste Group B and C sherds. Darkest shading indicates closest similarity to Empedrada Paste A. Q= source of obsidian ............................ 802 LIST OF PLATES Plate I. View of the Manachaqui Cave site complex from the north side of Manachaqui Valley facing south . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 804 Plate II. Manachaqui Cave and 1988 trench looking east . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 0 5 Plate III. Manachaqui Cave 1990 work in progress ....... 805 Plate IV. Manachaqui Cave Unit 6 north profile 1988 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 806 Plate V. Manachaqui Cave Unit 13 north profile 1990 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 807 Plate VI. Manachaqui Cave Units 14-17 Floor CC, 1990 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 807 Plate VII. Manachaqui Phase Shape B jar (see Fig. 48c) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 808 Plate VIII. Manachaqui Phase Paste B2 (see Fig. 66c) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 808 Plate IX. Ground slate points. m: Manachaqui Phase. K, 1: Suitacocha/Colpar Phase. f, g: Colpar Phase. a-e, h-j: Empedrada Phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 809 Plate X. Suitacocha Phase Shape F jar (see Fig. 74) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 810 Plate XI. Colpar Phase. Paste B9 (see Fig. 94c-e) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 810 Plate XII. Empedrada Phase Paste C2 a kaolin sherds ..... 811 Plate XIII. Ernpedrada Phase projectile points . . . . . . . . . . . 811 xxix LIST OF TABLES Table 1. Soil sediments from Unit 15, Sector A ......... 813 Table 2 . Soil sediments from Unit 36, Sector B ......... 813 Table 3 . Hearth classification, measurements and chronology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 814 Table 4. Radiocarbon dates from Manachaqui Cave ........ 815 Table 5. Manachaqui Phase Shape A rims ................. 816 Table 6. Manachaqui Phase Shape B rims ................. 816 Table 7. Manachaqui Phase Shape Table 8 . Manachaqui Phase Shape D rims ................. 817 Table 9. Manachaqui Phase Shape E rims ................. 817 c rims ................. 817 Table 10. Totals pertaining to rib and flange embellishment techniques ...................... 817 Table 11. Suitacocha Phase Shape A rims ................. 818 Table 12. Suitacocha Phase Shape B rims ................. 818 Table 13. Suitacocha Phase Shape c rims ................. 818 Table 14. Suitacocha Phase Shape E rims ................. 818 Table 15. Suitacocha Phase Shape F rims ................. 819 Table 16. Empedrada Phase Shape E rims .................. 819 Table 17. Empedrada Phase Shape G rims .................. 819 Table 18. Botanical remains by phase .................... 820 Table 19. Fauna of the Lavas en Phase .................... 821 Table 20. Fauna in Lavasen/Manachaqui Strata ............ 821 Table 21. Fauna of the Manachaqui Phase ................. 822 Table 22. Fauna in Manachaqui/Suitacocha Strata ......... 822 Table 23. Fauna of the Suitacocha Phase ................. 822 Table 24. Fauna in Suitacocha/Colpar Strata ............. 823 XXX xxxi Table 25. Fauna of the Colpar Phase ..................... 823 Table 26. Fauna in Colpar/Ernpedrada Strata .............. 823 Table 27. Fauna of the Empedrada Phase .................. 824 Table 28. Animal size by phase. Sums are NISP (Number of Individual Specimens) .............. 825 CHAPTER 1 THE PROBLEM OF TROPICAL MONTANE FOREST PREHISTORY It is ... a zone of tropical conditions wherein Nature takes on terrifying proportions--terrifying luxuriance of vegetation, terrifying onrush of rivers, terrifying animal life. Never, at any rate so far as we now know, have either the lower portions of the ceja de la montana or the montana itself been the seat of any stable and advanced community; on the contrary, they are probably the habitat of arboreal man in an archaic stage of culture beyond which progress is not possible save under very definite suasion from the outside world. {Philip A. Means 1931:22-23) Pitt downed the last of his brandy. "It doesn't seem possible an advanced civilization existed in such a remote region without some kind of outside influence." {Dirk Pitt, action hero in Clive Cussler's novel Inca Gold 1994) This thesis presents a description and interpretation of new archaeological data from the edge of the eastern tropical montane forest in northern Peru. These high- altitude forests are referred to variously in published literature as "cloud forest," "ceja de selva" and "ceja de montana." The eastern tropical montane forest covers the slopes of the Central Andean cordillera between approximately 3,500 and 1,500 m, and marks the environmental and cultural transition from highland Andes to lowland Amazon jungles. In modern-day Peru, economic and cultural linkage between Andean and Amazonian regions is tenuously 1 2 maintained by air traffic, and by half a dozen dangerous, winding roads that are frequently rendered impassible by stormy weather and landslides. Western society conditions us to presume that without airplanes, chainsaws, bulldozers, dynamite and powerful economic incentives, substantial interchange between the Andes and eastern lowlands would be virtually impossible to sustain. This perception of the montane forest as a remote and inhospitable frontier has a long history, and it lingers still, as the quote from Cussler's best-selling tall tale of treasure and treachery in the montane forests of Chachapoyas attests. Traditional views of an empty eastern frontier have been based on assumptions like those listed by J. Scott Raymond (1976:205-206): 1) That the rugged terrain and dense vegetation of the ceja inhibited or prevented travel between the highlands and the Montana. 2) That disease, heat and pests made colonization of the tropical forest difficult, if not impossible. 3) That the poor soils, heavy rainfall and steep slopes of the eastern highlands severely restricted the agricultural potential. 4) That historic and present demographic patterns in the montana are a good indication of the pre-Columbian demography. As Raymond points out, these assumptions have no basis in archaeological evidence. The final assumption regarding demography has been particularly pervasive. Both popular and scholarly perceptions have had to overcome ambivalent and negative characterizations of the "largely unpopulated" montane forest (Steward 1948:508). Through pioneering studies of the 1960s and 70s that documented surprisingly 3 plentiful evidence of pre-Hispanic montane forest occupation, mainstream scholarly thought regarding the forested eastern slopes has been revolutionized such that Lyon (1981:8) has remarked, "Wherever a reasonable amount of research has been carried out we find continuous occupation and utilization of the land from the highlands into the montana." Despite the new perspective provided by a half-dozen projects over the past three decades, scholars still tend to regard the tropical montane forest alternately as: 1) a barrier containing the eastward spread of Andean civilizations, 2) a threshold (or doormat) for early prehistoric upslope and downslope population movements and 3) the intractable recipient of late prehistoric highland agricultural colonies. Each of these positions perpetuates stereotypes of sparse and inconsequential montane forest populations passively witnessing repeated incursions by "culturally superior" migrants bound for elsewhere (RenardCasevitz et al. 1988:5), or colonists from expansionist highland polities. The montane forest's potential for human settlement and utilization was first scientifically evaluated by Tosi (1960:148) who concluded that the montane wet forest (bosque muy humedo montane) constituted a nearly useless, and even dangerous, formation for farmers and herders. He attributed repeated and failed attempts to colonize the upper forest to 4 the excessively humid climate and unusually steep terrain (Ibid.:l54). The region's low economic potential was later reiterated by the Peruvian government's Oficina Nacional de Evaluaci6n de Recursos Naturales (ONERN 1976). Archaeologists have repeated the same themes. (1967:11), for example, wrote that: Lanning "Because of the combination of intense rains and steep mountain sides, the high cloud forest is a dangerous area which may not have been much exploited either for agriculture or as pasture." In reference to the eastern limits of Andean civilization, Lanning remarked that "the Incas failed to import their own civilization to the montana, the Spanish failed in their turn, and the modern Peruvians have failed in spite of repeated attempts" (Ibid.:197). Inca "failure" to conquer the forest lands to the east has been an oft-repeated theme (Means 1931:22-23; Meggers 1954:808; Gade 1972, 1979). Traditional views of historical demography fostered by the lack of documentary evidence for habitation of the upper forests have been further reified by the inability of modern Peruvians (excepting the recent coca boom) to intensively colonize the region. Archaeologists habitually cite this repeated failure as a measure of the montane forest's inhospitable disposition with regard to human utilization (e.g. Bonavia 1968:76; Kauffmann 1986:6-7). Despite slowly shifting perspectives, indigenous montane forest prehistory is still subordinated in the 5 interest of furthering "macro-scale analyses" (Lightfoot and Martinez 1995) of highland Andean cultural evolution and political economy. Scholars working within evolutionist paradigms have postulated Andean-Amazonian population movements to su~port universalist, ecologically-based explanations of cultural innovation and emerging sociopolitical complexity in the Andes, and in Amazonia as well. Andeanists effecting macro-scale analyses of highland political economy view montane forest colonization as an adjunct feature of imperial expansion in the adjacent highlands. The population movement hypotheses characteristically ignore the possibility of montane forest inhabitants, while the colonization models accord them parenthetical treatment at best. Although the aforementioned paradigms continue to shape the predominant perceptions of montane forest prehistory, a small group of archaeologists, mostly Peruvians, has worked outside of these macro-scale frameworks, documenting longterm montane forest occupations and crediting montane forest populations with a dynamic capacity in prehistoric interregional interaction. Ironically, some of these scholars have emphasized the unique trajectory of montane forest cultural development (Morales 1993). Lumbreras (1974:149), in his synthesis of Peruvian prehistory, even suggests that the eastern montane forest represents "a new culture area, distinct from the Central and Northern ones." 6 Such disparity in opinion has few parallels in world archaeology, and it exemplifies the considerable power that theoretical frameworks exert over data interpretation. Despite a recent modest resurgence, migration theories have fallen out of mainstream archaeological discourse. Nonetheless, hypothesized Andean-Amazonian population movements retain a measure of credibility because they highlight anomalous instances of similar material patterning in these highly dissimilar, yet juxtaposed regions (e.g. Tello 1942; Meggers 1971:146-149; Lathrap 1970, 1971, 1974; Rivera 1975, 1991). Documentary evidence supports models postulating montane forest colonial enclaves (Murra 1964, 1967, 1975) that are still formulated and re-formulated today. In effect, these proposed migratory episodes have come to represent the salient developmental events in eastern montane forest prehistory. Therefore, any archaeological treatment of this region must address issues of population intrusion. As archaeology is developing increasingly sophisticated frameworks for examining archaeological evidence from frontiers and boundaries (e.g. Schortman and Urban 1992; Lightfoot and Martinez 1995) a reevaluation of these montane forest models is appropriate and timely. With this thesis, I intend to weigh both new and extant montane forest evidence for hypothesized long-distance population movements and highland colonization against the 7 alternative hypothesis that the tropical montane forest harbored substantial indigenous populations with long histories of in situ cultural development. A re-analysis of evidence once utilized to infer prehistoric migrations can now be shown to demonstrate that montane forest societies participated in complex modes of economic interaction, acting alternately as primary purveyors and intermediary conveyors in long-distance exchange connecting distant Central Andean, Amazonian and Northern Andean highland and lowland regions. An analysis of continuity and change in cultural remains recently excavated from Manachaqui Cave constitutes the core of my thesis. Today, this small rockshelter at the upper edge of the montane forest infrequently lodges hunters and cattle herders from nearby villages. Excavations reveal that Manachaqui Cave was more intensively utilized during prehistory. Because the remains of a road that connected prehistoric highland population centers and tropical forest destinations lies only a few meters to the north, it is reasonable to suppose that Manachaqui Cave once sheltered travelers routinely crossing the ecotone. Manachaqui Cave's stratified deposits contain a record of human utilization beginning approximately 8,000 B.C. and the early Preceramic Period, and ending with the arrival of the Spanish conquistadors in 1532. Period, Initial Period, The Late Preceramic the first half of the Early Horizon, 8 the Early Intermediate Period and the Late Intermediate Period and/or Late Horizon are particularly wellrepresented. Conspicuous gaps in the sequence coincide with the Chavin horizon (500- 200 B.C.), the Middle Horizon (A.D. 700 - 900) and possibly the Late Intermediate Period. Study of the cultural sequence at Manachaqui Cave offers a window on the dynamics of Andean-Amazonian interaction, and related cultural processes before and after the emergence of Chavin civilization. The rockshelter's proximity to modern Ecuador renders a similar opportunity to examine the Central Andes' changing relationship to the Northern Andes. My interpretation of the Manachaqui evidence suggests that popular "verticality" or "zonal complementarity" models are inappropriate frameworks for determining modes and motivations involved in most of the interactions documented at Manachaqui. I call attention to the need for the development and application of alternative models of Andean interaction and exchange that avoid the a priori assumptions that interaction on the forested slopes was always vertical, or even ecologically complementary. It would seem more appropriate to view the so-called montane forest "frontier" as a dynamic "cultural interface" where populations and population segments consciously manipulated styles and symbols to express changing intra-group and inter-group boundary relations, and to facilitate certain kinds of interaction (Hodder 1982; Lightfoot and Martinez 1995:485). 9 Rather than remote peripheries, Central Andean montane forests were loci of creativity where interaction generated innovations with potential developmental ramifications in adjoining regions. The conclusion that autochthonous populations did develop independently in the eastern montane forests is significant for several reasons. First, the archaeological sequence invalidates assumptions that montane forest populations are always intrusive. Second, while theories of early population movements probably cannot be "disproved," they are rendered highly suspect as the burden of proof must now be borne by migration advocates. Third, the long record of interregional interaction suggests that analyses of cultural evolution emphasizing only internally generated processes neglect a significant and dynamic component of Central Andean cultural development. A new awareness of montane forest prehistory is especially crucial to the development of theories that accord interregional interaction a powerful role in stimulating the growth of cultural complexity in the Andes and Amazonia (e.g. Shady and Rosas 1979; Burger 1992). In the following section, I will describe the Central Andes, the Central Andean tropical montane forests, and the eastern tropical montane forest in particular, in terms of geography and environment. I summarize the predominant interpretive frameworks utilized by Andeanists that have 10 shaped our understanding of montane forest prehistory. Emphasized are the roles of evolutionary paradigms with unilinear assumptions of unitary origins, and "verticality" or "complementarity" models of Andean political economy commonly invoked to account for eastern montane forest settlement. Then, by specifying explicit sets of archaeological correlates for hypothetical population movement and colonization scenarios distilled from the extant theoretical literature, I outline a methodology for evaluating the explanatory effectiveness of both models. I point out some difficulties involved in correctly inferring migration from the kinds of evidence available, and I suggest a more fruitful manner of conceptualizing boundary interaction that partially reconciles the deep division between competing migration and interaction explanations. Chapter 2 will describe the proposed population movement and colonization scenarios postulated by Andeanists, as well as the larger theoretical agendas that their hypotheses served. Chapter 3 introduces the Pataz- Abiseo study area, presents a scheme of ecological zonation and describes modern economic activities. An analysis of Pataz-Abiseo culture history offers my interpretations of existing ethnohistoric and archaeological evidence for changing historic and prehistoric demography, settlement patterns, cultural boundaries and interaction networks. Here I present new documentary evidence for politically 11 autonomous montane forest societies, and show how postConquest events transformed the Pataz-Abiseo area from a node of interregional interaction into the "imaginary frontier" described by Lyon (1981). This background sets the stage for description of the Manachaqui site environment and delineation of the working hypothesis that Manachaqui Cave functioned as a wayside station servicing prehistoric interregional travel and commodity transport (Chapter 4) . Chapters 5 through 9 present the archaeological data recovered during the 1988 and 1990 field seasons. Chapter 10 will first present an evaluation of Manachaqui Cave's postulated wayside station function, and then show how the new data support an interpretation of long-term autochthonous development rather than the population movement or colonization hypotheses. In Chapter 11, I implement a theoretical framework emphasizing interregional exchange and boundary interaction to interpret Manachaqui Cave's archaeological record. Utilization of this alternative framework reveals that the tropical montane forests were loci of early Central Andean population nucleation, technological innovations and complex social formations. While migration explanations of culture change tend to stifle further inquiry, an interregional interaction framework provides a starting point for future research. Andean Environments and Culture Areas This thesis deals with an archaeological problem of 12 ample spatial and temporal dimensions, and so must include definition of pertinent geographic regions and cultural sequences for heuristic purposes. The regions are: the Central Andes, western Amazonia and the Northern Andes, but the following paragraphs will focus on environmental distributions in the Central Andes. Geographic, as well as cultural distinctions between the Central Andean, Amazonian and Northern Andean regions (Fig. 1) have been consistently recognized by scholars during the mid-twentieth century (e.g. Steward 1949; Steward and Faron 1959; Willey 1971) although their terminology has varied. The term "Amazonia" refers to the Amazon River Basin below 1,500 mas delineated by Meggers and Evans (1983). corresponds to Lumbreras' The term "Northern Andes" (1981) "Andes Septentrionales" and especially modern Ecuador (of course the prehistoric significance of modern political boundaries should not be assumed). Throughout this thesis, Rowe and Menzel's (1967) Central Andean prehistoric master sequence is employed as the least encumbered by evolutionary assumptions. The sequence is as follows: Preceramic Period (10,000-1,800 B.C.), Initial Period (1,800-900 B.C.), Early Horizon (900200 B.C.), Early Intermediate Period (200 B.C.- A.D. 700), Middle Horizon (A.D. 700-900), Late Intermediate Period (A.D. 900-1350), and Late Horizon (A.D. 1450-1532). Within the Early Horizon, Burger (1988, 1992, 1993) defines the 13 Chavin horizon (500-200 B.C.) as the apex of the Chavin cult's influence and a pivotal event marking the emergence of ranked socio-economic hierarchies and state-level political organization in the Central Andes. The Central Andes The term "Central Andes" as it is utilized in this study refers to the coastal and highland environments within the borders of modern Peru. The southern limits of the Central Andes area are not directly pertinent to this study, but Bennett's (1948) and Burger's (1984a) identification of a northern boundary or buffer zone separating the Peruvian Co-tradition from the Northern Andes justifies their separation as distinct units of analysis in the present study. The antiquity of this boundary and its permeability at different prehistoric moments are issues of current debate (Hocquenghem et al. 1993). The Sechura Desert in extreme north-coastal Peru, the Catamayo Valley in interior Ecuador and the low (2100 m) Andean pass at Porculla are landmarks within the inter-Andean area. Bennett regarded the tropical forest edges as the eastern limit of the Peruvian Co-tradition (Bennett and Bird (1949:95-96), but explorers (e.g. Savoy 1970) and archaeologists (e.g. Raymond 1985, Hastings 1985; Schjellerup 1992) have recently documented evidence of preHispanic Andean settlement deep within Amazonian tropical forests. The degree to which the pre-Hispanic Central 14 Andean montane forests covering the eastern slopes can be conceived of culturally as "Central Andean," "Amazonian" or distinct from either (cf. Lumbreras 1974:149, 1981:32; Kauffmann et al. 1989:6; Morales 1993:642-653) remains to be demonstrated. Problems regarding the eastern boundary's locations, antiquity and permeability are central to this thesis. The contrasting environments of the arid coast and seasonally-moist highlands are usually distinguished in published descriptions of Peru and the Central Andes (e.g. Moseley 1992; Burger 1992) because they afforded differing subsistence opportunities to human inhabitants from earliest times. The cold Humboldt Current which sweeps north against the Peruvian coast before veering out to sea near the Ecuadorian border conditions the coastal climate. The combination of the cold water current, easterly winds and tropical latitudes creates a temperature inversion which results in extreme aridity relieved only by winter fog and drizzle from May to November, especially between 300 and 800 m. The Humboldt Current also supports unusually high quantities of marine fa~,a that comprised a resource of primary dietary importance to coastal dwellers. Cyclic changes in the ocean currents every seven to 10 years generates heavy rains and occasional catastrophic flooding called El Nino. The dry coastal plain separating the steep western 15 flanks of the Central Andean cordillera and the Pacific Ocean narrows southward from 100 kms to less than 20 kms. Aridity correspondingly increases moving south from the humid mangrove swamps at the Ecuadorian border. This coastal desert stretches from Peru's northern border well into Chile, but is broken at regular intervals by river valleys oriented east-west and draining wet-season highland precipitation between October and May. Low, barren and rocky divides separate coastal valley oases where human populations have always clustered, especially since plant cultivation became a subsistence priority. A belt of steep, semi-arid slopes, canyons and thorn forests constitutes the transition from coastal to highland environments. The Andean highlands become higher, wider and drier from north to south. The geographer Troll (1958) distinguishes between the high "puna Andes" of central Peru and Bolivia, and the lower "paramo Andes" of Colombia, Ecuador and northern Peru. Salomon summarizes the distinction between the two Andean landscapes: In the former, the upper slopes are characterized by scant rainfall, strong insolation, and sharp diurnal temperature variation regularly producing nighttime freezes. In the latter, the upper slopes receive considerable rainfall and drizzle but little direct insolation. They do not freeze as regularly. The true "paramo Andes," that is, montane areas in which no puna occurs, extend northward from a line running approximately through Trujillo and Cajamarca, Peru. Some paramo may occur in the eastern cordillera south of this line as well ... (Salomon 1986:24). Frost conditions highland agriculture in the puna Andes, and Salomon documents a subsistence focus on tuber cultivation 16 that contrasts with the maize-based agriculture characteristic of the paramo Andes. Although the puna Andes environment appears to correspond to the Central Andean region, Burger (1984a) emphasizes the lack of a close fit between cultural affiliation and environment in northern Peru. As Salomon observed, the paramo extends well into the Central Andes. East of the Central Andean continental divide, highland waters feed tributaries of the Amazon River. The greatest ecological variation is encountered where the largest Amazon tributaries such as the Marafion, Mantaro, Apurimac and Urubamba Rivers cut deep canyons into the cordillera. Areas of dry, xerophytic vegetation in valley bottoms below approximately 2,300 mare called temple (Rosas and Shady 1974) or yunga in native terminology (Pulgar 1987). Relatively moist intermontane valleys of the quechua between 2,300 and 3,200 m support cultivation of many modern highland staple crops, but the greatest variety of prehistorically important tubers is grown above 3,200 min the jalca or suni. The higher and wider Andean cordillera farther south (below 10° south latitude) includes wide expanses of puna environments (above 4,000 m according to Pulgar 1987) where cultivation was undertaken at substantial risk and camelid pastoralism was a principal pre-Hispanic economic activity. The southern Central highlands surrounding Lake Titicaca constitute a broad plateau with 17 precipitous western and eastern flanks. These environments supported cultural developments of only secondary relevance for this study. The Central Andean Tropical Montane Forests The tropical montane forests are woodland environments found at a maximum altitude of approximately 3,700 m around the peripheries of the Central Andean region. Based upon "topography and predominant climatic and biogeographical influences" (Young and Leon 1993:237}, the greater tropical montane forest has been divided into highland, northern, western-slope and eastern-slope forests (Fig. 2}. In Peru, the implementation of the Holdridge System to classify ecological zones (Tosi 1960; ONERN 1976} has heightened awareness that tropical montane forests were once more extensive than they appear today (Guillet 1985}. At present, the montane forest is most clearly seen as two parallel, altitudinally-restricted belts extending north and south along the western and eastern flanks of the Andean cordillera. On the western side the forest is most clearly defined from the 7th parallel south latitude (near Peru's Chicarna River valley} to northern-most Colombia, while the eastern forest stretches from northernmost Argentina to Venezuela. The western montane "dry" forest has been divided into macrotherrnal (lower: below 1000 m}, mesotherrnal (middle: 1200-2000 m) and oligotherrnal (upper: 2500-3200 m} 18 forests (Valencia 1992). Mesothermal forests like the upper Zafia montane forest described by Dillehay and Netherly (1983) do not occur below the 7th parallel, while oligotherrnal cloud forest fragments extend farther south. The eastern montane forest shades into the lowland tropical forests of greater Amazonia below 1,500 m. Fragmented northern montane forest joins the eastern and western strips at the low saddle separating the Northern and Central Andean massifs. Of the four divisions, the highland forest has suffered the greatest impact from human depredation. Its extent prior to heavy anthropogenic deforestation in the distant past has not been adequately appreciated in the archaeological literature, mostly due to lack of paleoenvironmental field studies. Guillet (1985) describes several processes that contributed to their destruction during both historic and prehistoric eras. Young and Leon (1993:239) estimate that only five to ten percent of the original surface cover remains of Peru's highland Andean montane forests. Continuous expanses of northern and western-slope montane forest become increasingly fragmented south of the Ecuadorian border reflecting both human impact and natural processes (Ibid. :237; Valencia 1992). M. Shimada (1982, 1985) reports ethnohistoric and archaeofaunal evidence for more extensive woodlands in the Cajarnarca Basin during 19 prehistory. She emphasizes the importance of forest- browsing white-tailed deer in the intermontane valley's early subsistence economy. Perhaps little more than 25 percent now remains of the northern montane forest (Young 1992a:58). The western slopes of the northern Peruvian Andes show remnants of previously more extensive montane forest, and the best known forest relic lies in the upper Zafia River valley (Valencia 1992; Alva 1988a). Dillehay et al. (1979) report that the upper Zafia montane forests once supported endemic fauna that included parrots, monkeys, jaguars and boa constrictors. Valencia (1992: Fig. 2) notes that modern western-slope human populations cluster most densely in the oligotherrnal forests around 3200 m. The eastern-slope montane forest is of primary interest for this study. It differs both structurally and physiographically from the western "dry" forest, and it is separated from the western and northern montane forests by the Marafion River Canyon (Young 1992a) . The Eastern Montane and Premontane Forests While scholars in the past maintained simplistic views of the eastern montane forest as an environmental and cultural transition separating the highland punas from the lowland tropical forests, geographers and biologists have recently recognized it as an exceptional and singular 20 environment (Gentry 1992; Leon et al. 1992; Young 1992b, 1993; Young and Leon 1993; Young and Valencia 1992), worthy of study in its own right. Young (1992b:121) characterizes the Central Andean eastern-slope montane forest as "a complex, but fairly natural geographical and biological unit." Geographers and botanists view this Andean forest as a disappearing repository of unparalleled species diversity and local endemism (Gentry 1992:11). That so many floral and faunal species are endemic and restricted to the montane forest underscores its distinctive character and begs the question: what are the implications of this new awareness for interpretations of prehistoric human utilization and settlement? The highest elevations of the forested eastern slopes are often referred to as "cloud forests" because of the dense mists which shroud them almost daily. Moist air moving east to west over the Amazon lowlands rises and cools as it strikes the cordillera, dropping much of its moisture on the eastern flanks, rain forests. thus watering one of earth's wettest As previously noted, the eastern-slope montane forest lies between 1,500 and 3,500 min altitude (Young 1990, 1992b, Young and Leon 1993). Within Peruvian territory, the eastern montane forest stretches 1,500 km north to south, and its width varies from 50 to 250 km (Young 1992b). Young estimates a typical gradient as 3,000 to 4,000 vertical meters within a distance of less than 100 21 km. Only sixty to seventy percent remains of eastern-slope montane forest cover as a result of deforestation (Young and Leon 1993:239). Peru's eastern montane forest can be divided into upper (2,500 to 3,500 m) and lower (1,500 to 2,500 m) tiers (Young 1992b) . Together they are bounded by subalpine grasslands above 3,500 m, and premontane and subtropical forests below 1,500 m. These subalpine grasslands constitute the southern extension of Troll's paramo Andes that stretches down the eastern face of the Central Andean massif. Above this paramo lie barren and rocky alpine life zones, while below it, a belt of fragmented forest (often found between 3,400 and 3,700 m) provides the transition into continuous montane forest. The distinctions which researchers draw between the montane and premontane forests, like other life zones, are based on bioclimatic characteristics such as mean annual biotemperature and precipitation and how these vary seasonally (Holdridge 1967; Hartshorn 1983). While details of biogeographic taxonomies are of marginal importance to this study, Holdridge life zones do often correspond to environments preferred for certain economic activities. For example, the premontane forest between 1,500 and 500 m is the preferred environment for coca cultivation (Plowman 1984). The spatial distributions of eastern slope life zones are complex and locally variable. Many tributaries of the 22 Amazon draining the Peruvian Andes flow south to north (e.g. the Urubamba, Apurimac, Huallaga and Marafion Rivers) complicating what might otherwise be an orderly clinal west to east transition from highlands to lowland tropical forest. Traveling down-slope one expects mean temperatures to rise and the forest canopy to occur at progressively greater heights. However, dendritic river systems create extensive rain shadows (e.g. the upper Tarma and upper Huallaga River valleys) and localized, open highland-like environments. Peculiar juxtapositions of montane forest and dry forest {xerophytic) vegetation are not uncommon. Any study of eastern slope resource utilization or settlement must take many environmental variables into account and focus on local conditions particular to the area in question. Furthermore, present site environments in proximity to the ecotones may have differed substantially during periods of occupation because of ancient or modern deforestation practices and/or paleoclimatic change. For descriptive purposes, Young {1992b: Fig. 1) has divided the eastern slope montane forest of Peru into six "physiographic provinces or subregions" characterized by differing geologies and life zone distributions. From south to north, these are: Madre de Dios, Urubamba, Tambo, upper Pachitea, western Huallaga and Chachapoyas. One problem with adopting Young's subregions for heuristic purposes is the tendency of montane forest settlements to straddle river 23 divides, thereby straddling subregion boundaries. For this thesis, broader distinctions drawn between southern, central and northern subregions are sufficient. These not only correspond to subregions utilized by Leon et al. (1992: Fig. 1} to analyze the eastern montane forest's floristic composition, but they match the subregions originally recognized in the first synthetic descriptions of montane forest archaeology (Bonavia and Ravines 1967}. The southern subregion corresponds to the former departamentos (now politically realigned in regiones} of Puno, Madre de Dios and Cuzco; the central to Ayacucho, Junin, Pasco and Huanuco; and the northern to San Martin and Amazonas. To avoid awkward terminology in the remainder of this thesis, I propose to refer to the southern, central and northern subdivisions of the eastern montane forest simply as the southeastern, central and northeastern montane forests respectively (Fig. 2}. Along the length of the Central Andes, montane forest timberline has responded to global climatic events at the end of the Pleistocene epoch, and to local, periodic phenomena such as droughts (Rodbell and Hansen n.d.}. Within the forest, the upper slopes experience periodic landslides leaving scars that are rapidly colonized by secondary vegetation, while down-cutting by turbulent waters in the canyon bottoms frequently destroys the narrow river terraces during rainy-season floods (especially between 24 October and May). By the early Holocene, human agents had begun to contribute to the ecological dynamics of the eastern slopes. Although topography, elevation, and especially frost are commonly considered the principal environmental limiting factors determining the elevation and configuration of timberline, modern-day highland villagers living adjacent to the eastern montane forest periodically burn back the forest vegetation for a variety of reasons (Young 1990, 1993; Leo 1992). In Young's opinion (personal communication), few if any regions remain in which forest timberline can be observed in equilibrium and unaffected by burning and other human activities. Vegetation at the forest's edge is burned every few years in order to eliminate woody species and improve pasturage for cattle. He notes "that evidence of burning, while ubiquitously present in the form of scorched bases of tussock grasses and dead woody plants, is not necessarily obvious to a short-term visitor to the area" (1990:28). Young also believes that the first indigenous peoples to inhabit these grasslands burned them intentionally simply because clearing the scrubby vegetation aids foot transit so markedly. According to ONERN (1976:128), life zones of the upper montane rain forest are devoid of permanent human habitation because of insurmountable topographic and climatic obstacles to traditional economic activities. However, reconnaissance 25 by explorers and archaeologists has revealed evidence that prehistoric human activities dramatically altered the montane forest landscape. Abandoned agricultural terraces, ruined settlements and paved roads within the upper forest attest to the presence of substantial pre-Hispanic populations that cleared large expanses of forest for intensive agriculture. In much of the montane forest, native ecosystems have returned to some semblance of equilibrium following Spanish conquest and abandonment by most permanent forest dwellers. From the early Spanish occupation to the present day, however, highland villagers of the eastern cordillera have maintained access to forest resources (Brush 1977; Camino 1977; Gade 1972; Murra 1972; Thompson 1980). During the mid-twentieth century there has been a renewed effort to colonize the eastern slopes by landless Peruvians. Meanwhile, pioneering agricultural and lumbering activities moving up the western tributaries of the major lowland rivers have been altering forest composition where not eliminating the forest altogether. Young observes that the Tambo Province shows the most substantial impact from highland settlement and use. The Urubamba Province likewise shows heavy human impact. Interpretive Frameworks for Andean Migrations This section will provide a brief description of the popular frameworks that have shaped scholarly perception of 26 the montane forest. The "jungle barrier" vision of the montane forest that predominated until the mid-twentieth century was outlined in the previous section. Here I will briefly address the theoretical frameworks that underlie population movement and colonization hypotheses described individually in Chapter 2. These can be characterized as "ruling frameworks," analogous to Rouse's {1986:3) theories" or Snow's {1995:59) "ruling "controlling models," except that they are fundamental sets of assumptions at higher levels of abstraction from which ruling theories or controlling models are derived. The first framework is the unitary origins paradigm from which the postulated AndeanAmazonian population movements sprang. The second is the popular model utilized to describe Andean cultural ecology and political economy termed "verticality" or "complementarity." Unitary Origins and Population Movements This dissertation focuses on the span of centuries between 2000 B.C. and A.D. 700. During this time, Central Andean cultures experienced a sporadic series of transformations leading to increasingly complex social organization and the eventual emergence of politicallycentralized states. Specific cultural transformations that have captured the interest of scholars interested in the origins of South American civilizations include the adoption of agriculture, the appearance of sedentary villages, the 27 introduction of ceramic technology, the rise of ceremonial centers and cities, and the emergence of socio-political complexity and its associated art and ideologies. Eastern Asia (e.g. Meggers et al. 1965), Mesoamerica (Spinden 1917; Uhle 1922b; Rivet 1968 [1924]; Meggers and Evans 1963), the Peruvian coast (Larco 1941; Lanning 1967), the Central Andean highlands (Tello 1929, 1943, 1960) and the Amazon Basin (Tello 1942; Lathrap 1970, 1977) have alternately been touted as nuclear areas or "hearths" during the decades prior to 1970 when evolutionary theories invoking migration and diffusion were common. More recently, Lathrap (1977) has resurrected Spinden's (1917) hypothesis to advocate a "unitary model" for agriculture's invention and subsequent spread from Amazonia. Agriculture was invariably believed to be the foundation of complex cultural development, and population movements, along with diffusion, were considered principal modes of spatial transmission for such innovations. A basic but erroneous assumption underlying most migration theories was the belief that race, language and culture are not independently variable, but move across the landscape in fixed association. Hence, Uhle, Rivet, Meggers, Lathrap and others have proposed series or "waves" of long-distance population movement as pivotal events in archaeological sequences, and especially in the "founding" of civilizations. 28 Archaeological data gathered over recent years have led to the realization that Central Andean civilization was not a package of culture "traits" appearing and diffusing in synchrony (Lumbreras 1981; Moseley 1983). Evolutionary milestones such as the adoption of agriculture, pottery and sedentary lifeways often have independent histories, and cultural development proceeded in piecemeal rather than revolutionary fashion. That the accretion of the elements which together constitute "civilization" was a slow process in the Andes is illustrated by the span of nearly three thousand years separating the first construction of public monuments and the emergence of the first Andean regional states. Nowadays, few scholars believe that a single hearth for Andean civilization can be indicated. For example, many of the earliest pottery technologies found on the South American continent apparently developed independently (Raymond et al. 1994; Hoopes 1994). The earliest Andean socio-economic hierarchies probably arose at several localities participating in long-distance interaction sparked by the Chavin cult's spread from the north-central highlands (Burger 1992, 1993). The first expansionist states arose independently on the coast (Fogel 1993) and in the highlands (Isbell and Schreiber 1978). Migration-origin theories have thus been discredited first on empirical grounds, and thereafter for paradigmatic reasons. Their 29 legacy however, has been the identification of suggestive material patterning that still warrants archaeological analysis and interpretation. Andean Political Economy, Verticality and Colonization Since the mid-1960s, Murra's (1964, 1967, 1975) ethnohistorically-derived model of highland Andean political economy has dominated ethnographic and archaeological studies of Central Andean interaction and exchange. Murra observes that Andean societies developed singular techniques of procuring products from a variety of distinct, vertically-stacked ecozones in the highlands through direct colonization, rather than relying upon exchange systems that potentially compromised a community's economic selfsufficiency. The Chupachu and Lupaca evidence for colonization extracted from sixteenth century census documents provides Murra with two examples of Andean "ethnic groups" with differing degrees of socio-political complexity sharing this same economic strategy. The former occupied the high elevations above the upper Huallaga River valley, while the latter inhabited the shores of Lake Titicaca on the southern Peruvian altiplano. Both groups reportedly maintained distant colonies in warm lowland ecological zones. Squarely within the substantivist tradition of economic theory, Murra's model of "vertical control" or "zonal 30 complementarity" emphasizes non-market systems based upon traditional modes of reciprocal exchange at the community or "folk" level, and redistribution at complex politicaladministrative or "state" levels. More specifically, Murra hypothesizes that the desire of highland polities for access to maize, coca and other goods available only in the lowlands spurred the "conquest and colonization" of appropriate ecological zones (Murra 1975:59). To Murra, the imperative of establishing vertical control through colonization was a "pan-Andean" characteristic (Murra 1967), and that the well-known Inca practice of relocating populations as mitimae colonies represented "a late and altered manifestation of an ancient Andean pattern" (1975:60). Maintenence of community self-sufficiency through intra-ethnic exchange is considered a uniquely Andean alternative to the interregional, inter-community trade economies that characterize other prehistoric world areas like Mesoamerica and the Near East (Morris 1978:318; Murra 1981:52; Van Buren 1996:341). The degree to which occupational specialization and mercantile activities took place in the Central Andes (Rostworowski 1977) remains unresolved, and has received relatively little attention from archaeologists. Murra's conclusions that verticality was an ancient pattern common to Andean societies at all levels of sociopolitical organization from Colombia to Argentina has had 31 wide-reaching implications for studies of Andean cultural development and exchange. Archaeologists have applied Murra's model to interpretations of prehistoric Andean cultural evolution (e.g. Moseley 1983, 1992), especially in the south-Central Andes (e.g. Mujica 1985; Rice et al. 1989; Kolata 1983, 1991; Aldenderfer and Stanish 1993; Goldstein 1993). Moseley (1992:43) views verticality as an adaptive strategy supplying highlanders throughout the Central Andes with needed "marine salt, seaweed, fish, fruit, beans, maize, coca and cotton" from the lowlands. There is a consensus (e.g. Moseley 1983, 1992; Burger 1985a; Stanish 1992) that access to produce from lower elevations was and still is a biological imperative for Andean populations on the puna. Moseley (1992:46) hypothesizes that verticality strategies emerged during Preceramic times at the household and community level as a consequence of highland population growth coupled with an expanding diversity of domesticated lowland crops. Moseley theorizes that the domestication of the llama and alpaca as cargo beasts, and the application of irrigation technology within the intermontane basins signaled the appearance of "true verticality adaptations" by the beginning of the Initial Period (Ibid. :100, 142). However, Miller and Burger (1995) present evidence for the relatively late utilization of the llama for cargo transport, during the Early Horizon in most areas. In 32 Moseley's view (1983:194-196, 1992:46), the highlanders' continual drive to directly control lowland agricultural lands generated "a marked downward thrust" manifest in repeated prehistoric colonization and periodic military incursions into the lowlands. Thus, highland occupations frequently extended down to 1000 m on the Andean slopes (1992:45). He observes (Ibid.) that the verticality strategy is "least pronounced" in the north, and more intense and elaborate in the south-Central Andes. Whereas Moseley views verticality/complementarity as an adaptive strategy, Stanish (1992) regards it as a politicaleconomic strategy. In a recent study of south-Central Andean cultural development, Stanish (Ibid.:SO) argues that, "the most successful criteria for addressing the prehistory of the region are political and economic ones as understood within the general framework of zonal complementarity." Stanish is among many scholars (see Masuda et al. 1985; Lynch 1981) that have implicitly or explicitly advocated widening the concept of complementarity beyond "directcontrol" resource procurement through colonization, to include "indirect control" through trade between autonomous polities. In effect, all investigations of Central Andean exchange have become macro-scale analyses of highland political-economic evolution, emphasizing core-periphery relationships. Archaeologists have already registered complaints that 33 the concepts of verticality/complementarity had been stretched to include virtually all Andean interactions (e.g. J. Topic and T. Topic 1985:56; Patterson 1988:217). Burger (1985a:275-277) has contested Moseley's position and indicated the need for an "interregional model" for exchange. The Topics (J. Topic and T. Topic 1985:58) rejected the verticality model as a conceptual aid to interpret long-distance interaction between coastal and highland societies in their north-Central Andean study area. They observe that implementation of the verticality model obscures "the range of goods and services considered necessary for self-sufficiency, and the means used to obtain them ... at different times and places within the Andean area." Stanish acknowledges points of controversy surrounding the complementarity model (1992:6), but maintains that gathering "virtually all" Andean exchange under the umbrella of complementarity provides an internally consistent framework for explaining Andean cultural evolution, at least in the south-Central Andes. In the most recent theoretical examination of verticality, Van Buren (1996) questions some of the most basic assumptions underlying arguments by Murra, Moseley and many other Andeanists. Van Buren examines the south-Central Andean case of the Lupaca kingdom showcased by Murra as an example of verticality colonization, and observes that intensive archaeological investigations on the western 34 slopes have failed to produce solid evidence for a preHispanic colony-type occupation. Drawing on post-Conquest historical documents, she is able to show that the primary economic activities at the Colonial Period (and possibly Inca) colony of Terata Alta most clearly served the needs of a small group of highland political elite rather than the greater community. Thus Van Buren demonstrates that verticality colonization cannot be interpreted as the archetypal Andean ecological adaptation a priori. Further, she has sown seeds of doubt that verticality colonization even existed as a pre-Hispanic institution, political or otherwise. Archaeologists examining interaction in the northCentral Andean highlands have made limited use of verticality and complementarity conceptual frameworks, invoking these only for local-level exchange (e.g. Shady and Rosas 1979; J. Topic and T. Topic 1985). Network analysis (J. Topic and T. Topic 1983), "peer polity interaction" (Burger 1984a, 1993), and envoy or "expedition" exchange (Lumbreras 1993:364) are among alternative methods and models yet to be adequately explored for interpretation of regional and interregional levels of interaction. The Problem of Tropical Montane Forest Prehistory: Migration or Interaction? The research problem confronted in this thesis entails determining whether archaeological material patterning in 35 the eastern montane forest constitutes evidence of prehistoric migrations, or indicates local cultural development and interaction. According to Rouse, who has detailed the most extensive methodology for "inferring" migrations from archaeological data (Rouse 1958, 1986}, these are the two basic alternatives which must be weighed one against the other if a prehistoric migration involving population replacement is suspected. The challenge is to distinguish the movement of culture from the movement of people. Both kinds of movement were constant features of human prehistory world-wide, yet only the movement of culture through "diffusion" and "trade" has generated a large body of theoretical literature promoting diverse frameworks to guide interpretation (Adams et al. 1978; Schortman and Urban 1987}. Adams et al. (1978} have traced the intellectual history of migration explanations in anthropology, observing that "migration theory" is neither a paradigm, nor a formal interpretive framework. As they wrote, "migrationism" was said to be in full retreat as a viable explanation for change in the sub-fields of Archaeology and Physical Anthropology, while less so in Linguistic Anthropology. During the mid-twentieth century, many migration theories have failed to withstand critiques backed by enlarged data bases and improved analytical techniques. Other migration theories have been "obviated" by American Archaeology's 36 shifts toward nee-evolutionist and nee-Marxist agendas intent on toppling culture-historic interpretations deemed unscientific (Ibid.:1978:504; Trigger 1989:421). al. Adams et (1978:487) observe that most migration theories had never been much more than "ad hoc and somewhat mechanical explanations" for prehistoric change, especially for developmental discontinuities and anomalous trait and site distributions. On the other hand, they argue that the theoretical bias against "migrationism" is deleterious to the discipline. Two compilations of culture-historical studies bearing directly on the problem of discriminating prehistoric migration and diffusion were presented in the 1950s by Thompson (ed. 1958) and Lathrap (ed. 1956). Lathrap and his colleagues coined the term "site-unit intrusion" (usually evidence of migration) to contrast with "trait-unit intrusion" (usually evidence of diffusion) . However, Rouse (1958 [in Thompson ed.]) presented the first explicit set of guidelines for "inferring" prehistoric migrations. The fifth and final of his five analytical steps contended that "it is incumbent upon the person who wishes to demonstrate migration to consider and eliminate the possibility that some other hypothesis may better fit the facts at his disposal" (Rouse 1958:66). This statement called for rigorous scrutiny of the evidence and clearly placed the burden of proof upon advocates of migration explanations. 37 Almost 30 years later Rouse himself assumed "the burden" with the publication of Migrations in Prehistory, an explicit and detailed methodology for inferring prehistoric migrations with four case studies (Rouse 1986). In addition to reiterating and developing points argued in 1958, Rouse emphasizes differentiation of kinds of migrations, kinds of cultural development and kinds of interaction. He maintains that much of the disillusionment with migration explanations has stemmed from the failure to discriminate between kinds of migrations, and failure to systematically test hypotheses (1986:17-18). Thus Rouse delineates specific systematic techniques for controlling chronological and spatial aspects of the problem in order to establish temporal priority at hypothesized sources, and to identify stages and routes of movement. Also, issues of mobility including modes of transportation, environmental transitions and potential barriers must be considered. He maintains that evidence from historical linguistics and physical anthropology that archaeologists typically muster to corroborate their hypotheses must be utilized only as independent checks. Rouse's work sparked a flurry of migration literature authored by archaeologists representing American processual and European historical research traditions (as these are defined by Binford and Sabloff 1982). Anthony's (1990) processual approach borrowed concepts from contemporary disciplines of Geography and Demography to demonstrate how a 38 nomothetic, behavioral approach supports a postulated westward migration of Yamna Kergan pastoralists from the Eurasian steppes during the fourth millennium B.C. European archaeologists Otte and Keeley (1990) simultaneously called for the revival of migration and diffusion concepts to interpret upper Paleolithic Period culture change in the Old World, arguing that both are valid, historically documented processes. Ironically, Anthony's study was denounced as ahistorical by European scholars (Chapman and Dolukhanov 1992), while Otte and Keeley's position was criticized by processualists (Clark and Lindly 1992) who categorically reject analogies between historical processes and Paleolithic site formation processes. Meanwhile, Marxist critics (e.g. Patterson 1991; Sued-Badillo 1992) have condemned Rouse's methodology, at least as it has been employed in the Caribbean, as the reductionist trivialization of local historical processes typical of a "hegemonic framework" (Patterson 1991:4). Clearly, "migrationism" has not fit comfortably within any single contemporary archaeological paradigm. Rouse (1986:158) observes that the relative success ascribed to his four case-studies was owed in part to their prehistoric recency and the feasibility of tracing material patterning back through time and space using the directhistorical approach. Thus, Snow (1995) employs Rouse's 39 formula to reassert previously discredited theories of Iroquoian migration. He denounces the anti-migration bias of the "controlling in situ model," and contends that major anomalies in reconstructions of prehistoric Iroquoian language, kinship, ceramic manufacture and site distribution support a migration explanation. The oft-cited Four Corners-Rio Grande pueblo migration is currently being reexamined utilizing insights gained from the debates prompted by Rouse's work (Cameron 1995). Those concepts deemed useful for this particular study will be described in the following section. Concepts and Strategies for Evaluating Migration Hypotheses Selected methodological points and procedures developed by Rouse and others provide a theoretical basis for evaluating postulated montane forest population movements and colonization. Unfortunately, partisan scholarship has produced parallel sets of operative concepts and definitions. Consequently, the choice of concepts and terms has tended to reflect each scholar's inclination toward either culture-history or processual paradigms. For example, Snow (1995) relies on Rouse's methodology, while Cameron (1995) chooses to modify Anthony's criteria. The following theoretical discussion will rely principally on Rouse's original formulations, but will also draw from the most useful of Anthony's and Cameron's additions and 40 modifications. Operational definitions of basic concepts such as "population movement" and "colonization" are essential because, from an archaeological perspective, they entail investigation of distinct sets of material correlates. Rouse's term "population movement" is analogous to Anthony's (1990:902) "long-distance migrations" that characteristically cross ecological or cultural boundaries, and Cameron's (1995:113) "long-range, short-term movements of large segments of a population." Rouse (1986) observes that such population movements may transpire as the "peopling" of unoccupied areas, or as the replacement of extant populations by rapid processes of expulsion or absorption. Population movements resulting in replacement must be distinguished from imperial conquests with agendas of territorial expansion. Archaeologists ostensibly have the methodological tools to detect short-distance and long-distance migrations at the levels both of communities and societies (Adams et al. 1978:489). Rouse and proponents of Andean-Amazonian population movements are clearly concerned with the "society level." Anthony (Ibid.) remarks that "long-distance migration should result in changes that would have distinct effects on the archaeological record," although Adams et al. point out that these population movements are the least common "in the course of recorded history." Well-known 41 archaeological examples of postulated society-level migrations resulting in replacement include the "wave-ofadvance" spread of neolithic farmers across Europe (Ammerman and Cavalli-Sforza 1973) and the westward Indo-European migration of horse-mounted societies from eastern Europe (Anthony 1986, Renfrew 1987). Rouse notes that population movements should not be confused with "immigration" which is localized migration of individuals or social groups that results in assimilation rather than population replacement. Snow (1995:74-75) postulates a series of Iroquois community-level migrations that ultimately amount to a society-level population movement, while Cordell (1984, 1995) postulates series of family and community moves that likewise constitute a society-level population movement. Neither of these cases provide good analogies for this particular investigation as they both might be considered intra-regional from an ecological point of view. The Four Corners-Rio Grande migration as described by Cameron and Cordell appears to be a case of massive immigration and assimilation within a broad pre-existing interaction sphere. Because the term "colonization" is utilized in Andean studies to refer to at least three separate processes, a working definition for this study must be carefully specified. Andean colonies in Murra's verticality model appear on the landscape as "archipelagos" or enclaves 42 implanted far from the colonizing community's homeland. The colony is an intrusive settlement, perhaps accompanied by similar colonies from other homelands, surrounded by an indigenous population at its new locality. Confusion stems from additional academic and colloquial usages of "colonization" to refer to either: 1) the peopling of an unoccupied frontier, or 2) the frontier installation of administrative armature by a conquest state in order to exploit local natural and human resources. In this study, the terms "colony" and "colonization" refer strictly to Murra's archipelago variety in which colonists expend their own labor to exploit local resources. The other two so- called colonizing processes are referred to in this study simply as "peopling" and "conquest." Comparison of material patterning from suspected homelands, along routes of movement and at final destinations, is a procedure fundamental to the evaluation of population movement and colonization hypotheses. Some general guidelines serve equally for the archaeological identification of both processes. For example, there is a consensus that a culture-historical discontinuity or anomaly in both time and space is common to all hypothetical migrations. To evaluate similarities across temporal and spatial dimensions, Rouse (1986:160) and Stanish (1992) emphasize the importance of tracing contexts rather than isolated attributes or "traits." They agree that sites and 43 site "structures" with behavioral correlates must be the primary units of analysis. Such comparative analyses of site structures focus on activities, activity areas and specific behaviors associated with artifacts. Despite general resemblances in overall strategies, specific approaches for recognizing population movements and colonization must differ in details. Basically, this is because of the differing intentions of migrants in each process. The protagonists in most hypothesized prehistoric population movements are societies with organizational and/or technological advantages that encourage encroachment and predation (cf. Sahlins 1961; Davis 1973; Rouse 1986; Snow 1995). Archaeologists working within cultural- ecological frameworks often contend that migrants are "pushed" out of their core areas by population pressure. Cameron notes that southwestern archaeologists are increasingly considering "pull" factors such as economic and ideological incentives. One important variable that Cameron formally introduces into the archaeological migration literature is the "unit of migration." She (1995:112-113) argues that, because decisions to migrate were probably made by kin groups or social units larger than individuals or nuclear families, Anthony's application of modern demographic concepts may be inappropriate. In the pre-Conquest New World, extended kin groups served as the basic organizational units for most 44 social, economic, political, and religious institutions. The implications of such large units of migration are that: 1) population movements were relatively rare occurrences; 2) only under certain conditions of stress ("push") and social disintegration did individuals and nuclear families migrate and 3) under benign or favorable ("pull") conditions, large units of migration may be expected to have left evidence more closely resembling a site-unit intrusion. As Cameron suggests, it behooves the analyst to investigate specific social, political and economic circumstances promoting or preventing migration. Most postulated population movements cross relatively homogeneous landscapes with few natural barriers (Adams et al. 1978:501). Rouse argues that archaeological complexes cannot be simplistically followed across the landscape because migrants carry only a subset which alters concomitantly with migrants' comparing site assemblages, "adaptive radiation." In the analyst examines internal contextual relationships by noting the co-occurrence of artifacts and features. Most important is identification of the behavioral correlates of sites, complexes and assemblages relating to environmental habits, preferred site locations, subsistence strategies, settlement layouts, household organization and food selection (Rouse 1986:166174) . For Rouse, the selection of appropriate units for 45 analysis and classification is crucial to the success or failure of correct inference. Combining Willey's (1945) concepts of archaeological "horizons" and "traditions," Rouse and Cruxent (1963:23) developed the concept of the "series" to describe development of a style or complex across space as well as through time. The series concept has clear advantages for inferring population movements as a means of diachronically interpreting spatial dimensions of stylistic developments. Series and subseries labels carry the suffixes "-oid" and "-an", and Rouse (1987:14) likens these to language families and subfamilies. Because series and subseries under conditions of migration represent population segments carrying only a portion of their culture, he (1987:10) compares the establishment of a new line of development to the "founder's principle" concept of population biology. Rouse contends that archaeological classifications should be modeled after the linguistic classification schemata which have most successfully revealed historical relationships. Unlike migrating populations, Andean verticality colonies are deliberately implanted into ecological zones differing from the homeland. In theory, such colonies should be clearly distinguishable as site-unit intrusions. Morris (1978) was among the first scholars to consider the kinds of archaeological patterning expected under conditions of verticality colonization. He (1978:318) theorizes that, 46 "if [Murra's] model is correct, we should have a situation whereby interregional exchange of subsistence goods is handled by intra-community mechanisms." Morris suggests that such intra-community exchange should produce tightly circumscribed spatial distributions of pottery style attributes. Pottery styles in colonized regions should be interdigitated with others like "the islands of an archipelago" (Ibid.). In theory, each style represents one node within a single intra-community exchange system spanning multiple ecological zones. Despite the popularity of Murra's model for theorizing and interpreting archaeological data, little field research has been directed specifically towards the identification of Andean verticality colonies until recently. Archaeological work on the western slopes of the south-Central Andes by Aldenderfer and Stanish (1993; Stanish 1992) has attempted to identify such colonies through synchronic evaluation of similarity between contexts found in differing ecological zones. Domestic contexts, and especially households, they argue, are most appropriate units of analysis for identifying the special intra-"ethnic" relationship between colony and homeland. Kitchen middens, non-elite residences and storage structures are among domestic contexts said to directly reflect a resident population's ethnic affiliations. Non-domestic contexts such as tombs, elite residences and corporate architecture are considered more 47 likely to contain exotic elements reflecting external relations such as exchange and alliance systems (Stanish 1992:9). Stanish lays out a specific set of archaeological correlates for identifying Andean colonies. First, a colony may show non-local artifacts in both domestic and nondomestic contexts. A lack of any exotics, however, might suggest no complementary relationship. The domestic component of a colony should be identical to that in the homeland, showing only "ecological impact." Actually the theoretical literature includes little discussion of the potential effects of a colony's new environmental and cultural surroundings on households and domestic economies. According to Murra and Stanish, colonies are mono- or multiethnic enclaves that may be distributed among other mono- or multi-ethnic colonies in a particular valley. Thus, verticality colonization introduces heterogeneity at the level of sites, localities and regions (Ibid.:44-45; Morris 1978:318). Some of Stanish's criteria consider colonization in differing regional socio-political contexts. He notes that the degree to which social organization within colonies mirrors that of the homeland remains unknown. However, he hypothesizes that colonies established by centralized Andean states would house low-status individuals in a servile capacity. According to Bonavia (1978:400-401), colonies 48 established under the auspices of Inca or Huari imperial interests should show little or no evidence for occupational differentiation, social structure or political structure. "Absolute" intra-site functional homogeneity, he argues, characterizes settlements that he observed in the central and northeastern montane forests. Bonavia theorizes that these colonies were occupied by communities of full-time agriculturalists fully dependent upon their parent highland polities (Ibid.). Utilizing Stanish's and Bonavia's assumptions, we can hypothesize that montane forest colonies established under conditions of imperial hegemony would lack architectural features reflecting social heterogeneity and political autonomy. We might even expect a dearth of architectural evidence for the integration of family and other social units, and for socio-political hierarchies. Potential analogies might include the Inca settlements Callachaca A (Niles 1987: Fig. 2.1, 2.3) and Raqay-Raqayniyoq (Ibid.: Fig. 2.8) that apparently supplied housing for small family units (perhaps rendering their share of m'ita labor tribute to the Inca state). The lack of common courtyards, or even facing doorways, suggests that intra-community social interaction that potentially interfered with work or fostered organized resistance was discouraged. The Inca cases just described provide exceptionally clear examples of hypothetical colonies where colonists 49 divested of their ethnic identities (at least temporarily) provided labor tribute to the Inca state. As previously mentioned, Murra views these as late, altered variants of Andean verticality colonies. Ostensibly they conform to the verticality model because non-local populations were installed for production directly benefitting the homeland (in this case, Cuzco). Implicit in Murra's model is the assumption that, regardless of the colonizing polity's degree of socio-political integration, the implanted population is directly engaged in production rather than in the administration of local populations. Evidence that local rather than colonizing groups were the producers suggests that analytical frameworks of conquest and imperial administration are more appropriate than verticality colonization models. Stanish hypothesizes that colonies established by Andean polities at lower levels of social integration would house groups of status equal to their homeland counterparts (Ibid.). Stanish interprets the post-Tiwanaku highland colonization of the western-slope Otora Valley as multiethnic in composition. According to his reconstruction, initial colonization was followed by the emergence of a local elite and subsequent independent cultural development. In this case, colonization has, in effect, become a population movement. Stanish's data has potential for revealing much about general processes of migration and 50 consequential environmental impacts and social changes, and their archaeological interpretation. In cases of both hypothetical population movements and colonization, it is usually the degree of perceived similarity between assemblages at sources and destinations that supports or contradicts the migration argument. Of course the lack of any similarities render questions of migration moot. The evaluation of postulated population movements requires the diachronic analysis of change, while the evaluation of postulated colonization involves detailed synchronic comparisons. Because interpretations of population movements must take temporal as well as spatial variables into account, they ultimately require a greater number of theoretical assumptions. Hence, population movements, as they are presently understood by archaeologists, should be far more difficult to demonstrate than Andean colonization. Perusing the theoretical literature devoted to identification of prehistoric migration, readers are led to believe that, provided adequate knowledge of spatial and temporal variables, as well as a variety of behavioral contexts, archaeological evidence of population movements and colonization should be easy to interpret. Anthony (1990:897) insinuates that his methodology, unlike Rouse's, yields "conclusive" results. However, it would seem that only certain kinds of migrations can be inferred with any 51 confidence. Although Stanish has published only selected sets of data, it would appear that his implementation of context-specific strategies for comparative analysis has succeeded in identifying outlying western-slope settlements composed of highland social groups. His methodology should also be useful in identifying the early stages of population movements, especially given their purported propensity for bypassing nearby territories in favor of more distant areas, or "leap-frogging" (Anthony 1990:902-903). Contexts for Inquiry: Manachaqui Cave and the Greater Eastern Montane Forest In this thesis I attempt to evaluate the evidence for population movements and colonization from increasingly broad archaeological contexts beginning with Manachaqui Cave and extending to the greater eastern montane forest. Different levels of inquiry require different strategies appropriate to the kinds of data available. Furthermore, the utility of certain kinds of archaeological data may be regarded as questionable considering the emphasis on contextual approaches advocated in the literature. Immediately suspect is my primary source of data, Manachaqui Cave. Manachaqui Cave is, of course, a rockshelter and not a homeland or hypothetical "colony." However, the rockshelter lies within Chachapoyas, an archaeological culture area known from documentary and archaeological evidence to have 52 occupied montane forest environments. Manachaqui Cave is also strategically placed between two postulated migratory "highways," the Marafion and Huallaga River valleys (see Chapter 2); and beside an east-west pre-Hispanic road connecting highland and montane forest destinations (see Chapter 3). The archaeological remains at Manachaqui include deposits from time periods corresponding to the population movements and colonization postulated by archaeologists investigating montane forest prehistory. Regardless of the aforementioned features favoring the utility of Manachaqui Cave, issues surrounding the site's functions and assemblage comparability must be addressed. Stanish (1992) dev0tes extensive discussion to the potential for misinterpretation inherent in "artifact-based approaches" where contexts are not controlled. Archaeological assemblages recovered in the highland, montane forest and lowland regions surrounding Manachaqui Cave represent a variety of contexts, usually cemeteries, caves and rockshelters, early ceremonial structures and late prehistoric administrative complexes. Unfortunately, relationships between site function and assemblage composition are seldom addressed in useful detail. Nonetheless, potential sampling biases should be recognized to the degree possible and taken into account. In the Andes and elsewhere, caves and rockshelters are settings for limited sets of human activities. As Straus 53 (1990:279) points out, "the natural features in caves and rockshelters structure activities that were conducted at these types of places." Manachaqui Cave is a small an~ rockshelter that became smaller through time as natural cultural deposits accumulated under foot. The range of activities suited to such a space was always limited. From a broader perspective, Manachaqui Cave represents one "place" in a larger system of places, each corresponding to particular sets of activities or behaviors (Ibid.; Binford 1982). Likely functions suggested by Manachaqui Cave's small size, and by analogies found in the Andean ethnographic, ethnohistoric and archaeological literature, include: 1) a permanent habitation for a small family, 2) a temporary hunting camp, 3) a herders' camp, 4) a huaca or sacred site, 5) a wayside station, and combinations of any of these. A "domestic" assemblage representing the entire suite of household activities in a permanent cave habitation might be the most desirable of data sets for interregional comparisons. "base camps" However, only hypothesized Preceramic Period (e.g. Rick 1988) and one Initial Period herding camp {Lavallee et al. 1985) have been interpreted as permanent or semi-permanent habitations. Unfortunately, ceramic components from hunting and herding camps have been reported only superficially, but these assemblages might be characterized as specialized subsets of a larger domestic 54 repertoire. As repositories of offerings valued as exotic, sacred cave sites would presumably present the greatest danger for misleading interpretations. If Manachaqui Cave frequently served as a wayside station as its roadside location suggests, then an important question is whether it served migrating populations or groups engaged in interregional exchange. In either case, a wayside station would contain assemblages of a domestic, or "secular" utilitarian nature, although concerns with transportation and portability should impose constraints on the structure of pottery and other artifact assemblages. DeBoer {n.d.) reports that vessels utilized by modern Central Ucayali Shipibo travelers are simply smaller versions of the large vessels utilized within domiciles. Under conditions of population movement and/or colonization, there is no reason why archaeological assemblages from a wayside station should not reflect the stylistic norms of the migrant or colonizing populations. At Manachaqui Cave, a prehistoric population intrusion of any kind should leave a disconformity in the cultural sequence representing the sudden and complete imposition of an intrusive assemblage. Under conditions of population movement and replacement, Manachaqui Cave would be positioned to service the migratory counterstream {Anthony 1990), and the post-migration cultural sequence should reflect only the development of the new population. Ceramic 55 assemblages, while they present some interpretive problems (cf. Adams 1979), do yield useful information for evaluating migration hypotheses. Snow {1995) points to anomalous shifts in ceramic technology in support of a migration explanation, while Raymond et al. {1994) and Hoopes {1994) argue that salient differences in earliest New World pottery assemblages discredit migration theories. might d~scern Simultaneously we evidence for shifts in diet and/or subsistence technologies. Under conditions of colonization as Murra describes it, we might find a temporary co-occurrence of local and intrusive artifacts and artifact styles within single phase assemblages. Likewise, evidence for subsistence could appear as the co-occurrence of local and intrusive sets of food remains in single faunal and ethnobotanical phase assemblages. The interpretive challenge would be to distinguish such co-occurrences from assemblages of food remains left by single populations with generalized subsistence patterns. In order to approach such interpretive problems within a single archaeological site, high levels of stratigraphic and temporal resolution are crucial. The interpretation of co-occupation from food remains would rely on the identification of a temporary shift from a relatively focused subsistence pattern to an ostensibly generalized pattern and back again. Considering the harsh realities of site formation and assemblage 56 preservation in most sites, these interpretive avenues present considerable difficulties. In broader montane forest valley or multi-valley contexts, predictive statements by Morris, Stanish and Raymond concerning style distributions and settlement patterns are potentially useful for detecting prehistoric population intrusions. The reader will recall from previous discussion of criteria that verticality colonies in regional contexts should be identifiable by circumscribed, archipelago-like spatial distributions of distinctive domestic assemblages, artifact styles and style attributes (Morris 1978; Stanish 1992:43). Conversely, regional homogeneity should indicate local autonomy, and "a discrete 'culture area' definable by a relatively consistent set of archaeological indicators" (Stanish 1992:45). Substantial stylistic diversity exhibited within site ceramic assemblages (i.e. abundant "foreign" pottery) might constitute additional evidence for local autonomy and interregional exchange rather than verticality colonization. Influenced by Brush's (1974:280) use of Murra's (1972) framework, Raymond (1988:290) stresses the importance of local geography in determining other characteristics of montane forest colonization: "the permanency and size of the settlements in the montana are a function of the distance, relative economic importance of the resources, and the extent to which it is necessary to protect the lands from 57 neighboring ethnic groups - of both highland and lowland origin." He offers the following additional predictions based primarily on the distance separating highland and "montaiia" (montane or premontane forest) resource zones: If the resource zones are near one another, i.e., within one or two day's walk, then all the resources may be exploited from a single highland village, and settlements in the montana may be either non-existent or small and scattered. Resources which are widely spaced and require several days' walk lead to a pattern of settlement which Murra (1972) has called the Archipelago type. The principal villages are in the highlands but satellite communities are established in the montafia ... If the resource zones are large and continuous but strung out over a long valley system, then an extended, multi-ethnic settlement pattern may develop with products moved via a trade and market system ... Under such circumstances permanent settlements of highlanders would be established in the montana (Ibid. :290-291). Raymond's formula is based upon Murra's ethnohistoric descriptions of montane forest colonization by the Chupachu of Huanuco, and the Lupaca of Puno. Murra (1975:70-71) reports that Chupachu colonies consisted of three or four homes and were located within four day's walk from the highland nucleus. The highland Lupaca "kingdom" reportedly maintained colonies with over a hundred homes, ten or more days' walk distant (Ibid.:79-80). Raymond's third prediction for long, ecologically diverse valleys is based upon Brush's and Gade's (1975) interpretations of land-use in the Vilcanota (or Urubamba) Valley of eastern Cuzco Department. Such geographical and ecological formula, backed by ethnohistorical evidence serve Raymond to interpret the archaeology of the montane forest. Mayer (1985:66) observes that settlement locations are also partly determined by the 58 need to be close to the production zone requiring "the most intense care." In addition to interpretive utility, Raymond's predictive model provides a baseline from which to measure deviation from expected patterning both in the northern subregion, and in other eastern montane forest subregions. The recognition of such deviations should prompt a search for historical and other non-ecological variables. The Alternative Hypothesis: Local Development and Interaction In the remainder of this thesis, I hope to demonstrate that the archaeological data from both Manachaqui Cave and the Central Andean montane forests support interpretations of autochthonous cultural development by populations habitually engaged in long-distance exchange between neighboring Andean and Amazonian regions. In order to proceed with this study, some factors expected to complicate evaluation of the hypothesized migrations should be considered. The following discussion is most pertinent to the postulated Andean-Amazonian population movements, but implications for the evaluation of colonization hypotheses are likewise evident. Clearly, population movements with the highest probabilities of archaeological detection are those which "people" sparsely inhabited or uninhabited regions like Polynesia and east-Arctic North America (Rouse 1986). As 59 bounded gene pools with tightly delimited spheres of interaction, island settings seem ideal for tracing population movements, especially where migrants carry a distinctive and complex technology as Rouse's Caribbean migrants did. However, changes in artifact attributes and behavioral contexts that occur during slowly unfolding population movements (e.g. multi-directional population expansions, "waves-of-advance" or prolonged immigration) into varied social and physical environments may leave material patterning sufficiently complex to confound archaeological identification of the process. Of course the inability to document migration archaeologically does not mean that migration and population replacement have not occurred. Linguistic and physical anthropological evidence might still provide evidence for population replacement. Physical anthropological evidence for population replacement is frequently unavailable. Portentous for this study, is the fact that even the Caribbean Island evidence that Rouse showcases may leave ample room for contention (e.g. Drewett 1993:456). One even begins to suspect that prehistoric society-level population intrusions into occupied social and physical environments, at least as archaeologists have envisioned them, are ideal theoretical constructs with few counterparts in archaeological reality. Unlike the postulated migrations to be evaluated in this thesis, Rouse's four "successful" cases 60 had historical components and were therefore amenable to application of the direct-historic approach. Rouse warns that interpretation becomes increasingly tenuous moving backward through time, and Snow (1995:60} flatly states that "attempts to correlate archaeological and linguistic lines of evidence for periods prior to recent millennia are specious." It is not surprising that Anthony's and Otte and Keeley's postulated Copper Age and Paleolithic Period migrations received harsh criticisms. The postulated Andean-Amazonian population movements purportedly surmounted an imposing natural barrier. sheer magnitude, For the environmental contrast between the Andean and Amazonian regions has few if any peers. Perhaps unrivaled is the quantity and compaction of ecological regimes entirely replaced across short horizontal distances. Investigations of cultural geography by Steward (1948; Steward and Faron 1959} and other anthropologists assure us that the Central Andean-Amazonian environmental transition has coincided, and continues to coincide, with a major cultural boundary of long standing. The degree to which Central Andean and Amazonian populations were culturally, linguistically and biologically discrete at different moments in antiquity remains to be adequately addressed with archaeological evidence. However, this shifting cultural interface (or series of interfaces} has probably ranged within the approximate geographic limits of the montane and 61 premontane forests for at least two and a half millennia (since the Chavin horizon according to Lathrap 1971:97). The antiquity and persistence of Andean-Amazonian boundaries undoubtedly structured and conditioned every class of border interaction. Steward envisioned "migratory waves" breaking against the base of the Central Andes, creating what Lathrap calls "linguistic balkanization" at the western edge of Amazonia. A complex history of language divergences and convergences in this cul-de-sac is likely matched by complex archaeological and biological distributions. Cultural interaction has been sustained along the Andean-Amazonian interface for millennia as the archaeological data unequivocally demonstrates (Lathrap 1973a; Raymond 1988). This continuous interaction inevitably included processes of immigration, acculturation and replacement. The well-founded assumption that opportunities for the exchange of material goods, information and genes were constant, and always changing, strongly implies that postulated migrations would have occurred against a backdrop of intense and incessant social interaction and trade. Preliminary archaeological research (e.g. Allen 1968; Lathrap 1970; Raymond et al. 1975; Hastings 1985) indicates that the Andean-Amazon interface is characterized by complex distributions of styles and stylistic attributes. al. Adams et (1978:503), observe that migration theories are most 62 controversial "in areas where there is a diversity and complexity of cultural styles." Apparently, style is a frequently cited, but seldom convincing tool for migration advocates. This is because stylistic resemblances also serve archaeologists to identify networks or spheres of interacting societies. The frequent failure of migration theories to withstand close scrutiny strongly suggests that stylistic evidence for intense cultural interaction and exchange is prone to frequent misinterpretation as evidence of society-level population movements. Such interaction characteristically creates "noise" in what Rouse and Anthony imply should be a relatively straightforward procedure. Archaeologists have observed that this noise or "ambiguity" is typical of social boundaries where styles become fluid, often merging to render "hybrids," as they are manipulated to serve different ends (Ericson and Meighan 1984; Lightfoot and Martinez 1995:482). Lightfoot and Martinez advocate analyses of material culture, stylistic behaviors and information exchange at boundaries to elucidate cultural processes of potential significance both locally and in the adjoining regions. It would appear that there is seldom sufficient evidence to convincingly infer or entirely discount hypothetical prehistoric migrations in contexts of intense and sustained interaction. Indeed, exchange of goods and information along the Central Andean, Northern Andean and 63 Amazonian boundaries must have been accompanied by some population interbreeding and genetic exchange. Under these circumstances it seems most appropriate to adopt an inclusive definition of "boundary interaction" that encompasses the exchange of material goods, information, linguistic elements and genetic material. From this perspective, Manachaqui Cave's probable function as a wayside station offers a unique opportunity to take the pulse of interregional interaction at a major prehistoric crossroads. While migration explanations constitute convenient and occasionally elegant solutions to temporal and spatial problems, a framework emphasizing "boundary interaction" provides promising new avenues for future investigations. CHAPTER 2 CIVILIZING MIGRATIONS, VERTICALITY COLONIZATION AND THE EASTERN MONTANE FOREST "Los guerreros emplumados, los c6ndores y las cabezas clavas que ... en el Pajaten aparecen en forma primaria, en Chavin se replican y alcanzan su perfecci6n. cEntonces? Las investigaciones deben encaminarse en este sentido. En el Gran Pajaten debemos empezar a buscar las raices de Chavin, o sea las huellas del hombre cavernario que se transfor.ma en hombre-pensante, racional, humano ... ;Si hubo un Eden americano, lo hallaremos en el Gran Pajaten o cerca de el! Tal vez eso quiso decirnos el sabio medico llamado Julio C. Tello (Micerino the wise professor in Teofilo Maguina Cueva's Chavin: La Epopeya Jamas Contada 1990). This chapter constitutes a survey of the major theories that bear either indirectly or directly on tropical montane forest prehistory. In Chapter 1, I described the earliest notions which implicitly or explicitly characterized the montane forest as an empty frontier or barrier separating pre-Hispanic Andean and Amazonian worlds. I also described the "unitary origins" and "verticality" frameworks responsible for hypotheses invoking: 1) society-level population movements and 2) colonization. For descriptive purposes, these two migration scenarios can be labelled "civilizing migrations" and "verticality colonization," titles that reflect their theoretical underpinnings. In the first two sections of this chapter, I will describe prominent civilizing migration theories followed by the 64 65 major verticality colonization hypotheses. The third section will note alternative viewpoints expressed by a small group of "dissenting" archaeologists. A map is provided (Fig. 3) to locate the sites mentioned in the following text. Civilizing Migrations and the Eastern Montane Forest As contemporary scholars who frequently referred to one another's work, Meggers (1971) and Lathrap (1970) have been central figures formulating "civilizing migrations" that reputedly introduced "advanced" cultural innovations to their Andean or Amazonian destinations. However, the stage was set for the development of such frameworks by the early twentieth century theorizing of Spinden (1917), Uhle (1922b), Rivet (1968 [1924]), Steward (1949) and Tello (1967 [1921]). These and other scholars have worked under the assumption that a New World neolithic or "formative" stage complex of pottery, sedentism and agriculture had a single origin or hearth, and subsequent spread. Meggers and Lathrap are scholars who have most clearly labored under the influence of unilineal evolutionary "ruling frameworks." Meggers' framework is derived from her mentor Leslie White, the "intellectual heir" to L. H. Morgan (Willey and Sabloff 1974:154; Trigger 1989:290). Lathrap's is an eclectic blend of culture history and evolutionary theory (Oliver 1992) influenced by a diverse range of scholars including Charles 66 Darwin, Edward Sapir, Alfred Kroeber, Carl Sauer and Julio C. Tello. While working within the popular evolutionary paradigm of his era, Tello was inspired by nationalist and regionalist fervor to substantiate autochthonous origins for highland Peruvian civilization particularly to counter Uhle's (1922b} theory of "foreign" Mesoamerican origins (Lumbreras 1974:8). Although these population movement theories seem obsolete in the face of enlarged data bases and newer paradigms, Tello's inductively reasoned model reminds us that they persist as explanations for anomalous material patterning and hypothesized "site unit intrusions" in the eastern Central Andes and western Amazonia. Tello's theory persists in Peruvian popular culture as Maguifia Cueva's quotation at the beginning of this chapter attests. Rivera's (1975, 1991, Rivera and Rothhammer 1986} relatively recent presentation of evidence for a Preceramic Period "trans-altiplano" population movement from lowland Amazonia to the Chilean coast demonstrates that prehistoric population movements up and down the Andean slopes remain as viable hypotheses, and should not be rejected out-of-hand. Discussed below are major "civilizing migration" hypotheses that can be separated into two groups. Representing the first set are European and Peruvian scholars who argued for long-distance, society-level population movements. During the early twentieth century, 67 Tello and Rivet promoted theories envisioning population intrusions from the lowlands north and east of the Central Andes. Tello (1967 [1921]; 1942) postulated a precocious role for migrant Amazonian populations as pioneers of agriculture and donors of essential components of high culture to the Central Andes. Rivet (1968 [1924]) indicated the "oriente" along with more distant and diverse sources including Mexico, Central America and Polynesia for the succession of ancient Peruvian "civilizations." Both scholars relied on perceived distributions of material culture and native languages for supporting evidence. As Tello influenced generations of Peruvian archaeologists and inspired the more recent proposal of Amazonian origins by Lathrap, his hypothesis and its implications for montane forest prehistory will be outlined in some detail. Rivet's impact on montane forest prehistory has been less direct, and is apparent in subsequent theorizing by French scholars like Langlois (1939, 1940a, 1940b) and the Reichlens (1950), who first attempted to characterize ancient settlement in the northeastern montane forests. The French group represents a long tradition of European scholarship rooted in evolutionary assumptions (Binford and Sabloff 1982; Clark 1994). A more recent evaluation of northeastern montane forest prehistory by Peruvian archaeologist Kauffmann (Kauffmann et al. 1989) is explicitly rooted in L. H. Morgan's unilinear evolutionary 68 framework and its basic assumption of unitary origins (Kauffmann 1983:80-81). The more contemporary set of "civilizing migration" theories has been framed by the U.S. scholars Meggers, Lathrap and Lathrap's student Isbell (1974). Although Lathrap is known for his hypothesis of unitary Amazonian origins for the New World Formative Stage, he also proposed an additional, independent population movement for the peopling of the eastern slopes (1970:171-179) that was further elaborated by Isbell (1974). Isbell linked Lathrap's hypothetical movement of Quechua-speakers to the Central Andean developmental sequence and, coincidently, to the Middle Horizon expansion of the highland Huari &mpire. Although the intellectual pedigrees are clearly distinct, Lathrap's and Isbell's population movements, like Tello's, Rivet's and Meggers' all assume the synchronous movement of race, language and culture. Primordial Migrations and the Eastern Montane Forest Much of Julio C. Tello's writing may be regarded as a counterpoint to Uhle's thesis that Peruvian civilization resulted from migrations of Mayoid populations spreading southward from Central America (Uhle 1922a, 1922b) . Tello argued that the earliest civilizations originated within the modern-day borders of Peru, and that the mother civilization which gave birth to Mochica, Nazca and other later 69 "civilizations" developed in situ at the highland site of Chavin de Huantar (Tello 1942, 1943, 1960). However, Chavin's population and many of its cultural traditions derived directly from the tropical forest. During what Tello first termed an "Era Primordial," migrants of "low culture" (Tello 1967 [1921] :199) from the tropical forests arrived in the highlands, first entering the humid northeastern Andes and subsequently adapting to the dryer central Peruvian Andes as they moved west and south. These early forest migrants subsisting on agriculture, hunting and fishing, and bearing cult beliefs and myths featuring the jaguar-deity, formed a basal cultural stratum in the northeastern Andes (Ibid.:209, 1960). 1929; They also brought pottery and cultigens such as cotton, coca, maize, chili peppers, gourds, squash, beans, peanuts, sweet potatoes and manioc (1929:21, 1960:12). However, agricultural technologies and animal domestication first developed in the "healthy" highland environment. The fully-developed subsistence complex then descended to the arid and insalubrious western yunga and coast (1929:21, 166; 1960) . Tello presented ecological, archaeological, linguistic and iconographic arguments, coupled with observations based on eastern tropical forest ethnology and myth, to support his view that Chavin civilization drew constituent elements from the tropical forest. Like other scholars of his time, 70 Tello regarded the edge of the tropical forest as the natural limit of Andean civilization (1942:621). However, he was aware of the ecological complexity of the forest's edge (Ibid.:633), and believed that much of the cultural interchange between tropical forest and highland cultures took place in regions of ecological interdigitation, especially where warm tropical east-west, or "cis-Andean," valleys deeply penetrated the highlands. Tello's ecological argument for the transfer of cultigens from lowland to highland ecozones (1942, 1960) actually anticipated some of Sauer's ideas (1952:50) by a decade. In Tello's model, human colonization and the adaptation of lowland cultigens to colder environments occurred at moist intermediate elevations between the eastern lowlands and highlands where irrigation was unnecessary. The Marafion river valley in particular attracted Tello's attention as a strategic area for colonization by tropical forest populations exploiting both forest and cisAndean or yunga resources, and where lowland cultigens became adapted to colder and more varied environments {1960:18, 23). On both sides of the Marafion River, Tello initially identified settlements of a culture currently known as Recuay, which he alternately termed Recuay, RecuayPasto, Huaylas, Marafion and Archaic North Andean. This culture exhibited distinctive monumental stone architecture and ceramics, an elaborate lithic sculptural tradition and 71 shaft tombs. In his early work, Tello (1960) considered these as representative of the first Andean civilization. However, he eventually became aware of the stratigraphic superposition of Recuay over Chavin remains at Chavin de Huantar (Tello 1943:155), and his later writings emphasized the primacy of Chavin civilization (1942, 1943). To argue for the temporal priority of Chavin, Tello relied on stratigraphic and ceramic evidence. The fine, dark monochrome incised and polished pottery which he attributed to the Chavin style lay beneath temple constructions at Chavin de Huantar, and also underlay the archaeological remains of other cultures throughout both highland and coastal regions (1943:158-159). Tello also invoked ceramic evidence to advance his hypothesis of Chavin's tropical forest affiliations. He believed that sherds from the deepest strata at Chavin de Huantar resembled "the incised and carved types of the Amazon" (1943:152). The importance that he attached to the major Amazonian river valleys as hypothetical migration routes prompted Tello to visit the headwaters of the Marafion and Huallaga Rivers. At an elevation of approximately 2,000 m within the upper Huallaga River valley surrounding the town of Huanuco, Tello found evidence to support his tropical forest origins thesis. Artificial mounds at Kotosh and Shillacoto contained what he interpreted as "incised and carved" Amazonian pottery in strata below those containing 72 Chavin and Marafion types (Tello 1942:710). Tello was familiar with three pottery "types" found in "Anti Suyo," for which he provides cursory descriptions: 1) the early dark monochrome excavated at Kotosh, and reported from the localities of Matibamba (20 km downriver from Chavin de Huantar) and Ong6n (near present-day Tayabamba in southeastern La Libertad Department), Monzon (northeastern Huanuco Department) and Satipo and San Luis de Shuaro (Junin Department), 2) bichrome and polychrome pottery with geometric and cursive decoration from the upper Utcubamba and lower Marafion river valleys and 3) glazed, polychrome pottery like that produced by the Conibo and other Panoanspeaking groups in the Ucayali River valley (1942:635-636). A close look at Tello's schematic view of Andean cultural development (1942:715) reveals that he envisioned these three types as sequentially representative of the first, second and third ages or epochs. First and second epoch pottery styles cluster in and around the eastern montane forests. Other lines of evidence utilized by Tello to argue for tropical forest origins are of varied nature. He perceived historical relationships between linguistic groups stretched along the same latitudinal parallel like the coastal Muchik and the Amazonian Amuesha, Chol6n and Hivito languages as evidence for the westward expansion of Arawakan speakers {Tello 1942:629). Also indicative of the northeastern 73 Andean cultural substratum formed by Tello's early forest migrants are ethnographically recorded myths featuring a dragon, the moon and sun and their twin offspring (Ibid.:629), and the jaguar cult (1929:168). The tropical forest jaguar, Tello argued, is the principle deity depicted in Chavin sculpture (1960:160), and appears as the god of rain in the highland Wira-kocha myth (1929:168). Secondary deities in Chavin art include condors, bats, serpents, fish and other fauna common to Andean habitats (Ibid.:168). Tello also observed a number of Andean cultural traditions which he believed originated in the eastern forests. Among these were the taking of trophy heads and the attached belief systems, cranial deformation, urn burial and other funerary practices, the manipulation of feathers for sumptuary display and the use of stone axe heads. In summary, Tello argued that Andean civilization was conceived in the highlands. However, its basic constituents such as populacions, cultigens and technologies invariably originated farther east in Amazonia. A review of Tello's maps and diagrams (Tello 1943: Lam. I, p.645; Lam. VII, p.715), and the archaeological sites specified in his discussions, indicates that the area which he alternately referred to as the "floresta," "Andes Orientales," and "Antisuyo" lies in the sub-tropical forest at the eastern base of the Andes, or Lathrap's upper montana (1970). As Bonavia (1981) points out, Tello never actually worked in 74 the eastern montane forest proper. The middle elevations of the Marafion and Huallaga river valleys served as loci for the "colonization" or upslope expansion of lowland populations and cultigens in the process of adapting to new environments. The headwaters of the Marafion and Huallaga rivers ultimately yielded pottery and other kinds of evidence which Tello utilized to construct his argument. Criticisms of Tello's theoretical positions aimed at his interpretation of Chavin as the first pan-Andean civilization (e.g. Larco 1941; Kroeber 1944), rather than his population movement hypothesis. In fact, he makes reference to Rivet's support for his ideas regarding tropical forest origins (1942:633). Uhle, Jacinto Jij6n y Caarnafio (1927), Rivet and Tello all agreed on migrations from a northerly direction. Tello parted company with this group by proposing exclusively Amazonian, as opposed to Central American origins. Civilizing Chachapoyas The upper elevations of the northeastern montane forest between the Marafion and Huallaga Rivers correspond to the Inca Province of Chachapoyas. The name Chachapoyas has been retained by historians, anthropologists and archaeologists referring to the ethnographic and archaeological culture area surrounding the Utcubarnba River Valley. Unlike other "civilizing migrations" postulated to account for prehistoric events in the heart of the Central Andes, these 75 hypotheses reflect an early scholarly fascination with Chachapoyas. As we will see, hypothesized migrations converge on Chachapoyas from all directions. Major "civilizing migrations" impacting Chachapoyas from the north were postulated during the early twentieth century. An early consensus for northern and/or eastern origins, based largely on intuition rather than rigorous analyses, lasted until the mid-1960s when population movements from the west were proposed. These theories are described below. Population Movements and "Cuelap Civilization" Adolf Bandelier (1909) wrote the first monograph on archaeology and ethnography of the northeastern montane forest Utcubamba Valley. Like scholars arriving before and after him, Bandelier devoted considerable attention to the spectacular Chachapoyas site of Cuelap (also Kuelap) . Setting the stage for subsequent investigations in the Inca Province of Chachapoyas, Bandelier noted similarities between the Chachapoyan Quechua dialect spoken and the southern Andean Aymara language, as well as an abundance of native names of places and individuals which could not be related to Quechua. He concluded, "the country of Chachapoyas was once inhabited by a tribe or tribes, that belonged to a stock different from the Peruvian mountaineers, their western neighbors" (1909:10). After his survey of Chachapoyas archaeology, Langlois (1939, 1940a, 76 1940b) hypothesized population movements "from the north or the northeast" (1940b:226; this and subsequent translations from Spanish are mine) that recall Tello's claim for the importance of the northeastern Peruvian Andes as the entryway for the tropical forest populations ultimately responsible for Chavin civilization. However, Langlois' interpretations were most strongly influenced by Uhle's ( 1922b) ·and Jij 6n y Caamafio' s ( 1927) arguments for long- distance migrations of "archaic cultures" from Central America. Langlois utilized evidence (mostly toponymic) of intrusive linguistic elements from Ecuador and Mexico to bolster his suggestion. With the encouragement of Rivet in the 1940s, Henry and Paule Reichlen reported evidence for three successive "civilizations" spanning five centuries of continuous preHispanic occupation based on reconnaissance and excavations in and around Cuelap. They were most impressed by resemblances between pottery of the "Cuelap Civilization" and styles of the Amazon, Orinoco, Ecuador and Colombia (Reichlen and Reichlen 1950:240). Neither the Reichlens nor Langlois cite Rivet's theory of multiple migrations into the Central Andes from the north and east (Rivet 1968 [1924]), yet Rivet's vision likely influenced their interpretations. Nine years after the Reichlens' publication, Horkheimer's synthetic review of Utcubamba valley prehistory expressed support for Langlois' original thesis of "primitive 77 immigrations from the north or northeast" (Horkheimer 1959:89), although he declined to cite more specific points of origin. Regarding the "immigration" process, Horkheimer theorized: The topography shows us that the entrance to the Utcubamba valley is relatively easy from the north and that, in contrast, natural communication with the south and west is relatively difficult ... The immigrants stayed within the blind alley of the Utcubamba where the arrivals from lowland and tropical regions found a series of ecological elements to which they were accustomed and that other regions lacked. The difficulties of communication with the south and west caused, for centuries, an isolation during which the established immigrants developed their own cultural characteristics without much outside influence. Later, with increasing population and changing needs and aspirations, contact between the inhabitants of the Utcubamba and their neighbors, especially in the region of Cajarnarca, was accentuated (Ibid.:89-90). Apparently, Horkheimer's "immigrants" intruded upon local "civilizaciones desconocidas" (unknown civilizations) that he places at the bottom of his "scheme of cultural succession in the Utcubamba region" (Ibid. :Fig. 1). He postulates that "Cuelap Civilization," characterized by influences from nearby highland Cajamarca, was in place by A.D. 1000. Most recently, Lerche (1995} has restated the same hypothesis for which he finds supporting evidence in the province's restrictive geography and its ceramic styles, iconography and historical linguistics. Lerche believes that the "Chachapoyas cultural complex" intruded from the north or east around A.D. 900, following a 400 year period during which the province was only lightly inhabited (Ibid. : 3 7} . The undated Huayabamba ceramic complex 78 excavated in the central Huallaga Valley lowlands by Ravines (1978) purportedly represents the eastern limits of Chachapoyas (Ibid.:21). Like Ravines (1978:531; 1981a:178), Lerche believes that the lowland Huayabamba complex is ancestral to Cuelap styles (Ravines declined to specify migration or any other mode of style transmission) . A carved wood statue reputedly from a Chachapoyas tomb located east of Leimebamba bears the representation of a cayman, thereby providing Lerche with additional evidence for lowland origins (Lerche 1995:75). By "north and east," Lerche clearly refers to Amazonia. "Highlandizing" the Montane Forests Since 1980 Kauffmann has investigated Cuelap, Gran Pajaten (also called Abiseo) and other archaeological sites of the northern and northeastern montane forests between the Ecuadorian border and Huanuco Department. However, he refers to the greater northern and eastern montane forest from Ecuador to Bolivia comprehensively as the Andes Amaz6nicos or "Amazonian Andes" (Kauffmann 1986, 1987, 1992b; Kauffmann and Ligabue 1990; Kauffmann et al. 1989). During the past two decades Kauffmann has reported on field investigations funded by Italy's Centro Studi Ricerche Ligabue, and presented a "working hypothesis" that he terms "serranizaci6n de la selva," or the "highlandizing of the jungle, " in reference to a proposed prehistoric eastward expansion of both highland populations and anthropogenic 79 paramo environments produced by deforestation (1986, 1987, Kauffmann and Ligabue 1990; Kauffmann et al. 1989). The highlandizing process as envisioned by Kauffmann was preceded by a long period of prehistoric population growth and agricultural intensification in the Andean coast and highlands. Kauffmann subscribes to a hypothesis of unitary origins for New World "High Culture" in coastal Ecuador, and spread to the Central Andes (1992a:96-102). In numerous publications, Kauffmann has expressed his view that virtually every aspect of Central Andean cultural development was caused and conditioned by the constant state of crisis rendered by overpopulation, insufficient agricultural land and the shortage of food (Kauffmann 1976, 1983:81, 1991, 1992b; Kauffmann et al. 1989:8). The eastward population movement began when carrying capacity was reached in the Central Andes. Kauffmann observes that virtually nothing is known of indigenous montane forest inhabitants who were either absorbed, displaced or exterminated during the eastward population movement (1986:6; 1987:6). Contact with these populations, adjustments to the environment and relative isolation imbued the resulting cultures of the Amazonian Andes with Amazonian features and "singular formal aspects" although they remained fundamentally Andean in character (1987:5, 1992b:108). Kauffmann's eastward population movements reportedly 80 began by A.D. 1 (1992b:108; Kauffmann and Ligabue 1990:336; Kauffmann et al. 1989), and took place in "systematic form and administered by coercive powers" that extensive terracing systems, (1987:6). He theorizes "cultist-administrative centers" like Gran Pajaten, and elite cemeteries provide evidence for centralized political control of the highlandizing process (1987:6; Kauffmann et al. 1989:7-8). Rock art and lithic sculpture documented in the northern montane forests near Jaen indicate "pre-agricultural" and "formative" occupation respectively (1991:47). However, it remains unclear whether Kauffmann regards these earliest finds as indigenous or intrusive. He attributes the sculptural style of lithic carvings from the Tabaconas River valley (also near Jaen) to influence from "formative" highland centers like Kuntur Wasi (Ibid.). Kauffmann's hypothesis is a broad-brushed and panoramic scenario fashioned to a large degree for consumption by a lay audience, especially Peruvians and Italians. In support, he typically provides a general discussion of archaeological features encountered in the northern and northeastern Amazonian Andes. He does not differentiate the indigenous from the highlandizing intrusive populations utilizing archaeological evidence. largely hypothetical. These remain, therefore, It is the amalgamation of the two hypothetical populations, or its material remains, Kauffmann customarily describes. that His hypothesis dimly 81 echoes Murra's verticality-type colonization coordinated by "coercive" authorities, yet his unitary origins framework sets Kauffmann's theory squarely within the tradition of "civilizing migrations." Neolithic Migrations and the Eastern Montane Forest Most influential to the work of Lathrap were Tello's notions that lowland to highland migration characterized early relations between the Amazon and Andean areas, and Sauer's view that the botanically-rich lowland tropical forests were the likely hearth where agricultural systems first developed. Domesticated cultigens, sedentary agricultural economies, ceramic technology, a complex cosmology and related art, Lathrap argues, are among the tropical forest's contributions to Andean civilization. He assumed Tello's position that the first pan-Andean civilization sprang from the center at Chavin de Huantar, and that Chavin's residents were agriculturalist migrants from the eastern tropical forest, perhaps arriving via the Marafion valley (Lathrap 1973b:92-93). Lathrap also followed Tello in the belief that the migrants brought with them a complex mythology based on tropical forest cosmology. Yet while Tello argued for a highland crystallization of Chavin civilization, Lathrap (1970, 1971, 1973b, 1974, 1977) contended that the entire Chavin complex arrived with intrusive Arawak-speaking populations bearing a fully 82 neolithic economy and complex social organization developed many centuries prior on the Amazon River floodplains. Much of Lathrap's work aimed at discrediting the proposition advanced by Steward, but elaborated by Meggers and Clifford Evans, that prehistoric cultural evolution in Amazonia was retarded by the region's uniformly low agricultural potential (Meggers 1954). Meggers and Evans (1983:330) have consistently argued that "the lowlands of South America were the recipients rather than the originators of major technological innovations, especially agriculture and pottery making." From their perspective, Amazonian societies contributed little, if anything, to the development of Central Andean civilization. During his fieldwork on the Central Ucayali, however, Lathrap observed cultural and ecological processes on the river floodplains and proposed mechanisms for Amazonian population growth based on tropical forest subsistence and demographic patterns. A "phenomenally efficient system of cultivation of bitter manioc" supplied the motor, churning out populations in excess of local floodplain carrying capacities and driving them upstream into every available floodplain niche and up through the eastern montane forests (Lathrap 1970, 1977:744). Chavin's immigrant populations expanded from their Amazonian homeland southward "along the eastern face of the Central Andes" where they penetrated the north and south-trending Andean valleys of the Marafion and 83 Huallaga rivers (Lathrap 1982:312). The interpretation of Chavin iconography is central to Lathrap's argument for the "Chavin pattern's" tropical forest origins (1971, 1973b, 1974, 1982). The images evoking lowland Amazonian cosmology and staple foods depicted on the carved stone Tello Obelisk (1973b, 1982) represent "a celebration of the prior [tropical forest] agricultural system which allowed the Chavin elite to achieve its current dominance" (Lathrap 1973b:103). Thus, the depicted mythological event and inventory of lowland cultigens constitute a material record of the subsistence economy that fueled the up-stream expansion of floodplain populations, the intrusion into the eastern highlands and settlement at Chavin de Huantar. At the time of Lathrap's investigations (prior to Burger's reinterpretation of the Chavin horizon), scholars believed that Chavin's sphere of influence had disseminated as far as the south coast Ica valley by 1000 B.C. (Patterson 1971). To debates regarding Andean civilization's origins, Lathrap brought new ceramic evidence derived from the earliest and most complete stratigraphic sequence from the upper Amazon Basin (Lathrap 1958, 1962). His Early and Late Tutishcainyo Phases from the Central Ucayali drainage roughly correspond to the Central Andean Initial Period, while the Shakimu Phase (later sub-divided) was thought coeval with the spread of Chavin. Like Tello, Lathrap 84 sought to buttress his population movement hypothesis by searching beyond Chavin to find archaeological localities that served as migratory links. the Japanese He took special interest in excavations at highland Kotosh in the upper Huallaga Valley where the earliest ceramics comprised a large sample of the same early pottery first identified by Tello as "Amazonian" (Izumi and Sono 1963). Lathrap utilized the new evidence to restate Tello's argument for an early intrusive Amazonian population movement into the upper Huallaga River drainage. This contention hinged on a series of stylistic similarities shared between the Initial Period Wairajirca Phase pottery, Tutishcainyo Phase ceramics, and the geographically and stylistically intermediate Cave of the Owls fine wares from Tinge Maria (Lathrap and Rays 1963). The Wairajirca material directly overlay preceramic temples of the Kotosh Mito Phase. Meggers and Evans were likewise interested in the upper Huallaga and Ucayali sequences. They proposed that Tutishcainyo and Wairajirca were both manifestations of a cultural sub-stratum characterized by pottery featuring zoned-hachure decoration and flanged rims found beneath Amazon Basin sequences ranging from the Rio Napa of Ecuador to Maraj6 Islru1d of eastern Brazil (Meggers and Evans 1961, 1983; Meggers et al. 1965; Evans and Meggers 1968). The "Zoned-hachure Horizon/Tradition" was said to have originated during the third millennium B.C. Tesca Phase on 85 the Caribbean coast of Colombia. Meggers and Evans hypothesized that the tradition was carried from the Andes into Amazonia during an episode of increased aridity between 2,000 and A.D. 1 that facilitated population movement by opening forest-free corridors (Meggers 1975, 1977, 1979). Accordingly they viewed lowland Tutishcainyo as a derivation of the highland Wairajirca style (Evans and Meggers 1968:92). In response to Meggers and Evans' skepticism regarding his claims for Tutishcainyo•s great antiquity and temporal priority, Lathrap (1971) constructed a complex argument utilizing cross-dates in lieu of lacking radiocarbon evidence in order to demonstrate the temporal priority of the Tutishcainyo tradition. To demonstrate how the pottery styles reflected lowland to highland population movement rather than the reverse, he listed a series of Early and Late Tutishcainyo stylistic attributes that appear to be intrusive within the Wairajirca assemblage. Lathrap (1965, 1970, 1971) argued that, at the same time, Central Andean ceramic traditions had exerted no visible influence on the lowland Tutishcainyo pottery. Lathrap described the Wairajirca Phase assemblage as "a blending of two distinct cultural traditions, one of which was very like Early Tutishcainyo, and relates to an upward expansion of tropical forest peoples to the eastern slopes of the Andes; the other of these traditions appears to be 86 indigenous to the Central Andes or Peruvian coast" 1971:93). (Lathrap While the Wairajirca neckless ollas typify earliest Central Andean traditions, a hypothetical "collateral relative" of the Amazonian Tutishcainyo style donated carinated vessel shapes, and zoned-incised and postfire painted decorations deemed intrusive at Kotosh. Lathrap emphasized that all Wairajirca vessel shapes and decorative modes, except for the neckless olla, have Early Tutishcainyo "homologies." Similarly shaped double-spout- and-bridge bottles appear in both assemblages. This evidence, combined with the lack of neckless ollas and other signs of highland-derived influence in the Tutishcainyo assemblage, provided Lathrap with a powerful case for upslope rather than downslope intrusion (Lathrap 1970:106107, 1971:94). Lathrap aligned the following Kotosh Kotosh Phase with the final years of the Ucayali Late Tutishcainyo and subsequent Early Shakimu phases. Kotosh Kotosh radiocarbon dates are 890 ± 170 B.C., 920 ± 230 B.C. and 1120 ± 110 B.C. (Izumi 1971), while his Early Shakimu Phase yielded the date of 650 ± 100 B.C. (Lathrap 1971). The Early Shakimu Phase appearance of slipped and burnished bowls with excised decoration similar to those of the Kotosh Kotosh Phase suggested a historical relationship to the spreading Chavin art style (Ibid. :92) and a shift to downslope diffusion (Ibid.:88). Stylistic similarities between vessels of the 87 Kotosh Sajara-patac Period and Late Shakimu Phase ostensibly demonstrated continuing communication between the highlands and lowlands that ended abruptly around 200 B.C. with the Central Ucayali appearance of the Barrancoid pottery tradition (Lathrap 1970:117). The coeval Higueras Phase at Kotosh represents the Andean Early Intermediate Period. Lathrap theorized that the early Kotosh sequence documents a "site-unit intrusion" (1971:94) of Amazonian populations that dominated the indigenous Andean populations of the upper Huallaga River valley. for the Tello-Sauer hypothesis Hence, he found support (Lathrap 1970:107), explicitly rejected Meggers and Evans' downslope hypothesis, and concluded that, "The earliest sedentary communities with elaborate masonry structures so far discovered in the Central Andean highlands appear influenced by, and indeed partially derived from, tropical forest peoples" (1971:94). The fundamental assertions coursing through all of Lathrap's publications are: 1) neolithic economies and stratified societies had developed fully by 3,000 B.C. in the tropical lowlands and 2) that the fully-developed complex penetrated the Central Andes intact via population movements from Amazonia. Some of his bold and provocative assertions have been sustained, although far less dogmatically, by the small group of scholars that he trained. Support for Lathrap's interpretation of developmental events at Kotosh appeared with Kane's conclusions derived 88 from study of material culture excavated at Kotosh and nearby Shillacoto (Kano 1972, 1979). At Shillacoto, archaeologists uncovered a temple similar to the preceramic Kotosh Mito Phase temples, yet utilized during the Initial Period Wairajirca Phase (Burger and Salazar-Burger 1980:30). Kano utilized stylistic and iconographic evidence to demonstrate that the Chavin "feline cult" had a history of pre-Chavin stylistic development in the Huanuco basin, and still earlier roots in as yet undiscovered Amazonian styles. He admitted, however, a lack of supporting evidence for concurrent shifts in construction styles, subsistence economy and social organization expected to accompany such a population replacement during the Mito/Wairajirca transition (Ibid.). Most criticisms of Lathrap's position came from scholars arguing for Andean civilization's temporal priority on the Central Coast. Lanning dismissed the possibility of any historical relationships between Kotosh and Tutishcainyo pottery, suggesting instead "some interchange of stylistic ideas" (1967:87-88). Fung (1982 [1971] :457-458) remarked that evidence for socio-political complexity is lacking at Kotosh. Lathrap's lowland immigration hypothesis received close scrutiny from proponents of Moseley's theory of Andean civilization's maritime origins. Moseley questioned the validity of Lathrap's analytical procedure by dismissing Initial Period pottery styles as useful analytical units: 89 In ceramic assemblages from the early part of the Initial Period, ornamentation is generally based upon a limited repertoire of incised geometric and curvilinear motifs, and the vessel-shape categories are not numerous. Styles are therefore relatively unelaborated, and often share superficial resemblances because of their simplicity (Moseley 1983:209). Furthermore, Moseley contested Lathrap's conceptual linkage of pottery style, tropical forest agriculture and complex social organization, thereby rejecting the validity of utilization of pottery styles as indices reflecting "the invention and diffusion of the economic foundations of New World civilization" (Ibid.). Moseley pointed out that corporate architecture (which he regards diagnostic of complex social organization) pre-dates pottery in the Andes. Finally, he echoed Kane's observation that Kotosh produced no evidence that changes in subsistence practices and architectural styles accompanied the appearance of Wairajirca pottery during the Mito/Wairajirca transition (1983:209, 210). More recently, Moseley (1992) has acknowledged the temporal priority of the Amazonian Tutishcainyo tradition around 2000 B.C., but he provides an alternative to Lathrap's population movement hypothesis. To account for the stylistic resemblances, Moseley considers tropical forest influences on Wairajirca the result of verticality's "downward thrust" (Ibid.:46) and highlander exposure to lowland ceramic technology during regular forays into the lowlands (Ibid.:lOO, 144). Moseley's idea suggests highland "procurement" rather than lowland "bestowal," and highland 90 developmental superiority is unstated, yet strongly implied. Moseley's critical remarks recounted above reflect the relatively recent consensus that agriculture, sedentism and monumental construction each have their own developmental histories in the Central Andes dating from the Preceramic Period. Chavin civilization is presently characterized as an amalgamation of cultural traditions drawn from several regions,· the tropical forest included (Burger 1992; Lumbreras 1993). The current perception of Chavin as nexus has rendered moot any consideration of Amazonian site unit intrusions. Lathrap's interpretation of the Initial Period U-shaped coastal monuments as representations of the cayman's jaws (Lathrap 1985) constitutes a parting shot at Lanning's and Moseley's maritime origins theory. Many scholars acknowledge tropical forest sources for the imagery carved in Chavin stone sculpture, yet its significance has rarely been interpreted as more than evidence of Chavin's amply sized interaction sphere (e.g. Lumbreras 1981; Raymond 1988). Nevertheless, Raymond still regards Lathrap's postulated migration as a valid hypothesis (1988:293). Raymond emphasizes the temporal priority of tropical forest agriculture and pottery technology, and reiterates Tello's and Lathrap's views regarding the importance of the Marafion River valley as a likely route through which early lowland cultigens penetrated the highlands. Transhurnance or "limited migration" are cited as 91 possible modes of agriculture's early diffusion from the north highlands to the Peruvian coast (1988:294). Lumbreras (1993:354-355) also continues to emphasize Chavin de Huantar's auspicious location at the headwaters of the Amazon River's Marafion and Huallaga affluents. In a recent consideration of Chavin sculpture's tropical lowland imagery and Chavin de Huantar's origins, Burger (1992:154-155) rejects Lathrap's population movement hypothesis for two reasons. First, he questions Lathrap's use of sculptural iconography and its mythic representations to reconstruct historical sequences and population movements. Second, Burger submits that if Lathrap's migration scenario were accurate, then the earliest pottery styles from Chavin de Huantar's cultural sequence should reflect tropical forest ancestry. Instead, Burger interprets the earliest Chavin pottery, assigned to the Urabarriu Phase (Burger 1984b), as an amalgamation of diverse Central Andean styles that reflects Chavin's economic and social interaction with distant coastal and highland regions (1992:155). Burger also pointed out that subsistence data from Chavin de Huantar show a mixed highland agricultural economy, the product of a long history of local Andean development (Ibid., 1995; Burger and van der Merwe 1990; Miller and Burger 1995). To Burger, the tropical forest imagery that Lathrap utilizes to bolster his hypothesis most likely represents the importation of foreign 92 esoteric knowledge from the lowlands by leaders at Chavin de Huantar's Old Temple (1992:157-158). Such sacred knowledge could have been obtained during the long journeys undertaken by shamans, and its lowland source would have facilitated acceptance and respect as a source of supernatural power. Quechuas, Maize and the Eastern Montane Forest Like the Langlais-Lerche hypothesis of northeastern montane forest peopling, Isbell's (1974) elaboration of Lathrap's hypothesis postulates population movements originating in distant premontane or subtropical forest settings to the north. Isbell's hypothesis combines interpretations of archaeological, ecological and linguistic data to generate a theoretical model for the peopling of the eastern montane forest by Quechua-speaking populations. Isbell took Lathrap's model one step further however, describing a population intrusion with widespread impact throughout the Central Andes. His argument is complex and requires acceptance of a series of assumptions beginning with several proposed initially by Lathrap. At the onset of the Higueras Phase in the upper Huallaga valley near Huanuco, Lathrap noted a sudden shift of site locations from valley bottoms to defensible promontories between 3,500 and 4,000 m, accompanied by a shift to slope agriculture on artificial terraces. The simultaneous adoption of coarse, brown ware ceramics featuring large jars with flaring rims 93 and red slip, crude modeling and applique decoration purportedly indicates "a total replacement of population" sometime between A.D. 100 and 600 (Lathrap 1970:173). This intrusion of Quechua-speakers from the Bolivian montane forests far to the south ostensibly marked the beginning of the Central Andean Early Intermediate Period (Ibid. 174176). Examples of coarse wares purportedly associated with eastern-slope Quechua speakers include pre-Tiahuanuco Chulla Pampa ceramics from Bolivian Cochabamba, Late Intermediate Period Killke from highland Cuzco, Chacra de Giacomotti in lowland Junin, and pottery of the historic Chupachu of Huanuco. In effect, Isbell borrows Lathrap's theoretical approach linking Amazonian ecological zones, language groups and pottery styles, and applies it to the highland Central Andes. Of central importance to Isbell's model of Quechua expansion is Murra's postulated association between terracing systems and maize cultivation (Isbell 1974:142; Murra 1956 [dissertation cited by Isbell], 1980:7, 12). Maize served as the staple for an expanding Quechua population which developed terrace agriculture as one of several stages accommodating its agricultural technology to the Andean slopes. Second, Isbell reasons that, "the origin of the adaptation to the valley slopes should be found in an area where dry cultivation of maize was possible without the necessity of terraces" (Isbell 1974:150). Discounting the 94 dryer and colder regions south and west, he proposes that the evolution of maize cultivation systems should have taken place in the warm, moist valleys between Macas, Ecuador and Chachapoyas, Peru (Ibid.). Here dry maize cultivation could be adapted to gentle slopes with a simple system of banking earth, and to steeper slopes by employing masonry walls to retain the thin soils typical of the eastern flanks. The development of terracing techniques permitted initial Quechua settlement on the moist eastern slopes. The subsequent addition of irrigation to terrace systems allowed the system's transfer into dry Andean environments further south which, in turn, facilitated continued Quechua expansion. The principal types of evidence employed by Isbell are ceramic attributes, radiocarbon dates and historical linguistics. In order to conceptually link the coarse brown wares to Quechua population movements he implements Rouse's classificatory terminology to refer to a coarse brown or "CB series." In addition to jars with flaring rims, he judges strap handles and hemispherical bowls to be diagnostic attributes of the CB series. These features appear most notably in pottery from Lambayeque, Huamachuco (terminal Early Intermediate Period), Gran Pajaten, Huanuco (Kotosh Higueras Phase and historic Chupachu), Huancayo (mixed with Early Horizon San Blas), Middle Horizon Ayacucho, Cuzco (Killke and Inca), the Ica valley (Ocucaje Phases 8 and 9) 95 and Eastern Bolivia. Isbell describes radiocarbon evidence for the appearance of the CB series at eastern Andean sites noting dates of A.D. 70+200 for the Higueras Phase and an additional date "in the first century A.D." associated with the Bolivian Chullpa Pampa style. The presence of the CB series in the south coast Ica valley at the end of the Early Horizon (Ocucaje Phase 9) suggests to Isbell that the CB series "could have been introduced in the Central Andes towards the second half of the first millennium B.C." (1974:146). Isbell regards suggestions of a Central Andean protoQuechua hearth by linguists Parker (1963) and Torero (1964) as subject to alternative interpretation. Most important to the region and time period pertinent to Isbell's study is the present-day distribution of Quechua A speakers in the former Departments of Cajarnarca, Lambayeque, Amazonas, Ecuador, San Martin and Loreto lying north and east, and circumscribing Quechua B dialects in the central Peruvian Andes. According to Isbell, early eastern Andean occurrences of the proposed CB series such as the Higueras style relate to upslope population movement of proto-Quechua A/B speakers (1974:140). He proposes a subsequent divergence of the two Quechua sub-groups within the Central Andean highlands immediately prior to the Middle Horizon (Ibid. :148). Isbell's interpretation of the distributions of coarse 96 wares and Quechua languages was published four years after the hypothesis of his mentor. When later confronted with alternative models for the prehistoric spread of Quechua from the central coast (e.g. Torero 1974; Bird et al. 198384), Isbell restated his position (Isbell 1983-84:253) and pointed to corroborating linguistic evidence for eastern Ecuadorian Quechua origins presented by Louisa Stark (fully published in 1985). In a lengthy and complex discussion of the origins of Amazonian Shipibo art, Lathrap unreservedly endorsed Isbell's model which should therefore be viewed as superseding his own south to north hypothesis (Lathrap et al. 1985). For the first time, Lathrap explicitly outlines the assumptions central to virtually all of his work on Amazonian culture history in a series of nomothetic statements: " ... on a pre-state level of organization ... language does not spread geographically unless there is a concomitant movement of appreciable groups of people" (Lathrap et al. 1985:43). Furthermore, he states that populations (and their languages) do not move unless forced to migrate due to population growth or environmental degradation, and population movements are enabled by technological innovations. Finally, " ... people cannot preempt the land of their neighbors unless there is some kind of competitive superiority either in sheer numbers, more integrated organization, or in other factors giving a 97 military edge" (Ibid.). These same assumptions provide the foundations for Isbell's hypothesis. Like Isbell, Lathrap et al. (Ibid.:76, 83) places the hearth of proto-Quechua in southeastern Ecuador. Arguing for Quechua influence in Curnancaya assemblages of the upper Ucayali, Lathrap et al. utilize evidence from modern Shipibo-Conibo narratives referred to as the "Inca cycle." The Curnancaya style appearing on the Central Ucayali by A.D. 800 purportedly corresponds to the arrival of Quechuaspeaking intruders that subsequently dominated Panoanspeaking lowlanders (Lathrap et al. 1985:74). The authors regard Curnancaya as derived from the red-banded-incised pottery recovered near Macas in Ecuadorian Amazonia. Ostensibly providing further evidence of Quechua affiliation, the applique faces on the upper portion of Sivia jars purportedly link the Curnancaya tradition to "plain brownware styles of the northeastern Peruvian Andes," and especially "the style known as Cuelap" 1985:68). The authors conclude that, (Lathrap et al. "the hearth of lowland Quechua expansion must be placed in the deep archaeological sites of the Macas region of Ecuador, and the initial expansion toward Peru must have been down the Upano-Zarnora trough" (Ibid.: 83). The composite Isbell/Lathrap model postulates a hearth of proto-Quechua speakers on the major tributaries of the Santiago River south of Macas, and southward population 98 movements along the base of the Central Andes. Migration occurred either as a wide swath or as a multi-pronged advance which by A.D. 1 incorporated the eastern Peruvian highlands (e.g. Kotosh Higueras and Chulla Pampa styles). By A.D. 800 the Quechua-speakers reached the lowland Ucayali and Apurimac drainages (e.g. Cumancaya and Sivia complexes). The CB series constitutes the salient material evidence for the Quechua spread into the "previously unoccupied niche on the eastern slope of the Andes" (Ibid.:76), while pottery derived from the Ecuadorian red-banded-incised style accompanies the intrusion of a Quechua-speaking population deep into the eastern lowlands. Criticisms bearing on southward Quechua population movements target the conceptual linkage between language, pottery and mythology assumed by Lathrap and his colleagues (DeBoer and Raymond 1987:120-121). DeBoer and Raymond note that Stark's (1985) postulation of an eastern Ecuadorian hearth for proto-Quechua provides a point of support for the northern origins of the language, but they protest Lathrap's liberal application of untested presumptions. Lathrap's (Lathrap et al. 1987:237-238) rejoinder argues tersely that the proposed linkage of language and pottery styles is not only justified, but that every archaeologist should be manipulating linguistic data in conjunction with archaeological data analysis. An additional criticism of Lathrap's original 1970 argument by Hastings (1985:555) 99 finds the suggestion of north and south population movement "difficult to accept" and "counterintuitive" given the imposing natural barriers presented by eastern slope topography. The larger scholarly community seems to have shared little interest in these hypotheses and subsequent debate. Verticality, Colonization and the Eastern Montane Forest The previous section described how the eastern montane forest has been envisioned as a threshold traversed by waves of migrating populations moving upslope or downslope. Yet even prior to 1911 and the "discovery" of Machu Picchu it was clear that pre-Hispanic populations once inhabited this seemingly inhospitable region. The immense fortified settlement of Cuelap had already been visited during the nineteenth century by Bandelier (1906), Raimondi (1942), Middendorf (1895), Wiener (1884) and others. Concerted efforts to investigate archaeological occupations in and around Peru's eastern montane forest began only three decades ago. The pioneering explorations of Bonavia (1964, 1968b, 1968-69, 1972b), Thompson (1973), Lathrap (1970) and Lathrap's students Allen (1968), Isbell (1968) and Raymond (1972) marked an era of growing interest in the montane forests which had become known within the archaeological community by a number of terms, especially the Peruvian vernacular ceja de selva (Bonavia 1964) and ceja de montana 100 (Lathrap 1970). Bonavia and Ravines (1968:153) first conceived of the ceja de selva as synonymous with Tosi's (1960) "bosque muy humedo montana" or "very humid montane forest" Holdridge life zone (3,800 - 2,500 m). However, Hastings describes a ceja de montana between 1,500 and 4,300 m, while Lathrap employs the term "ceja" even more liberally to refer to habitats down to 300 m (1970:95). Regardless of the terms utilized, research in recent decades has demonstrated that prehistoric montane forest settlement typically straddles multiple ecological zones. Unlike the somewhat anachronistic "civilizing migration" hypotheses, colonization theories based upon Murra's verticality model have been on the cutting edge of Andean studies for the past two decades. The precedent for these studies was set by Bonavia's (1964, 1967-68, 1968b, 1972b) early investigations in the central and northeastern montane forests. His fieldwork provided the basis for his synthetic overview of eastern montane forest prehistory, and the formulation of a seminal working hypothesis, co-authored with Ravines, that attributed montane forest settlement complexes like Caballoyuq and Gran Pajaten to "tropas de colonizaci6n agricola" (Bonavia ar.~ (agricultural colonization troops) Ravines 1967:62; Bonavia 1968b:74). The theoretical statements expressed by Bonavia and Ravines incorporated core elements of Murra's verticality thesis, especially the "ideal economic pattern in which each group 101 had access to natural resources from other ecological levels" (Bonavia and Ravines 1967:62). The authors cited the Visita Hecha g la Provincia de Chucuito por Garci Diez de San Miguel en el Afio 1567 (Diez de San Miguel 1964 [1568]) published by Murra, along with Nufiez del Prado's (1958) study of the Q'ero. Bonavia later directly cited Murra's model of vertical settlement distributions to support his interpretations of field observations in eastern Ayacucho (Bonavia 1972b:29). Estimations of dates, site functions and cultural affiliations, coupled with a series of ecological assumptions, constitute the foundations of Bonavia and Ravines' theory. The authors (1967:67, 1968:155-156, 157) hypothesized that, prior to Inca imperial expansion, the ceja de selva was increasingly utilized on a temporary basis by highland communities expanding agricultural production into eastern slope ecozones. They characterized initial colonization as "gee-cultural" and "a spontaneous step," and categorically discounted demographic pressure as the principal cause (1967:67). Ceja de selva colonization by these same populations became permanent under the direction of the Inca state at the end of the fifteenth century (Ibid.:67). Bonavia and Ravines cite the Inca's interest in controlling access to coca as the most probable motivation (Ibid. :62) behind the sixteenth century expansion of the agricultural frontier to its maximum eastern limit 102 (Ibid.:68). The collapse of supporting infrastructure with the Spanish conquest of the Inca empire purportedly precipitated the abandonment of ceja de selva colonies which were, by definition, dependent. Analyzing broad patterns of prehistoric settlement along the eastern frontier of the Andes, Bonavia and Ravines consider a list of "ceja de selva" sites within varied eastern slope environments. They group these sites geographically into southern, central and northern categories that exhibit differing characteristics (Bonavia and Ravines 1967:61). Their southern group includes sites surveyed by Isbell (1968) in the upper Inambari River valley near the Bolivian border, and sites east of Cuzco examined by Tello (1942:633), Fejes (1944) and Bingham (1930). Sites in eastern Ayacucho Department surveyed by Bonavia (1964, 1967-68, 1972b) constitute the central group, and their northern group consists of Gran Pajaten (Bonavia 1968a), Bambamarca, Tantarnayo and Chachapoyas sites. Most of these ceja de selva sites are not confined to the humid montane forest, but occupy ecological zones on both sides of montane forest timberline (Church 1994:282-283). To Bonavia and Ravines, the term ceja de selva refers to cultural as well as environmental distributions. In repeated statements, Bonavia and Ravines concur with Tosi's negative assessment concerning the montane forest's soils, their stability and their potential to sustain 103 permanent agriculture (e.g. Bonavia and Ravines 1967, 1968:153-154; Bonavia 1972a:92). They specify terracing as the critical soil conservation technique that allowed agricultural production (Bonavia and Ravines 1967:67). To account for the perceived abandonment of settlements along the eastern frontier, they cite the diminished capability of maintaining the artificially created agricultural microenvironments after the fall of Cuzco: "Following the collapse, control over the peripheral centers naturally disappeared and meant the abandonment of these places in which, without certain conditions imposed by a strong organization, it is practically impossible to subsist (Bonavia 1968b:75). Bonavia and Ravines do not directly address the possibility of indigenous forest-dwelling ceja de selva inhabitants prior to the proposed colonizations episodes. They ambiguously mention "traditional groups settled on the eastern slopes" (1968:157), yet it remains unclear if these populations represent their postulated first wave of highland colonists maintaining temporary residences or to some antecedent indigenous population. Bonavia offers his first explicit statements on the nature and antiquity of pre-Inca settlement of the ceja de selva in later publications (e.g. Bonavia 1978:398-399) in which he endorses Lathrap's Quechua expansion hypothesis. The descriptions presented below demonstrate that 104 virtually all attempts to model settlement of the eastern montane forest have featured late prehistoric (post-Chavin horizon) downslope migrations and colonization from adjacent highland Andean regions. Scholarly views of montane forest prehistory contrast sharply with developmental scenarios proposed for the Andean coast and highlands where Preceramic Period populations began clustering in and around production zones of·increasing importance to their developing agricultural economies. Preceramic Period occupations within the continuous montane forests have not been documented. Rather, incursions from adjacent areas are invariably viewed as signaling the beginning of montane forest prehistory. Before describing individual hypotheses regarding eastern montane forest colonization, some common problems affecting the investigations and the data base should be noted. Because of the logistical difficulties of working in unpopulated areas, very little survey, systematic or otherwise, has been done. The primary sources for archaeological data include sketchy reports of brief exploratory forays into the montane forest. Rarely has sustained research focused on cultural remains within the continuous montane forest, and broad generalizations often stem from surface observations at sites situated at the subalpine-montane forest ecotone. The archaeological evidence most frequently described 105 within and adjacent to the montane forest is architectural. Ruined settlement complexes often consist of stone buildings set on prominent ridge tops and surrounded by artificially terraced slopes. Surface scatters of artifacts are scarce where sites are not heavily disturbed, and the harsh climate rapidly erodes the surfaces and decorations of exposed potsherds. If nothing else, the following discussions should demonstrate that, as a consequence of the aforementioned problems, theoretical discussion of eastern slope prehistory has far out-distanced the size and quality of the data base. The Southeastern Montane Forest According to Bonavia and Ravines, the southeastern montane forest (their southern ceja de selva division) includes the upper Inambari Valley where Isbell noted two architectural styles that he attributed to Inca and Late Intermediate Period occupations (Isbell 1968:110). dating of the Inambari sites ~s His largely conjecture, as pottery and other temporally diagnostic features were rare or absent. While settlements were distributed between 4,000 and 1,800 m, agricultural terracing continued down to 800 m. Isbell's views regarding the origin and antiquity of eastern montane forest settlement (described earlier in this chapter) were published several years after this fieldwork. In the Urubamba River drainage east of Cuzco, Bonavia and Ravines observe that imperial Inca architectural styles 106 attest to the cultural affiliations and dates of occupation of sites listed for the southeastern montane forest. Most of the sites lack published descriptions, nor have they been probed for earlier subsurface components. The authors note substantial functional variability between sites, and suggest that defensive considerations account for their strategic locations. The southern pattern differs from that of the central and northern groups which show more equivocal evidence for Inca involvement in the hypothetical colonization. The Central Montane Forest Bonavia's evidence for colonization in the central "ceja de selva" was gathered during three brief field excursions during 1963, 1964 and 1970 in eastern Ayacucho (Bonavia 1964, 1967-68, 1972b). Most of the data stems from surface evidence gathered at settlements and dispersed habitations featuring circular dwellings, vast terrace complexes and occasional funerary chambers called chulpas. Bonavia encountered low quantities of eroded sherds at these sites, and even test pits yielded disappointing quantities of diagnostic material. Nevertheless, his three reports document extensive and densely occupied settlements. For example, at Caballoyuq (3,570 m) Bonavia counted 289 circular structures in the portion of the site not concealed by forest, and estimated that the settlement complex extended five kilometers down the ridge (1964:13). These 107 structures ranged between 3.90 and 7.70 min diameter. The overall distribution of settlements like Caballoyuq extends from 4,000 m above the eastern banks of the Mantaro River down through humid subalpine paramo and montane forest eastern slope environments to perhaps 1,600 m (Bonavia 1972b:23, 29). Evidence of prehistoric intensive agriculture continues down to approximately 900 m where some highlanders presently cultivate coca destined for transport and eventual sale in Huanta (Ibid.). Based upon his architectural observations and his study of approximately one ~housand sherds and a few whole and reconstructed vessels collected during the three field trips, Bonavia concluded that all of the sites reconnoitered pertained to the same people ("gente") 68:256). (Bonavia 1967- An apparent lack of architectural evidence for intra-settlement functional variation led him to suggest that these were specialized agricultural villages ("villas"). Bonavia did not locate any unequivocal evidence for chronology on any of his explorations. After his third excursion to eastern Ayacucho however, Bonavia concluded that " ... the villages and other agricultural works discovered in the area are nothing other than products of the same phenomenon: the Inca colonization of an economically important area ... " (1972b:29). The data, he argued, conformed to Murra's model of vertical control (Ibid.). Bonavia recognizes the site of Taipi in the 108 subalpine paramo, which alone displays rectilinear architecture, as the lone exception to his pan-regional generalizations. A Huari-style ceramic figurine from a grave context (Bonavia 1967-68:Lamina 10), coupled with Isbell's finds at the Huari site of Jargampata (Isbell 1971) ostensibly supported Bonavia's suggestion that Taipi dates to the Middle Horizon. Consequently, the archaeology of the region was thought to reflect the Inca emulation of practices begun during the prior Huari imperial expansion (Bonavia 1972b:26). Colonization as Huari Political Economy Not far south from where Bonavia had described eastern slope settlement between the middle and lower reaches of the Mantaro River, two of Lathrap's students had begun survey and excavations in the Pampas and Apurimac River valleys by the late 1960s. In the Pampas River valley, Isbell (1971, 1977) excavated at Jargampata, a site that he describes as strategically situated at 2,500 m elevation between puna and eastern lowland production zones, and interprets as an administrative complex serving the Huari empire. Raymond (1972, 1976, 1985, 1992) meanwhile worked in the premontane forests (500 to 1500 m) of the Apurimac River canyon below. In his study of the Apurimac premontane and subtropical forests, Raymond utilized ethnohistoric references and ethnographically documented settlement distributions as well 109 as archaeological evidence to argue for "repeated, systematic colonization of the lowland area" by highland groups during the Middle Horizon (Raymond 1976:206, 1992). Raymond begins by citing Murra's model and its premise that "control of the eastern slopes of the Andes was a critical component of the economic systems" 1992). (1976:205, He identified three "late prehistoric cultural complexes" named Sivia, Quimpiri and Simariba that he characterizes in terms of their ceramic assemblages, settlement distributions and inferred economic orientation (Raymond 1985). Ceramic assemblages were found together and in isolation, and correlations between the three assemblages and site locations suggest Amazonian riverine, premontane forest interfluve and highland-derived occupations respectively. Modern-day Panoans like the floodplain- dwelling Conibo of the Ucayali provide an analogy for Sivia complex peoples. Shallow deposits on ridge and bluff tops set back from the river contain Quimpiri complex deposits believed to represent an orientation to interfluvial resources similar to the present-day Campa's. Quimpiri sites line both sides of the river below 600 m but extend farther upstream only on the eastern side. On the western side adjacent to the highlands, the Simariba complex is found at sites often lying distant from the river between 600 and 800 m on the same valley slopes extensively colonized by highlanders in historic times. Several lines 110 of evidence suggest a history of hostile relations between peoples of the Amazonian Quimpiri and Andean Simariba complexes in competition for interfluvial agricultural land. The interspersed Quimpiri and Simariba site distributions are interpreted as evidence of fluctuating territories and buffer zones. Among the sites containing Simariba pottery, Raymond located two especially large ones (over 30 hectares} with architectural evidence for highlander occupation (Raymond 1985}. At Palestina and Vista Alegre, stone alignments may have served as footings for adobe houses and/or field partitions while other constructions are clearly rooms with walls up to two meters high. Raymond cross-dates these sites "on the basis of the few diagnostic ceramics, settlement plan and architectural features" to the Middle Horizon and hypothesizes that they served as "outposts of the Huari empire" (1985:42}. He suspects that the Simariba complex also includes Early Intermediate and Late Intermediate Period pottery (Ibid., 1976:209}. The absence of Late Horizon material he attributes to the lack of chronological control and research. The Simariba pottery illustrated is a small sample from the Granja de Sivia excavations (Raymond 1972: Fig. 53, 54a; Raymond et al. 1975: 64a-d}. Both "formal characteristics" and "stylistic affinities" are said to ally this complex to Huari ceramics in the neighboring highland Andes (1985:42, 111 1976:208, 1992:28). Raymond (1985) notes the correspondence between Simariba site locations and the ideal coca growing elevations between 600 and 900 m. Palestina and Vista Alegre he interprets as "frontier settlements built to define a territorial boundary, administrate coca production and trade with tropical forest peoples (1988:298, 1992:29). He reasoned that these sites probably linked the highland Huari empire to Panoan-speaking lowland floodplain-dwellers 20 km away at the site of Granja de Sivia, which in effect afforded Huari access to commodities circulating within lowland riverine trade networks. Raymond offers two summary positions regarding the origin and antiquity of ceja occupation: one specific to the Apurimac valley, and a second broadly generalizing. Based upon his fieldwork, Raymond believes that there is little evidence that highland utilization of the forested slopes of eastern Ayacucho pre-dates the Middle Horizon (1988:298). However, he points out that the Ayacucho site complexes examined by Bonavia remain undated (1976:210-211). The distribution of some poorly preserved settlements with "circular, oval, D-shaped and rectangular structures" of unknown cultural and temporal affiliations reportedly resembles the Simariba distribution (1992:27), but these were only cursorily examined. Raymond's fieldwork included a "survey transect" up the western side of the Apurimac River valley in which he found "no signs of prehistoric 112 settlement" between 1,500 and 3,000 m (1985:40}. Although he does not specify its exact location, his transect was probably not far south of Bonavia's (1972b} route from the middle to the lower Mantaro River courses. The lack of evidence for "settlement" contrasts with the abundant evidence described by Bonavia, and the evidence recovered by Hastings (1985} farther north in the Tarrna River valley. Raymond's broad characterizations of tropical forest settlement systems outline features of three generalized types (Raymond 1988}, apparently based upon his interpretation of the Sivia, Quimpiri and Simariba cultural complexes. In addition to Amazonian floodplain and interfluvial settlement systems, Raymond notes a third occupying the ceja that, "is basically an extension of the highland agrarian system" and, "is not self-sufficient but is tied closely to social, political and economic networks in the highlands" (Ibid.:286). He concludes that such "colonization of the steep terrain of the upper montana apparently began during the Early Intermediate Period" (1988:296), and cites Lathrap and Isbell's assessment of eastern slope prehistoric time depth. Raymond theorizes that "the expansion and contraction of highland colonies in the Montana corresponds to the rise and fall of imperial states in the highlands" {1976:211, 1992). Yet even during the post-Huari period of political fragmentation and territorial contraction, highland polities maintained a 113 colonial presence in the forests because of the continued importance of coca and lowland produce (1988:299). His discussions of eastern slope human occupations include no specific references to indigenous ceja populations. To summarize Raymond's view, the intensity of eastern slope occupation, as well as Andean-Amazonian interaction corresponds closely to highland political developments. Lathrap assumed Raymond's position by considering the fleeting emergence of social complexity interpreted at lowland Cumancaya sites as a secondary development stimulated by interaction with the Huari empire (Lathrap et al. 1985:89). Bonavia (1991:380) also subscribed unreservedly to Raymond's interpretations, and especially his dating of systematic colonization of the eastern slopes to the Middle Horizon. Finally, Raymond views the Inca presence in the ceja as imperial conquest and co-option of pre-established highland-based political and economic systems (LeMoine and Raymond 1988:125-126). Verticality Colonies in the Tar.ma Canyon Sites of the lower Tarma River valley farther north also lie within Bonavia and Ravines' central ceja de selva group, yet they are physically separated from Bonavia's and Raymond's Apurimac valley sites by the substantial mountain range dividing the Apurimac/Ene and the Tarma/Chanchamayo/Perene river watersheds. This same range 114 combines with local topographic features to create a "weak rain shadow" (Hastings 1987:147) and considerable ecological complexity in the Tarrna River valley. Hastings' work focuses on the distribution of "ceja de montana" ( 1, 500- 4,300 m) settlements and their ecological and temporal contexts in order to reconstruct economic and political systems across eastern slope "vertical" gradients. A wide transect joining adjacent portions of previously surveyed highland puna and lowland tropical forest constitutes Hastings' study area (1985:270, Map 6-1). Thus he is presented with the opportunity to place his investigations within a broad regional context, and to examine evidence for prehistoric ethnic and culture area boundaries. Hastings is clearly influenced by Murra's work as the explicit incorporation of "verticality" as a conceptual framework for his study attests (1985:4). He also attempts to interpret evidence for economic articulation between the Andean "vertical" system and neighboring Amazonian economies. Based on interpretations of survey and excavation data, Hastings reconstructs two settlement patterns corresponding to two phases of occupation in the lower Tarrna valley montane forest termed the Tranca and Paraupunta Phases. Distinctive architectural styles and statistically-derived ceramic complexes characterize each of the two patterns/phases, which fall relatively late in prehistory. His lower Tarrna valley work failed to turn up unequivocal 115 evidence of human occupation prior to the late Middle Horizon. Therefore, Hastings surmises that the eastern-most distribution of the Early Intermediate Period Huacrapuquio ceramic complex in the upper Tarma valley (down to approximately 2,700 m, yet absent below present montane forest treeline) comprised Andean civilization's eastern frontier prior to the first occupation of the lower Tarma valley (1985:553). The Tranca Phase, beginning between A.D. 800 and 900, is said to represent "the first substantial and widespread colonization of the Lower Tarma ceja de montana" or montane forest (Ibid. :693). The Tarma Phase occupation features ridgetop settlements with square and rectangular masonry buildings, many with internal dividing walls, and pottery assigned to a complex termed Malambo. The Tranca settlement pattern (defined by 33 sites) emphasizes the subalpine zone near treeline, the ridge-tops between 2000 and 2300 m protruding into the forest, and to a lesser extent the slopes below. The inordinately large site of Tranca in the subalpine zone may have served as an administrative hub overseeing economic activities in the region. Hastings suggests that the paramo emphasis may indicate the dietary importance of potatoes, while relatively small herds could be tended nearby. Lower ridgetop settlements provided access both to lowland produce and to exchange with Amazonian groups. Hastings' survey revealed an eastward "nearly continuous site distribution 116 from the culturally Andean Tranca pattern into the more Amazonian-affiliated Camonal [ceramic complex] zone of occupation" (Ibid.: 697). Hastings did not recover stratigraphic evidence to bolster his argument for the Tranca Phase's temporal priority within his two-phase sequence. A few corrected radiocarbon measurements suggest dates of A.D. 855 to 1235 for the Tranca Phase (Ibid.:682). However, Hastings admits that efforts at both relative and absolute dating of the Tranca Phase encounter substantial problems that hinder precise temporal placement. To establish a basis for the relatively early placement of Tranca, he cross-dates the associated Malambo pottery complex using comparisons with Early Intermediate Period and Middle Horizon pottery from the upper Mantaro valley. He also notes the presence of a Huari style cup or kero recovered from one site tentatively assigned to the Tranca Phase (Ibid.:792). Hastings suggests that the Camonal Complex was Malambo's contemporaneous Amazonian counterpart during the Tranca Phase. The Malambo and Camonal Complexes overlap stylistically sharing such attributes as specific rim forms and strap handles (1985:520, 589, 674), but comparisons between the two are also hampered by problems of inadequate chronological control. Hastings characterizes Camonal as an "Amazonian-Andean hybrid" (1985:589), yet he is aware that it bears no explicit resemblance to any one Amazonian style. 117 He also grants the possibility of greater antiquity for the Camonal Complex (Ibid.:678). Unfortunately, ceramic sequences in both the neighboring highlands and lowlands also suffer inadequacies in chronological precision. Hastings declines to promote a single hypothesis regarding the origins of Tranca Phase "colonists" (1985:721). He observes that the predominant pottery, Malambo Luster Ware, appears to have been locally produced. Lack of clear antecedents in the surrounding regions leaves Malambo pottery "an anomaly" (Ibid.:553). Hastings notes that apparent similarities to pottery assemblages from the Chupachu territory of the Pachitea Andina "raise intriguing questions about possibly widespread colonization north and south along the eastern Andean margins" view, (Ibid.:721). In his the possibility of these two complexes representing "variants of a single general tradition prevalent for a time on the eastern flanks of the Central Peruvian Andes" warrants investigation (Ibid.:554). Hastings in effect departs from his verticality framework to entertain the possibility of a northern progenitor for the Malambo and Pachitea Andina complexes. He observes that similarities between the two are to some degree paralleled by architectural similarities (Ibid.:652). In addition to the Tranca architectural and settlement patterns, Hastings identified another "architectural pattern" labelled Peruruhuay. It is characterized by small, 118 semi-square, semi-subterranean structures. Although the sites remain undated, ceramics are said to be "weakly diagnostic of the Malambo Complex" (Ibid. :635). The settlement distribution of the relatively small site sample approximately matches the Tranca pattern (Ibid.: Table 115). Neither the Peruruhuay nor the Tranca architectural patterns have similar highland counterparts (Ibid.:642). The San Blas ceramic complex and an architectural tradition featuring circular masonry buildings characterizes "a second period of colonization" of the lower Tarma River valley during the Late Intermediate Period and Late Horizon termed the Paraupunta Phase. Because San Blas Red-on-Buff "seems to develop earlier on the puna than in the valleys" (1985:724), and because of similar architecture on the Junin plateau, Hastings interprets the Paraupunta Phase as the result of colonization from the adjacent puna (Ibid. :727). Various lines of evidence lead Hastings to identify Paraupunta settlement with the historically documented Chinchacocha, best known as puna herders. The Paraupunta settlement pattern is similar to the Tranca pattern with one clearly dominant site (the site of Paraupunta). Less ambiguous spatial and temporal resolution facilitates Hastings' interpretation of the Paraupunta economic system in terms of Murra's framework (Ibid.:731-732; Parsons and Hastings 1988:218). More masonry structures per hectare, rather than a greater number of sites assigned to the phase, 119 ostensibly indicates population growth (Ibid.:701). Fortification and defensible locations become a greater concern. The mixing of Tranca and Paraupunta assemblages at some sites suggests to Hastings that the Paraupunta settlement and subsistence system's replacement of the Tranca system was "a gradual process" (Ibid. :727). Hastings describes Paraupunta Phase Andean and coeval Amazonian settlements as separated by as much as 1000 vertical meters (Ibid.:702), and Paraupunta sites along the Andean-Amazonian cultural interface as heavily fortified (Ibid. :730). However, San Blas sherds in Amazonian site contexts dominated by the coeval Amazonian Chanchamayo Complex, and Chanchamayo sherds in San Blas contexts demonstrate exchange relations. Close resemblances between the Chanchamayo Complex and Raymond's Quimpiri complex prompt Hastings to propose a strong association between these two styles and the historic Campa or Amuesha. The overlapping distribution of the Camonal and Chanchamayo complexes precludes interpretations assigning either to floodplain or interfluvial ecological niches. Hastings also reports that sites with remains of these ceramic complexes invariably lack stone architecture. Hastings concludes that the Tarma ceja de montana occupations, especially during the Paraupunta Phase, show a continuous distribution of occupation down the eastern slopes spanning 3,000 vertical meters. His interpretations 120 recognize a substantial degree of political and economic autonomy obtained by the lower Tarma valley subgroup, and increasing economic centralization overall, yet he considers the possibility that the Tranca and Paraupunta systems developed locally within the ceja de montana unlikely: It might, however, be misleading to perceive structural changes at this subgroup level as exemplary of cultural evolution in the ceja de montana. The emergence of such subgroup centralization in the Andean margins could be as much an externally introduced change as a locally evolved process. The lower Tarma ceja de montana is now and always has been a remote, sparsely populated area with a foul climate for permanent occupation. It should not be surprising if local communities were more often recipients-or direct outgrowths--of regionally evolving change rather than instigators of such change. Cultural developments in the Lower Tarma can only be seen as part of a regional system, and for most of prehistory these developments have probably been frontier manifestations of that system (1985:730-731). Arguing against the possibility of local evolutionary processes, Hastings asserts that the "drastic shift in settlement locations" at the beginning of the Paraupunta Phase reflects "economic changes not previously underway during the Tranca Phase" (Ibid. :731). He believes that a perceived lack of developmental continuities from one phase to the next illustrates an absence of internally generated evolution. The Northeastern Montane Forest Bonavia's brief field investigation at Gran Pajaten in 1966 provided much of the evidence for tropas de colonizaci6n agricola that figured prominently in the theoretical statements co-authored with Ravines (Bonavia and Ravines 1967; Bonavia 1968a). However, the first scientific 121 reports describing Gran Pajaten's archaeology were published by Pimentel (1967) and Rojas (1966, 1967) who directed various aspects of the government-sponsored 1965 and 1966 expeditions to the site. At 2,850 m, Gran Pajaten is situated on a ridge top above the Montecristo River, a tributary of the Abiseo River which in turn empties into the Huayabamba (also Huallabamba) and Huallaga rivers. The site is relatively small, yet it has not been completely mapped nor the total number of its constructions tallied. While only eighteen buildings appear on published maps (Pimentel 1967; Bonavia 1968a), some of the circular stone structures are elaborately embellished with cornices, stone mosaic friezes, tenoned heads, and large staircases serving elevated entryways. Based on information gathered during both expeditions, Pimentel and Rojas (a student of Tello) concluded that Gran Pajaten belongs to the Huaylas culture that, according to Tello (1942:712-713), surrounded the Marafion River basin during the third epoch dated A.D. 800 - 1321. They both recognize an Inca (Late Horizon) component at the site identified by a copper knife and a fragment of Inca-style pottery. Rojas (1967:17) evaluates their significance: These two objects are strong evidence for contact with the Inca empire. Presumably this contact was of a commercial nature, since the architectural elements lack any trace of Inca influence ... The influence of this ancient community must have been great, perhaps dominating a society which occupied much of the upper forest land. Thus Rojas envisions Gran Pajaten as a local and independent 122 development that presumably began with the arrival of Tello's "primordial" Amazonian migrants. Ultimately, it participated in Late Horizon Inca-orchestrated interaction. Although Bonavia participated in the same governmentsponsored investigation in 1966, he did not address Rojas' hypothesis. Architectural observations, surface collections and small-scale excavations in and around two of the site's most elaborate structures (Building Nos. 1 and 2} provide the basis for Bonavia's independent conclusion that Gran Pajaten pertains to the Late Horizon. Bonavia postulates that, like the eastern Ayacucho sites, Gran Pajaten represents a process of eastward expansion directed by the Inca. His theoretical statements emphasize evidence gleaned from a sample of 1,645 sherds, 400 of which derived from a controlled 2 x 2 m excavation within Building No. 1. These sherds were classified into three types labelled A, B and C. Type A predominates and consists of a single locallyproduced coarse ware featuring two distinct decorative modes. To the local decorative style based on applique Bonavia gave the name "Abiseo," while he referred to a polychrome-painted decoration as "Inca-derived." Necked- jars with flaring rims and shallow bowls dominate the inventory of Type A vessel shapes which Bonavia classified according to categories established for Inca pottery of Cuzco by Pardo (1957}. A coarse ware distinguished from Type A chiefly by paste characteristics, Type B contains 14 123 sherds. Three sherds of fine kaolin ware pottery constitute Type C. Neither Type B nor C figures in Bonavia's interpretations. Based upon his ceramic analysis, Bonavia concluded that the two Type A stylistic modes represent the co-existence of two cultures at Gran Pajaten {Bonavia 1968a:70, 157-158). The Abiseo style "corresponds to a traditional style, linked with the highlands and that represents a local colonizing culture" (Ibid. :74). The Inca-derived style "would be the style of prestige that would indicate the Incaic imposition" (Ibid.). Bonavia regarded the Late Horizon date for the settlement as clearly demonstrated by the Cuzco-related vessel shapes and by the "Inca" polychrome decoration on sherds (Ibid.:107, Lamina 12). Bonavia recognized that evidence for various phases of occupation might be revealed by future investigations at Gran Pajaten (Ibid.:66). Building No.2 in particular was believed to show evidence of refurbishment by the haphazard use of carved, decorated sandstone slabs that presumably carne from older, dismantled structures. As a "working hypothesis", Bonavia suggested that the sandstone slabs which served as media for mostly curvilinear motifs, like the Abiseo style pottery decoration, represented a prior site occupation by people of a local highland cultural tradition (Ibid.:73-74). The rectilinear slate mosaic decoration may have represented the imposition of the "new 124 norm," manifest also in the presence of the Inca-derived pottery style. In effect, Bonavia was suggesting the possibility of pre-Inca construction at the site by a local "colonizing" highland culture. In a recent publication, Bonavia (1991:527) dates colonization of the northeastern montane forest to the beginning of the Late Intermediate Period. Bonavia's monograph on Gran Pajaten's archaeology generated critiques by two reviewers (Ravines 1967-68:335336; Isbell 1970:237) who questioned his methodology and his conclusions. Ravines noted that both the settlement layout and the ceramic sample showed potential for the differentiation of more than one temporal unit. However, Bonavia's analytical emphasis on technological aspects of the pottery had rendered his study insensitive to the detection of chronologically significant attributes. Ravines' criticism was echoed by Isbell who also complained that Bonavia's use of Pardo's vessel shape classification masked insights on ethnic groupings as well as temporal developments intimated by changing vessel shape preferences within Bonavia's stratigraphy. Isbell (1974:139-140) his aforementioned Quechua expansion hypothesis as an alternative to Bonavia and Ravines' (1967) postulated Late Horizon west to east expansion. Bonavia's interpretations of Gran Pajaten satisfied other scholars. An excerpt from Moseley's (1992:190) recent 125 synthesis of Peruvian prehistory shows support for the colonization interpretation: Refinement of high montane adaptations facilitated a northward thrust into the Ecuadorian sphere of tropicalinfluenced sierra populations, known as Kuelap. These colonized mouncain settings were in a sense more tropical than Andean. Remarks to a journalist by Meggers implicitly question the need for further research: A full-scale investigation will be interesting, but of predominantly local significance ... There is no indication that the culture at Gran Pajaten spread to other areas. The inhabitants of the city were probably recipients of cultural innovations from Inca centers in southern Peru, not originators (Science News 1985: 117). The views of Moseley and Meggers clearly reflect their respective "ruling frameworks" that dictate responses characterizing montane forest sites like Cuelap and Gran Pajaten as highland-derived implants. Alternative Viewpoints The interpretations of some archaeologists who have worked in the northeastern montane forest Utcubarnba Valley over recent years coincide with Langlois' suggestion of far northern linkages, although they argue unequivocally for autochthonous origins and in situ development (Ruiz 1972; Shady 1987a, 1987b) . Based upon analysis of pottery remains excavated from Cuelap, Ruiz emphasizes similarities between Cuelap's earliest styles and El Salado Phase styles from lowland Bagua, Garbanzal pottery from far north coastal Peru and Guangala pottery from southwestern Ecuador (1972:181). Ruiz and Shady work under the assumption that settlements 126 between the Huallaga and Marafion Rivers like Gran Pajaten and Cuelap "give evidence for singular and distinctive Central Andean cultural expressions in their urban and architectural design as much as in their diverse known manifestations; in no way can they be explained as derived from colonizing advances from the nearby highlands" (Shady 1987a:86). Also in reference to the northeastern montane forest, Lumbreras observes that: "this zone as a whole exhibits closer relationships with the little-known northern Andean area than with the Central Andes ... " (1974:149). He characterizes this area as the "focus of high-level social development ... where peoples flourished with a high level of development that now we measure partially with architectural accomplishments as important as those of so-called Pajaten in the northeast of PerU with its neighbors and relatives Cuelap and others in Chachapoyas ... " (Lumbreras 1981:31). Although Lumbreras does not directly address the issue of human occupation's time depth in the montane forest, he clearly envisions in situ development. In a more recent publication, Morales (1993:646-653) assumes an even more extreme posture favoring autochthonous cultural development within the greater northern and eastern montane forests. By labeling this region Amazonia Andina, or "Andean Amazonia," Morales implicitly contradicts Kauffmann's highland-centric interpretations of the Andes 127 Amaz6nicos. In Morales' view, ancient montane forest settlement cannot be comprehended utilizing Andean or Amazonian perspectives. He states that not only do scenarios eliciting conquest and/or colonization impose inappropriate analytical frameworks, but that applications of Andean periodization schemes presupposing Huari and Inca hegemony are spurious. Rather, Amazonia Andina societies engaged in interregional interaction independently and irrespective of political developments in the neighboring highlands since the late Preceramic Period. Cuelap, he regards as the fortified center of an upper Utcubamba Valley state-level polity (Ibid.:652). During later prehistory, encroaching Andean empires were forced to adapt to local montane forest social, economic and political conditions. Alternative views of local development are typically expressed in unambiguous terms, yet provocative assertions are seldom buttressed by more than perfunctory references to monumental architecture at Gran Pajaten or Cuelap. Not only is the montane forest data base rarely addressed in systematic fashion, but archaeologists rarely address each other's hypotheses. Langlois' hypothesis of northern origins has been ignored by all but the European scholars. Bonavia (1978, 1991) and Hastings (1985) refer to Lathrap's Quechua-expansion hypothesis despite its replacement by Isbell's, and Tello's postulated migrations from the northeast is seldom cited by anyone concerned with montane 128 forest settlement. As a result, potentially viable yet conflicting hypotheses for montane forest settlement have tended to accumulate, rather than to supersede one another. This chapter has summarized numerous distinct hypotheses that conceive of the prehistoric montane forest as an empty migratory corridor, a migratory cul-de-sac, a sparsely populated buffer zone, a remote ecological zone servicing highland polities and an archaeological culture area. Isbell's, Moseley's and Morales' interpretations could scarcely be more at odds with one another. The following chapters will present documentary and archaeological data confirming the presence of autochthonous and autonomous polities in the northeastern montane forest, and evaluate each of the aforementioned hypotheses in light of the new evidence. CHAPTER 3 THE PATAZ-ABISEO STUDY AREA: PRESENT AND PAST The Chachapoyans, whom I thought to be the builders of the [Gran Pajaten] ruins, promised to be an exciting new culture of Amazonia. Could they have been indigenous or had they come from the eastern lower jungles? Did they come from the northern Andes to the west? Only further exploration and study could answer these questions. The site appeared to be a ceremonial center or important temple building of a civilization far more advanced than that of the builders of Vileabamba the Old. The circular buildings, more difficult to construct than rectangular structures, were vastly superior to anything I had seen in the montanas of southern Peru. If the ruins did belong to the Chachapoyas civilization, then they certainly had a culture far higher than that of the Incas during the same period. (Gene Savoy 1970:143) The previous two chapters outlined the hypothetical prehistoric migrations most frequently cited as salient developmental events in eastern montane forest prehistory. This chapter introduces the study area and provides background information on present and past cultural geography and economic activities. The study area straddles the high cordillera separating the upper Marafion and central Huallaga river watersheds (Fig. 4), and includes northeastern montane forest life zones. Both rivers originate in the snow-capped peaks of Huanuco and Pasco departments and flow north-northeast roughly parallel to one another until the Huallaga swings east and exits the Andean 129 130 foothills near 7o south latitude. eastward farther north at so The Marafion veers south latitude. More precisely, this study will focus upon the narrow strip of intervening cordillera between 7c and go 30' south latitude, just south of where the two rivers diverge. At these latitudes, the forested foothills of the Andes flank the narrow central Huallaga floodplain, but the entrenched Marafion does not emerge from its rocky inter-Andean canyon until it approaches its confluence with the Utcubamba River much farther north. Scholars usually conceive of societies above the eastern banks of the Marafion, and on the crest of the Marafion-Huallaga divide, as "Andean" (e.g. Rowe 1946:Map 3), while those along the west banks of the Huallaga are regarded as "Amazonian" (e.g. Steward 1948:Map 5). However, little is truly known of ancient population distributions in the area, especially within the unexplored montane forests of the Huallaga valley. The immensity of the Marafion Canyon is truly aweinspiring. It is over 3,000 m deep from the summits of the Marafion-Huallaga divide (4,500 m) to the valley bottom (1,200 m), and approximately 40 km wide. Less than 20 km separate the course of the Marafion and the top of the divide, rendering an average gradient between 15 and 20 percent for the Marafion valley slopes. Along this stretch the Marafion is not a reliable means of transportation as it is navigable by canoe or balsa raft for short distances only 131 during the brief dry season. Its gradient within the lower canyon is three percent (Pefiaherrera del Aguila 1986:131). Below the mouth of the Utcubamba, the Marafion enters a 150 km stretch of cataracts (or pongos), punctuated at beginning and end by the great pongos of Rentema and Manseriche. None of the Marafion's east-bank tributaries south of the Utcubamba are navigable. At the same latitudes, the Huallaga River lies much lower at approximately 350 m. Its banks are 70 to 80 km distant from the Marafion-Huallaga divide. While the gradient of the Huallaga valley's slopes averages around five percent, most of the vertical drop occurs abruptly in the precipitous upper montane forests. The river descends a relatively mild gradient of approximately one percent (Ibid. :143). Historically, canoes and balsas plied the central Huallaga below Tingo Maria (elevation 650 rn) Casas 1930, 1934; Pefiaherrera del Aguila 1987). {de las From Tingo Maria down to the mouth of the western tributary Huallabamba River (also Huayabamba), riverine traffic contends with numerous rapids including the pongos of Cayumba, Sabaloyaco and Cachihuafiusca (Raimondi 1876:423). Below the Huallabamba, the widening river has accumulated a floodplain up to 5 km wide {DeBoer 1984), and traffic flows relatively unimpeded until reaching the next series of whitewater obstacles below Shapajo {near Tarapoto) . Below Chazuta, the river passes into the lowlands via the great pongo of 132 Aguirre. Some small pongos along the central Huallaga have been "pacified" with dynamite in historic times (de las Casas 1930:182-183). Many of the west-bank tributaries of the Huallaga south of 7c latitude can be navigated for short distances on their lower reaches. However, Weberbauer (1945:98) characterizes the largest of these, the Huallabarnba, as "the axis of a great fluvial system." The Huallabarnba River may be ascended in canoe at least 60 km to the abandoned mission at Jesus de Pajaten (Savoy 1970), and perhaps up to 100 kms (Torres Calderon 1903:296). include, from north to south, The Huallabamba's tributaries the Huarnbo, Jelache, Huabayacu, Bombonaje, Jepelache, Pajaten and Abiseo (or Apisoncho) rivers. Since the beginning of this century, geographers and engineers have recognized the Huallabamba as an extraordinary westward intrusion of lowland waterway that, if reached by road from the Andes, would most efficiently establish a commercial artery connecting the economies of Andean northwestern Peru and the Amazon River port city of Iquitos (Torres Calderon 1903; Weberbauer 1920; de las Casas 1934). Despite such early interest, the Huallabamba drainage remains lightly occupied and virtually unexplored by outsiders. Politically, the study area corresponds to the districts of Pataz (La Libertad Department) on the west side of the divide, and to Huicungo District (San Martin 133 Department) on the east side. Manachaqui Cave lies at Pataz District's northern border with Condormarca district. Since 1983, the Abiseo river drainage in Huicungo District has been administered as the Rio Abiseo National Park (RANP) . Together the Pataz and Huicungo districts (including the RANP) encompass a west to east transect beginning at the bottom of the Marafion canyon, spanning the cordillera and ending in the Huallabamba valley lowlands. This wide transect will be referred to as the "Pataz-Abiseo area" (Fig. 5). The eastern addition of Pachiza District on the banks of the Huallaga completes this transect. The borders of these modern political-geographic units are not expected to coincide with past cultural or ethnic distributions. Studies of modern, historic and prehistoric highland Andean economies rely on classification schemes to differentiate ecological zones that served to provide desired resources. I begin this chapter by detailing a scheme of biogeographic zonation based on botanical inventories compiled in the Pataz-Abiseo area, and I describe modern land-use within each zone. Analogies derived from modern ethnographic sources can be indispensable for archaeological interpretation as long as their utility is evaluated rather than assumed. Ethnohistoric evidence presented next portrays rapidly changing demographic and interregional economic conditions during the Colonial Period. A third section marshals 134 archaeological evidence for pre-Hispanic settlement distributions and boundaries in the Pataz-Abiseo area, and relates them to their biogeographic (ecological) contexts. While this chapter provides basic background information regarding the study area environment and socalled Chachapoyas culture, it also emphasizes change and the need for an informed historical perspective in order to approach problems of prehistoric economic activities in the area. As the first section on modern land-use demonstrates, the study area is relatively isolated from the mainstream Peruvian national economy. Readers informed about Peru's modern political and economic geography may regard the selection of this area for studies of prehistoric highlandlowland relations as misguided. In recent years, opponents of llregionalizaci6nll and the proposed administrative union of II Andean II La Libertad and II Amazonian II San Martin departments point to the complete lack of economic articulation across the natural barriers separating two worlds with disparate natural, social, economic and political histories. At first glance, there is no reason to suspect that conditions have ever been otherwise. However, a review of early Colonial Period documents demonstrates that the study area's present economic conditions are not indicative of those of its past. fact, In the Pataz-Abiseo area was a locus of particularly intense highland-lowland exchange, and several historical 135 processes and events that contributed to its present "marginal" economic status can be identified. By contrasting present and Colonial Period information on cultural geography with what is known archaeologically of prehistoric settlement distribution, we may reconstruct major economic and demographic transformations. The early Colonial Period stands out as a time of rapid, even chaotic, change although thin strands of economic ties survived only to dissolve slowly but completely by the early twentieth century. The following narrative will demonstrate how modern ethnographic observations may fail to illuminate prehistoric conditions, and thereby illustrate the weakness of archaeological studies that rely uncritically upon ethnographic analogies. Present economic articulations {or lack thereof) between Andean highland and Amazonian lowland regions along the length of the eastern slopes cannot be mechanically projected into the prehistoric past. The widespread idea among archaeologists of a highland-lowland dichotomy reflects the current national social, political and economic reality of Peru. A brief summary will highlight points interpreted as crucial for understanding Pataz-Abiseo history and prehistory, and the dynamics of long-term regional economic transformation. 136 Life Zones and Land-Use in the Pataz-Abiseo Area Natural and anthropogenic fragmentation of the montane forest renders the Marafion-Huallaga divide an ecologically complex orographic region. From south to north summit elevations diminish, and fragmented montane forest covers increasing percentages of the valley slopes. The east-west distribution of ecological zones found between the headwaters of the Rio Utcubamba and the modern town of Huancaspata (including the study area) conforms to Brush's "compressed" type of zonation (Brush 1977) . The repetitive sequence of dry Marafion-side to moist Huallaga-side life zones is most notably broken where the alpine grasslands broaden north of Bolivar, and south where the intermontane valley created by the Cajas River partitions the cordillera near Tayabamba. Unfortunately, our understanding of eastern slope environments is hindered by the lack of field studies and meteorological stations which would normally provide useful information for evaluating the agricultural potential of given life zones. It is my intention in this study to clearly distinguish between kinds of site environments, yet recognize that the ecotone separating the highland grasslands from the montane forests is wide and complex, and that it shifted over time. Andeanists studying historic and prehistoric regional economies on the eastern slopes frequently apply taxonomies derived from folk tradition (e.g. Pulgar Vidal 1987), 137 western biological sciences (e.g. Holdridge 1967) or hybrid combinations of these to differentiate study area ecological zones (Mayer 1985:49-50; Church 1994:282). In the Pataz- Abiseo area, information regarding pre-Hispanic production zones cannot be obtained through modern land-use analogies because all but the western-most portion of the area is virtually uninhabited today. Furthermore, an ethnographically-derived zonation scheme would be of questionable validity for dealing with past situations in which unknown socioeconomic systems organized production of unknown crops. Because it does not impose biases on interpretation of pre-Hispanic land-use patterns, Young's (1992b, 1993) scheme developed for the Pataz-Abiseo area utilizing the Holdridge system provides the foundation for archaeological investigations (Church 1994). Young's (1990) classification of ecological zones encountered in the west to east transect between the Marafion and Huallaga Rivers was developed in part to examine issues of conservation. His categories have heuristic value for biological inventories and biogeographic taxonomy, but they do not necessarily coincide with pre-Hispanic eastern slope production zones utilized for potatoes, maize and coca. Exactly how the extreme environmental conditions within the continuous montane forest of the central Huallaga watershed affected the distribution of pre-Hispanic production zones requires further investigation. Young observed, inventoried 138 and collected specimens of plant species encountered within Pataz District life zones, and from the upper reaches of the Rio Abiseo National Park from 4,200 m down to approximately 2,300 m (Leon et al. 1992; Young 1990, 1992b, 1993; Young and Leon 1988, 1990). Therefore, his descriptions cover the ecological zones from which the archaeological data pertinent to this study derive. Within the Pataz-Abiseo area, Young classifies environments into four "macro-ecological zones" 6) (Figs. 5 and loosely corresponding to the folk taxa used locally: temple, Quechua, jalca and puerta del monte (Young 1993). From west to east these are the Dry Forest, Moist Montane, Tropical Alpine and Montane Rain Forest zones (Ibid. :271). For each of the four macro-ecological zones, Young provides inventories of native plant species and common crops, as well as a brief summary of land-use practices and the agricultural calendar for Pataz. Published descriptions are not yet available for each of the montane forest life zones that Young has identified (Young, personal communication). Also, the lower montane forest and premontane forests of the Abiseo drainage remain mostly unexplored by western scientists. Virtually all of the following biogeographic information specific to the Pataz-Abiseo area has been distilled from Young's doctoral thesis and recent publication of Pataz's ecological zonation (Young 1990, 1993). 139 The Dry Forest Zone At the western extreme of the Pataz-Abiseo area, the Marafion valley rain shadow produces the Dry Forest macroecological zone (1200 - 2300 m) . The Dry Forest zone corresponds to Holdridge Tropical Thorn Steppe and Tropical Dry Forest life zones. year. Rainfall averages only 500 mm per According to Young, human exploitation has eliminated the original character of the forest, leaving cactus and scrub vegetation to further proliferate on the steep slopes of the Marafion canyon. Land-use patterns today probably bear little resemblance to those in prehistory because of the introduction of sheep and goat grazing, and industrial gold and silver mining. Ecological degradation has eliminated wild populations of large mammals. Only five percent of the land area is suitable for cultivation, and irrigation is necessary to grow maize, avocados, onions, tomatoes, limes, pineapples, bananas, passion fruit, guavas, cherries, mangos, oranges, papayas, sweet potatoes, manioc, Trujillo coca, sugar cane as well as assorted crops of minor economic importance (Young 1993:271, Table 2). Informal placer mining by native villagers on the shores of the Maranon yields small amounts of gold. Residents of the Dry Forest Zone sell products such as fruit to passing truck drivers, and coca to highland villagers. The Moist Montane Zone Travelling eastward and upslope one encounters the bulk 140 of the human population between 2,300 and 3,400 meters in the Moist Montane macro-ecological zone. Young notes that all of the staple crops except for coca are cultivated between 2,400 and 3,400 meters (Young 1990). The town of Pataz (incorporated in 1770) is situated on the southern slopes of the Frances River valley for easy access to mineral resources at 2,700 m. Village residents still utilize shafts abandoned by industrial mining interests in the 1940s. Pataz interacts commercially with Trujillo where gold is exchanged for cash. Agricultural produce from the highlands surrounding Huamachuco makes up the deficit of foodstuffs not obtainable from the cultivated slopes above and below Pataz. Pataz and adjacent districts have long been recognized as among Peru's richest in mineral resources (Tarnawiecki 1927). Above town, miners have followed veins bearing gold and silver sulfides within a quartz and pyrite matrix deep into the mountainside. Modern mining activities have traditionally centered on the mountainsides from Zarumilla south to Buldibuyo. Mines located between 3,500 and 4,000 m above Parcoy and Llacuabamba have produced extremely high yields of over 100 grams of gold and silver per ton of processed mineral (Ibid. :182). According to Tarnawiecki (1927:178) prehistoric and early Colonial Period mining activities took advantage of gold and silver oxides (placer deposits) occurring 40 to 60 meters below primary exposures 141 on the lower valley slopes of the Moist Montane and Dry Forest zones ( "sitios de climas templados") . Los Alisos and other agricultural hamlets above Pataz, like the self-sufficient towns of Condorrnarca and Uchucmarca farther north, occupy locations at approximately 3,000 m for efficient access to principal crop production zones (Brush 1977:82-83). The Moist Montane Zone receives 750 - 1000 mm of precipitation annually, most between November and May. Native vegetation has been completely altered by human exploitation, and now consists of shrubs and small trees, and forest remnants in gullies, ravines and along hedgerows. Only approximately 10 percent of the steep terrain can be cultivated. Common crops include maize, beans (P. vulgaris), broad beans (V. faba), squash (Cucurbita maxima), peppers (Capsicum annum), wheat, alfalfa, barley, tarwi, peas, potatoes, olluco, oca, agave, chirimoya, prickly pear, figs, apples, peaches, elderberries, cabbage, carrots, cayenne and oregano. From Pataz, the Marafion can be reached after a four to five-hour walk west, and the edge of the continuous eastern montane forest in the Abiseo River watershed lies a long day's walk over the cordillera. Above Pataz between 3,300 and 3,600 m lies a discontinuous belt of dense closed forest which is usually fog-bound during the rainy season (Young 1993:272. Table 4). It is a habitat for deer and provides a source of firewood. 142 The Tropical Alpine Zone Young describes the Tropical Alpine macro-ecological zone, occurring above 3,600 m on the western side of the Marafion-Huallaga divide, as divided equally between rocky peaks and slopes, and wide U-shaped valleys. Tussock grasslands characterize the valley floors with sedges and shrubs thickly covering areas of poor drainage. The western side of the cordillera (Tropical Subalpine Wet Paramo life zone) receives 1,500 mm of precipitation, but is relatively dry between May and September. The eastern side (Tropical Subalpine Rain Paramo life zone) receives over 2,000 mm of rain and experiences fog almost daily. Frost, sleet and snowstorms are not uncommon at the peak of the dry season. Highland villagers utilize the Tropical Alpine zone for cattle grazing today. Cattle owners typically maintain rustic huts, tend a potato patch and hunt waterfowl, deer and predators such as Andean foxes, spectacled bears and mountain lions during periodic visits to the area. Within the past 25 years, economic activities in the Tropical Alpine zone have been altered by the reduction of cattle-herding resulting from 1970s agrarian land reform, and by the introduction of trout which are caught with a baited hook or by hand. The voracious trout eliminated much native aquatic life prior to identification by specialists (Leo 1992), although there were probably few native vertebrate species. Economic activities on the eastern side 143 of the divide were again transformed in 1983 by the creation of the Rio Abiseo National Park and new prohibitions of traditional activities such as cattle grazing and hunting. A state-sponsored agricultural program recently introduced a herd of alpacas tended by the Comunidad Campesina of Los Alisos. Long range plans include harvesting and exporting the alpaca wool as a cash-generating industry. The Tropical Montane Rain Forest Zone Scientific studies in the forested Abiseo River drainage have long been hindered by the absence of pertinent and reliable cartographic information. Modern map coverage by the Institute Geografico Nacional (IGN) cartographic (1:100,000 series) does not include the eastern montane forests of San Martin Department. Comparison of recent LANDSAT photos to the 1:500,000 departmental map (1985) reveals that the latter errs in its location of several rivers. Aerial photographs have proved useful for ground surveys, but coverage by the AF-60-17 series (1962-63) currently available from the IGN does not extend much below the upper montane forest. These mapping problems await rectification by the Peruvian government. Although technically belonging to the Tropical Subalpine Rain Paramo life zone, Young and Leon's (1988) "fragmented" montane forest constitutes the ecological transition between 3,700 and 3,400 minto "continuous" 144 montane forest. Young utilizes the term "timberline forest" to indicate both the fragmented forest and the uppermost strip of continuous forest. Forest patches occur on the slopes of the U-shaped valleys, especially under cliffs, around boulders, within gullies and surrounding standing water. In order to catalogue vegetation, Young and Leon (1990) divide the continuous montane forest into approximations of the lower (1,500 - 2,500 m) and upper montane forest (2,500 - 3,500 m) zones described earlier in this chapter. The lower montane forest is less important for the purposes of this study because it remains archaeologically unknown. Young identifies two principal Holdridge life zones within the upper montane forest. These are the Tropical Montane Wet Forest life zone (2,500 - 2,900 m) and the Tropical Montane Rain Forest life zone (2,900 - 3,400 m). He believes that generally low temperatures and frost, low light levels, acidic soils with high aluminum levels, and the very high humidity are factors which individually or in combination discourage crop cultivation in the Tropical Montane Rain Forest life zone (Young, personal communication 1992). Hastings (1985:70, 73-74) likewise reports that the "cloud forest sub-zone" at elevations between 2,500 and 3,300 m in the Tarma River canyon has only minimal economic importance. Rather than indicating that agriculture never extended into this zone, these observations suggest that 145 agricultural production focused on more favorable higher and lower elevations. The boundaries, and even the presence and absence of the Tropical Montane Rain Forest and other life zones should be expected to vary regionally and temporally. A higher forest canopy allows easier travel within the Tropical Montane Wet Forest life zone. drainage, Within the Abiseo this zone contains abundant evidence for dense prehistoric settlement and economic utilization that will be outlined later in this chapter. Bonavia noted maize thriving during his visit to Gran Pajaten, and in 1986 University of Colorado archaeologists found potatoes that locals had planted in their 1985 excavation backfill. While the lower montane forest remains virtually unexplored, it has attracted small groups of highland agriculturalists that have recently colonized Jucusbamba (2,000 m) on the Abiseo (UC and FPCN 1991:34; Young et al. 1994). Agricultural systems of the lower montane forest remain undescribed. Equally little is known of the premontane forest life zones between 500 and 1,500 m, except that similar environments north and south of the RANP currently produce much of the world's coca crop. Many non-agricultural resources may lie concealed beneath the montane forest vegetation. During his attempt to reach the Huallaga from Pataz in 1919, Weberbauer (1920:8) complained that some residents of Pias jealously withheld information regarding the upper forest location of 146 a source of salt. According to Tarnawiecki (1927:177}, the gold and silver-bearing deposits mined above the Marafion continue into the "montaiia," yet their extent can only be crudely estimated. De las Casas claims that mineral deposits at the abandoned settlement of Achiras (1,000 m} in the Abiseo valley premontane forest yielded high gold content (1935b:342}. Some western affluents of the Huallaga also contain placer deposits (de las Casas 1930:173}. The Huallabamba and Hua1laga Valley Premontane Forests De las Casas documented the predominant agricultural activities in the premontane and subtropical forests of the central Huallaga valley lowlands in a series of monographs promoting the region's economic potential (de las Casas 1930, 1932, 1933, 1935a, 1935b}. He details a wide range of cultivated produce and plant resources including beans, potatoes, maize, manioc, assorted fruits, cotton, coca, vegetable dyes and medicinal plants (de las Casas 1932}. The town of Pachiza was founded above the banks of the Huallabamba by Franciscan Padre Sobreviela in 1790 as a place for canoe travelers to rest and re-supply during their descent to the Huallaga from the now-abandoned mission of Jesus de Pajaten (Unanue and Sobreviela 1963 [1791] :154}. How far canoes may navigate up the Abiseo river system within the study area proper remains unknown. At the time of de las Casas' report (1935b:346}, Pachiza villagers 147 exported agricultural staples including livestock via the Huallaga to Iquitos in canoes and balsas. Numerous sources of salt are available along the course of the central Huallaga. These include several near Tocache, and a source named Cachihuafiunca near El Valle (de las Casas 1930:179). A source producing salt that circulated widely in pre-Hispanic and Colonial Period Amazonian trade lies farther down the Huallaga at Pilluana and Chazuta (DeBoer 1984; Reeve 1994). Rubber was of substantial economic importance at the time of de las Casas' study, but DeBoer noted that coca has now become the area's principal cash crop (DeBoer 1984). The social and political problems that have accompanied narcotics trafficking presently render this region virtually inaccessible for scientific study by outsiders. Future investigations of east slope zonation and human land-use within the study area might include matching life zones and crop requirements, thereby suggesting a range of hypothetical prehistoric crop production zones optimal for potato, maize, manioc and coca under present climatic conditions. Such an effort would be severely hindered by the lack of information pertinent to soil conditions as well as prehistoric crop varieties and agricultural techniques. Of primary interest is the relationship between the vertical distribution of potential production zones on the eastern slopes and the cultural boundary separating the 148 monument-building, nucleated montane forest groups known to have once occupied the study area (Lennon et al. 1989) from the upper Amazonian groups dwelling in the Huallaga valley bottom. Ethnohistory of the Pataz-Abiseo Area The following discussions utilize the available documentary information to explore how late prehistoric and Colonial Period populations were distributed over this landscape. According to Espinoza's (1967) ethnohistoric analysis of Chachapoyas, prehistoric populations on the Marafion-Huallaga divide were organized in numerous autonomous political units. Utilizing Colonial Period administrative and ecclesiastical records, it can be demonstrated that some of these polities interacted habitually with neighboring Amazonian groups of the central Huallaga River valley. The latter may be thought of as border intermediaries (Reeve 1994) connecting Amazonian and Andean interaction spheres. The documentary evidence also reveals that the early Colonial Period was a time of demographic chaos in both highlands and lowlands. Sudden depopulation and demographic disintegration contributed to the breakdown of traditional communication linkages, and ultimately led to the complete disarticulation of the two areas by the mid-twentieth century. This decline was merely one aspect of the pan- regional economic collapse that accompanied the Spanish 149 imposition of new political and economic orders. The descriptions and interpretations to follow first focus on the highland populations occupying the crest and upper slopes of the Marafion-Huallaga divide, and then on tropical forest groups occupying the lower slopes of the Huallaga valley. Southern Chachapoyas The Pataz-Abiseo area lies within the southern half of the Inca province of Chachapoyas (Cieza de Leon 1976 [1553]; Garcilaso de la Vega 1966 [1609]; Rowe 1946; Schjellerup 1984). According to Espinoza's ethnohistoric analyses of several surviving documents (1967:333), so-called Chacha or Chachapoyas ethnic groups were distributed north to south between Levanto near the modern city of Chachapoyas and Huancaspata at the modern departmental border of La Libertad with Huanuco. The Marafion canyon separated Chachapoyas from groups subsumed within the Inca provinces of Cajamarca, Huamachuco and Conchucos to the west. Espinoza draws an eastern boundary incorporating the upper montane forest. Inca provincial boundaries have traditionally provided the principal basis for determining the spatial extent of Chachapoyas ethnic groups. While accepting Espinoza's frontiers for the purposes of discussion, it should be remembered that the Inca provinces were political constructs imposed upon population sets for the purpose of efficient administrative control. Local ethnic boundaries were 150 disregarded where they were considered irrelevant to imperial interests (Zevallos 1987). The term "southern Chachapoyas" will be used to refer to the Spanish-administered province (or corregimiento) of Cajamarquilla. The Spanish sub-divided Cajamarquilla into five repartimientos (or encomiendas) listed as Leimebamba y Cochabamba, Chilchos y Laya, Cajamarquilla y Condormarca, Sucos y Puymal and Collay from north to south. The three southernmost encomiendas are of primary interest to this study. Ethnohistoric research in southern Chachapoyas suffers from the failure to locate the pertinent Toledo era Visita compiled between 1572 and 1574 by Diego Alvarez (Espinoza 1967:225). Thus we lack the kind of early detailed census information utilized to examine demographic and organizational problems in other regions like Huanuco. Also, the synchronic ethnographic profile offered below should be regarded as specific only to Chachapoyas on the eve of Inca conquest. Chachapoyas Culture Documentary evidence confirms archaeological reports regarding Chachapoyas settlement patterns. Chachapoyas natives utilized all four of Young's macro-ecological zones and the bulk of the population centered on the upper slopes and high promontories near the interface of the Moist Montane and Tropical Alpine zones around 3,500 m (Brush 1977:82-83). In this respect, Chachapoyas settlement 151 patterns were little different from coeval patterns in other Central Andean regions where defense and efficient access to both tuber and pastoral production zones were primary considerations (Ibid.:83). Chiefly interested in maximizing control over indigenous labor resources, both the Incas and Spaniards centered their administrative activities at highland Chachapoyas settlements. Inca-installed mitimaes from the coast inhabited the Dry Forest Zone near Balsas (Espinoza 1967:230), and both Thompson (1976:99) and Curtin (1951:60) report archaeological evidence of occupation (albeit undated) at Pusac and Matibamba respectively. Chachapoyas occupation of the Tropical Montane Rain Forest Zone will be addressed later within this chapter. Current conceptions of Andean commodity movement across a "vertical" landscape lead us to expect a linear west-east distribution of intra-ethnic exchange partners in the PatazAbiseo area. Even today, Bolivar (Cajamarquilla) interacts commercially with Cajamarca, just as Pataz communicates almost exclusively with Huamachuco and Trujillo. Yet considering perceived similarities between archaeological assemblages distributed north-south along the eastern slopes of the Central Andes (e.g. Lathrap 1970; Isbell 1974; Hastings 1985:554) the longitudinal orientation of shared Chachapoyas cultural characteristics is significant. While topography may dictate the elongated shape of Chachapoyas to some degree (see Rowe 1946:185, Map 3), continuous pre-Inca 152 interaction along a north-south axis is also a likely factor. Zavallos (1987) observes that "there should have been powerful reasons for the Lima government to consider the zone [Pataz and Cajamarquilla] as part of the polar jurisdiction of the city of Chachapoyas since the 16th century." The fact that Chachapoyas' continued to exercise jurisdiction over Pataz until well into the 19th century is probably an artifact of ancient north-south interrelations. Although some have hypothesized state-level political integration for pre-Inca Chachapoyas (Ravines 1972:218; Brush 1977:43-44; Morales 1993:652), most analyses of ethnohistorical and archaeological evidence have concluded that Chachapoyas populations had never been politically unified prior to Inca conquest (Langlois 1939a:229; Espinoza 1967:233-234; Zevallos 1987; Schjellerup 1990:237; Lerche 1995). Regarding socio-political organization, Espinoza (1967:233-234) argues that the primary unit was the ayllu led by a chief (or curaca) and a council of elders. Although each ayllu was independent, hostile threats prompted the creation of temporary alliances. Each ayllu maintained a principle village with secondary hamlets comprising a small-scale village hierarchy which applied the name of the ayllu to the entire group. Espinoza reports that ayllu leadership was traditionally inherited, and that ethnic unity throughout Chachapoyas was expressed in a common principal deity named Curichaculla, as well as a 153 common language, dance and music. Chachapoyas settlements typically consist of clusters of circular stone buildings between six and 15 m in diameter, arrayed on hill or ridge tops in a seemingly random fashion. family. Each building likely housed a nuclear Some ruined settlements such as Cuelap are famed for their elaborate systems of walled fortification. A few structures bear stonework mosaic friezes similar to, yet less elaborate than, those reported at Gran Pajaten. Espinoza remarks that ample land for cultivation engendered pacific relations between Chachapoyas ayllus. He views the fortifications as providing defense against the abundant jaguars, pumas and bears (Ibid.:234), and against lowland marauders from the north and east (Ibid.:235-236). However, Bandelier noted oral traditions describing hostilities between settlements such as Cuelap and Levanto (1907:19). An oral history of hostilities between the Chachapoyas settlements of Cunturrnarca and Coben survives today in Condorrnarca (personal observation, 1986). Espinoza concurs with Langlois' (1939a:233) conclusion that, because there was no political unity, there were no "organized temples" nor attendant religious specialists, and that coordination of religious activities was charged to shamans (Espinoza 1967:235). Jesuit priest Jose de Acosta (1958 [1577] :236) commented on the activities of powerful Chachapoya shamans: "grandes hechizeros y herbolarios," some 154 of whom were coaxed into giving up their practice of poisoning their enemies. Bandelier observed shamanic curing with cornmeal, coca and an herb called shayr at the end of the nineteenth century, and remarked that "these medicinemen are for good as well as for evil" (Bandelier 1907:13-15). Fifty years later, the Reichlens (1950:221-222) found little remaining of native curing practices and other indigenous traditions. Chachapoyas Under the Inca Following the Inca conquest led by Tupac Inca Yupanqui, the Chachapoyas Province was divided into hunos or decimal units of 10,000 tributaries for the purpose of expedient administration. A provincial Inca administrative center was built at the site of Cochabarnba and a curaca installed to govern the huno of Cochabarnba y Leimebarnba (Schjellerup 1979-80:302). Tupac Inca Yupanqui charged the curaca Chuquipundio of Cunturmarca with the governance of the southern Chachapoyas huno of Cunturmarca y Collai (Espinoza 1967:240). The seat of administration may have been located at approximately 3,500 m several kilometers upslope from the modern village of Condormarca where local residents refer to a ruined settlement of circular and Inca-style stone buildings as "Cunturmarca" or simply "Pueblo Viejo" (personal observation, 1986). At Cunturrnarca, a geometric frieze on the facade of one rectangular structure illustrated by Izaguirre (1923:194) was still partially 155 preserved in 1986. Schjellerup (1984) describes Inca dominance of Chachapoyas as always tenuous. Conquest was achieved only after a series of fierce battles at several Chachapoyas settlements like Cunturrnarca and Cajamarquilla, and heavy losses suffered by Inca militia (Garcilaso de la Vega 1966:479). Huayna Capac quelled one rebellion (Ibid.:554) and natives were punished by the relocation of some populations, several of which were sent to Cuzco. The Inca simultaneously implanted foreign mitimaes composed of Huanca, Cajamarca and Chimu-Lambayeque groups from the south and west. In 1532, the unforgiving Chachapoyas offered their military services to Pizarro's troops in order to rid themselves of Inca domination. The Spaniards rewarded them with independence until Alonzo de Alvarado and his troops marched into Chachapoyas in 1537 to claim it for the Crown. Spanish Conquest and Administrative Units in Southern Chachapoyas The complexities involved in utilizing colonial administrative units such as encomiendas or repartimientos to locate places and interpret ethnic and/or political boundaries in such a poorly understood region are substantial. Nevertheless, the utility of this analytical approach has been demonstrated by Julien (1985) and will be pursued in the following discussion. The earliest well- known published maps that provide detail on Chachapoyas and 156 the Pataz-Abiseo area were drawn by Franciscan missionaries in the late 18th century (e.g. Sobreviela's and Amich's maps in Izaguirre 1923:105, 237). forest place names. These provide few montane Sorely lacking are detailed maps and other published documentation pertinent to Jesuit activity in the 16th and early 17th century. Unfortunately, Espinoza's map of ethnic groups and places on the MaraftonHuallaga divide (1967:333) lacks sufficient supporting bibliographic information with which to evaluate its accuracy. Maps currently printed by the Institute Geografico Nacional (IGN) furnish few clues to locate the pertinent places. According to pre-Toledo era census information compiled in 1538, 1548 and 1561 (Rivera Serna 1956-57; Loredo 1958; Hampe 1978), two southern Chachapoyas encomiendas of primary importance to this study can be isolated, although their boundaries are difficult to establish. The encomienda of Sucos y Puymal charged to encomendero Honorato Juan Bautista Esteban includes places and/or ayllus named Sucos, Puymal, Piax and Baldeboyo. The modern highland towns of Pias and Buldibuyo appear on modern maps on the Marafton side of the cordillera bordering the Rio Abiseo National Park. The encomienda of Cajamarquilla y Condormarca held by Ines Nieto and Juan Garcia Sanmames abuts Sucos y Puymal to the north, and Collay held by Juan Montenegro lies to the south. That settlements at Sucos and Puymal were located deep 157 within the Tropical Montane Rain Forest will be demonstrated below. During the Viceroy Toledo's population relocation program in the 1570s, at highland "Vchupiax" "Andaraca" the forest inhabitants were resettled (probably modern highland Pias) and (location unknown) where presumably they were easily manipulated (Miranda 1906) . Spanish administrators torched the abandoned villages so that resettled families would not return (Espinoza 1967:237). Nevertheless, some families did surreptitiously return to the montane forest by 1593 as Mogrovejo's visita implies (Ibid.). Demographic Collapse Garcilaso de la Vega (1966:476) reports an estimate of more than 40,000 inhabitants in the Chachapoyas Province prior to Inca conquest, a sum which he most likely took from Chachapoyas-native Blas Valera's lost manuscript. This figure seems far too conservative given Julien's calculation of over 43,000 individuals for the Colonial Period corregimientos of Cajarnarquilla, Pacllas and Luya y Chillao in the 1570s (Julien 1985:25-26, Table III). The basis for Garcilaso's estimate remains unknown, but the sum of 40,000 probably refers only to tax-payers (tributaries). Lerche's (1995:36) estimate of 300,000 to 500,000 individuals in preInca Chachapoyas seems more reasonable. Cook (1981:110) confirms that "major epidemics swept Peru in the 1520s, 1530-2, 1546 and 1558-60," and estimates a total population decline in the Andes between 1520 and 1620 as "approximately 158 93 percent" (Ibid.: 114) . According to Cook (1981:195), the northeastern Peruvian Andes experienced a demographic rate of decline "generally more than double the north highland average, " and "paralleled and in some cases exceeded the rates of the coast." Regarding the population decline in the southern Chachapoyas corregimiento of Cajarnarquilla, Cook (1981:183) observes that, "the populations of the corregimiento were relatively large, but the tributary population was declining at a rate much higher than that of the other 'highland' repartimientos of the north ... " During the 27 years between 1575 and 1602, the number of individuals paying tribute in the repartimientos of Cajarnarquilla and Buldibuyo {corresponding to the encomiendas of Cajarnarquilla y Condorrnarca and adjacent Sucos y Puyrnal) fell 45 percent and 56 percent respectively {calculated from Cook 1981:185, Table 40). In 1576, Jesuit priest Cristobal Sanchez died of fevers after administering last rites to victims of a smallpox epidemic in nearby Leimebarnba {de Acosta 1958:238). "Few Indians" were left in Moyobarnba that year because of "high mortality" (Ibid. :234). To summarize the fate of southern Chachapoyas populations, Cook cites Chronicler Vasquez de Espinoza's observation: Caxamarquilla y Collay ... has at present few Indians, because many have died, and others have retreated to the land of war that they call los Aucaes; all of the province is heavily forested and very warm ... (Vazquez de Espinosa 1969 159 [1626] :281). The term Auca derives from the Quechua "Auka" meaning "enemy" or "rebel" (Cooper 1946:690). In Ecuador it was reportedly used by the Quijos to refer to "the pagan tribes of the lower forest regions" (Steward 1948:653). Espinoza's report (1967:237) of Toledo-era refugees in the montane forest lends support to the idea of an eastward flight from the area of Spanish control. Espinoza (1967:230) and Zevallos (1987) both emphasize the population decline caused by the abuses of early Spanish landlords responsible for organizing native laborers. Most of the Spaniards controlling Chachapoyas in 1548 held mines (Loredo 1958), but a scarcity of labor ultimately left many of these idle (Vasquez de Espinoza 1969:281). (1921 [1593] :70) counted eight "negros" Mogrovejo (probably African slaves) and some "yanaconas" working the mines of Bartoleme Gutierrez near Condormarca. Slave labor at the mines was not unusual in Colonial Peru, but here the importation of valued laborers from elsewhere may reflect local scarcity. Cook reports that demographic collapse precipitated the decline in gold production in the nearby region of Jaen (Cook 1981:191). Apparently, such logistical difficulties overshadowed and ultimately extinguished Spanish interest in Chachapoyas' mineral wealth as well. The Central Huallaga Valley The prehistory and early colonial history of the 160 central Huallaga River valley is of secondary importance to this study only because Manachaqui Cave lies squarely within southern Chachapoyas. Unfortunately, most historical descriptions portray Amazonian societies already altered by decades of disruptive European influence. Epidemic diseases, and Spanish and Portuguese slaving reportedly prompted demographic instability and intensified hostil1ties throughout the region (Reeve 1994:109-111). Differing colonial administrative policies exercised by the Spaniards mean that much information pertinent to the years preceding Spanish conquest commonly provided to civil census-takers in the highland Andes will never be forthcoming for most of upper Amazonia. Much of our knowledge of 16th to 18th century central Huallaga populations in and adjoining the study area is gained from written reports by missionaries and soldiers, but the period prior to Franciscan missionary interest beginning in 1670 remains obscure. The following account provides only a cursory glimpse of pre-contact central Huallaga cultures abutting Southern Chachapoyas. Early Missionary Contacts According to the earliest documentary reports (Mogrovejo 1921 [1593]), the Cholon and their northern neighbors the Hivito occupied the central Huallaga lowlands within and adjacent to the study area. Seventeenth century missionaries describe their homelands as behind the corregimiento of Cajamarquilla (ARSI 1636-37:fos.116v, 121: 161 transcription by Mary-Elizabeth Reeve). By 1600, itinerant priests, apparently working out of highland Condorrnarca, had "reduced" substantial numbers of Cholones and Hivitos, and established churches at five of nine listed settlements (Mogrovejo 1921:68-69; de la Riva Herrera 1907 [1655] :290). According to seventeenth century Jesuit documents, the Cholones and Hivitos were conquered by the Incas (ARSI 163637:fos.116,121). Subsequent to the Spanish conquest and administrative partition of Peru, the two groups were included within the Chachapoyas encomienda and doctrina of Cajamarquilla y Condorrnarca (Ibid.). pastoral visita in 1593 (1921) Archbishop Mogrovejo's lists the number of tributarios at the nine Chol6n and Hivito settlements. These lowlanders were among the first upper Amazonian groups to be "missionized" by virtue of their incorporation within the southern Chachapoyas encomienda/doctrina. Information regarding Chol6n and Hivito residence patterns is also provided by Mogrovejo•s pastoral census. At the Chol6n "pueblo" of Quisupay (where there was a church), the archbishop encountered at least 120 natives living in three houses ("Casas apartadas"). At Chamal (where there was no church) , two houses sheltered at least 88 persons. same. Hivitos residence patterns were apparently the Historically such residence patterns are common to tropical forest societies of the western Amazon basin in which extended families reside communally in malocas, large 162 houses constructed of perishable materials (Steward 1948; Meggers 1971). Mogrovejo may have been describing typical mission dwellings in which long houses were divided into single family apartments (Metraux 1963:649), but apparently the Chol6n and Hivito villages were not proper mission settlements (reducciones). All of the settlements except for San Joan de Ulat retained their solely indigenous names. Mogrovejo's description of an individual named Don Joan Momuman as headman of the "montafia" provinces ("cacique de toda la dicha montafia") indicates the presence of some overarching native authority. According to Reeve's analysis of documentary evidence (1994:112-113), the Cholones and Hivitos acted as "border intermediaries" mediating exchange which linked lowland floodplain-dwellers such as the Tupian-speaking Cocama to highland Chachapoyas populations. prior to Spanish contact, Reeve observes that, "the Chol6n and Hivito appear to have represented the uppermost extent of tropical forest peoples into the Andean region along the Upper Huallaga River" (1994:112-113). The desire to control this well- established commercial conduit for Amazonian produce likely constitutes the rationale behind Inca conquest of this forested region. The first century of Spanish and Portuguese contact brought an undetermined number of epidemic diseases from both the highlands and lowlands (Cook 1981, 1992; Myers 163 1988; Denevan 1992; Reeve 1994). Studies by Myers (1988:65- 66) conclude that the Cocama on the lower Ucayali and Huallaga rivers suffered an epidemic between 1558 and 1560. Measles and small-pox spread along the course of the Huallaga between 1638 and 1645. Contact period trading activities evident all along the eastern slopes undoubtedly accentuated the transmission of contagious diseases between the Andes and Amazonia. Intensive interaction at particular locations would have created an effective nexus of disease vectors with dire local consequences (Dobyns 1992). Denevan (1992:220-221) regards both the population densities and the impact of disease as relatively light in the Amazonian "forest uplands" (excluding the ceja de montafia). Considering the uplands as a whole his low counts may accurately reflect demographic change, but his level of generalization does not take into account local environmental, social and historical factors that rendered some populations more vulnerable to contagion than others. Forest upland groups linked into disease vectors centered in the Andes and upper Amazon River floodplains by habitual interaction probably suffered disproportionately. The sudden demographic disintegration at Bagazan on the route connecting San Juan de la Frontera (Chachapoyas) and Moyobamba documented by Cook (1981:195-196) may illustrate the heavy price paid for intermediating highland-lowland exchange. Population clustering around the churches 164 established early in Cholon and Hivitos country must have provoked similar consequences. Early seventeenth century competition for converts between Jesuit missionaries from Borja near the bend of the Marafion, and Franciscan missionaries from Huanuco widely disseminated coveted trade goods which acted to amplify and distort patterns of trade and warfare in the intervening area (DeBoer 1981; Reeve 1994). Increased raiding to obtain slaves to exchange with the Spanish for iron tools came to characterize contact between the central Huallaga, the lower Ucayali and the upper Amazon (Reeve 1994:112). Already in 1593, missionary observers noted that travel between settlements was dangerous, and that the Cholones and Hivitos defended themselves from attacks by the Motilones (or Lamas) in a "fortress" (Mogrovejo 1921 [1593] :69). Missionary records state that the Cholones and Hivitos suffered attacks by the Chocoltos (the Choltos listed by Steward 1948:600) from the north, and also that the two groups warred with one another (ARSI 1636-37:fol.116v, 121). Between 1630 and 1636, Jesuit missionaries from Trujillo reportedly catechized upper Amazonian natives east of Moyobamba and Cajamarquilla (Vargas Ugarte 1941:29-30). Missionaries devoted to newly proselytizing the Cholones and Hivitos were deemed particularly successful. In three years the Jesuits documented native languages and established two mission settlements: La Concepcion with 709 Cholones and San 165 Javier with 624 Hivitos. By 1640, the Hivitos had endured hostilities perpetrated by the Cocama and Aguano of the lowland river floodplains to the northeast (Reeve 1994:111; Lopez de Alvarado 1907 [1656]:362). At some point, the lowland chain of violence spilled into southern Chachapoyas where raids from the central Huallaga "destroyed" the Chachapoyas settlements of Condormarca and Collay (Amich 1854 [1768] :75; Steward 1948:601). At about this same time, the Trujillo Jesuits abandoned the central Huallaga doctrina settlements (de la Riva Herrera 1907 [1655]:290), perhaps because of violence, epidemics and/or neglect by central offices. Some Cholones and Hivitos later testified that the priests were driven out by "some old caciques" who threatened to kill them (Ibid.). The names of the settlements listed by Vargas Ugarte and Mogrovejo do not appear in subsequent documents. In August of 1654, General Don Martin de la Riva Herrera began a campaign to "pacify" the lowland populations along the central and lower Huallaga river course. De la Riva Herrera planned to enter the northeastern montane forest via Condormarca and Capillania with over two hundred soldiers. According to de la Riva Herrera's (1907) reports, a party of sixty lowlanders anticipated his arrival and climbed to welcome the soldiers at the highland "puesto de Cillangat foods (?)" [sic] with gifts of cassava and other lowland (de la Riva Herrera 1907:286-287). Five days' march 166 north from Capillania, this "puesto" is clearly the Tarnbo de Callangate (3,500 m) in the Tropical Alpine Zone between Bambamarca and Cajamarquilla (Raimondi 1900a:36). With his lowland guides, De la Riva Herrera likely descended a route later identified by Raimondi (1900a:39) that parallels the Tubaybal and upper Tepna rivers, affluents of the Jepelache River. A five day journey brought the company to a settlement of the Porontos. Steward (1948) does not mention the Porontos who may have occupied territory either north or west (upslope) of the Hivitos. De la Riva Herrera claimed the Porontos province for the king, ordered the building of a church and renamed the settlement San Antonio de Porontos. After installing a government and a resident priest, he distributed axes, machetes and knives (de la Riva Herrera 1907:288-289). Then he marched two days to Hivitos country. The Hivitos also peacefully received de la Riva Herrera, who renamed their village Limpia Concepcion de Jivitos and sent for Chol6n caciques to participate in the proceedings. After receiving assurance from the gathered natives that the trouble-making "old cacique" had since been eaten by a "tigre," de la Riva Herrera distributed more gifts of tools and left Spanish-sanctioned authority, including a priest, before departing. The soldiers built rafts in Hivitos country and, despite difficult passages, they descended the "Xivitos River" (the Jepelache and/or 167 Huallabamba?) 12 leagues (his estimate) to the Huallaga in approximately three days (de la Riva Herrera 1907:293). At this point, De la Riva Herrera marched on to lead a military fiasco in Jivaro country, where he eventually distinguished himself by his abusive tactics (Jouanen 1941:417-426; Reeve 1994:133). Ultimately these later Cholon and Hivito settlements also failed for unknown reasons.· To Jesuit authorities in Quito, the doctrina of Cundurmarca represented their most distant southern outpost, one which perhaps they no longer cared to support. Priests were obligated by law to accompany conquistadors regardless of an order's interest in supporting the endeavor (Reeve 1994:120). Any initial enthusiasm demonstrated by the newly installed missionaries must have been quickly dampened by epidemics reported in the nearby upper Huallaga region in 1662 and 1670 (Steward 1948:597), and by the widespread "reign of terror" incited by rebellious Cocamas who attacked mission settlements downriver between 1663 and 1666 (Reeve 1994:122). In 1670, lowlanders once again emerged from the forest to request missionaries and resumption of peaceful trade with the highlands. This time the Franciscan order based in Huanuco filled the void left by the Jesuits and established their own permanent missions among the Cholones and Hivitos in 1676 (Amich 1854:76; Steward 1948:600-601). The Hivitos were settled at Jesus de Pajaten and Jesus de Monte-Sion, 168 and the Cholones further south at San Buenaventura del Valle and Pampa Hermosa (Amich 1854:78). Amich provides sketchy descriptions of the Cholones and the Hivitos who he found scattered throughout the forest by small-pox epidemics and no longer answering to any single political authority (Ibid.:19). He reports that fishing and the cultivation of typical lowland crops provided central Huallaga subsistence, while cotton and coca were also grown. Little documentary information exists regarding customs and ritual activities. Steward (1948:605) states that, "the Cholones were reputed to be powerful doctors in 1830," but he provides no indication of his source for this information. Ethnohistoric Evidence for HighlandLowland Boundaries An important question formulated during Julien's research (1985) on Chachapoyas ethnohistory concerns the explicit justification of "highland" boundaries in the eastern tropical forests. Documentary evidence may be utilized not only to justify Espinoza's placement of Chachapoyas' eastern boundary, but also to show that the upper montane forest served as the locus of social and political activities in the Pataz-Abiseo area. Two historical references identify Sucos and Puyrnal as forest settlements on the Huallaga side of the cordillera. To conduct his pastoral visita of 1593, Archbishop Mogrovejo entered the montane forest from "Cundumarca" (presumably 169 Condormarca) and traveled to, "some villages inside the montana, that they call los montes de Puymal" 1921:68). (Mogrovejo Mogrovejo's report claims that "all of the Indians of the montana" gathered at the "pueblo" of "Yare de Puymal" or "Yare que llaman Puymal," to be counted. Espinoza (1967:237) speculates that Yare is the original name of the site dubbed Gran Pajaten by Savoy (1970), and Abiseo by Bonavia (1968b). After descending from Yare de Puymal an arduous 16 leagues ( "se va con ha.rto trabajo"), Mogrovejo's entourage reached the first of several Hivito settlements where the inventory of residents and converts continued. The brief report provides no information regarding settlement in the intervening country. The second geographic reference to Sucos and Puyrnal is contained in writs dated 1653 and addressed to the king touting Condormarca as the most appropriate point of entry for General de la Riva Herrera's lowland pacification campaign (Bautista de Escobar 1899 [1653] :81-83). Entering "tierra de montafia" from Condormarca, Bautista de Escobar notes, one encounters the abandoned "pueblos" of "Xucos, Puymal, Cholones, Xivitos, San Geronimo de Chinch6n" and others. The details provided in Mogrovejo's and Bautista de Escobar's testimonies confirm that Sucos and Puymal were montane forest settlements situated above territory occupied by Cholones and Hivitos, and that early southern Chachapoyas encomiendas included both highland and upper montane forest 170 territories and/or populations. If Espinoza is correct in assuming that the ayllu constitutes the principal Chachapoyas socio-political unit from which the principal settlement takes its name, then it follows that Sucos and Puymal were forest-centered ayllus and not colonies dependent upon larger, more powerful highland polities. Pre-Toledo era lists of corregimientos and encomiendas within and beyond the province of Chachapoyas are likely to contain the names of other eastern montane forest settlements and/or ayllus. Pinpointing their locations with the available documentary information constitutes a challenge for future ethnohistoric investigation. The native populations settled in and around the PatazAbiseo area probably belonged to the encomienda of Sucos y Puymal rather than to Cajamarquilla y Condormarca. This conclusion is based simply upon the distribution of known locations of place and/or ayllu names grouped within each. That Mogrovejo and Bautista de Escobar speak of gaining entrance to the forested sites of Sucos and Puymal via highland Condormarca seems to contradict this inference. However, the Condormarca entry to the forest likely provided the direct access to the settlements at Sucos and Puymal from the north (and Cajamarca) prior to Spanish colonization. Once within the montane forest, Mogrovejo's entourage may have journeyed southeast to encounter Yaro de 171 Puymal in the upper Montecristo River where ruined settlements including Gran Pajaten are found. Alternatively, Mogrovejo may have found Yaro de Puymal where modern Condormarca residents claim that there are ruins, due east in the upper Pajaten River valley. Unambiguous evidence that the lower montane forest was occupied during the late 18th century derives from Franciscan documents. In 1791, Padre Sobreviela listed 28 "Indios y Neofitos" at Jucusbamba in the Abiseo River valley (Unanue and Sobreviela 1963:157). The information at hand does not clarify whether the missionaries established the settlement with a local population or brought in natives from above and/or below. The ruined foundations of a church found in association with Inca structures near La Morada (2,100 m) in the Bombonaje River valley by Schjellerup (1992:359) may represent an earlier (16th or 17th century) reducci6n. Pre-Hispanic settlement on the eastern slopes of the Marafion-Huallaga divide may have been continuous, or perhaps broken by lightly occupied territory utilized for coca production in a manner like Raymond (1985) and Bonavia (1972b) describe for the premontane forests of the Apurimac and lower Mantaro valleys. The evidence just recounted places the Porontos, Hivitos and Cholones within the premontane forests where the latter two groups reportedly cultivated coca during the seventeenth century (Amich 172 1854:78-79; Steward 1948:601). Central Huallaga groups may have moved upslope to occupy desirable land vacated by beleaguered highlanders. At the same time they likely absorbed many highland refugees. Also, increased raiding by the Aguanos and Cocama along the course of the Huallaga early in the seventeenth century probably rendered river terrace locations undesirable for settlement. In sum, paltry documentary information distorted by high disease mortality, population dislocations and forced relocations does not offer an adequate basis for definitive interpretations of highland-lowland frontiers in the study area. Ethnohistoric Evidence for HighlandLowland Interaction Much documentary evidence for pre-Hispanic commercial relations between southern Chachapoyas and the central Huallaga groups is indirect. First, it seems that the incorporation of the montane forest settlements and lowland "provincias" of the Cholones and Hivitos into the highland- based Jesuit "doctrina de Cundumarca" prior to 1593 (Mogrovejo 1921:67-70) indicates early, perhaps preHispanic, linkage between these Andean and Amazonian populations. Priest Rodrigo Alonso's doctrina stretched from Calemar and Bijos (now Vijos) on the banks of the Marafion into the central Huallaga valley lowlands, and included Condormarca and Puymal. The Condormarca doctrina 173 is apparently unique as lowland forest communities do not figure into any other highland doctrina census counts south of Moyobamba. Other clues suggesting linkages between southern Chachapoyas and central Huallaga populations have been offered.by Julien (1985). Based on her interpretation of Loredo's pre-Toledo era (1548) list of tributaries, Julien believes that the terms Imigas, Ancimgas and Animigas (Loredo 1958:259-260) refer to lowlanders. She points out that these terms are contrasted in the lists with "Serrano" (highlander). Juan Lopez Montenegro (or Montero) held "unos Imigas" in his highland encomienda of Collay (Ibid. :260) . One of the two caciques subject to the encomienda of Chilchos y Laya on the upper Huallabamba is referred to as an Ancimga (Ibid.:259). Encomendero Juan de Rojas was charged with "otros tantos Imigas" in the lowland Bagua area (Ibid. :264). Highland Collay and nearby Tayabamba sit atop a route descending the Mishollo and Tocache Valleys to the Franciscan mission of Pampa Hermosa. An approximation of the route between Tayabamba and lowland Tocache was utilized by Franciscan missionaries (Raimondi 1876:428; Unanue and Sobreviela 1963:154), and is still travelled today. After 1670, the Franciscans maintained access to Cholon and Hivito settlements by two additional trails. One connected Capillania above Condormarca and the mission of Jesus de Pajaten (approximately 400 m) where travelers 174 embarked down the Pajaten River in canoes. It probably fell into disuse when the mission burned in 1801. Accounts of lowland traders appearing in Condormarca persist in oral narratives (personal observation, 1986) . The other route descended the Abiseo River valley from the Franciscan convent at Huaylillas, passed by the installment at Jucusbamba, and terminated at the mission settlements of El Valle and Si6n near the banks of the Huallaga. Only its upper portion is traveled today by ten highland families that have recently resettled Jucusbamba (Young et al. 1994). Along these routes, the Cholones and Hivitos from the Franciscan missions habitually traded lowland produce, especially coca, to highlanders for tools and clothing (Amich 1854:78-79). Nineteenth century commerce brought coca and tobacco to the highlands (De las Casas 1935a:345). However, Raimondi (1900a:40) noted that traditional trade between southern Chachapoyas and the lowlands had ceased despite the potential economic advantages of obtaining cheap salt from central Huallaga sources. According to Rios et al. (1982:88), the Abiseo River route through Jucusbamba and Achiras was abandoned with the opening of the central highway connecting Huanuco and Tinge Maria during the mid20th century. This leaves only the southernmost Tayabamba- Tocache route still in sporadic service to foot traffic. 175 Archaeology of Southern Chachapoyas and the Central Huallaga Valley This section presents archaeological evidence for ancient settlement on the Marafion-Huallaga divide. Discussion focuses on regional chronologies, subsistence economies and exchange. Clear distinctions between archaeological assemblages from southern Chachapoyas and the central Huallaga valley correspond to expected differences between Andean and Amazonian cultural traditions. A narrowed focus on the Pataz-Abiseo area highlights data pertinent to site distribution by ecological zone, and evidence for chronology and economic activities. Most important to note is the evidence for greater population concentration on the eastern slopes of the divide, a pattern that contrasts sharply with that observed today. Southern Chachapoyas has attracted few systematic archaeological studies to date. Because of material similarities shared within greater Chachapoyas, there is a temptation to assume that southern Chachapoyas' developmental trajectory differed little from that of the better-known northern region. It should be useful therefore to consider briefly the results of investigations throughout Chachapoyas prior to focusing on the Pataz-Abiseo area. The central Huallaga has received only three archaeological studies to date, but only one of these bears directly on prehistoric developments within the Pataz-Abiseo area. A summary overview of archaeological work on both sides of the 176 divide will offer background information that should ultimately aid evaluation of data from the study area. The history of investigations into the archaeology of Chachapoyas province reflects archaeologists' unwavering fascination with Chachapoyas architecture, especially the spectacular "fortresses" of the upper Utcubamba. However, the bias that this emphasis on architectural contexts has imposed upon reconstructions of the region's prehistory remains largely under-appreciated. Some investigations have been confined to cleaning and recording architectural details, while others describe excavated contexts within and around buildings. The cultural sequences for which there is stratigraphic evidence derive from investigations in and around Cuelap (Reichlen and Reichlen 1950; Horkheimer 1959; Ruiz 1972). Other projects center attention on late prehistoric settlements around Uchucmarca (Thompson 1973, 1976) and Chuquibamba (Schjellerup 1990, 1992). These projects have rendered some appreciation of architectural variability and settlement patterns, and an approximate age for Chachapoyas architectural traditions. To date there has been no concerted attempt to investigate the antiquity of human occupation in Chachapoyas territory. With the exception of the recent report of archaeological investigations in and around the Rio Abiseo National Park (Lennon et al. 1989), the archaeological literature from Chachapoyas lacks references to possible 177 Preceramic Period occupations. The cultural sequence outlined for the upper Utcubamba valley begins with Ruiz's Early Intermediate Period Cancharin Phase (Ruiz 1972). Archaeological remains dating to the Initial Period and Early Horizon have not been reported on the Marafion-Huallaga divide between Bagua and Tantamayo in Huanuco Department (Bonnier 1983). The Reichlens (1950:300) noted that, although Tello's 1937 explorations on the divide stopped just south of the Utcubamba drainage, he placed Chachapoyas within the "area of distribution" indicated for his ChavinKotosh Civilization (Tello 1942:711, Lam. III). Ruiz (1972:181-183) perceived a close historical relationship between Cuelap's Cancharin Phase styles and Early Horizon styles from lowland Bagua, but did not attempt to evaluate Tello's migration hypothesis. Ruiz's interpretation of the excavated Cuelap sequence emphasizes local cultural development and stylistic continuity in pottery design attributes from the Cancharin Phase through the Late Horizon Cuelap-Inca Phase. A coarse brown culinary ware dominated by simple jar and bowl shapes often exhibits applique decoration and incised or notched rims. Cancharin Phase bowls with negative resist and white- on-red geometric painting may indicate connections farther north. Schjellerup has identified coeval "Cajamarca I" ceramics between Cuelap and Pataz in Chuquibamba (1992:362). The architectural tradition at Cuelap featuring 178 circular stone dwellings appeared by the end of the Middle Horizon (A.D. 800 according to Narvaez 1988:139). At this time Cajamarca cursive style painted bowls were locally imitated (Ruiz 1972:188), and widely imported across the Marafion between Cuelap and Cajamarquilla (now Bolivar) farther south (Ibid.:187; Vega 1978:19). Lerche's (1995:24- 25) suggestion of a Middle Horizon hiatus in the Cuelap sequence is not supported by Ruiz's data. During the 1967 clearing activities at Cuelap, Ruiz (1969) recovered sherds that he considered Middle Horizon 2B imports from Ayacucho or Ica. Shady (1987a:86) assigns these same sherds to Nazca Phase 9. Surface architecture at the most elaborate Chachapoyas hilltop settlements seems to date to the Late Intermediate Period and Late Horizon (Narvaez 1988; Kauffmann 1991; Schjellerup 1992; Lerche 1995), as do chamber tombs set into cliff faces. Shady (1976:585, 1987a:86) reports that pottery of the corresponding Cuelap phase has been found in Cajamarca and Bagua, indicating long-distance exchange with the highlands to the west and lowlands to the north. Schjellerup's (1985) studies of agricultural earthworks such as slope terraces between 3,200 and 3,800 m postulate extensive cultivation of potatoes and other high altitude tubers in the Chuquibamba area. These activities, in combination with camelid herding, constitute a subsistence orientation much like that of other highland Central Andean 179 regions south and west. Its prehistoric development in Chachapoyas remains unclear. Excluding the study area, virtually all of the radiocarbon dates from Chachapoyas have been reported by Schjellerup (1992), and none pre-date the Late Intermediate Period. These and other ancient settlements span the upper elevations of Young's Moist Montane Zone and the lower elevations of his Tropical Alpine Zone. On a visit to the village of La Morada near 2,000 m within the lower montane forest, Schjellerup located a small complex with classic Inca architectural features and a bath that she named Pukarurni. Lerche (1995) provides a sketchy description of additional Inca remains at similar elevations within the Huallaga watershed. Most of the central Huallaga river valley remains archaeological terra incognita. DeBoer (1984) surveyed selected areas of the floodplains and dry tropical forest along the valley bottom north of the study area between Bella Vista and the Pongo of Aguirre. Significant for this study is his interpretation of several sherd scatters located not far from the Huallaga-Huallabarnba confluence (HUA-4, HUA-5) as remains left by a multi-family house, or maloca. Late ceramics with crushed-sherd temper are much like the material that central Huallaga potters produce today. An older style from Chazuta (Ibid.:114-115) and Tarapoto (Myers 1981b) shows some design similarities to the 180 Curnancaya ceramics described further south (Raymond et al. 1975). DeBoer concludes that central Huallaga styles are typical of an ancient and wide-spread upper Amazonian potting tradition. featuring bowls, A shared tri-partite shape inventory jars and ollas probably has "great time depth," and may be "indigenous to the montaiia" 1984:114). (DeBoer Excavated collections from the study area described by Ravines (1978, 1981b) will be addressed in the following section. Archaeology of the Pataz-Abiseo Area Before attempting to derive some understanding of population distributions and regional economic relations from archaeological data so far recovered in the PatazAbiseo area, the data's limitations should be understood. The study area has not been extensively surveyed on either side of the divide, although some site localities have been recorded prior to Rio Abiseo National Park Research Project (RANPRP) studies in 1985. The sites that we visited in Pataz and in the upper montane forest are typically those with standing stone architecture well-known to the local villagers. The villagers are intimately familiar with sites on the Marafion side of the divide, but, collectively at least, they know little about sites within the Huallaga-side montane forest. The chief biases distorting perceptions of the local prehistoric panorama are those posed by certain construction materials (perishable vs. stone), and forest 181 cover. The following paragraphs represent the current understanding of selected features of Pataz-Abiseo archaeology. Pataz District Several archaeologists passing through Pataz (usually en route to Gran Pajaten) have observed some archaeological features such as pictographs, chamber tombs and terraces. During investigations in areas higher up and farther east between 1985 and 1990, RANPRP archaeologists described and mapped some of these features, and photographed the collections of local farmers. The following observations on the archaeology of Pataz District stem primarily from these observations. In the Dry Forest Zone, only one site discovered fortuitously in 1990 has been documented between the Marafion and the town of Pataz. A series of small terraces leading up to a single ruined stone enclosure lie at approximately 1,400 m, 200m above the Marafion. A few thin sherds from fine painted bowls may indicate an Early Intermediate Period occupation. Clearly the Dry Forest Zone was utilized, but it still awaits systematic research. The lower portion of the Moist Montane Zone below the town of Pataz has not yet received attention from archaeologists. In the surrounding heights, however, villagers are familiar with two small clusters of ancient stone structures crowning promontories that extend westward 182 into the Marafion canyon on either side of the Frances River valley. Rojas' A visit by RANPRP archaeologists in 1990 confirmed (1967:9) report of ruins above Zarumilla on the crest of Cerro Cuyrnuy (3,380 m). This site, called El Ushnu by Zarumilla residents, consists of four oblong ascending terraces reinforced by low field-stone walls that delimit a total area of approximately 1.2 hectares. Remains of at least seven semi-circular stone walls can be discerned through the vegetation, but recent fence construction seems to have borrowed much of the ancient construction material. All but one of the 24 sherds collected from the surface pertain to large jars with long flaring rims. The paste is sand-tempered, and rim morphology is identical to that of Late Horizon pottery illustrated by Bonavia (1968b) at Gran Pajaten. The odd rim sherd belongs to a neckless olla. Zarumilla residents also report sherd scatters farther up the ridge at Pampa Verde near 4,000 m. El Ushnu's extraordinary view up and down the length of the Marafion canyon suggests that it was one of several sites established at strategic locations for geo-political control. From El Ushnu, the stepped profile of Cerro Alto Las Pircas (4,100 m) across the Frances valley (behind Pataz) can be seen. Pataz residents report that a small cluster of stone walls is found there, but archaeologists have yet to examine the site. The Institute Nacional de Cultura's inventory of archaeological monuments (INC 183 1983:Hoja 16-g) specifies late dates of occupation for a ruined settlement on Cerro Tinajera (3,102 m) above the south side of the Lavasen River. RANPRP archaeologists did not visit the site, nor could additional published information on the settlement be located. Of the known sites within the study area, only the site on Cerro Tinajera at the northern extreme seems to be a settlement, and therefore potentially indicates something of pre-Hispanic Pataz population distributions. Pottery that has surfaced in fields surrounding Los Alisos (3,100 m) appears to be late prehistoric in age. Nearby funerary chambers and pictographs previously described by Perez (1969) and visited in 1990 on the slopes of Cerro Colpar may also date to the Late Intermediate Period or Late Horizon. The preliminary assessment of prehistoric settlement in the Moist Montane Zone of Pataz District is that it may have been lighter than expected. However, the bias favoring reports of intact stone constructions may encourage specious impressions. Pataz settlement also appears sparse compared with information from adjacent southern districts gathered during surface surveys by Curtin (1951). Pre-Hispanic settlements documented in Parcoy, Buldibuyo and Chillia further south typically feature circular stone constructions crowning hilltops and ridgetops between 3,200 m and 3,800 m. Among several classes of observed architectural features, Curtin 184 observed several small terrace systems. Building counts are not provided, but some of these sites covered an area of several hectares. Stylistic evidence from fine painted kaolin bowls and relief sculpture on stone slabs at Nufiamarca suggest communication linkages south and west with the Callej6n de Conchucos, the upper Santa Valley and the Santiago de Chuco region during the Early Intermediate Period (Campana 1988). Aside from the Pataz-Abiseo area, the archaeology between Parcoy and Uchucmarca remains unstudied. Some late sites in Cajamarquilla/Bolivar Province have been briefly described (Vega 1979; Lynch 1992). Virtually nothing is known of the archaeology surrounding Bambamarca and Condormarca. For the Tropical Alpine Zone at the crest of the Marafion-Huallaga divide, Lennon et al. (1989) describe four classes of sites, none of which represent proper settlements. The first class consists of administrative structures and bridge remains associated with the paved "Inca" road purportedly linking Huanuco and Chachapoyas (Von Hagen 1955), and the road spurs striking eastward toward the montane forest. These structures date to the Late Horizon, but the road system is probably much older. Rock shelters such as Chirimachay (Cueva Negra) and Manachaqui Cave comprise the second class. Travellers and herders used these for protection into the pre-Inca past. Highly dispersed and isolated house footings and windbreaks, 185 usually appearing as low stone circles or semi-circles, dot the valley slopes and remain undated, as do the low (less than 50 em) stone-faced platforms constructed near water sources that constitute the fourth class of archaeological remains. The Abiseo Drainage and Central Huallaga Lowlands East of the Marafion-Huallaga divide only some upper portions of the Abiseo's tributary Montecristo river valley have been explored, but these upper montane forests apparently harbored the densest pre-Hispanic settlement in the Pataz-Abiseo area. Sites such as Gran Pajaten, Cerro Central, Las Papayas, La Playa and El Encanto (Pimentel 1967; Rojas 1967; Bonavia 1968b; Savoy 1970; Deza 1975-76; Kauffmann 1983; Lennon et al. 1989) exhibit architectural and ceramic styles similar to those reported elsewhere in Chachapoyas. While La Playa dates only to the Late Horizon (Deza 1975-76; Cedr6n 1989), recent excavations by RANPRP archaeologists at Las Papayas have produced radiocarbon evidence for a Late Intermediate Period occupation (Church 1994:293). Simultaneous investigations at Gran Pajaten yielded pottery and radiocarbon dates confirming occupation during the Early Intermediate Period, and perhaps the Early Horizon (Ibid.: Table 1). Maize was consumed, and perhaps cultivated in the site vicinity, as evidenced by charred 186 maize kernels recovered from a charcoal lens dated A.D. 40 ± 60 (uncorrected) within Building No. 1's construction fill (Ibid.). Fine kaolin ware bowls imported from adjoining highland regions indicate early long-distance interaction. Culinary pottery from both Early Intermediate Period and Late Horizon contexts features the necked jars with flaring rims and semi-hemispherical bowls. It has been suggested that similar ruined settlements and terraces should be found on all of the hilltops surrounding Gran Pajaten (Savoy 1965:4; Rojas 1967:11) RANPRP archaeologists were able to confirm the presence of an extensive settlement with an estimated 150 to 200 circular constructions across the valley at Cerro Central (Lennon et al. 1989). Between La Playa and Gran Pajaten lies Las Papayas with an estimated 100 buildings resting on a series of large terraces. The system of terraces below Gran Pajaten covers over 50 hectares, but the total extent of terrace and settlement systems in the valley cannot be estimated without more systematic survey coverage. Also, the Montecristo valley below 2,600 m remains unexplored by archaeologists. Ravines' (1978) survey and excavations in the lower Abiseo, Huallabamba and Pachicilla valleys constitute the only archaeological investigations in the central Huallaga valley premontane forest. The report provides little information regarding site environments and how the survey 187 and excavations were conducted. and 440 m. Most sites lie between 390 Ravines excavated three ceramic "complexes" and stratigraphic evidence for a relative chronology at the site of Santa Rosa (S-Huay-5). Sand-tempered pottery with excised decoration lay beneath sherds from vessels with applique decoration assigned to the Huayabarnba Complex. The Huayabarnba Complex is divided stylistically and temporally into Huayabarnba 1 and the more recent Huayabarnba 2. Ravines regards Huayabarnba 1 pottery ancestral to both the overlying 16th century Jerusalen Complex, and Andean styles of Chachapoyas (1978:531). Crushed-sherd temper characterizes the later complexes, while potters utilized fine sand temper to fabricate the earliest complex that Ravines and DeBoer (1984:117) agree is related to the Central Ucayali Shakimu styles (circa 650 B.C.). Limited information on a small sample of vessel shapes impedes judgment regarding whether or not the undated Huayabarnba 1 and 2 complexes participated in the upper Amazonian tradition described by DeBoer. The Huayabamba shape inventory appears to exhibit variability exceeding the upper Amazon tradition's. Three of the four most clearly identified vessel shapes belonging to the 16th century Jerusalen Complex (Ravines 1978:Laminas l:a, 2:1 and 2:2) more strongly suggest inclusion within the upper Amazonian tradition, but the small sample size still hinders evaluation. 188 Considering the Pataz-Abiseo area as a whole, it seems clear that settlement density by the time of Spanish conquest was weighted to the Huallaga watershed Tropical Montane Rain Forest Zone. The extensive and sophisticated terrace and building constructions at forest sites like Gran Pajaten represent investments in time, labor and materials that were not expended in potentially productive areas of Pataz. Nor does the topography of the Montecristo valley's upper montane forest favor settlement, at least for agricultural purposes (Church 1994). Sampling problems restrain attempts to place the Pataz-Abiseo settlement distributions into a regional context. Most important is the question of whether or not Montecristo valley settlement density is representative of densities in neighboring valleys north and south. Summary This chapter has considered events during particular time periods in the Pataz-Abiseo area in an attempt to identify economic and demographic transformations that have obscured perception of this area as an active node in preHispanic Andean-Amazonian commerce. Most remarkable is the demographic shift of populations from the Tropical Montane Rain Forest Zone of the Abiseo drainage to the Moist Montane Zone of Pataz that accompanied the European incursions. Achieved against a backdrop of high mortality and eastward flight, Toledo's forced relocation of these populations into 189 nucleated highland settlements probably constitutes the single most important demographic event. It also seems to represent the moment at which previous highland-lowland linkages were rendered structurally untenable. The following paragraphs will attempt to abstract other salient points from the ethnographic, ethnohistoric and prehistoric data recounted within this chapter. First, these data confirm that a highland-lowland boundary did separate cultures with markedly disparate developmental features on the eve of Spanish conquest. The review of ethnohistoric and archaeological evidence provides a view of loosely organized highland societies with small-scale settlement hierarchies, residence patterns emphasizing nuclear family households, a subsistence pattern typical of the Central Andes, and a pottery style commonly associated with highland Quechua-speakers. While the socio-political organization of central Huallaga groups like the Hivitos remains virtually unknown, both ethnohistoric and archaeological evidence demonstrate that residence patterns emphasized extended and/or multi-family households and that subsistence activities focused on manioc and other lowland produce. Ravines' Jerusalen Complex may represent the associated lowland Amazonian pottery. Early church interest in extending the highland doctrina of Cundurmarca into the eastern lowlands likely indicates a desire to tap into the pre-Hispanic tradition of 190 interaction between southern Chachapoyas and central Huallaga societies as the Inca had previously. Reports that both the Inca and Spanish officials similarly incorporated the Amazonian Cholones and Hivitos within the administered Andean province of Chachapoyas support the notion of an ancient and durable exchange linkage between highlands and lowlands in the northeastern montane forest. Scholars suggesting that the Inca failed to conquer the Amazonian forests overlook the fact that the Inca province of Moyobamba (Rowe 1946:187) is covered by tropical montane and premontane forests, and Moyobamba lowlanders probably had similar ancient traditions of trading (DeBoer 1984:15) and mediating trade (Reeve 1994). Montane forest settlements such as Sucos, Yaro de Puymal and Gran Pajaten seem to have been politically independent nodes in a pre-Hispanic communication network that linked Central Andean populations to the expansive lowland interaction spheres described by Lathrap, Reeve, DeBoer (n.d.) and others. Savoy's account of finding circular stone constructions at the abandoned Franciscan mission of Jesus de Pajaten, and a paved road near the junction of the Jepelache and Pajaten Rivers at approximately 400 m (Savoy 1970:148) provides provocative but sketchy evidence suggesting "highland" settlement at a strategic point of river navigation situated on the Huallabamba River waterway. The Huallabamba, easily reached 191 by Huallaga River trade expeditions originating downstream, probably served as an important destination offering access to a large number of potential highland "consumers." For Central Andean societies west and south of the study area, the Huallabamba could have provided the closest and most direct access to lowland Amazonian riverine interaction networks. Thus the Hivitos and southern Chachapoyas natives assumed ostensibly lucrative intermediary positions channelling goods to these Central Andean societies. Demographic collapse in the study area undoubtedly sparked breakdowns in local political economic structures and kinship-based exchange obligations, contributing to a rapid disintegration of long-distance exchange networks. The prospect of administering a sparse and isolated native population diminished Spanish economic interest in the area, despite its purported mineral wealth. Nor did the Jesuits see fit to continue support for the distant doctrina of Cundurmarca. Ultimately, the sudden population decline exacerbated the seemingly complete acculturation of remaining southern Chachapoyas populations into Peruvian mestizo culture, and the virtual disappearance of native languages. While Colonial Period southern Chachapoyas populations attempted to cope with sudden soaring mortality, the political authority within Cholon and Hivito territory noted by Mogrovejo in 1593 was likewise undermined by contagious 192 diseases, intensified warfare and slave raiding. The first Franciscan missionaries in 1676 found the Cholones and Hivitos "scattered throughout the forest neither recognizing cacique nor superior other than their elders to whom they confer special respect" (Amich 1854:76). Hopes that highland-lowland commerce could be restored dimmed slowly during the Colonial Period. Bautista de Escobar's 1653 attempt to direct De la Riva Herrera's route of conquest through Condormarca, Sucos and Puymal represents one such effort. The importance that the central Huallaga lowlanders attached to regaining access to highland trade is demonstrated by their emergence twice in the highlands (in 1654 and 1670) attempting to secure a missionary presence in their homeland. By then however, the Hivitos' highland trade partners had either died, fled or been forcibly removed from the upper montane forests. After the burning of the mission at Jesus de Pajaten in 1801, both traditional and newly forged communication routes bypassed Condorrnarca, Sucos and Puymal. Highland-lowland interaction consequently diminished to a pale remnant of its former character by the beginning of the 19th century. Epidemics that continued to strike highland communities like Huaylillas and Cajamarquilla during the mid-nineteenth century (Raimondi 1900a:37, 1900b:l39) denied highland populations the opportunity to recover. According to Mariategui, the marginalization of Peruvian Amazonia was exacerbated by the 193 end of the rubber boom after World War I (1971:162, Note 4). Mid-twentieth century construction of roads descending the eastern Andean slopes through Moyobamba to the north, and Huanuco to the south, finally assured the consignment of this region to an economic oblivion mitigated only periodically by mining booms. Information gleaned from the small quantity of available Spanish documentation does not clearly elucidate the nature of pre-Hispanic nor early Colonial Period economic activities within and adjacent to the study area. However, interaction at the time of Spanish conquest probably occurred on several scales. Local and regional exchange, which are usually imbedded in kinship networks in the Central Andes (Morris 1978), emphasized the east-west flow of goods. In the Pataz-Abiseo area, the geographic distribution of life zones which, like the cordillera, are aligned north to south, promotes movement east and west as the most efficient means of accessing the widest variety of ecozones apt to serve as desirable production zones for certain crops or other resources. The east-west economic interaction recorded by Brush (1977) at Uchucrnarca may be considered a modified version of prehistoric local or regional exchange that once characterized the MarafionHuallaga divide. North-south interaction for which there is indirect evidence probably involved the exchange of information and goods unrelated to subsistence. Within the 194 study area, montane forest populations at Yaro de Puymal and Gran Pajaten may have maintained production zones both deep in the forested lowlands and in drier Marafion-side ecological zones, although confirmation requires further research. Opportunities for controlling interregional exchange along an important traditional trade route probably represented a powerful inducement to settle the montane forests east of Condormarca, although exploitation of montane forest resources was also a likely consideration. Interregional interaction extended beyond kinship bonds and connected centers of political power across boundaries separating coastal, highland and tropical forest cultural traditions. With Spanish conquest, the collapse of Andean political and economic structures led to the virtual elimination of interregional interaction. Networks across the Andes suddenly contracted or disappeared leaving atrophied fragments of local and regional exchange that mostly served domestic economies. Ancient settlements like Sucos, Puymal and Gran Pajaten that figured importantly in interregional interaction rapidly became functionally irrelevant in the scheme of emergent Spanish hegemony. CHAPTER 4 MANACHAQUI CAVE AND ITS ENVIRONMENTAL CONTEXT There is said to be a superb view of the Montana from the summit, but the clouds {almost within reach of the hand) boiling up from the great deep below, effectually cut it off and we could see nothing. When we had got some distance down and obtained a view through an opening in the thick growth of the mountainside, we looked down upon the most rugged country I have ever seen. There seemed to be no order or regularity in the hills which were thickly covered with forest; but the whole had the appearance of the surface of a vast boiling caldron suddenly stricken motionless. {William L. Herndon 1952 [1854] :66) Archaeological investigations at Manachaqui Cave represent the most recent stage of continued research in the vicinity of the Rio Abiseo National Park begun in 1985 by archaeologists from the University of Colorado and the National University of Trujillo, Peru under the direction of Thomas Lennon, Miguel Cornejo Garcia and Segundo Vasquez Sanchez. During 1985 and 1986, survey, mapping and excavations focused on Montecristo valley sites situated between 2,600 and 2,900 m within the Tropical Montane Rain Forest Zone. During these campaigns, Manachaqui Cave served as a convenient place for archaeologists, field crews, muledrivers and cargo animals to spend the night near the crest of the Marafion-Huallaga divide during the two-day trip between Pataz and La Playa staging areas. Most of this route follows pre-Hispanic paved roads that pass close to 195 196 rockshelters such as Chirimachay (Cueva Negra) and Manachaqui Cave. Raimondi (1900b:127) obliquely refers to Manachaqui Cave as a local landmark within his description of travel from Condormarca to Pataz in 1860, and the Gran Pajaten expedition of 1965 sponsored by the Peruvian government was the first of many to camp at Manachaqui (Rojas 1967:9; Deza 1975-76:45; Kauffmann 1980:27). One of the research priorities of the Rio Abiseo National Park Research Project was to unearth stratified cultural sequences in order to date the antiquity of human presence in the montane forest. While the archaeological potential of the many rockshelters in the Tropical Alpine Zone was always suspected, Manachaqui Cave's promise became particularly obvious in 1986. Between the 1985 and 1986 University of Colorado-University of Trujillo field seasons, villagers from Pataz cut sod blocks from the ground surface in front of the cave in order to plug openings at the cave rear, thereby creating a dryer, more comfortable interior space to sit, cook and sleep. The sod removal exposed sherds and lithic remains that the archaeologists observed and collected at the end of the 1986 season. Cursory examination of the artifacts confirmed Manachaqui Cave's potential to yield a sequence of human occupation. Although the project would have preferred recovery of an archaeological sequence from within the continuous montane forest, budget constraints prompted 197 Lennon to target Manachaqui Cave and other rockshelters for test excavations in 1988. Despite the success of these excavations in recovering a surprisingly early sequence dating from the early Preceramic Period, the project was indefinitely suspended because of Lennon's competing commitment to archaeological research in the U.S.A. I had participated in the Rio Abiseo National Park Project from its inception as a graduate student at the University of Colorado-Boulder, and in 1989 Lennon offered me the opportunity to take over the 1990 Manachaqui Cave investigations for my dissertation research. This chapter focuses on Manachaqui Cave's immediate natural surroundings, and its present and past usage as a shelter. The first of two sections situates Manachaqui Cave within the Pataz-Abiseo study area, and characterizes its changing natural surroundings, especially the Holocene geology, climate, flora and fauna. Modern floral and faunal data serve as important baseline information with which to identify ecological zones and habitats from which the archaeological flora and fauna of Manachaqui Cave were procured. The second section includes a description of the Manachaqui Cave site complex and the network of pre-Hispanic paved roads with which it appears to be associated. f~~ctionally The problem of Manachaqui's prehistoric function or functions is more precisely defined, and a working hypothesis presented. 198 The Changing Manachaqui Valley Environment Archaeological investigations at Manachaqui Cave benefit from several years of multi-disciplinary fieldwork in and around the Rio Abiseo National Park. Young's contribution to study area zonation has already been described. Between 1985 and 1990, the Rio Abiseo National Park Research Project also sponsored geological, palynological and zoological field studies. These provide a rich set of data on Manachaqui Cave's surrounding environment lacking from most archaeological localities. Most importantly, they provide baseline information with which to evaluate evidence for past environmental alterations. The Manachaqui Cave site complex lies in the Tropical Alpine Zone at 3,625 m, 7= 42' south latitude and 77c 30' west longitude, above the south bank of the Manachaqui River (Fig. 7). The 4,300 m peaks of Cerro Suitacocha and Cueva Negra flanking the Valley to the north and south are typical of the jagged Marafion-Huallaga cordillera which Weberbauer (1945:98) describes as weathered into fantastic shapes of "towers, horns and needles." Pleistocene epoch glaciation has carved Manachaqui into a "text-book" example of a shaped valley. u- The Manachaqui River descends gradually for 13 kms from Laguna Brava at the Marafion-Huallaga divide and continues westward through small lakes and marshes to the terminal moraine at Laguna Baja. Below Laguna Baja, the 199 river enters the steep, V-shaped valleys below and combines with other streams tumbling out of the Suitacocha and Chirimachay valleys to form the Lavasen River tributary of the Marafion. Geology and Pleistocene Glaciation Despite some efforts by Wilson and Reyes (1964) to map the geologic strata in Pataz quadrangle during the 1960s, Birkeland et al. (1989) regard the bedrock of the Manachaqui valley as poorly known. Apparently Wilson and Reyes devoted minimal attention to the less accessible Tropical Alpine Zone east of the Marafion, and none at all to the Tropical Montane Rain Forest Zone east of the Marafion-Huallaga divide. Wilson and Reyes' stratigraphic profileD includes part of the Tropical Alpine Zone in the western portion of the Pataz-Abiseo study area (1964:Fig. 5). According to these studies, Manachaqui valley bedrock belongs to the Lavasen Formation and consists of Tertiary age volcanic materials. The Lavasen Formation caps both the pre-Cambrian age Marafion Complex, which consists of mica, schist, metaandesite and shale, and upper Cretaceous granodiorite intrusives that can be seen in outcrops in the Moist Montane Zone below. The Geological Map of Peru (Institute de Geologia y Mineria 1975) shows the Lavasen volcanics and the mostly metamorphic Maranon Complex as predominant bedrock strata beneath the montane forests of the Abiseo River drainage. However, the observation of sedimentary 200 formations containing beds of limestone, sandstone and slate surrounding Gran Pajaten at 2850 m attest that Huallaga basin geology remains for all intents and purposes unknown. The Lavasen Formation upon which Manachaqui Cave sits is composed of pyroclastic material, especially dacitic and rhyolitic tuffs with andesites and volcanic breccias (Wilson and Reyes 1964:48). However, Birkeland et al. (1989:56) observed mostly granites and quartzites during their work in the glacial valleys of the Tropical Alpine Zone. A pebble count from Manachaqui valley alluvium by Rodbell (1991: Fig. 2.1) showed a predominance of quartz accompanied by granodiorite and volcanic rocks. These disparate observations await reconciliation and exploration by systematic geological survey. Landforms of the Tropical Alpine Zone are largely the product of Pleistocene epoch glacial scouring and Holocene epoch glacial deposition and erosion. Terminal Pleistocene glaciation covered Manachaqui Cave's current location and extended down to elevations of 3,600 in western-facing valleys, but it was more substantial on the eastern side of the divide, reaching down to 3,200 m. Radiocarbon dating of organic material in Manachaqui Valley glacial deposits by Birkeland et al. (1989a, 1989b) documents the first retreat of glacial ice by 12,100 ± 190 B.P. Based upon additional indices, Rodbell (1991:16) estimates that deglaciation began by 13,500 B.P., and that all cirques (and, hence, all 201 valleys) were ice-free by approximately 10,000 B.P. Deposition of loess and the formation of soils followed rapidly after deglaciation, and landforms have remained relatively stable to the present (Birkeland et al. 1989:63). Holocene Environmental Change According to ONERN (1976), the uppermost slopes west of the Marafion-Huallaga divide are classed within the Tropical Subalpine Wet Paramo life zone. However, fragments of Abiseo watershed forest spill westward over the divide into the upper Manachaqui valley and surround the Laguna Brava cirque. Young's studies of forest relicts thriving in protected Tropical Alpine Zone micro-environments indicate that "the spatial distribution of these timberline forests is essentially produced and maintained by the grassland fires, which can kill unprotected trees and shrubs ... " (Young 1991:28). He estimates that, because of past and present burning practices, "the elevational limit of closed forest on most valley bottoms has probably been lowered by about 500 m" (Young 1993:277). If the hypothetical limit of closed forests reached 3,800 or 3,900 m, then Manachaqui Cave lay within tropical montane forest prior to anthropogenic deforestation. The question then becomes, When exactly was Manachaqui Valley deforested? To reconstruct the local paleoenvironment, sediment cores were extracted from Laguna Baja at Manachaqui Valley's terminal moraine by Rodbell and Hansen (1990; personal 202 communications 1993-94}. Hansen's analysis of pollen samples from the cores provides new information on upward and downward shifts in ecological zonation during the late Pleistocene and early Holocene epochs. Preliminary results from these unpublished studies indicate that Montane Wet and Montane Moist Forest life zone species advanced into the valley immediately following the glacial retreat and climatic amelioration around 12,000 B.P. A reduction in forest species accompanied by an increase in paramo species between 12,000 and 10,000 B.P. may signal a period of drier, cooler conditions. During the early Holocene (after 10,000 B.P.} Montane Wet Forest species again predominated in pollen counts until approximately 6,000 B.P. A subsequent decrease in forest species coincides with a period of greater regional aridity reported from the eastern lowlands of Ecuador (Bush and Colinvaux 1988}. Evidence from ice cores recently extracted from Mt. Huascaran in the north-Central Andes near Huaraz document similar periods of early Holocene Amazon forest expansion followed by a cooling trend and lowering timberline after 5,000 B.P. (Thompson et al. 1995}. After 4,000 B.P,. the Manachaqui core shows increased quantities of charcoal co-occurring with pollen indicators of agricultural activities that provide evidence of burning and extensive anthropogenic environmental alteration. Pollen evidence for subsequent environmental changes is not easily 203 interpreted because human activities appear to have increasingly influenced the composition of local vegetation. While timberline forest may have periodically expanded and contracted, it is reasonable to conclude that Manachaqui Cave has stood in relatively open subalpine paramo since 4,000 B.P./2,000 B.C. A record of later Holocene environmental change is lacking for the Pataz-Abiseo area. Cardich (1985} considers the effects of Eddy's (1977} postulated global climatic oscillations on the upper limits of cultivation and cultural development in the Central Andes. In Cajamarca, frost limits the present elevational limit to 3,700 m although Cardich (1985:305} noted evidence of past cultivation up to 3,900 m. In Cajamarca and the Pataz-Abiseo area, warmer temperatures may have facilitated cultivation up to 3,900 m during the Late Preceramic, Initial, Early Intermediate and Late Intermediate Periods (Ibid.: Fig. 6.14}. How severely cultivation limits would have been depressed during relatively cool intervals is more difficult to determine. Cardich's studies concur with investigations farther south in Junin Department that suggest periods of cooling during the Middle Horizon and terminal Late Intermediate Period (Wright et al. 1989; Seltzer and Hastorf 1990}. Flora and Fauna The recent floral and faunal inventories in the Rio Abiseo National Park were the most systematic and thorough 204 yet attempted in Peru's montane forests. General characteristics of Young's Tropical Alpine macro-ecological zone were described in the previous chapter, and its floral composition is detailed elsewhere (Young 1993:Table 5). Graminoids and shrubs typical of Andean paramo environments are the most conspicuous elements of the zone's flora. Trees found in protected micro-environments are Gynoxys spp. and Weinmannia microphylla. These species presently occur in closed montane forests on both sides of the MarafionHuallaga divide. Grasslands, forest remnants and other habitats recently sampled by biologists harbor a diverse fauna, much of it endemic to the eastern slopes, as well as numerous species new to science. Fauna of the Tropical Alpine and Tropical Montane Rain Forest zones in and around the Rio Abiseo National Park was inventoried by biologists in 1987, 1988 and 1989 (APECO 1988a, 1988b, 1991). Collecting by various methods took place at broadly defined locations such as lakes and river valleys along a transect beginning on the western side of the divide and terminating near 2,000 min the lower montane forest. Specialists worked independently each season to sample birds, mammals, amphibians and reptiles in a variety of habitats such as dry grasslands, floodable grasslands, rocky areas, isolated forest patches and continuous forest. Fish were sampled along the same transect. Overall methodologies have always been geared to logistical 205 feasibility in this remote region. The authors emphasize that a number of constraints precluded the collection of truly representative samples from each locality. The Manachaqui Valley was not extensively sampled although collections at nearby Chirimachay and La Plap valleys augment the total sample from the Marafion-side Tropical Alpine Zone. During his 1919 expedition to Pataz, Weberbauer (1920:6) noted the abundance of bear and deer for hunting in the "puna." Among the most conspicuous open grassland fauna today are deer, fox, armadillos and a wide variety of rodents and herpitofauna. Biologists did not tabulate species according to ecological zone although species distributions were occasionally estimated. They conclude in their final report that the large data base obtained during the three field seasons remains insufficient to adequately characterize the region's fauna (APECO 1991:52). The Manachaqui Valley and Manachaqui Cave Between 1986 and 1990, RANPRP archaeologists surveyed the Tropical Alpine Zone valleys of Chirimachay, Suitacocha, Manachaqui, La Plap and Chochos at the northwest corner of the park. The complex of archaeological features at Manachaqui Cave has been referred to in reports as M-1, the first of a series of sites recorded during systematic archaeological reconnaissance of the Manachaqui Valley's bottom and accessible lower slopes (Lennon et al. 1989). 206 The most frequent site type occurring in all but La Plap and Chochos valleys are rockshelters. In Manachaqui Valley, survey during the 1988 and 1990 seasons documented 12 sites. Six are rockshelters. Two of these (M-7 and M-8) were sampled by test excavation, but their cultural remains have yet to be analyzed. The six additional sites in Manachaqui Valley include the remains of a shrine or "huanca" (M-5); a small, naturally occurring niche between stones on the valley slope that contained a nearly complete Abiseo Style vessel (M-3); the remains of a bridge foundation (M-4); and a pair of low platforms on either side of the Manachaqui River where it breaks through the moraine that dams Laguna Baja (M-9). In 1986, Birkeland reported a band of gray and green chert crossing the pre-Hispanic road approximately 2 kms east of Manachaqui Cave. M-11 in 1990. We recorded this suspected quarry site as Also in 1990 we designated M-12 the stone masonry foundations of a bridge crossing the Manachaqui River's deep ravine below Laguna Baja. crossing this bridge (1900b:127). Raimondi describes Data from these sites will be detailed only when deemed pertinent to interpretations of Manachaqui Cave's archaeology. Manachaqui Cave The rockshelter traditionally referred to by local villagers as Manachaqui Cave is formed by an accumulation of large granite boulders that apparently fell from the heights 207 of the valley's southern slopes during late Pleistocene glacial ablation (Plate I) . These boulders sit on a mound of gravely alluvium beneath a network of gullies draining the south slope. The principal gully runs with water year- round. Because Manachaqui Cave consists of fortuitously clustered boulders, it differs from well-known highland Andean caves like Guitarrero Cave (Lynch 1980), Pacharnachay (Rick 1980) and Pikimachay (MacNeish et al. 1981) that were shaped by mechanical or chemical weathering of exposed bedrock. The boulders are clustered in such a way that large interstices between and beneath them form three small "shelters" labeled M-1A, M-1B and M-1C facing north, west and east respectively (Fig. 8). M-1B and M-1C both yielded small surface collections of sherds, but M-lA constitutes the most spacious and intensively utilized of the three shelters. It therefore received virtually all of our attention. The mouth of M-1A opens to the northeast over an ample berm (Plate II). M-1A's interior is approximately four meters wide and three meters deep with a surface area of nearly 13 m~. The shelter's ceiling is presently little more than one meter from the ground at its highest point. Spaces between the boulders leave openings at the rear of MlA and on the eastern side. Deeper within the cluster of boulders is a large open space partially exposed to rain and 208 dripping water. A 1 x 1 m excavation unit revealed shallow deposits and few cultural remains, indicating infrequent use. During prehistory, M-1A's inhabitants periodically sealed the principal or anterior shelter from wind, rain and draining water with stones and sod. The shelter's berm (or talus slope) approaches two meters in height and spreads over an estimated area of 130 m2 • It contains trash midden, stone rubble and the scant remains of eroded living surfaces. In total, M-1A's cultural deposits cover a horizontal area of nearly 145 m2 • The pre-Hispanic road and a rockshelter designated Site M-1D (formerly M-2) lie approximately 40 yards due north of M-1A. A great 9 x 7 m boulder nearly 10 m high sheltered brief occupations sampled with a 1 x 1 m excavation unit in 1990. Between M-1D and the base of M-1A's berm, an area eight meters in diameter was leveled, and two concentric stone circles or semi-circles were laid. referred to as Site M-5, Previously this feature was re-labelled M-1E in 1990 because of its presumed functional association with Site M-1. Its construction apparently occurred late in M- 1A's history of occupation. The builders cut into the base of the berm to lay the outer stone ring. Excavations uncovered a long, prismatic stone lying inside the inner circle that once stood upright in the center of the stone arrangement as the previously mentioned shrine or huanca. Other nearby and somewhat enigmatic archaeological 209 features may have been functionally associated with Site M1. On the slope due west of M-1A, a small, but prominently situated shelter commands clear views up and down the valley. Several stone alignments approximately 40 m north and west of M-1A are difficult to delineate and interpret, but suggest that portions of the slope below the M-1A were modified for unknown reasons. Across the pre-Hispanic road and situated on the valley bottom nearly 100 m northeast of M-1A is a low L-shaped wall. The long portion running east and west continues approximately 15 m, and from its western end a much shorter wall extends northward. been partially dismantled. These walls have About 100 meters west-northwest of the L-shaped wall is an isolated 2 x 1 m rectangular stone platform less than half a meter in height. Without excavation the functional assessment of these features is highly speculative. Some of them may not be prehistoric. The Pre-Hispanic Road System For decades Andean scholars have been aware that a preHispanic paved road (commonly referred to as the Inca Road or Highway) runs along the crest of the Marafion-Huallaga divide (Raimondi 1900a:36-40, 1900b:123-130; Weberbauer 1920:7; Von Hagen 1955; Vega 1979:25). This north-south artery may have once connected centers of Inca administration at Huanuco Pampa and Levanto near the modern city of Chachapoyas (Von Hagen 1955). For modern villagers the pre-Hispanic road remains the principal route of travel 210 between towns strung along the divide such as Pias, Pataz, Condormarca, Bambamarca and Bolivar. Residents of Condormarca utilize the road to run herds of cattle in the Tropical Alpine Zone, and they continue to maintain it through periodic communal labor. RANPRP scientists and personnel have habitually utilized the stretch above Pataz to travel north as far as Chirimachay. Hyslop (1984:233) observes that "stone paving is primarily a water-control devise on Inka roads." Accordingly, roads crossing the study area are more often paved than not. The main pre-Hispanic road and associated spurs now show varied states of preservation depending to a large degree upon whether or not they have been refurbished, regularly used and maintained following the Spanish conquest. The principal north-south artery has received the most attention from modern villagers, but bridges have not been repaired. The degree to which the road's pre-Hispanic width has been maintained is not clear, but it generally does not exceed three meters. Narrow drainage channels of original Inca design cut the road into segments where drainage is needed. Artificial causeways have not been observed but the road often takes advantage of glacial moraines to cross swampy terrain. The ancillary roads branching eastward toward the forest are more difficult to follow as they are typically covered by vegetation and often by soil. 211 At least two ancillary roads or road systems departed eastward from the Huanuco-Levanto artery in the Pataz-Abiseo area. Weberbauer (1920:8, map on p.13) mentions a road east of Pias that descends the valley of Pampa Hermosa to the upper Tumac River valley. While this road has not yet been examined by archaeologists, RANPRP scientists have travelled a second road system composed of two east-trending spurs that join at Laguna Empedrada before continuing down the upper Montecristo valley, ultimately to disappear under dense vegetation. These two spurs provide access to the Montecristo valley from the north and south. From the north, the spur henceforth referred to as Tragaplata-Empedrada departs from the Huanuco-Chachapoyas artery at Laguna Tragaplata and descends southeastward into the Pefia Blanca Valley. Entering the Chochos Valley, it curves due south and gradually climbs to the junction at Laguna Empedrada. From the south, the Chirimachay-Empedrada spur departs from the artery northeastward at Chirimachay [Cave]. Then it crosses a high ridge and descends into Manachaqui Valley, passes Manachaqui Cave and climbs over the Marafion-Huallaga divide to Laguna Empedrada. Remains of rectangular stone buildings are situated along the road at points of high visibility. Site C-8, Los Paredones and Mirador (Lennon et al. 1989) are "control sites" as defined by Hyslop (1984:315-316). Like the north coast sites that Hyslop views as "the best evidence for a series of control 212 points" (Ibid.) these stations are found at the entrances of valleys. They are invariably positioned to view movements along the roads from optimal vantage points. Test excavations at C-8, Los Paredones and a larger complex of rectangular structures, C-1 near Chirimachay, produced small numbers sherds exhibiting late prehistoric local and Inca styles (Lennon et al. 1989:47) suggesting that the entire complex of control installations was built under Inca direction. Indirect evidence hints at greater antiquity for the paved road. The confirmation of camelid remains in Manachaqui Cave's Early Intermediate Period strata prompts the observation that loaded caravans could not have repeatedly crossed the swampy alpine grasslands and entered the muddy montane forest without the benefit of paved surfaces like those that characterize the extant "Inca" road. Hypothetical Functions of Manachaqui Cave One of Manachaqui Cave's most striking characteristics is the integrity of its cultural deposits. Until 1985-86 they remained essentially undisturbed since the shelter was last utilized during the Late Horizon. In contrast, rockshelters along pre-Hispanic roads examined in the Pefia Blanca valley and east of Bolivar have been shoveled out and "renovated" (personal observation, 1989). The most altered rockshelters are those in Tropical Alpine Zone valleys where particular families have long held grazing rights. At M-1A, 213 the only artifacts recovered from the upper layers and clearly post-dating the Late Horizon were left by recent scientific and tourist expeditions camping en route to Gran Pajaten. Lack of Colonial Period artifacts suggest that the Chirimachay-Empedrada spur was not habitually utilized by missionaries or traders between the 16th and 20th centuries. In fact, the Montecristo valley road may descend the eastern slopes no further than the upper montane forest, although it conceivably connects to other road networks reported deep within the forest (Savoy 1970; Muscutt et al. 1994:6; Muscutt, personal communication 1994) . Modern Uses of Highland Rockshelters In recent decades, villagers from Pataz have sheltered at Manachaqui Cave while tending livestock, hunting, fishing and guiding outsiders on the three-day journey to Gran Pajaten. Between the modern towns of the Marafion-Huallaga divide, a number of well-known rockshelters serve as convenient refuges for travelers. The names of some appear in the accounts of Raimondi and Weberbauer, although none have been investigated archaeologically. In north highland Peru only Guitarrero Cave in the Callej6n de Huaylas (Lynch 1980) has been intensively studied. In the puna Andes farther south, caves and shelters above 4,000 rn are still sporadically, seasonally and even permanently utilized by herders, and occasionally by their animals (Ravines 1971:18; 214 Cardich 1973:16; Lavallee 1977:69). Flores-Ochoa (1979:104) notes that traveling herders leading alpacas in the southern highlands of Puno Department habitually camp in caves during trading excursions. During fieldwork in the Junin puna, Rick investigated a small shelter known to modern herders as Hotelmachay de Pampacancha (literally "Hotel Cave of Pampacancha") in reference to its hotel room-like appearance (Rick 1980:240) . While the name "Hotelmachay" implies overnight visits by journeying highlanders, the cave at Pampacancha is also known to herders as Pintadomachay because of pictographs within. The painting and/or pecking of rock surfaces to shape images with sacred meaning is a particular form of ritual activity performed in and around highland Andean caves during prehistory. Rick (1980:56) reports that natives traditionally left ritual "offerings of coca leaves and liquor to the huamanis (cave spirits)" at the rockshelter Pachamachay. Caves figuring in Central Andean mythology are pacarinas, considered portals to Ucu Pacha, the underworld from which humans, plants and animals once emerged to populate the surface world or Cay Pacha (Arriaga 1968:24; Valcarcel 1964:137, 144). In Pataz, residents refer to a cave in nearby Suitacocha valley as the "cueva del bruj o" (the sorcerer' s cave) . Oral traditions recall its secret use by local shamans for rituals, and local villagers still will not willingly enter. 215 Ancient Oses of Highland Rockshelters Highland Andean rockshelters may have served prehistoric functions similar to those of today. However, during the intervening centuries, Andean societies have undergone such massive structural transformations that substantial qualitative and quantitative changes in material patterning associated with functions like herding and hunting should be expected. The previous chapter documented sudden and radical 16th century changes in southern Chachapoyas local and regional economies. During prehistory, differences in material culture and animal species associated with hunting and herding activities should correspond to periodic modifications of subsistence priorities. Changing socio-political contexts presuppose differing behavioral correlates for hunting during the Preceramic Period, Late Horizon and late 20th century, and we may expect concomitant alterations in Manachaqui Cave's material patterning. Most useful interpretive accounts of hunting and herding activities during the Initial Period and subsequent periods derive from excavations of rockshelters in the Junin puna (e.g. Lavallee 1977, Lavallee et al. 1982; Matos and Rick 1981). Lavallee's studies incorporate analyses of the entire range of material and associated activity areas to generate conclusions, and therefore contribute useful 216 comparative information for this study. At a glance the relatively small faunal collections, and paltry lithic assemblages from Manachaqui Cave's earliest ceramic-bearing levels suggest activities other than those commonly performed during comparable time periods at Telarmachay. Salient differences between collections of organic remains and artifact assemblages from the Junin caves and Manachaqui Cave will be detailed in Chapter 10. Unfortunately there are no detailed studies of Andean caves interpreted as shrines and wayside stations to provide comparisons for this analysis. Manachaqui Cave as a Wayside Station Manachaqui Cave's roadside location and its modern function as a shelter for villagers traveling to and from the upper montane forest present an ethnographic analogy that serves as a working hypothesis for this study. Evaluation of this hypothesis offers a special challenge because Andean archaeologists have yet to investigate rockshelters with a comparable primary function. In fact, my search of the world-wide archaeological literature failed to locate any references to caves or rockshelters utilized specifically as wayside stations in similar contexts of regional or interregional transit. Hence, adopting the wayside station working hypothesis necessitates comparisons of Manachaqui Cave's archaeological remains to those from highland cave sites interpreted as hunting and herding camps 217 in search of analogous artifact assemblages and activity areas. Reports of excavations at Chobshi Cave, Guitarrero Cave, Telarrnachay, and Pachamachay offer serviceable data for such comparisons. The latter two especially may render useful functional analogues because they are interpreted as camps for hunting and herding, and also because they span the Preceramic Period-Initial Period transition when the occupants adopted the use of pottery. However, the comparisons drawn may provide negative correlations at best. They can not directly confirm the working hypothesis. Whether or not Manachaqui Cave was regarded as a pacarina may never be determined, but the associated huanca (M-1E) clearly indicates past ritual activities at the site during late prehistoric times. Excavations of M-1E revealed shallow deposits and a small sample of late ceramics indicating that the huanca was probably erected during the Late Horizon. That the name Manachaqui (or mama-chaki) might be translated literally in Quechua (using the popular dictionary compiled by Lira 1979) as "mother-foot" or "mother-dry" (dependent upon whether a glottal stop distinguishes ch'aki) also hints at the attachment of religious meaning to the rockshelter. The name Manachaqui may thus have signified "mother of the traveler." In the Andean literature, Inca way-stations, or tambos, represent the closest functional analogues to Manachaqui Cave. However, their free-standing masonry construction and 218 the internal diversity consonant with their complex administrative functions (Hyslop 1984:279-280) hinder comparisons to Manachaqui Cave. The Tambo de Tunscucancha in Huanuco, the most intensively studied to date, consists of 10 or 11 buildings with artifact distributions suggesting domestic and ritual functions, as well as status differences (Morris 1966). Less elaborate complexes like Callangate and Frailetambo have yet to be thoroughly investigated. Rowe (1946:270-271) reports that the Inca tightly regulated travel and trade. Pre-Inca interaction in southern Chachapoyas was undoubtedly governed by other sets of constraints and opportunities left for archaeologists to discover. The basic technologies involved in trade and travel were probably quite simple. Personal belongings and goods for trade may have been wrapped in a mantle and slung onto the traveler's back as Rowe (1946:237) describes. Llamas carried up to 100 lbs. in cloth saddle bags (Ibid.:239) or perhaps net bags. Such technologies are probably as old as textile technology and camelid domestication. The former was in place by the Late Preceramic Period (Bird 1963; Adovasio and Maslowski 1980; Grieder et al. 1988), but intensive utilization of domesticated camelids in north highland Peru may only date to the Chavin horizon (Miller and Burger 1995). Material remains of transport technologies may be difficult to identify. Textile remains 219 in particular do not normally preserve in humid tropical archaeological contexts. Manachaqui's Ethnographic Analogy Consideration of how Manachaqui Cave is utilized today aids the generation of a working hypotheses to model Manachaqui's ancient functions. literature search, Again, a world-wide this time for ethnographic descriptions of modern cave and rockshelter utilization, failed to locate useful analogies for Manachaqui Cave. Shanidar Cave in northern Iraq is a well-known example of a modern cave habitation, but it is not functionally comparable to Manachaqui as its enormous interior seasonally shelters a community of approximately 45 Kurds, their huts and even their livestock (Solecki 1979). Comparisons of Manachaqui Cave's present surface activity areas to those within deeper strata should help support or refute a wayside station working hypothesis. Today up to eight individuals sleep huddled together for warmth on a bed of ichu straw against the eastern interior wall of the shelter where they find shelter from damp easterly winds (Fig. 9). Occupants maintain a hearth for cooking and/or warmth against the western wall and close to the cave mouth where the same winds carry smoke out into the open. Young spent long periods of time traveling with Pataz villagers through the Tropical Alpine Zone at the western 220 border of the Rio Abiseo National Park, and reports most of the following observations (personal communication, 1994). Nowadays, a trip to the "jalca" (Tropical Alpine Zone) social occasion greatly enjoyed by the men. is a Young noted that meals are invariably prepared prior to departure, thus obviating the need for cooking fires, collecting or transporting fuel and the great patience required for boiling food at high altitudes. On the first day, several varieties of small previously-boiled potatoes are eaten. Sufficient portions of cooked guinea pigs (or cuy) are brought so that each traveler has his own portion, and chili peppers are added to most foods. Young's description agrees with ethnographic observations from the northeastern Andes reported by Morales (1995). According to Morales (1995:131), pre-prepared guinea pig serves as the usual travel "fiambre" or "cooked food provision." After consuming the guinea pig on the trail, its skull may be carried horne to assure successful propagation of the household livestock, or it may be offered with coca to the mountain spirits for protection from storms and other travel hazards. Morales also observes that travelers may keep the right foot to carry as an amulet for good luck and to give strength for walking. Parched maize (or cancha) and/or nuna beans accompany the main fare. Gade (1975:126) reports that cancha is habitually carried as "trail food" in Cuzco. Cobo's (1890- 221 93, I:343; Coe 1994:222) chronicles of traditional native customs note that portions of cancha, or cancha flour mixed with cold water, were carried as travel meals. Nowadays, food is typically carried in a pair of metal or plastic bowls lashed one over the other and transported in a small woven saddle bag on mule-back, or over the traveler's shoulder. Travelers also catch trout and hunt white-tailed deer (outside of the park boundaries) if time permits. we observed that they leave nothing when they abandon Manachaqui Cave except matted straw and ashes from a small fire maintained for warmth. Andean foxes would be expected to scavenge any discarded bones or other food remains. Working Hypothesis In order to evaluate the working hypothesis that Manachaqui Cave functioned as a wayside station, negative evidence for hunting and herding functions may be combined with ethnographically and theoretically-derived sets of expectations for archaeological material patterning. First we would expect that most, if not all, artifacts would have been produced elsewhere. "imports." In a sense, they would be At a wayside station frequented by itinerant individuals and groups, the range of functional types might constitute only a fraction of the total range of functional categories produced or used at home. In view of the ethnographic analogy, we might hypothesize that large cooking vessels would be rare or absent from a wayside 222 station assemblage. When ceramics occur, small sizes should be characteristic of culinary pottery. Portability would be a primary concern, and small vessels also suffice to service fewer individuals. Vessel types associated with special ritual activities should be rare or absent. Such types include decorated bottles presumably utilized for ritualized consumption of corn beer (chicha) (Burger 1992:155, 166). Identification and quantitative analyses of Manachaqui Cave's organic remains can provide insights into the shelter's ancient functions. Prehistoric "trail food" might consist of grains, legumes or tubers either pre-cooked or easily prepared at the site. While travelers presumably hunted and trapped game for consumption en route, a wayside station's deposits would not include the great quantities of bone typically unearthed at hunting and herding camps (cf. Miller 1979). Hunting and herding refuse at Chobshi Cave, Guitarrero Cave, Telarmachay and Pachamachay included many kilograms of bone material, mostly from large game animals like deer and camelids. At a wayside station we might instead expect to see the remains of small animals or portions of larger animals that could be conveniently packed. Large game animals like deer may have been preferred targets, but most of the meat would likely have been prepared for transport and carried away, perhaps to feed families at home. Consequently, Manachaqui Cave's refuse should reflect disposal of small portions of large 223 animals consumed at the shelter. In her analysis of Manachaqui Cave's botanical remains, Pearsall (Appendix F) lists a set of assumptions governing expected patterns of food charring and deposition within a rockshelter setting. Based on these assumptions, and the known regional distributions of ecological zones and plant habitats, she suggests some indicators for distinguishing the site·' s function. The first concerns the abundance of charred food remains relative to the abundance of wood charcoal, and aims to determine the degree to which cooking of plant foods was done at Manachaqui. Food remains should be more plentiful relative to charcoal at habitations where food is inadvertently charred during repeated cooking activities. Disproportionately large quantities of charcoal would result from fires built for warming rather than cooking. The utility of this indicator rests upon the assumption that travelers typically carried pre-prepared food. Changes in charred food:wood ratios may signal shifts in Manachaqui's function. Of course warming fires could fortuitously char small amounts of food. Additional indicators suggested by Pearsall concern the quantity of food items gathered or grown locally relative to the quantity of "exotic" food items brought from distant ecological zones. Changes in local:exotic food ratios through time may indicate shifts in Manachaqui's function as a habitation (relatively high values for local foods) or 224 wayside station (relatively high values for exotic foods). Caveats to be considered include the fact that hunting and plant gathering in the rockshelter's vicinity may have provided fresh food when distances to be traveled were great. Also, the tending of some high altitude crops could have routinely occupied Manachaqui Cave's habitual users. A third indicator proposed by Pearsall relates to the abundance of corn kernels relative to cobs. probably "exotic" to Manachaqui Valley. Corn is A higher proportion of kernels relative to cobs may indicate that corn was shelled and parched at the site. Pearsall hypothesizes that shelling and parching may more likely occur in habitations. The higher percentage of cobs relative to kernels may indicate a tendency of travelers to consume corn directly from the cob. However, the inadvertent carbonization of cancha brought to Manachaqui Cave by travelers carrying preprepared food could result in misleading ratios and erroneous interpretations. The results of this study suggest that the kernel:cob ratios are not effective indicators of Manachaqui Cave's function, but they are included so that readers may judge for themselves. All of the ratios are susceptible to alteration by a variety of outside factors as well as site formation processes. Regional subsistence shifts with the introduction and abandonment of certain plant foods more or less likely to be preserved during Manachaqui's occupational sequence 225 potentially alter ratio values, as would changes in food preparation and consumption practices. Therefore, the changes in this and other ratios must always be evaluated in conjunction with other lines of evidence. Furthermore, Manachaqui Cave may have served multiple functions at any given time. The various kinds of material patterning hypothesized above presuppose ideal situations that may seldom conform closely to prehistoric reality. Summary The two most important problems addressed within this chapter involve the identification of changing paleoenvironmental contexts during periods of Manachaqui Cave's intensive prehistoric use, and the determination of what those uses were at given prehistoric moments. Interpretations of function are to some degree dependent upon knowledge of corresponding paleoenvironmental contexts if subsistence-related functions such as herding and hunting are to be considered. Timberline forest environments may be appropriate to sustain game such as white-tailed deer, paca and armadillos, but herds of llamas and alpacas are best adapted to open puna grasslands. Paleoenvironmental reconstruction suggests that open grasslands predominated in the Manachaqui Valley during the time period under study, although fragmented and continuous upper montane forest habitats always remained within reach by short (no more than a half-day) excursions from 226 Manachaqui Cave. Timberline forest probably encroached on the M-lA locality periodically under favorable climatic conditions, and in the absence of sustained burning. Manachaqui Cave's ecotone location offered favorable circumstances for hum~~ utilization of a great diversity of resources, yet poor conditions for the sustained exploitation of single species. The working hypothesis of Manachaqui Cave as a wayside station was sketched in the preceding paragraphs. M-lA's modern usage provides one basis for this hypothesis, and the presence of the pre-Hispanic road provides additional plausibility. Also, the distribution of southern Chachapoyas settlements in the highlands and montane forests described within the preceding chapter invites the assumption that people frequently traveled between the two ecological zones. As a node in east-west transit, Manachaqui Cave conceivably sheltered migrants, pilgrims, trading expeditions, war parties, hunters and herders during the periods under study. Because this study of function has few if any theoretical predecessors, the data must be interpreted through series of expectations or multiple working hypotheses, each taking environmental and historical contexts into consideration. However, precise identifications of Manachaqui Cave's prehistoric functions are not necessarily pivotal to evaluations of other hypotheses regarding Andean-Amazonian interactions addressed 227 within this thesis. The dearth of ancient settlements in these Tropical Alpine Zone valleys permits the conclusion that many individuals sheltering in Manachaqui Cave arrived from one side of the Marafion-Huallaga divide for the express purpose of crossing to the other. CHAPTER 5 EXCAVATIONS AT MANACHAQUI CAVE As noted in previous chapters, excavations at Manachaqui Cave were undertaken in 1988 and 1990 in order to remedy the dearth of detailed cultural sequences in the northeastern Peruvian highlands. As repositories of sequentially stacked primary contexts, caves and rockshelters have served as "the essential data source for the construction of cultural stratigraphies and histories in many regions of the world" (Straus 1990:265). Like sites with architecture, cave sites with occupation floors may provide diachronic evidence of economic or other activities. Unlike sites with architecture, it is less probable that cave floors were altered subsequent to deposition by renewed construction and the addition of potentially confusing mixed fill brought from elsewhere. Under ideal conditions, the stratigraphic order encountered reflects consecutive stages of human occupation. But in reality, caves and rockshelters present problems such as "imperceptible depositional hiatuses, stratigraphic disturbances" and "high rates of palimpsest deposition" (Ibid. :256). Fortunately, resulting interpretive ambiguities can be minimized by implementing certain methodological adjustments which will be outlined 228 229 within this chapter. Analysis of Manachaqui Cave's stratified ceramic remains provides the basis for the chronological sequence that serves as the core of this thesis. Andean archaeology has a long history of reliance upon pottery styles as indicators of temporal and spacial relationships (Adams et al. 1978:492). At Manachaqui Cave, ceramics constitute by far the most abundant and potentially informative cultural remains. The primary depositional contexts of Manachaqui's strata offer opportunities to define archaeological phases with a high degree of temporal precision. The regional data base compiled by other investigations affords possibilities for cross-dating and the identification of interaction networks. Hence, this thesis emphasizes Manachaqui ceramics, although it also integrates information rendered by analyses of lithic and other cultural remains. Cave and rockshelter sites frequently yield wellpreserved organic remains that are rapidly weathered or destroyed by exposure at open sites. The dry soils within Guitarrero Cave (Lynch 1980) and the Ayacucho caves investigated by MacNeish et al. (1981) contained a diversity of botanical and faunal remains, and even relatively intact assemblages of Preceramic Period basketry and textiles. Excavations at Pachamachay (Cave) in Junin (Rick 1980) encountered damp soils in which carbonized macro-botanical remains were useful for identification and sourcing by 230 ecological zone and habitat (Pearsall 1980). Like Pachamachay's, Manachaqui Cave's soils are humid. Nevertheless, bone and carbonized botanical remains are preserved throughout the site, although the interior deposits yielded larger, less weathered samples. The study of Manachaqui Cave's stratigraphic record of organic remains presents an unprecedented opportunity for the reconstruction of subsistence activities and changing patterns of cultural interaction in the humid northeastern Peruvian Andes. This chapter begins by outlining the techniques employed for 1988 and 1990 excavations at Manachaqui Cave. A second section provides descriptions of the stratigraphy, occupation floors, and associated hearths and architectural elements. The radiocarbon evidence from these contexts permits the bracketing of archaeological phases with dates. Interpretations of site formation are offered where pertinent to the understanding of Manachaqui's cultural sequence. A third section details changes in hearth morphology, and the final section briefly describes analytical methodologies utilized for the study of Manachaqui's artifacts and organic remains. The 1988 Excavations The 1988 test excavations at Manachaqui Cave codirected by Lennon and Segundo Vasquez Sanchez aimed to determine the depth of the shelter's deposits, evaluate material preservation and assess the site chronology. Work 231 took place during 21 days between February 20th and March 12th. This is the rainy season in the eastern Andes, and working conditions were therefore less than ideal. On many days it was necessary to work under tarpaulins, although these were removed in order to examine and draw features and stratigraphy. In order to expose a cross-section of the site's strata, as well as to sample both the interior and exterior deposits, a trench comprised of six 1 x 1 m2 units (Units 16) was situated to extend outward from the rear of the shelter along the eastern interior wall (Fig. 10). Two additional units aligned with the trench were excavated near the base of the berm (Units 7 and 8), and a single unit (Unit C-1) was placed downslope beyond the berm to examine culturally sterile soil deposits from the surrounding slopes. Digging by trowel then proceeded utilizing 5 ern arbitrary levels measured from a datum assigned the arbitrary value of 100.0 m. Levels for all units except 7 and 8 maintained a strictly horizontal orientation irrespective of surface slope. In Units 7 and 8, each arbitrary level did maintain the slope and undulation of the berm surface in hopes of approximating tilting sub-surface cultural strata. Unlike the level numbering system employed for Units 1 through 6 which reflects absolute elevation with reference to a fixed and independent datum, the system utilized for Units 7 and 8 differs by simply assigning level 232 numbers in relation to the surface beginning with level number one. Soil samples were systematically collected from each level of selected units prior to passing all of the excavated dirt through 1/4" or 1/16" screens. The excavation of Units 1 through 8 provided information with which to formulate plans for future work. First, Manachaqui Cave was found to contain cultural deposits to a depth of one meter within the shelter interior, and two meters in the deepest part of the berm. The berm deposit appeared equally divided between strata with ceramic period remains, and layers with Preceramic Period remains beneath (Plate IV). Second, there were few visible strata extending throughout the site with which to differentiate occupations and/or phases. For the site as a whole, visible stratigraphic distinctions consisted of differing shades of black and dark brown soil, the presence or absence of color mottling, and varying sediment size. The more visibly complex interior stratigraphy showed horizontally distributed bands and lenses of charcoal, ash and a yellowish brown soil as ephemeral indications of layered floors, hearths and activity areas. The berm excavations yielded fewer visible strata, but changing patterns in artifact content became obvious as the excavations proceeded. During 1988, preliminary analysis of the excavated pottery confirmed that, despite the paucity of visible 233 stratigraphy at Manachaqui Cave, a temporal sequence of pottery styles does remain intact. The principal methodological concern resides in how to identify and isolate the temporal components. The method employed here to distinguish the cultural sequence resembles that utilized by Lathrap for ceramic complexes of the Central Ucayali site UCA-6 that he ultimately assigned to the Early and Late Tutishcainyo Phases (Lathrap 1962). During excavation, we observed that particular pottery attributes were characteristic of some levels and entirely absent in others. The quantity of lithic artifacts per level also varied in a suggestive manner. Consequently, the frequency of those pottery attributes were charted by excavation unit and level in order to illuminate potentially discrete vertical distributions. Lathrap (1962:46-47) refers to this utilization of frequency distribution as a "macrochronology." Further refinement of each isolated cultural component produces what he terms a "microchronology." Pottery attributes chosen for a macrochronological study of Manachaqui Cave's cultural sequence were sherds with 1) basal, medial and shoulder carinations and/or applique ribs, 2) fine-line incised decorations and 3) fine kaolin pastes. Observations from ceramic sequences in neighboring northern Andean regions suggest that these isolated components most likely pertain to the Initial 234 Period, Early Horizon and Early Intermediate Period respectively. The resulting macrochronological distributions appear in Appendix D. Units 4, 5 and 6 show the clearest separation of temporal components, while Units 1, 2 and 3 demonstrate far less. This can be explained in part by the tendency of the 1988 horizontal, arbitrary levels to approximate the orientation of the horizontal stratigraphy within Units 4, 5 and 6, and their tendency to cross-cut the sloping stratigraphy within Units 1, 2 and 3. A fourth temporal component within the uppermost layers at Manachaqui lacks any of the three ceramic attributes, and probably corresponds to the Late Intermediate Period and/or Late Horizon. A primary objective of the 1990 excavations at Manachaqui Cave was to confirm, or to modify and complete this tentative and sketchy macrochronological sequence, and ultimately to produce a more refined chronological sequence. The 1990 Excavations The 1990 investigations at Manachaqui Cave (directed by the author) took place during a period of 38 days between September 25th and November 4th. The project's start was delayed almost two months by Peru's major economic crisis and the reduced purchasing power of the dollar. Consequently, we were forced again to work during the rainy season. Again we were obliged to use tarpaulins for protection. Troweling through black soils under tarpaulins and leaden-gray skies challenged everyone's eyesight, and 235 the illumination provided by a Petromax gas lantern greatly facilitated the identification of soil details in the shelter interior. The preliminary analyses of Manachaqui Cave's 1988 pottery sample suggested that the east-side interior deposits were highly disturbed. A concavity in the east- side ground surface matched by undulations in the underlying strata indicates that soil may have been repeatedly scooped out or moved around, probably to prepare sleeping surfaces. Because the west profile of the test trench showed evidence of more intact occupation floors, it was decided to devote special attention to careful excavation and sampling of the interior's west side. The 1988 excavations had produced small quantities of bone and charred botanical remains that promised useful economic information if larger samples from less disturbed contexts could be obtained. Utilizing the same 1988 grid system, additional 1 x 1 m units were selected for excavation. With the goal of understanding deposition and site formation, we proceeded to expose stratigraphic profiles by excavating trench-like sets of 1 x 1 m squares oriented east to west. A strategy of maximizing the number of exposed profiles has been recommended by Straus (1979, 1990). This technique facilitates correlations and comparisons of stratigraphy across the site, although admittedly we were simply searching for more useful stratigraphic evidence than that 236 which the 1988 excavations had exposed. The trench-like sets consisted of Units 12-13, 18-23 and 29-32 (Plate III). Units 35, 38 and 39 connected 1988 Units 1-6 and 7-8. The subsequent opening of additional squares completed the sampling and exposed architectural features. In order to discuss procedures and results, it becomes expedient to differentiate between the shelter interior and the exterior berm deposits with the labels Sector A and Sector B respectively (Fig. 8). In Sector A, the exposure of relatively intact strata visible in Unit 12's and 13's north profiles (Plate V) led to the decision that Units 14, 15, 16 and 17 should be excavated simultaneously as a 2 x 2 m block with the hope of uncovering successive occupation surfaces. It was suspected that hearths and evidence of economic activities might be found on the west side of the interior just as they occur on the west side of the site surface today. Thus, the 2 x 2 m block was excavated by stratigraphic levels or occupation surfaces where possible, and by 1, 2 or 5 em arbitrary levels where neither strata nor surfaces could be detected. Stratigraphic excavation of the block relied on tactile as well as visible observation because occupation surfaces were invariably more compact than surrounding layers. Additional information on Sector A excavations and the cultural stratigraphy will be offered later within this chapter. The decision to excavate unit levels that maintained 237 the slope and undulations of the ground surface constitutes an important change in 1990 excavation procedures for both sectors. In 1988, only Units 7 and 8 were excavated in this fashion. This technique was intended to diminish the likelihood that sloping sub-surface strata would be crosscut and "mixed." Where it was discovered that the slope of sub-surface strata did not match the surface slope, we attempted to adjust excavation levels accordingly. Arbitrary 5 em levels were most frequently employed in Sector B, although stratigraphic levels were preferred when they could be discerned during excavations. Readers should be advised when consulting the appendices that, because levels within some units were adjusted to accommodate natural strata, floors and localized features, they do not always correspond to similarly numbered levels in other units. Two methods were utilized to sample organic remains from both sectors. First, unscreened soil samples were systematically collected from selected units. A second method first implemented in 1988 involved passing soil samples through 1/16" screens with water at the site. The stream located a short distance west of the rockshelter supplied running water for this procedure which was applied in 1990 only to samples from Sector A's occupation surfaces. After surfaces were exposed, mapped, photographed and bulk sampled, they were "peeled off" by troweling and then passed 238 through the 16" screen at the stream to remove fine sediments. The remaining screen contents (both the heavy fraction and charred botanical material) were then bagged. In Trujillo, selected screened and unscreened soil samples were processed by water and chemical flotation techniques after removing charcoal for potential use in radiocarbon dating. The heavy fractions typically contained stones, lithic detritus and bone fragments that were sorted and bagged separately. The contents of hearths encountered during excavation were bagged in their entirety. In Trujillo, these samples underwent water and chemical flotation after removing some wood charcoal for dating. Heavy fractions were then sorted and bagged by class of material. A final category of soil sample was collected from each stratum for laboratory evaluation of pH value and sediment characteristics. Those results will be presented below. Site Formation and Stratigraphy This section will provide descriptions of Manachaqui Cave's stratigraphy and depositional sequence. As previously noted, it is helpful to conceptualize the shelter interior and the exterior berm separately as Sectors A and B. Sector A contains Units 1-3, 9-17 and 25. The division separating Sectors A and B approximates the shelter's drip line. The stratigraphy and the ceramic period cultural sequence generally repeat themselves across the site. 239 Excavations in both Sectors A and B encountered occupation surfaces, hearths and architectural features, yet the two sectors represent differing behavioral contexts. Sector A served as Manachaqui Cave's principal focus of activity, while Sector B in effect constitutes Sector A's trash midden. Much of its broken pottery and lithic detritus originated in the shelter interior where relatively clean living surfaces were preferred and maintained. The organic remains and activity areas preserved in Sector A offer the most useful information with which to reconstruct economic activities at Manachaqui Cave. Sector B's deposits have suffered weathering by exposure to rain and moving groundwater, but they remain less disturbed by trampling, hearth construction and other post-depositional cultural activities. cultural sequence. Sector B contains the most intact Together, sectors A and B offer complimentary information with which to approach problems of Manachaqui Cave's occupational sequence and changing functions. Site Formation The following constitutes a general description and interpretation of Manachaqui Cave's deposits in terms of their constituents, and their culturally and naturally induced transformations. The term "transformation," coined by Schiffer (1983), refers to the cultural and natural processes which affect archaeological remains and their 240 distributions following initial deposition. At Manachaqui, both sectors were subjected to characteristic sets of transformations, although some particular transformations were common to the archaeological deposits throughout the site. These are discussed following a brief description of Manachaqui's depositional processes. The sediments comprising Manachaqui Cave's deposits may be conceived of as both "endogenous" and "exogenous" (Farrand 1985:21). Endogenous sediments originate in the shelter's ceiling and walls as the result of frost action and collapse. They frequently occur in Sector A's deposits as stones and gravel. Exogenous sediments include those introduced by wind, water and animals (including humans). The natural sediment matrix in Manachaqui Cave, and the principle constituent of post-glacial soils in Manachaqui Valley, is loess in the form of dark-colored silt and clay sediments. On Manachaqui Valley's Pleistocene-age moraines, Rodbell (1991:93-95) reports caps of loess 20 to 40 em thick overlaying glacial till. Our excavation of Unit C-1 a few meters below the rockshelter revealed 15 em of dark, finegrained sediments between a five em thick humic layer above, and bedrock below. Small amounts of inorganic sediment must have been inadvertently introduced to Sector A by human carriers. Also a light brown silty soil layered among the floor deposits appears to have been intentionally laid down. It may have been brought from the stream bed west of the 241 site. Sector A, the shelter interior, contains a complex stratigraphy comprised of thin and tightly stacked layers corresponding to prepared and unprepared occupation surfaces or "floors." These are occasionally interspersed with layers of disturbed and undifferentiated floor remains. The first series of floors identified at Manachaqui Cave was found in Units 12 and 13, and designated Floors 1 through 4. Floor remains also appear in Unit 25. A careful excavation of the adjacent 2 x 2 m block {Units 14-17) revealed the complex series of occupation floors identified by the letters A through Z, and AA through GG from top to bottom. Sector A's archaeological deposits range in thickness from 0.80 min the rear to 1.1 m near the mouth. That the shelter's interior and exterior stratigraphic and cultural sequences maintain good vertical correspondences throughout the ceramic period suggests that Sector A has never been subjected to massive soil disturbances resulting from attempts to clean or renovate the interior. Sector B's deposits range from 1.1 m deep at the mouth of the shelter to nearly 2.0 mat the crest of the berm, and 0.90 m deep in excavations near the berm's foot. In the deepest portions of Sector B, the ceramic and pre-ceramic deposits share a nearly equal thickness of approximately one meter. Sector B's deposits contain less organic material than Sector A's {1.54% of total bulk sediment for the 242 average level in Unit 36 compared to 4.13% in Unit 15 levels), but a much greater volume in artifacts. The five hundred cubic centimeters of sediment within each of Sector B's arbitrary levels frequently yielded over 250 sherds, or at least one sherd for every two cubic centimeters. Thus, the thickness of Manachaqui's ceramic period strata, and the height of the berm in general, are primarily functions of artifact, rather than soil deposition. Straus (1990) emphasizes the importance of understanding the rates of sedimentation at caves and rockshelters for stratigraphic interpretation. Utilizing data from Unit C-1 and Rodbell's (1991) observations regarding local rates of loess deposition, a clearer understanding of Manachaqui Cave's site formation processes may be approached. A radiocarbon date of 9700 ± 270 B.P. (Birkeland, Rodbell and Short 1989:58) from organic materials lying atop "glacial or fluvial sediments" located less than half a kilometer up-valley from Manachaqui Cave provides a minimum estimate for glacial retreat above the shelter. If we assume that the 20 em of loess and humus have accumulated on top of Unit C-l's gravel deposits over the following span of ten thousand years, then we may derive an average deposition rate of 2 em for each thousand years, or 2 mm for each one hundred years. Certainly the simplicity of this formulaic description of Manachaqui Cave's natural sedimentation masks the 243 dynamics of a complex depositional process. For example, early Holocene flooding would have governed both erosion and deposition of the lowest strata. deforestation ~n Also, burning and the Manachaqui valley probably accelerated loess deposition after 4,000 b.c. Regardless, it is safe to assume that the tempo of cultural deposition at the shelter outpaced that of natural deposition during most of Manachaqui's stratigraphic history. As Straus predicts (1990:269), this results in the close juxtaposition and complex stacking of activity residues characteristic of Sector A. Where occupation floors can be isolated, they are difficult if not impossible to interpret in terms of longevity and intensity of use. Clearly both factors operated to determine the volume and density of Manachaqui Cave's deposits, but their relative importance remains obscure, at least at present. Furthermore, with such a slow natural sedimentation rate it should not be surprising if we cannot stratigraphically identify occupational hiatuses lasting several hundred years. may seem directly juxtaposed. Pre- and post-hiatus strata At Manachaqui Cave the processing of series of radiocarbon dates as recommended by Straus facilitates interpretation of the changing dynamics of site formation. Water was an important agent of sediment transport at Manachaqui Cave. Having seeped into the shelter interior through the numerous interstices between boulders, it has 244 acted as agent of both sediment deposition and postdepositional transformations. Prior to human efforts to seal the openings in Sector A's rear and east walls, runoff generated by periodic heavy rains would have periodically washed, and perhaps scoured, the shelter interior. The earliest evidence for attempts to seal the interior consists of piled field stones. Modern practices suggest that sod brought in from outside was also utilized to plug the openings. This procedure would have transferred exogenous soils (and perhaps cultural materials) into the shelter interior. Subsequent erosion of the sod plugs would have introduced secondary deposits into Sector A's otherwise primary depositional sequence. This kind of process may be responsible for much of the depositional mixing evident in Sector A, and especially in Unit 11 which abuts a gap between two boulders comprising the interior east wall. While no evidence has appeared to suggest massive efforts to clean or renovate the shelter interior by the removal of soils, there are indications that it was kept relatively free of heavy trash accumulations. Sherd counts for units within Sector A are considerably lower than those for Sector B. Laboratory retro-fitting of Sector A and B sherds suggests that trash such as broken routinely pitched out onto the berm. potte~J was Other culturally induced transformations at Manachaqui Cave include sequential episodes of activity area preparation and 245 trampling. The preparation of hearths and sleeping areas certainly caused the translocation of sediments and associated cultural remains, from earlier strata. typically pulling deposits up Trampling is the most ready explanation for the small size of potsherds at Manachaqui Cave, particularly in Sector A. Other disturbances include a large basin-shaped pit that was·dug into Unit 16's upper strata, and a tubular vertical hole, perhaps from a post, found in the northern half of Unit 13. Leo {personal communication 1992) reports that rodent species frequenting Manachaqui Valley do not burrow, although they do take advantage of pre-existing burrow-like spaces. Several burrowing centipedes were encountered during excavations at Manachaqui Cave. Most were in Sector B, but the extent of disturbance that they may have caused is difficult to evaluate. Stratigraphy During the excavation of Manachaqui Cave's archaeological deposits, several factors hindered the precise identification of individual strata. All of the strata vary subtly in chromas and hues from dark brown and gray to black, and difficulties distinguishing these were aggravated by the aforementioned lighting problems. Color variation within strata further complicated characterization. In many cases sediment grain size and the presence or absence of color mottling served as defining 246 characteristics. The high number of large stones impeded stratigraphic differentiation in the lowest levels, especially in Sector B. Soil sediment analyses performed at the National University of Trujillo Department of Chemical Engineering document stratigraphic change by measuring percentages of clay, silt, sand and gravel (Tables 1 and 2). Samples analyzed from Units 15 and 36 show rising quantities of gravel and diminishing amounts of silt and clay with increasing depth. Soil pH was also measured for each stratum. Manachaqui Cave's deposits are divisible into four principal stratigraphic units. through 4 (Figs. 11-21). These are labeled Strata 1 Stratum 4 is bedrock. In certain locations Strata 1, 2 and 3 are sub-divisible by visual criteria. Stratum 3 corresponds to Manachaqui's Precerarnic Period occupation and its gravely sediments throughout the site confirm that water was an important depositional and transformational agent. Stratum 3 and its contents will not be described in detail, except as necessary to illuminate the sequence of site formation. The upper, ceramic-bearing Stratum 2 and Stratum 1 provide the data for this study. The following paragraphs constitute descriptions and interpretations of Manachaqui Cave's stratigraphic units. Strata and sub-strata present throughout the site are detailed, but floors, hearths and architectural features are described by sector. Figures 22 and 23 illustrate the 247 relationship between the stratigraphy and arbitrary levels in typical Sector B unit profiles. The reader may use these in conjunction with the pertinent macrochronological distributions presented in Appendix D to locate temporal components. Radiocarbon evidence is also introduced. Stratum 4 Bedrock at Manachaqui Cave is a yellow-brown color and consists of large rocks embedded in a compact, gravely soil matrix. Many loose rocks lie directly on the bedrock soil, especially near the foot of the berm. Most of these rocks may have been removed from the shelter interior and tossed downslope during the earliest occupations. Stratum 3 Stratum 3, the gravely sediments overlaying bedrock and containing Preceramic Period remains, is distributed throughout both Sectors A and B. However, the densest cultural deposits left by Preceramic Period occupations at Manachaqui (mostly lithic refuse) lie within the thicker Stratum 3 layers of gravely soil at the bottom of Sector B, the berm. Although Stratum 3 within the shelter contains some lithic material, there are virtually no diagnostic pieces to compare with those from the artifact-rich Preceramic levels in Sector B. Radiocarbon evidence presented below suggests that only the final centuries of the Preceramic Period are represented in the shelter 248 interior while the thicker berm deposits yield much earlier dates. The differences between these two depositional contexts can be explained as the result of flooding and precipitation runoff that repeatedly washed Sector A's earliest deposits into Sector B until the shelter's users began habitually to divert water away from the interior. Therefore, Sector B's Stratum 3 deposits, especially the lowest and most gravelly portions, should be regarded as mixed with secondary deposits from Sector A. Of course we might also expect Manachaqui's users to perform wasteproducing activities such as stone tool working outside, rather than within the confined sleeping and cooking areas. Toward the rear of Sector A, Stratum 3 can be subdivided by a subtle difference in color into Stratum 3A and Stratum 3B. While Stratum 3A is black (lOYR 2/1 black), Stratum 3B has a brownish chroma (7.5YR 3/2 dark brown), probably from mixing with the bedrock soil. The distinction between the two strata blurs and becomes imperceptible toward the shelter mouth, but again becomes evident outside of the shelter. A report of the cultural content of Stratum 3 falls beyond the purview of this thesis. Stratum 2 As previously noted, Stratum 2 contains the organic and inorganic remains of first importance for this study. Its deposits pertain to cultural occupations from the terminal Preceramic Period through the Late Horizon. Stratum 2 249 features a higher silt to gravel ratio than Stratum 3. However, like Stratum 3, it can be subdivided on the west side of the shelter interior into a nearly black Stratum 2A above a dark brown Stratum 2B. It is also sub-divisible north of Unit 6, beyond the crest of the berm. Although Strata 2A and 2B cannot be distinguished from one another at the deepest part of the berm, a layer of angular stones ranging from five to ten centimeters in length or diameter lies just beneath the projected 2A/2B interface and probably represents an eroded occupation floor or surface. This layer is most clearly visible in the east and west profiles of Units 4, 5 and 6 where lies between 70 and 75 ern below the ground surface. An associated hearth (Feature R-4) will be described below. Why the stones should appear so concentrated on this surface remains unclear. Stratum 2A varies from 35 to 45 ern in thickness and has the highest percentages of silt and clay of the Manachaqui strata. Where it can be isolated, it contains Early Intermediate Period materials. Within Stratum 2B the percentage of gravel steadily increases with depth. Its thickness ranges from 45 to 55 ern, and its contents include the late Precerarnic, Initial Period and Early Horizon materials. Stratum 1 Stratum 1 is the uppermost of the Manachaqui Cave 250 strata, consisting of grasses, rootlets and humus. Associated cultural remains pertain to the Late Intermediate Period and/or the Late Horizon. Stratum 1 is only discernible around the perimeter of the berm because a thick layer of sod from the crest was removed during the villagers' 1986 efforts to "clean" the site. Some of this sod was cut into bricks and crammed into the open interstices at the rear of the shelter. Soil was also tossed over the edge of the berm to form the layer designated Stratum 1A. Stratum 1B constitutes the original pre-1986 superficial layer. find Stratum 1 intact. Only in Units 18 and 19 did we Stratum 1 can be distinguished from Stratum 2 because of its organic content rather than because of a markedly different color. In stratigraphic profile it has been observed that this layer dries more rapidly than Stratum 2. Archaeological Features, Floors and Hearths Within this and the following section reporting Sector B's cultural deposits, several classes of non-portable artifacts will be described. walls. Most are floors, hearths and The floors within Sector A are labelled as such. Labelled as "features" are several prehispanic walls, isolated fragments of partially preserved floors and a single hearth in Sector B, and a drainage duct in Sector A. To help the reader locate floors and features in threedimensional space, pertinent unit and level numbers are 251 offered, as well as the stratum number and depth below ground surface (bgs). For each floor described, the depth below the ground surface is reported at the vertical division between Units 1 and 2 seen in the west profile of Units 1-3 (Fig. 11). This is the common point shared by Units 1, 2, 13 and 15. Figure 19 indicates the location of some floors in relation to Unit 15's stratigraphy. Sector A Features and Floors The floors excavated within the 2 x 2 rn square formed by Units 14, 15, 16 and 17 were difficult to identify and to excavate as discrete entities. Many appeared as compact lamina 0.5 to 2.0 ern thick that we attempted to isolate and peel off in succession. Straus summarizes the interpretive problem presented by tightly stacked occupation surfaces, what he refers to as "complex palimpsest cave deposits" (Straus 1990:264): The stacking effect, often with few or no discernible sterile stratigraphic zones, means that archaeologically even the finest levels are palimpsests ... It is unreasonable and unnecessary to try to isolate individual occupation residue surfaces (Ibid.:265 with original italics). While the isolation of "moments in time" may not be theoretically possible or even necessary, Sector A's "palimpsest" strata did serve our purposes as stratigraphic alternatives to arbitrary excavation levels. Given the soil conditions however, this excavation method still did not enable us to discern whether or not some tightly stacked Sector A floors had truncated earlier floors, or if minor 252 cleaning episodes had eradicated portions of one or more floors. In excavating the so-called floors of Manachaqui Cave's Sector A, the primary goals were to evaluate: 1) the chronology of site occupation through stratigraphic sampling and 2) the site's functions through diachronic analysis of the behavioral contexts that the floors and other archaeological features represent. Straus promotes the continued use of cave sites to gather information on regional cultural chronology as long as careful stratigraphic excavation methods acknowledge the palimpsest problem. In order to understand the shelter's functions, he councils that caves and rockshelters be understood as "places" within larger systemic contexts where limited sets of activities repeatedly took place: "even if our thinnest archaeological levels and lenses in caves and rockshelters are palimpsests, it is this redundant, structurally conditioned use of space that allows us to obtain meaningful behavioral information from these kinds of sites" 1990:279). (Straus Behaviorally meaningful levels, Straus argues (Ibid. :268), are best isolated by "anchoring" them with features such as hearths and other activity areas. To interpret every floor within tightly stacked series as a separate occupational event would imply that each can be isolated, that its rate of cultural deposition can be accurately accessed and, therefore, that each individual 253 "floor" has knowable temporal and behavioral meaning. Since this is clearly not the case, the only Stratum A floors reported individually within the following paragraphs are those which appeared in stratigraphic isolation from other floors. These are the lowest floors in Stratum 2B. The series of floors in Stratum 2A occur as stacked and virtually indistinguishable palimpsests, and many were only discerned in the southern half of the 2 x 2 m block (Units 14 and 15). Beyond the shelter's drip line, which crosses Units 16 and 17, erosion has eliminated most traces of ancient surfaces. Other interpretive problems related to floor excavation will be discussed during description of the floors and features to follow. Only Feature R-5, a small drainage canal, and the remains of one floor are associated with Stratum 3 in Sector A. The canal constitutes the first evidence of human modification of the shelter interior. The preservation of the first identifiable floors may be attributed to its efficacy in diverting water. Feature R-5 (Stratum 3 [70 em bgs]; Level 16 in Unit 12) Feature R-5 is a small drainage channel shaped with stones and clay resting directly on Sector A's bedrock (Fig. 24). This drain was found in Unit 12 and designated "Rasgo 5" or R-5. It was built to divert runoff entering the rear of the shelter away from activity areas on the east side of 254 the shelter. R-5 begins at the rear southwest corner under a ceiling protuberance that we observed collecting and dripping water into the interior during each storm. A semi- circular ridge of clay apparently forms part of a catchment basin. The basin drained through a narrow channel comprised of parallel ridges of stones embedded in clay. The water ran through a gap between two large boulders protruding from Sector A's bedrock that functioned as a natural conduit directing the water out of the western side of the shelter's mouth. Floor GG (Stratum 2/3 [99 em bgs]; Level 68 in Units 15 - 17) One layer designated "possible" Floor GG was identified as a compact surface in three of the four units comprising the 2 x 2 m block (excavation of Unit 14 had previously exposed bedrock). Like the floors above, GG was probably compressed by prolonged human activity. It lies at the interface of Stratum 2 with Stratum 3 and contained abundant charred organic material, but lacked artifacts except for micro-debitage recovered in the 1/16 inch screens. No radiocarbon dates were run because of lacking associated diagnostic material. Floor FF (Stratum 2B [93 em bgs]; Level 66 in Units 15-17) This compact surface was uncovered within Stratum 2B in Units 15, 16 and 17. An irregularly shaped rock-filled 255 hearth straddles Units 15 and 17 (Fig. 25a). Its length is 50 em, its width is 34 em and it varies from 4 to 7 em deep. Charcoal from the hearth rendered the date 3520 ± 100 B.P.: 1570 b.c. (all dates reported in the text are uncalibrated). Several angular stones lie embedded within its fill which was bagged for transport and processing in Trujillo. One chipped-stone scraper and one intrusive sherd from overlaying strata lay on FF's surface. Floor EE {Stratum 2B [88 em bgs]; Level 63 in Units 15 - 17) Floor EE includes a large, at least partially disturbed, rock-filled hearth in Units 16 and 17 {Fig. 25b). It appears to have been roughly circular with a diameter of approximately 50 em, and 10 to 13 em deep. The western edge was probably blurred by exposure to rain and roofline dripping. Hearth charcoal produced the date 3830 ± 100 B.P.: 1880 b.c. Floor DD {Stratum 2B [85 em bgs]; Level 60 in Units 14 - 17) At the eastern edge of Unit 15 lies a circular hearth cut during excavation of Unit 3 in 1988. Its diameter measures 26 em and it is only three em deep. Like Floor EE's hearth, Floor DO's also contains numerous angular stones. 256 Floor CC (Stratum 2B [81 em bgs]; Level 57 in Units 14 - 17) Floor CC yielded a large rock-filled hearth 50 em in diameter and five em deep in Unit 17 (Fig. 26a; Plate VI). Floor BB (Stratum 2B [77 em bgs]; Level 54 in Units 14 - 17) Floor BB appears to be the latest of the Preceramic Period floors preserved at Manachaqui Cave. An oblong hearth 38 em long by 24 em wide in Unit 15 is only two em deep (Fig. 26b). Several stones found resting above the eastern and western rims of Floor BB's hearth may have been pushed aside during hearth preparation, or extracted from the hearth prior to re-use. The hearth's fill yielded the date 3670 + 100 B.P.: 1720 b.c. A few sherds collected at this depth (only in Unit 15) are most likely the product of stratigraphic mixing toward the eastern side of Sector A. Feature R-7 (Units 9, 10, 11, 25) Feature R-7 is a stone wall uncovered in Units 9, 10 and 25 that appears to be continuous (Figs. 27, 28). The excavation of Unit 11 revealed the remains of a similar and probably contemporaneous wall. The term "wall" is applied here to connote architectural function. In reality, Feature R-7 is a pile of deliberately placed fieldstones. The base of R-7 rests directly upon the aceramic/preceramic deposits at the bottom of Stratum 2B, yet the presence of potsherds embedded within the stone masonry indicates Initial Period 257 construction. As mentioned previously, this wall impeded the flow of water into the shelter interior from the slopes behind. It may have been refurbished repeatedly for centuries, but we found it completely covered by soil. Floor AA (Stratum 2B [58 em bgs]; Level 43 in Units 14 - 17) Floor AA is the earliest Initial Period floor, although sherds begin to appear in substantial quantities in all four units of the 2 x 2 block above level 50. A circular hearth in Unit 17 measures 26 em in diameter and two em deep (Fig. 29a). A date from the hearth charcoal is 2850 ± 90 B.P.: 900 b.c. Floor Z (Stratum 2B [56 em bgs]; Level 41 in Units 14 - 17) Floor Z contains two hearths labeled Hearth 1 and Hearth 2 in Units 16 and 17 respectively (Fig. 29b). 1 is the largest encountered at Manachaqui Cave. Hearth In plan view it is slightly quadrangular, and it is 48 em long (east to west) and an estimated 45 em wide. The northern extreme was removed during the excavation of Units 20 and 21 leaving this and Hearth 2 both visible in profile (Fig. 20). Hearth 1 measures 11 em deep and, unlike the other Sector A hearths, it contained many sherds. Hearth 1's charcoal rendered the date 2800 ± 90 B.P.: 850 b.c. Hearth 2 was also cut by prior excavations, but it has a circular shape, measures approximately 30 em in diameter and has a depth of 258 nearly five em. The 1988 profile (Fig. 11) revealed a stone in the center of Hearth 2's charcoal fill. Floor Y (Stratum 2B [50 em bgs]; Level 38 in Units 14 - 17) Floor Y was identified only by its compaction and the presence of horizontally dispersed sherds. There is no hearth, but charcoal gathered from the floor produced the date 2450 ± 90 B.P.: 500 b.c. Floor X (Stratum 2B [46 em bgs]; Level 35 in Units 14 and 15) Floor X was only identified in the southern half of the 2 x 2 m block. Floor charcoal yielded the date 2560 ± 100 B.P.: 610 b.c. (this charcoal sample was small and required additional counting time during processing). Sector A's palimpsest strata between Floor X and Floor H (Level 11) in Units 16 and 17 were not sufficiently preserved to permit stratigraphic excavation. The corresponding levels in these two units were excavated arbitrarily in five centimeter increments. Floor X shows increasing color mottling, perhaps from the infiltration of light brown soils found in more substantial quantities within overlying strata. Floor W (Stratum 2B [38 em]; Level 30 in Units 14 and 15) Floor W appears as a compact layer best preserved in Unit 15. It corresponds to the lowest of the four floors first identified in Units 12 and 13 (Floor No. 4). A 259 radiocarbon date from Floor W charcoal is 2110 ± 80 B.P.: 160 b.c. Floor V (Stratum 2B [35 em bgs]; Level 28 in Units 14 and 15) Floor V was also identified by its compaction, and otherwise lacks distinguishing cultural features. Because of its proximity, it might be grouped with the series of Floors U through H just above. Within Floor V and the overlaying floors, a narrow fissure in the soil trending northwest to southeast becomes visible in Unit 15. The crack closely traces the outline of the large rock protruding from bedrock below. It does not correspond to any stratigraphic divisions that we could detect visually, but is clearly a post-depositional soil feature. The mechanics of this "micro-faulting" remain unclear but may have something to do with the differential rates of soil drying and matrix contraction and/or compression originating above. This fissure is a likely perpetrator of stratigraphic mixing. Floors U through H (Stratum 2A [17-34 em bgs]; Levels 26 - 11 in Units 14 and 15) This series of 16 "palimpsest" floor layers seems to represent centuries of intensive, repeated use of the shelter. During these centuries, Manachaqui Cave's users periodically laid down the coarse yellowish-brown fill described earlier. Subsequent disturbances have left brown 260 mottles and large brown patches in otherwise black, charcoal-laden silty soil. We excavated hearths in Floors T, P, L, J and H in Units 14 and 15. The shape of Floor P's hearth may indicate repeated uses (Fig. 30). Excavation of Unit 16 by arbitrary level uncovered two hearths. One is most likely associated with FloorS (Fig. 31a), while the other is less securely matched to Floor M (Fig. 31b). In Units 12 and 13, floors identified as 3, 2 and 1 roughly correspond to Floors T, Nand I in Units 14 and 15. these floors are associated with Stratum 2A. All of Hearth dimensions and pertinent radiocarbon dates from Sector A are listed in Table 3. Floors G- A (Stratum 2A [2-16 em bgs]; Levels 10 - 2 in Units 14 - 17) Floors G through A constitute a continuation of the same series of palimpsest occupations as Floors U through H. They are distinguished only because of change in the attributes of its associated pottery. contain six hearths. Floors G through D Five of these are irregularly shaped and no more than one ern deep, indicating haphazard preparation and perhaps only a single usage. The sixth hearth, in Floor D (Unit 16), is 21 ern in diameter, six ern deep and contains an irregular stone embedded in the underlying soil. Floors C through A have three irregular, shallow hearths. Floors G through A yielded the largest amounts of ceramic and lithic remains of all of Sector A's 261 floors, and they most likely pertain to the Late Intermediate Period and/or the Late Horizon. Sector B Features and Floors As previously mentioned, Sector B constitutes Sector A's trash midden. Although potsherds in Sector B are most often found vertically or diagonally arrayed, isolated "pavements" of horizontally-oriented sherds representing fragments of eroded occupation surfaces indicate that refuse was also generated by activities on the berm surface. Where these surface fragments were preserved, we assigned them a feature number, exposed them as feasibility permitted, and described their salient characteristics. Possible correlations between Sector B's and Sector A's floor fragments and architectural elements are discussed later in this chapter. Features R-6 (Level 16 [47 em bgs] in Units 23 and 28) and R-8 (Unit 18) Features R-6 and R-8 are crude stone walls that extend from the mouth of Manachaqui Cave onto the berm, presumably for additional protection from wind and rain (Fig. 32a, b, 28). Directly abutting the shelter exterior, Feature R-6 projects 1.2 m northward from the east side of the shelter's mouth. It is 1.1 m high and 0.80 m wide. The wall's size and form were obtained by piling large and irregularly shaped stones three courses high and two courses wide. base of the wall corresponds to Unit 22's Level 16 (the The 262 lower portion of ceramic-bearing Stratum 2), and therefore indicates Initial Period construction. In Unit 18, Feature R-8 seems to be a similar yet poorly preserved construction projecting from the western side of the cave mouth. The wall was formed by placing several large stones over a boulder that protrudes from bedrock. The top of Feature R-8 is 30-40 em higher than the top of R-6. Because only a small area of the stone feature was exposed, its dimensions could not be assessed. It also remains unclear whether or not the wall abuts the shelter exterior, although it likely does. Feature R-8's stratigraphic associations are likewise difficult to establish. Its construction dates to the late Preceramic or the Initial Period. Feature R-4 (Stratum 2 [69 em bgs]; Level 16 in Unit 31) Feature R-4 represents the only post-Preceramic Period hearth found in Sector B (Fig. 33a, b). It is associated with the layer of stones just beneath the projected 2A/2B interface previously described. A circle of stones with a diameter ranging from 30 to 35 em delimits the hearth's circumference. Poor preservation and stones within the surrounding soil matrix hindered precise definition of the base and walls of the basin-shaped hearth, but its depth is estimated at three to four ern. Charcoal from the hearth contents produced the radiocarbon date 2810 ± 100 B.P.: 860 263 b.c. At the bottom of the hearth lay a horizontal cluster of sherds from the same carinated vessel. A thin layer of charred organic remains lies in association with the hearth surface. Feature R-3 (Stratum 2 [41 em bgs]; Level 8 in Unit 21) Feature R-3 is a small horizontal distribution of sherds that represents the remains of an occupation surface within the lower half of Stratum 2. Many of the fragments belong to two vessels with incised-line decoration and assigned to the Early Horizon. Feature R-2 (Stratum 2 [35 em bgs]; Level 6 in Units 26, 29, 30) Feature R-2 is a scatter of sherds and one ground slate point distributed horizontally in Stratum 2. concentration lay in Unit 30. The principal The association of kaolin wares suggests and Early Intermediate Period date. Feature R-1 (Stratum 2 [15 em bgs]; Level 4 in Unit 19) Feature R-1 is another horizontal distribution of sherds within Stratum 2. Remnants of this floor may have been removed in all but Unit 19 during the 1986 cleaning episode. Material remains of this occupation were probably incorporated within sod bricks utilized to seal the shelter interior, or were tossed off of the berm's edge with loose soil. Because of ancient surface undulations, R-1 may 264 correspond to the uppermost floors A or B in Sector A, and perhaps to Feature R-9, an identical distribution of horizontal sherds uncovered within the first excavation level in Unit 26. further detail. Feature R-9 need not be described in Both Features R-1 and R-9 date to the Late Intermediate Period or Late Horizon. Changing Hearth Morphology and Function The 24 hearths excavated within the 2 x 2 block may be classed into three groups based upon salient attributes. It is likely that each class served a different function. These classes are: 1) rock-filled hearths, 2) hearths with a single stone embedded in the bottom, and 3) simple basinshaped hearths with no stones. The five rock-filled hearths uncovered in Floors FF, EE, DD, CC and Z are generally largest with a mean diameter of nearly 44.8 em and mean depth of 7.8 em. Archaeologists have interpreted hearths like these as roasting pits for meat based upon ethnographic analogies (e.g. Lavallee 1977:87-88). The traditional native practice of roasting meat in a pit filled with hot rocks is called a pachamanca. The six "single-stone" hearths in Floors D, H, M, P, S and T average 30.2 em in diameter and 6.4 em in depth. The embedded stones are irregularly shaped and vary from 10 to 15 em in length or diameter. Floor Z's Hearth 2 may be the earliest of this variety, but its accompanying stone appeared to be floating in the hearth fill rather than 265 embedded within the soil beneath. In the absence of ethnographic analogies, it may be difficult to establish whether or not the single-stone hearths were used for cooking. The stone may have absorbed heat, acting to radiate warmth for the shelter's inhabitants after the fire's coals had cooled. The 13 simple basin-shaped hearths found in Floor BB and floors above are highly variable in shape and surface extent. Their depth averages only 1.9 em. Perhaps these served merely to warm chilled wayfarers for short periods. A functional analysis of Manachaqui Cave's hearths should provide clues regarding the shelter's changing function, and changes in the site function may reflect transformations in regional economic systems. In viewing the relative stratigraphic positions of each hearth class, the most striking change is the virtual disappearance of rock-filled hearths above Floor CC. Floor Z's Hearth 1 may constitute a late exception, although it contains few rocks in comparison to earlier examples. Associated radiocarbon dates suggest that the shift in hearth preference occurs during the Preceramic Period/Initial Period transition and coincides with the introduction of pottery technology in the Peruvian north highlands. The study of organic remains associated with Manachaqui Cave's three hearth types and their corresponding floors potentially provides information regarding relationships between subsistence and technology. 266 These data are addressed in later chapters. Radiocarbon Evidence for Chronology In order to date the occupational sequence at Manachaqui Cave, a strategy of processing radiocarbon dates from stratigraphic columns in each sector was implemented. The series of radiocarbon dates from the 2 x 2 m block (Units 14-17) in Sector A provides a preliminary chronological assessment of the shelter's history of use and disuse. Radiocarbon dates from Sector B aid stratigraphic correlation of the two sectors, and simultaneously serve to date the macrochronological components within their most intact stratigraphic contexts. The radiocarbon dates from each sector are discussed below, and all of the dates are listed in Table 4. Sector A Radiocarbon Evidence Radiocarbon dates from Sector A show consistent overall trends. Some ordinal inconsistencies may be discounted by considering the standard deviations of the corrected dates (Pearson and Stuiver 1993; Stuiver and Pearson 1993). For example, the corrected sigmas of the seemingly inverted dates from Floors H and L (A.D. 544 - 660 and A.D. 646 - 776 respectively) overlap. The same is true of the seemingly inverted dates from Floors X and Y (A.D. 801 - 413 B.C. and 761 - 391 respectively) . Recall that Floor Y supplied a small sample of charcoal requiring special treatment. 267 More problematic are inconsistencies between dates rendered by Floor P (110 b.c.) and Floor T (a.d. 490). The corrected sigma from Unit 14 Floor P's earlier date (A.D. 129 - 341) does not overlap with that from Unit 15 Floor T (A.D. 595- 672), even when taking two sigma for both dates into account. This problem may stem from post-depositional transformations such as the partial removal of layers while preparing or "cleaning" activity areas. The Floor P date was extracted from a hearth, and the Floor T date was processed from charcoal embedded in the floor. An effort at refurbishment that involved the removal of soil from the center of the shelter's interior could result in such a stratigraphic disconformity. It might be hypothesized that the "cleaning" occurred sometime between circa a.d. 110 and circa a.d. 450, and involved the destruction of Floor P remains that would have been contained within Unit 15. On the other hand, the early Floor P hearth date was conceivably produced by the burning of comparatively old wood. In general, we found disturbances extremely difficult to detect and delimit given Sector A's stratigraphic conditions. Dating inconsistencies could also have resulted from a tendency of our stratigraphic excavation to inadvertently cross-cut tilting and undulating prehistoric palimpsest surfaces where they are not well-preserved. is improbable that alternative techniques utilizing It 268 arbitrary levels would have alleviated such mixing problems. A single radiocarbon date that requires explanation because it is later than dates from the two floors above belongs to Floor FF. The corrected sigma of FF's date (1913 - 1677 B.C.) overlaps with BB's (2135- 1789 B.C.), but not withEE's (2397- 2038 B.C.). The stratigraphy at this depth shows every visible indication of being intact, and it seems most logical to disregard Floor FF's date as aberrant. Manachaqui Cave's Sector A radiocarbon dates may be grouped into five sets corresponding to the Andean periods and horizons defined in Rowe and Menzel's sequence (1967). The sets of dates also corroborate the periodization proposed for the 1988 ceramic macrochronology. Accordingly, phase names taken from local toponyms have been assigned to the terminal Preceramic Period component as well as the four ceramic period macrochronological components. Radiocarbon dates from Floors BB through FF (1880 to 1520 b.c.) represent the Lavasen Phase, the final centuries of the Late Preceramic Period. Dates from Floors AA and Z (900 and 850 b.c.) pertain to the Initial Period, and the associated ceramic complex featuring vessels with carinated and semicarinated wall profiles is assigned to the Manachaqui Phase. Radiocarbon dates from Floors Y and X (610 and 500 b.c.) fit within the Early Horizon, and the corresponding ceramic complex with its fine-line incised decoration pertains to the Suitacocha Phase. Dates from Floors W 269 through H (160 b.c. to a.d. 570) coincide with the Early Intermediate Period when the use of kaolin ware pottery became widespread throughout the north-central Andes. These dates can be separated into an earlier component represented by dates from Floors W and P (160 b.c. and a.d. 110) and a later component represented by dates from Floors H, L and T (a.d. 450 to 570). These temporal components are termed the Colpar and Empedrada Phases respectively. Manachaqui Cave's radiocarbon dates and their relative stratigraphic positions also suggest extended periods of the shelter's disuse, but the problem of sequential hiatuses will be considered in the following section. Sector B Radiocarbon Evidence and Stratigraphic Correlations with Sector A Three Preceramic Period dates were processed from Sector B's berm deposits. An AMS date derived from a small charcoal sample from Stratum 3C in Unit 31 Level 35 is 10,270±60 B.P.: 8,320 b.c. However, the problem of early Preceramic Period occupations at Manachaqui Cave awaits future study. Two radiocarbon dates from Stratum 3 in Sector B fall fully within the third millennium B.C. These dates are 4120 ± 130 B.P.: 2170 b.c. and 4280 ± 110 B.P.: 2330 b.c. from Unit 31, Levels 25/26 and Level 30 respectively. The associated deposits include abundant lithic material with several tool types (e.g. projectile points, scrapers, 270 burins, etc.). The Preceramic dates from Sector B do not overlap with the Lavasen Phase dates from Sector A which represent only the last few centuries of the Preceramic Period. The paucity of Sector A's (Floors BB through FF) lithic remains confounds comparisons of the shelter interior and exterior artifactual components. For these reasons, it may be imprudent to assign both to a single Preceramic Period phase. In the following chapter, the probability of a transitional stage separating the Preceramic and Initial Periods will be addressed. A date from Feature R-4, the single Sector B hearth in Unit 31's Level 16, of 2810±100 B.P.: 860 b.c. (I-17,320) serves to correlate Stratum 2's rock layer with Manachaqui Phase Floors z and AA in Sector A. All three Manachaqui Phase dates fall within the fifty year (uncorrected) span between 850 and 900 b.c. A stratigraphic association is suggested by Floor AA's depth of 61 em below the surface (in the northeast corner of Unit 17), Feature R-4's depth of 69 em below the surface, and the projected slope of M-lA's strata from Sector A to Sector B. The approximate contemporaneity of the three contexts is corroborated by ceramic evidence to be described in the next chapter. Two dates from Unit 22 Level 11 and Unit 6 Level 25 directly date the Suitacocha Phase ceramic component. These dates are 2630 ± 100 B.P.: 680 b.c. and 2740 ± 90 B.P.: 790 b.c. respectively, and both of these Sector B contexts may 271 be regarded as roughly coeval with Sector A's Floors X and Y. At 41 em below ground surface, the Suitacocha Phase sherds comprising Feature R-3 were probably deposited during activities that contributed to the formation of Floor Y at 44 em below the surface (in the northwest corner of Unit 17). Radiocarbon dates to correlate Sector B's Early Intermediate Period levels with those identified in Sector A have not yet been processed. Based solely upon the "negative" evidence provided by Manachaqui Cave's series of calibrated radiocarbon dates, it may be argued that gaps in the occupational sequence lasting three hundred years or more occurred during the mid-Early Horizon, the middle Early Intermediate Period and the Middle Horizon. Of these three, it seems that the hiatus during the middle Early Intermediate Period may be more apparent than real. Additional radiocarbon dates may fill the possible gap between the Early Intermediate Period Colpar and Empedrada Phases, and also provide evidence for Middle Horizon occupation apparently lacking in Manachaqui Cave's ceramic assemblage. Some of the best evidence that occupational hiatuses occurred within Manachaqui Cave's occupational sequence is the erosion of sherd surfaces presumably resulting from extended periods of exposure on the site surface. In Sector B, all of the phase assemblages have suffered severe erosion except for that belonging to the Manachaqui Phase. This 272 condition suggests that the Suitacocha, Ernpedrada and Poblano phases were each followed by decades or centuries of Manachaqui Cave's comparatively light usage or abandonment. For example, radiocarbon dates and sherd erosion patterning suggest that a gap in the shelter's sequence corresponds to the mid-Early Horizon {coeval with the Chavin horizon) . Manachaqui Cave's occupational hiatuses require explanation as reflections of regional and pan-regional social, political and economic events. Materials Analysis The analyses of ceramic, lithic and other cultural remains from Manachaqui Cave are designed primarily to extract chronological information. Thus, emphasis is placed upon relating artifacts and their salient attributes to the stratigraphy and associated radiocarbon dates, thereby establishing relative and absolute sequences for the shelter's occupation. Functional consideration of pottery vessels, stone tools and other artifacts receives sufficient attention only to address questions of Manachaqui Cave's changing functions and to assess the comparability of Manachaqui's assemblages to others in neighboring regions. Clearly there is a need for raw material sourcing studies, use-wear analyses and additional specialized investigations, but these await future opportunities. 273 Ceramic Analysis The ceramic study presented in subsequent chapters is a simple attribute analysis much like those performed by Patterson (1966) on the Central Coast of Peru, and Burger at Chavin de Huantar (1984b) and Huaricoto (1985b). The attributes of primary importance relate to ceramic technology, vessel shape and decoration. A total of 72,310 sherds were recovered during the 1988 and 1990 excavations and subsequently examined for the analysis of Manachaqui's occupational sequence. This thesis will examine the ceramic sequence only through the Early Intermediate Period. However, the absence of visible, discrete stratigraphic correlates for proposed phase components combined with the lack of clear differentiation between Early Intermediate Period and later wares preclude estimates of the total sherds corresponding to each of the four established phases. Of the 72,310 total sherds, approximately 7,500 sherds exhibiting selected diagnostic morphological, decorative and technological features were culled and sorted by established categories. Many sherds deemed too fragmentary or eroded to supply reliable information were separated out, leaving 6,889 sherds that supplied the data for the following analysis. The study of the Manachaqui artifact assemblages differs from other analyses because there are few useful stratigraphic indices to isolate occupations and independently support proposed phase designations. 274 Necessity requires the utilization of the artifacts and their horizontal and vertical relationships to one another to generate a sequence based solely on internal evidence. At Lathrap's Ucayali sites and at Manachaqui Cave, ceramics provide the most plentiful and informative material from which to derive the sequence. The archaeological phases presented within the next three chapters represent the product of macrochronological analysis of ceramic attributes (what Lathrap termed "ceramic features"). The analysis focused mostly on the units excavated in Sector B judged through the course of study to retain the most intact depositional sequence. Those units cluster around the crest of the berm. Attempts at establishing ceramic chronologies usually begin by distinguishing between wares presumed to vary in frequency through time. The majority of Manachaqui's pottery sample from all levels consists of plain brown wares. Angular grains of igneous rock comprise the predominant non-plastic inclusions within ceramic pastes throughout the occupational sequence, and the pastes lack other useful diagnostic characteristics with which to consistently distinguish wares. Hence, wares are regarded as insensitive analytical units for the purpose of evaluating change in the bulk of the pottery sample. Nevertheless, it becomes apparent to the analyst that the variety of wares increases in later levels of Manachaqui 275 Cave's deposits. The creation of large "paste groups" based on easily recognizable gross features constitutes a response to this problem. Wherever possible, distinctions are made between the predominant brown ware and other wares with clearly observable distinguishing characteristics. Among these others are wares with abundant quartz crystal and micaceous inclusions, or white, pink and gray wares made with kaolinerich clays. Within the following chapters, these three sets of wares will be termed Paste Group A (predominant brown wares), Paste Group B (deviant, often micaceous wares) and Paste Group C (kaolin wares). Group A wares were likely produced at settlements close to Manachaqui Cave by the populations that most frequented the road and shelter. Group B and C wares, which appear in relatively small quantities, represent pottery originating at greater distances. They cannot be termed "exotic" because all pottery was, in a sense, exotic to Manachaqui Cave. Nor can they be labelled "trade wares" because they may have arrived by means other than trade. The erosion of sherd surfaces precluded classification of pottery remains by surface treatments (e.g. polished, burnished, matte, etc.). However, clearly identifiable surface treatments and wares can be useful attributes for analyses provided they are weighted equally with the presence of other attributes like fine-line incision, red- 276 painting, rim-beveling, etc. The problem of erosion also hinders efforts to quantify the frequency of certain superficial decorations, especially slipping and painting. Due to these analytical methodologies, and of course the stratigraphic conditions, sherd totals pertaining to every phase can only be estimated by an assortment of techniques, each one frought with difficulties. Some estimates are necessary in order to determine proportions of paste classes within each phase assemblage. Resulting paste group ratios provide one index indicating the relative intensity of in~eraction from phase to phase, and these ratios are obtained using simple rim counts. The assumption that rim-type frequency ratios provide approximations of vessel-type frequency ratios is based upon quantitative experimentation by DeBoer and Carroll (1992). Three aids which readers can utilize to visualize the vertical distributions of ceramic attributes are included in the appendices. Appendix A includes a list of the illustrated artifacts showing the provenience of each. Appendix D shows the macrochronology which supplied the initial basis for the creation of phases. Appendix E presents the distributions of vessel shapes and rim forms by provenience, and Tables 5 through 19 summarize vessel/rim frequencies and dimensions and other ceramic data. Computer data base and spreadsheet programs such as Paradox 5.0 and Quattro Pro 5.0 aided the compilation and sorting of 277 ceramics by provenience and attributes. Lithic Analysis The collection of chipped and ground-stone artifacts from Manachaqui Cave's ceramic period deposits (including Late Horizon deposits) consists of an estimated 9,500 pieces. In the absence of use-wear studies, the chipped stone lithic analysis is admittedly cursory and strictly technological. Refuse from a simple core-flake industry, with the addition of a few unifacially flaked tools and projectile points, marks the time periods in question. Functional interpretations are more easily rendered for the limited yet familiar varieties of ground-stone implements unearthed at Manachaqui. Of these, the ground-slate points constitute a set of tools for which a more penetrating technological analysis is warranted, especially given the presence of slate blanks, partially worked preforms and slate detritus. The provenience of each stone artifact category is matched against ceramic macrochronological distributions in order to determine its pertinent archaeological phase. Analyses of Organic Remains Organic remains recovered during excavations at Manachaqui Cave are either charred botanical remains, or burned and unburned bones. These provide direct evidence of plant and animal products consumed at the site, and indirect 278 evidence of the site's functions (see the previous chapter). In the case of the botanical remains, only the material accidently charred is preserved. Nearly all of the unburned faunal remains from the shelter interior survived, although trampling and weathering throughout the site contributed to a generally poor state of preservation. All samples had to endure the rigors of transportation to Pataz by muleback, and to Trujillo by truck. Botanical Analysis Through a two-stage process of water and chemical flotation employing the IDOT technique (Pearsall 1989:3552), virtually all botanical remains were extracted from the soil and hearth fill samples recovered from Manachaqui. Sources of sampling error include the loss of minute remains from floor samples wet screened with 1/16th inch mesh at the site, and the destruction of remains too fragile to withstand the wet screening and subsequent rigors of transport. Samples from Sector A Unit 15's floors and the hearths were sent to Dr. Deborah Pearsall at the University of Missouri for analysis. After agreeing to emphasize the earlier components, Pearsall selected a total of 42 samples for sorting, remains. floors, identification and quantification of botanical Analyzed were 14 samples from all Lavasen Phase 4 samples from both Manachaqui Phase floors, 4 samples from both Suitacocha Phase floors, 18 samples from 17 Empedrada Phase floors, and two samples from late 279 prehistoric floors. Some of the larger samples were sub- sampled, and frequencies for all taxa identified were tabulated on a computer spreadsheet program. These are tabulated within Pearsall's report included here as Appendix F, and in Table 18. Pearsall then calculated a series of ratios to aid interpretation of the rockshelter's changing functions. The bulk of the botanical remains consisted of wood charcoal, which was weighed but not identified. Faunal Analysis All of the bones and bone fragments recovered during 1988 and 1990 excavations at Manachaqui were sent to Dr. Jonathan Kent at the Salango Laboratory in Salango, Ecuador for analysis (Kent 1994). At Salango, Kent and student assistants weighed and counted the faunal materials. The bone sample totaled 3,673 bones of which 2,400 (65.34%) come from Sector A, and 1,273 (34.66%) come from Sector B. The remains were sorted in terms of "number of identified specimens" (NISP). These specimens were derived from single excavation samples (i.e. they share common provenience), and may be comprised of from one to 85 fragments of a single bone type and animal size. Kent's bone types are: Long bones: humerus, radius, ulna, femur, tibia, fibula, fragments of phalanges and metapodia that were not recognized as such due to fragmentation; and Irregular bones: recognized phalanges and metapodia, recognized vertebral parts, carpals, tarsals and hyoid; Flat bones: scapula, ribs, cranial portions, and vertebral spines and processes not recognized as such; 280 Teeth: crowns, roots and fragments of enamel; Scrap: any bone not recognized as any of the other categories. Scrap specimens may contain up to 200 tiny bone fragments. Animal size was broken down into categories of small, medium and large, and into intermediate categories when necessary. According to Kent, these are: Small: rodents such as mice and guinea pigs, and most birds. Medium: most carnivores, pacas (agouti), a few birds and pudus, the smallest of the deer (cervidsl. Large: carnelids and other cervids. After subtracting bones of the scrap category, the remaining 2,319 bones (63.14 of the sample total) served to analyze: 1) bone types from Sector A vs. Sector B; 2) bone types from the east vs. the west side of Sector A; and 3) bone types and animal sizes by period. Specimens identified and classified into the above categories were then tabulated by unit and level or floor, and entered into a spreadsheet computer program. Results of the faunal analysis are summarized in Tables 19-28. The high total of bones is misleading as the entire collection weighs 450.8 grams (less than 1/2 kilogram). Nineteen bones found outside of M-1A comprise 20 percent of the total weight. Kent observes that the high degree of fragmentation can be observed within Unit 15 where 1,850 bones weigh 165.3 grams together, and average 0.089 grams each. Kent's taphonomic study observes that bird activities, "trampling, digestion, weathering, human 281 smashing, and both rodent and carnivore gnawing are all evident." Scrap bone tallies that are higher outside than inside the shelter (44.32% compared to 31.59%) support the conclusion that Sector B's deposits were exposed to greater attrition. Kent concludes that the depositional patterns in evidence "seem largely the results of human activities with minor contributions made by birds and mammalian carnivores." Again, ceramic and radiocarbon evidence provided the chronological framework for assigning the faunal remains to particular periods. Results from analyses of organic remains for each phase are summarized within each of the following chapters. CHAPTER 6 THE MANACHAQUI PHASE In this and the following two chapters, the material remains representing each archaeological phase are described in detail. As noted in Chapter 5, the Manachaqui Phase corresponds to the Initial Period in Central Andean prehistory, when the production and use of pottery first becomes important in Andean daily life. Not surprisingly, pottery constitutes the most abundant material remains recovered from Manachaqui Phase strata. Ceramic and stone artifacts will be described in turn below, followed by sections reporting non-artifact remains such as floral and faunal material. Manachaqui Phase Antecedents While the Preceramic-Initial Period transition on the coast coincides with documented shifts in subsistence and settlement patterns {Lanning 1967), economic stability may be more characteristic of coeval highland development {Burger 1992:104). However, our view of economic change in the highlands may be constricted by preservation problems as well as the archaeological focus on monumental architecture and on rockshelters of the central Peruvian puna. 282 As the 283 only north highland rockshelter occupied during the Preceramic-Initial Period transition studied to date, Manachaqui Cave potentially supplies new and valuable information. The transition at Manachaqui warrants analysis beyond the purview of this thesis, but a few preliminary observations can be offered here. At Manachaqui Cave the Preceramic-Initial Period shift apparently included a transitional stage that might be thought of as the Terminal Preceramic Period. In Sector A's Stratum 2B, Lavasen Phase Floors BB through FF represent this intermediate stage. The shift from rock-filled to simple basin-shaped hearths marks the end of the stage. Underlying Sector B's Manachaqui Phase levels, Lavasen Phase remains are contained within Stratum 2B's lowermost 10 to 15 centimeters. These layers yield lesser quantities of chipped stone artifacts than the Preceramic Period levels of Stratum 3, and none of Stratum 3's projectile points nor finely worked tools. The Lavasen Phase raw stone materials in both sectors are coarse grained, and sufficiently isotropic only for a simple core-flake industry. The early Initial Period Manachaqui Phase is distinguished by a remarkable dearth of any chipped-stone tools (see Appendix D) • This proposed stage was also identified during Kaulicke's excavations in highland Cajamarca where "Final Preceramic Period" material evidence underlay Initial Period 284 Pandanche Phase A deposits: In the Final Preceramic that forms the basal part of the cultural sequence we have a poorly documented phase whose definition is based on the absence of ceramics, and welldefined layers with clear cultural associations such as hearths ... There are few lithic artifacts and they are not clearly distinguishable from others found in overlaying layers with ceramics. One can even speak of an impoverishment of the industry as raw materials such as chert [silex] do not appear ... (Kaulicke 1975:40). A single charcoal sample from two hearths (one simple and another delineated by stones) that apparently belongs to Kaulicke's Terminal Preceramic Period provided two dates: 2010 b.c. and 2125 b.c. (Kaulicke 1981:388). The earliest radiocarbon dates from the overlaying Initial Period Pandanche Phase A deposits are 1395 b.c. and 1490 b.c. (Ibid.). At Manachaqui Cave, the Lavasen Phase dates of 1720 b.c., 1880 b.c. and 1570 b.c. from Floors BB, EE and FF represent this Terminal Preceramic stage. If we are dealing with the same developmental phenomenon at both Pandanche and Manachaqui Cave, then we might hypothesize that a Terminal Preceramic in the northern Peruvian highlands began between 2,200 and 2,100 b.c. (recall Sector B's Late Preceramic Period date of 2170 b.c.) and ended around 1500 b.c. Intensive analysis of Preceramic Period levels and their lithic contents will be crucial to better understand the Preceramic-Initial Period transition at Manachaqui. Analysis of changing frequencies of raw lithic materials and tool types through the transition provides one avenue for continued research. Changes in material culture during these shifts reflect alterations in the rockshelter's 285 function which in turn indicate modifications in the local and regional economies. Manachaqui Phase Ceramics The Manachaqui Phase deposit is approximately 35 em thick in Sector B's excavation units at the edge of the berm, and contitutes 35 percent of the berm's meter-thick ceramic period deposits. Above Sector B's Manachaqui Phase layers, the Suitacocha, Colpar, Empedrada and Poblano Phase deposits are an estimated 20, 10, 20 and 15 em thick respectively. Examples of Manachaqui Phase levels from the berm are Unit 5, Levels 26 through 33; Unit 28, Levels 14 through 22; Unit 30, Levels 13 through 20; and Unit 32, Levels 13 through 20. Of the 6,889 diagnostic sherds culled (from the total 72,310 sherds) for this study, 3697 (53.7%) belong to the Manachaqui Phase. Ceramic Paste A During the Manachaqui Phase, Paste Group A is composed of a thin brown ware represented by 3,588 diagnostic sherds. These constitute 97 percent of the total 3,697 diagnostic Manachaqui Phase sherds. The precise geographic origins of the Paste A ware will remain unknown until archaeological work on a regional scale reveals more about Initial Period population and ceramic distributions. At present there is no reason to suspect that the large quantity of Manachaqui Phase Paste A pottery was transported from distances farther 286 than the intermontane valleys flanking both sides of the Marafion-Huallaga divide. Manachaqui Phase Paste A is compact, rather than porous. Vein quartz comprises the predominant non-plastic constituent, and angular grains reach a maximum 1.5 mm in size. Analysis with the scanning electron microprobe at Yale's Geophysics Laboratory permitted the identification of hornblende, ilmenite and numerous igneous rocks of heterogenous mineral composition. Unlike the quartz, these grains are more homogeneously sized, and seem to have been carefully sorted and added as crushed rock and sand temper. Like other Initial Period pottery of the Peruvian north highlands, Manachaqui Phase vessels exhibit thin walls between 2 and 4 mm in thickness, a characteristic that doubtless contributed to the severe fragmentation of the collection. construction. A few sherds bear evidence of coil Restricted vessels with unfinished interior surfaces occasionally show evidence of wiping with vertical strokes. Potters burnished vessel surfaces with a smooth implement, perhaps a pebble, leaving horizontal tracks one to two centimeters wide. The firing was well-controlled. The pottery is hard, and over 90 percent of the sherds show complete oxidation to a brown, reddish brown or dark brown color. Unintended firing smudges, or smoke clouds, are present but not common. described below. Other technological details will be 287 Morphology A total of 3,387 sherds were utilized to interpret and classify Paste A vessel shapes of the Manachaqui Phase. These were either pieces of vessel rims (N=2,129), or angled and decorated sherds from vessel mid-sections (N=l,258). Represented by 2,017 rim sherds that constitute 94.6 percent of the Manachaqui Phase Paste A rim sherd collection, restricted vessel shapes predominate. rounded, or nearly rounded bases. All vessels had Despite this apparent conformity, there is a considerable range of variation in formal details especially evident within the collection of rims. The Manachaqui Phase style might be considered highly "sculpted." sections. Clay is added to rims, shoulders and midRims are typically reinforced by folding, and then beveled, notched or deeply incised. Applique ribs, flanges and adornos attached to vessel mid-sections serve to alter body profiles. Such details blur taxonomic distinctions between shape and decorative characteristics, presenting challenges for establishing consistent classifications and terminologies for ceramic attributes. These same troublesome qualities give the phase assemblage its stylistic unity and unique character. Several other aspects of the Manachaqui Phase Paste A assemblage hinder the application of schemes traditionally employed to classify vessel shapes in the Andes. The first is the severe fragmentation of the collection which hampers 288 the reconstruction of vessel shapes (sherds are characteristically less than five em in diameter) . Trampling, vessel breakage patterns and other factors frustrate attempts to associate rim shapes with body profiles. Consequently, this analysis focuses on rim form as the attribute exhibiting greatest variability and most suggestive of each vessel's gross morphological characteristics. Through the painstaking exercise of retro- fitting sherds, the general tendencies of Manachaqui Phase pottery shapes became evident. A further hindrance to classification relates to how the ancient inhabitants conceived of vessel shapes and functional categories. Of course the problem of vessel functions is embedded within the larger problem of site function addressed in the final chapter of this thesis. The Paste A assemblage lacks consistently occuring sets of attributes that normally form the bases for establishing analytical dichotomies between coarse vs. fine wares and cooking vs. serving wares. Native, or ernie, formal and functional categories lack rigid standardization, and some variables of crucial importance for interpreting formal and functional classes are virtually continuous. The most nettlesome include rim angles, rim thickening, wall curvature and degree of interior surface finishing. In short, there are tendencies, but few inviolate rules, that governed construction and decoration of pottery utilized 289 during the Manachaqui Phase. In an attempt to approximate functional categories as closely as possible, this ceramic classification prioritizes rim morphology and interior surface treatment as independent variables. Given the limited evidence provided by the rim sherds, five vessel shape categories were delineated employing strictly objective criteria. Each shape category is interpreted as having functional significance. The presence or absence of some attempt to smooth vessel interior surfaces provides a basic dichotomy between vessels interpreted as "jars" utilized for food preparation, storage and perhaps transportation, and "bowls" utilized for serving, eating and perhaps transportation. While the interior surfaces of jars remain rough and unfinished, bowl interiors show evidence of smoothing and/or burnishing. Functional implications of each of the vessel categories will be further considered in later discussions. Rim Forms The following three-tiered hierarchical classification is based on 2,129 classifiable rim sherds grouped into the five vessel shape categories. The categories are Shape A: neckless jars and jars with incipient necks; Shape B: shortnecked jars; Shape C: restricted bowls; Shape D: unrestricted bowls with concave walls; and Shape E: unrestricted convex bowls. A finer level of classification separates rims of differing thickness, orientation, aperture 290 and general treatment. Thus, each rim form is given a number designation (e.g. Rim 1, Rim 2, etc.). A third level occasionally highlights relatively subtle distinctions deemed significant, and these carry letter designations. This heavy emphasis on rim morphology is appropriately employed here because the Manachaqui Phase rims are so distinctive. With few exceptions, they can be distinguished from rims of later phases, and, hence, temporal indices. they do serve as The rims were shaped by careful trimming and beveling, and often finished by polishing. Manachaqui Phase vessel necks are invariably short where present. The degree to which particular rim forms consistently correlate with specific body shapes and dimensions remains to be determined with a less fragmented study sample. Other singular features of Manachaqui Phase rims will be described below. Body Shapes For the purposes of this analysis, Manachaqui Phase jar and restricted bowl shapes (Shapes A-E) are considered either 1) globular, 2) semi-carinated or 3) carinated. The bodies of Shape E unrestricted bowls are, of course, semihemispherical. The globular vessel bodies of jars and restricted bowls are either spherical, ellipsoid or ovaloid. Semi-carinated and carinated shapes may feature convex or concave upper walls. True carinated vessels exhibit a sharp corner point or angle separating the shoulder and base, 291 thereby creating a composite silhouette. The corner point dividing upper and lower sections of carinated vessels is often referred to in the literature a shoulder or basal angle (e.g. Meggers et al. 1965; Lathrap 1962). Perhaps the latter term (basal angle) better describes the Manachaqui Phase assemblage where most joints or points of maximum width occur comparatively low on the vessel profile (Fig. 34a-h) . · Because the term "medial" seems to best characterize the location of carination angles at, or just below the vessel's mid-point, it will be utilized in this study. The term "shoulder" will signify carination angles above the mid-point. Vessel mid-sections may be embellished with a medial applique rib or flange, lending a carinated appearance to an otherwise globular or semi-carinated profile (Fig. 34i-p). The angularity of the vessel mid- section is a continuous variable, and Figs. 34a-h illustrate the graded series from globular to semi-carinated to carinated profiles. An additional complex body shape (Fig. 34c) remains poorly understood, but probably resembles twotiered Machalilla vessel shape 12: Shouldered-neck Jars from coastal Ecuador (Meggers et al. 1965: Fig. 90 and Table F) Little is known of Manachaqui Phase vessel heights. Not one vessel from Paste Group A could be sufficiently reconstructed to permit measurement. However, vessel diameters were calculated from pieces of medial carinations 292 and sherds with medial applique ribs and flanges using a formagauge and a rim diameter template. 1,097 such medial sherds were separated, and 208 (19%) of these provided reliable measurements. Diameters (always measured on the sherd interior) range from 12 to 24 em. is 16 em (N=38). The mode diameter These statistics conflate formal and functional vessel categories, and are not particularly useful. However, they do show an absence of large vessels with maximum diameters of 30 em or greater. In order to view statistical breakdowns by inferred functional categories, an attempt was made to examine metric differences between medial carination sherds with unfinished vs. finished interior surfaces. The former are most often associated with the two previously mentioned jar Shapes A and B, while the latter typically belong to bowl Shapes C and D. The mode diameter provided by 120 examples of the sherds with unfinished interior surfaces is again 16 em. A small sample of 33 sherds with finished interior surfaces inhibited comparability of distributions, but there are indications that the bowl category has multiple modes, a greater average diameter (19.6 compared to 16.9 em) and a larger standard deviation (3.0 compared to 2.65 em). These figures only verify with certainty that, on the average, bowls are wider and more variable in size than jars. It should be reiterated that secure associations between Manachaqui Phase rim forms and vessel body shapes 293 are few. Generalizing statements such as, "all short-necked jars featuring Rim Form 2 had carinated bodies," are rarely possible, although some possibilities are noted. The separate discussions of rims and body shapes just presented are a response to this dilemma. Shape A: Neckless Jars and Jars with Incipient Necks A total of 175 rim sherds grouped into 13 distinctive rim forms can be attributed to neckless jars, and jars with incipient necks. These constitute 8.2 percent of the Manachaqui Phase rim collection. Jars with incipient necks feature inflected, or concave, upper body contours that approach or end at a point of vertical tangency near the vessel orifice. The interior surfaces of these vessels have been left scraped or wiped in cursory manner. Exteriors show careful smoothing and often burnishing. Of the five Manachaqui Phase shape categories, Shape A encompasses the most variability in rim form. The majority (65 percent} of vessels exhibit everted or direct rims. The exceptions are the semi-spherical neckless jars with inverted rims represented by Rims 2, 4 and 5. Within the Shape A class, functional variability is suggested by the presence of two principal size categories. A small category is represented by vessels with Rims 1, 3, 4 and 7-13 (N=71, x=12.17 em, :=1.98 em}, and a large category by vessels with Rim 6 (N=lO, x=23.7 ern}. Adding the Rim 2 and Rim 5 vessels with orifice diameters greater 19 ern to the large category 294 renders an average of 24.07 em with standard deviation of 2.35 em (N=15). The Shape A rims are described below, and within Table 5, in order of decreasing frequency. R±m 1 (Figs. 35a-h). jars with incipient necks. Rim 1 (48 examples) belongs to The point of vertical tangency is reached at the lip (or end point) and some rims flare slightly outward. The rim interiors have been gradually thickened to a maximum thickness of 4 mm near the lip. While the lip interior edge is rounded, the exterior edge ends in a dull point. The abrupt curvature below the vessel orifice may indicate predominantly globular body shapes (Fig. 35i). The exterior surfaces of rims were burnished to a low luster. R±m 2 (Figs. 35j-p, 36a-g). Rim 2 (39 examples) exhibits convex walls and an inverted rim with interior thickening. Maximum thickness occurs at the lip which is squared or semi-squared with slightly rounded edges. The maximum thickness ranges from 5 to 12 mm and tends to increase with rim diameter. 2 to 3 mm thick. Vessel walls below the rim are These rims suggest globular bodies of variable height (Fig. 36h). The bi-modal distribution of measured rim diameters (N=29) correspondes to the two size categories described above. Rim diameters of the smaller class (N=25) range from 10 to 19 em, average 14.56 em and yield a standard deviation of 2.42 em. Eight of these rims 295 (e.g. Fig. 36a,b) are decorated with a notched applique fillet of clay adhered to the lip exterior (the decorative technique is described later within this chapter). One displays notches cut directly into the exterior edge of the rim (Fig. 36g). Four undecorated rim sherds (Fig. 36c-f) represent the larger vessel size category. Rim 3 (Figs. 37a-c). Rim 3 (15 examples) pertains to neckless jars with walls 3 to 5 mm thick. The rims are direct, and lip exteriors have been thickened to a maximum of 11 mm by applying a clay fillet which is subsequently decorated by notching in a manner similar to the decorated variants of Rim 2. Only one sherd is undecorated. The vessel body may exhibit any of the three possible profiles. Rim 4 (Figs. 37d-f). to globular jars. The 12 examples of Rim 4 belong Their characteristic orientation is sharply inverted and approaches the horizontal plane. They show gradual, often uneven, thickening on the interior surface. The end of the rim has been "pulled" inward to leave a 1 to 2 mm overhang or clay cornice, and the lip exterior edge has been squared. applique decorations. Six of the rims bear The total rim thickness reaches a maximum of 6. 5 mm. Rim 5 (Fig. 37g-i). The 12 examples of Rim 5 are direct or slightly inverted. Clay has been applied to 296 reinforce the lip exterior, and the top of the rim has been flattened. The diameters from five measured rims indicate sample partition into the small and large size categories. Maximum thickness at the lip is 10 mm, and vessel walls vary between 2 and 3 mm in thickness. Like Rim 2 vessels, these may have had simple globular body shapes. Rim 6 (Fig. 38a-e). Ten rim sherds pertain to neckless jars with slightly concave upper walls. The rim form and treatment are similar to those of Rim 2 although the exterior edge of the lip terminates in a dull point. Maximum thickness ranges from 6 to 10 mm. Rim diameters are among the largest of the Manachaqui Phase assemblage ranging from 22 to 29 em and averaging 24.3 em (N=8) in diameter. Two rims represent particularly deep vessels, and body profiles were likely globular rather than carinated. Rim 7 (Fig. 38f-j). Eight sherds comprise Rim 7. These rims end just short of reaching the point of vertical tangency at the orifice. They are gradually thickened to a maximum of 6 mm just under the lip. with rounded edges. The lip is semi-squared Several retro-fitted sherds reveal a semi-carinated body profile (Fig. 38j). Vessel walls are 2 to 2.5 mm thick. Rim 8 (Fig. 38k-m). Eight rims are unthickened or slightly thickened to 3 or 4 mm. The lip was trimmed flat 297 while the clay was quite dry, leaving small cracks still visible on the surface. Rim 9 (Fig. 38n-p). Rim 4 has seven examples. They are thickened at the lip which is rounded and 6 mm thick. An overhang of clay is folded over the interior edge of the lip. The vessels' upper walls are 3 to 3.5 rnm thick. Rim 10 (Fig. 39-e). Six rim sherds represent Rim 10 which features gradual curvature below the point of vertical tangency at the lip. The rims are unthickened and tapered toward the lip which has been lightly flattened. The vessels' upper walls are 4 mm thick. Rim 11 (Fig. 39f). Five rim sherds may belong to a single neckless jar exhibiting a direct rim. The lip shows exterior thickening and has been beveled flat on the top and exterior surface. Its orifice diameter is 11 em. The upper walls are 5 em thick. Rim 12 (Fig. 39g). Five rim sherds with gradual exterior thickening show a short, squared protrusion and a flattened lip. The maximum thickness is 8 rnm. The wall below the rim is 2 rnm thick. Rim 13 (Fig. 39h). Two rim sherds show exterior thickening in the form of a short, beveled flange. Including the flange, the rim's thickness is 7 rnm. The 298 vessel's upper walls are 3 mm thick. Shape B: Short-necked Jars Represented by 1,728 rim sherds, short-necked jars are the most common Paste A vessel type, comprising 81.1 percent of the Manachaqui Phase rim collection. The few partially reconstructed short-necked jars show semi-carinated and carinated body shapes. The occurrence of simple globular shapes is probable but remains undemonstrated. Statistics relating to diameters of rim forms with a sample size of 10 or greater are presented in Table 6. Orifice diameters range from 9 to 24 em, and most rim forms yield average diameters between 13 and 14 em. Interior surfaces remain unfinished while exterior surfaces were smoothed to a matte finish. Potters nearly always burnished the tops and interior surfaces of rims. This seems to be an important feature of Manachaqui Phase short-necked jars, perhaps related to the vessels' functioning. Two groups of 298 decorated rims are presented separately at the end of the Shape B rim form descriptions as Rims 9 and 10. They were not included within the existing morphological categories because decorative treatment frequently obscured classificatory criteria. Their larger average rim diameters may indicate different vessel size and shape tendencies. R~s la-c {Fig. 39i-aa). With 503 examples, Rim 1 is 299 the most numerous of the Manachaqui Phase Paste A rim forms. Examination of broken sherd edges reveals that the potters simulated vessel necks by reinforcing the exteriors of otherwise direct rims. The technique involved either bending the upper wall outward, then curling and tucking the end underneath the bent clay, or simply doubling the end back over the vessel exterior 5 to 7 mm. The folded clay was then joined to the vessel wall exterior, often haphazardly, with the fingers. Most surface traces of the "fold-over" technique were removed by beveling the rim surfaces flat, and then wiping the wet clay smooth. Wiping the interior surface of the rim often left a tiny cornice of clay protruding over the top of the rim which potters did not remove while burnishing the rim surfaces. The total thickness of the wall and folded portion ranges from 7 to 16 mm. While Rim la (250 examples) shows a beveled end and a rounded exterior (Fig. 39i-p), Rim lb (245 examples) was beveled flat on the exterior surface as well (Fig. 39q-x). Rim lc (eight examples) was beveled to create more than two surfaces, rendering a "prismatic" effect (Fig. 39y-aa). The larger sherds show that interior surfaces were invariably left rough and unfinished during manufacture. The rim illustrated in Fig. 40a shows the only handle found in the Manachaqui Phase assemblage. A hypothetical Shape B Rim lb vessel with a globular body is depicted in Fig. 40b. 300 Rims 2a, b (Figs. 40c-p, 4la-i). Rim 2 and the remainder of the Manachaqui Phase jar rims described below have true curved, albeit short, necks. Rim 2a (454 examples) is characterized by exterior thickening, often achieved by the fold-over technique just described, and a semi-squared lip (Figs. 40c-p, 4la, b). These rims typically lack the sharp edges left by beveling. An unusually large sherd with Rim 2a has a semi-carinated profile and an applique flange has been added to the vessel mid-section (Fig. 4la, b). The maximum thickness of Rim 2a varies between 5 to 10 mm. Rim 2b (41 examples) is left unthickened, but the lip is similarly semi-squared (Fig. 4lc-i) . It is 4 to 5 mm thick. Rims 3a, b (Fig. 4lj-w). Rim 3a (224 examples) is thickened, often by fold-over, and rounding of the lip frequently renders a bulbous appearance (Fig. 4lj-p). thickness varies from 5 to 10 mm. Its Rim 3b (N=67) is unthickened with a similarly rounded lip (Fig. 4lq-w). It is typically 4 mm thick. Rim 4 (Fig. 42a-h). Rim 5 (51 examples) is gradually thickened, reaching maximum thickness at the lip which has been trimmed flat. A highly variable neck curvature suggests that some of the corresponding vessel shapes were more "open" than others, perhaps functioning as "bowls." Fig. 42h shows a large rim sherd with applique decoration. 301 The maximum thickness of these rims varies from 5 to 8 mm. Rim 5 (Fig. 42i-p, 43a-c, 44a-d). These flanged rims (37 examples) project horizontally or slightly upward 8 or 9 mm from a vertical or insloping wall. The undersides of the flanges often present smooth curves from the pointed tips of the rims to the vessel walls. lightly flattened. The tips are pointed or Some examples bear applique decoration (Fig. 43a, b), while another has been decorated by incising large squared notches (Fig. 44a). Still others show deep incisions, occasionally combined with applique (Figs. 44bd) . Evidence suggesting that at least some of these vessel shapes were semi-carinated (Fig. 43c) consists of matching rim and medial flange decorative motifs. Decorative motifs and medial flanges are described in the section on decoration. Rim 6 (Fig. 45a-g). Rim 6 (25 examples) has been bent outward and its interior surface has been flattened. Most are lightly thickened, but taper toward the lip which may be rounded or pointed as the result of the interior flattening. At the neck angle, the rims are approximately 6 rnm thick, and they extend 10 to 18 mm. Rim 7 (Fig. 45h-l). Rim 7 (19 examples) has been thickened on the exterior surface or underside. The lip comes to a point where a tiny clay cornuice often protrudes 302 as a result of smoothing the interior surface. The maximum thickness ranges from 6 to 8 mm. Rim 8 (Fig. 45m-o, 46a). Rim 10 (nine examples) is distinguished by thickening, a beveled, horizontal upper surface and a pointed or lightly flattened tip which approximates a flange (like Rim 5). A partially reconstructed vessel shows a semi-carinated profile (Fig. 46a) and bears applique decoration. The walls are 3 mm thick and the maximum thickness reaches 7 mm. Rim 9 {with notched decoration) 48a-c). (Fig. 46b,c, 47a-f, Decorated Rim 9 (189 examples) includes profiles similar to those of Rims 2 through 4, 7 and 8. There are only two examples of Rim 1 vessels decorated by notching. Three rim sherds with particularly large diameters (21, 22 and 24 em) may represent a special size class of this vessel. Rim 10 {with incised decoration) {Figs. 48d,e, 49a-c). Rim 10 (109 examples) includes rim forms similar in profile to Rim 4, but unthickened variants are best compared to Rim 2b. Where clay was applied to thicken the rim exterior, it was usually haphazardly smoothed to the vessel wall. The process of cutting incisions deep into the lip displaced clay outward further "thickening" the rim. thickness varies from 5 to 9 mm. The resulting Rim interiors were always 303 burnished down to the point of maximum neck constriction, and a tiny clay cornice left during the wiping stage extends from the interior surface 1 to 2 mrn beyond the end of the rim. Fig. 49a shows a carinated body. Some rim and wall angles suggest that Rim 10 includes relatively open shape variants (e.g. Fig. 49c). Shape C: Restricted Bowls Six restricted bowl shape variants are suggested on the basis of rim morphology. The Shape C restricted bowl category consists of 113 rim sherds representing 5.3 percent of the Manachaqui Phase rim collection. The majority of these bowls (Rims 1 and 2) are believed to exhibit globular, semi-carinated or carinated body profiles although attempts to retrofit entire body profiles were again unsuccessful. Rim diameters normally range from 7 to 19 em, while the two most popular classes show diameters averaging approximately 14 and 14.5 em. Interior and exterior surfaces are carefully burnished. Statistical data are presented in Table 7. R~s la, b (Fig. 50a-q). A total of 68 rim sherds representing Rim 1 probably belong to convex bowls with globular body profiles. Both interior and exterior surfaces of these rims have been burnished to a low luster. walls range from 2 to 4 mrn in thickness. Vessel Typical examples of Rim la gradually thicken to a maximum 6 mrn (Fig. SOa-k). 304 Lips are rounded or show rounded interior edges and pointed exterior edges. The tip of Rim 1b's (11 examples) lip has been trimmed flat (Fig. 501-p). Fig. 50q offers an hypothetical Shape C Rim 1a bowl with a globular body profile. Rims 2a, b (Fig. 51a-o). The concave upper walls of Rims 2a and 2b (27 examples) suggest that these sherds belong to semi-carinated or carinated bowls. Like Rim 1 sherds, they show careful polishing, similar angles of orientation, occasional light thickening of the interior surface toward the lip and similar rim and wall thicknesses. The lip of Rim 2a (Fig. 51a-i) exhibits a sharp exterior edge and rounded interior edge. Rim 2b's lip (Fig. 51j-n) has been lightly flattened by trimming. Fig. 51o shows an hypothetical Shape C Rim 2a vessel with a semi-carinated body profile. Rim 3 (Fig. 52a-b). Eight examples of Rim 3 belong to small, shallow convex bowls with well-smoothed interior and exterior surfaces. The rims have been gradually thickened to 5 mm, and the lips show semi-squared profiles. Rim 4 {Figs. 52c-e). Seven rim sherds from small globular bowls are strongly inverted like those of Shape A Rim 4 vessels. The thickness of these rims ranges from 4 to 8 mm, and lips are rounded or semi-squared. Surfaces are 305 burnished and these bowls frequently carry applique decorations on their rims. Rim 5 (Fig. 52f). Two unthickened, direct rims with squared lips also belong to bowls. They are smooth, yet unburnished, and the interiors show remains of a thick white paint. They are 3 mm thick. R~ 6 (Fig. 52g). The sole example of Rim 6 belongs to a large globular bowl with well-burnished surfaces. It shows slight, gradual thickening to a maximum of 6 mm and the lip is semi-squared with rounded edges. Shape D: Unrestricted Bowls with Concave Walls There are 87 examples of unrestricted bowls with concave walls representing four percent of the Manachaqui Phase rim collection (Table 8). This shape class accommodates rim forms that apparently pertained to open vessels with concave upper walls and semi-carinated or carinated profiles. Again, none of the vessels could be adequately reconstructed to confirm suspected lower body contours. Rim diameters vary between 11 and 26 em, and are greater on average than diameters of Shape B short-necked jars. Both the rim and upper wall interior and exterior surfaces were usually burnished to a low luster. Because of the severe fragmentation of the collection, it could not be confirmed whether or not a small number of Shape B short- 306 necked jar variants belonging to Rims 4 and 10 might fit within this category. It should be recalled that, for the Manachaqui Phase shape inventory as a whole, the degree of neck constriction may be a continuous variable. The following rim forms were confidently assigned to this category. Rim 1 (Fig. 53a-h). These 32 everted rims are unthickened with squared, and occasionally semi-squared, lips. The rim orientations range from nearly vertical to approximately 45 degrees, and rim diameters are likewise highly variable. Sherd thickness varies from 4 to 5 mm. Rim 2 (Fig. 53i-l). Rim 2 (14 examples) is slightly everted and lightly thickened. The interior and end surfaces of the lip have been flattened to create a dull point. Their maximum thickness reaches 8 mm. Rim 3 (Fig. 54a-d). Rim 3 slightly everted and thickened. (13 examples) is also The lip exterior has been rounded rendering a dull point on the lip interior. thickness varies between 6 and 7 mm. Its Fig. 53m shows a hypothetical reconstruction of a whole vessel. Rim 4 (Fig. 54e-g). Rim 3 (13 examples) is also slightly everted and thickened, but the lip interior has been rounded leaving a dull point on the lip exterior. is 6 to 7 mm thick. It 307 R~ 5 (Fig. 54h-j). Rim 5 (13 examples) shows flange- like thickening on the lip exterior. The lips have been flattened on their upper and exterior surfaces. They are approximately 7 mm thick. R~ 6 (Fig. 54k,l). Two sherds belong to a single small bowl with a globular or semi-hemispherical profile and walls as thin as 1 mm. mid-section. An applique rib surrounds the vessel The rim interior is thickened to 5 mm, and the lip is pointed. Shape E: Unrestricted Convex Bowls The 26 rim sherds comprising this shape category represent one percent of the Manachaqui Phase rim collection and suggest a wide variety of vessel shapes and treatments (Table 9). Seven distinct rim profiles can be distinguished, and meager sample sizes for each range from one to five sherds. There remains some question whether Rim 3 represents an intrusive example of later non-Manachaqui Phase bowl rims. R~ 1 (Fig. 55a, b). Rim 1 (6 examples) belongs to a convex bowl with direct, vertical walls and rim. thickness varies from 2.5 to 5 mm. Wall The lip exterior has been thickened to a maximum 8 mrn, and the top of the lip has been flattened. R~ 2 (Fig. 55c, d). The six examples of Rim 2 pertain 308 to bowls with straight or slightly curved walls. The walls' angle of orientation varies from slightly insloping to slightly outsloping. The rim shows gradual thickening and the lip exterior has been thickened by applying a fillet of clay which was subsequently incised in a manner similar to Shape B, Rim 10. Maximum thickness at the lip ranges from 9 to 13 mm. Rim 3 (Fig. SSe-h). The five Rim 3 examples are direct, nearly vertical, and belong to simple convex bowls. They show gradual but slight interior thickening below an in-sloping, Rim 4 tapering lip. Thickness is 3 to 4 mm. (Fig. SSi, j). Three examples of Rim 4 have slightly concave walls and the interior surface has been thickened toward the lip. The lip is flat but slightly in- sloping, has rounded edges and is 7 mm thick. The lower walls are 3.S mm thick. Rim 5 (Fig. SSk, S6a). These three rims show thickened interior surfaces and rounded lips. well-burnished. Their surfaces are All three examples carry applique button decorations on the lip. Rim 6 (Fig. S6b). or two large open bowls. The maximum rim thickness is 8 mm. Two examples of Rim 6 belong to one The walls have been gradually thickened from 3 to 8 mm at the semi-squared lip. 309 R~ 7 (Fig. 56c). A single rim representing Rim 7 appears to belong to an large open bowl with polished surfaces. The rim is gradually thickened and ends at a squared lip. A small decorated ridge of clay protrudes from the interior surface beyond the lip. Miscellaneous Shapes Of the three sherds (Fig. 57a-c) from Manachaqui Phase strata that do not pertain to vessels, only the spoon handle depicted in Fig. 57a could be identified with respect to function. The dish-shaped end of the spoon is broken and does not show evidence of polishing or any special attention to surface finish. Such spoons occur with considerable frequency in Early Intermediate Period contexts. This example may be the earliest yet recovered in the Central Andes. It is doubtful that the decorated sherd shown in Fig. 57b belongs to a vessel because of its irregular shape. The interior surface is rough and unfinished. The third piece (Fig. 57c) has a smooth tabular shape with rounded edges backed by an irregular broken fragment. R~ Decoration Vessel decoration during the Manachaqui Phase is based upon plastic modeling techniques. It is additive in the sense that applique provides the basis for most decorative modifications. Potters applied ribs, buttons and adornos (defined here as representational protruberances with 310 zoomorphic or anthropomorphic features) to vessel rims, shoulders and mid-sections. Applique ribs and buttons were usually embellished by notching or deep incision. Fine-line incised designs and painting occur only rarely in the sample, and pattern burnishing appears on only one sherd. Weathering may have affected the quantification of sherds with these kinds of superficial decoration, but even the best-preserved Manachaqui Phase sherds from Sector A rarely show such techniques. A total of 1,816 rim and non-rim sherds showing decorative modification can be attributed to the Manachaqui Phase, and each one of the five shape categories described above is represented. It is tempting to assume that vessels with notched rims carried notched applique ribs on their mid-sections, and that incised rims and incised ribs were similarly associated. However, the evidence at hand shows plain rims with incised ribs and flanges (Figs. 38j, 41a) and notched rims with incised medial ribs (Fig. 48c; Plate VII) . It still may be reasonable to deduce that rims and flanges both bearing rare, identical adornos belong to the same vessel. Only a less fragmented sample of Manachaqui pottery can serve to disclose the rules or tendencies governing possible combinations of decorative elements. The following is a description of the techniques utilized to decorate Manachaqui Phase vessel rims and bodies. Confirmed associations between techniques and vessel shapes will be 311 offered wherever possible. Of the 2,129 Manachaqui Phase rims, 360 (16.9%) exhibit some kind of decoration. Decorations are notching, incision, applique buttons, adornos and red paint. Rim Notching The most common technique employed to embellish vessel rims during the Manachaqui Phase (216 examples) is notching of the lip. The inner and outer edges of squared lips, and the ends of rounded or pointed lips, are notched at intervals (3 to 20 mm) with a flat or rounded section of a stick or cana to produce 2 to 3 rnm wide U or V-shaped notches (Figs. 46b,c, 47a-f, 48a-c). Often, the aforementioned clay cornices left during manufacture were notched by the same technique (e.g. Figs. 46b, 47a and f). Where notches occur close together, "serrated" appearance. the rim presents a One Shape B Rim 5 sherd (Fig. 44a) shows unusually wide notches, while unusually thin notches appear on Shape D Rim 5 (Fig. 54j) and Shape BRim 8 (Fig. 45n) . The technique of directly notching the lip occurs on Shape B Rims 2, 4, 7 and 8 necked jars, and on Shape D Rims 2 and 5 and Shape E Rim 7 unrestricted bowls. On Shape A Rims 2 and 3 neckless jars, the notching is executed on an applique band attached to the lip exterior (Figs. 36a, b, 37a-c). 312 Rim Incision The second most common technique with 133 examples involves jabbing large punctations directly into the end of a thickened lip with the end of a stick or cana (Figs. 44bd, 48d,e, 49a-c). The gouging tool repeatedly struck the surface at an angle of approximately 45 degrees as the potter rotated the vessel. The jabbing motion either originat·ed directly over the rim, or angled toward or away from the vessel interior. The resulting punctations vary in depth from 2 to 5 mm, reach 9 mm long and 4 mm wide and range in distance from one another from 1 to 9 mm. Gouging of large punctations frequently gave the rim exterior an uneven, swollen appearance. However, the interior of the rim and neck was always maintained even and well-smoothed. Incised rim embellishment occurs on Shape B necked jars (Rims 5 and 10) and Shape E unrestricted bowls (Rim 2). Occasionally the technique occurs in combination with applique appendages like buttons and adornos. Applique Rim Appendages Eleven Manachaqui Phase vessel rims were also embellished with buttons and adornos. Seven sherds bear buttons that appear as small disks approximately one centimeter in diameter. They may be notched or incised to render a "slashed" appearance. A button on the tip of neckless jar Shape A Rim 6 (Fig. 38a) was grooved by impression with a thin implement like a stick, while a 313 button perched on the edge of restricted bowl Shape C Rim 4 (Fig. 52d) shows two incisions cut with a sharp point. Buttons on the flat top of necked jar Shape B Rim 5 (Fig. 44c and a similar example not illustrated) were fashioned into doughnut-like circles before fastening to incised rims. The buttons on the tips of three Shape E Rim 5 unrestricted bowls (two illustrated in Fig. 56a,b) appear to have been altered by pressing a small, nearly square-shaped implement into their centers. Four rim sherds are festooned with adornos. In addition to buttons, Shape B Rim 5 exhibits thin, undulating rolls of clay suggestive of serpents (Fig. 43a, b). A Shape B Rim 5 sherd with incision carries a broken zoomorphic or anthropomorphic adorno with only a preserved set of arms and hands (Fig. 44d). Fingers were rendered by incision, and the arms show circular punctations. off. The head has broken The small modelled face attached to the lip of Shape D Rim 1 probably represents a snake (Fig. 53a). Vessel Body Decoration The Manachaqui Phase ceramic assemblage includes 1,456 decorated body sherds. For the purposes of description, the 1,418 sherds with applique are divided into categories of: 1) medial and shoulder ribs and flanges, 2) notched and incised bands and 3) other applique such as buttons, adornos and lugs. Described last are 31 sherds with fine-line incised decoration. The 1,258 examples of sherds with 314 applique medial and shoulder ribs and flanges are either unembellished, notched or incised. Rolls of clay modeled into ridges with triangular cross-sections between 4 and 9 rnm wide were applied horizontally to encircle the exteriors of vessels with globular, semi-carinated and carinated profiles. Potters fastened ribs by smoothing the upper edge to the vessel walls. edge and the wall. Less care was taken to join the bottom Sherds exhibiting rib ends indicate that medial and shoulder ribs did not always completely surround the vessel. Appendages identified as flanges are discontinuous, short in length and protrude higher from vessel surfaces than ribs. The final burnishing of vessels followed the application of ribs and flanges. Unembellished Ribs Unembellished applique ribs are simply affixed coils of clay that rise 2 to 7 rom above the vessel wall surface (Fig. 58a-d}. There are 146 examples. The high percentage of sherds exhibiting both unembellished ribs and unsmoothed interior surfaces (Table 10} suggests a close association with Shape A and/or B jar shapes. Only rarely do unembellished ribs appear on vessel shoulders above the midsection (Fig. 58e-g}. Notched Ribs Notched medial and shoulder ribs are modified with the basic technique of impression utilized on notched rims 315 (Shape B Rim 9). Of the 350 notched ribs, the majority (196) exhibit notches impressed directly and vertically into the clay (Fig. 58h, i). Oblique, vertical strokes produced characteristic notches on a total of 55 sherds with ribs (Fig. 58j, k), direct and diagonal on 24 (Fig. 581, m), and oblique and diagonal on 37 (Fig. 59a, b). One of two notched rib variants, Notched A refers to ribs averaging approximately 4 mm wide and 2.5 mm high which invariably occur in the medial position. Three out of four Notched A sherds show smoothed or burnished interior surfaces, suggesting association with bowl shapes C and/or D. Thirty- eight sherds with Notched B feature relatively small and shallow notches on higher ribs protruding an average of 5.5 mm above the surface (Fig. 59c, d). Even higher percentages of Notched B sherds exhibit smoothed or polished interior surfaces, also suggesting association with bowl shapes. Because most Notched B ribs occur on the vessel shoulder rather than in the medial position a specific association with Shape C bowls is likely. Examples of Notched B ribs in both medial and shoulder positions can be observed on Shape C Rim 4 and Rim 6 restricted bowls (Fig. 52e and 52g respectively) . Incised Ribs and Flanges Incised applique ribs and flanges on a total of 617 sherds were decorated by two techniques. The most common with 532 examples and termed Incised A in Table 6, involved 316 cutting large gashes in the upper surface of medial ribs with a long, thin blunt or hollow-ended implement (Fig. 59eg) . The potter rotated the vessel (usually clock-wise) while rhythmically cutting into the rib surface with a jabbing motion. Drag marks left on the vessel walls by the jabbing implement are still visible where not erased by subsequent burnishing. curling tail (Fig. 60a). One Incised A rib terminates in a Incised A ribs occur with Shape A Rim 7 neckless jars (Fig. 38j) and Shape B Rim 9 (Fig. 48c; Plate VII) and Rim 10 (Fig. 49a) necked jars with notched rims. Eighty-five examples of the Incised B technique feature deep parallel incisions cut into flanges up to 8mm high (Fig. 60b-e). The incisions were produced either by cutting down and away from the vessel body from above with a slashing motion, or by cutting across the upper surface of the flange towards the vessel wall with a stabbing motion. This latter motion often left round impressions where the implement penetrated the vessel wall. Most, or perhaps all, Incised B flanges are discontinuous, and some resemble bird wings or fish fins. An Incised B flange decorates a Shape B Rim 2a jar in Fig. 41a and b. Applique Bands A total of 149 sherds show that Manachaqui Phase potters also applied bands of clay vertically and diagonally to vessel shoulders, often in conjunction with medial ribs 317 (e.g. Figs 48c; 49a). Here the term "bands" distinguishes these decorative elements from the horizontal medial and shoulder ribs and flanges just described. Of these, 118 (84%) were embellished with the same Notched A technique observed on medial ribs. Seventy of these sherds (59%) exhibit smoothed or polished interior surfaces (Fig. 60f-i), while surfaces of the remaining 48 (41%) were left unfinished (Fig. 60j, k; 61a, b). Rim sherds and partially reconstructed vessels show associations with Shape B Rim 8 (Fig. 45m, n; 46a), Rim 9 (Fig. 48c; Plate VII) and Rim 10 (Fig. 49a) short-necked jars, and Shape C Rim 1a (Fig. 50ik), Rim 2a (Fig. 51i), Rim 2b (Fig. 51n) and Rim 4 (Fig. 52c) restricted bowls. Fourteen sherds bear applique bands with circular punctation (Fig. 61c, d). These bands protrude approximately 6 mm from the vessel surfaces. They appear on the shoulders of Shape A Rim 4 neckless jars, a Shape C Rim 1b restricted bowl, as the tail of an Incised A medial rib from a vessel with a burnished interior (Fig. 61d), associated with an Incised A medial rib on a Shape D Rim 6 unrestricted bowl (Fig. 54k) and as the "arms" of the broken adorno of Shape B Rim 5 (Fig. 44d). There are six examples of applique bands with ovoid incisions. surfaces. These bands protrude 5-6 mm from vessel Two sherds, one from a rim (Fig. 42h) and another bearing an otherwise unembellished shoulder rib (Fig. 61e), 318 may belong to the same Shape B Rim 4 short-necked jar. One of two nearly identical sherds with burnished interior surfaces shows association between bands (perhaps parallel) with ovoid punctation, buttons with ovoid punctations and Incised A medial ribs (Fig. 61f). Nine examples of applique bands from Manachaqui Phase stratigraphic contexts are too eroded to further identify. A sherd on which the applique band has broken off features what appears to be a serpent head (Fig. 61g). A single applique band which might be a "renegade" Incised A medial rib meanders serpent-like across the shoulder of a vessel with an unfinished interior surface (Fig. 62a). Other Applique Appendages Other applique appendages that Manachaqui Phase potters affixed to vessel bodies include buttons, adornos and a single lug. The single lug appears as a knob where a vertical notched band meets a Notched A medial rib on a sherd with a burnished interior (Fig. 62b). Of four body sherds with buttons, two were described above in association with applique bands (Fig. 61f). Another body sherd with an unfinished interior surface features a button with two parallel slashes resting atop an unembellished medial rib (Fig. 62c). The fourth sherd has a finished interior, a button with four parallel slashes and an Incised A medial rib (Fig. 62d) . Of eight adornos from vessel bodies, six are broken 319 from mid-sections of vessels of unknown shape, while the seventh adorned the shoulder of a Shape A Rim 4 neckless jar (Fig. 37c). This seventh adorno appears as a protruding snout with applique eyes affixed to either side. Like other Manachaqui Phase adornos, it is zoomorphic, rendered in abstract fashion, and cannot be easily identified as to species. Three sherds with snake adornos on unembellished medial ribs (two are shown in Fig. 62e, f) probably belong to the Shape B Rim 5 necked jars with identical adornos atop their rims (Fig. 43a-c). Fig. 63a illustrates the modeled head of an amphibian or fish. Like the Shape A Rim 4 adorno just described, it was formed using small bits of shaped and incised clay to render facial features. This adorno, and another which may represent a bird wing or fish fin (Fig. 63b), come from vessels with unfinished interior surfaces. The remaining two adornos are knob-like protrusions, apparently representing heads, one with only a mouth (Fig. 63c) and another with only eyes (Fig. 63d). Lack of detail inhibits speculation with regard to species represented. Incised Decoration A total of 31 sherds from three Manachaqui Phase vessels exhibit incised decoration on their exterior surfaces. All of the sherds show burnished interior surfaces. Incisions were cut while the clay was still moist after smoothing or burnishing the entire exterior. None of these were included in Appendix D totals for incised sherds. 320 Xncised Vessel 1 (Fig. 64a-c). Twenty sherds with zoned punctation pertain to a single vessel. Despite the burnished interior surfaces of this "bowl," the rim form is most comparable to Shape B Rim 4 necked jars. This inconsistency demonstrates the potential pitfalls of applying a strict "bowl" vs. "jar" dichotomy to the Manachaqui Phase collection. The design exhibits parallel vertical lines 16 mm apart that serve to create zoned fields alternately filled with triangular or dashed punctations. The entire exterior of this vessel was burnished prior to incision. The motif begins just under the rim and ends at the basal angle or corner point. Xncised Vessel 2 (Fig. 64d, e). Ten sherds probably pertain to a single restricted vessel with a carinated profile and unknown rim form. Two partial reconstructions show zoned triangular and step-shaped fields filled with thin, slash-like punctations. The exterior surface was smoothed to a matte finish prior to incision. Lines discernible on one of the reconstructions (Fig. 64e) suggest that the triangular and step-shaped fields may alternate around the exterior of the vessel. Finally, a sherd from another vessel displays deep and relatively wide incisions (Fig. 64f). Two vertical parallel lines border a series of short horizontal, parallel lines. 321 Pattern Burnishing and Painting Pattern Burnishing and painting are two techniques which may be under-represented in the Manachaqui Phase collection because of its weathered condition. A single sherd (not illustrated) exhibits exterior surface burnishing with a spider web-like pattern similar to that on a Pacopampa bowl illustrated by Rosas (1976:572, Lamina Sa). Six sherds exhibit remains of red paint. One is a Shape B Rim 1b necked jar rim sherd with red paint or slip on its exterior. Four articulating sherds from a Shape D Rim 5 unrestricted bowl with concave walls (Fig. 45h) show red paint on both surfaces, except on the rim proper. One sherd found in Sector A and pertaining to a Shape D Rim 1 unrestricted bowl bears traces of post-firing red paint on both surfaces. The sherd's relatively dry Sector A context allowed preservation. Ceramic Paste Group B The Manachaqui Phase ceramic assemblage includes 109 sherds assigned to Paste Group B. Twenty-one Paste Group B rims comprise one percent (0.98%) of the total 2150 Manachaqui Phase rims. As a group the Paste B pottery is distinguished by glittery surfaces, yet as many as three pastes or wares may be represented. Manachaqui Phase Paste B pottery differs from Paste A not only in the pastes employed, but in details of vessel shape and decorative techniques. These Paste B vessels may have been traded to 322 the Paste A-producing population, or they may have been carried in by persons that only rarely frequented the shelter. Possible sources will be discussed later. In general these pastes are even more compact and finegrained than Paste A. A cursory microprobe analysis identified quartz, alkali feldspar and ilmenite mineral constituents. Clear crystalline quartz and mica produce the glittery effect. Non-plastic inclusions are well-sorted except for occasional large grains of vein quartz. The pottery colors range from a light tan to a dark grayish brown. Many sherds show incomplete oxidation and, because the fabric is softer than Paste A pottery, sherd surfaces tend to be more eroded. Nevertheless, a glossy finish observed on well-preserved specimens attests to burnishing. Paste 8 1 Fifty-six sherds from a minimum of five vessels share the same paste referred to as Paste B1 • Surface colors are 7.5YR5/6 (strong brown) and 10YR7/3 (very pale brown). Vessel 1 (Fig. 65a) is a short-necked jar with an everted rim, mouth diameter of 13 em, and body (exterior) diameter of 16 em. The rim is lightly thickened and the body profile is carinated. Vessel 2 (Fig. 65b) is a neckless jar with a mouth diameter of 14 em and an unknown body profile. The rim exterior has been thickened, beveled and grooved below the lip. Vessel 3 (Fig. 65c) is a carinated, restricted bowl with a direct, unthickened rim and a mouth diameter of 323 13 em. Red, post-firing paint (10R4/4 weak red to 10R4/6 red) covers the exterior but, like the Shape D Rim 5 bowl described earlier (Fig. 45h), the end of the rim was left unpainted. Vessel 4 (Fig. 66a) is a restricted bowl represented by small rim sherds with an unusual profile. The lip exterior has been thickened and brought to a dull point. interior shows a shallow groove under the lip. The rim The five sherds that constitute Vessel 5 (Fig. 66b) probably belong to a double-spout-and-bridge bottle, although they may also represent an asymmetrical bottle with a single spout and strap handle (DeBoer n.d.). The hypothetical reconstruction illustrated here is based upon a double-spout-and-bridge bottle illustrated by Lathrap (1962: 935, Fig. 49e; 1970:91, Fig. 9j), and assigned to the Late Tutishcainyo Phase. Paste B2 The 31 sherds consisting of Paste B2 belong to a single carinated vessel (Fig. 66c; Plate VIII) that was almost entirely reconstructed. The paste is distinguished from Paste B1 by larger non-plastic grains and a less compact fabric. The rim is unthickened and sharply everted, or flanged, with a diameter of 9 em. The lip has been squared. Beneath the lip exterior, two notched applique bands run vertically down the upper vessel wall to end at the corner point of the basal angle. The basal angle has been decorated with small nicks encircling the entire vessel. 324 Paste ~ Paste B3 contains less mica and larger quantities of crystalline quartz. Its color is brownish gray. These sherds are more difficult to distinguish from Paste A, but sixteen were isolated to reconstruct a jar with an angled neck and an everted rim 13 em in diameter (Fig. 67a). The body profile is globular although it carries a notched medial rib. The notching is deeper and spaced at wider intervals than Notched A decoration observed on Paste A vessels. Arch motifs rise from the medial rib at opposite sides of the vessel (only one of the two is depicted in Fig. 67a) . Lithic Remains Because of the stratigraphic conditions at Manachaqui Cave, the assignment of stone artifacts to particular phases presents some difficulties. As previously noted, the ceramic macrochronology serves as a reference for situating lithic remains within the established chronological framework. Examination of the macrochronology reveals that chipped-stone tools and debitage virtually disappear from the archaeological record after the Terminal Preceramic Period, and do not reappear in substantial numbers until the Early Intermediate Period. Of the approximately 9,500 lithic artifacts from the ceramic-bearing deposits, only an estimated 50 (or 0.5%) may truly pertain to the Manachaqui Phase. None of the retouched chipped stone tools from the 325 rockshelter's ceramic-bearing deposits can be confidently assigned to the Manachaqui Phase. In contrast, nine ground stone artifacts and cobble tools were retrieved from unmixed Manachaqui Phase levels. Ground and polished points of shale and slate constitute the most numerous class of ground stone artifact during the Manachaqui Phase. Although 72 such points were recovered from the excavations (Plate IX), not one remains intact. Like the pottery, these brittle points probably suffered breakage from trampling. Four point mid-sections and a tip were unearthed from Manachaqui Phase deposits. Mixed Manachaqui and Suitacocha Phase strata yielded seven more point fragments. Evidence that points were manufactured at Manachaqui Cave consists of small stone chips and preforms. Twelve chips from Lavasen Phase strata, and four preforms or blanks from mixed Lavasen and Manachaqui Phase strata provide evidence of experimental shale and slate working at the end of the Precerarnic Period. Other slate artifacts include a disk with two drilled perforations and incised lines on both sides (Fig. 67b), and a tiny tubular bead recovered while screening Floor Z soil samples. Of six ground stone tools and tool fragments, three were recovered from unmixed Manachaqui Phase contexts in Sector B. These include a porous sandstone cobble (from Unit 24, Level 15) faceted and grooved by abrasion that 326 would have been an appropriate implement for shaping the shale and slate points. It was found close to a disk-shaped pebble (from Unit 24, Level 16) with a concave surface that might have served for polishing or sharpening the points. The third piece is a granite cobble fragment that lacks clear signs of wear, but may have been shaped by pecking. The three remaining pieces from mixed layers were unearthed in Sector A. These include a hand-size cobble of felsic igneous rock with wear suggesting use as a milling stone and hammer (Unit 11, Level 7). The edges and one surface show polish, and the ends show chipping from blows. Two flakes from a finely ground and polished tool like an axe bit may have served as cutting tools. Rocks and Minerals Numerous rocks and minerals brought from outside the Manachaqui Valley were recovered during excavation of Manachaqui Cave. Seven chunks of white, angular vein quartz appear in Manachaqui Phase deposits and the layers immediately above and below. Uses for these minerals remain unclear, but quartz deposits in and around Pataz frequently bear sulfide ores such as gold and silver. No such ores were observed, and the chunks may simply have been appreciated as hard stone with sharp edges for cutting and chopping. A chunk of pegmatite, a small fragment of prismatic quartz crystal and two platy pieces of mica were also recovered from Manachaqui Phase levels. 327 Botanical Remains Because only the samples extracted from Sector A were analyzed by Dr. Pearsall, ethnobotanical information predating the Terminal Preceramic Period is lacking. The botanical taxa identified from Manachaqui Phase Floors AA and z are summarized in Table 18. These include fruit rinds of the Sapotaceae family similar to lucuma (Pouteria spp.) These and other unidentified fruit rinds may represent "exotic" foods brought from the lower elevations of the montane forest to the east. Foods that probably originated locally in the Tropical Alpine Zone near Manachaqui include tuber and root fragments that likely belong to sedge (Cyperus or Scirpus). Seeds identified are Festuca and either Chenopodium or Amaranthus, all native to the paramo grassland formation. All of these taxa, plus Lupinus; Ribes; Rubus; and Polygonum/Rumex seeds occur in samples from Terminal Preceramic Lavasen Phase Floors, which suggests continuity in the rockshelter's economy into the Initial Period. Faunal Remains A total of 72 faunal specimens were recovered from unmixed Manachaqui Phase levels in both sectors, and 52 were identified as Mammalia by Kent (Table 21). Represented are guinea pigs, mountain paca, white-tailed deer and armadillo. That only two of 72 specimens could be positively identified to the species level is a function of the faunal material's 328 • severe fragmentation. Comparison of the Manachaqui Phase inventory with the Terminal Precerarnic Period inventory is problematic because Sector B's unmixed Terminal Preceramic Period levels are difficult to isolate without further artifact analyses. Sector A's Lavasen Phase strata yielded only seven of the 97 identified taxa recovered from all of Site M-1A's Preceramic Period deposits (Table 19). All seven specimens are Mmmnalia, two are identified as Artiodactyla, and one as Cavia. Of the additional 90 Preceramic Period specimens from Sector B, 88 are Mammalia and two are Aves. One Mammalia specimen belongs to the Caviidae Family, while one Ave specimen is of the Strigiformes Family. Specimens from mixed Lavasen and Manachaqui Phase strata were also tabulated (Table 20). Manachaqui Phase Chronology Unfortunately, radiocarbon evidence processed so far does not offer a clear idea of the Manachaqui Phase's beginnings. The three radiocarbon dates of 850, 860 and 900 b.c. date only the upper portion of the Manachaqui Phase deposit which, on the whole, is thicker and richer in artifact density than deposits of any subsequent phase. Radiocarbon dates for the Lavasen Phase (1880, 1720 and 1570 b.c.) and the Suitacocha Phase (790, 680, 610, 500 b.c.) bracket the Manachaqui Phase occupation approximately between 1500 and 800 b.c. Also, future research may allow 329 the subdivision of the Manachaqui Phase into early and late sub-phases. Potentially late in the phase are the incised vessels of Fig. 64, and most of the Paste Group B vessels (Figs. 65-67a). Although present evidence is suggestive, my efforts to sub-divide the phase have not succeeded. The dates of 1500 to 800 b.c. for the Manachaqui Phase generally agree with radiocarbon evidence for the chronological placement of other early Initial Period phases such as Kotosh Wairajirca (Izumi and Terada 1972:308), Yesopampa (Terada 1979:173-174), Early Huacaloma (Terada and Onuki 1985:267-268; Onuki 1993:Fig. 15, p.91) and Pandanche A (Kaulicke 1981:388). Comparisons to be offered in the following sections suggest that pottery styles associated with these north highland phases are coeval. The sets of attributes shared with southern Ecuadorian Formative Machalilla and Chorrera Phase styles, during the Manachaqui and Suitacocha Phases respectively, suggest a later beginning date for the Manachaqui Phase around 1200 b.c. The shared attributes will be described below and in Chapter 7. Manachaqui Phase Ceramic Relationships This examination of relationships during the Manachaqui Phase will focus on the shelter's abundant pottery. The evaluation of migration hypotheses requires detailed comparative analyses and the identification of specific shared attributes. At issue is whether the Manachaqui Phase 330 Paste A assemblage can or should be regarded as a derivative of ceramic traditions in neighboring regions to the west, south, east or north. This comparative analysis of Manachaqui and other early Initial Period pottery styles also seeks to elucidate aspects of cultural geography, spheres of interaction and sources of innovation during the Preceramic-Initial Period transition. The Manachaqui Phase Paste A style should contain evidence for: 1) the cultural affiliation of Manachaqui Cave's users (e.g. Andean or Amazonian) and 2) cultural interaction on a variety of scales. Of course Manachaqui Phase Paste B pottery provides the most direct evidence for interaction, and its abundance provides one gross measure of the interaction's intensity. The interpretation of the Manachaqui Paste A style's relations requires assumptions regarding the relative significance of certain attributes. One reason that Lathrap and Meggers can derive conflicting interpretations of historical relationships from the same data stems from their disagreements regarding which shape or decorative attributes are diagnostic of migrations or diffusion (Meggers and Evans 1983:330-331}. Simple guiding principles utilized in the interpretations offered here and in subsequent chapters are that: 1} a relatively high number of shared attributes is most likely to indicate particularly intense interaction, and 2} statements regarding directionality are credible only where temporal priority has been demonstrated for one of the 331 ceramic complexes in question. As a general rule for the purposes of this study, the co-occurrences of single attributes, unless they are complex (e.g. double-spout-and-bridge bottle shapes), will be accorded little significance. Co-occurrences of multiple attributes will be granted greater validity, especially when similarities in both vessel shape and decorative attributes converge. It must also be remembered that, because of Manachaqui Cave's presumed special functions, its pottery sample most likely represents a subset of a more diverse "local" tradition. Therefore, the absence of ceramic attributes at Manachaqui (negative evidence) must be carefully considered before it can be accorded significance. The final chapter of this thesis will synthesize the evidence for interregional interaction during this and subsequent phases in order to illuminate trends that accompany the emergence of complex Central Andean societies. Paste A Relationships: the Central Andes The Manachaqui Phase Paste A assemblage has surprisingly little in common with the relatively simple pottery styles found in the heart of the Central Andes immediately west and south of Manachaqui Cave. Early Guafiape (Strong and Evans 1952), Las Haldas Phase 1 (Fung 1972), Yesopampa (Terada 1979), Toril (Burger 1985b) and other Central Andean assemblages feature neckless ollas and 332 simple convex bowl and bottle shapes. The globular variants of Manachaqui neckless jars (e.g. Shape A Rims 2 and 6) do resemble the ubiquitous Central Andean neckless ollas, but these constitute less than eight percent of the Manachaqui Phase Paste A assemblage. Manachaqui Cave's special function may account for the scarcity and relatively small size of these vessels. However, Manachaqui neckless jar shapes also tend to be shallower and more open than their Central Andean counterparts (further discussion of this shape will follow below) . Convex open bowls typical of the Central Andes are rare at Manachaqui where gourds could have been used as dishes for food or drink. Perhaps the dissimilarity between the Manachaqui Style and styles to the west and south is most clearly pronounced in the absence of necked jar shapes, reinforced and beveled rims, and carinated body profiles in early Initial Period Central Andean assemblages. Most early Central Andean pottery styles feature simple decorations rendered by rudimentary incision, punctation and applique techniques. In contrast, Manachaqui Style decoration is based almost exclusively on relatively elaborate applique and modeling techniques. Many of the Manachaqui Phase techniques can be found in isolation in the Central Andes. For example, rim notching occurs at Caballo Muerto in the Meche Valley (T. Pozorski 1983: Figs. 13, 17). However, it appears on neckless olla rims and the 333 technique's frequency and temporal position remain to be assessed. Also, gouged incisions and rim notching can be observed on purportedly "early" neckless ollas and convex bowls in Santiago de Chuco (Perez C. 1988: Lam. 1). Manachaqui Phase pottery blends these and other singular decorative features to render a unique and highly distinctive style. One·ceramic assemblage in the heart of the Central Andes to the southwest deserves special note because of a number of attributes shared with the Manachaqui Style. La Galgada in the middle Santa Valley yunga yielded a small sample of Initial Period ceramics including vessels with carinated profiles (Grieder 1988: Fig. 156gg), medial applique ribs (Ibid.: Fig. 156aa), incipient necks (Ibid.: Figs. 156ee, ff, gg and ii) and flattened and beveled rims. These La Galgada pieces are strikingly similar to the earliest Manachaqui ceramics. Grieder (Ibid.:185) reports that pottery appears at La Galgada between 1790 b.c. and 1565 b.c. Nearer to Manachaqui Cave and directly west across the Marafion Canyon, several investigations in the Huarnachuco area have failed to isolate early Initial Period occupations. Surface finds assigned to the Initial Period Marnorco Phase by Thatcher (1979: Figs. 5-8) include sherds from necked vessels with thickened rims resembling Manachaqui Phase Shape B jars. However, recent efforts by 334 Topic and Topic failed to corroborate Thatcher's Mamorco Phase, even at Cerro Mamorco (Theresa Topic personal communication 1987). Neckless ollas and brushed pottery recovered during the Topics' surface reconnaissance at Cerro Huachac may date to the early Initial Period, but later materials predominate (T. Topic personal communication 1994). None of these authors report early carinated vessels or other design features diagnostic of the Manachaqui Style. The Manachaqui style most closely resembles the earliest pottery at intermontane valley and western slope sites to the northeast in Cajamarca Department. The best- documented assemblages come from Huacaloma (Terada and Onuki 1982, 1985), Cerro Blanco (Terada and Onuki 1988) and Pandanche (Kaulicke 1981) in the intermontane highland valleys, and Montegrande in the middle Jequetepeque valley yunga (Ravines 1982; Ulbert 1994). Typical Central Andean neckless ollas predominate in these north-Central Andean assemblages, yet the presence of necked vessels with carinated and semi-carinated body profiles distinguishes them from more southerly Central Andean assemblages. Early Huacaloma Phase neckless olla, carinated jar, concave wallbowl, spherical bowl and open bowl shape categories loosely correspond to Manachaqui Shapes A through E. Cerro Blanco's La Conga Phase shape inventory is more closely analogous to Manachaqui's because of the predominance of carinated vessels and the dearth of open bowls. Open bowls are 335 entirely absent during Phase A at Pandanche where restricted vessel shapes were likewise preferred. In general, the shape proportions of Manachaqui Phase vessels differ from those found in the Cajamarca region. Manachaqui vessels tend to be narrower, necks are consistently shorter and carination angles occur one or two centimeters lower on vessel bodies. True carination angles are far more common in Cajamarca than at Manachaqui where potters mimicked them by attaching applique ribs to the midsections of semi-carinated and globular vessel bodies. At this early stage of research in northern Peru, it is difficult to evaluate the effect of site function on shape inventories and design features. Ceramic assemblages at Huacaloma, Cerro Blanco, Montegrande, Pandanche and Manachaqui ostensibly reflect ceremonial, domestic and traveling activities. Cajamarca's repertoire of decorative techniques is larger and more diverse that Manachaqui's, yet many of the applique and modeling techniques common to both areas differ in details. Montegrande potters habitually notched clay strips applied to vessel rims in addition to notching the rim directly. Some decorative techniques such as grooving and cane-stamping (Terada and Onuki 1982: Pl. 73-3, 10; Ulbert 1994: Lam. 36-1, 2) are not found at Manachaqui Cave and may be unique to the Cajamarca sites. Among other decorative techniques rare or entirely absent at Manachaqui 336 are the simple incised line motifs, zoned punctation and reddish slip at Montegrande and Huacaloma, and the brushed decoration at Pandanche. Because the rarity or dearth of some of the Cajamarca techniques at Manachaqui may be due to the rockshelter's special functions, the Manachaqui Phase assemblage's singular qualities are more significant. Unique to Manachaqui are techniques of rim treatment and lip incision, while rim embellishment is more common. Also characteristic of Manachaqui are the inordinate emphasis on "false" carination angles rendered with medial ribs, and the varied incision techniques employed on rims and applique ribs. Modeled adornos occur at Montegrande, but apparently not in the Cajamarca Basin. Curiously, Manachaqui also shares distinctive rim thickening techniques and varieties of carinated bowl shapes (Ibid.: Lams. 9-13), two-tiered shoulders (Ibid.: Lam. 1) and applique flanges and notched strips (Ibid.: Lams. 33-35, 17) with Montegrande that it does not share with the closer Cajamarca highlands. Some of Montegrande's zoned-incised decorative motifs (Ibid.: Lams. 14-3760, 14-3863, Taf. 37-7) are virtually identical to those of the Manachaqui Phase. Unlike Manachaqui's, the Montegrande shape inventory is still dominated by neckless ollas (Ibid.:58). Jar shapes typically show longer necks and all are assumed to have globular bodies. Montegrande pottery's resemblance to the 337 Manachaqui assemblage is difficult to evaluate because assemblages belonging to more than one phase appear to be conflated. Ulbert points out (Ibid.:149) that all of the Cajamarca sites share carinated vessel shapes not found farther south. South of Manachaqui Cave lie eastern Andean sites in similar geographic and ecological circumstances that have also produced early assemblages. At 3,800 m elevation on the Marafion-Huallaga divide in Huanuco Department, Piruro (Bonnier 1988) occupies a location analogous to Manachaqui's near the montane forest edge. Illustrations (Rozenberg 1982: Planche 6) show that the classic Central Andean neckless olla again constitutes the predominant vessel shape, and Rozenberg (Ibid.:135) expresses puzzlement at not encountering similarities to the coeval Kotosh Wairajirca Style nearby, nor any clear traces of Amazonian influence. Approximately 90 km farther south in Huanuco Department, local topography creates an arid rain-shadow, and highland-like environmental conditions extending down the eastern slopes below 2,000 m. At the site of Kotosh, necked jars with thickened and beveled rims comprise a sizable portion of the Wairajirca assemblage. Wairajirca Red Plain Form 6 (Izumi and Terada 1972: Plate 119, Figs. 9- 13) constitutes the second most popular plain ware vessel shape after the neckless ella. Necked jars also occur in significant numbers as Wairajirca Red Line-burnished Form 3 338 (Ibid.: Plate 125, Figs. 11 and 12), Wairajirca Zoned Hachure Form 1 (Ibid.: Plate 122, Figs. 1-4) and Wairajirca Black Line-burnished Form 2 (Ibid.: Plate 124, Fig. 6). These shapes have identical counterparts in Manachaqui's Shape B Rim 1 and Rim 2 globular jars. Wairajirca spherical bowls match Manachaqui's Shape C Rim 1 bowls, although this somewhat generic shape is less useful for comparative purposes. Despite the Kotosh potters' familiarity with carinated vessel shapes, Wairajirca short-necked jars reportedly have only globular or "boat-shaped" bodies. Eccentric vessel shapes lend Wairajirca a slightly more varied shape repertoire than Manachaqui's. The most conspicuous differences between Wairajirca and Manachaqui ceramics are found in modes of surface treatment. Wairajirca potters emphasized incised and post-firing painted decoration, and eschewed applique. However, they did affix zoomorphic adornos to vessels (Kano 1979). features are unique to Kotosh. Numerous other design These include the eccentric shapes, and modelled zoomorphic and anthropomorphic faces (Kano 1979) . Continuing southward along the eastern highland edge, earliest ceramics from the upper Mantaro Valley puna of Junin Department have been only superficially studied. However, Silva (1988:Figs. 13-21) describes a collection from Pachamachay that includes short-necked jars with 339 thickened, and often beveled, rims. He regards the collection as no earlier than 1200 b.c., and likens the jar rims to Kotosh Plain jar Forms 5, 6 and 7 {Izumi and Terada 1972: Pl. 114:20-30). The Pachamachay jar rims, and the few Initial Period jar rims from nearby Telarmachay {in the upper Tarma River drainage) illustrated by Lavallee {1977: Lam. 6b, d), are comparable to the previously cited Wairajirca Red Plain Form 6 and Manachaqui Shape B rim forms (e.g. Figs 42a-g this thesis). Apparently, neither the carinated vessel shape, nor the diversity of decorative modes that characterize Manachaqui and Kotosh appear in the earliest Junin assemblages. At Waywaka in the upper Apurimac River drainage radiocarbon dates ranging from 1600 to 710 b.c. justify comparisons between the Manachaqui Phase assemblage and the Initial Period Muyu Moqo Style {Grossman 1972, 1985). The Muyu Moqo A and subsequent Muyu Moqo B style are based on simple neckless olla and convex bowl shapes typical of the Central Andes. Only notching and incising on occasional rims {and one possible Phase B sherd from a carination angle) are techniques reminiscent of the Manachaqui Style (1972: Figs. 37, 41, 45, 46, 52, 145). During Muyu Moqo Phase C-D, which Grossman (1985:59) dates to the second half of the second millennium B.C., a series of new formal and decorative features appears. Notched carination angles characterize Muyu Moqo C-D, and notched "ridges" of clay 340 decorate the upper shoulders of neckless ollas approximately two ern below the rim (Grossman 1972: Figs. 79-81, 89-96). Muyu Moqo C-D necked jars with thickened and beveled rims (Ibid.: Figs. 131-135) are termed "New, Non-Traditional vessel forms." Their consistent association with novel decorative features including incised and punctate applique fillets and buttons (Ibid.: Figs. 126, 127, 136, 171, 173, 174) leads Grossman to regard the decorated jars as "the result of ideas being introduced from outside the area rather than the result of local experimentation" (Ibid.:120). The attributes that constitute Grossman's evidence for an intrusive complex are among the diagnostic features of the Manachaqui Style. The necked jar appears again as the predominant vessel shape at Initial Period Marcavalle (Mohr-Chavez 1980, 198la, 198lb) at 3314 rn elevation in Cuzco's upper Urubamba River drainage. Mohr-Chavez documents Marcavalle's changing interregional relationships throughout its four or five century occupation between 1100 and 600 b.c. Ceramic evidence suggesting relationships between the Marcavalle Phase A and B assemblages and the Manachaqui style includes the aforementioned necked jars and necked ollas, some of which bear folded-over, flanged and beveled rims (MohrChavez 198la:128), and carinated bowls. Marcavalle necked jars (Ibid.: Figs. 27b, c, r) and carinated bowls (Ibid.: Figs. 331, rn) both show the ridged shoulder or two-tiered 341 profile as well. One Marcavalle Phase A jar with "IV.A. stepped element" rendered in red paint and zoned by incision (Ibid.: Fig. 32k; 198lb: Fig. 36) bears a remarkably close resemblance to Manachaqui Phase Incised Vessel No. 2. Mohr-Chavez (1981b:327-328) likens the abundant incised and punctate applique flanges and lugs (1981a: Fig. 35) to similar features observed at Kotosh and Tutishcainyo, and many also have Manachaqui counterparts. "fillets" occur rarely at Marcavalle. Otherwise, applique The use of red specular hematite paint appears as early as Phase A. Mohr- Chavez (1981b:331) notes that design features including carinated vessel shapes, applique decoration and doublespout-and-bridge bottles with short spouts (Ibid.: Fig. 33a) link Marcavalle to Muyu Moqo C-D and the south coastal sites of Hacha and Erizo. Recent work on the eastern slopes far south of Marcavalle in Bolivian Cochabamba Department (Brockington et al. 1995) has yielded short-necked globular jars with everted, thickened and occasionally notched rims (Ibid.: Figs 7a, 7b [top], 18a, 18b) contemporaneous with the Manachaqui Phase and comparable to Shape B. Paste A Relationships: the Amazonian Lowlands In the lowlands directly east of Manachaqui Cave, the limited archaeological investigation performed within the lower Hual1abamba and Abiseo, and central Huallaga valleys has failed to locate evidence for such early human 342 occupation. One fragment of a "nicked shoulder flange" (DeBoer 1984: Fig.6h) found at Site HUA-3 and grouped with relatively late surface finds recalls the Manachaqui Phase notched applique ribs, but little can be inferred from a single sherd. With the Tutishcainyo styles in the Ucayali basin east of Huanuco, Manachaqui Cave shares the carinated vessel profile and concave upper walls that Lathrap considers diagnostic of Tropical Forest Culture (Lathrap 1970:110112). Late Tutishcainyo Phase pottery in particular bears a number of specific attributes in common with Manachaqui such as rim reinforcement, notching (referred to as "nicking") of carination angles, and the attachment of similarly incised flanges on basal angles (Lathrap 1962: Figs. 41i; 42d, e; 43a, c) . Less common at Tutishcainyo are notched applique bands (Ibid.: Figs. 44c, e) and notched rims (Ibid.: Figs. 47p, 48h). Morales (1992:156) considers the notched applique technique indigenous to Peruvian Amazonia, but it appears in coastal Ecuador's Valdivia Phase VI {Hill 197274:16) prior to its occurrence in Late Tutishcainyo. The more unrestricted Manachaqui Shape B carinated jars bear a general resernbance to common Early (Lathrap 1962: Fig. 20 and Late Tutishcainyo (Ibid.: Fig. 40) vessel shapes. Overall, the Tutishcainyo styles' emphasis on open vessel shapes and complex incised decoration contrasts with Manachaqui's repertoire of restricted shapes and modelled 343 decorations. Other early Amazonian assemblages from the base of the eastern Andes exhibit some formal attributes characteristic of the Manachaqui style. Near Tingo Maria in the Upper Huallaga premontane forest, Lathrap and Roys (1963) have documented pottery from the Cave of the Owls (660 m) that might be considered coeval with and stylistically intermediate between Late Tutishcainyo and Kotosh Wairajirca. Only surface decorations link this assemblage to highland Wairajirca, while the most common shape, a vessel with globular or carinated body profile, concave upper wall or neck, and thickened rim resembles rare Late Tutishcainyo shapes (Ibid.:15; Lathrap 1970:103). The Cave of the Owls profiles illustrated in Lathrap and Roys' Fig. 5a-k can also be likened to those of Manachaqui Phase Shape B Rims 4 and 10 jars, and Shape D unrestricted bowls with concave walls. The Cobichaniqui and Pangotsi Phases in the Upper Pachitea Basin (elevation 300 m) span the centuries from 1600 to 800 b.c. (Allen 1968:347, 351), and their assemblages include carinated bowls and globular neckless jars and bowls. A Pangotsi Phase Form 3 semi-carinated "bowl" with a weak neck and thickened, everted rim (Allen 1968: Fig. 7; Lathrap 1970: Fig. 14e) is comparable to Manachaqui Shape D bowls, while Form 6 is a neckless jar (Allen 1968: Fig. 8). The Pangotsi assemblage features 344 numerous restricted vessel shapes and a tendency to thicken vessel rims. Allen describes the style as an Andean- Amazonian composit (1968:350). Below the northeastern fringe of the Central Andes, archaeological sites of lowland Bagua investigated by Shady (1987b) occupy tropical thorn forest between 600 and 800 m. Shady (Ibid.:464) considers Morerilla Phase necked jars with applique decoration contemporary with, and closely akin to, Pandanche A, Early Huacaloma and Montegrande styles. Morerilla's similarity to the Manachaqui Phase assemblage is difficult to access because the former is represented by only 20 sherds (Ibid.: Fig. 2a-e). Like Montegrande, Morerilla is a low-elevation site assemblage displaying features of highland Cajamarca styles. Northeast of Bagua and Manachaqui in the Ecuadorian Oriente, the shapes and proportions (especially the low positioning of basal angles) of Yasuni Phase vessel Forms 7 and 8 (Evans and Meggers 1968: Fig. 8) closely parallel some Manachaqui Shape B and Shape D variants. Evans and Meggers find the single radiocarbon date of 50±90 b.c. for Yasuni acceptable. Lumbreras (1981:11), however, believes that the Yasuni materials are coeval with the oldest South American ceramic assemblages. Yasuni Form 7 and 8 flanged rims resemble those of Manachaqui Shape B Rim 5 and especially Shape D Rim 5, while Yasuni Form 6 rim profiles recall those of Manachaqui Shape B Rim 10. Yasuni rims and body 345 protrusions also show notching embellishment. The step motif is shared between Manachaqui Incised Vessel 2 (Fig. 31e) and Yasuni Incised and Punctate sherds (Ibid.: Figs. 11f, 12b), but most incised-line decorations on Yasuni vessel walls have no equivalents at Manachaqui. Although other Yasuni Phase vessel shapes (all open bowls) have no Manachaqui counterparts, the correspondences constitute compelling evidence for a close relationship between Yasuni and Initial Period Manachaqui. Some early Amazonian assemblages like Chambira (Morales 1992, 1993:636-641), Jauari (Hilbert 1968) and Taperinha (Roosevelt et al. 1991; Roosevelt 1995) show infrequent and scattered resemblances to the Manachaqui Phase assemblage. The co-occurrence of globular bowls and neckless jars at these sites (and rim notching at Taperinha) may be historically significant or fortuitous. Like Manachaqui's neckless jars, the Amazonian vessels (including Upper Pachitea Pangotsi Phase Forms 5 and 6) are shallower and less restricted than Central Andean neckless ollas, with more nearly vertical walls. The carinated vessel form constitutes a more useful indicator of stylistic relationships. Vessels with embellished basal angles are usually viewed as diagnostic of Amazonian ceramic traditions (Lanning 1967:85-88; Lathrap 1963, 1970, 1971, 1974; Burger 1985b:528). Vessels with prominent and highly embellished carination angles are 346 widely distributed across the northern continental lowlands of Ecuador, Colombia and Venezuela by approximately 1000 b.c., and they persist throughout widely separated Amazonian occupational sequences from the upper, middle and lower Amazon (Evans and Meggers 1968: Fig. 79; Hilbert 1968). Nevertheless, and despite Lathrap's judgement in favor of Amazonian temporal priority, i t remains uncertain whether or not Amazonian carinated vessels like those of Early Tutishcainyo predate those from late Valdivia Phase VI through VIII contexts of coastal Ecuador. The earliest Amazonian pottery from downriver at Taparinha (Roosevelt 1995) apparently lacks carinated profiles, but large spatial and temporal lacunas separate Taparinha and the Central Ucayali. Paste A Relationships: the Northern Andes Ceramic remains from sites in extreme northwestern Peru and southern Ecuador are closely allied with the ValdiviaMachalilla necked jar tradition of the Northern Andes defined by Lanning (1967:85). In Piura, necked jars, carinated and semi-carinated body profiles, rim notching, and notched applique ribs on carination angles occur in an "extremely small" Paita Phase A sample, a "poorly documented" Paita B sample (Lanning 1963:156-157, Pl. 1a, g), and a supplemented Paita C and D sample (Ravines 198687: Lams.1 and 2). Radiocarbon dates of 1660 and 1440 b.c. 347 associated with Paita ceramics have been reported by Richardson (1973:203), but Lanning's early phase subdivisions require further corroboration. Lanning (1963: Table 22) cross-dates Paita A and B with terminal Valdivia phases, Paita C with Machalilla and early Chorrera, and Paita D with late Chorrera. In the semi-arid southern Ecuadorian interior between 1200 and 1350 m elevation, some Catamayo Phase B (1300 to 900 b.c.) globular, necked jars from Loja carry beveled rims (Guffroy et al. 1987: Fig. 12a-c) virtually identical to Manachaqui Shape BRims 1 through 7. Some rims are lightly thickened, but apparently not by the fold-over technique. Bowls are absent from Guffroy's Phase B sample. Aside from similar rim burnishing techniques, the austere Catamayo Phase B pottery shows few additional resemblances to the Manachaqui assemblage. The Valdivia, Machalilla and Cerro Narrio assemblages from the Ecuadorian coastal lowlands and southern highlands best exemplify Northern Andean styles that feature necked jars and carinated vessel shapes. The Valdivia and Machalilla styles in combination (Meggers et al. 1965) share the greatest number of parallels in both shape categories and decorative techniques with the Manachaqui collection. Despite initial arguments for a Machalilla site unit intrusion and partial contemporaneity with Valdivia (Ibid.:171; Lathrap 1963), recent investigations have 348 concluded that Machalilla developed gradually in situ from Valdivia antecedents (Lathrap 1971:85; Hill 1972-74:24-26). Hill's (1972-74:21-24) and Lippi's (1983:351-355) analyses of available radiocarbon evidence concur with Lanning's (1967:85) estimate of 1500 to 800 b.c. for Machalilla's temporal placement. Thus, Machalilla is coeval with the Manachaqui Phase. Manachaqui vessel shapes with counterparts in both Valdivia and Machalilla assemblages (Meggers et al. 1965: Figs. 54 and 89, Tables A and F) include Shape A Rim 2 globular neckless jars (cf. Machalilla Form 15), Shape B necked jars (cf. Machalilla Forms 11, 12 and 13; Valdivia Form 19) and Shape D unrestricted bowls with concave walls (cf. Machalilla Form 9 and Valdivia Form 10). Manachaqui Shape A Rim 1 jars with incipient necks closely resemble Valdivia Form 18 (Ibid.: Fig. 54 and Table A), and Manachaqui Shape C Rim 1 restricted bowls parallel Valdivia Forms 4 and 5. Also, most of the Machalilla and Valdivia vessel shapes are restricted like Manachaqui's. Machalilla necked jars occur in both globular (Meggers et al. 1965: Fig. 83-5), two-tiered (Ibid.: Fig. 85-12) and semi-carinated variants (Ibid.: Fig. 86-6). However, the proportions of the semi-carinated Machalilla variants differ from Manachaqui's which have lower basal angles more comparable to Amazonian vessel proportions (e.g. Yasuni Phase Shape 8 in Evans and Meggers 1968: Fig. 8). In 349 contrast with heavily thickened and folded Manachaqui vessel rims, Machalilla's rims are unthickened or only lightly thickened. A fold-over thickening technique frequently observed on Valdivia jars (e.g. Meggers et al.: Fig. 35-1) was abandoned prior to the onset of the Machalilla Phase. Finally, like Manachaqui neckless jars, the Machalilla variety more closely resembles shallower and less restricted Amazonian neckless jars than Central Andean neckless ollas. Along with remarkably precise vessel shape parallels, the Manachaqui Phase assemblage shares some decorative techniques with the Machalilla assemblage. Recent investigations at La Ponga have documented rim embellishment by notching (Lippi 1983: Fig. 87: Nos. 875, 1018), incision (Ibid.: buttons. Fig. 87: No. 1612) and the addition of applique The co-occurence of rim incision at La Ponga and Manachaqui is especially remarkable because this particular technique rarely appears elsewhere. The wide-mouthed variant of the necked jar (Machalilla Form 13) exhibits a notched (or "nicked") carination angle {Meggers et al.: Fig. 78-7) . Despite such similarities in modeling techniques, line- incision and painting become the preferred decorative techniques by Machalilla times. The rim notching observed during early Valdivia phases (Ibid.: Plate 95) may be rare in Machalilla. In fact, Manachaqui's most common decorative techniques are more characteristic of Valdivia. Applique 350 bands and "fillets" occur frequently in late Valdivia phases (Ibid.: Plates 27-29). However, neither Valdivia nor Machalilla potters embellished vessel profiles by the Manachaqui technique of affixing applique medial and shoulder ribs. In the southern Ecuadorian highlands, pottery from Cerro Narrio and other sites includes necked jars and notched applique band decorations (Collier and Murra 1943: Plate 18, Figs. 3, 5, 7-12; Uhle 1922a: Fig. 12; Bennett 1946). The full range of vessel shape details is not clearly conveyed by the available illustrations, but some Red-on-Buff Type jars feature carination angles decorated with notched applique ribs (Collier and Murra 1943: Plate 18, Figs. 1, 2, 4 and 6). Provenience information at Cerro Narrio suggests their association with the early portion of the sequence. Uhle (1922a: Fig. 22 middle) illustrates a notched rim comparable to Manachaqui Phase Shape B Rim 9, and also notes the presence of applique serpent motifs (Ibid.: 211, Fig. 58A), and a "non-functional" handle (Bennett 1946: 55 and Fig. 8L; Uhle 1922a: 210, Fig. 6a top). Manachaqui, Early Cerro Narrio and Huancarcuchu vessels bear zoomorphic adornos (e.g. Collier and Murra 1943: Plate 24, Figs. 1-6; Bennett 1946:55). The most obvious difference between the assemblages is that, as on the ~eighboring coast, south highlanders emphasized painted and line-incised decoration. 351 A sober analysis of the data fails to support Braun's (1982) long Cerro Narrio sequence (Bruhns 1989:57). Early Cerro Narrio and the early phase at nearby Pirincay (Bruhns et al. 1990) are probably coeval with the Manachaqui Phase. To date, only a few illustrations of early Pirincay pottery have been published. The more securely dated early Cotocollao assemblage from the northern Ecuadorian highlands also shows considerable similarity to Manachaqui Phase ceramics. Globular neckless jars (Villalba 1988: Figs. 106, 123) and necked jars with carinated profiles (Ibid.: Fig. 109; Porras 1982: Lam. 20, Form 14) resemble Manachaqui Shape A and Shape B vessels respectively. Short-necked jars illustrated by Porras (1982: Lam 20, Form 17) show thickened and beveled rims much like Manachaqui Shape B's. Porras' (Ibid.: Lam. 18) vessel Forms 5, 7A and 7B match Manachaqui Shape D. Shared decorative attributes include notched rims and carination angles (Villalba 1988: Figs. 102, 107; Porras 1982: Lam. 3, Nos. 7-9) ~~d notched applique (Villalba 1988: Fig. 94, 99, 100; Porras 1982: Lam. 1). Rim and carination embellishment at Cotocollao was executed by incision directly into the vessel's surface rather than into applique fillets. With the early Ecuadorian highland styles, then, Manachaqui shares numerous attributes among which zoomorphic adornos, notched rims and notched carination angles are salient. 352 In summary, we may conclude that Manachaqui Phase Paste A ceramics again share most design attributes with northCentral Andean Cajamarca styles. Together, these assemblages merge the basic Central Andean ceramic industry with Northern Andean expanded repertoires of vessel shapes and decorative techniques. Perhaps the northern aspect is not surprising since Formative Period Ecuadorian styles have been favorite workhorses for diffusionist explanation in mid-twentieth century New World Prehistory (e.g. Meggers et al. 1965; Ford 1969; Lathrap et al. 1977; Paulsen 1977; Porter Weaver 1981:493-501). The Manachaqui style in particular combines general shape attributes of Machalilla, Cotocollao and Yasuni with decorative attributes typical of Machalilla, Valdivia and Early Cerro Narrio. affinities are particularly strong. The Machalilla Low carinated vessel profiles typical of Amazonian vessels distinguish the Manachaqui Phase Paste A assemblage from the Cajamarca styles. Rare at Manachaqui are the elaborate incised and painted decorations of Amazonia and the Formative Period Northern Andes, although some negative evidence may be a product of the rockshelter's special function. Yet despite its far-ranging and far-flung affinities, the Manachaqui style is unified, coherent and unique. Paste B Origins Manachaqui Phase Paste B pottery is most productively compared to Amazonian assemblages to the east. Manachaqui 353 Paste B1 Vessel 1 (Fig. 63a) and Vessel 3 shapes (Fig. 63c) resemble Pangotsi Phase Form 3 (Allen 1968: Fig. 7; Lathrap 1970: Fig. 14e) and Cobichaniqui Phase Form 1 vessels (Allen 1968: Fig. 5; Lathrap 1970: Fig. 13b) from the Upper Pachitea Basin at the foot of the Central Andes. The post- firing paint on Vessel 3 likewise suggests an eastern origin. Both Allen and Lathrap view Upper Pachitea populations as Amazonian representatives of early Tropical Forest Culture. However, because Paste B's non-plastic inclusions are primarily igneous rocks and minerals much like Paste A's, it is not unreasonable to suggest that the Paste B pottery was produced in the eastern Andean foothills, perhaps in the Premontane Forest Life Zone bordering the Central and Upper Huallaga valley. The medial applique rib on the Paste B3 vessel suggests a close relationship to the Paste A style and origins relatively close by. Most extraordinary is the resemblance between the Manachaqui Paste B1 Vessel 5 bottle spout (Fig. 64a) and Lathrap's Late Tutishcainyo spout fragment (1962: Fig. 49e) and hypothetical double-spout-and-bridge bottle reconstruction (1970: Fig. 9j). Equally striking is the similarity between the Manachaqui Paste B2 vessel shape (Fig. 64b) and Yasuni Phase Form 8 (Evans and Meggers 1968: Fig. 8) . The Paste B2 vessel's "nicked" embellishment of the carination angle is uncharacteristic of the Manachaqui 354 Paste A style, but typical of Late Tutishcainyo pottery as well as Ecuadorian Machalilla Phase ceramics. These two examples provide the clearest evidence for interaction with Amazonian societies. The Manachaqui Phase Paste B1 double- spout-and-bridge bottle provides support to postulate Paste B origins in the premontane foothills and lowlands of the Amazon Basin east, or perhaps northeast, of Manachaqui Cave. CHAPTER 7 THE SUITACOCHA PHASE The Suitacocha Phase coincides with the first centuries of Rowe's Early Horizon (800 B.C. - A.D. 1). Analysis of Manachaqui Cave's ceramic macrochronology, primarily the units in Sector B, provided a basis for the assignment of 1,831 diagnostic sherds, as well as associated lithic, ethnobotanical and faunal remains, to the Suitacocha Phase. Examples of Suitacocha Phase-bearing levels from the berm area include Unit 5, Levels 22 through 25; Unit 22, Levels 8 through 12; Unit 28, Levels 9 through 13; Unit 30, Levels 10 through 13; and Unit 32, Levels 9 through 12. The Suitacocha Phase deposit in Sector A appears thin in comparison to the Manachaqui Phase deposit below, and especially the Empedrada Phase deposit above. As discussed in Chapter 5, this may be attributed to alterations of the shelter interior space during later occupations. Suitacocha Phase Ceramics As noted in previous chapters, the Suitacocha Phase artifact sample is smaller and more weathered than its underlying Manachaqui Phase counterpart. Analyses of the stratigraphy, macrochronology and associated radiocarbon 355 356 dates suggest that cultural deposition was continuous during a seemingly rapid stylistic transition from the Manachaqui Phase to the Suitacocha Phase. Suitacocha Phase vessels, like their Manachaqui Phase predecessors are predominantly restricted. Vessel shape categories A through E remain basically the same, except that Shape D (unrestricted bowls with concave walls} drops out. The popular Manachaqui Phase carinated and semi-carinated jar shapes are replaced by globular shapes, and only a small percentage of Suitacocha Phase bowls have carinated profiles. Again, virtually all of the vessels had rounded bases. Shape A neckless jars dwindle in importance during the Suitacocha Phase, while Shape B short-necked jars continue to dominate the vessel shape inventory. Like their Manachaqui Phase antecedents, the Shape B jar rims were often heavily thickened or reinforced by folding the ends out and down. A groove left by the process of smoothing the folded end to the exterior surface is a diagnostic attribute during the Suitacocha Phase. Shape c restricted bowls and Shape E open bowls likewise exhibit reinforced rims. especially open bowls, are still rare. Bowls, Shape F is a new jar shape distinguished from jar Shapes A and B by long, direct and usually vertical rim forms. The Shape F rims might have been lumped into the Shape B category, but the lack of morphological overlap between the two categories suggests functional differentiation. The Shape X category was 357 conceived to contain rim forms that cannot yet be confidently assigned to any of the established shape categories. Suitacocha Phase decorations include a combination of Manachaqui Phase holdovers and new techniques. Present, but diminished in importance, is notched applique and appendages like flanges, buttons and adornos. Design emphasis shifts from applique modelling to geometric renderings utilizing incision, punctation and a few texturing techniques either individually or together. During the Suitacocha Phase, potters frequently embellished vessel rims by burnishing, or painting and burnishing them prior to firing, or by adding incised decorations. The slips or paints vary in color from maroon to blood red to pale iridescent red and pink. Red slip is also used to create contrasts with unpainted fields filled with incised decoration. On Shape B and F jars, the rim and bottom portion of the vessel body was frequently painted and burnished. The unpainted shoulder then served as a field for incised decorations. Weathering of the sherd surfaces becomes a serious impediment to the identification and analysis of decorative techniques, and especially to the quantification of painted sherds. Ceramic Paste Group A While there is clearly a single ware that predominates during the Suitacocha Phase, other similar wares 358 insufficiently distinctive to be consistently differentiated with the unaided eye are also present. Thus, Paste Group A consists of more than one brown ware less red in hue, and darker in color value than Manachaqui Phase Paste A. The Paste Group A sherds utilized to characterize the Suitacocha Phase total 1,792, representing 97.9 percent of the whole phase diagnostic sherd collection of 1,831 sherds. That Suitacocha Paste A wares are not easily distinguished from the predominant Manachaqui Phase ware reflects the minimal change in basic ceramic technology. In general, the Suitacocha Paste Group A contains greater quantities of vein quartz, but lesser quantities of quartz crystal, than Manachaqui Paste A. Microscopic and microprobe analyses confirm that grain sizes and size variability differ little from those of the Manachaqui Paste A. The microprobe identified only quartz and volcanic glass among the angular grains of complex igneous rocks. Surface finishing techniques utilized also differed little from those employed earlier. One sherd shows cord impressions on its interior surface (Fig. 67c). The impressions were likely produced by the technique termed "netting" observed on Urabarriu Phase sherds from Chavin de Huantar and illustrated by Burger (1984b: Figs. 143, 144). Apparently comprised of twined fibers, the 3-4 rnrn thick cords probably wrapped cloth pads that were pressed against vessel interiors during the burnishing process. Vessel walls are 359 slightly thicker and more variable in thickness than those of the preceding phase, ranging from 3 to 6 mrn. rims tend to be thinner and less massive. However, A slip with red pigment was often applied to vessels, or only to their rims before burnishing. Well-controlled firing left nearly all of the pottery completely oxidized. Morphology A total of 979 rim sherds, 19 sherds from vessel basal portions and nine miscellaneous body parts supplied information regarding Suitacocha Phase vessel shapes. The Suitacocha Phase sample is just as fragmented, and much more eroded than the Manachaqui Phase sample. Yet despite a smaller sample, greater consistency in technological treatment of more standardized vessel shapes permits clearer analytical distinctions between so-called jar and bowl shapes. In other words, the vessels classified as bowls consistently exhibit the smoothed interiors that jars invariably lack. This shift suggests that vessel shape classes were now more functionally specialized. Thus, the shape characterizations offered in the following pages contain fewer caveats. The most troublesome feature of the collection is the similarity shared between some jar and bowl rim forms that necessitated the creation of a "sink trap" category, Shape X, described below. Rim forms again constitute the focus of the following classification. The same three-tier scheme developed for 360 Manachaqui Phase rims is extended to accommodate Suitacocha Phase rim forms and interpreted body shapes (Appendix E) . Suitacocha Phase rims can almost always be distinguished from Manachaqui Phase rims based solely on morphology. Differences in paste qualities, the presence or absence of red paint, and the degree of erosion provide additional evidence to confirm or question initial determinations. Shape A: Neckless Jars A total of 30 sherds grouped into four rim forms belong to neckless jars (Table 11}. Together, these comprise only three percent of the Suitacocha Phase rim collection. The corresponding vessels were apparently small in size, and all but one example exhibits either exterior thickening, or an incised groove below the lip exterior. Rim 14 (Fig. 68a-c). Rim 14 (14 examples) thickens gradually to reach maximum thickness (between 6 and 8 rnrn) at a flattened and squared lip. All but one slightly upturned rim (Fig. 68c} are direct or lightly incurving. An incised groove encircles the rim exterior one centimeter below the lip. Vessel walls approximate 4 mrn in thickness. These rims exhibit red paint on the lip and exterior surface. Rim 15 (Fig. 68d-f). Rim 15 (8 examples) belongs to incurving rims that have been thickened on their exterior surfaces to a maximum of 6 to 9 mrn, and grooved below the 361 thickened portion. Their lips are rounded, and vessel walls range from 4 to 6 mrn thick. R~ 16 (Fig. 68g, h). The seven examples of Rim 16 are similar to Rim 15, but they lack the groove below the thickened exterior. Vessel walls show a thickness of 5 to 6 mrn. R~· 17 (Fig. 68i). This single sherd has a thickened (7 rnrn) interior surface. point. The lip exterior ends in a dull It belongs to the largest of the neckless jars, and its walls are 4 mm thick. Shape B: Short-necked Jars The 612 short-necked jar rims constitute 62.5 percent of the Suitacocha Phase rim collection (Table 12). Like their Manachaqui Phase antecedents, these jar rims are typically thickened on their exterior surfaces, and lips are often beveled or flattened. They range from 7 to 18 ern in diameter, averaging between 11.5 and 12 em. Rim orientations vary from sharply everted with an angular neck, to nearly vertical with a negligible neck. Lip exteriors were usually reinforced by folding the end of the rims over. Reinforced rims may feature a thin (1 rnm) groove directly beneath the thickened portion. Also, some Suitacocha Phase Shape B rims show a light outward inflection where the neck meets the shoulder 1 to 1.5 em below the everted rim. 362 Rims lla-g (Figs. 68j-s, 69, 70). The 585 examples of Rim 11 are divided into seven variants. Each of the variants represents only a morphological tendency. they constitute a continuous series. Together The reader will observe that the average rim diameter for each variant is nearly identical, at least among the larger samples. Corresponding vessel walls average 4 or 5 mm in thickness. Rim 11a (Figs. 68j-s, 69a-c) exhibits a thickened exterior (maximum 7 mm) and a flattened lip. Grooves below the reinforced lip exteriors lend these rims a "squared" profile. Three examples (e.g. Fig. 68s) show thickened protrusions, or lip flanges, to securely grasp the vessel by the rim. Rim 11b (Fig. 69d-k) examples are everted, with rounded lips and exteriors thickened to a maximum of 6 mm. Rim 11c (Fig. 691-r) differs little from 11b, but the thickened exterior appears more bulbous reaching a maximum of 7mm, and the lip ends at a dull point. Rim 11d (Fig. 70a-h) is vertically oriented, but otherwise it resembles Rim 11a because of its flattened lip and the groove frequently found below its thickened exterior (maximum thickness 4 to 7 mm) . The sherd illustrated in Fig. 70h bears iridescent pink paint. Rim 11e (Fig. 70i-l) is only slightly everted and nearly vertical and direct. The lip has been flattened and slopes toward the vessel exterior. 6 to 7 mm. Its maximum thickness is One sherd (Fig. 70i) exhibits pink iridescent 363 paint. Rim 11f (Fig. 70m-s) reaches a maximum thickness near 7 mm at the lip which has a dull point at its interior edge. The rim's thickened exterior surface forms a right angle with the neck, which also tends to be inflected again where it merges with the vessel shoulder (e.g. Fig. 71g, s). Rim 11g is essentially the same as Rim 11e, except that its flattened lip slopes toward the vessel interior. Rims 11e and 11g might show the same shoulder inflection as Rim 11f, but the fragments are too small to verify the possibility. One Rim 11g rim (Fig. 70v) shows traces of pale red iridescent paint. Rim 12 (Fig. 71a-f). These 16 examples are unthickened everted rims with rounded or semi-rounded lips. Curvature at the neck is smooth and nearly uninterrupted. The rim, neck and vessel walls are only around 3 rnm thick. In some instances, these rims might be mistaken for Manachaqui Phase Shape B Rim 3b, and their pastes best serve to distinguish them. Rim 13 (Fig. 71g-l). The 11 sherds representing Rim 13 are lightly thickened and slightly everted. The maximum thickness at the rounded lips reaches 7 or 8 rnm. Shape C: Restricted Bowls The 54 rim sherds from restricted bowls represent 5.5 364 percent of the Suitacocha Phase rim collection (Table 13). There are four variants of which all but one have thickened lip exteriors. Rim diameters range from 9 to 20 em. Despite weathering, some sherds show traces of burnishing on exterior surfaces. R~ 7 (Figs. 71m-72a). There are 37 examples of Rim 7. The top of the rim is thickened (7 or 8 mm maximum) and lightly grooved below the flattened, reinforced lip to render a "squared" profile reminiscent of Shape B Rim lla. Maximum thickness reaches between 7 and 9 mm at the lip, but vessel walls are thin, between 2 and 3 mm. The rim illustrated in Fig. 71u exhibits a pale red iridescent paint, while the bowl depicted in Fig. 72a shows only traces of red paint on the exterior surface. R~ 8 (Fig. 72b-d). Nine Rim 8 sherds pertain to three bowls that may be considered morphologically related to the Rim 7 bowls. Rim thickening is less pronounced, and all are decorated with a band of zoned punctation. approximately 3.5 mm thick. Vessel walls are Two of the three are entirely painted red (Fig. 72b and c), and one shows a lug handle on the rim (Fig. 72b) analogous to the lugs observed on Shape B Rim lla jar rims (see Fig. 68s). The bowl's rim lug has been flattened like the rest of the rim. Twenty-four sherds belonging to this particular bowl were recovered from Feature R-3 and surrounding unit levels. The lip of the Rim 365 8 example viewed in Fig. 72c is rounded rather than flattened. R~ 9 {Fig. 73a-c}. Seven rim sherds from round bowls with rounded, lightly thickened lips represent Rim 9. Maximum thickness at the lip varies from 7 to 9 mrn. Vessel walls are 5 rnrn thick. R~ 10. 10 {Fig. 73d}. A single rimsherd constitutes Rim The rim is gradually, and only slightly thickened from 2 to 3 mrn. The lip has been flattened and a tiny clay "cornice" protrudes from the lip's interior edge. Shape E: Unrestricted Convex Bowls Only 15 sherds belong to the Shape E category, which constitutes 1.5 percent of the Suitacocha Phase rim collection {Table 14}. There are four rim forms, which are represented by a single sherd. two of Like other Suitacocha Phase vessel rims, Shape E bowl rims typically show some sort of thickening treatment. R~ members. 8 {Fig. 73e-h}. The Rim 8 category has seven The rims vary in overall thickness, but all share a thickened and rounded lip that protrudes on the vessel exterior. Maximum thickness at the lip ranges from 3 to 6 mm. R~ 9 {Fig. 73i}. Six examples of Rim 9 are assigned 366 to the Suitacocha Phase. As with the Manachaqui Phase examples, they may have intruded from later deposits above. Rim 9 represents a simple, convex bowl with a rounded lip. Rim 10 (Fig. 73j). The single example of Rim 10 belongs to a large bowl with a thickened rim that has been rendered by folding 2 em of the rim end down over the vessel exterior. The maximum thickness attained is 10 rnm. Rim 11 (Fig. 73k, 1). Rim 11 is one rim sherd that shows gradual light thickening and a rounded lip. thickness reaches 5 mm. Maximum A sherd with matching paste, color and decorative technique probably corresponds to the base of the same vessel. Rim 11 is unique in representing the only Paste Group A vessel with a flat base. Shape F: Jars with Vertical Rims A total of 272 sherds represent nearly 28 percent of the Suitacocha Phase rim collection (Table 15). If combined with the 612 examples of Suitacocha Phase Shape B shortnecked jars, the total number of necked jars increases to 884, or 91 percent of the phase collection. The Shape F rims range from 4 mm to 6 em in height, averaging approximately 2.5 em. The majority are between 2 and 3 em high and oriented vertically, although inverted and everted varieties exist. A diagnostic feature of Shape F rims is an angular junction where the base of the rim/neck meets the 367 vessel shoulder. angle. Rim fragments nearly always broke at this Rim diameters range from 5 to 21 em, and average between 11 and 14 em. Vessel bodies were globular, often spherical, and walls varied from 3 to 4 mm in thickness. Virtually all of the Shape F rims were burnished or slipped red and burnished. Rim la, b (Figs. 74, 75a-v). Rim 1 sherds total 138. They may be gradually thickened toward the rim end, or unthickened. Lips of variant Rim 1a (Fig. 75a-n) are rounded, or end in a dull point. slightly everted. They may be vertical or Rim 1b (Fig. 75o-v) may exhibit an inverted, vertical or everted orientation. The lips are squared, or "semi-squared" with rounded edges. Together, the Rim 1 variants range from 1.5 to 3.5 em in height, and 4 to 6 mm in maximum thickness. Some Rim 1 sherds feature a narrow, incised groove at the union with the vessel body. Rim 2 (Fig. 76a-j). Rim 2 is represented by 79 sherds. The lip exterior has been thickened at the top to a maximum of 7 mm. The rims range from 2 to 4 em high and are often slightly inverted. incision. Occasionally these are decorated by The sherd illustrated in Fig. 74b bears iridescent paint. The incised sherds illustrated in Fig. 76j and k may belong to the same vessel. Rim 3 (Figs. 761-n, 77a-f, 78a). There are 26 examples 368 of Rim 3. These resemble Rim 2, except that the tops of the rims are bent outward and their undersides are thickened. The lips may be rounded or squared. The average height of Rim 3 cannot be determined because of breakage. Most of the Rim 3 sherds show embellishment by painting and/or incised decoration, and the average rim diameter is greater than those of other Shape F rim variants. Rims 4a, 4b (Fig. 78b-f). divided into two variants. Ten examples of Rim 4 are These rims are invariably short and vertically oriented, ranging from 4 to 7 mm high. Rim 4a (Fig. 78b-d) has been thickened to 6 mm and has been grooved at the rim's base. The lip has been squared. example illustrated in Fig. 78b bears iridescent paint. The Rim 4b's (Fig. 78e, f) neck is negligible, and its lip has been slightly thickened before rounding. Rim 5 (Fig. 78g, h). Six examples of Rim 5 have been thickened by folding 1 to 2 em of the rim down over the rim exterior. The vertically-oriented rim ranges from 2 to 3 em high, and the lip has been rounded. The folded over portion may extend over halfway down the height of the rim, and has not been smoothed to the rim exterior wall. Rim 6 (Fig. 78i, j). The six sherds representing Rim 6 are bent at an angle approximately half way below the lip. They are 2 em high, unthickened and the lips have been 369 squared. Rim 7 (Fig. 78k). comprise Rim 7. Four rim sherds from the same vessel It is the largest of the Shape F rims, and represents the largest Suitacocha Phase Shape F vessel. It is 6 em high, slightly convex and thickened to 1 em at its rounded lip. The mouth diameter is 21 em. Rim 8 (Fig. 79a-c). The three sherds representing Rim 8 are all vertically-oriented and slightly everted. ends have been thickened to 8 mm and squared. The Rims illustrated in Fig. 79b and c have shallow grooves beneath the thickened portions of the lip exteriors. All three are painted red. Shape X: Undetermined Vessel Shapes The rim forms grouped under the Shape X category cannot be confidently assigned to any of the previously described vessel shapes, although informed guesses may be offered in some cases. They are a diverse lot, and are described below in order of decreasing abundance. Rim 1 (Fig. 79d-j). Rim 1 belongs either to a restricted bowl similar to Shape C Rim 7, or to jars with insloping rims like Shape B Rims 11g or 1le, or Shape F Rim 2. The Rim 1 problem typifies the dilemma that frustrates attempts to determine the vessel shapes for all of the rims grouped under Shape X. Shape X rims often feature carefully 370 burnished and painted exterior and interior surfaces. Lacking evidence for a neck base or shoulder beneath the rim, it is virtually impossible to confidently group the rims in question with either bowls or jars. The 25 examples of Rim 1 are all thickened on the lip exterior to a maximum of 9 mm. The lip interior edge ends in a dull point. Thirteen rim sherds measured range from 9 to 14 em in diameter, averaging 11.54 em (S.D.=1.28). is 12 em. The mode diameter These dimensions are more typical of the smaller Shape B and F orifices than those of Shape C. R~ 2 (Fig. 79k-n). The eight rims are similar to Rim 1, but the thickened portion of the lip is smaller, attaining a maximum thickness of 7 mm. range from 11 to 20 em in diameter. Six sherds measured Three of them, probably from bowls, are decorated by incision. R~ 3 (Fig. 79o-r). These seven rim sherds all thicken gradually to a maximum 9 mm at the flattened lip which slopes down toward the interior edge. Five sherds measured range from 11 to 16 em in diameter. R~ 4 (Fig. 80a). These three rim sherds belong to one large Shape C or Shape F vessel. to 8 mm below a rounded lip. R~ 5 (Fig. 80b). The rim gradually thickens The diameter is 20 em. Two Rim 5 sherds also belong to either Shape C or Shape F. They are insloping and thickened 371 to 1 ern at the flattened lip. Rim 6 (Fig. 80c). The rim diameter is 12 ern. The one Rim 6 sherd probably pertains to a Shape E vessel. It is vertical and gradually thickened to a maximum of 8 mm below a flattened lip which slopes slightly toward the vessel exterior. Its diameter is 19 ern. Basal Sherds Of the 19 sherds from the basa: portions of vessels, three sherds might belong to a recessed jar or bowl base with raised painted and polished walls (Fig. 80d). vessel shape is not well understood. The The remaining 16 sherds pertain to at least four carinated bowls with protruding basal angles (Fig. 80e-h). Two of the bowls exhibit red slip on both interior and exterior surfaces. Diameters measured from the interiors of Figs. 80e and 80g are 17 em and 22 em respectively. The basal portions protrude 3 to 7 mm from the upper walls. The angles of orientation illustrated represent best estimations. Miscellaneous Shapes Nine sherds (Fig. 8la-i) pertain to ceramic vessels or objects of unknown shape. Three of these (Fig. 8la-c) are interpreted as pieces of "mammiform" vessel legs. Five (Fig. 8ld-h) seem to belong to handles, and a single sherd with incised decoration and tubular cross-section (Fig. 8li) 372 remains unidentified. Decoration A total of 785 sherds serve to evaluate Suitacocha Phase decorative techniques. The principal techniques may be grouped under the headings: painting, applique and incision. these. All of the shape categories show at least one of Apparently the Shape A neckless jars lack decorative treatment other than red paint and incised grooves on rim exteriors. The Shape F jars, and probably the Shape B jars, show all three of the major techniques. The Shape F Rim 1 jar illustrated in Fig. 74 represents a conventional Suitacocha Phase design layout in which the unpainted upperthird of the vessel body provides a field for applique and incised decorations, while the rim and lower two thirds are red slipped and burnished. Incised lines and applique ribs or flanges may separate the decorated jar shoulder from the painted lower hemisphere. Incised lines also delineate zones filled with red paint or punctation. punctation occur individually on rims. Incision and Additional observations regarding jar decorations will be offered below. The small sample of Suitacocha Phase bowls also exhibits frequent decoration. Shape C restricted bowls may be painted and incised while, of the Shape E open bowls, only the incised Rim 10 is decorated. Examination of body sherds confirms that applique occurs on bowls, but precise 373 shape and rim associations remain uncertain. As with the Manachaqui Phase collection, severe fragmentation limits the capacity to deduce tendencies governing associations between decorations ~~d vessel shapes; however, the overwhelming majority of decorated sherds pertain to jars (89%). It should also be borne in mind that this particular Paste Group A sample may represent more than one style or set of stylistic norms. Conversely, it probably fails to include techniques, motifs and combinations that might be observed within an expanded Suitacocha Phase sample. Painting Pre-firing red paint or slip, and traces of these, were observed on 461 sherds assigned to the Suitacocha Phase. Painted body sherds total 225, while 236 rim sherds (24% of the phase total) exhibit remains of red paint. The proportion of unpainted to painted sherds is, of course, distorted by the severe weathering of the Sector B sample. For example, none of the rim sherds from Units 7, 8 and 39 at the base of the berm show paint. Within the sheltered Sector A, proportions average close to 2 to 1 which suggests a similar proportion for the entire Suitacocha Phase rim collection, and a substantially higher total for painted body sherds as well. Of the 225 painted body sherds, 106 display red paint with no other techniques except polishing, and 16 of these carry red paint on interior surfaces. The remaining 90 sherds show red painted exterior surfaces, and 374 the majority belong to polished and painted jar bases like that of the reconstructed Shape F Rim 1 example (Fig. 74). Red paint is associated with applique and incision on 119 sherds. On many Suitacocha Phase vessels polished red paint serves to contrast with, and therefore highlight, unpainted decorative panels or fields filled with incised and punctate designs. The red paint utilized for vessel decoration varies from dark red (10R3/6) to a pale red or pink (10R5/3). It is actually a slip composed of water, clay and iron oxide which achieves a metallic luster and iridescent effect when the slip is especially dilute (Lathrap et al. 1975:53-55; Sonin 1977). The iridescence is best observed by splashing water onto a painted surface. Burnishing tracks left on rims and jar bases particularly stand out with the addition of water. The slip's use on vessel rims and jar bottom exteriors strongly suggests that the iridescent effect was incidental rather than intentionally produced. Faint remains of white paint appear on the exterior surfaces of two sherds from Sector A's Floor Y, and one sherd to be described bears two thin black-painted lines. Applique A total of 171 sherds with applique decorations, including three flanges, can be assigned to the Suitacocha Phase. Applique ribs and bands appear independently, as well as in combination with the other techniques of painting 375 and incision. Applique may be used to frame fields with painted or incised decoration. Techniques of applique embellishment include notching and punctation. Notched applique. A total of 94 sherds bear applique ribs and bands embellished by notching (Figs. 82a-g}. These can be divided into sherds with 1} single, isolated notched ribs or bands, 2} parallel notched ribs and 3} high relief notched ribs and bands. The Suitacocha Phase notching technique involves cutting or incising with a sharp pointed tool, thereby differing from the Manachaqui Phase technique. Only the high relief notched ribs and bands are still notched by impression, probably with sprigs of cana. Sherds with single notched ribs or bands (Figs. 82a-d} from sound Suitacocha Phase contexts total 52. Notched applique decorations also appear in later chronological components. An additional 15 sherds with notched applique were recovered from mixed Suitacocha Phase and Empedrada Phase levels. Suitacocha Phase sherds isolated and parallel notched ribs and bands are invariably straight. Bands are vertically oriented while framing design panels on jars (e.g. Fig. 74}, while ribs run horizontally around vessel mid-sections. 12 mm. These applique strips vary in width from 6 to On 20 sherds, two and three notched ribs run horizontally and parallel to one another (Fig. 82e}. pertain to one or more bowls with red-slipped interior All 20 376 surfaces. These parallel ribs are 5 to 7 mm wide, and separated from each other by 2 to 4 rnrn. The 22 examples of "high relief notched ribs and bands" may be straight or curving, and all belong to jars (Figs. 82f and g). rnrn high. These strips are typically 4 mm wide and 3 to 5 Notches are spaced 2 to 5 rnrn apart. High relief ribs may separate the unpainted upper from the painted lower hemispheres of jars (Fig. 82f), or bands may undulate sinuously around vessel shoulders (Fig. 82g). Additional examples of applique demonstrating associations with incised decorations are illustrated later within this chapter. Punctate applique. A total of 71 sherds bear applique ribs and bands embellished with either ovoid or round punctation. The ovoid punctation is rare, appearing on only 4 sherds from jar shoulders or mid-sections (Fig. 82h and i), and one sherd from a vessel leg (Fig. 81a). The 66 sherds bearing bands with round punctation are 3 to 4 rnrn wide, and the punctations are usually 2 mm apart (Figs. 82j and 83a). 3 neck. In Fig. 78a, an arching band adorns a Shape F Rim Punctate applique buttons appear only on one sherd (Fig. 82i). The motif depicted in Fig. 82j deviates from the majority of these bands which usually run sinuously around jar shoulders (e.g. Fig. 83a). They are frequently associated with incised designs and red paint as later illustrations show. 377 Unembellished applique. There are only three eroded examples of applique bands with no embellishment (Fig. 83b and c). All three pertain to jars. Flanges. Three sherds represent flanges (Fig. 83d-f). A short, curving, "shelf-like" flange has three small incised nicks on one side (Fig. 83d). A flange fragment from a jar mid-section protrudes with a triangular crosssection (Fig. 83e). It is grooved at both top and bottom junctures with the vessel wall. The upper surface shows short incisions, while the bottom is covered with red paint. The third flange is an incised tab projecting out from the upper shoulder of a jar, and sloping downward (Fig. 83f). Incision and Surface Texturing A total of 525 sherds bearing incision and/or surface texturing are assigned to the Suitacocha Phase. The decorative techniques grouped under this category include: cross-hatching, stamped circles, punctation, combing, rouletting, wet incision or inciso cortante, engraving, rocker-stamping and fabric-impression. Design motifs comprised of simple rectilinear patterns appear on jar rims, necks and shoulders, or on bowl exteriors. As previously noted, zones delineated by incision and/or applique may be filled with red paint, punctation, stamped circles, or incised line motifs. Patterns outlined with incised lines include triangles, arches and inverted rectilinear U-shapes. 378 A total of 174 sherds show straight or curved incised lines appearing alone, in sets or as parts of simple line motifs. Lines are narrow, and were created by cutting into leather-hard clay with a sharp-tipped implement. never more than 1 mm wide or 1 mm deep. They are Incised lines are found on Shape B jar necks (e.g. Figs. 69c, 83g), Shape F jar rims (Figs. 76i-l, 77a, e), Shape C bowl exteriors (Figs. 72a, 73d, 83e) and jar shoulders (Fig. 83h, i). Divergent arrays of parallel lines commonly occur on both bowls (Fig. 72a) and jars (Figs. 77e, 83j), while cross- hatching ornaments at least one jar represented by seven sherds (Fig. 83k and 1). Stamped circles. Stamped circles, concentric circles and circle-dots were observed on 76 fragments of jar rims (Figs. 69a, 77b-d), necks (Fig. 84a) and shoulders (Figs. 84b-l, 85a-f). They appear in association with red, and in one case black, paint (Fig. 84b), with divergent arrays of parallel lines (Fig. 84d-f), within incised zones (Fig. 84i), with zoned punctation (Fig. 84j, k), with applique ribs with round punctation (Fig. 85a-c) and high-relief notched ribs (Fig. 85d, e). Two sherds offer fragmentary glimpses of designs featuring stamped semi-circles intersected by incised lines (Fig. 841), and zoned by unusual applique ribs embellished with crowded rows of small punctations (Fig. 85f). 379 Punctation. A total of 226 sherds show punctation. In the following paragraphs, the punctation shapes differentiated include: round, ovoid, oblique, dash, triangular and crescent. Punctation may be zoned by incision or applique, but also occurs unzoned. Unzoned rows of punctations may run independently or parallel to incised lines. Infrequently they sketch linear patterns. Thirty- seven of these sherds belong to Shape C Rim 2 restricted bowls, while 188 pertain to Shape B and F jars, and one is assigned to Shape X. The thirty-seven Shape C sherds with zoned punctation derive from three bowls. The first features a single row of oblique incisions zoned by incised lines comprising a band one em wide under the rim exterior (Fig. 72b). A second bowl has a similar band filled with two offset rows of ovoid punctations (Fig. 72c). On both bowls, the exterior surfaces above and below the bands have been painted red and polished. The third bowl features a band with irregular rows of crescent-shaped punctations and, except for the band, has been polished inside and out. Shape B and Shape F jars feature punctation on necks and shoulders, and some sherds are sufficiently large to reconstruct incised and punctate motifs that likely repeat around the vessel circumference. For example, Fig. 85g illustrates a Shape F jar with a design field on the shoulder zoned on the top by red paint and on the bottom by 380 incised lines and red paint. Parallel incised lines outline triangles separated by fields with minute dash punctations. The punctations form diagonal rows running parallel to the incised lines. A Shape B jar (Fig. 85h) neck features oblique punctations within pendant triangular fields zoned by an incised line above, and incised lines and red paint below. Another sherd from a jar shoulder (Fig. 86a) depicts four parallel arching incised lines bordered above by a field with large triangular punctations and below by incision and polished red paint. A fourth jar shoulder fragment (Fig. 86b) shows oblique punctations within a zone (perhaps triangular) formed by a single curving incised line. The remaining illustrations of sherds with punctation decorations portray the variety of designs and design elements rendered on jar necks and shoulders. Many of these incorporate rows of punctations zoned by parallel incised lines. Fig. 86c exhibits a row of oblique punctations zoned by parallel incised lines which in turn delimits the bottom of a larger field filled with oblique punctations. Single rows of ovoid and widely spaced punctations may occupy large zones created by incision (Fig. 86d). Incised lines may be marked (unintentionally?) by rows of perpendicular dash punctations (Fig. 86e). Three sherds (Fig. 86f-h) show small, oblique punctations in arching rows zoned by single or parallel incised lines. A similar design element 381 incorporates parallel diagonal and horizontal rows of punctations (Figs. 86i-l, 87a, b). An unusual sherd (Fig. 87c) bears punctations that may trace an undulating pattern around the vessel circumference. A miscellany of sherds demonstrate other decorative uses of punctation. Punctations surrounding jar necks frequently appear independent of incision (Figs. 69b, k; 77f; 78f; 79n; 87d-h). Fig. 87h shows rare round punctations, while another single sherd demonstrates that punctation can be zoned or bordered by an applique band (Fig. 87i). Finally, the most complex motifs are rendered by combinations of incision, zoned punctation and notched applique bands on panels framed by red paint and repeating around jar shoulders (Figs. 87j; 88a, b). Combing and Brushing. Sets of very closely spaced incised lines were created either by combing with an implement with rigid bristles, or by brushing with an implement with flexible bristles. The techniques, which are not easily distinguished from one another, were observed on 14 sherds. One rim assigned to Shape X (Fig. 80a), and perhaps belonging to a large bowl, shows vertical sets of six very fine incised lines that on some sherds can only be observed under certain lighting conditions. A coarse combing technique was applied to a vessel neck depicted in Fig. 78c., and to a jar mid-section shown in Fig. 88d). 382 Fig. 88e illustrates extremely fine lines created by brushing, but utilized to render design elements typically executed with conventional incised lines. Four thin, eroded sherds from a jar show Rouletting. rouletting beneath parallel, curving incised lines (Fig. 88f) . The lines are wide and shallow, and were produced with a blunt-tipped tool. polished. The gray exterior surface is Although classed within Paste Group A, this pottery may have originated in more distant regions. Incised boss. A boss from a jar has a conical shape and exhibits the end portions of two parallel incised lines (Fig. 88g). Traces of red slip and polish remain visible. Fine-line Scratched. A single sherd (Fig. 88h) shows an incised line and light scratches made by apparently haphazard strokes with a fine-pointed implement. Squiggles. Eight sherds exhibit sets of shallow, wavy lines and could all belong to the Shape X Rim 2 vessel illustrated, probably a restricted bowl (Fig. 79k). Some of them (e.g. the sherd at left) may have been rendered by a combing technique. Inciso cortante. Seven tan sherds show inciso cortante, a technique of incising deep lines into wet clay with a sharp pointed implement (Fig. 73j, k). The technique 383 leaves wider (1.5 mm}, deeper lines than is usual for Suitacocha Phase pottery. Small ridges on both sides of each incision are left by the implement as it pushes aside the wet clay. All of these sherds probably pertain to the same flat-bottomed open bowl designated Shape E Rim 10. Engraving. Five sherds exhibit engraving, or incising the vessel after the clay is dry and hard. leaves grooves with rough, chipped edges. The technique All of the sherds seem to come from the same Shape C Rim 7 vessel (Fig. 71t} which also shows red paint below the incisions. Rocker-stamping. Rocker-stamping can barely be discerned on three very eroded sherds which could not be usefully illustrated. Isolated non-dentate rocker-stamping lines zig-zag vertically down the vessel shoulder. They belong to a jar with a red-slipped exterior. Mat and/or fabric impressed. Three sherds show traces of having been decorated by pressing or adhering a fabric or woven mat to their exterior surfaces. Two of the these are unusually thick and show small, chevron-like marks occurring in parallel, vertical rows (Fig. 89a}. A third has faint, white-stained marks, apparently left by a gauze or net-like fabric (Fig. 89b}. Adornos There are four adornos in the Suitacocha Phase 384 collection {Figs. 89c-e, 90a). Three are hollow, while one is solid; all are zoomorphic and anthropomorphic heads that probably adorned jar shoulders or mid-sections. Fig. 89c depicts a bird, most likely a parrot, with incised eyes and mouth, and a distinctive hooked beak. It was painted red. Fig. 89d shows an unidentified animal with inset, applique eyes and prominent snout reminiscent of the opposurn's. small appendage under its cheek may be a paw. A Fig. 89e, a lug-like appendage with deep incisions and a notched applique rib vaguely resembles a bat's head and face. Fig. 90a is an anthropomorphic head which originally projected from the vessel at an angle. Ceramic Paste Group B The Suitacocha Phase ceramic assemblage includes 39 sherds assigned to Paste Group B. These represent 2.13 percent of the total phase sherd collection. The three Paste Group B rim sherds represent only 0.305 percent of the Suitacocha Phase collection of 982 rims. This figure under- represents the quantity of Suitacocha Phase pottery from relatively distant sources. Most likely, some Paste Group A sherds with rare decorative techniques {e.g. rouletting, inciso cortante) originate from sources other than that of the bulk of the assemblage. Paste B4 Thirty-eight sherds assigned to the Paste B4 category 385 are light tan or beige in color (7.5YR6/6 reddish yellow), and contain a sandy temper. Electron microprobe analysis identified grains of quartz, alkali feldspar, chlorite, hornblende and iron oxides. be crumbly. The semi-compact paste tends to A darker core indicates that the pottery was incompletely oxidized during firing. At least two necked jars are represented by rims and decorated sherds. One rim (Fig. 90c) from a short-necked jar was thickened by the fold-over technique and grooved where the folded end meets the neck exterior. Were it not for its distinctive paste and Suitacocha Phase context, this rim would have been classified as Manachaqui Phase Shape B Rim 2a. The other rim is slightly thinned toward the lip and sits atop a longer neck (Fig. 90d). Shape BRim 12. It somewhat resembles Paste Group A The decorated sherds show a thick, notched flange (Fig. 90b), horizontal rows of round punctations zoned by parallel incised lines, and unzoned vertical rows of punctations (Fig. 90e-g). Paste B5 Paste B5 is represented by a single sherd from the basal angle of a flat-bottomed bowl (Fig. 90h). Its Sector A context (Unit 11 Level 8) is mixed, and assignment to the Suitacocha Phase is based on regional comparisons described at the end of this chapter. The paste is very compact and microprobe analysis identified angular grains of plagioclase feldspar, apatite, quartz and an unidentified glass. The 386 core of the sherd is dark brownish gray, and the surface is black (2.5Y3/0 very dark gray). The interior of the bowl is highly polished to a metallic luster. The uneven exterior surface has been scored by short strokes with a sharp implement. Lithic Remains As noted in the previous chapter, stone tools and debitage do not appear in abundance at Manachaqui Cave until the Early Intermediate Period. An estimated 100 stone artifacts, representing one percent of Manachaqui Cave's total lithic remains, belong to the Suitacocha Phase. It is unlikely that all of the chipped stone flakes from Suitacocha Phase layers intruded from earlier and later deposits. However, only ground and polished shale and slate artifacts are confidently assigned to this phase, while three chipped stone artifacts are tentatively included. Of the ground and polished stone artifacts, one preform, five flakes and ten broken segments of ground shale and slate points were retrieved from Suitacocha Phase levels. Another preform, seven flakes and ten point fragments are from mixed Suitacocha, Colpar and Empedrada Phase levels. The three chipped-stone artifacts include two from Unit 4 level 21. One is a tabular core of felsic igneous rock from which flakes were removed around the edges leaving some of the cortex. The other is a similar tabular piece of felsic rock, but flakes have been stricken from 387 many directions. One short side has been retouched to create a concave working edge. The third chipped-stone artifact is a gray chert biface with ovoid cross-section. Lustrous surfaces indicate that the stone was thermally altered, perhaps intentionally, prior to flaking. Numerous large hinge and step fractures suggest that the chert continued to present insuperable flaws despite heating. Rocks and Minerals Rocks and minerals from Suitacocha Phase levels include one chunk of unworked vein quartz, and a small piece of mica. Another piece of mica and a broken piece of a prismatic quartz crystal were recovered from mixed Suitacocha and later levels (37-7 and 18-6 respectively). Botanical Remains The reader may refer to Table 18 for the list of botanical taxa identified by Pearsall in Suitacocha Phase Floors Y and X in Sector A. Again, Festuca and Chenopodium/Amaranthus represent potential food sources obtainable within the immediate site environs. The Sapotaceae and other unidentified fruit rinds recovered may have been brought in from lower elevations. The most significant new food counted among those exotic to the Tropical Alpine zone are maize (Zea mays) and beans (Phaseolus). These both appear in Floor Y, the earliest of the two Suitacocha Phase floors. Maize also occurs in Floor 388 X although beans do not. Faunal Remains A total of 41 faunal specimens were retrieved from unmixed Suitacocha Phase levels {Table 23), and 32 of these could be identified by Class {Kent 1994). In addition to Mammalia, the inventory includes one Ave {bird), one Reptilia {reptile) and two Osteichthyes {fish). Seventeen of the 28 Mammalia specimens could be identified at finer taxonomic levels. Of eight Rodentia, one belongs to the Muridae Family, and three Caviidae were identified to the Genus Cavia. None of the Suitacocha Phase remains could be positively identified to the species level, although one White-tailed Deer (Odocoileus virginianus) specimen was retrieved from Unit 25 Level 7, a mixed ManachaquiSuitacocha Phase context {Table 22). One Artiodactyl specimen recovered in deposits between Suitacocha Phase Floor X and Colpar Phase Floor W {Unit 14, Level 34) is tentatively classified as Camelid {Table 24). This specimen would be the earliest occurrence of Camelidae excavated from Manachaqui Cave. Suitacocha Phase Chronology Given the acceptable radiocarbon dates of 850, 860 and 900 b.c. for the uppermost layers of the Manachaqui Phase deposits, and 500, 610, 680 and 790 b.c. for the Suitacocha Phase deposits in both Sectors A and B, 800 b.c. presents 389 itself as an appropriate beginning date for the Suitacocha Phase. Some of the vessel forms and virtually all of the decorative techniques that characterize the Suitacocha Style are likewise found in late Initial Period/Early Horizon occupations at sites like Ancon, Kotosh, Huacaloma, Cerro Blanco, Kuntur Wasi, Kotosh and Chavin de Huantar. Radiocarbon evidence from these sites (Burger 1981:599, 1992:230-233) suggests that mid-Initial Period stylistic shifts at Manachaqui Cave lagged one or two centuries behind major restyling events across the north-Central Andes. However, some of these sites suffer dating problems, and only additional well-dated cultural sequences will modify or reify the current panorama of uneven patterning in Central Andean cultural development. Discerning the end of the Suitacocha Phase is more problematic. The next ceramic component clearly isolated utilizing Manachaqui's macrochronology features kaolin ware pottery, and can be confidently cross-dated to the Early Intermediate Period. In Chapter 5, an occupational hiatus from 450 b.c. to 200 b.c., hypothesized. evidence. (during the Chavin horizon) was This hypothesis is sustained only by negative There are no clear indications of the Chavin cult's stylistic influence (as defined by Burger 1984b, 1988, 1992) on Manachaqui Cave's pottery, nor is there ceramic evidence of any style change prior to the mid-Early Intermediate Period. The centuries between 500 and 200 b.c. 390 are not represented within Sector A's floor sequence, nor in Manachaqui Cave's radiocarbon sequence. Most likely, the shelter continued to serve local populations infrequently during the mid-Early Horizon, but played no significant role within Chavin interaction spheres. The problem of isolating evidence for a ceramic component corresponding to the first centuries of the Early Intermediate Period will be addressed in Chapter 8. Paste A Relationships: the Central Andes Early Horizon Central Andean pottery styles are characterized by expanding inventories of vessel shapes and decorative techniques. Many Central Andean pottery traditions now feature polished wares, either red-slipped or smudged black, on which incised, punctate, engraved, stamped, line-burnished and post-fire painted decorations were executed. While the principal vessel shapes remained neckless ollas and open bowls, short-necked jars with everted and sometimes beveled rims like Manachaqui and Suitacocha Phase Shape B appear in variable amounts along the eastern edge of the Central Andean cordillera at Piruru (Rozenberg and Picon 1990: Figs. 2f, j), Kotosh (Izumi and Soto 1972:203), Chavin de Huantar, Pachamachay, Telarmachay, Waywaka and Marcavalle. Rim lobes are common features on such eastern highland jars. Rarely are the rims reinforced, however, and none exhibit profiles like the distinctive 391 Suitacocha Phase Shape B Rim 11 variants. Jars with tall, direct rims identical to Suitacocha Phase Shape F begin to appear in the Early Horizon Central Andean heartland in relatively small percentages. Examples are Middle Guafiape Form 4 (Strong and Evans 1952: Fig. 51), La Pampa Brown Form 3 (Terada 1979: Pl. 103:18-21) and perhaps Chavin de Huantar Urabarriu Phase Jars 2 through 5 (Burger 1984b: Figs. 48-51). In the Chicama Valley, globular jars with vertical rims and decorated shoulders constitute minority elements of the Cupisnique Culture as defined by Larco (1941). An illustrated example (Ibid.: Fig. 78 top, far right) presents an zoned-punctate motif of pendant triangles resembling Suitacocha decorated sherds (Fig. 85h). Neither systematic ceramic studies nor analyses of absolute chronology have been effected for Cupisnique to date. At Chavin de Huantar, jars with vertical rims constitute only five percent of the Urabarriu Phase shape inventory (Burger 1984b:44), and they apparently lack Shape F's characteristic right angle joint at the base of the neck. Jars identical to Suitacocha Shape F first appear during the Chakinani Phase (500 to 400 B.C.) as Jars 3B, 6 and 7 (Ibid. :Figs. 168-172). Shape F-like jars are entirely absent from Early Horizon Kotosh (Izumi and Sono 1972), Huaricoto (Burger 1985b) and the central coast. Thus, the shape's distribution is uneven, but largely restricted to 392 the north-Central Andes. Most importantly, Central Andean necked jars rarely, if ever, carry decoration other than polish or slip. Central Andean potters mostly embellished bottles, cups and open bowls, shapes absent or rare in the Suitacocha assemblage. Decorative techniques like zoned punctation and redslip can be found throughout the Early Horizon Central Andes. The manner in which slip is utilized at Huaricoto for creating contrasts between the painted vessel body and the unpainted, zoned decoration on the vessel shoulders (Burger 1985b) resembles the Suitacocha Phase use of slip painting. At Chavin de Huantar however, slipping rarely serves as more than a finishing technique (Burger 1984b:51) Suitacocha incised motifs found elsewhere in the Central Andean heartland include pendant triangles filled with punctation, and zoned rows of punctations. Kotosh seems to share the greatest number of motifs with Suitacocha, including arches and rectilinear U-shapes (Izumi and Sono 1963, 1972). The latter motif is more typical of the earlier Kotosh Wairajirca Phase, and it is also found in an early Nepefia Valley coastal context (Proulx 1985: Pl. 3c). Closer to Manachaqui Cave, the Huamachuco area shows evidence of Early Horizon occupation at several sites (Thatcher 1979; T. Topic and J. Topic 1987), but the collection of diagnostic material assigned to the Colpa Phase is small (T. Topic personal communication 1994). Some 393 "ollas with straight outflared rims" (T. Topic and J. Topic 1987:14) from Site 102 apparently confirm the presence of necked jars. Thatcher's Shape 3B (1979: Figs. 18, 19) may represent Suitacocha Shape F's Huamachuco counterpart. His bowl Shape 16 resembles Suitacocha Shape E Rim 8. Nevertheless, neckless ollas and open bowls with red slip comprise the bulk of the Colpa Phase assemblage. Thatcher reports short-necked jars during his Mamorco Phase (1972: Fig. 4n-r) but, as previously noted, the existence of the Mamorco Phase has been placed in doubt (and the pertinent rim sherds thus consigned to temporal limbo). Incised decorations on Colpa Phase bowls show little resemblance to Suitacocha Phase examples. With the exception of neckless ollas with grooved rims like Shape A Rim 14, the unusual Cerro Pelon style (Zaki 1983) displays few resemblances to Suitacocha. Again during the Suitacocha Phase, the most salient similarities are shared with the northernmost Central Andean sites in Cajamarca. Suitacocha Phase Shape F jars are common, although still outnumbered both by neckless ollas and open bowls at Huacaloma, Cerro Blanco and perhaps Pacopampa. Some Huacaloma and Cerro Blanco examples are identical to Shape F Rims 1 and 4a, even showing the characteristic red slip and polish (e.g. Terada and Onuki 1982: Pl. 79:4, 1985: Pl. 60:11, 1988: Fig. 28:7-8). A few Huacaloma Line Burnished Type Form 3 vessels are identical 394 to Shape B Rim 11f, including the distinctive inflection at the base of the neck (Terada and Onuki 1980: Pl. 80:5, 1985: Pl. 58:6). Such jar rim forms occur in minor proportions at the Cajamarca sites. Pacopampa and Pandanche jar shapes also have Suitacocha Style counterparts. Pacopampa Group II rims, some of which bear red slip, closely resemble Suitacocha Shape F Rim 2 (Fung 1975: Lams. 9:3, 10:23-25, 11:26-37), Shape BRim 11b (Ibid.: Lam. 10:12, 13) and Shape BRim 11e (Ibid.: Lam. 10:15). Likewise, Pandanche Phase B1 rims parallel those of Suitacocha Shape F (Kaulicke 1975: Lam. XII: second row from top, five at left), and Shape B (Ibid.: top row, left). five at Kaulicke (Ibid.:45) reports that some short-necked jars were decorated. The popularity of Inciso cortante like that observed on the solitary Suitacocha Shape E Rim 10 eclipses other decorative techniques of the Pacopampa Pacopampa Phase. Burger suspects an Urabarriu Phase bowl sherd with inciso cortante of being foreign to Chavin de Huantar (Burger 1984b:67). The Suitacocha Phase inciso cortante example, and the Shape X Rim 4 sherd with combed decoration (cf. Rosas and Shady 1970: Fig. Sb) could be "imports" from Cajamarca. Kaulicke (1975:45, 53) postulates that both non-local influences and non-local pottery in the Phase B1 assemblage derive from the lowland tropical forests and Ecuador. Cajamarca decorative techniques also display close 395 Suitacocha Phase parallels. Among these are the use of red slip to contrast with zoned incised and punctate fields (Terada and Onuki 1988: Figs. 27:2, 28:13, 14, 16), restricted application of red slip to the lip and upper portions of jar rims (Ibid.: Figs. 27:18, 22; 28:7, 8), and parallel notched applique ribs on red-slipped bowls (Ibid.: Figs. 30:9, 12). The red-slipped, polished and incised bowls illustrated in the Huacaloma (Terada and Onuki 1985: Pl.56:10) and Cerro Blanco (Terada and Onuki 1988: Fig. 27:2) monographs provide the closest matches for comparison to distinctive Suitacocha Shape C Rim 8 decorated bowls. the other hand, On there are no polychrome bowls or jars (e.g. Terada and Onuki 1988: Fig. 32:4) at Manachaqui Cave, and white slip is likewise absent. In the western slope yunga of Cajamarca, middle Jequetepeque valley sites have also yielded Shape F type rims (Ravines 1982: Fig. 110:7; Alva 1986: Figs. 407 and 411, 413, 414; Ulbert 1994: Lam. 29) identical to the previously cited Cupisnique example (Larco 1941: Fig. 78). The Montegrande assemblage described by Ulbert continues to display stylistic correspondences to Manachaqui' Cave's early pottery assemblages. Some jars from private collections illustrated by Alva, especially Fig. 411, might be almost indistinguishable from Suitacocha Phase Shape F Rim 1b examples. The middle Jequetepeque publications illustrate no Shape B analogues, but in the upper Zafia 396 valley montane forest, Alva recovered a small sample of "short-necked jars with reinforced rims" identical to Shape BRims 11c (Alva 1988a: Figs. 37:7 and 38:19 and 25). Red slip, zoned red painting, geometric zoned punctation and incised line geometric motifs common to the Jequetepeque and Zafia valleys (Ulbert 1994: Lam. 14; Alva 1988a:347) are among diagnostic Suitacocha attributes. Montegrande basal angles from red-slipped carinated bowls (Ulbert 1994: Lam. 15: Bowl A7) provide potential analogies for the Suitacocha Phase carinated vessels. Kaulicke (1975:45) unearthed one decorated sherd from a similar vessel (Ibid.: Lams. XII bottom right, XIIIb) in a Pandanche B1 context. The Early Horizon Jequetepeque, Zafia and Pandanche materials all lack associated radiocarbon dates. While the Cajamarca assemblages are clearly the closest stylistic peers of the Suitacocha Style, some significant differences should not be overlooked. striking, yet they are sporadic. Affinities may be At Huacaloma and Cerro Blanco, like elsewhere in the Central Andes, neckless ollas are still far more important than necked jars. Decorated, often polychrome, open bowls outnumber both olla and jar shapes at these Cajamarca sites. In contrast, the Suitacocha neckless jar is the least popular of three jar shapes. Shape BRim 11 short-necked jars with distinctive folded and beveled rims constitute 60 percent of the 397 Suitacocha Phase rim collection, yet similar rims are rare or absent elsewhere in the Central Andes. Open bowls comprise only one percent of the Suitacocha Phase collection, while open and restricted bowls together encompass only 6.6 percent. Undoubtedly, Manachaqui Cave's prehistoric functions were largely responsible for the site's vessel shape inventory. Nevertheless, neckless jar shapes, which differed morphologically from the Central Andean variety during the Manachaqui Phase (see Chapter 6), may never have been very important in the northeastern montane forests. Kaulicke neither mentions nor illustrates neckless ollas/jars in his discussion of Pandanche Phase Bl vessel shapes (Kaulicke 1975:45) One might also consider that only its smaller size and the addition of its short, everted rim distinguishes Manachaqui Cave's Initial Period shortnecked jars (especially Shape B Rims 1 and 2) Central Andean neckless olla. from the The neck addition may be deemed a stylistic detail, and the two vessel shapes' respective functions may consequently be regarded as interchangeable if not redundant. The rarity of the ubiquitous Early Horizon Central Andean decorated bowls also likely reflects Manachaqui Cave's specialized functions. Such bowls are typically associated with ritual contexts such as Chavin de Huantar's Ofrendas Gallery (Lumbreras 1993). They are thought of as 398 having functioned, at least on occasion, to convey ritual offerings (Mohr-Chavez 1981b:343). Ulbert (1994:142) finds that red-slipped and incised bowls are closely associated with the monumental constructions at Montegrande. They also frequently accompanied the Early Horizon Central Andean dead into the after-world as grave goods (e.g. Larco 1941, Alva 1986) . The Suitacocha Shape C Rims 7 and 8 vessels must represent a larger, more diverse decorated bowl tradition on the Marafion-Huallaga divide. Also, decorated Suitacocha Phase jars may have functioned in some of the same social and religious contexts as Central Andean bowls. Regarding regional comparisons, however, it is most significant that there is no evidence for the Cajamarca post-firing or polychrome painting techniques on Suitacocha Phase bowls or jars. Manachaqui Cave's divergence from eastern slope styles farther south may be deduced from the lack of shape analogues, and painted and applique decorations at Kotosh. Paste A Relationships: the Amazonian Lowlands East of Manachaqui Cave in the Central Huallaga valley the first documented human occupation is coeval with the Suitacocha Phase. At S-Huay-5 on the lower Huallabamba, Ravines excavated red-slipped and excised pottery that he assigns to the Shakimu Style first identified on the Central Ucayali River course by Lathrap (1962, 1971). Lathrap offers the radiocarbon date of 650 ± 100 b.c. to which 399 DeBoer (1972-74:95) adds a date of 830 ± 135 b.c. for the closely related Bin6 Style of the upper Ucayali. DeBoer's illustrations of rim treatment on Bin6 cups and jars (Ibid.:Figs. 12-17; also Figs. 46 and 50) show close resemblances to Suitacocha Shape BRim 11 and Shape C Rim 7. At variance with the Suitacocha Style, however, large open vessels appear to be the most important Shakimu shape class. All of the common Shakimu vessel shapes, with the exception of the convex excised bowls, have sharp basal angles and carinated profiles. Shakimu carinated bowl shapes (Ibid.: Figs. 5-8; Lathrap 1970: Fig. lOa-d, k) more closely resemble the earlier Manachaqui Phase Shape D unrestricted bowls with concave walls, but temporal disparity makes such similarities difficult to evaluate. Red slip seems to be the only decorative feature shared between the Shakimu and Suitacocha assemblages. Although larger ceramic samples and radiocarbon dates obtained from the Central Huallaga may prove otherwise, the two styles do not appear to be closely related. Following the Pangotsi Phase in the Alto Pachitea drainage, the Nazaratequi Phase may be coeval with, or slightly later than the Suitacocha Phase (Lathrap 1970:98). From the Pangotsi Phase, Nazaratequi retains a large, necked jar shape with an unthickened rim (Allen 1968: Fig. 15, Forms 12 and 13), but carinated bowls dominate the phase assemblage. The presence of these necked jars distinguishes 400 Nazaratequi from the neighboring Shakimu Style and loosely allies it with styles to the west and north (i.e. the Central and Northern Andes). Allen (Ibid.:352) again notes the composite nature of the assemblage. Nevertheless, there is little, especially in the way of decorative attributes, to suggest a specific relationship to the Suitacocha Style. Far more arresting are Suitacocha's formal and decorative similarities to Bagua I and II assemblages from the Marafion and Utcubamba Valley lowlands to the north. Neckless ollas are also rare at Bagua (Shady and Rosas 1979:112), but a polychrome bowl tradition links the lower Utcubarnba region to highland Cajarnarca. Bagua's vessel shapes include necked jars with folded rims (Shady 1987b: Fig. 2f-i) like Suitacocha Shape B Rim 11a, and necked jars with vertical rims (Ibid.: Fig. 4a-d, Sg-i, 6a-c) like Suitacocha Shape F Rims la, 1b and 3. Bagua Style jar rims are much taller on average, but they do commonly bear red slip. The jars also carry simple fine-line incised decorations on the rims (Ibid:. Fig. Sg, h) and upper shoulders (Shady and Rosas 1979: Fig. 130). Zoned punctation diagnostic of the Suitacocha Phase only becomes important during Bagua II. Carinated bowl shapes of Shady's Alenya Complex (Shady and Rosas 1982: Fig. 4b) provide analogies for the Suitacocha basal angles from carinated bowls. Alenya rims match Suitacocha Shape X Rim 3. Intrusive trade sherds at Bagua indicate long-distance 401 exchange networks, and support the Bagua Phase's temporal alignment with the Kotosh Kotosh Phase in Huanuco, the Urabarriu Phase at Chavin de Huantar and the Pacopampa Pacopampa Phase in Cajamarca (Shady 1987b:481}. Limited reconnaissance and test excavations in the lower Utcubamba and nearby Chinchipe areas have yielded early examples of necked jars (e.g. Rojas 1969: Lam. VI top a-c, e), and jars or bowls with reinforced rims (Ibid.: Lams. VI top d, g and bottom a). The basal angle from a carinated bowl illustrated by Bushnell (1966: Fig. 1 top) is similar to the Suitacocha examples, while the distinctive rim from a "massive vessel" (Ibid.: Fig. 2 bottom) is virtually identical to Shape E Rim 9. These parallels are striking given the small size of the Utcubamba-Chinchipe samples. So far this lowland region lacks radiocarbon evidence for absolute chronologies. Farther afield, on the eastern slopes of the Ecuadorian cordillera near Macas, early necked jars are found at the Sangay mound complex (Porras 1987, 1989}, and at the Yaunchu site originally excavated by Harner (Rostoker 1988). Despite numerous radiocarbon measurements (Porras 1987:296}, the absolute dating of these Upano Valley finds remains problematic. Based on Porras' dates, Athens (1986:112) estimates a maximum antiquity for the Sangay Upano Tradition of 400 b.c. Rostoker (1988:75-76) employs several lines of evidence to suggest greater antiquity for the stylistically 402 related Yaunchu remains. Although these ceramic samples might post-date the Suitacocha Phase by several centuries, some similarities are too specific to be overlooked. Interpreted Upano Tradition jar shapes from Sangay closely resemble Suitacocha Shape F Rims 1, 2 and 3 {Porras 1987: Figs. 76a-ch, 77a; see also Ramp6n 1959: Fig. 2). Many Sangay jars are polished and bear incised-line and redbanded-incised decorations on the rims and vessel shoulders {Porras 1987:199-202). Yaunchu also yielded necked {or "collared") jars {Rostoker 1988: Figs: 65-70), one of which has a distinctive lip treatment identical to Suitacocha Shape F Rim 8 (Ibid.: Fig. 68). These Yaunchu rims carry red paint or incised-line decorations, and the upper shoulders may be incised as well (Ibid.: Figs. 69, 71). From a private collection reportedly from Macas, Bushnell {1946: Fig. 2f) illustrates a necked jar with vertical rim and incised decoration on the shoulder. These Upano rims are on average about one or two em. higher than their Suitacocha Phase counterparts. Porras (1989: 378) notes that 89 percent of the Upano Tradition pottery at the Sangay mounds consists of shallow bowls and plates. The only attributes paralleled in the Suitacocha bowl assemblage might be the basal angle on these Upano style vessels {Porras 1987: Figs. 70ch third from left, and 73c far right; Ramp6n 1959: Fig. 3, Nos. 15, 16), and mamiform legs {Porras 1987: Fig. 99). Rostoker also 403 illustrates carinated bowls (1988: Figs. 25-29, 37). His Fig. 28, 29 and 37 bowls carry rims identical to Suitacocha Shape X Rim 3. Another unusual bowl rim with a raised surface (Ibid.: Fig. 54) may provide an analogy for the sherd illustrated in Fig. 80d (this thesis). Upano decorative motifs most similar to Suitacocha's include simple incised lines, arrays of divergent lines (Porras 1987: Figs. 47c, 48b, 54a), incised arches (Ibid.: Fig. 51i) and zoned punctation (Ibid.: Figs. 58 and 59). None of these motif correspondences is very precise. More intriguing are parallels in subject matter of Upano Tradition and Suitacocha Phase adornos. These are zoomorphic and anthropomorphic heads that seem to have been attached to the shoulders of jars. Porras describes three zoomorphic adornos as an anteater or opposum (Ibid.: Fig. 116b), a bat (Ibid.: 116c) and a bird (Ibid.: Fig. 116d). The similarities in detail shared by the actual pieces cannot be evaluated using Porras' illustrations. One anthropomorphic adorno not discussed by the author (Ibid.: Fig. 107e) is a human head bent sideways at the neck much like its Suitacocha Phase analogue. A sketch of an additional zoomorphic adorno representing an unidentified animal head is presented by Bushnell (1946: Fig. 2g). Many resemblances shared between Suitacocha and the distant Ecuadorian Oriente styles are precise and, therefore, difficult to dismiss regardless of what criteria 404 archaeologists believe paramount for comparative analyses. Necked jars, especially jars with vertical rims and decorated shoulders, are diagnostic elements in early eastern Ecuadorian pottery styles. Necked jars illustrated for the Pastaza Phase (Porras 1975b: Fig. 8) dated somewhere between 100 b.c. and a.d. 700 (Athens 1986), and Tivacundo Phase jars (Evans and Meggers 1968: Fig. 19 Shape 5) radiocarbon dated at a.d. 510 apparently represent the shape's continued popularity in the Oriente. Morphologically analogous vessels do not appear in southwestern Amazonia until approximately a.d. 800 when Cumancaya and related styles spread throughout the Central Ucayali (Raymond et al. 1975) and Central Huallaga valleys east of Manachaqui Cave (DeBoer 1984:114-115; n.d.), puportedly carried in by Lathrap's Quechua migrants. Paste A Relationships: the Northern Andes In the coastal lowlands separating the Central Andes of Peru and the Northern Andes of Ecuador, coeval pottery styles exhibit features diagnostic of both areas. Necked jars and elaborate incised bowls dominate Formative Period assemblages in which neckless ollas are rare or absent. Necked jars with direct rims have been recovered from Batan Grande (Shimada et al. 1982: Fig. 39; 1990: Fig. 13b-e), Lanning's (1963: Fig. 1f) and Ravines' (1986-87: Lam. 1:3) Paita Phase C and D sites and Nafiafiique's Panecillo Phase 405 contexts (Guffroy 1989: Fig. 8a-c). These typically resemble the tall and slightly flaring Bagua and Pacopampa rims that rise to between 4 and 8 em, much higher than the vertical Suitacocha Shape F rims. Paita Phase D, Panecillo Phase and Pechiche Phase jars mostly continue Valdivia and Machalilla tendencies toward gently curving, everted necks and unthickened rims (e.g. Lanning 1963: Figs. 7a-c, 8; Ravines 1986-87: Lam. 1; Guffroy 1989: Fig. 8; Izumi and Terada 1966: Pl. 32:1-13). One rim assigned to Paita Phase B is identical to Suitacocha Shape BRim 11a (Lanning 1963: Fig. 21b). The distinctive inflected upper shoulder diagnostic of Suitacocha Shape B Rim 11 jars is illustrated by Lanning (Ibid.: Fig. 4b) and Guffroy (1989: Fig. 8e). Incised zoned punctation and/or zoned red slip and painting appears on all of the phase assemblages coeval with Suitacocha. Some jars exhibit the tendency to locate design motifs on the upper shoulders (e.g. Lanning 1963: Fig. 7; Izumi and Terada 1966: Pl. 32:15; Guffroy 1989: Fig. 8d. c). Farther inland in the Loja Province of southernmost Ecuador, the transition from Catamayo B short-necked jars with everted rims and beveled lips (Guffroy et al. 1987: Fig. 12) to Catamayo C necked jars with vertical rims (Ibid.: Fig. 13) mirrors an identical shift in jar rim morphology at Manachaqui Cave. Catamayo C rims closely parallel Shape F Rim 1 (Ibid.: Fig. 14f, i), Rim 2 (Ibid.: 406 Figs. 13a-c, 14a-c) and Rim 5 (Ibid.: Fig. 14d) in details of thickening and average height. Loja jar bodies and lips are frequently slipped red (Ibid.:90). Simple incised line decorations ornament rim exteriors (Ibid.: Figs. 13a, b, 14a, g, k), and notched applique, anthropomorphic and zoomorphic adornos are also associated with the jar shapes (Ibid.: Figs. 16d, e). Other decorative techniques such as zoned punctation are apparently absent. Nonetheless, the Catarnayo C Style's correspondences to Suitacocha Phase ceramic attributes are perhaps more precise than the parallels shared between Suitacocha and Bagua. Archaeologists working in the frontier between the Central and Northern Andes often note similarities between their materials and the widespread Chorrera style (800 - 300 b.c.) of coastal Ecuador (e.g. Izumi and Terada 1966:88; Shimada et al. 1982:146-147; Guffroy et al. 1987:113). Information regarding Chorrera material culture is found scattered in numerous publications (e.g. Bushnell 1951; Estrada 1958; Meggers et al. 1965; Meggers 1966; Lathrap et al. 1975; Evans and Meggers 1982 and Bischof 1982). The unambiguous resemblances between Chorrera and the Suitacocha style are only surprising if one overlooks the numerous design features shared by the Valdivia/Machalilla and Initial Period Manachaqui assemblages. Bushnell (1951: Figs. 37a-e) illustrates two Engoroy (the Santa Elena Peninsula Chorrera variant) classes of 407 necked jars with vertical rims. The wide-mouth class in particular (Ibid.: Figs. 37c, e) mirrors Suitacocha Shape F. Oddly, Bischof's (1982) Engoroy jars do not include this shape. Estrada illustrates jars of the same narrow and wide-mouth classes from the coastal site of Ayangue (1958: Fig. 48:1 and Fig. 47:4 respectively). The coastal Formative Period sequence established by Meggers et al. (1965) not only demonstrates the great antiquity of decorated necked jars in the Northern Andes, but documents their persistence into the Chorrera Phase. Red-slipped and polished vessels date to the beginning of the Valdivia sequence, as does the tendency to decorate jar shoulders. The Chorrera counterpart of Suitacocha Shape F is "Machalilla Polished Red" Vessel Form 6 (Ibid.: Fig. 84:6). The Form 6 rim grades into a gently curved everted form, but an angular junction of the neck and body visible on the interior surface is a diagnostic attribute. Red slip covers the jar exterior and the rim interior down to the angle. Machalilla Striated Polished Plain Vessel Form 9 is the same shape without red slip. Both of these "Machalilla" pottery types figure prominently during the subsequent Chorrera Phase (Ibid.: Fig. 89). At Pefion del Rio, a Late Formative Period habitation site in the Guayas Basin, vessel Forms 15A-C mirror Suitacocha Shape F. The sharply angled joint at the base of the rim is diagnostic of the Chorrera jars. However, like other western Ecuadorian jars, the Pefion del 408 Rio varieties carry strictly everted rims of varying lengths and thicknesses. Suitacocha Shape F rims show more variable rim orientation angles and rim thickening techniques, thus they more closely resemble the Upano Tradition jars from the Oriente. Machalilla Polished Plain Jar Form 4, Machalilla Striated Polished Plain Jar Form 10 and Machalilla DoubleLine Incised Jar Form 6 with reinforced lips vaguely recall Suitacocha Shape B Rims 11d-g. The Suitacocha Shape B short-necked jars really have no precise Chorrera match. Chorrera shallow carinated bowls from La Ponga illustrated by Lippi (1983: Fig. 53b [bottom row, second from right]) exhibit basal angles analogous to the Oriente and Suitacocha examples. Even more precise resemblances between the Chorrera and Suitacocha styles are found in shared decorative techniques. A Suitacocha Shape C Rim 7 sherd with engraved decoration (Fig. 71t this thesis) displays a faithful duplication of the Double-line Incised technique and motif illustrated by Meggers et al. (1965: Pl. 137a). Lippi (1983: Fig. 63, #1506) illustrates the same engraving technique and motif. The Punctate, and Incised and Punctate techniques (Meggers et al. 1965: Pl. 144w-z, 147b-i) mirror the Suitacocha examples depicted in Figs. 86c-k, 87a-h. Red slip creates contrasting painted and punctate fields separated or "zoned" by incision (Lippi 1983: Fig. 73). Bosses embellish 409 shoulders of other Machalilla/Chorrera vessels (Meggers et al. 1965: Pl. 143), and Suitacocha incised motifs common in coastal Ecuador include simple arrays of divergent parallel lines (Ibid.: Pl. 144h), zoned superimposed X motifs (Ibid.: Pl. 133n; Lippi 1983: Figs. 64, 85, 88), cross-hatching (Meggers et al. 1965: Pls. 131, 133), arches (Ibid.: Pls. 145e, 148; Lippi 1983: Figs. 71, 72) and pendant triangles (Meggers et al. 1965: Pl. 132h). Delineating, or "zoning" red-painted fields with incision gained popularity during the Machalilla Phase (Ibid.: Pl. 145a-i) and became a diagnostic attribute during Chorrera (Ibid.: 139 and Pl. 157b, c; Meggers 1966:58; Evans and Meggers 1982), as do polished red rims, incised rims, iridescent paint, applique rib, plain rocker stamping, and some techniques unknown for Suitacocha. Interpretations of Suitacocha's external relations must take into account that many of the shared Northern Andean incised motifs: 1) are engraved, 2) occur on bowls as a preferred medium rather than jars, 3) are distributed widely over the Northern Andes and the Oriente, and 4) may appear at different prehistoric moments in different regions (they rarely appear in welldated stratigraphic contexts). Furthermore, the temporal and stylistic relationships between Machalilla and Chorrera remain to be clarified with stratigraphic evidence. In the Ecuadorian highlands, necked jars with redslipped and polished vertical rims and incised shoulders 410 appear with the earliest documented pottery styles at Monjashuaico (Bennett 1946: Figs. 10, 11), Huancarcuchu (Ibid.: Figs. 6, 7), Cerro Narrio (Collier and Murra 1943: Figs, 10 top two rows left, 12 top row) and Cotocollao (Villalba 1988: Fig. 128). Some Huancarcuchu jars with fields of punctation and red paint zoned by incision (Bennett 1946: Fig. 71, 9o) closely resemble decorated Suitacocha jars in general design layout. Unlike Suitacocha, however, Huancarcuchu red paint was habitually applied in parallel bands. The manner in which jar shoulders from Cafiar Province carried zoomorphic and anthropomorphic adornos (Uhle 1922: Lams. X-XII) provides an analogy for the placement of the Sangay and Suitacocha adornos. Incised line motifs common to Suitacocha and south highland Ecuador (Ibid.: Figs. 9, 11, 13; Collier and Murra 1943: Pls. 19: 12, 13) likewise match motifs on the Machalilla/Chorrera pottery types just discussed (Meggers 1966:64). Similar motifs, occasionally combined with punctation, are common at Cotocollao (Villalba 1988: Fig. 88, 90, 92, 96) and in the northern Ecuadorian province of Irnbabura (Myers 1976: Figs. 2, 3). Porras graphically demonstrates these pan-Ecuadorian resemblances with illustrations (Porras 1982: Lams. 38-53). Observe the likeness shared by incised sherds from Machalilla (Meggers et al. 1965: Pl. 144h), Sangay (Porras 1987:54a), Charnbira 411 in the Peruvian Amazon (Morales 1992: Fig. 4a) and Cerro Narrio (Collier and Murra 1946: Pl. 19, Figs. 12). Clearly there is a need to sort out the temporal and stylistic relationships among these assemblages. The reader will recall that Early Cerro Narrio shares affinities with the preceding Manachaqui Phase style as well as Suitacocha. Meggers (1966:62-64) observes that early south highland pottery combines Machalilla incised motifs and Chorrera vessel shapes. assemblage. So does the Suitacocha Phase Unfortunatley, reliable temporal information regarding Northern Andean ceramic attributes most pertinent for a comparative analysis of the Suitacocha Phase assemblage is not yet available. Nonetheless, it should be apparent to the reader, especially in light of the relatively solid data base for Formative Period south coastal Ecuador, that the Suitacocha Style is in large measure comprised of diagnostic attributes that possess great antiquity and wide geographic distribution in preHispanic Ecuador. Paste B Origins The Suitacocha Phase Paste B sample offers additional evidence for interaction over very long distances. Paste B4 pottery likely originated far to the north in the Ecuadorian Oriente. The sources for Paste B5 probably lay on the Central Andean north coast. These identifications have not been substantiated by hands-on comparisons with sherds from 412 the suspected source areas, although such comparisons are warranted. Paste B4 The Paste B4 sample of 38 sherds shows similarity to several coeval styles with respect to decorative attributes. Rows of large punctations arrayed in zoned-incised fields occur on Kotosh Phase bowls at Kotosh (Izumi and Terada 1972: Pl. 116), and Urabarriu Phase neckless ollas, bowls and cups at Chavin de Huantar (Burger 1984b: Figs. 65-69, 21, 99B). Sherds with these kinds of motifs also appear intrusively in a Machalilla context at Salango (Norton et al. 1983: Fig. 12a-e). Because the Paste B4 sherds belong to decorated necked jars, origins farther north than Kotosh or Chavin are indicated. The Paste B4 sherds approximate descriptions and illustrations of Upano Punteado (Upano Punctate) Type pottery recovered from excavations at Sangay (Porras 1987:153-158). Porras describes non-plastic inclusions in the paste as "river sand, feldspar and quartz." Porras does not specify associated vessel shapes for his small sample of 49 decorated sherds, but necked jars are well-represented at Sangay. The B4 decoration matches his description of Motif 3: "Circular punctations inscribed in double horizontal line, double curved line in some cases, and double oblique line in others" (Ibid. :154). Porras' Fig. 58fi provides a basis for comparing Upano zoned punctation with Suitacocha 413 Paste B4 sherds, and the B4 applique flange may be compared with Upano flanges illustrated in his Figs. 65e and f. Origins in the Amazonian montane or lowland forests of southern Ecuador or northern Peru are most likely. Paste B5 The single black polished sherd of Paste B5 displays a texturing technique observed at Chavin de Huantar, and termed "rough scratch," by Bennett (1944:87). Bennett recovered sherds with these scored applique nubbins from pit CH-10 (Ibid.: Fig. 30c) which also yielded numerous Chakinani Phase (500 to 400 B.C.) sherds (Burger 1984b:167). Similar surface scoring appears in Qotopukyo Style pottery of Lumbreras' Ofrendas Phase (800 to 700 B.C.} at Chavin (Lumbreras 1993: Lam. 45). vessels are bowls. Apparently, none of the Chavin It is also possible that Paste B5 originated in a Cupisnique context on the Pacific coast directly west of Manachaqui. In either case, the sherd is typical of polished black wares of the Early Horizon Central Andes. Prehistoric Cultural Development and Interregional Interaction in the Tropical Montane Forests of Peru Volume 2 A Dissertation Presented to the Faculty of the Graduate School of Yale University in Candidacy for the Degree of Doctor of Philosophy by Warren Brooks Church Dissertation Director: Richard L. Burger December, 1996 CHAPTER 8 THE COLPAR PHASE The Colpar Phase was established during the final laboratory analyses of Manachaqui Cave's radiocarbon dates and ceramic macrochronology. Chapter 7 presented radiocarbon evidence from Sector A (the shelter interior) suggesting at least two distinct occupations (or series of occupations) during the Early Intermediate Period. Colpar Phase represents the earlier of the two. The Analysis of stratified remains from Sector B resulted in the isolation of several vessel shapes and rim forms which did not fit comfortably into either the Suitacocha nor the ostensibly subsequent Empedrada Phase. These shapes and forms constitute a weakly represented, yet nonetheless discernible, component labeled the Colpar Phase. Future radiocarbon analyses will attempt to correlate Sector B's evidence for Early Intermediate Period occupations with Sector A's radiocarbon and ceramic sequences. An additional indication that a temporal component between the Suitacocha and Empedrada Phases can be isolated consists of several "intermediate" levels containing neither incised decoration nor kaolin ware sherds. particularly evident in Units 28 and 30. 414 These levels are The levels that 415 provide most of the Colpar Phase remains include Unit 28 Level 9, Unit 30 Level 9 and Unit 31 Level 9, but it cannot be demonstrated that any of these levels are completely unmixed. The assignment of vessel shapes and rim forms to the Colpar Phase required careful consideration of pertinent sherd proveniences. There is considerable room for error, especially because a small sample is involved. Most difficulties were encountered when attempting to separate shapes/rims belonging to the Suitacocha and Colpar phase assemblages. Colpar Phase ceramics show closer stylistic affinity to Suitacocha Phase pottery than to Empedrada Phase pottery. Finally, the difficulty in isolating unmixed Colpar levels hampers the assignment of lithic, rock and mineral, botanical and faunal remains to the Colpar Phase, and frustrates attempts at even rough quantitative estimates for some of these assemblages. Hopefully, continued archaeological investigations in this region will locate unmixed deposits with which to define more clearly the Colpar Phase. Colpar Phase Ceramics The Colpar ceramic collection is the smallest of Manachaqui Cave's diagnostic phase assemblages, represented by only 134 diagnostic sherds. In most respects, the Colpar Phase represents a continuation of stylistic norms established during the preceding Suitacocha Phase. Most 416 notably, restricted vessels with reinforced rims, globular bodies and round bases continue to dominate the shape inventory. Notched applique and red slip painting are the only decorative techniques confidently assigned to the Colpar Phase. Because of the small sample and the analytical problems just cited, the apparent absence of attributes such as line-incised and textured decoration, punctate applique, flanges and adornos during the Colpar Phase may be apparent rather than real. Shape A neckless jars, Shape F jars with vertical rims and carinated body profiles also drop out of the shape inventory. The highly weathered condition of the small collection further reduces the number of rims that can be securely classified and accurately illustrated. It also hinders evaluation of the role of painted decoration. During the Colpar Phase, ceramics of Paste Group B increase in number and diversity, especially in proportion to the Paste Group A ceramics. appear for the first time. Paste Group C kaolin wares Both paste groups can be considered constituents of a nascent category of "fine wares" contrasting with "coarse wares" of Paste Group A. As in the previous phases, these assemblages are assumed to consist of pottery produced at greater distance from Manachaqui Cave than Paste Group A pottery. It is primarily the stylistic diversity that strongly indicates that both Paste Groups B and C are "intrusive" relative to Paste Group 417 A which, in contrast, exhibits stylistic unity,. The so- called "fine wares" of the pre-Abiseo Phase assemblage at Gran Pajaten show the same lack of stylistic coherence (Church 1994:288). Ceramic Paste Group A The Paste Group A category is comprised of 93 sherds representing 69.4 percent of the total Colpar Phase assemblage. Like its Suitacocha Paste A predecessor, it may consist of more than one ware. Because the Colpar Phase was first distinguished and defined after the collections were no longer available for study, pertinent Paste Group A sherds were not purposefully included in the hands-on analyses. That the overall similarity of Colpar Phase sherds to Suitacocha Phase sherds obstructed recognition of the phase assemblage during the hands-on pottery analysis in Trujillo demonstrates that the basic ceramic technology remained virtually unchanged (at least from a macroscopic perspective) . Morphol.ogy Because of the difficulties in isolating the Colpar Phase ceramic assemblage, it should be kept in mind that the totals (e.g. of rims and decorated sherds) are understated and, consequently, the percentages representing attribute frequencies are distorted approximations. Only 71 rim sherds are assigned to the Colpar Paste Group A assemblage. 418 Some vessel shape and rim form categories seem to have carried over from the Suitacocha Phase. B Rim 11b and Shape X Rim 1 may persist. persistence are less certain. For example, Shape Other cases of The contours and proportions of Colpar vessels remain unknown, although comparison with earlier Suitacocha Phase vessels and coeval vessels from neighboring areas may provide analogies. Shape B: Short-necked jars The 22 rims identified as Colpar Phase short-necked jars represent 31 percent of the Paste Group A rim collection. Like earlier Shape B rims, these are beveled and reinforced by adding clay to the lip exterior, or by folding over the rim end and smoothing it to the vessel exterior. Gone is the narrow groove under the thickened lip that characterized many Suitacocha jar rims. Jar mouth diameters range from 10 to 16 em in diameter, but only ten shape B rim diameters were measured. Rims llh-j (Fig. 91a-j). into three variants. These 22 rims are divided The 13 examples of Rim 11h (Fig. 91a- e) resemble Suitacocha Phase Shape B Rims 11b or 11c, but the pointed lip tends to be bent further outward, and the thickened underside is less pronounced and joins the vessel neck in an unbroken curve. mm. Its thickness ranges from 5 to 9 Five examples of Rim 11i (Fig. 91f-h) resemble Suitacocha Phase Shape B Rim 11d, but the flattened lip is 419 thicker (7 to 10 mmj, and the rim may be more vertically oriented like Shape F rims. The three measured specimens range in diameter from 10 to 12 em. The four examples of Rim 11j (Fig. 91i, j) are vertically oriented, lightly thickened to 7 or 8 mm and the lip is beveled to a dull point. Both of the examples measured showed diameters of 16 em. Shape C: Restricted bowls Colpar Phase restricted bowl rims number 26, representing 37 percent of the Colpar Paste A rim assemblage. All but one of the rim variants are thickened, and all likely pertain to low, rounded vessels. As expected, rim diameters are more variable than those of jars, measuring fr,om 12 to 19 em (N=12). R~ 11 (Fig. 91k-o). The 16 examples of Rim 11 show rounded lips reinforced by rolling the rim end down over the exterior. Thickness at the lip ranges from 5 to 8 mm. Diameters range from 12 to 19 em (N=6). between 3 and 4 mm in thickness. Vessel walls vary The only unweathered sherd from Sector A shows highly burnished interior and exterior surfaces, and clearly visible tracks approximately 1 mm wide left by a blunt-edged burnishing tool. R~ 12 (Fig. 91p-s). The six examples of Rim 12 are lightly thickened to 6 or 7 mm and have flattened lips. 420 Diameters range from 12 to 16 em (N=4). 4 to 5 mm thick. The walls vary from An eroded rim (Fig. 91p) exhibits applique nubbins. Rim 13 (Fig. 91t). Two examples of Rim 13 are unthickened or lightly thickened and show a shallow groove running around the lip exterior. Both fragments were too eroded to obtain a diameter measurement. The walls are approximately 3 mm thick. Rim 14 (Fig. 91u). The single example of Rim 14 is thickened to 8 mm, has a flattened but slightly concave lip, and diameter of approximately 18 em. Its walls are 4 mm thick. Rim 15 (Fig. 9lv). The lone Rim 15 sherd is unthickened with a dull point at the lip exterior. It shows a diameter of 12 em and the walls are 3 mm thick. Shape E: Unrestricted convex bowls The 23 Shape E rims constitute 32 percent of the Colpar Paste A rim collection. the lip. Again, most exhibit thickening near Two of the five rim forms (Rim 3 and Rim 9) occur in the earlier phases (e.g. Fig. 92a-c). The other three thickened and beveled rims are unique to the Colpar Phase. Together the Shape E rims range from 11 to 20 em (N=l3) diameter. in 421 Rim 12 (Fig. 92d-j). The 13 examples of Rim 12 are direct rims with a reinforced exterior extending approximately one em below the rounded or pointed lip, and ranging from 4 to 7 mm in thickness. measure only 2 mm in thickness. Unusually thin walls The rim depicted in Fig. 92j has a small notched lobe projecting about 3 mm above the lip. Rim 13 (Fig. 92k, 1). Two sherds represent Rim 13. They thicken gradually to 6 mm toward the lip which is beveled to leave a dull exterior point and a flat sloping interior surface. A shallow groove runs around the lip exterior 5 mm below the tip. The bowls' walls are 3 to 4 mm thick. Rim 14 (Fig. 92m). The single Rim 14 sherd resembles Rim 13, but the lip is beveled to create two flat surfaces culminating in a point. There is no exterior groove. The maximum thickness reaches 7 mm near the lip, and the walls below are 4 mm thick. Decoration As previously mentioned, the difficulty encountered in stratigraphically isolating the Colpar Phase assemblage coupled with its severe erosion results in the underrepresentation of decorative techniques and inexact sherd counts. At least 22 decorated sherds belong to the Colpar 422 Phase assemblage. The collection's extreme fragmentation hinders the identification of associations between shapes and rims and particular techniques. The following observations are offered with these limitations in mind. Painting Pre-firing paint or slip is usually observed only on sherds from Sector A's sheltered, but often mixed deposits. Based upon these observations, however, it would seem that all of the Colpar Phase vessel shapes and most of the rim forms may bear paint. Red is clearly the most common paint color, although its frequency cannot be accurately assessed. Five of the Shape B Rim 11h and 11i examples (including Fig. 91a, c-e) show traces of red paint. Two of these also show traces of black paint on the lip (Fig. 9la, c). One Shape C Rim 12 sherd (Fig. 91r) exhibits red paint on both surfaces. The surfaces of the single Shape C Rim 14 (Fig. 9lu) example have traces of white slip. Applique Colpar Phase applique may occur as small nubbins as seen in Fig. 91p. The notched applique lobe previously described adorns one Shape E Rim 13 example (Fig. 92j) I and the tradition of decorating jar bodies with notched applique strips apparently continues. Fifteen sherds bearing wide (8 to 12 mrn) strips with cut or incised (not impressed) notches were recovered from suspected Colpar contexts (Fig. 92n-p) I 423 and mixed Suitacocha, Colpar and Empedrada contexts. Probably only a fraction of these sherds truly belonged to Colpar Phase vessels. Their specific location on vessel bodies remains unknown although they were probably medial ribs surrounding jar mid-sections. Ceramic Paste Group B Paste Group B sherds are more numerous and diverse in design details than those of the two preceding phases. Again, because Manachaqui Cave's ceramic macrochronology provides only a small window on the Colpar Phase, it is difficult to assign small samples of Paste B and C wares (especially those from Sector A) to their proper phases. Wares were assigned to phases based primarily upon macrochronological position. However, information regarding regional stylistic developments was an additional consideration for temporal placement. Some 38 sherds representing five sub-groups (or wares) made conspicuous by atypical pastes have been assigned to Colpar Phase Ceramic Paste Group B. In some cases, sherd surface erosion impaired the identification of attributes judged diagnostic of each paste sub-group during the ceramic analysis. Therefore, these totals are also conservative. This may be the case for Paste B10 • The 38 Paste Group B sherds represent 28.4 percent of the Colpar Phase diagnostic sherd total. Of the 86 Colpar Phase rim sherds, 15 or 17.4 percent belong to Groups Band 424 C. This percentage is artificially high because of the difficulties in separating the Colpar Phase assemblages from the preceding Suitacocha Phase and the following Phase assemblages. R~pedrada Recall that some Colpar Phase Paste A Shape B Rim 11 sherds were counted among the Suitacocha Phase rims. One or more of the Paste B wares listed below conceivably belong in the Empedrada Phase Paste B assemblage. Paste B6 Fourteen highly eroded sherds of Paste B6 pottery probably pertain to a single bowl with flat base, direct divergent walls and unknown rim diameter (Fig. 93a, b). The rim shows an unusual grooved or recessed exterior surface. The soft paste is completely oxidized a brick red color (SYR 5/6 yellowish red). Quartz, mica and angular grains of igneous rock comprising non-plastic inclusions vary greatly in size. Although the paste matrix is very fine, some grains reach up to 4 mm in diameter. Paste B7 The thirteen sherds representing this pottery exhibit a dark brown paste with a dark gray core left by incomplete oxidation during firing. They are covered by a brownish black veneer (5YR3/2 dark reddish brown) created by smudging the fired pot in a carbon-rich atmosphere. The dark surface glitters because of abundant mica and smaller amounts of 425 quartz. The sherds pertain to open convex bowls (Fig. 93c- f) with slightly thickened rims and flattened, in-sloping lips closely resembling Shape E Rim 13. A deep incised groove runs around the vessel circumference just lip exterior. ~~der the One bowl interior (Fig. 93f) exhibits decoration with a wide curving line of pink iridescent paint (lORS/3 weak red) that may have been applied with the finger. A small amount of paint is also visible within the groove on the rim exterior. Paste B8 Five hard sherds from three bowls show light gray surfaces (10YR6/2) with smoke clouds and black, unoxidized cores (Figs. 93g; 94a, b). The paste is micaceous with angular bits of quartz and sand. Three articulating sherds (Fig. 93g) comprise a fragment from a simple convex bowl covered with pink slip (lORS/6 red) and painted with three thin, curving white painted lines. The other two bowl fragments are undecorated, but one displays a carinated profile (Fig. 94a) while the shape of the third (Fig. 94b) resembles that of the slipped and painted bowl. Paste B9 Four rim sherds belong to one or two bowls or jars with vertical rims (Fig. 94c-e, Plate XI). The incompletely oxidized paste is fine, light gray (10YR7/2 to 10YR7/4} on the surfaces and dark gray within. Non-plastic inclusions 426 are mostly angular grains of igneous rock. A red (10R4/4 weak red) painted design repeating around the rim consists of Xs and pendant triangles on the interior, and Xs, pendant triangles, lines and dots on the exterior. Paste B~ The two Paste B10 sherds belong to a single open bowl with a thickened rim interior (Fig. 94f). compact and fired hard. The paste is A dark gray core and brown surfaces (5YR3/4 dark reddish brown) indicate incomplete oxidation, and a small smoke cloud covers part of the rim exterior. The non-plastic inclusions are angular grains of igneous rock. Orangish-red paint (2.5YR5/8 red) covers much of the exterior surface, and the small amount of paint visible on the interior surface shades into brownish tan with a metallic luster. The entire bowl is well-polished. Ceramic Paste Group C The Colpar Phase marks the first appearance of fine, well-made pottery rich in kaolin clay at Manachaqui Cave. The wares united under the category of Ceramic Paste Group C in this and the following Empedrada Phase assemblage vary greatly in color and texture. Undecorated or eroded kaolin sherds pertaining to the Colpar Phase could not be isolated for the reasons previously cited. Only three Paste Group C sherds (2.2% of the phase collection of diagnostic sherds) are assigned to the Colpar Phase. 427 Paste C1 Three sherds of extremely fine, fully oxidized (7.5YR7/ light gray) pottery probably belong to an open bowl with a lightly thickened, flaring rim (Fig. 94g). inclusions are invisible to the unaided eye. Non-plastic Two of the three sherds are rims, while the third is a body sherd. The lip exterior is rounded and the interior edge forms a point. Vessel walls are 3 mm thick. The lip exterior shows traces of red paint, and the wall exteriors exhibit a motif composed of thin continuous and broken painted lines. Lithic Remains By viewing the frequency of lithic remains per level within the Colpar Phase macro-chronological contexts, especially those in Units 28 and 30, a pattern of increasing utilization of stone tools during the Early Intermediate Period can be discerned. The elevating numbers mostly reflect the increasing deposition of chipped-stone cores and manufacturing detritus left by a simple core-flake industry. Some flakes of gray and dark green igneous rock and chert were retouched with percussion blows to render steep working edges but, again, no specific totals for tools, cores or waste flakes can be cited for the Colpar Phase. Balancing the suspected increased rate of lithic debris deposition with the overall paucity of artifacts within this rather effemeral phase assemblage, a total of 500 lithic 428 artifacts (about five percent of the Manachaqui Cave total) can be offered as a reasonable estimate. Of the ground- stone tools, only polished slate and shale points occur in Colpar Phase contexts. These points occur in abundance during both the preceding Suitacocha Phase and the subsequent Empedrada Phase. Excavations in Sector B recovered 10 point fragments, one possible preform and seven laminar pieces of slate or shale debitage from levels approximating the Suitacocha-Empedrada phase interface. Rocks and Minerals One piece of mica (Unit 37 Level 7) and one fragment of a prismatic quartz crystal (Unit 18 Level 6) may have been deposited during the Colpar Phase, but could also belong to the earlier or later deposits. Botanical Remains The only floor radiocarbon-dated to the Colpar Phase is Floor W, although Floors U and V might also constitute Colpar Phase deposits. Flotation of Floor Wand V samples produced food remains, while the Floor U sample yielded only charred wood (Table 18). kernels and cob fragments, Floor W contains abundant maize cf. Sapotaceae fruit, fruit rinds, tuber and root fragments, Lupinus, Festuca and Chenopodium and/or Amaranthus. fruit rinds and cf. Ribes. Floor V contains maize, 429 Faunal Remains Only the Sector A specimens recovered from Floors W, V and U are described here (Table 25). Thirty Floor W specimens are classed as Mammalia and one as Osteichthyes (fish). Three of the Mammalia belong to the order Artiodactyla of which two are Cervidae cf. Mazama, and and one is cf. Camelidae. Eleven Mammalia are Rodentia or cf. Rodentia, mostly Caviidae. Two Dasypodidae specimens are probably remains of armadillo. One Osteichthyes from Floor V (probably Colpar Phase but tabulated in Table 26 as mixed deposit) belongs to the ct. Eleotridae family of sleepers, a fresh water fish found in warm estuarian river environments. Nineteen Mammalia specimens were identified from Floor V, among which six Cavia were identified. Floor U (probably Ernpedrada Phase) yielded one specimen identified only as Ave and nine Mammalia, one of which is assigned to the genus Cavia. Radiocarbon Evidence for Colpar Phase Chronology The Colpar Phase is established on the basis of: 1) the two radiocarbon dates from Sector A representing the first centuries of the Early Intermediate Period, 2) small, but perceptible gaps separating the two macrochronological components bearing incised-line decoration (Suitacocha) and abundant kaolin fine wares (Ernpedrada), and 3) an ephemeral, yet distinguishable, ceramic stylistic component straddling 430 the macro-chronological interface between the Suitacocha and Empedrada Phases. In the following sections, cross-dates to this same time period will be intimated by the resemblances shared between Colpar ceramics and pottery styles in neighboring regions. The Colpar Phase awaits further definition through additional work in this area. The radiocarbon dates of 160 b.c. from Unit 15's Floor W, and a.d. 110 from Unit 14's Floor P constitute evidence for Manachaqui Cave's utilization immediately following the Chavin horizon. The temporal lacuna of 340 radiocarbon years (359 calibrated radiocarbon years) separating the latest of the Suitacocha Phase dates from Floors X and Y (500 b.c.) and the Floor W date of 160 b.c. is interpreted as a period of the shelter's relative disuse. Manachaqui Cave probably sheltered travelers infrequently during the Chavin horizon. The Sector A radiocarbon evidence supports a beginning date of approximately 200 b.c. for the Colpar Phase. The determination of an ending date is more problematic. Ceramic evidence for Colpar Phase chronology will be discussed at the end of this chapter. Paste A Relationships: the Central Andes Some scholars have characterized the first centuries of the Central Andean Early Intermediate Period as a stylistic horizon marked by pre-firing "White-on-Red" painted decoration (e.g. Willey 1971; Lumbreras 1974). The 431 proliferation of this decorative technique follows on the heels of the Chavin horizon at many sites previously incorporated within Chavin's sphere of influence, especially in the Callej6n de Huaylas and on the north-central coast. Coastal assemblages such as Bafios de Boza, Puerto Moorin and Salinar are relatively well-dated and described, while coeval highland styles generally lack detailed descriptions and/or absolute dates. At the onset of the Early Intermediate Period, jar shapes most often feature short necks, everted rims and, occasionally, tall spouts. first time on the coast. Carinated jars appear for the Convex bowls exhibit both restricted and open shapes, and both jars and bowls commonly feature vertical and horizontal strap handles. The stirrup- spout bottle shape is most common in the north-coastal valleys directly west of Manachaqui Cave, while doublespout-and-bridge bottles predominate in south-coastal valleys like Ica and Nazca. In the Central Andes, the white-on-red technique is relatively short-lived and eventually eclipsed by negative resist painting. Jar shapes increasingly feature long, flaring collars with direct and tapered rims. Details of design, especially surface decoration, vary from assemblage to assemblage during these centuries, although pre-firing slips and painted geometricline motifs are characteristic. Ceramic variability at this time in the Central Andes is difficult to evaluate given the 432 current lack of knowledge. The clearest similarities to the Colpar Phase assemblage are found one long day's walk eastward to the montane forest site of Gran Pajaten. All three Colpar Shape B rim forms have precise counterparts in collared bowl rims of the Pre-Abiseo Phase pottery assemblage {Church 1994: Fig. 12, 13a-g). (Ibid.: Fig. 12t). Some Gran Pajaten examples show red slip These parallels raise the issue of whether some of the eroded Colpar Phase Shape B vessels should be considered bowls rather than jars. Only comparative analyses of surface treatments on better preserved samples from this period will settle the question. It should be kept in mind, however, that the dichotomy between so-called jar and bowl morphologies is never very rigid during Manachaqui Cave's early ceramic phases. Gran Pajaten's pre-Abiseo Phase assemblage includes cultural remains left by occupations between 420 b.c. and a.d. 460 (Ibid.: Table 1). Based upon the new chronological information from Manachaqui, it is reasonable to suppose that Gran Pajaten's collared bowls are temporally associated with the set of four early dates {420 b.c., 250 b.c., a.d. 20, a.d. 40) recovered within Building No. 1's construction fill. Colpar Phase Shape E Rim 3 open bowls and Gran Pajaten's simple bowls (Ibid.: Fig. 131, u) are also alike, but such common bowl shapes are less useful for comparisons. More significant are resemblances between Colpar Shape E Rim 433 13 and Gran Pajaten beveled-rim bowls (Ibid.: Fig. 40b). The Colpar assemblage apparently contains numerous vessel shapes that are absent at Gran Pajaten, and Gran Pajaten yielded no evidence of notched applique decoration. Nevertheless, the similarity between the Colpar and preAbiseo Phase assemblages is close enough to consider them local variants of a single stylistic tradition. The montane forest site of Cuelap farther north may have been occupied this early, but so far lacks absolute dates for the Early Intermediate Period Cancharin Phase (Ruiz 1972). Cancharin deposits predating Cuelap's impressive architecture include open bowls with beveled rims similar to Colpar Shape E Rim 13 (Ibid.: Lam. XXXIVa-d). These rims and some bowl fragments with white-on-red and negative resist decoration may indicate that at least some of the Cancharin Phase collection dates to the first half of the Early Intermediate Period. However, most of the Cancharin Phase materials, especially the jar shapes, are more akin to later Early Intermediate Period pottery assigned to Manachaqui's Empedrada Phase. Future investigations at Cuelap may permit sub-division of the Cancharin Phase. Looking west from Manachaqui to the Central Andean coast, the Puerto Moorin and Gallinazo Periods in the Viru valley should be roughly coeval with the Colpar Phase (Strong and Evans 1952; Fogel 1992). Huacapongo Polished 434 Plain Shapes 3/4 and 5 represent jars with reinforced rims somewhat analogous to the Colpar variety. Shape 4 persists into the subsequent Gallinazo Period (Strong and Evans 1952: Figs. 37, 40). Thus, coeval north coast assemblages do share reinforced rim forms with the Colpar Style. However, the associated vessels are very large and ovoid in shape with mouth diameters ranging from 20 to 40 em during Puerto Moorin, and 15 to 40 em during Gallinazo. Unlike Colpar, the Virli reinforced rims only occur on a limited range of vessel shapes during each period. Like Puerto Moorin pottery, Salinar Phase ceramics from the site of Cerro Arena in the Moche Valley are characterized by the white-on-red technique. The inventory of vessel shapes illustrated by Brennan (1978: Fig. B-1) includes neckless ollas and jars with both short and long necks. While the short-necked variety might be considered analogous to the Colpar jars, their everted rims are unthickened. In sum, most of the features that typify Early Intermediate Period coastal assemblages are absent in Manachaqui's Colpar Phase Paste A assemblage. These include the white-on-red technique, negative resist decoration typical of Gallinazo, strap handles, neckless jars and jars with long, flaring rims. Thus, the overall degree of similarity between the Colpar and coastal assemblages is negligible. In the highlands separating Manachaqui Cave and the 435 coast, the precise definition and dating of early sites and ceramic assemblages in Huamachuco, Santiago de Chuco and the upper Chicama valley remains problematic due to the dearth of excavated and radiocarbon-dated stratigraphic sequences. The recent Huamachuco investigations by T. Topic and J. Topic (1987:23) substantiate Thatcher's (1979) Sausagocha Phase which opens the Early Intermediate Period. The corresponding ceramic assemblage is dominated by neckless ollas and open bowls (T. Topic and J. Topic 1987:23). Thatcher (1979: Figs. 33-35) illustrates jars with long, flaring rims. Colpar-like short-necked jars appear to be uncommon or altogether missing in Huamachuco at this time. Short-neck jar rims from McCown's (1944: Fig. 20q, r) excavations at the Sausagocha Phase settlement of Cerro Ca~pana East, and from Thatcher's Mamorco Phase collections (1972-74: Fig. 5oq) are heavily reinforced and more massive than Colpar Phase examples. Only during the Sausagocha Phase does rim notching become important in Huamachuco. The Colpar style notched applique decoration seems to be entirely absent. Red slip is common to both Colpar and Sausagocha assemblages, yet neither exhibit incised decoration nor white-on-red decoration. Nevertheless, it would seem that the only significant parallel between the two assemblages is found in the co-occurrence of beveled-rim bowls. Rims from Cerro Campana East parallel Shape C Rim 12 bowls (cf. McCown 436 1945: Fig. 20qq), Shape E Rim 13 (Ibid.: Fig. 2lf, t) and Shape E Rim 15 (Ibid.: Fig. 21d). In each case the similarity is precise and, therefore, significant despite the many differences between the Sausagocha and Colpar assemblages. In the neighboring highlands of the upper Chicama valley, Krzanowski (1986: Table 11.1, 1986: Table 1) characterizes the Totorapamba Phase as a local variant of the Sausagocha Phase, and the final phase of the Early Horizon Pelon Tradition. The Totorapamba pottery has not yet been fully reported, but Krzanowski (1986:245) remarks that it "is almost identical to that of the Sausagocha Phase" as Thatcher defined it. Farther south in the upper Santa drainage, the Quinu Period at Pashash (Grieder 1978) has not been directly dated, but it is probably coeval with the later portion of the Colpar Phase. Vessel shapes include the neckless olla and some necked jars devoid of similarity to Colpar Phase shapes. An isolated, yet arresting parallel is again the presence of beveled-rim bowls including two forms which resemble Colpar Shape E Rim 13 (Ibid.: Fig. 341) and Shape C Rim 14 (Ibid.: Fig. 34m). These are the same rim forms reported from the eastern montane forest sites, and Huamachuco collections just described. Beveled-rim bowls resembling Colpar Shape E Rim 13 appear a short distance south of Pashash at La Pampa. 437 Terada (1979:179) proposes a sub-division of the undated Tornapampa Phase into early and late components. The beveled-rim bowls (Ibid.: Pl. 98-6, 14; 99-6, 7, 9; 101-5) belong to the earlier component which also features whiteon-red decoration recalling the Huaras style at Chavin de Huantar. Bowls and cups with beveled rims date to the late Initial Period Urabarriu Phase at Chavin de Huantar. None of these, not even the later Janabarriu Phase open bowls and cups, exhibit rim profiles like Shape E Rim 13. However, Janabarriu Jar 10 rim profiles (Burger 1984b: Fig. 293-295) are identical to Suitacocha Shape B Rim llc, and Colpar Shape B Rims llh and llj necked jar rims. Some other Janabarriu bowl rims with beveled lips somewhat resemble Colpar Shape C Rim 12. None of the Chavin decorative features appear in the Colpar assemblage. The Colpar Phase's closest overall stylistic affiliations with Central Andean assemblages found to the west are once again in Cajamarca. At a glance these similarities are not readily apparent as Layzon Phase ceramics are best known for distinctive red-on-white (as opposed to the more widespread white-on-red) painted decorations that appear abruptly, yet remain absent in Colpar Paste A pottery. Also unlike the Colpar Style, Layzon retains notched rims and incised decoration from earlier phases. In fact, the only decorative techniques 438 shared by the two styles are the application of red slip to vessel rims and the limited use of notched applique (Terada and Onuki 1985:92). Despite the contrasting decorative traditions, a comparison of vessel shapes reveals remarkably close correspondences. Layz6n neckless ollas occur in greatly diminished numbers, and the shape inventory now features globular jars with long, flaring rims, short-necked jars with short rims, wide-mouthed jars and open, semispherical bowls (Ibid.). There are clear morphological parallels between Colpar short-necked jars and Layz6n wide-mouth jars. In particular, Colpar Shape B Rim 11h jars find counterparts in Layz6n Red-on-White and Layz6n Red Slipped Form 4 jars (Ibid.: Pl. 67-5 and Pl. 68-17). Colpar Shape BRim 11i mirrors Layz6n Red-on-White Form 4 jars (Ibid.: Pl. 67-9), while Shape B Rim 11j jars resemble Layz6n Brushed Form 2 wide-mouth jars (Ibid.: Pl. 70-3). Also indicative of stylistic relationships are the Huacaloma Black Polished Form 10, Huacalorna Red Polished Form 7 and Huacaloma Brown Polished Form 3 open bowl variants with reinforced lips (Terada and Onuki 1985: Pl. 59-15, 59-28, 60-5) identical to Colpar Shape E Rim 13. According to Terada and Onuki (Ibid.: Tables 9-11, 28-30, 53-55, etc.), these shapes occur during the Late Huacaloma Phase and persist through the EL Phase. The EL Phase is coeval with the Chavin horizon (Onuki 1993: Fig. 15). The 439 common Layz6n convex bowl shapes resemble Colpar Shape C Rim 15 and Shape E Rim 9, but again, these shapes are not particularly diagnostic. Pottery of the following Initial Cajamarca Phase is said to retain much of Layz6n's stylistic character with the addition of fine, slipped and painted kaolin ware bowls. However, the Colpar-like short-necked and wide-mouthed jars drop out of the sequence. Initial Cajamarca non-kaolin ceramics feature neckless ollas and globular jars with either "swollen" rims (Terada and Onuki 1982:122-123, Pl.9411-14) or long, flaring rims (Ibid.: Pl. 95-1-9). These two jar shapes persist into subsequent phases of what Matsumoto (1993) terms the "Cajamarca Tradition." It is Layz6n that lays a foundation for this tradition characterized by a distinctive set of pottery attributes that persist throughout Cajamarca's Early Intermediate Period. The only Cajamarca Tradition non-kaolin shape with a Colpar analogue is the open or slightly restricted bowl with beveled rim (Terada and Onuki 1982: Pl. 91-1, 2). Also in Cajamarca and showing some remarkably similar features with Colpar is the Pandanche Phase CII ceramic assemblage from the highlands farther north. In contrast to Layz6n, earlier incised-line techniques disappear after Phase CI and the Chavin horizon (Kaulicke 1975:47-48). It is in Pandanche Phase CI that bowls with beveled rims like Colpar Shape E Rim 13 and others previously cited occur 440 (Ibid.: Lam. XIX: top row, fourth and fifth from left). These do not persist into Pandanche Phase CII, and they are conspicuously missing from the Layz6n assemblages described above. Wide incisions and red lines rendered in pre-firing paint decorate Phase CII vessels, most of which are reportedly globular jars with vertical necks. Like Colpar's however, Phase CII rims are reinforced on the lip exterior, and occasionally beveled. Some short-necked jar rims have clear Colpar Phase counterparts (e.g. Ibid.: Lam. XXI top row third from left, bottom row third from left). neither mentions nor illustrates neckless ollas. Kaulicke That he does not differentiate between jars and bowls in his descriptions may indicate the same ambiguities present at Manachaqui Cave and Gran Pajaten. Phase CII is the final phase in Kaulicke's published sequence. Finally, the Colpar Phase Shape E Rim 13 beveled-rim bowl also finds an analogue at the western montane forest site of Poro-Poro (Alva 1988a: Figs. 38-17). On the eastern slopes of Huanuco farther south, the Kotosh Sajara-patac Phase pottery displays few resemblances to Colpar Phase ceramics. Sajara-patac vessel shapes include neckless ollas and open bowls. vertical or flaring rims. Jars have short Izumi and Terada (1972: Pl. 102) illustrate a Sajara-patac jar strikingly similar to Suitacocha Phase Shape F jars with incised decorations laid 441 out in panels on the shoulders. However, this shape has already come and gone at Manachaqui Cave. Sajara-patac "short-necked jars" exhibit longer necks than their Colpar counterparts, and they lack the reinforced lip. Sajara-patac broad-line incision, punctation and applique techniques are absent in the Colpar assemblage, while Kotosh potters employed no painted decoration other than an occasional red slip. The continuing tendency to bevel vessel rims remains one of the few pottery attributes common to both styles. In fact, beveled-rim bowls are not only ubiquitous during the Kotosh Sajara-patac Phase (e.g. Ibid.: Pl. 100-5, 105-16-24), but they have a long history that may pre-date the Chavin horizon (e.g. Ibid.: Pl. 118- 11-17). The following Kotosh Higueras Phase at Kotosh shows even less similarity to Colpar pottery than Kotosh Sajara Patac. Onuki (1993: Fig. 15) dates the beginning of the phase at A.D. 1, and the argument that the phase ushers in new, different and perhaps "foreign" cultural patterns at Kotosh is well-known (Izumi and Terada 1972:205, 311; Lathrap 1970:173; Isbell 1974). The beginning of the Kotosh Higueras Phase signals the introduction of new pastes, vessel shapes and decorative techniques. Globular jars with long flaring rims (Izumi and Terada 1972: Pl. 98, 99-1-5), strap handles, notched applique and white-on-red and negative resist decorations co-occur. Of these attributes, 442 only applique characterizes Manachaqui's Colpar Paste A assemblage. To recapitulate, the greatest degree of overall similarity is found between the Colpar Phase assemblage and the earliest ceramics from nearby Gran Pajaten. Colpar's relationship with Layzon is difficult to evaluate given that the dissimilarities between their respective repertoires of decorative techniques are as striking as the similarities between their respective inventories of vessel shapes. It is more difficult to lend substantial weight to the parallels in jar rim morphology shared with Janabarriu Phase Chavin de Huantar. On the other hand, they are difficult to dismiss given the absence of parallels between Colpar Shape B rims and jar rims at any other Central Andean site south of Cajamarca. During the Colpar Phase another indication of linkage between Manachaqui Cave and neighboring highland groups to the west is the presence of the distinctive beveled-rim bowls. The appearance of these bowls within north highland sequences serves as a marker indicating cross-dates, or temporal overlap, between the pre-Abiseo Phase assemblage, Cancharin Phase, Sausagocha Phase, Quinu Period, Tornapampa Phase, Early Cajamarca Phase, Pandanche Phase CI and Kotosh Sajara-patac Phase as each is defined by its respective author. Again it is also difficult to evaluate the significance of these parallels, since this rim profile cross-cuts otherwise disparate stylistic traditions. 443 Paste A Relationships: the Amazonian Lowlands In the Amazonian lowlands to the east, Barrancoid pottery has been found over a wide geographic area extending from the lower Orinoco in the north to the Central Ucayali in the south. Lathrap dates the arrival of populations bearing the Barrancoid Style to 200 b.c. on the Central Ucayali {1970:117}. Lathrap's Barrancoid Hupa-iya Phase assemblage includes wide-mouthed vessels with flaring, unthickened collars {1962: Figs. 84, 85}. dominate the vessel shape inventory. Open bowls Neither shapes, nor the typical incised and modelled Barrancoid decorations suggest a relationship to Manachaqui's Colpar Style. Lathrap reports that, by a.d. 100 the Central Ucayali Barrancoid populations populations were replaced by producers of the Yarinacocha Style. The predominantly open, neckless vessels of the subsequent Yarinacocha Phase likewise show no relationship to Colpar pottery. Morales {1992} excavated Barrancoid pottery in northeastern Peru's Chambira River valley but his Siamba Phase assemblage has not yet been fully described or dated. Ravines' {1978} brief field campaign in the Central Huallaga valley directly east of Manachaqui Cave did not uncover any traces of a Barrancoid presence. His stratigraphic sequence includes two related ceramic complexes that overlay the Shakimu remains, but still lack absolute dates. The Huayabamba Complexes 1 and 2 include 444 short-necked "ollas" with everted, reinforced rims (e.g. Ibid.: Lam. 4b). However, a cross-date with the later Empedrada Phase to be described in Chapter 11 appears more likely. At the foot of the Andes to the south, globular, necked jars of the Nazaratequi phase (Allen 1968: Fig. 16, Forms 12 and 13) cited in the previous chapter may be contemporaneous with the Colpar Phase and suggest continued ties to the west and /or north. In the Amazonian lowlands north of Manachaqui Cave, Shady's (1987b) Bagua sequence concludes with the El Salado Phase that she dates to the end of the Early Horizon. She reports that El Salado vessels frequently bear exterior reinforced rims painted red and occasionally exhibiting an incised groove beneath the lip (Ibid.:476). In profile, the two rims illustrated in the upper right corner of Fig. 10 are not unlike Colpar Phase Shape BRim 15 examples. Shady (1987b:474) notes that such rims are also popular in the forested Tabaconas valley of Jaen where they appear to have been reinforced by the fold-over technique (e.g. Miasta 1979: Lams. 9a, 12a, 24c, 31b, 45f, SOb). Some of the Tabaconas rims (e.g. Ibid.: Lams. 28b, 57a, 61d) resemble those Colpar restricted bowl Shape C Rim 11. Miasta's undated assemblages also contain vessel shapes comparable to earlier Suitacocha Phase Shape F jars with vertical necks (e.g. Ibid.: Lams. 39a, 49b, 59e). The wide diversity of decorative techniques utilized in 445 the Bagua-Jaen region and elsewhere across the Amazon Basin might support arguments against relationships to the seemingly austere Colpar Style. Farther north in the Ecuador Oriente, none of the styles potentially coeval with the Colpar Phase (e.g. Sangay, Pastaza, Kamihun, Tivacundo) show similar decorative or formal attributes. More detailed information regarding vessel shapes is not yet available for the ceramic complexes of the Bagua-Jaen region, and the lack of absolute dates likewise hinders efforts to relate them to Colpar remains. However, comparisons utilizing available descriptions and illustrations indicate that Colpar Phase ceramics maintained a stylistic relationship with styles at the base of the eastern Andes directly to the northeast that is primarily reflected in distinctive rim treatments. Paste A Relationships: the Northern Andes In the lowland regions of extreme northern Peru, archaeological assemblages coeval with the Colpar Phase continue to display Northern Andean features. In the Piura River valley, assemblages of the coastal Sechura Phases A-D and mid-valley Chapica, Tamarindo A and Tamarindo B Phases are now dominated by jars with flaring rims, and vertical rims like the earlier Suitacocha Shape F examples (Lanning 1963: Fig. 23a-c, f, g, t; Ravines 1986-87: Lam. 5; Guffroy 1989: Fig. 7e, f; Kaulicke 1991: Figs. 14, 15). exhibit angled joints at the base of the neck. Many Lanning 446 (1963:209-210) regards white-on-red and negative resist decorative techniques in Sechura A through D as introductions from southern Ecuador. These same techniques characterize assemblages from the Tumbes area where Izumi and Terada (1966:81, 88) conclude that the closest stylistic affiliations of the Garbanzal style are found in the Ecuadorian Jambeli and Guangala styles. Precise similarities between the Colpar Phase Paste A assemblage and the Piura and Tumbes collections are few. In Piura, notched and plain applique strips can be observed on body sherds until the end of Sechura B (Lanning 1963: Fig. 22:i-m; also Guffroy 1992:110). Perhaps most significant is the presence of bowls with beveled rims during Sechura Phases C and D (Lanning 1963: Figs. 11i, 12a-b), and perhaps as early as Sechura A (Ravines 1986-87: Lam. 3-16), identical to those described from the Peruvian north highlands and montane forest earlier in this chapter. They also occur during the mid-valley Tamarindo Phase B (Kaulicke 1991: Fig. 15 lower right) which Kaulicke likens to Sechura PhaseD (Ibid.:417). From Loja in the southern Ecuadorian interior, the reinforced rims on necked vessels (Lecoq 1987) bear a general resemblance to those of the Colpar Phase, but the necks tend to be longer and Loja decorative techniques continue to emphasize incision. Lecoq points to Late Cerro Narrio and Jambeli as Loja's closest stylistic affiliates. 447 The Late Cerro Narrio assemblage from the southern Ecuadorian highlands is likely coeval with the Colpar Phase. It includes necked jars, bowls and compoteras. Most of the jar shapes feature unthickened short and long everted rims (Collier and Murra 1943: Fig. 10 bottom row, Fig. 14). One of the fifteen illustrated jar profiles closely matches Colpar Shape BRim 11i or 11j third from left). (Ibid.: Fig. 10 bottom row, Otherwise, the Late Cerro Narrio assemblage shows little resemblance to the Colpar pottery. Guangala and Jambeli are North Andean archaeological cultures assigned to the Regional Developmental Period in Meggers' (1966) schematization of Ecuadorian prehistory. She perceives these assemblages as outgrowths of a Chorrera stylistic heritage. Necked jars, convex bowls, compoteras and vessels with multiple leg supports typify many collections (Ibid.: Fig. 18). White-on-red and negative resist decorative techniques are characteristic of Guangala (Bushnell 1951) and Jambeli (Estrada et al. 1964) pottery. The only close morphological analogies to Colpar shapes might be found in the Early Guangala beveled-rim bowls' likeness to Colpar Shape E Rim 13 (Bischof 1982: Fig. 6d; Stothert 1993: Fig. 55e). These bowls with beveled lips and grooved exteriors are especially numerous in the southwestern Ecuadorian lowlands where early Engoroy (Chorrera) contexts indicate temporal priority (cf. Bischof 1982: Fig. 2a, f, 3b, d). 448 Of all of the reported northern assemblages reviewed for this study, Porras' (1975a) Cosanga collections from the montane forest east of modern-day Quito bear the strongest overall resemblances to the Colpar Phase assemblage. Porras (Ibid.:189) dates the Cosanga "Phase" to the centuries between 500 b.c. and a.d. 800. Although there are no indications of Cosanga-type compoteras at Manachaqui, the rim profiles illustrated for Shapes 3, 5 and 9 (Ibid.: Figs. 19 and 20) closely parallel Colpar Shape C Rims 12, 15 and 11 respectively. Although the Colpar Shape B rims find no precise matches among the Cosanga jar rims, the fold-over technique of rim reinforcement is a particularly diagnostic shared element. This technique is so distinctive that, coupled with the bowl rim resemblances, a stylistic relationship between the Colpar and Cosanga assemblages is clearly indicated. Paste B Origins The suspected sources of Colpar Phase Paste Group B ceramics are not so far-flung as Suitacocha Phase sources, but when considered as a group, they suggest emerging patterns of interaction that foreshadow the intensive interaction of the subsequent Empedrada Phase. Again, confirmation of suspected sources awaits substantiation through additional analyses. 449 Paste B6 Open bowls with modeled rims are found throughout the north-Central Andean highlands, in far northern coastal Peru and in southern Ecuador during this time period. McCown {1945: 21g) illustrates a rim form similar to Paste B 6 's {Fig. 93a) among a collection of modeled rims from Cerro Campana East, but offers no specifics regarding its paste characteristics. During his surface survey, Thatcher recovered several sherds of a distinctive pottery from Cerro Campana East and other sites with Mamorco and Sausagocha Phase remains: The paste cross section is a brick-red color with inclusions that look like gravel, including a noticeably large percentage of mica. The inclusions range up to more than 2 mm. in size and make up 40-50% of the rather soft paste. The surface is the same color as the cross section and the inclusions are quite noticeable. The surface is grainy and matte. No decoration is present (Thatcher 1979:92-93). That these sherds are intrusive at Thatcher's sites is implied by their scarcity. Paste B6 's place of origin remains unknown, but a Peruvian north highland source seems likely. Paste B7 Paste B7 sherds {especially Fig. 93f) unite features diagnostic of a classic Chorrera or Guangala bowl decorated with iridescent paint. The profiles of these southern Ecuadorian beveled rims are comparable to Colpar Shape E Rim 13 and, by extension, to the northern examples cited within the above discussion of Colpar Paste A relations. A group 450 of four non-local sherds of a fine, black micaceous pottery from Gran Pajaten's pre-Abiseo Phase assemblage also consist of Paste B7 (Church 1988:216-217). A single rim sherd represents a simple, convex bowl with unthickened rim and rounded lip (Ibid.: Fig. 51b). An Ecuadorian source is suggested by the presence of iridescent paint and the practice of smudging, both made popular during Chorrera. Bowls with beveled rims and iridescent paint persist during Guangala (Bischof 1982; Stothert 1993). Vessels with micaceous paste at Early Guangala Phase sites have been considered intrusive (Paulsen 1970:72), or fabricated with imported micaceous tempering material (Masucci 1992). It is difficult, therefore, to offer a more precise source location than "southwestern Ecuador." Paste B8 The Paste B8 sherds include the only example of whiteon-red decoration at Manachaqui Cave (Fig. 93g), although nothing truly identical could be located among other published collections of white-on-red style pottery. The motif of thin, roughly parallel lines looping below the rims of simple convex bowls is common in Cajamarca Layz6n Phase collections However, (Terada and Onuki 1985: Pl. 67-11, 67-12). in Layz6n the motif is executed in red paint on a white slip. From Huamachuco, Thatcher (1979: Fig. 38) illustrates a Sausagocha Phase bowl with the same motif 451 rendered in red paint on orange slip. The carinated bowl shape (Fig. 94a) also occurs in Huamachuco (McCown 1945: Fig. 19e). These Paste B8 sherds probably all originated in the neighboring highlands. Paste B9 The literature search failed to turn up a match for the rims with red painted geometric designs representing Paste B9 (Fig. 94c-e). Similar motifs rendered in white diagonal and crossing painted lines and dots appear on red Garbanzal bowl rims (Izumi and Terada 1966: Pl. 13a-15 and 16) and jar rims (Ibid.: Pl. 13b-8) from the far north coast of Peru. Perhaps these sherds originated in extreme northern Peru. Paste B10 Both McCown (1945: Fig. 17h, j) and Thatcher (1979: Fig. 42) illustrate rims with profiles resembling the distinctive Paste B10 rim (Fig. 94f), but in their Sausagocha Phase assemblages the decoration is executed with red paint on a light orange paste. The Paste B10 sherd probably represents pottery brought from neighboring highlands. Paste C1 Origins The Colpar Phase Paste C, rim sherd (Fig. 94g) is identical in both form and decoration to Layzon Red-on-White pottery from Cajamarca (cf. Terada and Onuki 1982: Pl. 8912, 13 and 14). However, neither the Layzon Phase wares, 452 nor similar red-on-white pottery at Cerro Arena (Brennan 1978:609; Mujica 1984:12), are fashioned in pure kaolin paste although exotic kaolin sherds do occur at Cerro Arena (Brennan 1978:599). Because of their isolated occurrence on the coast at Cerro Arena, Mujica argues that the red-onwhite style ceramics are imports from Cajarnarca. Despite the difference in clays employed for Layz6n Red-on-white and the C1 sherds, it still appears likely that the Colpar Phase examples were brought from Cajarnarca. Ceramic Evidence for Colpar Phase Chronology At present, the Colpar Phase is envisioned as the span of centuries between 400 b.c. and a.d. 200 during which Manachaqui Cave was only sporadically used. Manachaqui Cave's artifact assemblages lack attributes suggestive of participation in the Chavin cult, and the isolated match between Colpar and Janabarriu jar rims appears to be weak evidence of stylistic relationship. However, this popular rim form "with no antecedent for its short everted neck" is either "innovative" (Burger 1984b: 127), or intrusive, perhaps as a form borrowed from eastern montane forest pottery styles like Suitacocha or Colpar. Note that this co-occurrence does not necessarily imply a Chavin horizon cross-date for the Colpar Phase. On the other hand, a cross-date suggested by Colpar Shape E Rim 12's EL Phase counterpart in Cajamarca does suggest Manachaqui Cave's use 453 during the Chavin horizon. Shape E Rim 13 beveled-rim bowls also appear in the Central Andes during the Chavin horizon, but because the shape persists into the Early Intermediate Period, this evidence for Manachaqui Cave's use during the Chavin horizon remains less conclusive. Colpar Phase ceramic evidence, especially the close formal correspondences to post-Chavin horizon Cajamarca assemblages, most clearly indicates a beginning date of 200 b.c. It appears likely that the shelter was also infrequently utilized during the Chavin horizon, hypothetically advancing the beginning of the Colpar Phase to 400 b.c. Estimating the end of the Colpar Phase with ceramic evidence is also problematic, mostly because the nature and timing of style changes in the neighboring highlands during the Early Intermediate Period remain poorly understood. Because a paucity of fine kaolin ware pottery seems to characterize this phase, it is unlikely that it extends beyond a.d. 200. Radiocarbon dates for the subsequent Empedrada Phase range from a.d. 450 to 570. Whether or not the Colpar and Empedrada Phases are separated by an occupational hiatus will be considered in the next chapter. CHAPTER 9 THE EMPEDRADA PHASE Cultural occupations at Manachaqui Cave during the midEarly Intermediate Period are assigned to the Empedrada Phase. As described in Chapter 5, the phase is distinguished in the ceramic macrochronology by the abundant remains of fine, painted bowls fashioned from kaolin-rich clays. Again, levels from the deep Sector B berm deposits comprise the most intact stratigraphic contexts with which to isolate the Empedrada Phase materials. Representative and relatively unmixed samples come from Levels 4 through 7 in Units 27 and 28; and Levels 5 through 8 in Units 30, 31 and 32. Empedrada Phase deposits occur virtually at the surface near the mouth of the shelter (e.g. Units 20 and 21} where the 1986 "cleaning" activities removed the uppermost soil layers. In Sector A, numerous compact, superimposed floors rich in organic contents (Table 1, Stratum 2A} attest to the intensive utilization of Manachaqui Cave at this time. These floors contain the best-preserved samples of ceramics and organic remains, and charcoal from Sector A's hearths rendered all of the Empedrada Phase radiocarbon dates. Future radiocarbon analyses will target the kaolin ware454 455 bearing layers in Sector B in order to correlate the stratigraphic sequences from the two sectors, thereby confirming or modifying the absolute dates proposed for this and other Manachaqui Cave phases. An estimated total of 5,977 artifacts constitute the Ernpedrada Phase collection. The inventory of non-ceramic remains is more diverse than those of previous phases at the shelter. The large sample of lithic remains now includes an assortment of tool types. The full range of Manachaqui's faunal and botanical remains are found in Ernpedrada Phase strata. Empedrada Phase Ceramics A total of 1227 diagnostic sherds represents the Ernpedrada Phase. The transition from the Colpar to the Ernpedrada Phase marks the most substantial stylistic shift in Manachaqui's ceramic sequence. A nearly two thousand year-old pottery tradition characterized by short-necked jars and restricted bowl shapes ends abruptly, and is replaced by new vessel shapes such as globular jars with long, flaring rims and compound jars with bowl-shaped rims. The restricted bowl shape drops out entirely, and bowl rims rise to nearly 50 percent of the rim assemblage. The Ernpedrada Phase ceramic assemblage shows fewer shape categories than any previous phase assemblage at Manachaqui. Empedrada Phase jars with flaring rims are morphologically, and perhaps functionally, analogous to the 456 Suitacocha Shape F jar shape. Shape F vanishes from the Manachaqui Cave shape inventory by the beginning of the Colpar Phase, but its disappearance might be attributed to modifications in the shelter's functions over time. It should be kept in mind that the Colpar Phase sample is small, and Shape F constituted only a fraction of the jar shapes during the earlier Suitacocha Phase. In other words, Empedrada Phase jars might be functional and stylistic outgrowths of Shape F jars. The ancient tradition of fold- over rim reinforcement persists in variants of the Empedrada Phase jars. Also enduring virtually unchanged are decorative techniques featuring only red slip or paint, and notched applique strips. Again, severe weathering of the ceramic sample, especially in Sector B, frustrated aspects of the analysis. The assemblages of Paste Group B and Paste Group C pottery expand to unprecedented size and diversity during the Empedrada Phase. In sherd frequencies the Paste Group C now eclipses Paste Group B. The identification and chronological placement of these sherds met with the same problems previously described for the eroded assemblages of the previous phases. The most intact sherd surfaces are found in Sector A, while the largest samples and most intact stratigraphy are found in Sector B. The weathered condition of the Sector B sample frustrates endeavors to subdivide the Empedrada Phase. 457 Ceramic Paste Group A The Empedrada Phase Paste Group A assemblage consists of 359 diagnostic sherds, representing 29.3 percent of the phase diagnostic sherd collection. During this phase, Paste Group A encompasses an unprecedented diversity of pastes or wares. Again, the basic ceramic technology changes little. Microprobe analysis of one Empedrada Phase rim sherd enabled identification of angular grains of quartz, alkali feldspar, hornblende and ilmenite. It would seem that, although the selection and processing of clays and non-plastic additives changes from phase to phase, the criteria for raw material selection remains the same. The impression gleaned from handling the entire Manachaqui Cave ceramic collection is that the number of distinct pastes that constitute Paste Group A increases with each phase. The Empedrada Phase represents a burst of paste diversity second only to the extraordinary profusion contained in the Late Horizon Poblano Phase Paste A assemblage. Morphology A total of 342 sherds rendered information with which to reconstruct Empedrada Phase Paste Group A pottery shapes, 338 of which are rims. The shapes of bowls are typically semi-spherical and jar bodies are globular. There is still no evidence for vessel bases that are not rounded. 458 Shape E: Unrestricted convex bowls A total of 163 Shape E rim sherds represent 48.2 percent of the Empedrada Phase Paste A rim collection (Table 16). Most are unthickened or slightly thickened. None of the bowl rims are reinforced in the manner of the previous phases, and beveled varieties now constitute a small minority. Rim 3 and Rim 9 persist from the Colpar Phase (Figs. 95a-c and 95d-k respectively). Together the Empedrada Phase bowls range in diameter from 6 to 23 em (N=79), although one bowl diameter reaches 30 em. The diameters average 14.22 em with a standard deviation of 3.68. R~ 15 (Fig. 951-s). The 26 examples of Rim 15 belong to semi-spherical bowls with slightly thickened, unthickened and tapering rims that end in rounded and semi-flattened lips. The form is identical to Rim 9 except that it is slightly in-curving. Walls range from 4 to 7 mm in thickness. R~ 16 (Fig. 95t-w). The 14 examples of Rim 16 belong to semi-spherical bowls with unthickened rims and flattened lips. Wall thickness ranges from 4 to 5 mm. R~ 17 (Fig. 96a-d). The 10 examples of Rim 17 pertain to carinated bowls with vertical or nearly vertical rims and beveled, in-sloping lips. The walls are 4 to 5 mm thick. 459 R~ 18 (Fig. 96e-h). Nine examples of Rim 18 represent bowls with diverging walls and unthickened, direct rims. The lips are rounded. The vessel walls range from 4 to 8 mm thick. R~ 19 {Fig. 96i-k). The six examples of Rim 19 belong to relatively shallow open bowls with unthickened rims flattened at the lip. R~ Wall thickness ranges from 3 to 6 mm. 20 {Fig. 961-n). Four Rim 20 sherds are from open bowls with unthickened, sharply inverted rims. The lips are rounded and wall thickness varies from 3.5 to 4.5 mm. R~ 21 {Fig. 96o). The four Rim 21 examples all probably belong to the same large bowl with a 30 em diameter. The rims are vertical and unthickened with flattened lips. R~ The maximum wall thickness is 8 mm. 22 {Fig. 96p, q). Three Rim 22 examples represent relatively shallow bowls with divergent walls and thickened, bulbous rims. Wall thickness is 3 rnm, but maximum thickness at the rounded lips reaches 7mm. R~ 23 {Fig. 96r). Two eroded rims may belong to the same bowl of unknown diameter. They are thickened to 4 mm on the exterior surface and near the lip which ends in a dull point. R~ 24 {Fig. 96s). These two eroded rims may also 460 belong to a single bowl with unknown diameter. The walls reach 10 mm in thickness and the unthickened rim ends at a rounded lip. Shape G: Jars with Long, Flaring Rims A total of 152 Shape G rims (45 percent of the Empedrada Phase rim collection) pertain to globular jars with short, constricted necks and long, flaring rims. The jar rims are separated into 11 categories based upon morphological characteristics (Table 17). Many are thickened, and some even exhibit the folded over, reinforced lip so characteristic of Paste Group A pottery of previous phases. Overall, the vessel orifices vary from 7 to 24 em in diameter, and show great size variability. Collectively they (N=69) average 15.32 em in diameter with a standard deviation of 3.80 em Rim 1 (Fig. 97a-m). The 56 examples of Rim 1 belong to jars with long, unthickened rims. Thickness varies from 4 to 9 mm. The lips are either rounded or form a dull point. Some of the rims bear a flanged lobe (Fig. 97i, Rim 2 (Fig. 98a-h). j). These 43 examples of Rim 2 represent jars with long, flaring rims that have been reinforced by folding over the end. The union of the folded end and the rim exterior is typically left uneven and unsmoothed. Apparently the length of the rim from neck to 461 lip is highly variable. Rim thickness below the folded portion varies from 3 to 7 mm, and the reinforced end ranges from 7 to 12 mm thick. Rim 3 (Fig. 98i-k). The 21 Rim 3 examples belong to jars with gradually thickened, bulbous r1ms. invariably rounded. The lips are Below the thickened portion, the rim measures 3 to 4 mm in thickness, and the thickened end measures 7 to 9 mm. Rim 4 (Fig. 981-n). Twelve sherds constitute Rim 4 which shows a thickened, rounded interior surface and a nearly flat underside. The lip ends in a dull point. Maximum thickness reaches 8 mm. Rim 5 (Fig. 98o). Eight Rim 5 sherds pertain to jars with thick, direct rims. The illustrated rim tapers from 14 mm thick at its sharply angled neck to 6 mm at its semisquared lip. Rim 6 (Fig. 98p, q). Four rims represent one or more jars with unthickened rims exhibiting slightly concave interior surfaces. The lip ends in a dull point. Thickness ranges from 4 to 5 mm. Rim 7 (Fig. 98r). The three Rim 7 sherds represent one or more jars with unthickened rims. The necks are sharply angled and the lip has been beveled flat on top leaving a 462 dull point at the exterior edge. Rim 8 (Fig. 98s, t). Two rims, apparently from the same jar are nearly identical to Rim 7, except that the end has been reinforced and "squared" by beveling. The rims are 4 mm thick and the reinforced ends measure 6 mm thick. Rim 10 (Fig. 98u). One Rim 10 sherd represents a jar with an unthickened, direct rim. The lip is rounded and rim thickness is 8 mm. Rim 11 (Fig. 98v). One Rim 11 sherd belongs to a jar with a straight neck approximately 2.5 em high and an everted, unthickened, direct rim 1.5 em long. The neck and rim are both 5 mm thick. Shape H: Jars with Bowl-shaped Rims A total of 22 Shape H rims represent 6.5 percent of the Empedrada Phase rim collection. The Shape H category was established to distinguish these compound vessels with bowlshaped rims (also commonly referred to as S-shaped) , although there is no evidence that body shapes differed from those of Shape G jars. Four rim form categories represent variability in rim profiles and sizes. Mouth diameters range from 10 to 16 em, although one is 25 em. Only nine rim diameters were measured, and the mode and mean are both approximately 14 em. 463 R~ 1 (Fig. 99a-d). Nineteen sherds represent Rim 1. The carefully modeled S-shaped profiles display vertical or slightly inverted upper walls and everted ends. The lip of one rim (Fig. 99d) is slightly thickened or "beaded." The others are generally unthickened, ranging from 3 to 7 rnm in total thickness. Six measured orifices exhibit diameters from 10 to 16 em. R~ 2 (Fig. 99e). The single Rim 2 sherd is identical to Rim 1, but a mouth diameter of 25 em indicates a very large vessel. R~ The rim is 7 mm thick. 3 (Fig. 99f). bowl-shaped profile. rounded lip. One Rim 3 sherd shows a simple, The rim is unthickened and ends in a The mouth diameter is 11 em, and the rim walls are 5 mm thick. R~ 4 (Fig. 99g). Rim 4 is represented by a single bowl-shaped rim that has been slightly thickened to 4 rnm approximately 1 ern below the lip. The orifice measures 10 em in diameter. Miscellaneous Shapes Two of five sherds are believed to be broken spoon handles (Fig. 100a, b). Fig. 100a is 3.5 ern long, The handle fragment illustrated in 1.1 ern thick and tipped with a zoomorphic head featuring incised eyes, mouth and nostrils. It probably represents a llama. The other handle fragment 464 (Fig. lOOb) is 4.5 em long, 1.5 em at its thickest point, and has a long, conical shape. Three articulating sherds represent an unknown object (Fig. lOOc). It has a concave, scoop-like shape. Decoration Only 23 sherds supply information on Empedrada Phase decorative techniques. Again, we can be certain that painted decoration is under-represented due to the weathered state of the collection. Eight of nine body sherds with red paint or traces of paint on their interior surfaces apparently belong to bowls. The illustrated example (Fig. 96h) shows that potters still preferred to apply the paint to limited areas of the rim. One Shape E Rim 16 bowl (Fig. 95t) exhibits red paint on the interior and white paint on the exterior surface. A single sherd bears traces of white paint on the exterior surface. Fourteen sherds from Empedrada Phase levels display plastic decorative techniques. Nine bear notched applique strips identical to those of the Colpar Phase (Fig. lOOd, e). The surface of the sherd depicted in Fig. 100e shows shallow incisions where the potter scratched the vessel surface while "nicking" a discontinuous applique fillet, probably with long, downward strokes. Again, the applique strips seem to have been affixed to jar mid-sections. Shape GRim 1 jar rims carry notched lobes (Fig. 97i, Two j). Three Shape G Rim 1 jar rims feature rows of small round 465 punctations 1 to 2 rom in diameter on their interior surfaces (Fig. 97k-m) . Ceramic Paste Group B The Paste Group B category consists of 153 sherds representing 12.5 percent of the Empedrada Phase collection of diagnostic sherd collection. paste sub-groups. They are divided into eight Eighty percent of the Paste Group B sherd collection belongs to two sub-groups of particularly micaceous pastes. The remaining 20 percent consists of sub- groups with between one and eight sherds. Eleven (2.7%) of the total 410 Empedrada Phase rims belong to Paste Group B. Paste B 11 Ninety-one sherds representing Paste B11 may all belong to a single, large necked jar. Firing left the paste incompletely oxidized as indicated by the dark grey core sandwiched by light brown interior and exterior surfaces. Sand was added as temper and the abundant mica could be a natural constituent of the clay. The thickened, flaring rim of the jar is nearly 7.5 em long, 11 rom thick near the lip, and shows an orifice diameter of 30 em (Fig. 100f). constricted neck is sharply angled. exhibit traces of red paint. Its Unweathered sherds The sherd interiors show striations left by a bristled tool utilized to finish the surface. 466 Paste Bu Twenty-three heavily eroded sherds may have all belonged to a single vessel, perhaps an open bowl with everted rim (Fig. 100g). The paste exterior is 5YR5/6 yellowish red, and a dark grey core shows incomplete oxidation. Mica and quartz are principle constituents of the non-plastic inclusions that comprise more than 50 percent ·of the paste. The great quantity of mica renders the sherds very brittle, but when intact it must have lent this vessel a spectacular glittery appearance. The vessel has a tapered, flaring rim, and the orifice measures 15 em in diameter. The lip is rounded. Paste Bu Eight sherds come from a bowl fashioned from a fine paste with well-sorted particles of mica, pyrite and igneous rocks. It was fully oxidized to an orange color (5YR5/6 yellowish red) . The lightly thickened rim corresponds to a simple open bowl with an annular base (Fig. lOla-c). A geometric motif now partially eroded was rendered in prefiring red paint (10R4/6 yellowish red) on the bowl exterior. Paste Bu Eight sherds comprised of a distinctive coarse, sandy paste belong to two bowls with red (10R4/4 weak red) and brown (7.5YR5/6 strong brown) painted lines. The red paint 467 approaches purple in hue. Also distinctive is the exterior surface which displays wide tracks left by the burnishing tool. One large rim exhibits brown lines criss-crossed by short, red strokes on the interior surface (Fig. lOld). brown paint also covers the lip exterior. The Two sherds from a second bowl (Fig. lOle, f) show dots and alternating brown and red lines on the interior surface. Paste B 15 Four sherds grouped as Paste B15 also pertain to a bowl (Fig. 102a, b). The paste is coarse and sandy with a light brown (10YR7/4 very pale brown) color. smooth, but undulating and unburnished. The surface 1s A reconstructed rim indicates that the bowl has a carinated shape with slightly inverted walls and unthickened rim ending in a dull point at the lip exterior. It apparently rested on an annular base. The rim measures 20 em in diameter and the base, 13 em. The bowl interior shows a brown (10R3/6 dark red) painted design comprised of wide uneven lines. The exterior was slipped with a pale orange (5YR6/6 reddish yellow) wash, and then painted with two parallel lines, each approximately 3 mm wide. Paste B 16 Three sherds belong to a bowl with fine, sandy, dark brown (5YR4/2 reddish gray) paste and evidence of three drilled repair holes (Fig. 102c). Two of the sherds are 468 lightly thickened rims with flattened lips, indicating a simple semi-spherical shape. diameter. They measure 14 em in On one rim sherd a painted decoration rendered in iridescent specular hematite paint is barely visible on the exterior surface. The design beneath the lip consists of a stepped motif underlined by two parallel horizontal lines. Below these, the motif is probably a reclining "scroll-S" with another horizontal line beneath it. The second rim shows only traces of paint. Paste B 17 Two body sherds from jars with similar decorative treatments (Fig. 102d, e) may represent two distinct wares. The pastes are very fine with fine sandy inclusions. One sherd is fully oxidized to a light tan color (Fig. 102d), while the other shows a gray core indicating incomplete oxidation (Fig. 102e). thickness. They range from 3 to 4 rnm in The exterior surfaces of both sherds are covered with a dark brown slip (7.5YR4/4 brown and 5YR4/4 reddish brown respectively), wide orange painted lines (2.5YR5/8 red and 5YR6/6 reddish yellow) and a thick white paint. Paste B 18 A single rim sherd from a bowl consists of a hard, relatively coarse pinkish gray (10YR7/4 very pale brown) paste with well-sorted igneous rock temper (Fig. 102f). The walls are 5 mm thick, while the rim is gradually thickened 469 to 7 mm ending at a rounded lip. The polished bowl exterior displays an orange slip (2.5YR5/8 red), and the potter employed the negative resist technique to produce a pattern of circles surrounded by smudged black (perhaps triangular) fields. Ceramic Paste Group C The Empedrada Paste C category consists of 715 sherds representing 58.3 percent of the Empedrada Phase collection of diagnostics. Sixty-one (14.9%) of the 410 Empedrada Phase rim sherds belong to Paste Group C. If combined with Group B's 11 rims, the total 72 rims comprise 17.6 percent of the Empedrada Phase rim collection. Empedrada Phase Paste Group C characteristics are essentially the same as those noted for the Colpar Phase. Because of their unusually high diversity and poor preservation, wares (or individual pastes) represented within Ceramic Paste Group C cannot be consistently distinguished from one another with the unaided eye. Where the fired clay remains unoxidized, all kaolin-rich pastes may appear dark grey. The fully oxidized pastes show a continuum of colors and hues ranging from snow white to cream to beige to bright orange, depending upon the impurities in the clay. Therefore, the entire collection of 715 Empedrada Paste Group C pottery was lumped into Paste Group C2 of necessity. 470 Paste C2 a-h Empedrada Group C2 pastes are usually fine-grained, compact and hard, without visible non-plastic inclusions (Plate XII) . Finely sorted sand was added to some of these pastes as temper. The pottery was often slipped, polished and painted before firing to produce decorated thin-walled bowls, many with annular bases. The average rim diameter is 14.68 em (N=59) with a standard deviation of 2.69 em and mode of 14 em (N=29). Wall thickness ranges from 2 to 6 mm. Distinctions within the Paste C2 category are made according to salient technological, formal and/or decorative attributes, or combinations of attributes. Some of these most basic categories (indicated by lower case letters, e.g. Paste C2 a, C2 b, C2 c ... etc.) may contain more than one ware. It should be kept in mind that the sole purpose behind this categorical scheme is to distinguish groups of sherds by easily recognizable features, and comparative analysis. thus facilitating description Analysis of the fine wares recovered from excavations at Gran Pajaten was handled in the same manner (Church 1988, 1994). A total of 553 Paste C2 sherds show no signs of painted decoration, although some off-white sherds carry white slip or traces of white slip. An unknown number of these are highly eroded sherds from painted vessels. From this collection, 14 rims are illustrated (Fig. 103a-m, p), along with two sherds from annular bases (Fig. 103n, o). These 471 portray a representative variety of characteristic vessel shapes and rim treatments. A total of 170 Paste C2 sherds are distinguished by clearly recognizable (usually decorative) attributes or clusters of attributes and these are described in the following paragraphs. Paste C2 a (Red Painted) A total of 124 Paste C2a sherds show decoration, or traces of decoration, with red paint. with a white slip prior to painting. Some were covered Profiles of seven of the 12 rims with traces of red paint are presented in Fig. 103q-w. Figs. 103x, 104a-e and 105a-e show painted designs on rims sufficiently preserved for illustration. Fig. 106a- e shows five of a total 23 body sherds with painted red designs. Six sherds (including Fig. 106b-e) bear decorations on bowl interior surfaces. Most painted motifs consist of straight and curving lines tracing simple geometric patterns. Line width is highly variable. Twenty- three sherds are from vessels covered with an orange (5YR7/6 reddish yellow) or reddish orange (2.5YR5/8 red) slip. Both bowl rims (Fig. 104g-j, m, n) and annular bases (Fig. 106f, k, 1) are represented. Paste C"b (Black Painted) Twenty-two sherds grouped as Paste C2b exhibit dark brown or black painted designs. The paint color is typically 5YR3/2 and SYR 3/3 (dark reddish brown). Six 472 sherds consist of white (oxidized) and gray (unoxidized) kaolin paste. One rim with black paint applied to the lip area (Fig. 107a) represents a rare jar shape in Manachaqui's Paste Group C collections. Three rims (Fig. 107b-d) belong to bowls with simple linear painted designs. A body sherd shows a design painted on the exterior with both thin and wide lines (Fig. 107e). Sixteen of the Paste C2 b sherds consist of a fully oxidized orange paste. One rim exterior displays a painted row of pendant triangles above a set of two parallel, A sherd from an annular base horizontal lines (Fig. 108a). exhibits crossed lines on the bottom interior (Fig. 108b). The profiles of two eroded rims with indistinguishable designs are also illustrated (Fig. 108c, d). Paste C2 c (Red and Black Painted) Seven sherds exhibit exterior surfaces decorated with both red and black paint. Two rims bear simple red and black painted lines (Fig. 108e, f). Two sherds show that black lines may delineate red painted zones (Figs. 108g, 109a) . Others show black and red alternating with the natural paste color (Fig. 109b, c). Paste C"d (Tan on Buff) Five sherds with a tan colored paint (7.5YR6/6 reddish yellow) constitute Paste C2 d. Two similar bowls with in- curving rims may be represented (Fig. 109d-f). The motif 473 consists of simple geometric lines and a step pattern. The rims portrayed in Figs. 109d and f probably belong to the same bowl. Paste C2 e (Black on Orange Slip) Two sherds, both rims from separate bowls, have decorated exteriors showing black painted lines (2.5YR3/ very dark gray) on an orange slip (SYRS/6 yellowish red). One bowl rim (Fig. 109g) has parallel sets of divergent diagonal lines, while the other (Fig. 109h) exhibits parallel horizontal lines. Paste C2 f One body sherd with a gray core left incompletely oxidized during firing belongs to a decorated jar (Fig. llOa) . The sherd is 5 mm thick. Over the natural polished light brown surface, a brown slip (2.5YR3/6 dark red) was applied leaving an unslipped horizontal band 3 em wide. Black pigment appearing in the upper part of the band was probably produced by resist smudging. A row of concentric circles and circle-dots rendered in thick white paint repeat within the band which likely surrounded the jar's shoulder or mid-section. Smears of white paint obscure the motifs, especially the circle-dot pattern at right. Paste C2 g The sherd with a gray core depicted in Fig. llOb pertains to a restricted vessel with thin walls left 474 incompletely oxidized during firing. On its light brown, highly-polished exterior surface, a design rendered in red (2.5YR4/8 red} paint is barely visible. Paste C2 h One eroded sherd of grayish white kaolin paste has cracked and eroded surfaces (Fig. llOc}. It belongs to a jar neck and rim closely resembling Shape GRim 2. The end of the rim has been reinforced by folding it over. Maximum thickness is 8 mm, while the neck walls are 4 rnm thick. Lithic remains An estimated 50 percent of the Manachaqui Cave ceramic era lithic remains (approximately 4,750 artifacts} belongs to the Empedrada Phase. The 43.5 percent of the collection remaining after accounting for the four phases under study pertains to the Late Horizon Poblano Phase. Of the 4,750 Empedrada lithic artifacts, 4,706 (99%} are chipped-stone products of the same simple core-flake industry that characterized the Colpar phase. Thirty-four chipped-stone artifacts exhibit indications of retouch, leaving an estimated 4,672 non-retouched cores, flakes and blocky shatter. Of seven cores, six consist of mafic igneous rocks including basalt, and the seventh is made of gray chert from the outcrop crossing the ancient road to the east (Site M11). Four of the cores have battered surfaces. More frequent than in any of the previous phase 475 assemblages are retouched flakes. Twenty-nine exhibit somewhat haphazard unifacial retouch on one or two edges. These same "scraper" edges show tiny step fractures and crushing from heavy use. Fourteen consist of igneous rock, nine pieces are of the local gray chert and six consist of chert from unknown sources. Three obsidian flakes were recovered from Empedrada Phase contexts. Two of these are small pressure flakes picked out of the fine screens used to sort the Ernpedrada Phase floors (Floors I and K} in Sector A. A third is the proximal end of a thick, broken blade with a prominent dorsal spine and parallel edges showing evidence of heavy use. A piece of blocky obsidian shatter was collected from a mixed Empedrada and Poblano level (Unit 31 Level 5} . Two narrow biface fragments appear to be the basal portions of broken un-stemrned projectile points made from non-local chert (Plate XIII}. Both points were long, narrow (1.8 and 2.4 em} and relatively thick (7 and 11 rnrn} with rhomboidal cross-sections. The point from Unit 23 Level 8 exhibits a crudely-shaped tip, while the other from Unit 23 Level 10 shows a semi-squared base. In size and basic outline they are not unlike the ground slate points, and they may have served a similar function. The 44 ground stone artifacts from Ernpedrada Phase levels are pieces of ground and polished slate and shale. One oblong pebble (5.1 x 2.4 x 2.1 em} shows no signs of 476 use, but was certainly carried to the shelter from elsewhere. The ground and polished point fragments from Empedrada Phase levels number 24. Five fragments from mixed Empedrada and Poblano Phase levels are assigned to the Empedrada Phase because there are no such points in unmixed Poblano Phase levels. Thus, the Empedrada Phase ground and polished points total 29. An additional piece of slate with a convex, polished surface apparently split off a larger ground and polished slate artifact. An ovoid slate disk ground smooth, but otherwise unworked, x 0.3 ern. from Sector A Unit 16 Level 8* measures 2.7 x 2.1 Two similar ovoid disks from Unit 10 Levels 2 and 3 at the rear of the shelter may date to the Empedrada Phase or to the Poblano Phase. x 2.2 x 0.4 em. The disk from Level 3 measures 3.2 The disk from Level 2 measures 2 x 1.5 x 0.25 em, but a perforation 3 mm in diameter has been drilled into its center, and the wide end is notched or serrated {Fig. 108d). The presence of two blanks and 11 flakes of the same slate or shale provides evidence for the manufacture of pendants, projectile points and perhaps other items at the shelter during the Empedrada Phase. Rocks and Minerals A small quantity of rocks and minerals brought to the site during the Empedrada Phase include two small pieces of mica, a chunk of quartz combined with galena and one complete prismatic quartz crystal. 477 Botanical Remains Unit 15's Floors G through T (but excepting Floors LL and 0) provided samples of the Empedrada Phase carbonized botanical remains examined by Pearsall. Maize kernels and cob fragments continue in abundance, while beans occur relatively rarely. Fruit rinds, tuber/root fragments and all of the taxa identified in the earlier phase assemblages appear in varying quantities during the Empedrada Phase (Table 18). Faunal Remains With the Empedrada Phase, the variety of taxa, most of which is presumed to represent food remains, continues to expand (Table 27). Bird bones, some identified as owl, are more frequently represented than in previous phases, and reptiles appear for the first time. Finer taxonomic resolution reveals the presence of the llama, white-tailed deer, Andean fox, squirrel and guinea pig. A Homo sapiens premolar from Floor 0 in the shelter interior directly represents the human protagonists under study. Radiocarbon Evidence for Empedrada Phase Chronology Three radiocarbon dates from Sector A represent the midEarly Intermediate Period. Dates from Unit 14 Floors L and H are a.d. 570 ± 80 and 450 ± 80 respectively, while a date from Unit 15 Floor T is a.d. 490 ± 80. The Unit 15 date is associated with levels bearing the kaolin sherds diagnostic 478 of the Empedrada Phase in Manachaqui's macrochronology. However, as with the Colpar Phase, future radiocarbon dating should seek to correlate Sector B's stratigraphic chronology with Sector A's. At present, the radiocarbon dates do not provide clear resolution of the phase's beginning or ending dates. In the previous chapter it was suggested that a hiatus may separate the Colpar and Empedrada Phase utilization of Manachaqui Cave. Two undated occupation floors separate the radiocarbon dates of 160 b.c. and a.d. 490 in Unit 15, and three undated floors separate the dates of a.d. 110 and a.d. 450 in Unit 14. Additional radiocarbon dating may close this 340 year gap (only a 180 year gap if the sigmas of Floor P and Floor H dates are factored in) . On the other hand, evidence for a similar break exists in the radiocarbon sequence from construction fill at nearby Gran Pajaten where more than four centuries separate the uncalibrated dates of a.d. 40 and a.d. 460. At present, the radiocarbon evidence indicates a dating of a.d. 400 to 600 for the Empedrada Phase, but ceramic cross-dates providing additional chronological information will be discussed at the end of this chapter. Paste A Relationships: the Central Andes For the first time in the Manachaqui Cave sequence, Paste A pottery shows virtually all of the characteristic 479 features of Central Andean assemblages to the west and south. Large, globular jars with long, flaring rims and semi-spherical bowls constitute the two most basic vessel shapes in the Central Andes from the mid-Early Intermediate Period until the Spanish invasion. so generic and widespread, Because these shapes are they are poor indicators of specific relationships by themselves. However, a brief review of specific parallels between the Empedrada assemblage and some of the better-known assemblages from the neighboring highlands enables sufficient resolution for determining which of the Central Andean archaeological cultures show clearest stylistic affinities to Manachaqui Cave's Early Intermediate Period assemblages. Similarities between Empedrada Phase ceramics and Gran Pajaten's pre-Abiseo Phase assemblage are predictable given the ancient settlement's proximity to Manachaqui and its radiocarbon date of a.d. 460. Shape G Rims 1, 3, 4 and 6 are identical to jar rims encountered in Building No. 1's construction fill (Church 1988: Fig. 36h-m; 1994: Fig. 11f- k, Fig. 11b, c; 1988: Fig. 40g respectively). The simple bowls are likewise indistinguishable (Church 1994: Fig. 13hu), although some of the more restricted Gran Pajaten examples recall Colpar Phase bowl shapes. Red slip is common to both sites, but rim punctation and notched applique decorations have not been observed at Gran Pajaten. Gran Pajaten's pre-Abiseo Phase rim notching (Church 480 1988:168) is absent from the Empedrada assemblage. Empedrada Phase pottery also finds analogous counterparts in the Cancharin Phase assemblages from the vicinity of Cuelap. Profiles of Shape G Rims 1, 2 and 6 resemble those depicted by Ruiz {1972) in Lams. VIII:1-31, X:a-q and IX:c-g respectively. Bowl rim profiles are also similar, and carinated varieties occur at both sites. Cuelap's·bowls with everted rims and thickened interiors do not occur at Gran Pajaten nor at Manachaqui Cave. Directly west in Huamachuco, the dates A.D. 400 to 600 coincide with the Early Huamachuco Phase {T. Topic and J. Topic 1987: Fig. 13). Only the most generic similarities between the profiles of flaring jar rims suggest some relationship to Empedrada Phase ceramics. Unlike the Empedrada Phase, the Early Huamachuco Phase is characterized by the presence of fine wares with cursive-style decoration {Thatcher 1972-74: Figs. 23-36), both imported from Cajamarca to the north and imitated locally {T. Topic and J. Topic 1987:23). In Cajamarca, the production of cursive- style pottery marks the beginning of the Middle Cajamarca Phase {Matsumoto 1993:188). Matsumoto's {1993: Tabla 1) absolute chronology for Early Intermediate Period Cajamarca is based upon Shimada's {1987:140) suggestion of an a.d. 450 cross-date for Moche Phase IV and the Middle Cajamarca Period. The dates for both the Early Huamachuco and Middle Cajamarca assemblages are ostensibly supported by the 481 association of Early Huamachuco Phase cursive-style pottery with radiocarbon dates averaging a.d. 405 from the principal constructions at Marcahuamachuco (J. Topic and T. Topic 1982:21, 1983b: Table 3}. The points just outlined in the previous paragraph lead to the suppositions that either: 1} the Empedrada Phase pottery predates its associated radiocarbon dates as well as the Early Huamachuco and Middle Cajamarca Phases featuring cursive-style pottery or 2} Manachaqui Cave's interactions at this time were focused on Central Andean regions south of Huamachuco and outside of Cajamarca's interaction sphere. Evaluation of the first supposition requires an examination of earlier phase assemblages in Huamachuco and Cajamarca. These would be the Purpucala Phase in Huamachuco dated from A.D. 1 to 400 (T. Topic and J. Topic 1987: Fig. 13}, and the Early Cajamarca Phase dated from A.D. 200 to 450 (Matsumoto 1993: 188} . Although a few resemblances to Huamachuco's Purpucala Phase remains have been observed in the pre-Abiseo Phase assemblage at Gran Pajaten (Church 1994:288-289}, there is little in the Empedrada Phase assemblage that presents similar correspondences. Thatcher's Empedrada Shape GRim 6 matches (1972-74: Fig. 19a-h} Jar 8 rim, and red and white paints were also applied to Purpucala Phase coarse wares. In fact, Thatcher (1972-74:112} regards the decorative use of red and white paints as the only 482 remarkable similarity between Purpucala and contemporary Cajamarca styles. Unlike their Empedrada counterparts, Early Cajamarca Phase jar rims are often modelled, especially around the lip (Terada and Onuki 1982: Pls. 90, 92, 94-96). In terms of shared design features, the Empedrada assemblage bears little more than superficial resemblance to Purpucala and Early Cajamarca Phase ceramics. The closest stylistic match for the Empedrada Phase assemblage can be found south of Huamachuco in the upper Santa River drainage best known for Recuay Style pottery. Based on his finds at the site of Pashash, Grieder (1978) has published the only Recuay ceramic sequence. The centuries between a.d. 300 and 600 correspond to Grieder's Quimit, Yaia and Huacohu Phases of the Recuay Period (Ibid.:63, Table 10). Empedrada Shape G Rims 5 and 6 resemble Recuay jar shapes B-1 and B-4 (Ibid.: Fig. 35k, s) of the Quimit Phase, while Empedrada Shape G Rims 1, 2 and 4 belong to jars identical to Recuay shape B-5 (Ibid.: Figs. 38f, l; 36i) of the Huacohu and Yaia Phases. The distinctive Empedrada Shape H Rims 1, 2 and 3 find precise parallels in later Usu Period (a.d. 600-700) vessel shapes B-10 and B-8 (Ibid.: Figs. 40i, 38j, k). That these latter shapes only appear late in the Pashash sequence (Ibid. :62) supports Thatcher's (1979:99) argument that identical rim shapes at Cerro Campana East (McCown 1944: Fig. 18a-g) represent a post-Sausagocha Phase re-occupation of the site 483 (J. Topic and T. Topic 1982:20 present a contrary point of view). In the eastern Andes of Huanuco to the south of Manachaqui Cave, archaeological survey and excavations have not yet turned up mid-Early Intermediate Period occupations (Rozenberg 1982:135; Bonnier et al. 1985:96). Farther south however, Hastings (1985) has grouped two sets of mid-Early Intermediate Period assemblages on the Tarma Valley eastern slopes into the Malambo and Camonal Complexes. The former is distributed from the Tropical Alpine and Subalpine Life Zones down through the eastern Montane Forests (Hasting's "Lower Tier") and into the Premontane Forest (his "Montana Uplands"). The Camonal Complex is mostly restricted to the Montana Uplands and Bottomlands at the foot of the Andes (Ibid.: Table 11-11). However, Hastings (Ibid. :590) notes that shared attributes occasionally render the two complexes virtually indistinguishable. Both of Hastings' complexes share design features with the Empedrada assemblage. Empedrada's resemblances to the Malambo Complex are general, as jars with flaring rims, semi-spherical bowls and notched applique decorations are typical (Ibid.: Fig. 9-11, 9-13). Parallels with the Camonal Complex (Ibid.: Figs. 9-14, 9-15) are more specific as Empedrada Shape G Rims 1, 2 and 4 mirror the rims depicted in Hastings' Fig. 9-14t-v, z-ae and k-1 respectively. Jar rims of both Tarma Valley complexes 484 feature rims with exterior thickening. The "folded" rim, however, is almost exclusive to Carnonal (Ibid.:590). Distinctive bowl-shaped rims like Empedrada Shape H Rims 1 and 2 also occur in the Camonal Complex (Ibid.: Fig. 9-14hk). Carnonal's rare corrugation and frequent use of strap handles are not shared by the Empedrada assemblage. The regional comparisons offered in the above paragraphs demonstrate that the Empedrada Phase Paste A ceramic assemblage is closely affiliated with coeval Central Andean pottery styles, yet maintains specific design features in common with early assemblages from eastern slope localities such as Gran Pajaten, Cuelap and the Tarma River valley. The most compelling resemblances to Empedrada Phase Paste A pottery are found in the upper Santa valley Recuay region and, to a far lesser extent, in Huamachuco. Both Grieder (1978:75) and Thatcher (1972-74:112) have noted the similarities shared between their respective regions of study. The complete absence of Huamachuco and Cajamarca cursive-style pottery at Manachaqui Cave draws attention south to Recuay as the Empedrada assemblage's closest stylistic companion in the highland Central Andes. Paste A Relationships: the Amazonian Lowlands Below Manachaqui Cave to the east Ravines' (1978) Central Huallaga sequence presents close stylistic affinities to highland pottery traditions for the first 485 time. The Huayabamba I Complex recovered from three sites on the lower Huallabamba River below its confluence with the Abiseo and Jelache Rivers displays several parallels with the Empedrada assemblage. Ravines' (Ibid.: Lam. 3a, b) Forms A and B are jars with bowl-shaped rims conceptually similar to distinctive Empedrada Shape H Rim 1. The likeness may not be immediately apparent because the end of the Huayabamba vessel's rim (see Lam. 3a profile) has been folded over, and the rim's carination angle is sharper. A second formal correspondence is evident between Huayabarnba Form G necked jars (Ibid.: Lam. 4d, e) and Empedrada Shape G Rim 2 jars with folded-over rims. Note that these are the same formal resemblances that characterize the stylistic relationship between the Empedrada assemblage and Hastings' Camonal Complex. The Empedrada Phase pre-dates the appearance of the Cumancaya Tradition on the river floodplains east of the Central Andes. It is presumably coeval with Lathrap's (1962) Pacacocha Phase on the Central Ucayali, and perhaps with Allen's (1968) Naneini Phase from the foot of the eastern Andes. While these lowland assemblages carry design features which are unmistakably Amazonian, it may be significant that the jars with bowl-shaped rims like Empedrada Shape Hare found in both (Lathrap 1962: 113a, c, e; Allen 1968: Fig. 25: Form 4). Lathrap (1962:351) maintains that they may have early Amazonian antecedents in 486 Hupa-iya (Barrancoid) shapes. Allen (1968:357) observes that Naneini is the only Alto Pachitea assemblage that can be "linked" to lowland Amazonian pottery traditions to the east. However, Naneini retains a globular, necked jar shape from earlier phases (Allen 1968: Fig. 26: Form 6) suggesting continued ties to the Central Andes. Hastings (1985:590-591) regards the rim folding technique as a common thread connecting Malarnbo, Carnonal and Alto Pachitea styles, and the Ernpedrada and Cancharin (from Cuelap) assemblages should be added. Otherwise, coeval lowland Amazonian assemblages, and the Curnancaya assemblages that succeed them, are stylistically dissimilar to Ernpedrada and other Central Andean assemblages, especially with respect to decorative techniques and motifs. Paste A Relationships: the Northern Andes The Empedrada Phase at Manachaqui Cave corresponds to the centuries spanning the Regional Developmental PeriodIntegration Period transition in the neighboring Northern Andes (Meggers 1966). Some of the Ernpedrada features do have northern correlates, yet it is not necessary to range beyond the Central Andes to document Ernpedrada's closest resemblances and hypothetical affiliations. Nevertheless, a stylistic relationship apparently does exist with ceramics from the analogous Ecuadorian montane forest setting of 487 Cosanga (Porras 1975a). Empedrada jar Shape G Rims 1, 2, and 3 match illustrated rim profiles for Cosanga Shapes 7 and 10 and 11 (Ibid.: Figs. 20, 21), while Cosanga Shape 5 represents less diagnostic semi-spherical bowls much like Empedrada examples. As noted in the previous chapter, the co-occurrence of the distinctive fold-over rim reinforcement technique constitutes a commonality among styles found along the eastern slopes of the Central Andes. Based on her analysis of Cosanga/Panzaleo pottery from northern Ecuador, Bray (1995:146) contends that folded rims are the unique "hallmark" of this eastern slope style. Additional Cosanga/Panzaleo features occurring in the Empedrada assemblage are the parallel rows of punctations on rim interior surfaces (Porras 1975a: Lams. 29g-l, 30a-m). At Cosanga, the punctation occurs more typically on short rims of wide-mouthed vessels. Nevertheless, the technique has a limited distribution in the Northern Andes and has not been reported this far south. Bray (1995:140) considers this punctation a diagnostic feature of Cosanga/Panzaleo pottery during early centuries of the Integration Period. Paste B Origins Paste B,, Useful clues in determining hypothetical sources for the micaceous pottery of Paste B11 may be found in Porras' (1975a) report on Cosanga ceramics, and Bray's (1995) 488 summary of Cosanga/Panzaleo vessel morphology. All three of the Cosanga pastes are sandy and micaceous, and Bray (Ibid.:137) observes that "thin vessel walls, ash-colored paste, and micaceous inclusions" distinguish Panzaleo pottery found in archaeological assemblages throughout much of northern Ecuador in a variety of temporal contexts. The single Paste B11 jar rim profile (Fig. 100f) precisely mirrors Porras' (1975a: Fig. 21) Shape 10 profile, and Bray's (1995: Fig. 7g) Form III/7 profile. Porras (1975a:115) reports similar surface treatments, and observes that Shape 10 jars frequently exhibit red slip (Ibid.:127). The Manachaqui Paste B11 rim diameter of 30 ern exceeds the Shape 10 diameter range of 10 to 25 measured by Porras, but is smaller than the Cosanga maximum of 40 ern (Ibid. :107). Paste B12 This pottery is too eroded and fragmented to determine the shape of the corresponding vessel or vessels. The micaceous paste suggests some relationship to Paste B11 pottery from the eastern slopes of the Northern Andes. Little more can be inferred without a larger, better preserved sample. Paste B13 Again, a micaceous paste suggests a possible relationship with the Paste B11 and B1 :: pottery just described. Porras (1975a:ll3-118) writes that fully 489 oxidized Cosanga pastes often appear orange or yellowish red. The Paste B13 bowl shape (although admittedly generic) matches Porras' Shape 5 (Ibid.: Fig. 20). Its painted design (Fig. lOla, b) resembles the motif depicted in Porras' Lam. 36a, although the red-painted Cosanga design is rendered on white slip. A similar triangular scroll motif can be found closer to Manachaqui on Cajamarca Tradition kaolin bowls of the Cajamarca Linear-painted Type dating to the Early Cajamarca Phase (Terada and Onuki 1982: Pls. 44c5, 102-25; Matsumoto 1993). Taking paste characteristics into account, it is not unreasonable to suggest origins on the eastern slopes of the northern Peruvian or Ecuadorian Andes. Paste B14 The eight sherds representing Paste B14 pertain to two bowls that originated far to the south in the highland Central Andes between Lake Junin and the modern city of Ayacucho. The use of purplish red paint is a diagnostic characteristic of mid-Early Intermediate Period pottery reported from the Salinas de San Blas (Morales 1978:331) and the upper Tarma and Mantaro valleys (Hastings 1985:525-528; Browman 1970; Benavides 1971). The sherds illustrated by Morales (1978: Lams. 5, 6) show similar decorative panels painted on the rim interiors. The same red and black cross- hatch motif (Fig. lOld) appears on the bowl depicted in Morales' Lam. 6. 490 Ceramics from Pachamachay described by Silva (1988:2628 I Figs. 56-58) may also be related to Paste B14 pottery. Silva observes that highland populations in Huancayol Ayacucho 1 Junin~ Huancavelica and Huanuco shared a common decorative tradition yet maintained regional autonomy during the Early Intermediate Period. Huacrapuquio 1 Browman has termed the style and suggested dates of A.D. 500 to 600/650. Hastings (1985:592-595) reports that this eastern highland style penetrates the upper edges of the Tarma canyon montane forests as an imported ware (Ibid.: Fig. 9-12). The Junin sherds described by Morales and Hastings provide sufficiently close resemblances to conclude that the Paste B14 sherds likewise represent this south-central highland stylistic tradition. Paste B15 Sherds from decorated bowls like the Paste B15 example have been recovered from Cancharin Phase contexts at the site of Cuelap (Ruiz 1972). Ruiz's Cuelap Pintado Tosco Type (Ibid. :77-79) is likewise unburnished and frequently shows horizontal striations left after smoothing the surfaces. Horizontal red painted lines (Fig. 102a) may appear on bowl lips and exteriors (Ibid.: Lam. XXXVIIa-d) I while vertical lines and haphazard dots commonly decorate bowl interiors (Ibid.: Lam. XXXVIal c). The match between Paste B:s and Ruiz's written descriptions and illustrations of the Cancharin pottery is remarkably close. The "finas 491 capas roj izas" {fine, reddish layers} {Ibid. : 77} probably refer to the thin orange wash observed on the exterior surface of the Empedrada example. Ruiz does not mention annular bases, but his sample consists of only 43 sherds, all of which come from the "fortaleza" or main settlement of Cuelap. Paste B16 Decorated bowls such as these (Fig. 102c} have been recovered during the investigations at Cerro Ochoconday on the western slopes of the Andes below Huarnachuco {Alfredo Melly and Theresa Topic personal communications 1990}. The Cerro Ochoconday examples are better preserved, and show the same horizontal lines, step motifs and scroll-S motifs rendered in iridescent paint on the exterior surfaces of brown, polished bowls. The only published illustration of this ware might be Wilson's {1988} Fig. 212p. Paste B17 The two sherds with white and orange paint on polished brown slipped surfaces {Fig. 102d, e) fit Strong and Evans' {1952:344-347 and Fig. 80} descriptions of Castillo White, Red, Orange Type ceramics recovered from Gallinazo contexts in the lower Viru valley. Strong and Evans regard this pottery as intrusive on the coast, and Theresa Topic {personal communication 1987} has found larger amounts in the upper Viru drainage. Topic {personal communication 492 1991} has examined these two particular sherds and believes that they are not the same wares that she observed during the upper Virli surveys. In the highlands above Virli, Perez (1988: Lam. 4) illustrates similar pottery from Santiago de Chuco. Grieder's Pashash publication includes a photograph of Huacohu Phase (late Recuay Period a.d. 500 - 600} "white and red on orange" sherds which might be related to the Paste B'- 7 wares. Actually, the Paste B17 sherds match Strong and Evans' descriptions and illustrations so closely that it may not be imprudent to conclude that they pertain to styles akin to Castillo White, Red and Orange that originated in the north highland Recuay area, perhaps near Santiago de Chuco. Paste B18 The Paste B18 rim sherd with negative resist decoration (Fig. 102f) invites comparison to Gallinazo Negative, Decorative Sub-type VIII from Viru (Strong and Evans 1952:302 and Fig. 59L, M}. However, Gallinazo Negative utilizes natural or white-slipped surfaces as a base. The presence of an orange-red slip may indicate origins farther north. Ruiz (1972: Lam. XXXVc-g} assigns similar red- slipped and negative resist decorated sherds at Cuelap and/or Pumahuanchina to his Cuelap Polished pottery type and the Early Intermediate Period Cancharin Phase. The quantity of these negative decorated sherds is unspecified, but the 493 pottery may be exotic at Cuelap as well. Estrada et al. (1964: Figs. 30, 31) illustrate similar negative dot motifs on red slip for the Jambeli Culture of southwestern Ecuador. Cafiar Polished and Tuncahuan pottery from Cerro Narrio (Collier and Murra 1943: Pls. 26, 38-41) frequently displays similar dot motifs in negative, but the vessels are typically unslipped. Izumi and Terada (1966:84) urge caution in attaching great temporal or spatial significance to this widely distributed motif. Nevertheless, we might speculate that the Paste B18 example originated in extreme northern Peru. Paste C Origins Interaction spheres characterized by the circulation of fine kaolin wares in the north-Central Andes expand during the Early Intermediate Period. Cajamarca and Recuay are the best known archaeological cultures producing this pottery, but decorated kaolin ware pottery has also been recovered during excavations at nearby Gran Pajaten. The Gran Pajaten sample was judged to be intrusive, non-local pottery because of its unusual variety of pastes, decorative techniques and motifs, coupled with its corresponding lack of stylistic cohesiveness (Church 1994:288). The Empedrada Phase sample at Manachaqui Cave should be regarded as intrusive (relative to Paste A ceramics) for precisely the same reasons. It should be reiterated that the variety of Empedrada kaolin ware ceramics was subsumed for descriptive purposes within a 494 single category (Paste Group C) in response to the multiplicity of separate, yet visually indistinguishable, kaolin wares. Sequences from the Cajamarca and Recuay areas have been developed, although the Cajamarca region has received more extensive study. Chronology and intra-regional variability of the Recuay tradition remain to be fully investigated. This study is handicapped especially by the lack of fieldwork in the upper Marafion valley or Callejon de Conchucos to the southwest of Manachaqui Cave where much of the Empedrada Phase Paste C pottery may be originating. The hypothetical sources for Paste C described below do not include Cajamarca which, judging by the absence of cursivestyle pottery, does not seem to interact directly with the study area during the Empedrada Phase. Paste C2 Virtually all of the Empedrada Phase Paste C2 decorated kaolin wares can be subsumed within the highland Recuay pottery style as described by Bennett (1944:99-106}, Grieder (1978} and Wegner (1981}. While illustrations of the most elaborate Recuay pottery featuring negative-resist decoration have been widely published (e.g. Larco Hoyle 1962; Bankmann 1977; Grieder 1978}, the simpler positivepainted bowls are more often represented as sherds encountered during surface surveys performed outside of the upper Santa Valley core area (e.g. Strong and Evans 1952; 495 Thatcher 1972-74; Gambini 1984; Proulx 1985; Wilson 1988). Hence, it appears that the positive-painted variety was more widely distributed than the Recuay negative-resist variety. Unfortunately, the decorative variability of the positivepainted style remains one of Recuay's least-known aspects. Despite the problems cited above, some illuminating comparisons between Ernpedrada and Recuay positive-painted wares can be offered. The semi-spherical bowl shapes and simple rim profiles representative of the highland kaolin ware traditions are not particularly diagnostic. Decorative techniques, and especially motifs, constitute the most useful attributes for revealing stylistic relationships although some of these too were widely shared. The following comparative analysis focuses primarily on decorative attributes. Paste C2 a-h The Paste C2a motif rendered in red paint on white kaolin, and depicted in Figs. 103x and 104a, also occurs in Thatcher's Sausagocha Phase sample from Cerro Campana East which, as previously noted, may contain ceramics from multiple Early Intermediate Period occupations. Apparently depicting interlocking serpent heads, the motif also appears in Nepefia Valley contexts (Proulx 1985: Pl. 7A: PV 31-59p). An illustration of a sherd from an Early Intermediate Period context at Cerro Campana West, and remarkably similar to the Empedrada example, was shown to me by T. Topic (personal 496 communication 1994). The Nepefia example appears to be rendered in red and black on white kaolin, while the Cerro Campana West sherd exhibits black on buff-colored kaolin. A motif composed of a red band with zig-zag edge and thin parallel lines identical to Fig. 104b was unearthed in the Virli Valley (Strong and Evans 1952: Fig. 81c). Both thin and wide lines painted on kaolin pottery resembling Empedrada Paste C2 (Figs. 104c-e, 105a-e, 106a-e this thesis) are frequently illustrated in the Recuay-related publications cited here. The painted wavy line on the bowl depicted in Fig. 104e also appears at Chavin de Huantar (Bennett 1944: Fig. 31a), in the Santa Valley (Gambini 1984:138-139; Wilson 1988: Fig. 212b) and on an illustrated sherd from the Callej6n de Conchucos site of Cashajirca shown to me by Richard Burger (personal communication 1993). Decorative bands of pendant triangles above horizontal lines like that depicted in Fig. 108a occur at Gran Pajaten (Church 1994: Fig. 15a-c), at Cerro Campana East in Huamachuco (Thatcher 1979: Fig. 46) and again in the lower Santa Valley (Wilson 1988: Fig. 212i-k). Similar designs ornament kaolin bowls recovered from late occupations at Chavin de Huantar (Tello 1960: Fig. 174 upper left). Black horizontal stripes on red paint or slip (Fig. 109h) occur in Thatcher's (1972-74: Fig. 14, 15) Purpucala Phase collection. The Paste C2 f sherd (Fig. 110a) shows both the negative-resist and positive-painting techniques. From 497 Pashash, Grieder (1987:228, No. 6) illustrates a Recuay jar with negative resist decoration and an upper border executed in "cream circle dot on black." The Empedrada sherd probably originated in the Recuay core area. Many less specific correspondences between Empedrada Phase Paste C sherds and other north-Central Andean ceramics of the Early Intermediate Period could be listed here in an extended, but belabored, discussion. the Recuay area is clearly evident. Close interaction with Only one Empedrada Phase Paste C sherd identified, the single Paste C2 g jar rim sherd, was probably produced outside of the Recuay and Cajamarca core areas despite the utilization of kaolin-rich clay. The rim resembles Empedrada Shape B Rim 2 and may represent a vessel fabricated within the study area. The clay may have originated to the west, but because of the lack of accurate geological information from the Huallaga basin montane and premontane forests, the existence of kaolin deposits in the study area cannot be ruled out. Lithic Evidence Lithic materials from ceramic era sites in the Andes are often ignored as archaeological evidence. However, the Empedrada Phase chipped-stone lithic assemblage contains a few formal elements such as projectile points that can be productively compared to the few reported finds in neighboring areas. While similar points are found in the north highlands, none of these date to the mid-Early 498 Intermediate Period. Possible stylistic analogues have been recovered in Cajamarca (Terada and Onuki 1982: Pl. 54-6, 7}, but they date to the Layz6n and Initial Cajamarca Phases (200 b.c. to a.d. 200}. Similar "willow-leaf-shaped" points at La Pampa date to the Late Horizon, and possibly to the early Initial Period (Terada 1979: Pl. 81a}. Urabarriu Phase points from Chavin de Huantar (Burger 1984b: Figs. 384-387} also resemble the Empedrada Phase points. Whether or not projectile points appear in a particular phase assemblage at a particular archaeological site depends upon site functions and the kinds of activities that took place over time. The regional panorama of seemingly scattered appearances of similar projectile points during different time periods at different sites described in the literature leads one to suspect that these kinds of projectile points were produced continuously throughout most of the prehistoric ceramic era. Where and when they appear in site reports can be viewed as the result of a series of factors, sampling and reporting procedures utilized not being the least important. Unequivocal evidence for long-distance interaction at Manachaqui Cave has been obtained by trace element analysis of the Empedrada Phase obsidian artifacts. The three obsidian pieces from Empedrada and mixed Empedrada and Poblano Phase contexts were subjected to neutron-activation analysis by Burger et al. (1995}, and were found to consist 499 of stone from the Quispisisa source in Huancavelica. This identification compliments the evidence for linkage between Manachaqui Cave and interaction spheres far to the south represented by the Paste B14 Huacrapuquio-like sherds. Ceramic Evidence for Empedrada Phase Chronology In the previous chapter, it was observed that the Colpar Phase may not have extended beyond a.d. 200. Ceramic and stratigraphic evidence attest to a major style shift in the Manachaqui Cave sequence occurring at a.d. 200 or shortly thereafter. The cross-dating of pottery is crucial to further define the Empedrada Phase in the absence of more radiocarbon evidence. Unfortunately, the north highland ceramic sequences do not offer particularly high chronological resolution for the early centuries of the first millennium A.D. Nevertheless, there are comparative data with which to estimate a beginning date for the Empedrada Phase, and to determine whether or not a hiatus exists between a.d. 200 and 400. The chronological contexts in which the widely dispersed Central Andean kaolin wares occur all correspond to the Early Intermediate Period. The most solid evidence with which to temporally situate much of the Empedrada Paste C pottery comes from the Recuay site of Pashash and the Viru valley where non-local Recuay sherds are found in Gallinazo Period contexts (Strong and Evans 1952). The Gallinazo 500 occupation of the Virli Valley may have lasted until approximately a.d. 400 (Shimada and Maguifia 1994: Cuadro 1.2). Based upon his radiocarbon evidence from Pashash, Grieder (1978:63) posits a.d. 300 to 600 as dates for the Recuay style. However, the beginning of the Recuay Period at Pashash has not yet been directly dated, and Wegner's (1981:6) suggested a.d. 200 beginning date is more probable in view of the Gallinazo-Recuay temporal overlap documented by Strong and Evans in the Virli valley. Manachaqui's ceramic cross-dates do not provide evidence of a hiatus between the Colpar and Empedrada Phases. On the contrary, the presence of Pastes B17 , B18 and C2 a implies Gallinazo cross-dates. These, coupled with Recuay cross- dates and the radiocarbon evidence, indicate continuous occupation at Manachaqui Cave between a.d. 200 and 600. An eighth century cross-date to the Usu Period suggested by the co-occurance of Shape H Rims 1 and 2 at Manachaqui Cave, Pashash and perhaps Cerro Campana East extends the proposed ending dates for the Empedrada Phase to a.d. 700. A terminal Early Intermediate Period occupation during the Empedrada Phase is further supported by the presence of Paste B14 Huacrapuquio-like imported ceramics. Finally, absence of cursive, floral or Huari style pottery in Manachaqui's deposits provides negative (and therefore inconclusive) evidence that the Empedrada Phase ended by a.d. 700 and the Middle Horizon. the CHAPTER 10 MANACHAQUI CAVE AND MONTANE FOREST MIGRATIONS The previous four chapters presented evidence for four archaeological phases constructed primarily on the basis of stratigraphic correlations between ceramic attributes and radiocarbon dates. This chapter implements Rouse's (1986) three-pronged strategy for inferring migrations by surveying evidence supplied by historical linguistics and physical anthropology in conjunction with Manachaqui Cave's archaeological sequence. Because relatively little anthropological research has been conducted in this geographic area, none of these data sets is sufficient, and sufficiently detailed, to successfully argue for the postulated prehistoric migrations. The scant linguistic evidence is highly equivocal, while there is little evidence from physical anthropology with which to gain useful insights into prehistoric biological populations. As usual, archaeological evidence is the most abundant, albeit prone to the interpretive problems discussed in Chapter 1. Evaluation of the archaeological evidence for population movements and colonization entails the investigation of Manachaqui Cave's prehistoric functions, 501 502 and assessment of the wayside station working hypothesis. Once Manachaqui Cave's functional contexts are understood, then full attention can be devoted to critical examination of diagnostic indices of migration such as anomalous disconformities in assemblage structures or contents. Gaining an understanding of the rockshelter's functional contexts is likewise crucial to the formulation of alternative hypotheses. Re-interpretation of the "migration evidence" is the subject of the final chapter. Discussion of the evidence for and against migration from the rockshelter and from the study area will proceed phase by phase. Unfortunately, the utilization of diagnostic indices requiring high stratigraphic and temporal resolution described in Chapter 1 is infeasible given the nature of Manachaqui's palimpsest occupation floors. Discussion of the Manachaqui evidence will be followed by evaluation of migration hypotheses from other localities in the eastern montane forest. First, however, brief summaries of study area pre-Conquest language and human biological distributions will be presented. Evidence from Historical Linguistics Chapter 3 presented evidence that southern Chachapoyas societies and their Amazonian Cholon and Hivito trading partners in the Central Huallaga valley were brought under the control of the Inca. During successive periods of Inca and Spanish domination of the Marafion-Huallaga divide, the 503 Quechua language was imposed as lingua franca in order to facilitate administration and control of the local populations. This imposition, coupled with the catastrophic depopulation of both highland and lowland areas, ultimately left only mestizo, Spanish-speaking populations and a lack of scholarly consensus regarding the area's pre-Inca, and perhaps pre-Quechua, languages. The linguistic problem encountered here is complex because it is the language "substrate" that is sought rather than the last native language documented by European observers. In the studies described below, place names and personal names that often survive language replacement provide scholars with clues, but conclusive evidence has not been forthcoming. Tello's and Lathrap's hypothesis would be served by the identification of an Arawakan sub-stratum, and Isbell's hypothesis requires the substantiation of a two thousand year development of Quechua A in the study area. Verticality scenarios might be served by finding evidence of language archipelagos to match colony archipelago distributions, all of the migration hypotheses would be strengthened if pre-Inca Chachapoyas and Central Huallaga languages were identical. A survey of opinions for the Chachapoyas and Central Huallaga areas is recounted below. Chachapoyas Languages While evidence for pre-Inca Quechua dialects would clearly serve Isbell's and Lathrap's hypothesis of early 504 Quechua expansion from the northeastern Peruvian or Ecuadorian lowlands, some scholars have considered the "modern" Chachapoyas dialect of Quechua A (Torero's Quechua 2) only a late prehistoric veneer spread as consequence of Inca imperial expansion (Rowe 1946:185; Torero 1974; Bird et al. 1983-84). As noted in Chapter 2, Isbell challenges Torero's and Bird et al.'s hypothesized proto-Quechua coastal Central Andean homeland (Isbell 1974:151, 198384:245), but neither he nor Lathrap provide evidence for Quechua origins based upon analysis of linguistic data from Chachapoyas or from the Ecuadorian Oriente. The question remains: did Inca Quechua replace other Quechua dialects in Chachapoyas, or some other unknown non-Quechua language or languages as Bandelier (1907), Zevallos (1987) and others have suggested? Culle (Torero 1974) and the Cajamarcas division of Puruha-Mochicas (Jijon y Caamafio 1943) are extinct languages believed to have been spoken in the Huamachuco and Cajamarca northern Peruvian highlands prior to the expansion of southern Quechua. Little is known of Culle. Jijon y Caamafio (1943:463, Map III, No. 82) includes Chachapoyas within the domain of Cajamarcas, which belongs to his proposed Northern Andean Macro-Chibchan phylum. (1950:194, Map 18) limits Puruha-Mochicas' Mason (he calls it Yunca-Puruhan) range to the northern coast of Peru and southwestern Ecuador, and regards affiliations of 505 Chachapoyas languages as undetermined. Zevallos (1982) contends that Yunca toponyms were introduced into the Chachapoyas province by Yunca-speaking mitimae groups from the north-central coast relocated by the Inca. Based on an onomastic analysis of 645 names taken from 16th century Chachapoyas documents, he argues against their affiliation with any language group in the adjacent Andes. He suggests further inquiry into northern and eastern lowland, especially Carib, relationships (Zevallos 1982:4). In summary, there is no agreement whether an affiliate of Quechua, Yunca, Chibchan, Carib or some other language family (or families) was spoken among the ethnic groups ultimately subsumed within the Inca province of Chachapoyas. The adoption of a lingua franca (whether Inca Quechua or any other) to facilitate interaction probably masks a complex mosaic of indigenous languages in pre-Incaic Chachapoyas. Chol6n and Hivito Languages The extinct Chol6n and Hivito languages remain as mysterious as other aspects of their pre-Hispanic culture. Steward (1948:514, 598) reports that much Quechua spoken in the central Huallaga during the nineteenth century was a post-Spanish contact or late prehistoric introduction. The Amazonian partial assimilation of Andean Quechua supports suggestions of habitual highland-lowland interaction in the study area. Regarding pre-Quechua language distributions, 506 Steward characterizes upper Amazonian language distributions as extremely complex (Steward 1948:507). Mason's map of South American linguistic distributions shows both Hivito and Chol6n as language isolates (Mason 1950: Map 18). Chol6n has been more adequately documented (by de la Mata 1923; Tessmann 1930:546-547) than Hivito, but most scholars regard them as phylogenically related. As noted in Chapter 2, Tello (1942:629) regards both, with Arnuesha and Muchik (Yunca), as belonging to a larger family of Arawakan languages representing Central Andean civilization's demographic sub-stratum. Jij6n y Caarnafio (1943:465) groups "Cholona" along with Chachapoyas in his Macro-Chibchan language phylum, although Mason (1950:192) views Cholona's relation to Chibchan as "doubtful." Tessmann (1930:626) finds similarities between Chol6n and Quechua. The degree of Quechua infusion remains unknown, but it may be attributed to commercial and social modes of interaction dating to the Late Horizon, if not earlier. Evidence supporting a hypothesis of substantial borrowing from Quechua by the Cholones and Hivitos comes from the foot of the eastern Andes farther south at Cerro de la Sal where Wise (1976) observes that the Amuesha's assimilation of Quechua elements has hindered efforts to classify their indigenous language. Much of the lexicon borrowed from the Quechua highlanders is related to "social and commercial dealings" (Ibid.:358), including words for 507 buying and selling. The Inca are presumably responsible for many intrusive features of Amuesha language and mythology, and commerce sponsored by Colonial Period missionaries that employed Quechua trade jargon also exerted a lasting impact. Most importantly, Wise observes that lexical borrowing from Quechua has led to transformation, and even partial replacement of Amuesha Arawakan (Ibid.). Wise (1985:208) has recently listed Cholon and Hivito as both "unclassified" and "extinct." Other recent classifications group them (Cholon specifically) with other Andean languages (Landar 1977; Greenberg 1987; Kaufman 1990) including Quechua. According to Tessmann (1930:626), Hivito may have been mixed with Pano-Ge. Pefia Meza (1935:344) reports that during the 19th century, an isolated band of refugee Hivitos and Panoan-speaking Conibos from the Ucayali inhabited the forested slopes between the Huallabamba and Jelache Rivers (north of the Abiseo) . Consequent intermarriage may account for the mix later observed by Tessmann. The central Huallaga region's role as a refuge for natives fleeing Spanish slavers (Reeve 1994:123) and retribution for rebellious activities (de la Riva Herrera 1907) undoubtedly contributed to additional linguistic mixing since the sixteenth century. Hivito and Cholon apparently absorbed multiple linguistic influences and we may not attain a clearer picture of Central Huallaga indigenous languages unless more illuminating Colonial missionary documents turn 508 up. The evidence at hand both here and in southern Chachapoyas, however, cannot be easily utilized to support of refute any of the postulated migrations. Evidence from Physical Anthropology Needed for systematic studies of prehistoric population distributions are large skeletal samples from welldocumented geographic and temporal contexts. The semi-arid micro-climates in which Chachapoyas dead were often interred favor preservation of skeletal material, and such studies based upon field investigations are feasible. For example, a study of 154 skulls from five funerary sites in Chuquibarnba by Jakobsen et al. (1986-87) suggest the presence of two quantitatively identifiable populations. The local Late Intermediate Period population tends to be dolichocephalic and rather typical of Central Andean highland regions of nearby Cajarnarca, while skulls from the Late Horizon site of Salsipuedes may pertain to an intrusive population from the south, perhaps arriving as a result of Inca-directed population movements (Ibid. :154). Examination of stature produced evidence of a similar north-south dichotomy. Salsipuedes individuals were on average 4 ern shorter. Such data may be useful to evaluate the origins of Chachapoyas populations (Ibid. :154-155), especially since some comparative information is available from studies elsewhere on the continent (Stewart and Newman 1950). Truly 509 necessary, however, are human remains representing prehistoric populations in surrounding lowland and highland regions. Studies of modern populations in these areas may offer clues, but these have not been conducted either. Demographic shuffling under the Inca, followed by chaos and depopulation under the Spaniards, will have undoubtedly produced complex distributions and numerous interpretive problems. On a larger scale, paleodernographic relationships between the upper Amazon and Central Andes are impossible to examine while skeletal samples from the prehistoric humid lowlands are scant. Evidence from Archaeology: Manachaqui Cave and the Pataz-Abiseo Area In this section, the material evidence from Manachaqui Cave will be brought to bear on the population movement and colonization hypotheses for each phase. A point of departure for consideration of the postulated migrations will be provided by the assessment of the rockshelter's function. The reader will recall from Chapter 1 that some functional contexts are better suited than others to effectively perform the comparative analyses required to evaluate the postulated migrations. Specifically, domestic or quotidian contexts are more desirable than ceremonial contexts. Discussion of botanical, faunal and artifactual evidence for site function from each phase will be followed by detailed evaluation of pertinent migration hypotheses. 510 Prior to discussing the phase data however, some general observations can be offered regarding the spatial patterning of cultural remains throughout the Manachaqui sequence. Based on her analysis of the botanical remains from Sector A's sequence of floors, Pearsall (Appendix F) concludes that her study "generally support(s) the hypothesis that Manachaqui Cave was a camp for travelers for much of its history." As evidence she notes the low quantities of food remains relative to wood charcoal and the presence of "exotic" food items during the entire sequence. Problems with the interpretations of maize evidence were cited in Chapter 4. In the ratio histograms presented within Appendix F (Pearsall's Figs. 1-7), several floors show marked deviation from relatively even distributions, and these will be noted below. One of the most striking patterns discovered during the excavation of Manachaqui's Sector A, was the consistency maintained by the spatial distributions of features hearths), artifacts a~d surface down to bedrock. (e.g. organic remains from the site The ethnographic analogy and surface distributions appear to be effective tools for interpretation of subsurface prehistoric patterning and identification of activity areas. Virtually all of the hearths were prepared on the west side of the shelter to the extent permitted by the shelter's wall and bedrock protrusions. The two exceptions are visible in the south 511 profile of Unit 1 (Fig. 18) and the east profile of Unit 3 (Fig. 14). Evidence that east side surfaces were repeatedly scooped out to create a concave sleeping surface is seen in the dipping strata visible in Unit 1 and 2 profiles (Figs. 11 and 14). The lack of orderly sequences seen in the Unit 1, 2, 11 and 25 macrochronologies (Appendix D) attest to heavy disturbance of the east side's stratigraphic integrity. Kent's faunal analysis provides additional evidence for repetitive spatial patterning throughout Manachaqui Cave's prehistoric occupation. His tallies of specimens on the west side (Units 12, 13, 14, 15, 16, 17) and east side (Units 1, 2, 3, 11, 25) of Sector A reveal that east side units yielded only 74 specimens while west side units contained 2,284. Nearly a third of the west side faunal material (745 specimens or 32.6%) is classified as unidentified scrap, while east side excavations produced no scrap at all. Kent concludes that most of the bone deposition, activity and resultant trampling took place on the west side while the east side was kept relatively clean. Kent's conclusion ostensibly supports the inference that the shelter's occupants habitually cooked and ate around the west side hearths, while the east side consistently served for sleeping. However, ceramic and lithic remains are more concentrated on the east side, away from the hearths, where they average 79.6 sherds per 512 excavated 5 em level compared to the west side's 60.6. sherds. If the samples from Units 16, 17 and 3 at the shelter's mouth are omitted, the east side average rises to 85.8 sherds per level and the west side average drops to 38.3. Lithic remains show similar patterning with 12.7 artifacts per level on the east side compared to 3.9 on the west. These patterns contrast with artifact distributions documented at the Initial Period rockshelter herding camp at Telarmachay where lithic debris and soot-caked potsherds clustered together with faunal remains around the hearths (Lavallee 1977:87). While cleaning activities conceivably account for the Manachaqui Cave distributions, the separate distributions of hearths and food refuse on the west side, and pottery and lithic refuse on the east side may also indicate that pottery was not necessarily utilized for cooking and stone tools were not utilized strictly for food processing. This co-variation is consonant with the interpretation that preprepared food was routinely brought into the shelter and consumed around the hearth where the shelter's occupants warmed themselves. In this scenario, personal pots and stone tools might have been kept close to their owners. Because of stratigraphic mixing in Sector A's east side, the spatial distributions of the artifacts cannot be broken down by phase, but other kinds of archaeological patterning will be examined phase by phase below. 513 Pre-Lavasen occupations and the Lavasen Phase Pre-Lavasen and Lavasen Phase occupations at Manachaqui Cave await future opportunities for more thorough analysis. Some general observations are offered here because there are preliminary indications of major changes in the rockshelter's functions prior to the Initial Period Manachaqui Phase. The reader will recall from Chapter 5 that there are no apparent pre-Lavasen strata in Sector A, and Sector B's pre-Lavasen Phase strata yielded a variety of functional stone tool types along with two radiocarbon dates of 2120 and 2330 b.c. Associated projectile points, scrapers and burins indicate that activities at Manachaqui included hunting and the processing of animal products. The importance of hunting and butchering at Manachaqui Cave is difficult to evaluate given the poor preservation of the pre-Lavasen faunal sample. Additional evidence for hunting activities is provided by the statistical distribution of animal sizes represented among the pre-Lavasen remains (Fig. 111). Although poor preservation may bias the pre-Lavasen distribution to some extent, a prevalence of medium to large-sized animals (e.g. pacas, pudus and cervids) is documented in Kent's analysis. Despite the occasionally small sample sizes, the changing distributions of faunal remains from pre-Lavasen times to the Early Horizon Suitacocha Phase clearly show a 514 diminishing utilization of large animals and a corresponding increase through time in the utilization or consumption of small animals (Table 28, Figs. 111-116), perhaps as preprepared travel food like that described by Young and Morales (see Chapter 4). Similar emphasis on relatively large game is evident at Early and Middle Preceramic Northern Andean sites like Chobshi Cave in Ecuador (Lynch and Pollock 1981: Cuadro 1, p.98) and Nemocon and Sueva Rockshelters in Colombia (Correal 1979: Cuadros 1, 4) where large and medium sized animals (white-tailed deer, pudu and paca) were most frequently utilized. Faunal assemblages from Preceramic Period Central Andean puna rockshelters consist of primarily cervids and camelids (e.g. Rick 1980: Table 10.1, p.234; Wheeler 1985: Tableau 2). and manes II and II II Hundreds of ground-s tone slabs small mammals, rodents and birds II dominate lithic and faunal assemblages from Middle Preceramic house remains clustering just below the Upper Zafia Valley western montane forests (Dillehay et al. 1989:749-750), but these were apparently left by semi-sedentary horticulturalists rather than by mobile hunters and/or herders. Stone grinding tools are rare or absent in Manachaqui Cave strata, and the few pre-Lavasen Phase macrobotanical samples recovered from Sector B have not yet been studied. Nearly all of the animals represented in Manachaqui's faunal collection for all phases are native to paramo and forest 515 edge habitats. During the Lavasen Phase we find the first evidence for conditioning Sector A, presumably for habitual and/or prolonged use. The catch-basin and drainage canal (Feature R-5) unearthed in Unit 12 would likely have been destroyed by runoff entering from the rear of the shelter interior (where Units 9 and 10 are located) had it not been prepared in combination with some sort of sod or stone wall. It remains unknown whether Feature R-5 represents the first attempt to condition Sector A, or if it is simply the first to be preserved by site formation processes. Lavasen Phase artifacts in both sectors include stone tools produced by a simple core-flake industry, but analyses of raw materials and tool functions have not yet been attempted. Much of the selected stone has poor flaking qualities and was probably opportunistically gathered from the surrounding subalpine valleys. The paucity of faunal remains suggests that the tools served functions other than meat processing. It may be significant that small animals constitute a larger proportion of the faunal sample (Fig. 112) Pearsall's botanical evidence for Manachaqui Cave's functions appears in Appendix F as food:wood ratios with (Appendix F, Fig. 1) and without Festuca (Appendix F, Fig. 2), and local:exotic+local counting unidentified fruit rinds as local (Appendix F, Fig. 3), and excluding fruit rinds 516 altogether (Appendix F, Fig. 4). Charred Festuca seeds may have been left by burnt straw bedding. However, an estimated 344 of the 404 (85%) Festuca seeds from Floor CC were recovered from its rock-filled hearth. quantity of 3,132 The inordinate (estimated) Festuca seeds from Floor BB were all found concentrated within its associated simplebasin (but perhaps re-utilized, rock-filled) hearth. This patterning strongly suggests that these locally gathered seeds were being cooked as food. High food:wood ratios in Floors FF through BB (Appendix F, Fig. 1) support a habitation scenario rather than the wayside station working hypothesis. The local:exotic+local ratio (Appendix F, Fig. 3) also supports hypothetical habitation, but the appearance of maize drastically lowers the ratios above Floor z. The stone-filled hearths in Floors FF through CC suggest cooking, perhaps by stone-boiling in vessels fabricated with perishable materials, or roasting seeds and/or meat among the hot rocks. Neither possibility seems more likely than the other given: 1) the difficulty of stone-boiling at high altitude and 2) the scarcity of animal bones during this phase (N=21). The animal sizes represented in the small Lavasen faunal sample reflect utilization of the full spectrum of fauna characteristic of an ecotone setting. Animal sizes show a distribution similar to that of the larger pre-Lavasen assemblage, but the number of small animals increases. The only other 517 Andean rockshelter to show relatively high proportions of small animal specimens is Guitarrero Cave (2,580 m), but Wing (1980:152-153) reports that the profusion of unburned bones of opossums and "small rodents" is "incidental" and unrelated to subsistence. Needed for future reconstruction of Manachaqui Cave's changing behavioral contexts is an analysis of changing animal consumption patterns featuring anatomical breakdowns by species for each phase. If we also consider Rodbell and Hansen's palynological evidence for anthropogenic disturbances in the area after 4000 b.c., then it seems likely that the beginning of the Lavasen Phase at Manachaqui Cave was marked by the local introduction of crop cultivation, especially of puna tubers which are seldom preserved in archaeological botanical assemblages but might have been roasted in the rock-filled hearths. Wild Festuca seeds also contributed to the diet. Thus, evidence for the utilization of mostly Tropical Alpine Zone resources suggests that Manachaqui Cave likely served as a short-term habitation for individuals or small families engaged, at least seasonally, in seed gathering and tending crops. The botanical taxa listed are mostly the same as those utilized at the "hunting base camp" of Pachamachay (Pearsall 1980). Manachaqui may have served intermittently as a wayside station for travelers, possibly carrying preprepared meals that included guinea pigs and small game. At Manachaqui Cave there is no reason to suspect that 518 any changes at the onset of the Lavasen Phase should be attributed to migrations. The Late Preceramic Period throughout the Central Andes was marked by ongoing economic and social change (Quilter 1991). It should be observed, however, that Preceramic Period human occupation of the ceja de selva is not anticipated by many migration theories recounted in Chapter 3. Colonization scenarios often fail to address the problem of indigenous eastern montane forest populations. Lathrap et al. (1985) refer to the eastern montane forest prior to mid-Early Horizon migrations from the Ecuadorian Oriente as an "empty niche." The evidence just recounted suggests utilization of the Tropical Alpine Zone by groups fully acquainted with and adapted to the local environment. The Manachaqui Phase Occupation The Manachaqui Phase remains provide evidence for the intensification of the rockshelter's use, perhaps alternating as a temporary habitation and trailside waystation. Suggesting habitual use is the attention given to refurbishing the shelter with masonry plugs at the rear (Feature R-7), and walls (R-6 and R-8) mouth. flanking the cave The occupants of Floor AA would have had ample space to move about beneath a ceiling 1.60 m high. Although unequal in height, R-6 and R-8 may have functioned along with a simple roof of branches or cana and ichu to enclose an effective sheltered space extending a meter or two from 519 the cave mouth across the berm. The great fallen stone slab visible in Unit 26 and the northern edges of Units 18 through 21 (Fig. 28) provided a low wall partially enclosing the front of the shelter. In this "enclosure scenario" the leveling and maintenance of interior floors, combined with movement within the confined space, may be responsible for the turbated condition of Manachaqui Phase levels beneath Floor AA (Levels 43 to 50 in Units 15-17). Of the two Manachaqui Phase floors, Floor Z stands out as possessing two hearths, one of which is the largest unearthed at the rockshelter. Floor Z's Hearth 1 is also noteworthy because it was centrally situated at the mouth of the shelter. Here the shelter's dripline, coupled with partial exposure to the elements, would have likely disturbed or obliterated at least part of the hearth if not for protection by a roof. The partially scattered charcoal fill on the western side of Lavasen Phase Floor EE's hearth (Fig. 25b) was probably dispersed by exactly this kind of exposure to heavy rain and roofline dripping. Evidence from Botanical and Faunal Remains Floor Z's Hearth 1 is also the only Sector A hearth containing potsherds. Although the organic content of Floor Z's hearths has not yet been examined, the Floor Z sediments contained high food:wood ratios (Appendix F, Figs. 1 and 2) 520 and a high local:local+exotic food ratio if the problematic fruit remains are regarded as local (Appendix F, Fig. 3). Floor AA's two sets of ratios present conflicting evidence. High local:local+exotic food ratios contrast with very low food:wood ratios. Overall, the Manachaqui Phase taxa remain unchanged from the previous Lavasen Phase, except that Lupinus and Chenopodium/Amaranthus are now absent. The Manachaqui Phase faunal remains show the diminished importance of the larger animals utilized during the Precerarnic Period and a predominance of medium and smaller animals (Fig. 113). Initial Period travelers may have consumed guinea pigs as pre-prepared trail meals as modern villagers do today. However, Kent (1994:25-26) points out that guinea pig (Cavia porcellus) remains might be regarded as evidence of the shelter's function as a habitation since the animals frequent kitchen areas of Andean households where they find warmth and food scraps. They do not survive left alone at high altitudes for more than a few days. Hence, guinea pig remains alone do not constitute sufficient evidence for either wayside station or habitation functions. Nonetheless, the shift to smaller game is uncharacteristic of coeval Andean faunal assemblages as is the sudden disappearance of the chipped-stone tool industry. The use of such tools may have been obviated by habitually bringing pre-prepared meat to Manachaqui Cave. Pre-prepared meat could have been carried by travelers sheltering for the 521 night, but it also may have accompanied individuals camped for a number of days while dedicated to specific activities such as tending a tuber crop. In sum, the faunal assemblage provides only equivocal evidence of site function, perhaps thereby suggesting multiple functions, including both habitation and expedient use as a wayside station Evidence from Artifacts Although the use of Lavasen Phase rock-filled hearths remains unknown, the shift from stone-filled to simple basin hearths at Manachaqui Cave might be likened to evidence from Chiapas where Clark and Grosser (1995) describe an inverse relationship between decreasing quantities of fire-cracked rock and the increasing frequency of pottery vessels presumably utilized for cooking. At Manachaqui Cave however, it is not clear which of the Manachaqui Phase vessel types might have been utilized for cooking. None of the five shape classes show patterned fire-blackening, soot or other organic residues like the Initial Period Telarmachay pottery assemblage (Lavallee 1977:87), although food may still have been warmed in pots prior to consumption. Additional analysis of Manachaqui Cave's Preceramic Period remains may help resolve these questions. The Manachaqui Phase addition of pottery offers another interpretive tool for evaluating the rockshelter's function during the Initial Period. My literature search failed to turn up archaeological "transport assemblages" other than 522 mass-produced containers for long-distance commercial shipping of oil or wine described for the Old World. In order to formulate a series of expectations with which to generate reasonable inferences regarding vessel functions, I consulted several references, among which Rice's (1987) sourcebook, and Henrickson and McDonald's ethnographic survey (1983) were most useful. In her discussion of vessel functions, Rice (1987:225) notes that the many decisions made during pottery production are structured by considerations of the vessel's projected capacity, stability, accessibility of contents and transportability. Using her observations and assumptions, the Manachaqui Phase's potential utility as a transport assemblage can be assessed. According to criteria emphasized by Rice (1987:125141), Manachaqui Paste A ware is ideal for transport. The pottery is thin, fired very hard and is therefore lightweight (although perhaps brittle) . Restricted shapes provide adequate containment of vessel contents during transport. The round base is appropriate for carrying in a cloth bag, as are the rough exterior surfaces which prevent slippage. Recall that only the rims of Shape B jars are consistently burnished. Flanges provide purchases for handling and also prevent slipping. Thickened lips strengthen everted rims and furnish a means of fitting a skin or lid over the orifice for greater security during 523 transport. As expected, the assemblage is mostly composed of small vessels. With some exceptions, Manachaqui Phase vessel sizes are uniformly small, while many Central Andean assemblages feature both small and larger, bulky vessels intended for use around the home. Early Guafiape neckless ollas range from 30 to 50 em in maximum diameter (Strong and Evans 1952:253). Wairajirca Phase neckless ollas at Kotosh are likewise very large. While Manachaqui Phase Shape A Rim 2 and Rim 5 neckless jar categories include large variants, there are relatively few of these vessels. The dual size modes for the Shape A jars may reflect the site's alternate functions as habitation and wayside station. Perhaps vessel size comparisons should center on Manachaqui Phase Shape B short-necked jars which dominate the Paste A assemblage. With a modal rim diameter of 12 ern, and modal waist diameter of 16 em, Shape B jars are smaller than their Wairajirca counterparts with rim diameters between 15 and 19 em. On average they are smaller than analogous Huacalorna Coarse Brown jar Forms 4 and 5, and Brown Smoothed jar Forms 1, 2 and 3 which show larger rim and waist diameters and greater size variability. They do seem to share similar rim diameters with Machalilla Form 11 necked jars, but Form 11's rim and body proportions are unknown. Of course the Machalilla assemblage includes a much wider diversity of vessel shapes. 524 Specific references to vessels designed for transport include Henrickson and McDonald's {1983:634) description of a "true canteen" for water transport showing a maximum diameter between 10 and 20 em, and maximum height of 20 em. With a modal width of 16 em, Manachaqui Phase Paste A vessels meet these criteria. Only thirty-two of 208 sherds {approx. 15%) from carinated and semi-carinated vessel midsections measured show diameters greater than 20 em. Although vessel heights for the fragmented Paste A assemblage could not be measured, none of the hypothetical reconstructions project maximum heights greater than 15 em. DeBoer {n.d.) describes historic Amazonian Shipibo vessels utilized as portable cook pot, (kenti vacu), traveler's canteen (chomo vacu), an individual drinking mug (kenpo vacu) and a small food bowl (kencha vacu), and provides volume calculations for each. Manachaqui volummetric data are unavailable for comparison, but DeBoer's illustrations of all four Shipibo vessels depict sizes well within Henrickson and McDonald's range for canteens. Rice {1987:241) suggests that small orifices are preferable for transporting liquids and that the addition of vessel necks facilitates pouring. She observes that liquid containers usually show burnished and/or slipped (or glazed) interior surfaces to minimize absorption into the vessel walls, but this may not be true in the Andes. Manachaqui Shape B vessels exhibit variable attention to interior 525 surface finish, and vessel proportions may be a convenient compromise taking both secure containment and easy access to dry foods into account. The Manachaqui Phase assemblage's embellished rims (especially the notched rims) are impractical for pouring liquids. Archaeologists have found that vessel sizes correlate roughly with the size of groups serviced (Turner and Lofgren 1966; DeBoer n.d.). The previously noted bi-modal distributions of some Manachaqui Phase vessel types may prolonged stays by larger groups. Among the Shape B necked jars, larger orifice diameters (up to 24 em) lend decorated Rim 9 and 10 vessels an unusually wide range of sizes. Numerous individuals, perhaps a small family of semipermanent shelter inhabitants might be serviced by the larger, more elaborately decorated vessels. Inordinately large Shape A neckless jars (e.g. with 19-29 em rim diameters) would have been awkward for transport on long journeys, and may instead provide ceramic evidence for the rockshelter's alternate function as a permanent or semipermanent habitation. A few ground-stone artifacts may also serve as evidence for prolonged stays at Manachaqui Cave. Slate point fragments and chips in all of the phase assemblages suggest that points were habitually worked at the site, but Manachaqui Phase ground-stone tools appropriate for shaping and polishing the points represent an additional stage in 526 tool manufacture that may indicate extended occupations. A rare combination edge-ground cobble and hammer from probable Manachaqui Phase deposits in Unit 11 may have been utilized for processing food within the shelter. These items provide potential evidence for tool working and food preparation, activities that might take place in an Andean habitation. Only the fragment of quartz resembling other crystals found in Andean ritual contexts (Shady 1983; Burger 1984b) offers evidence of ritual activities. In general, the evidence for Manachaqui Cave's Initial Period functions suggests both habitation and use as an overnight shelter. However, contradicting interpretations that the rockshelter may have housed permanent agriculturalists, or that it was functionally connected to intensive agricultural activities at any time, is the virtual absence of milling stones, hoes, clod-breakers and other such artifacts. Consequently, the wayside station function appears predominant. The five-fold Manachaqui Paste A vessel shape inventory resembles the five-part breakdown of shape categories at Cajamarca sites like Cerro Blanco. It is most probable that, like the Shipibo (DeBoer n.d.), Manachaqui Cave's Initial Period users chose to travel with small-sized pots representing the same shape categories available for quotidian use in their domiciles. The Paste A assemblage should therefore be amenable to the kinds of comparative analyses required for this study. 527 Manachaqui Phase Evidence for Migrations Chapter 3 detailed the various migration theories postulated to explain eastern montane forest settlement and utilization. hypotheses, According to Meggers' and Kauffmann's the Manachaqui Phase Paste A ceramic assemblage should belong to Andean populations moving eastward from the highlands to the lowlands, while a Central Andean highland source is also presupposed by Moseley's highland-based verticality model. Theories of upslope population movement postulated by Tello and Lathrap dictate that Manachaqui Phase Paste A pottery, and other local styles along the Marafion-Huallaga divide between Bagua and Chavin de Huantar, should be derivatives of early upper Amazonian styles. These migration hypotheses can be evaluated empirically by considering the spatial and temporal patterning of early highland and lowland artifact styles. In Chapter 1 it was noted that the identification of ostensibly supportive stylistic distributions is fundamental to the plausibility of migration theories. Highland to Lowland Migrations Of the proposed migrations, eastward downslope incursions into the Pataz-Abiseo study area, whether by expanding highland populations or by highland colonists, are the least plausible. The distinctive Manachaqui Phase Shape 528 B necked jars, carinated body profiles and elaborate incised-applique decorations are rare or absent from early assemblages in the adjacent Central Andean highlands to the west and south. Initial Period Huamachuco appears to represent a demographic and developmental lacuna (Burger 1992), hardly the source of surplus populations or archipelago colonies. Even Manachaqui Phase Shape A jars or "neckless ollas" appear more closely allied to northern and eastern lowland neckless jars than to their Central Andean counterparts. While we expect that a special function site like Manachaqui Cave would contain only a subset of the full range of vessel shapes and sizes potentially available to the site's users, the Paste A assemblage actually shows a wide assortment of relatively complex vessel shapes and a rich decorative style with no clear highland antecedents. As the Manachaqui Phase assemblage finds its closest resemblances in the coeval Pandanche A, Early Huacaloma, La Conga and Montegrande styles of Cajarnarca, proponents of migration scenarios may point to the far northern Central Andes as a likely wellspring for Pataz-Abiseo area populations. However, these styles cannot be regarded as indigenous to either the highlands or the lowlands. The lack of unequivocal evidence that any of these Cajamarca styles predates the others precludes conjecture on origins. Also, the irregular distributions of shared attributes among these closely-related styles contradict expected wave-like 529 patterning left in the wake of population movements (Rouse 1986:177). All of the early Cajamarca ceramic assemblages differ from one another in design details. Yet even if all of these conditions were met, migration advocates would be forced to explain how it behooved Cajarnarca montane forest populations to colonize similar montane forest environments. Meggers has ascribed resemblances between the coastal Ecuadorian Valdivia/Machalilla styles and Initial Period Kotosh ceramics (e.g. tiered shoulders, zoned punctation, dot-ended incision, zoned dentate, dot-in-ring and zoned parallel hachure) to stylistically and temporally intermediate "stepping stones" still to be discovered in the intervening Andean highlands (Meggers et al. 1965:174). Neither Manachaqui Cave nor other recently investigated Peruvian highland sites show evidence of the postulated tell-tale route of descent to the Central Ucayali and surrounding western Amazonian lowlands. Unfortunately, Meggers' postulated migratory forest-free corridors providing passage to the lowlands cannot be credibly substantiated or discounted with the new Manachaqui Valley pollen evidence. Locating such hypothetical eastern slope corridors in prehistoric time and space presents a daunting challenge for paleoenvironmental reconstruction. Lowland to Highland Migrations As noted in Chapter 2, Tello, Lathrap, Lanning and others have emphasized the importance of the Marafion River 530 valley as an Initial Period migratory and/or diffusionary conduit leading from the Amazon Basin to the Central Andean highlands, and especially to Chavin de Huantar. However, the deeply incised Marafion Canyon has negligible bottomland for settlement by more than dispersed agricultural populations. While it is not impossible that cultigens, technologies, ideologies and their bearers worked their way up the valley, the upper elevations of the Marafion-Huallaga divide (above the Dry Forest Zone) are more likely to have attracted agriculturalist settlers. Traditional travel and transport routes in the precipitous eastern Andes tend to follow ridgetops rather than the circuitous, and often impassible, river bottoms. The pre-Hispanic Huanuco to Chachapoyas road described in Chapter 4 and the Suarez ridge reported by Hastings (1985) provide examples. For upslope Initial Period population movements from lowland Amazonia, the valley slopes and ridgetops above the Utcubamba river would have provided a likely entry as it purportedly did for Langlois' Utcubamba migrants nearly a thousand years later. It was to the Chachapoyas province that Tello had gravitated searching for Chavin civilization's origins when his 1937 expedition to the Marafion ended prematurely (Tello 1938; Mejia X. 1956). According to Lathrap (1970:107), archaeological evidence for a "collateral relative" of the Tutishcainyo and Wairajirca styles should be found on the narrow Central 531 Huallaga floodplains adjacent to the Pataz-Abiseo area. He also indicated the importance of the bend of the Marafion in the Jaen-Bagua area (Ibid.:107-109) investigated by Rojas (1969), Miasta (1976) and Shady (1987). Evidence for both Tello's and Lathrap's postulated migrations should therefore be found near the mouth of the Utcubamba River as well as distributed along the Marafion-Huallaga divide. At the mouth of the Utcubamba, Bagua's sequence of early assemblages consistently reflects stylistic development and patterned interaction occurring widely across the adjacent Cajamarca montane forests, rather than derivation from Central Ucayali or other reportedly early Amazonian styles. Some recent and potentially important investigations of early ceramic occupations in the upper Amazon potentially bearing on this problem have not yet been fully reported (Morales 1992; Ledergerber-Crespo 1995). At Manachaqui Cave, the most conspicuous potential evidence of upslope population movements from the adjacent lowlands is the carinated vessel profile, hallmark of the Central Ucayali Tutishcainyo phase assemblages. Lathrap's claim that Ravines had recovered Late Tutishcainyo ceramics from the Central Huallaga (Lathrap et al. 1984:46) is apparently a mis-reading or misinterpretation of Ravines' reports (e.g. Ravines 1978, 1981). Lathrap does not elaborate on claims that Pandanche A and Cotocollao are derivatives of the Late Tutishcainyo style. Tutishcainyo- 532 like low-positioned carination angles and embellished flanges do impart an Amazonian aspect to the Manachaqui Phase Paste A style that is rare or absent in coeval Cajamarca assemblages more frequently compared to late Valdivia styles. However, the principal Manachaqui Phase vessel, the restricted Shape B necked jar, has no Amazonian counterparts with the exception of Bagua's montane forest Morerilla assemblage and perhaps the deep, carinated Yasuni bowls. The applique and incised-applique techniques that typify the Manachaqui style contrast with the complex, incised-line decorations characteristic of early Amazonian styles. The Manachaqui Phase evidence, especially when combined with aforementioned evidence for the dearth of lowland impact on the Initial Period pottery further south at Piruro (Chapter 6), fails to support Tello's and Lathrap's hypothesized lowland to highland population movements. The Manachaqui Phase Paste A assemblage is stylistically unique, rich in singular details and internally coherent. described above, Regardless of scattered resemblances the assemblage cannot be considered a derivative of any known Central Andean or Amazonian tradition. Manachaqui Phase Paste A shares more attributes in common with distant Valdivia-Machalilla and perhaps Early Cerro Narrio styles to the north than with styles immediately to the west and east. Migrations from these 533 Northern Andean areas are improbable given salient differences between the Manachaqui assemblage and early ceramic assemblages from intervening areas such as the Loja and Catamayo valleys. The Suitacocha Phase Occupation The Suitacocha Phase occupation at Manachaqui Cave lacks the range of associated architectural evidence that was available for evaluating the rockshelter's function during the previous Manachaqui Phase. Two Sector A floors, Floors Y and X, and Feature R-3, a horizontal sherd scatter in Unit 21 Level 8, represent occupation surfaces, but information regarding specific activities must be gleaned from the organic remains and artifact assemblages. The lack of evidence for intentional reconditioning of the shelter interior may be considered evidence for the cave's expedient utilization strictly as a wayside station, but negative evidence alone is not convincing. Evidence from Botanical and Faunal Remains Important Suitacocha Phase additions to the inventory of Manachaqui Cave's organic remains are maize and beans, both "exotic" to the Tropical Alpine Zone. Otherwise, the utilization of local plant resources remains unchanged. An examination of Pearsall's food:wood and local:exotic+local ratios (Appendix F, Figs. 1-4) for Floors X andY shows relatively low values consistent with the wayside station 534 interpretation. Suitacocha Phase faunal evidence now includes fish, birds and reptiles occur, but quantities of these are not significant and none of the specimens are clearly exotic to the Manachaqui Valley. Paca and armadillo specimens in the Manachaqui Phase assemblage do not occur among the Suitacocha Phase remains. Sector A's Manachaqui Phase deposits are more substantial and richer in faunal remains than the Suitacocha Phase deposits (Tables 21 and 23), yet the latter contain more rodent bones than the former. The Suitacocha Phase animal size distribution (Fig. 114) is now clearly weighted toward utilization of small animals. With the addition of maize, the Suitacocha Phase organic deposits may now reflect the patterned consumption of the travel meals of parched maize and pre-cooked guinea pig for which Cobo, Young and Morales provide ethnographic evidence. Evidence from Artifacts Artifact remains from Suitacocha Phase deposits also provide support for the wayside station hypothesis. The scarcity of lithic remains has already been described as potentially supportive. The Paste A pottery continues to show a preference for small, restricted vessels with round bottoms, constricted necks and reinforced rims. A comparison between Suitacocha Phase and Manachaqui Phase rim measurements presented in Tables 5 through 15 shows a number of significant trends. First, the Suitacocha Phase Shape A, 535 B and F categories generally do not include large variants with rim diameters over 18 ern. On average, more constricted Suitacocha jar orifices are one centimeter smaller than their Manachaqui Phase counterparts, and this may be equally true for the smaller sample of Shape C vessels. The addition of Shape F jars during the Suitacocha Phase suggests that the two jar Shapes B and F may have served separate functions. Shape F's constricted neck and extended rim or "collar" would be more appropriate for transporting liquids although occasional similarities in Shape B's and Shape F's rim morphologies may indicate at least some functional overlap. Unfortunately, the reconstruction illustrated in Fig. 74 with an unusually large rim diameter of 15 ern and waist diameter of 21 em offers the only clue for interpreting jar sizes and proportions. If jar shape proportions are constant, however, then vessels with rim diameters of 11 ern (the Shape F mode) should exhibit modal waist diameters under 16 ern. Recall that sixteen centimeters is the modal and the average waist diameter of Manachaqui Phase Shape B carinated and semi-carinated vessels. From these calculations, we can deduce as a working hypothesis that Suitacocha Phase jars had volurnmetric capacities similar to Manachaqui Phase jars. Finally, the Suitacocha Shape F rims may be shorter than Bagua and Loja counterparts simply to minimize breakage during transport. 536 The Suitacocha Phase remains indicate changes in Manachaqui Cave's functions, and the trend seems to be toward increased functional specialization. Only a fragment of a prismatic quartz crystal from mixed Suitacocha and Colpar Phase deposits (Unit 18 Level 6), and a tiny flake of mica might be interpreted as indicators of ritual activities. In the absence of additional evidence of ceremonial activities, the wayside station working hypothesis seems the most probable. Suitacocha Phase Evidence for Migrations Migrations postulated for the Early Horizon occupation of the eastern montane forest are said to have originated either in the Central Andean highlands, or in the Amazonian lowlands northeast of Manachaqui Cave. The Suitacocha Phase assemblage shows even less stylistic resemblance to neighboring coeval Central Andean and western Amazonian assemblages than that of the preceding Manachaqui Phase. Stylistically, the Suitacocha assemblage might be likened to a Northern Andean finger extending southward deep into the Central Andean eastern montane forest. Populations in Huamachuco apparently remain dispersed, while nucleated Cajamarca populations to the northwest continue to produce large quantities of neckless ollas with the addition of polychrome bowls. In sum, hypothesized intrusions from the adjacent highlands and lowlands find no support from 537 Suitacocha Phase Paste A assemblages at Manachaqui Cave. The seemingly abrupt onset of the Suitacocha Phase, the appearance of maize, and the northerly aspect of its pottery style ostensibly bolster Lathrap and Isbell's postulated intrusions of Quechua-speaking migrants from the Ecuadorian Lathrap's (Lathrap et al. 1985) expression of full Oriente. support for Isbell's (1974) postulated Quechua expansion conjoins the two complex hypotheses, although Isbell's Quechua migrants were presumably entering the Central Andes at Huanuco nearly a thousand years before Lathrap's migrants supposedly arrived on the Central Ucayali bearing the Sivia style. According to this scenario, Suitacocha Phase Paste A should represent the CB Series, pottery diagnostic of Quechua-speaking intruders. Manachaqui Cave lies directly in the path of Lathrap et al. 's migration arrow (1985: Fig. 1) connecting Macas to the Central Ucayali by way of the Marafion-Huallaga divide. Despite evidence for the rockshelter's functional specialization, the diagnostic features of Lathrap's and Isbell's CB Series are sufficiently generalized across an assortment of functional vessel types that effective comparative analysis of the Suitacocha Paste A pottery is feasible. If the Suitacocha Phase Paste A pottery is a member of the CB Series, then its stylistic derivatives would include the Central Andean Kotosh Higueras and Amazonian Sivia styles among others. I intend to evaluate 538 the Quechua migration hypothesis empirically utilizing the archaeological data, but some problems with the assumptions common to both Lathrap's and Isbell's variants should be noted first. While some of his most basic theoretical assumptions regarding relationships between language and culture were challenged during Lathrap's lifetime, his convictions remained mostly unaltered (see Lathrap et al. 1987). Here it may be more appropriate to point out problems with assumptions specific to the Quechua expansion hypothesis. First, the eastern slopes were not an "unoccupied niche" (Lathrap et al. 1985:76) as the evidence from Kotosh and Manachaqui Cave demonstrate. Scholars such as Murra (1980), Lathrap (1970) and Bonavia (Bonavia and Ravines 1967; Bonavia 1978) have argued for associations between maize, terraced-slope agricultural systems and Quechua-speakers, yet evidence of ostensibly pre-maize occupations at Manachaqui Cave suggests that maize (the alleged technoeconomic key to the opening of the new niche) was not the determining factor that permitted eastern slope settlement. Furthermore, Schjellerup (1985:119) has described large terracing systems in Chachapoyas at altitudes between 3,200 and 3,800 m, too high for maize cultivation but ideal for high altitude tubers and grains. Admittedly, it is not impossible that natural or artificially-created microenvironments permitted cultivation of the most frost- 539 resistant varieties of maize. Regarding the location of the proto-Quechua homeland there is still no consensus. Stark (1985) has proposed an Ecuador Oriente homeland and an initial westward expansion around 400 b.c. She links this expansion to Porras' (1975:154-155) postulated upslope migration of populations that purportedly left Cosanga-Panzaleo pottery distributed throughout the highland Northern Andes. A later split is said to have sent populations to the north-Central coast of Peru by a.d. 800, and the final pan-Andean spread of Quechua occurred with the Inca conquest. Most importantly, there is no published evidence that Quechua was spoken in Chachapoyas prior to Inca conquest. Recall that non-Quechua language substrates have been suggested by Bandelier (1907) and Zevallos (1987). The ceramic evidence is even more damaging to the Quechua expansion hypothesis than the weakness of its basic assumptions. The Lathrap-Isbell hypothesis dictates that Kotosh-Higueras should be derived from {or at least historically related to) Suitacocha, yet the Suitacocha Phase Paste A assemblage more closely resembles the Kotosh Sajara-patac assemblage to which, by Lathrap's {1970:173) own reckoning, Higueras is decidedly not related. An assemblage morphologically comparable to Isbell's CB Series does not appear at Manachaqui Cave until circa a.d. 200-400 and the Empedrada Phase. Empedrada Phase pottery generally 540 has thicker walls, but Empedrada pastes lack the diagnostic "very coarse temper" (Isbell 1974:139). The purportedly diagnostic CB Series necked jar shape occurs in Northern Andean Valdivia, Bolivian Cochabamba's Formative phases and all around the eastern Central Andes long before the Early Horizon. Finally, there is no evidence for CB Series-like strap handles from the study area until the Late Horizon Quechua presence (Bonavia 1967; Church 1988). If Lathrap's hypothesis is correct, then the Sivia style must count both early Upano valley and Suitacocha styles as progenitors. Of the 10 Sivia vessel forms illustrated by Raymond et al. (1975: Fig. 53), only Form 8 convex bowls might find a Suitacocha counterpart. Lathrap's Quechua-speaking migrants purportedly carried pottery decorated with red-banded-incised techniques, stepped-fret and locking scroll motifs and applique face motifs. None of these elements occur on Suitacocha Phase Paste A pottery. To the above observations, DeBoer and Raymond's judgment can be added: "if Sivia were brought in by a progressive movement of Quechua-speakers colonizing the eastern slopes and valleys, one would expect some evidence of a Sivia-like complex to have turned up in the surveys of Hastings, Lathrap, Lathrap et al. in the Alto Perene valley, or of Ravines and DeBoer in the Huallaga valley" Raymond 1987:127). (DeBoer and Based upon their analysis of the evidence, these authors view Sivia as a lowland riverine 541 intrusion into the eastern slope premontane forests, rather than an Andean intrusion into the lowlands and the Central Ucayali. The Colpar Phase Occupation As Chapter 8 attempted to convey, the Colpar Phase occupation at Manachaqui Cave is rather ephemeral, and most clearly substantiated by Sector A's Floor W, and the hearth in Unit 14's Floor P. The paucity of evidence for intensive use of any kind between 500 b.c. and a.d. 200/400 is not inconsistent with the working wayside station hypothesis. By 200 b.c., travelers probably utilized Manachaqui Cave en route eastward to Gran Pajaten and other Montane Rain Forest Zone settlements. Pearsall's food:wood and local:local+exotic ratios buttress the wayside station interpretation. Floor W's soaring kernel:cob ratio stands in contradiction but, considering potential problems with this last ratio noted in Chapter 4, agreement between the other two provides more convincing evidence, in this case, favoring the wayside station hypothesis. Following the Chavin horizon, camelid remains predominate within faunal assemblages at north-Central Andean sites like Huacaloma where llamas assumed great importance as cargo animals (Burger 1992: Miller and Burger 1995). Camelid remains occur in small quantity during the Colpar Phase, but the presence of camelids alone does not substantiate either wayside station or habitation 542 interpretation. The Colpar Phase distribution of animal sizes (Fig. 115} displays the same preference for small animals that characterized the Suitacocha Phase. Colpar Phase evidence suggests that Manachaqui Cave continued to be functionally specialized, probably as a wayside station. Because artifacts of the Colpar Phase are difficult to isolate in Manachaqui Cave's macrochronology, caution warrants only the observation that Colpar Phase vessel sizes are comparable to those of the preceding Suitacocha Phase. Colpar Phase Evidence for Migrations Despite an apparent two or three hundred year gap in the local sequence corresponding to the Chavin horizon, the Colpar Phase assemblages provide clear evidence for cultural continuity in the Pataz-Abiseo area. Jar shapes and rims are altered only slightly from the preceding Suitacocha Phase. Only a few new bowl shapes could be assigned to the phase, and these sustain the long study area tradition of folding over and beveling rims. The few Colpar Phase decorative techniques all have Suitacocha Phase antecedents. This lack of marked change supports the interpretation that Pataz-Abiseo societies remained outside of the Chavin horizon interaction sphere. Changes in population cannot be postulated with support of the Colpar assemblages which instead provide evidence for local cultural stability. Proponents of verticality colonization scenarios may 543 contend that the ceramic linkages between Manachaqui Cave and Gran Pajaten indicate a highland-based colonization of the Montane Rain Forest Zone. In the Pataz-Abiseo area, Moseley's (1992:100) suggestion that the domesticated llama is the harbinger of "true verticality" is undermined by the observation that Gran Pajaten is poorly situated for agricultural colonization (Church 1994:292). Preliminary investigations indicate that primary economic activities at Gran Pajaten and Manachaqui Cave involved participation within interregional interaction networks (Ibid.). The Empedrada Phase Occupation Material remains from Manachaqui Cave's Empedrada Phase occupation suggest that the rockshelter served as a wayside shelter, and perhaps occasionally as a temporary habitation. Architectural elements include Sector A's palimpsest Floors T through G, some of which were apparently prepared with a light brown sandy silt. The high organic content and rich macrobotanical sample in these Sector A strata (Tables 1 and 18) attest to the shelter's intensive utilization during this period. Three of six Empedrada Phase hearths are of the embedded "single-stone" variety. Near Eastern pastoral nomads utilize the stones in similar hearths to prop up cooking pots (Frank Hole, personal communication 1996). It may be difficult to establish whether or not the singlestone hearths were used for cooking or warming food in the 544 absence of sooty potsherds or macrobotanical evidence from the still unstudied hearth fill. The embedded stone absorbed heat, perhaps serving to radiate warmth for the shelter's inhabitants after the fire's coals had cooled. Evidence from Botanical and Faunal Remains If fruit rinds are counted as local foods, then Pearsall's food:wood and local:exotic+local ratios may indicate that Floors T and S were deposited during habitation. The kernel:cob ratios stand in contradiction, especially for Floor S. Ratios from the remaining Empedrada Phase floors tend to support the wayside station hypothesis. Like the faunal assemblages from preceding phases, the Empedrada Phase assemblage contains a preponderance of small animal remains (Fig. 116). Kent observes that a Sciuridae (squirrel) specimen is "exotic," having been brought from a forest habitat. Larger animals representing greater percentages of the faunal assemblage may indicate increased large game hunting and the rockshelter's alternate function as a habitation. However, the increase in animals classified as "large" more likely reflects the presence of domesticated llamas, and it is only surprising that identified camelid specimens still do not outnumber cervid specimens (Table 27). Even while llamas are clearly in evidence, a herder's camp site function hypothesis is untenable. 545 Evidence from Artifacts The large quantity of unretouched and unifacially retouched chipped-stone artifacts may be interpreted as evidence of animal processing during extended habitation at Manachaqui Cave. The two chipped-stone point fragments were both recovered in Unit 23 inside the eastern wall (Feature R-6) extending from the cave mouth. At least 29 (42%) of the 69 ground and polished slate points belong to the Empedrada Phase assemblage. Worked chunks, disks and flakes suggest slate-working at the site. A complete quartz crystal again would seem to indicate ritual activities at the site. Major structural changes in Manachaqui Cave's ceramic assemblage have implications for interpretation of the rockshelter's Empedrada Phase functions. First, the unprecedented predominance of bowls suggests that these shapes were more important within Manachaqui's functional context than jars. While this may be true, it is also possible that the jar had been at least partially supplanted by some other container equally or more practical for transport by cargo animals, perhaps cloth satchels or alforjas (saddlebags). The larger rim diameters of the Empedrada Phase jars represent proportionately larger jars (Table 17) less practical for human transport, yet appropriate for storage use within a habitation, or for 546 securing within cloth or net alforjas as Guaman Poma's sketch demonstrates (Fig. 117). Empedrada Phase Evidence for Migration The abrupt shift in Manachaqui Cave pottery from a relatively localized to a Central Andean-affiliated style may be cited as evidence for mid-Early Intermediate Period migrations from the highlands. The inordinate number of specific similarities between Empedrada Phase and Pashash assemblages even permits conjecture of an upper Santa valley or Conchucos source for the invading populations. Such migrations ostensibly explain the presence of a developed Chachapoyas stone sculptural tradition (see Curtin 1951; Kauffmann 1983:525; Campana 1992) that is related in technique and motif to the Recuay tradition as defined by Schaedel (1948). Can the Empedrada Phase Paste A and Paste C assemblages be attributed to Recuay colonization of the study area and perhaps even greater southern Chachapoyas? Despite the striking qualitative change in the PatazAbiseo Paste A pottery style, a migration hypothesis is difficult to sustain. Style attributes that persist from anterior phases include the use of red paint on vessel rims, notched applique applied to the vessel waists, folded rims and rim lobe "handles." In fact, the eastern montane forest tradition of rim reinforcement by folding appears to have been introduced at Pashash at the beginning of the Recuay 547 Period. Other new Empedrada Phase features are not Central Andean such as the Panzaleo-like punctations on rim interiors. Furthermore, similar episodes of pottery re- styling were occurring at other locations around the perimeter of the Central Andes around this time, but I will return to this subject later. Given the controversies that always surround migration hypotheses generated on the basis of "traits," in this case strictly stylistic evidence (Adams et al. 1973: 488, 503), it is appropriate to weigh other kinds of evidence against the expectations of a colonization hypothesis. Some of the problems identified below pertain equally to other migration hypotheses addressed in this thesis. First, there is no clear evidence that Grieder's Recuay Period is temporally prior to the Empedrada Phase, a necessary condition to sustain migration hypotheses. Second, potential motivations for Recuay population movement or colonization must be considered. Incursion into a distant, occupied territory with an unfamiliar ecology is no small undertaking. Recall that Adams et al. 's (1973) world survey of migration theories determined that society-level population movements across major environmental transitions are rare indeed. Lack of evidence for highland demographic pressure or any other motivating condition also renders population movement hypotheses highly suspect. Interpretations generated within the verticality or 548 "complementarity" theoretical framework would likely invoke ecological symbiosis or "diadic interdependence" (cf. Salomon 1985:515, Fig. 23.1) to explain the EmpedradaPashash relationship. In this case, strictly ecological hypotheses are implausible considering geographic relationships between the upper Santa valley and the study area. At Pashash, access to nearby lower valley ecological zones lies in the surrounding Cabana, Santa and Marafion drainages. Proulx (1982) has documented Recuay control of western slope ecological zones down to 300 m in the coastal Nepefia Valley. The Abiseo drainage montane forests across the Marafion Canyon and over 100 km distant ostensibly offer few if any advantages for agriculture that cannot be found closer to home. The Marafion is known to have coincided with major cultural boundaries by the Late Intermediate Period (Garcilazo 1966; Espinoza 1969; C. Julien 1985; D. Julien 1993). If Gran Pajaten was the site of mid-Early Intermediate Period verticality colonization as ceramic and radiocarbon evidence from the study area might suggest, then we still must confront the aforementioned problem of the site's unfavorable location for agriculture. But perhaps the most serious problems with the colonization hypothesis are exposed by considering expected archaeological patterning under conditions of archipelago colonization or "direct control." Scholars regard economic self-sufficiency engendered by colonization and kin-based 549 exchange as the raison d'etre and/or the result of "vertical control of a maximum of ecological tiers" (Lynch 1981:226; J. Topic and T. Topic 1985:58; Moseley 1992:44). Because colonization establishes "ideal" conditions of direct control intended to obviate external exchange, we should therefore expect minimal material evidence for external exchange relationships. Contrary to this expectation, both Gran Pajaten's preAbiseo Phase assemblage and Manachaqui Cave's Empedrada Phase assemblage contain substantial quantities of pottery (7% and 18% respectively) not only exotic to the study area, but exotic to the north-Central Andean highlands as well. Empedrada pottery's contextual associations with styles of the Junin puna and Northern Andean montane forests undermine verticality arguments. These data do not conform to patterning consistent with the verticality model. It should be added, Van Buren's (1996) evidence that the Lupaca colony of Terata Alta was indeed engaged in external exchange further undermines Murra's model. Summary of Evidence for Migration at Manachaqui Cave Both the long sequence of southern Chachapoyas cultural development excavated at Manachaqui Cave, and the shorter northern Chachapoyas ceramic sequence excavated by Ruiz (1972) in the upper Utcubamba Valley show remarkable stylistic continuity through time. Common to both northern 550 and southern areas ~s an ancient decorative tradition based on applique and incised applique frequently rendering serpent motifs. Cuelap sequence. Serpent motifs occur throughout Ruiz's In the Pataz-Abiseo area, they appear on the earliest pottery assemblages at Manachaqui Cave and throughout the rockshelter sequence, as well as on late preColumbian pottery recovered from Gran Pajaten (Bonavia 1968). The population movement hypothesis first posed by Langlois, and Moseley's colonization hypothesis are likewise crippled by archaeological evidence for long-term continuity. Taken together, the two Chachapoyas sequences from the Marafion-Huallaga divide render hypothetical migrations highly improbable. The great irony inherent within the migration hypotheses addressed in this thesis is the heavy burden placed upon stylistic evidence to support fundamentally materialist frameworks that typically assign "style" to the stochastic or epiphenomenal realms of culture. Meggers' evolutionary scenario equates ceramic zoned hachure incision, incised rims and polychrome decoration with biological populations, just as Lathrap identifies neolithic Arawakan populations by their pottery's carinated basal angles. A conservative materialist approach might focus more narrowly on evidence for changes in diet and subsistence technology potentially indicating transformations in means of production. 551 Evidence of an intrusive "neolithic" population should be recognizable, although Manachaqui Cave is probably not the most appropriate functional context in to find such evidence. While root-crop cultigens are always important seldom preserved subsistence components, it is significant that, from the Lavasen Phase through the Empedrada Phase, the inventories of botanical and faunal taxa recovered from the excavations remain unchanged excepting the predictable Early Horizon additions of maize and beans (Pearsall 1992), and Early Intermediate Period inclusion of camelids (Miller and Burger 1995). While an argument for the impermeability of the Andean-Amazonian ecotone cannot be maintained in the face of contrary historical, ethnographic, linguistic and archaeological evidence, the magnitude of the transition between radically disparate environmental and cultural worlds cannot be dismissed as historically inconsequential. Pearsall's (1992, based on Harlan 1975) Agricultural Complex" "Lowland (below 1500 m), and "Andean Agricultural Complexes" (1500 - 3500 m) show only four cases of taxa overlap or co-occurrence. In other words, a "neolithic" society migrating across the ecotone would have to replace its means and modes of production completely. Archaeological evidence for such fundamental change at the crucial momencs in prehistoric time has not been forthcoming. 552 Evidence for Colonization in the Greater Eastern Montane Forest In Chapter 1, additional indices of intra-site functional variability, regional stylistic distributions and settlement patterning were described as potential predictive aids with which to identify archipelago colonies implanted in the eastern montane forest by highland-based, vertically distributed economies. Unfortunately, the intra-site functional homogeneity that Bonavia cites as characteristic of these "dependent settlements" cannot be fully evaluated without appropriate sub-surface sampling strategies that have yet to be implemented in the eastern montane forest. Bonavia's view that it is impossible "to establish differences in the social structure by means of samples from the sites" (Bonavia 1978:400) is unduly pessimistic in view of the limited fieldwork undertaken to date. The Southeastern Montane Forest Apart from the Late Horizon Inca occupation of the montane forests east of Cuzco and Lake Titicaca, there is little published information bearing on indigenous settlement and cultural development in the southeastern montane forest. Scant details of a Middle Horizon Huari or Tiahuanaco presence are found in Savoy's (1970:99, 117) report of explorations in the lower montane or premontane forests of the lower Urubamba Valley. Also in the lower Urubamba Valley, Kendall (1976) provides some descriptions 553 and plan maps of Late Intermediate Period settlements consisting of round and oval constructions and yielding Killke or Killke-like pottery. However, the extent of the lower Urubarnba occupations and their ecological contexts have not yet been analyzed. The late prehistoric settlements within the Inambari Valley east of Puno cursorily described by Isbell (1968) have yet to receive more intensive study. Clearly the southeastern montane forests were occupied long before the Late Horizon, but the lack of archaeological data prohibit serious consideration of hypothetical prehistoric migrations or verticality colonization. The Central Montane Forest The reader will recall from Chapter 1 that Bonavia (1978) regards absolute functional homogeneity a chief characteristic of central and northeastern montane forest site architecture. In the central subdivision, Hastings' (1985) more recent work contributes a relatively extensive description and analysis of prehistoric settlement patterns in the Tarma Valley. At the Early Intermediate Period or Middle Horizon site of Tranca, Hastings (Ibid.:632-633) identified approximately 71 mostly rectangular stone buildings, some of which are arrayed in rows or "welldefined patio clusters." The Late Intermediate Period site of Paraupunta, which succeeds Tranca at the apex of a similar hierarchical settlement distribution, contains 554 approximately 150 round buildings showing a bi-modal size distribution. Like Tranca buildings, they occasionally articulate with one another around patios (Parsons and Hastings 1988: Fig. 7.4). As Bonavia observes, some aspects of social structure are apparently difficult to detect from central montane forest site architecture. However, such intra-site distributions of habitations and patio groups are not restricted to sites like Caballoyuq, Tranca and Paraupunta. They are also characteristic of the highland Upper Mantaro Valley Wanka centers of Hatunrnarca and Tunarnarca (Earle et al. 1978, 1987). This observation suggests that, based upon surface architecture, Tranca, Paraupunta, Tunarnarca and Hatunrnarca all exhibit similar degrees of intra-site functional heterogeneity which may reflect similar forms of social organization among the populous. Apparently, surface architecture is not a reliable indicator of settlement dependency or autonomy at central montane forest sites. In support of his position disputing local autonomy for Tarrna Valley settlement systems, Hastings argues that: "the drastic shift in settlement locations ... suggests economic changes not previously underway during the Tranca Phase" (1985:731). However, he describes the transition from the Tranca to the Paraupunta settlement system as "a gradual process" (Ibid. :727). Most damaging to arguments for colonization is his observation that the early Tarrna Valley 555 settlement systems have no highland counterparts (Ibid.:642}. Hastings' statement that administration and decision-making were "rendered locally rather than in a hypothetically higher-order center somewhere in the sierra" (Ibid.:732} may be interpreted as contradicting his theoretical position that such systems reflect "externally introduced change" (Ibid.: 731}. In essence, Hastings' archaeological data equally support interpretations of fully-autonomous Tarma Valley settlement systems. Bonavia's (1972} Mantaro Valley site of Taipi may be functionally analogous to Hastings' Tranca and Paraupunta type-sites at the apex of indigenous, two-tiered montane forest settlement systems. The postulated Huari affiliation of Taipi and Raymond's Lower Apurimac Valley colonies can also be called into question on the basis of scant evidence. Raymond (1976:206) judges Bonavia's eastern Ayacucho collection of artifacts "meager and non-diagnostic," and admits that stronger correlates of Huari attribution are needed to verify his own colonization hypothesis (Raymond 1976, 1985}. The hypothetical Huari outposts of Vista Alegre and Palestina remain undated and only superficially described. importantly, Most the Simariba area ruins that Raymond {1992:27) observed, but could not investigate under the difficult field conditions, almost certainly pertain to indigenous montane forest populations. 556 Even granted Huari attribution for the premontane forest sites of Palestina and Vista Alegre, it remains to be demonstrated that models of Huari imperial conquest are not better suited than colonization to interpret the montane forest Simariba archaeological data. In a framework invoking political hegemony, Palestina and/or Vista Alegre potentially functioned as administrative centers established to control local populations as well as to manage interregional exchange. Neither of these administrative functions falls within the purview of verticality colonies as defined by Murra. The Inca sites recently reported by Schjellerup (1992) deep in the lower montane forests above the Central Huallaga may be functionally analogous to the lower Apurimac sites. The Northeastern Montane Forest As in the central and southeastern subdivisions, the problem of northeastern montane forest intra-site functional homogeneity cannot yet be addressed with published archaeological data from sub-surface contexts. However, surface architecture at Cuelap (Narvaez 1988) clearly exhibits functional heterogeneity. The site is clearly divided into upper and lower barrios or sectors, and constructions include "ceremonial" rectangular buildings, Ushaped structures and the unusual "Tintero." buildings occur in a wide variety of sizes, Circular show a diversity of stone inlay ornamentation and group themselves in four 557 distinct configurations, two of which articulate with open patios. The functions of numerous massive rectangular structures that dominate the ancient settlement of PirkaPirka above Uchucmarca (Vega 1982} remain to be interpreted through sub-surface sampling. At Gran Pajaten (Rojas 1967; Bonavia 1968} the differential arrangement and decoration of buildings likewise· strongly suggest functional diversity. The most recent excavation data pertinent to the problem of building functions has not yet been published. However, Bonavia's hypothesis that Gran Pajaten's rectilinear slate mosaic motifs and sandstone curvilinear motifs represent foreign colonists and local populations respectively cannot be sustained in view of the motifs rendered in both techniques illustrated by Pimentel G. (1967: Fig. 3}. Stylistic continuity evident in Gran Pajaten's pre-Abiseo Phase and Abiseo Phase ceramic remains reflects autochthonous cultural development since the Early Horizon, and likewise constitutes evidence contradicting hypothetical site-unit intrusions. Forthcoming publications of excavation data from the nearby site of La Playa will also shed further light on the problem of intra-site functional heterogeneity in the northeastern montane forest. Although little regional sampling has yet been effected in the northeastern montane forest, the distributions of settlements, architectural attributes and ceramic styles 558 exhibit considerable uniformity, rather than archipelagolike heterogeneity. The northeastern montane forest has been singled out as a distinctive archaeological region with unique characteristics that suggest a long period of local and autonomous cultural development (Church 1994:283). Among these characteristics are architectural elements such as cornices, staircases and stone frieze decorations, and the previously described pottery with applique decoration featuring serpent motifs. These same widely distributed attributes prompted Lumbreras (1974) to identify the region as a "new culture area," and motivated Ravines and Brush to propose state-level socio-political organization for preIncaic Chachapoyas. If Raymond's set of settlement pattern expectations for eastern slope verticality colonization is applied to the Pataz-Abiseo area, then we should find comparatively light montane forest settlement corresponding to the highly compacted ecological zonation. Yet in Chapter 3, it was noted that prehistoric populations in the Pataz-Abiseo study area were clearly concentrated within the forested Huallaga watershed, rather than on the unforested slopes of the Marafion drainage. Demographic patterns, coupled with the documentary evidence for autonomous forest populations subsumed by the encomienda of Sucos y Puymal, cannot be reconciled with the expectations of a verticality colonization framework. Neither the early settlement at 559 Gran Pajaten, nor the early Tarma Valley settlements correspond to any archaeologically visible highland polities capable of subsidizing their establishment. The prehistory of the Central Andean eastern montane forests suffers from many of the sampling biases that plague the archaeology of the Amazon lowlands. Yet while the scarcity of archaeological data from Amazonia has resulted in polemic debates, the scant information from the eastern montane forests has occasioned mostly pessimistic assessments of their potential productivity and habitability. Just like the Amazon Basin and the Intermediate Area, the montane forests have been offered as a case study in stunted cultural evolution. However, because the eastern montane forests correspond to the boundary between contrasting, yet juxtaposed culture areas or co-traditions, they may also be alternatively viewed as loci of cultural creativity and archaeological possibilities. In the following final chapter, archaeological data from Manachaqui Cave and the Central Andean montane forests are re-interpreted utilizing a framework emphasizing local cultural development and interregional interaction. CHAPTER 11 MONTANE FOREST CULTURAL DEVELOPMENT AND INTERREGIONAL INTERACTION All pristine developments are secondary developments dependent on outside resources, alliances and events (Clarke and Blake 1994:20). In the preceding chapter, I attempted to show that the archaeological data from Manachaqui Cave fails to support migration scenarios formulated to account for stylistic distributions and patterns of cultural development in the Central Andes. In the remainder of this thesis, I will present an alternative interpretation of this material patterning utilizing evidence from Paste Group A phase assemblages at Manachaqui Cave, and archaeological daca reported from surrounding regions, prehistoric "interaction spheres" to reconstruct (Caldwell 1964). Caldwell's term is implemented here to refer to broad fields of stylistic similarity among the ceramic assemblages of interacting societies distributed across the landscape. The ceramic interaction spheres are reconstructed on the basis of shared clusters of pottery design attributes that reflect limited kinds of interaction in which ceramics presumably functioned as visible media for symbolic expression. Thus, they reveal only one of numerous nested, overlapping and 560 561 complementary fields of interaction that occurred simultaneously across space. Early Andean ceramic interaction spheres cross-cut other interaction spheres defined by co-occurrences of architectural features like conical adobes, sunken circular courts and ceremonial hearths. As heuristic devises, they are units of analysis that provide one kind of evidence for frequent, if not habitual, interaction. In the Central Andes, as in most parts of the world, they are small pieces in very large puzzles. Ceramic interaction spheres correlate with archaeological phases because the same ceramics serve to define each phase. However, it should also be emphasized that interaction spheres, like other aspects of culture, were in constant flux. Archaeologically detected interaction spheres may be comprised of smaller interaction spheres that did not necessarily operate simultaneously. The utilization of stylistic similarity to reconstruct fields of interaction in the Central Andes is rare, but not unprecedented. MacNeish et al. (1975) and Browrnan (1975) have contributed interpretations of early exchange networks, but they did not fully implement stylistic similarity as a criterion. The studies of Mohr-Chavez (1981b) and Burger (1984a, 1988, 1993) have made the most explicit use of style and sets of style attributes to interpret early Andean interaction within and between regions. Inherent within 562 their analyses is the view that shared clusters of pottery design attributes facilitated cultural interchange by conveying a sense of sameness or common social identity, and open inter-group boundaries (Mohr-Chavez 1981b:343; Hodder 1982; Burger 1984a:45). This chapter's point of departure is the assumption that ethnic, linguistic and genetic boundaries may coincide with interaction sphere configurations, but such correspondences cannot be presumed solely on the basis of archaeological evidence. Rather, ceramic interaction spheres merely signify the intent to exchange goods and information across social, political and/or ecological boundaries. Whether they provide a window on ethnic affiliation, social interchange, ritual interaction and/or political alliance remains to be determined by the evaluation of complementary data sets and the contexts of interaction. Apparently, some modes of interaction can produce precisely the same kinds of material patterning that archaeologists have cited as evidence for migrations and the second half of this chapter will be devoted to discovering these. While Manachaqui Cave's Paste A assemblages provide indirect evidence enabling the reconstruction of interaction spheres, Paste Groups B and C provide complementary direct evidence for exchange networks. These networks overlay and cross-cut the aforementioned interaction spheres. At 563 Manachaqui Cave, small Paste Group B vessels probably mark the physical presence of travelers from distant localities. The quantities of Paste Group B and C sherds might constitute a gross measure of interaction's intensity, but they might not. Based upon Burger's (1992, 1993; Miller and Burger 1995} interpretation of the Chavin horizon as a widespread "socioeconomic metamorphosis" accompanied by the establishment of new political boundaries, we might hypothesize that Paste Groups B and C are more likely to reflect long-distance travel prior to the Chavin horizon, and mediated exchange following the Chavin horizon. Pre-Lavasen and Lavasen Phase Evidence for Interaction Finding Preceramic occupations at the edge of the northeastern montane forest, a purportedly "empty niche," should not be surprising as forested terrain did not deter Central Andean Preceramic Period populations from making a living within and adjacent to northern (Kaulicke 1981} and western (Dillehay and Netherly 1983; Dillehay et al. 1989} montane forest ecological zones. In the Northern Andes, Salazar (1983:102-104} argues that El Inga and Chobshi Cave lay at the paramo-montane forest ecotone during the early Holocene occupations. He considers the ecotone a preferred habitat that permitted access to resources in both ecological zones, as well as to specialized edge species (Ibid.:108-110}. While habitation can be inferred from 564 Lavasen Phase material patterning, there is little evidence indicating the occupants' degree of permanence. Paleoenvironmental studies in the north-Central Andean highlands are needed to determine the past extent of montane forests before early prehistoric subsistence patterns can be accurately reconstructed. One Preceramic Period find at Manachaqui Cave merits discussion here, because it portends long-distance interactions characteristic of the rockshelter's ceramic era occupations. Burger et al. (1995) report that a tiny pressure flake (Unit 2 Level 30) and the distal portion of a unifacially retouched flake (Unit 6 Level 36) made of obsidian from the Lavasen Phase deposits constitute unusual evidence for far-flung exchange linkages. Not only do these pieces represent the first documented occurrences of Alca obsidian (from northern Arequipa) in a Preceramic Period context, but they also constitute the most northerly discovery of Precerarnic Period obsidian artifacts in the Central Andes. Burger et al. observe that both Alca and Ecuadorian obsidian sources lie approximately 1000 km from Manachaqui, and the Cuzco Quispisisa source lies 655 km distant. They note that "the natural routes of transport from the north present less natural obstacles, but this was apparently less significant than cultural forces that shaped patterns of interaction." While the Lavasen assemblages offer few 565 additional clues with which to reconstruct interaction spheres closer to horne, Shady and Rosas (1979:127} point out that earliest Initial Period ceramic styles provide a view of Terminal Precerarnic Period interaction spheres. Manachaqui Phase Evidence for Interaction By 1200 b.c. and the mid-Initial Period, Manachaqui Cave's users were carrying pottery to the rockshelter. Although they may have been involved in subsistence-related activities involving habitation at the rockshelter, and perhaps limited transhurnance, the broad field of ceramic similarity shared by western, northern and eastern montane forest populations indicates that they were simultaneously engaged in long-distance exchange activities. The Manachaqui Phase Paste A assemblage's closest, or primary, relationship is clearly with the earliest Cajarnarca and Bagua assemblages. Shady (1987b} and Onuki (1985} associate these styles with the north-Central Andean yungas, while Onuki further identifies them with a hypothetical Initial Period subsistence strategy termed the "Yunga Tradition." The Montegrande, La Conga, Early Huacalorna, Pandanche A, Morerilla and Manachaqui styles are distributed along an altitudinally-restricted swath approxirnati~g past distributions of northwestern, northern and northeastern montane forests. This "Northern Montane Forest Ceramic Interaction Sphere" describes a wide arc around the northern 566 rim of the Central Andean region and was apparently positioned along a broad front to facilitate interaction with Northern Andean societies (Fig. 118). Like its companion Central Andean styles of the Northern Montane Forest Interaction Sphere, the Manachaqui Phase Paste A assemblage incorporates numerous Northern Andean formal and decorative attributes seen in the coastal Valdivia and Machalilla styles, and in the highland Cerro Narrio and Cotocollao styles. populations Unlike styles exhibited by migrating (Anthony 1990:903), Manachaqui Paste A's northern influence cannot be attributed to "a highly restricted point of origin." The archaeological evidence indicates that montane forest environments were attractive for early Central Andean population nucleation. Rather than from a demographic push, as Kauffmann postulates, nucleation resulted from the pull of desirable opportunities for interaction with groups in neighboring regions, and with other nucleated groups situated at the Central Andean boundaries. Despite the suitability of Huamachuco's intermontane valleys for a variety of productive subsistence systems, it's populations apparently remained scant and dispersed within a demographic "vacuum" left in the wake of attraction to the boundaries. Based upon their fieldwork in the early 1980s, J. Topic and T. Topic {1985:60-62) concluded that early Central Andean north-south interaction routes bypassed Huamachuco by 567 following coastal routes. La Galgada participated in this same interaction sphere which in effect surrounded Huamachuco. Grieder and Bueno (1985:106), Burger (1992:53) and Shady (1993:116), describe La Galgada as a node in early interregional exchange that connected the coast and tropical lowlands. The Manachaqui Paste A assemblage shows a weaker, yet clearly distinguishable relationship to other Initial Period assemblages recovered from sites in the eastern Central Andes. Manachaqui, Wairajirca, "early formative" Pachamachay, "formative" Telarmachay, Muyu Moqo C-D and Marcavalle A and B are connected to early Northern Andean, especially Valdivia-Machalilla, styles by co-occurrences of short-necked jar shapes and scattered features such as carinated body profiles, notched applique, specular hematite paint, stepped or two-tiered shoulders, rim beveling and complex bottle spouts. Lathrap has argued that these similarities are diagnostic of early migratory (Lathrap 1963) or exchange (1971:95) links between the Northern Andean Machalilla and Amazonian Late Tutishcainyo styles. Archaeologists investigating the earliest ceramicbearing occupations on the eastern slopes of the Bolivian Andes have emphasized similarities between Cochabamba Formative pottery exhibiting pattern-burnishing and complex incised decoration and the earliest assemblages from far northern Peru and southern Ecuador (Ibarra and Querejazu 568 1986:152-163; Brockington and Sanzetenea 1989; Brockington et al. 1995:164-166). Evidently, this eastern-slope interaction sphere extended more than 2,500 km from southern Ecuador to southern Bolivia, and perhaps beyond. Some Northern Andean, Amazonian and south-Central highland features are found on the south-Central coast of Peru at Hacha (Robinson 1994: Fig. 14) and Erizo, suggesting that the interaction sphere penetrated south and westward in a manner analogous to its inclusion of the inter-Andean littoral at the modern border of Ecuador and Peru. Considered together, patterns of stylistic similarity indicate that prominent Initial Period interregional interaction networks were longitudinally oriented north and south along the Central Andean flanks. Apparently, principal north and south regional communication and exchange networks followed montane forest routes on the slopes of the cordillera rather than descending to the coast as the Topics propose. The stylistic discontinuity between the Jequetepeque and Viru valleys noted by Ulbert (1994:149) reflects the permeable southwestern boundary of the Northern Montane Forest Ceramic Interaction Sphere matched on the southeast by a similar boundary represented by salient differences between the Initial Period assemblages at Manachaqui and Kotosh. Cross-cutting longitudinal montane forest interaction spheres, more narrowly targeted east-west interaction 569 spheres linked Upper Huallaga Kotosh and upper Urubarnba Marcavalle populations to Amazonian Central Ucayali societies. While the longitudinal sphere features design attributes found among the Valdivia and Machalilla styles, the eastern montane forest interaction sphere (or spheres) features complex line-incised decoration, post-fire paint and flanges observed in highland Wairajirca and Marcavalle A ceramics and coeval lowland styles like Cave of the Owls Fine Ware and Tutishcainyo. These relatively localized and intense interaction relationships excluded the nearby Amazonian Upper Pachitea societies. In Chapter 6, I suggested that the Paste B vessels, along with the two Paste A vessels exhibiting incision (Fig. 64), may occur late in the Manachaqui Phase. They are stratigraphically associated with the uppermost Manachaqui Phase deposits corresponding to Floor AA, Floor z and Feature R-4 (radiocarbon dated 900 ± 90, 850 ± 90 and 830 ± 100 b.c. respectively), and seem to reflect developments in adjacent regions east and north. Based upon known spatial and temporal distributions of bottles with asymmetric spouts in upper Amazonia, DeBoer (n.d.) proposes a "flash horizon dated to about 800 BC." He observes that the spatial distribution of these bottles mirrors later upper Amazonian interaction spheres that correspond to prehistoric Curnancaya, historic Shipibo and 17th century Jesuit missionary travel routes and interaction networks. The 570 Manachaqui Phase B pottery apparently arrived at the rockshelter during what appears to represent a brief period of increased long-distance foot travel. Suitacocha Phase Evidence for Interaction The terminal Initial Period flash horizon can be conceptualized as a time when Central Andean, Amazonian and Northern Andean societies became acutely aware of new possibilities for interaction with one another. Ultimately, it foreshadowed abrupt stylistic change in the Pataz-Abiseo area and the beginning of the Suitacocha Phase around 800 b.c. During the Early Horizon, it is more difficult to evaluate Paste Group A's stylistic relations by degrees of similarity. Styles are more locally distinctive, and the spatial distribution of resemblances is uneven, and perhaps typified by the spotty appearance of Shape F-like rims in Central Andean assemblages. The irregular spatial distributions of many shared attributes not only constitute evidence against wave-like migrations, but they suggest specific modes of interaction which target distant localities while bypassing some that are closer. The most extraordinary example of this stylistic leap is the Suitacocha Paste A assemblage itself. Rather than displaying closest resemblances to the Kotosh, Huamachuco, Bagua or Shakimu styles flanking Manachaqui Cave at the four cardinal points, the Paste A assemblage finds its closest 571 relationships to more distant Northern Andean so-called "Chorreroid" styles (Fig. 119). As they did during the Initial Period, northern montane forest societies in Cajamarca remain at the forefront of interaction with Northern Andean societies. Evidence for sparse populations in Huamachuco continues to suggest a centrifugal attraction to the northern edges of the Central Andean massif. Archaeological data from the Early Horizon northCentral Andes demonstrate that the Suitacocha Phase transpires against a backdrop of accelerating construction of public monuments and long-distance interaction between growing population centers. In breadth the Early Horizon interaction sphere transcends the earlier Northern Montane Forest Ceramic Interaction Sphere which has splintered into a number of smaller discrete and overlapping interaction spheres distinguished by relatively localized decorated bowl and bottle styles. The local spheres correspond to Early Horizon centers acting as nodes in redistributive systems controlled by emergent elite (Shady and Rosas 1979; Burger 1984a, 1992; Topic and Topic 1985). The Early Horizon interaction sphere was consolidated by an over-arching exchange network that distributed "trade pottery" among local centers such as Naiiafiique, Kuntur Wasi, Cerro Blanco, Bagua, Kotosh and Chavin de Huantar. Burger likens this integrative web to Renfrew's "peer-polity interaction," involving exchange of exotic "prestige items 572 with a high information content" (Burger 1984a:46, 1993:74). Decorated bowls were apparently perceived as particularly well-suited for conveying gifts between population centers (Mohr-Chavez 1981b:343). The elaborately decorated bowl represented by the single Paste B5 sherd is typical of vessels likely to circulate in the Early Horizon northCentral Andes. In the Early Horizon context of nested or concentric Central Andean boundaries, outlying centers with complex local styles and a relative abundance of associated exotic wares like Nafiafiique and Bagua seem to have developed as "gateway communities" (as defined by Hirth 1978), channeling interregional exchange between the lowlands and the Central Andes (Burger 1992:101). The data pertinent to Manachaqui Cave's function at this time, and especially the extraordinary resemblance of the Suitacocha assemblage to Ecuadorian highland and lowland styles suggests that the rockshelter also served within a similar context of channeled interregional exchange. Despite its Central Andean geographic context, the Suitacocha Phase Paste A assemblage exhibits a primary relationship to Northern Andean and Oriente styles, rather than to neighboring northern montane forest Cajamarca and Bagua styles. While Suitacocha is clearly more "Chorreroid" than Cajamarca's northern montane forests styles, all of these seem to reflect articulation with the Chorrera interaction 573 sphere. Especially illustrative of Suitacocha's "Chorrera connection" is the appearance of indisputably Northern Andean incised motifs and iridescent paint. The Shape F jars with decorated shoulders and the distinctive basal joints of Suitacocha carinated bowls (Fig. 80e-h) are likewise unequivocally "Ecuadorian." The overall paucity of Suitacocha Phase Paste Group B pottery is perhaps less supportive of the interaction interpretation, although it is likely that such important vessels were protected from breakage during transport. Evidence for interregional interaction between the Central and Northern Andes in the form of intrusive pottery is exceedingly rare in both regions. If the Suitacocha Phase Paste B4 sherds do originate in Ecuador's eastern montane or premontane forests, then they might constitute the first documented incidence of Northern Andean pottery recovered from Early Horizon stratigraphic contexts in the Central Andes. Direct contact between societies in these two regions may likewise be indicated by the presence of intrusive Early Horizon sherds at Chaullabamba recently reported by Hocquenghem et al. (1993: Fig. 3). Based upon perceived similarities between the Bagua and Cerro Narrio styles, Shady (1987b:480) contends that the Marafion Valley is a locus of Early Horizon inter-Andean interaction. The Suitacocha evidence lends support to her observation. Shady and Rosas (1979), and Burger (1988, 1992) point 574 out that socio-economic status differentiation becomes most evident in this context of heightened north-Central Andean interaction. The rich Early Horizon tombs excavated at Cerro Blanco, Poro-Poro and Shillacoto may represent interment of a socio-economically privileged class. Ultimately, it is upon the north-Central Andean Early Horizon interaction Sphere that the Chavin horizon is superimposed. At Manachaqui Cave, maize first appears within this increasingly politicized field of interaction that emphasizes public gathering and ritual at monumental centers. It may be that Suitacocha Shape F jars with tall, spout-like polished rims are associated with maize and chicha. Mohr-Chavez (1981b:345) observes that Marcavalle Phase C-D jars "behave as bowls" in prestige spheres, and perhaps within south-Central Andean gift exchange. More evidence is necessary to test these functional associations. There are indications that Initial Period east-west interaction networks conjoining the Central Andes and Amazonia persisted during the Early Horizon. Lathrap has discussed specific similarities between Kotosh-Kotosh and Shakimu styles (1971:87-88), and Mohr-Chavez (1981b) cites attributes shared between Marcavalle B-D, eastern-slope Cochabamba and lowland Amazonian styles. Other highland Central Andean assemblages like Urabarriu at Chavin de Huantar and the Huaricoto style at Huaricoto lack evidence of Amazonian affiliation (Burger 1992:154-155; 1985b:529). 575 Again, interchange with Amazonian Central Ucayali societies is apparently restricted to the upper Huallaga and Urubamba drainages. Material patterning indicative of these relationships and with the Suitacocha-Chorrera linkage suggests that Early Horizon interregional interaction was more narrowly focused, or point-specific than Initial Period interaction. Marcavalle A-D short-necked ollas and jars with zoned punctation on vessel shoulders provide suggestive evidence of a continuing Northern Andean connection in the southCentral Andes. Interregional exchange networks that distributed carved stone mortars like those found at San Isidro, Cotocollao, Huayurco, Bagua, Pacopampa, Kotosh and Marcavalle (Rojas 1969; Lathrap 1970:108; Peterson 1984; Zeidler 1988) probably trended north and south along the eastern face of the Andean cordillera, and trade in strombus and spondylus shells also followed similar over-land routes (Hocquenghem 1993). The same kind of decorated short-necked jars occurring in the south-coastal Rio Grande de Nazca Valley Tajo style (Silverman 1994: Figs. 7, 8) suggest similar affiliations although Silverman proposes Chavin horizon dates for these vessels. Of course later Paracas assemblages from this area are well-known for exhibiting Amazonian characteristics (Tello 1942:631-632). The Chavin Horizon Hiatus During the mid-Early Horizon, north-Central Andean 576 societies were united within a pan-regional sphere of interaction corresponding to the Chavin horizon. After 500 b.c., Manachaqui Cave seems to have fallen into relative disuse for two or three centuries. It might be hypothesized that societies in the northeastern montane forests were excluded from the Chavin interaction sphere, but additional evidence from the Chachapoyas area is needed to be sure. What is clear is that Manachaqui Cave no longer serviced systematic, or even frequent interregional interaction. Simultaneous with the spread of the Chavin cult, Burger (1984a; 1992:216) observes that a dearth of evidence for interregional interaction with the Northern Andes and Amazonia indicates the solidification of cultural boundaries surrounding the Central Andean or Peruvian "co-tradition." As interacting Central Andean societies turned inward, the gateway community at Nafiafiique was virtually abandoned, Bagua re-directed interaction toward other eastern lowland areas and the ancient tradition of highland-lowland economic interaction evident in the early Kotosh assemblages disappeared (Burger 1992:219). Likewise, Manachaqui Cave's role as a node in gateway interaction terminated until approximately 200 b.c. The beginning of the Early Intermediate Period marks the formation of new interaction spheres functioning in increasingly politicized contexts. Radiocarbon evidence from Gran Pajaten attests to the nucleation of populations at hilltop settlements within the 577 montane forest of the Abiseo drainage, probably for the express purpose of facilitating renewed interregional interaction (Church 1994:292). Colpar Phase Evidence for Interaction During the Early Intermediate Period Colpar Phase, Manachaqui Cave and Gran Pajaten yield evidence for occupation by local populations participating in the same interaction spheres. To judge by Paste A's limited repertoire of vessel shapes and surface decorations, and evidence provided by Pastes B6 , B8 , B10 and C1 , the study area's clearest connections are to montane forest societies east and north of Manachaqui Cave (Fig. 120). Preceding discussions of the Manachaqui and Suitacocha Phase data have demonstrated that this northwest-focused interaction sphere is ancient. The austerity of the Colpar Paste A assemblage, the lack of broadly distributed montane forest styles, and especially the reduced repertoire of surface decorations, indicates that montane forest populations around the study area remained largely outside of the realigned and more rigidly-bounded post-Chavin Central Andean interaction spheres, interacting primarily among themselves. Evidence from Manachaqui indicates that the rockshelter was still utilized expediently as a wayside station by infrequent travelers. Like their Suitacocha assemblage predecessors, the small Colpar Paste A necked-jars and 578 restricted bowls were well-suited for transport by individuals shouldering cloth bags. Camelids are clearly in evidence by this time, yet there are no clear indications that they were utilized as cargo animals, at least in any large-scale fashion. The lack of evidence for carnelid transport systems coincides with the paucity of ceramic evidence for study area interaction with neighboring Central Andean societies that had already adopted camelid transport technologies (Miller and Burger 1995). By the beginning of the Early Intermediate Period, populations in the far north-Central Andean highlands producing the Layzon style had developed systematic interaction with coastal Salinar societies (Mujica 1984}, probably utilizing llama caravans to transport commodities (Sechtin 1986:33) including bulky pottery vessels. The Colpar Paste A assemblage's formal resemblances to the Cajarnarca Layzon style are more likely residual than indicative of habitual or systematic interaction. In neighboring Huamachuco, land-use studies by McGreevy and Shaughnessy (1983) determined that evidence for intensive herding activities in the surrounding punas is scant until the Late Intermediate Period (after a.d. 900). The absence of corrals and the low frequency of carnelid bones in Early Intermediate Period faunal assemblages not only suggest that large herds were not kept (Ibid.:241), but that carnelid transport was unimportant to Huarnachuco's Early Intermediate 579 Period regional economy. During the Colpar Phase, foot traffic maintained contact between the Pataz-Abiseo area and populations with access to warm-water estuarine environments of the interAndean littoral. From mixed deposits, fish remains that Kent assigns to the Gobidae family of sleepers probably represent long-distance interaction with groups situated between the Piura and Guayas River basins, as do Pastes B7 ~1d B9 • Previously overlooked similarities between Pre- Abiseo Phase sherds from Gran Pajaten and finds in the inter-Andean area include an unusual modeled rim (Church 1994: Fig. 11m) identical to a Sechura B example (Lanning 1963: Fig. 231), and a body sherd with circles rendered by the negative-resist technique (Church 1988:41f). The circulation of beveled-rim bowls apparently transcended Central Andean boundaries and followed expansive networks established prior to the Chavin horizon. Evidence from Huacaloma and Pandanche intimates that these bowls of Ecuadorian origin began to first circulate at northern montane forest centers around the time of the Chavin horizon. By the post-Chavin Colpar Phase, the network had expanded to include the upper Zafia and Piura valleys, the upper Utcubamba, the upper Montecristo (Abiseo), Huamachuco and the upper Santa valley Cabana-Corongo area. The bowl shape persisted in the Northern Andes during the coeval Early Guangala Phase. In effect, the beveled-rim bowl's 580 distribution replicates preceding Early Horizon northCentral Andean interaction spheres. A functional interpretation is difficult without more information on the specific contexts of these finds, but the bowls were probably carried by travelers engaged in continued gift exchange, perhaps with political nuances. Empedrada Phase Evidence for Interaction The mid-Early Intermediate Period Empedrada Phase transpired against a background of hardened Central Andean political boundaries, inter-group conflicts, restricted travel and the systematic exchange of goods by llama caravan transport (Burger 1992:229). The comparative analyses presented in Chapter 9 demonstrated that mid-Early Intermediate Period Pataz-Abiseo area pottery styles shared a remarkably close relationship with Recuay styles from the upper Santa tributary Cabana River (Fig. 121). Overall resemblances between the Empedrada Paste A and Pashash's Recuay assemblages are no less striking than the well-known correspondences between pottery assemblages at highland Layz6n and coastal Cerro Arena. Pataz-~~iseo Additional survey in the and Cabana areas is needed in order to determine just how geographically specific this relationship is. More likely than verticality to account for the Pashash-Empedrada relationship, and the dense late 581 prehistoric Montecristo Valley settlement distributions described in Chapter 3, is the access that the Abiseo drainage offers to the lowland Huallabamba fluvial "axis" which brought Amazonian riverine trade to the doorstep of the north-Central Andes. The seminal event that precipitates the birth of this exchange linkage and the beginning of the Empedrada Phase appears to be the introduction of systematic llama caravan transport technology in the study area. Evidence for its debut at Manachaqui Cave consists of elevated quantities of camelid remains, and unprecedented shifts in the Paste Group A vessel shape inventory. The strong correspondences between Pashash and Empedrada pottery, coupled with the lack of evidence for herding at Manachaqui Cave and in the surrounding Tropical Alpine Zone, invites speculation that most of the llamas and the herders frequenting Manachaqui Cave during the Empedrada Phase were native to the Callej6n de Huaylas. The notion that "foreign" Central Andean societies widened their interaction networks by directly introducing llama-based transport technology into adjacent areas has important local and regional implications. Further exploration of this hypothesis at the local level requires more rigorous comparative analyses of Pashash, Empedrada and Gran Pajaten (pre-Abiseo Phase) ceramics and continued study of prehistoric study area land-use. In this case, the 582 Ernpedrada Phase Paste C assemblage may represent the physical presence of upper Santa Valley highlanders at the rockshelter. Although the Empedrada Phase Paste Group A assemblage clearly exhibits attributes that are locally derived, it does contain more than a single ware. It may be significant that Gran Pajaten's small pre-Abiseo Phase sample does not exhibit some important Empedrada Paste A attributes like applique decoration and Shape H vessels. Nor does it share other Ernpedrada Paste A affinities to the Pashash styles. New east-west interaction linkages utilizing cargo animals required the construction or refurbishment of paved roads in the Pataz-Abiseo area. It remains unclear how far into the montane forest llama caravans penetrated, but traces of paved roads observed in the forested Montecristo Valley apparently reflect linkage between Manachaqui and Gran Pajaten. Also, recall from Chapter 3 that Savoy (1970) located paved roads far below in the Central Huallaga subtropical forests surrounding the abandoned mission of Jesus de Pajaten. The lowland Huayabamba Complex at the eastern end of the Pataz-Abiseo study area cross-dates to the Early Intermediate Period, and its Central Andean aspect clearly reflects the linkage of lower Abiseo Valley societies into Central Andean interaction spheres. Nascent highland-lowland exchange systems incorporating Chachapoyas intermediaries cross-cut the ancient and 583 persistent interregional interaction sphere trending northsouth along the eastern face of the Central Andes. This "longitudinal" interaction sphere is evident in the Empedrada Paste A assemblage's similarities to the nearby pre-Abiseo phase assemblage directly east, the Tarma valley Malarnbo and Camonal complexes to the south and the Cancharin (at Cuelap) and Cosanga-Panzaleo styles to the north. The Huacrapuquio sherds and Quispisisa obsidian almost certainly reached Manachaqui Cave via eastern montane forest routes, as did the Paste B11 Panzaleo, or Panzaleo-like, pottery which originated from twice as far away. The large jar with orifice diameter of 30 em (Fig. lOOf) was probably transported to Manachaqui Cave by llama caravans on roads long ago lost beneath montane forest regrowth. Boundary Interaction and the Central Andean Tropical Montane Forests The previous section provided an interpretation of evidence for interregional interaction from Manachaqui Cave during nearly two thousand years of prehistoric occupation. The following paragraphs will take a broader view of evidence for interregional interaction in an effort to infer specific modes of interaction responsible for stylistic patterning in and around the Central Andean montane forests. The failure to find support for postulated society-level population movements at Manachaqui Cave is predictable if we consider the assumptions inherent within unitary origins 584 frameworks that material culture and styles react passively to the passage of time as do languages and gene pools. Archaeologists narrowly regarding the stylistic distributions as reflections of "cultural heritage" await the eventual discovery of evolutionary "stepping stones" (Meggers et al. 1965:174), and grandmothers" "collateral relatives," "mothers (Lathrap 1970:107, 1974:142), unknown donors (Braun 1982) and other missing links to corroborate their interpretive models. While in certain contexts style may behave passively as it is learned and re-learned from generation to generation, scholars investigating how style functions in broad contexts of inter-group interaction now recognize style's proactive functions in "informationexchange" (e.g. Wobst 1977; Hodder 1982; Weissner 1990; Hegmon 1992) . Material patterning previously cited as evidence for migration can be explained alternatively as conscious manipulation of styles and style attributes for the furtherance of a variety of social, religious, economic and/or political ends. This thesis assumes the position of Lightfoot and Martinez (1995) that the "ambiguity" in material patterning that attends processes of "boundary arbitration" is itself substance for productive investigation: "This ambiguity in material culture along frontiers [i.e. boundaries] provides an ideal opportunity for archaeologists to study the process of creolization - 585 how people modify, create and syncretize material objects in culture contact situations" (Ibid.:482). Thus, to account for Central Andean material patterning ostensibly supportive of migration hypotheses, the investigative emphasis shifts from study of the mobilization of people to study of the mobilization of style. Archaeological evidence recounted in the following paragraphs demonstrates that societal boundaries at the margins of the Central Andes were loci of stylistic innovations at crucial moments in prehistory. The term "boundary interaction," might be utilized to refer to inter-group interactions across mutually recognized social or socio-political boundaries that involve the overt manipulation of material culture, and especially stylistic innovation. The term "innovation" is employed here following Hantman and Flog's (1982:240) definition of the process as a systematic recombination of style attributes or attribute clusters. Of course all human interactions involve boundaries of some form or other, and most can be accompanied by some intentional manipulation of signs and symbols. At the edges of the Central Andes, however, the theoretical development of a boundary interaction concept can provide a useful interpretive tool for archaeological investigations because of: 1) the ancient and persistent coincidence of cultural boundaries with the major highlandlowland ecotones, 2) the unusually sharp juxtapositions of disparate cultural traditions, especially between Andean and 586 Amazonian areas, and 3) the particularly abundant archaeological evidence for style innovation in and around the Central Andean boundaries. Discussions of Andean and Amazonian cultural boundaries, the Andean-Amazonian ecotone and the contrast between Andean and Amazonian cultural traditions have been presented by numerous scholars including Lyon (1981), Raymond (1985), and LeMoine and Raymond (1987). The presence of culturally "intermediate," yet unique, autochthonous eastern montane forest populations rendered the pre-Hispanic juxtaposition of Andean and Amazonian cultural traditions far less severe than it appears today. Archaeological studies in the eastern montane and premontane forests have uncovered numerous instances of particularly obvious manipulation, or "hybridizing" of styles (e.g. Allen 1968:350, 352; Raymond 1976:209; Hastings 1985:589; Church 1994:290). Stylistic innovation characteristic of boundary interaction at the Central Andean margins can be classed into three closely related types identified during the course of this analysis. The first and most basic kind of style innovation has been repeatedly referred to in literature as style "hybridization." I prefer the terms "amalgam" and "amalgamation" to "hybrid" and "hybridization" to avoid an unintended biological analogy. Simple style amalgamation results in a syncretism of two or more styles 587 from identifiable sources. The second kind of style innovation might be termed "style displacement" because it is a particularly sudden and severe instance of style amalgamation. The third style innovation is "style amplification" because it involves style amalgamation in which selected foreign design attributes are deliberately amplified or exaggerated. These kinds of style innovation presumably accompany many of Renfrew's (1975} exchange," and perhaps Salomon's (1985} "modes of "institutions of Andean complementarity," but simple correlations between stylistic similarities and exchange relationships cannot be uncritically assumed (Hantman and Plog 1982}. One potentially important mode of boundary interaction not discussed by Lightfoot and Martinez (1995}, nor fitting comfortably within any of Renfrew's "modes of exchange," is the gateway community or gateway system (Hirth 1978}. Hirth's description (1978:37-38} emphasizes the importance of local geography, of major environmental and cultural boundaries, and of gateway interaction's pivotal role in the emergence of increasingly complex social and political formations: These communities flourish at the passage points into and out of distinct natural or cultural regions and serve as 'gateways• which link their regions to external trade routes ... They generally are located along natural corridors of communication and at the critical passages between areas of high mineral, agricultural, or craft productivity; dense population; high demand or supply for scarce resources; and, at the interface of different technologies or levels of sociopolitical complexity ... Gateway communities operate as commercial middlemen involving 'wholesale' activities. 588 Hirth also notes that gateway communities are often autonomous. Of particular importance to the interpretations offered within this thesis is Hirth's (1978:43) observation that mechanisms to express "cultural affiliation" between gateway communities and the societies situated along external supply lines are important to the network's functioning. Apparently, gateway communities and systems begin to appear at the Central Andean margins as long-distance, interregional exchange assumes increased importance to developing socio-political institutions during the Early Horizon. In theory, gateway communities should be a locus of style innovation, as information-exchange through style amalgamation is apparently a common mechanism for expressing cultural affiliation. Archaeologists examining problems related to style as information-exchange explicitly or implicitly concur that establishing the contexts of use for artifacts is an imperative step in formulating viable functional interpretations of style. In Chapter 10, it was determined that Manachaqui Cave is best understood as a wayside station, perhaps utilized on occasion for habitation by local populations. As such, the rockshelter serviced travelers engaged in interregional exchange during at least part of each phase, and associated pottery assemblages may therefore be viewed as active media for information- 589 exchange. The early Paste A assemblages are largely portable and thus, appropriate vehicles for conveying information germane to inter-group exchange protocol. This is perhaps even more true of Paste B and C assemblages which, by virtue of their intrusive presence, must have conformed to acceptable norms of protocol. In the following paragraphs, I hope to demonstrate that some pivotal events in the north-Central Andean prehistoric sequence previously attributed to population intrusions can be more accurately interpreted as local developments accompanied by style innovation in contexts of boundary interaction. These events include early population nucleation, initial acquisition of pottery and the early emergence of socio-economic stratification. They support Lightfoot and Martinez's (1995} assertion that cultural boundaries are particularly dynamic social environments where the manipulation of material culture attends the negotiation and re-negotiation of identities within and between groups. Because the spread of the Chavin cult had such a profound affect on interaction and boundaries, I will divide the following discussion into pre-Chavin horizon and post-Chavin horizon considerations of the evidence. Pre-Chavin Horizon Boundary Interaction Migration hypotheses spotlighting Amazonian sources and a Marafion valley entryway are not served by the late Initial 590 Period arrival of pottery at Manachaqui Cave, although opening dates for the Manachaqui Phase remain somewhat obscure. Interaction with pottery-using Northern Andean populations most likely precipitated the "dependent invention" (as described by Clarke and Gosser 1995) of ceramic technology on the northern slopes of the Central Andes. The linkage of the Manachaqui Phase Paste A pottery to this hypothesized dependent invention obviates the search for the style's "collateral relatives" and other common ancestors. The process presupposes rapid stylistic elaboration to facilitate entry into interaction relationships in which pottery was already functioning symbolically as a common medium. The acquisition of pottery technology among montane forest societies probably spread laterally along, rather than vertically up and down the Central Andean slopes. Although much montane forest exchange must have included small-scale, kin-based reciprocal exchange of "complementary" subsistence goods from varied ecological contexts, the verticality framework cannot account for the transverse dissemination of ceramic technology and common design attributes around the northern rim of the Central Andes. To account for the slope-specific configuration of the Northern Montane Forest Ceramic Interaction Sphere, models invoking limited subsistence-related transhumance and longdistance exchange would seem most appropriate. Hantman and 591 Plog (1982) implicate low population density as the governing factor responsible for similar homogeneous stylistic distributions in the U.S. southwest. However, an a..Tl.cient montane forest tradition of long-distance "lateral" exchange of primarily non-subsistence commodities along the Andean slopes constitutes a viable historical explanation for the broad and stylistically homogenous distribution presently observed among montane forest ceramic assemblages. Despite the difficult terrain traversed, modes of interaction approximated linear networks of "down-the-line" or "middleman" exchange like those ethnographically documented in Amazonia (Lathrap 1973a:172-173; Thomas 1981). Interacting groups either possessed access to highly localized resources or specialized in the production of perishable commodities. Osborn's (1989) ethnographic description of the mobile, eastern slope-dwelling Tunebo of Colombia suggests that the configuration of early montane forest interaction spheres responded to an absence of fixed lateral cultural boundaries along the slopes, relative to rigid and exclusive altitudinal boundaries drawn along vertical axes. Evidence from early western montane forest sites of Nanchoc (Dillehay et al. 1989) and Montegrande (Tellenbach 1981) may contradict hypotheses of high mobility. For most of the Central Andean montane forests however, a model combining limited transhumance and longdistance trade as basic features seems most appropriate in 592 the absence of additional data. In previous discussions I suggested that the lightweight, but well-fired Manachaqui Phase Paste A restricted vessel shapes were ideally suited for transport. These same characteristics describe other northern montane forest assemblages which are likewise well-adapted for use during activities requiring a degree of mobility. In other words, serviceability for transport may have been an important factor conditioning formal and technological aspects of the similar and widely-distributed northern montane forest styles. The everted rim stands out as a particularly useful feature for securing a vessel's contents with a fastened lid. be viewed as bo~~dary The northern montane forest styles can amalgams combining practical considerations for mobility with symbolic expressions of multiple exchange alliances with neighboring highland, montane forest and lowland societies. At the eastern edge of the Central Andes, different processes would seem to account for the appearance of the Initial Period Wairajirca pottery style at Kotosh. Lathrap's characterization of the Wairajirca Phase style as an intrusive welding of Central Andean and Amazonian ceramic traditions (Lathrap 1970:106} does not recognize a local, eastern slope stylistic component. Of the ~mall number of attributes shared between the Wairajirca and Manachaqui styles, the short-necked jar shape with reinforced rim is 593 most significant because, between Manachaqui Cave and Cuzco, it occurs only on the eastern slopes. Rather than a Type A3 site unit intrusion (Lathrap ed. 1956:15-18; Lathrap 1971:94), Wairajirca pottery can be viewed as a local montane forest assemblage demonstrating stylistic amalgamation in a context of boundary interaction. Clarke and Blake's (1994) model describing pottery's dependent invention within a Mesoamerican Formative Period context of factional competition may provide a useful analogy to explain the acquisition of pottery at Kotosh and the Wairajirca amalgamation of Amazonian style attributes. The Wairajirca style assemblages, especially those illustrated by Kano (1972, 1979) from Shillacoto, convey lowland Amazonian symbology even more explicitly than their presumed Tutishcainyo prototypes. The Wairajirca case demonstrates an exaggeration of foreign style attributes or "style amplification" that Lightfoot and Martinez (1995:485) would contend is not uncommon under conditions of competition between social segments in proximity to cultural boundaries. As in Chiapas, we might hypothesize that dependent invention occurred in the context of competitive attempts by one or more social factions to enhance their position by adopting a foreign technology and symbology for ostentatious use. In Chiapas, and probably at Kotosh, the manipulation of technology and style by emergent leaders announces exclusive access to foreign alliances, special 594 esoteric knowledge and supernatural power in a manner ethnohistorically and ethnographically documented in Central America, Amazonia and cross-culturally by Helms (1991, 1993) . While there is little direct evidence for social organization among Wairajirca Phase upper Huallaga Valley populations, Burger and Salazar-Burger (1986:77) have suggested that the more elaborate temples at Kotosh and Shillacoto were constructed with corporate labor directed by "recognized leaders" wielding coercive powers, and that some Kotosh temples appear to have been operating simultaneously. A trend of increasing temple size during the Preceramic Period Mito Phase culminates in the appearance of both Wairajirca pottery and the inordinately large temple at Shillacoto. Unusually elaborate, but mostly looted, Shillacoto Initial Period tombs (Izumi et al. 1972) were apparently prepared to inter these emergent elites. La Galgada provides another example of a ceremonial center with elaborate temples growing steadily through the late Preceramic and Initial Period transition (Grieder and Bueno 1985; Burger and Salazar-Burger 1986). Pottery again arrives in a context of architectural elaboration and comparatively rich interments. While Grieder (1988:190) finds no Amazonian aspects in the Initial Period assemblages, a relationship to northern montane forest styles was established in Chapter 6 of this thesis. Again, 595 we might hypothesize that ceramic technology and design attributes were being consciously manipulated to advance the position of competing social groups. Undoubtedly, important differences between the Chiapas, Kotosh and La Galgada processes would be revealed by more penetrating analyses beyond the purview of this thesis. Late Manachaqui Phase remains associated with the Initial Period "flash horizon" around 800 b.c. may be the earliest direct evidence for a long tradition of Central Andean pilgrimage. By now the rockshelter had begun servicing travelers, traders, religious specialists and other pilgrims, entering and leaving the Central Andes. By the Early Horizon, broad fields of style-sharing in montane forest interaction spheres had fragmented into competitive nodes within an expanding interaction network. Rather than reflecting migration, the Suitacocha Phase Paste A assemblage exhibits style displacement consistent with boundary conditions (Lightfoot and Martinez 1995:478) where positioning at a cultural interface augments a society's options for rapidly changing its social, economic and political alliances. The same dynamics may have been responsible for similar style shifts at Bagua that Shady (1987b:464) attributes to population invasions. The Suitacocha Phase amalgamation of styles originating in the Northern Andean and Oriente lowlands constitutes unprecedented evidence of a close articulation between 596 Central Andean Early Horizon and Northern Andean Late Formative (i.e. Chorrera) interaction spheres. Some archaeologists working in the north-Central Andes (Hocquenghem et al. 1993:253-254) have recently pointed out correspondences between the two regional chronologies, and postulated exchange relations facilitated by border intermediaries. At this time pottery styles in the two regions do share numerous formal and decorative features. Some featureshave temporal priority in one or the other region, while others may have appeared simultaneously during a synergistic co-development of the interaction spheres. Perhaps most illustrative of the amalgamation of Chorrera attributes evident in the Suitacocha style is the application of unmistakably "Ecuadorian" motifs and iridescent paint to the preeminent Paste A bowl shape (Shape C Rim 7 in Figs. 71t, u; 72a). The only Central Andean bowls comparable to the Suitacocha Phase carinated variety are Montegrande's Shape A7 bowls, and those too are likely derived from interaction with the north. Additional Central Andean amalgamation of Chorrera attributes can be discerned by comparing Montegrande's unusual Shape A4 bowls (Ulbert 1994: Tafel 10 top, 15 top) with Penon del Rio's Forms 9 and 14 (Nieves Zedefio n.d.: Figs. 17d, 19c,d, 24a) from the Guayas Basin. Kaulicke (1981: Abb. 8:5-14 Shape 5) has reported the same shapes from Pandanche, although they have not yet been reported from Huacaloma. Perhaps the most 597 "Chorreroid" Central Andean assemblages are found on the Andean slopes. At first glance, an absence of chonological agreement between the Suitacocha Phase, the Chorrera horizon and ostensibly earlier Cajamarca styles appears to contradict the preceding arguments for stylistic relationships. However, the Montegrande assemblages are probably later than the single 1800 b.c. thermoluminescence date (Ulbert 1994:150). Likewise, the Pandanche sequence begins no earlier than Kaulicke's 1490 ± 340 and 1390 + 340 b.c. radiocarbon dates. Suitacocha Paste A's relatively drastic amalgamation of Chorrera and Upano style attributes can be interpreted not only as a symbolic expression of exchange alliance with Northern Andean or Oriente groups, but as an instance of "style amplification" perhaps intended to offset the competitive edge held by Cajamarca groups situated closer to northern "suppliers." Again, verticality cannot account for regional demography or the absence of stylistic correspondence between the Paste A and adjacent lowland and highland coeval assemblages. Suitacocha Phase Paste A stylistic evidence suggests that study area societies were fully engaged in the mediation of north-south longitudinal exchange. Exchange mechanisms may not have differed much from prior "down-theline" or "middleman" modes, but regional stylistic distributions were probably more powerfully influenced by 598 the exchange of ritual paraphernalia and status markers by "home-base reciprocity exchange (Renfrew 1975:41). The flash horizon and Suitacocha Phase Paste B assemblages likely constitute direct evidence of long-distance expeditions (Lurnbreras 1993:364) that characterized Early Horizon interregional exchange networks. During the Suitacocha Phase, Manachaqui Cave apparently served as a conduit channeling interregional travel between northern lowland societies and Central Andean ceremonial centers connected within a web of intensifying and increasingly politicized ritual exchange of gifts. Scholars (Lathrap 1971; Lathrap et al. 1975; Raymond 1988:295; Burger 1992:102-103) have repeatedly observed that Early Horizon vessel shapes utilized for Central Andean ritual like the stirrup-spout bottle originated in the northern and eastern lowlands. Burger notes additionally that cylinder and stamp seals, napkin ring ear spools and modeled anthropomorphic and zoomorphic bottles found in north-Central Andean mortuary contexts "may have been inspired by Ecuadorian cultures." Lack of fieldwork around suspected boundaries and inadequate chronological control in both regions may be masking the impact of the articulation between Early Horizon and Chorrera interaction spheres in both areas. Ultimately, the north-Central Andean network of "peer polity interaction," many of its material accoutrements and a tradition of long-distance travel and pilgrimage were 599 subsumed within the Chavin horizon interaction sphere. The tropical lowland's "upward thrust" identified by Tello and Lathrap in Chavin de Huantar's iconographic expression has been explained by Burger (1992:155-156) as an intentional manipulation of foreign, esoteric images and symbols. That the overt manipulation of Amazonian imagery is limited to this single site during the mid-Early Horizon punctuates the focal position assumed by Chavin de Huantar within regional interaction networks. At this time Central Andean boundaries were not "closed" so much as neglected, as interest in foreign commodities was eclipsed by interest in the regional cult. Post-Chavin Horizon Boundary Interaction Manachaqui Cave seems to have infrequently served local populations for some seven to nine centuries during which time the earliest Central Andean expansionist states emerged, and boundaries became increasingly rigid and politicized (J. Topic and T. Topic 1983a, 1985; Burger 1992). During the Colpar Phase, at least some of the Paste Group B and C vessels were likely carried to Manachaqui Cave by travelers moving relatively freely around the northern and eastern peripheries of the bounded Central Andean interaction spheres. The Colpar Phase assemblages generally attest to a lack of intense boundary interaction and style innovation in the Pataz-Abiseo area. 600 By a.d. 400, Manachaqui Cave had entered into the service of an interaction alliance between highland Central Andean, local intermediary montane forest and lowland Amazonian societies involving llama caravan transport. Southern Chachapoyas societies assumed the role of montane forest intermediaries in linear interregional interaction. For the burgeoning Recuay political economy, this longdistance exchange served a need not so much for maize and coca which could be grown closer to home, but for exotic goods and information produced and traded by lowland Amazonian specialists. Raymond's (1988) list of lowland exotica likely traded to Central Andean highlanders during prehistoric times includes tapir's feet, feathers, pets, and hallucinogenic and curative herbs. Amazonian lowlanders facilitating this east-west exchange were border intermediaries (cf. Reeve 1994) on the Huallabamba providing a crucial link to Huallaga river trade. The research presented in Chapter 3 suggests that salt and valued minerals were also exchanged. Highland stone and metal tools, in great demand by lowlanders during the 16th and 17th centuries (DeBoer 1981), were likely traded deep into Amazonia during prehistory. The Recuay relationship with the Pataz-Abiseo area was probably severed by Huari incursion into the upper Santa Valley around a.d. 700. Within this scenario of east-west oriented interaction, the role of groups in the neighboring 601 Huamachuco area remains somewhat enigmatic. While the amalgamated Ernpedrada Paste A style can be thought of as the local expression of boundary shifts and incorporation within an expanding Central Andean interaction sphere, the assemblage also provides evidence of a pivotal shift in Andean transport technology. Changes in formal attributes signify new socio-economic alliances but, as in the case of Manachaqui Paste A, the ceramic technology is conditioned by the constraints of a new transport technology. Larger globular vessels made resistant to breakage by thickening the walls may have been judged most appropriate for llama transport. As the larger Ernpedrada Phase Paste A assemblage jars are no longer portable for long-distance foot transit, they likely ceased to function as primary media for informationexchange. This austere assemblage was given relatively little decorative attention. Paste B and C assemblages, on the other hand, continue to convey information related to exchange protocol. In fact, it seems likely that the primary function of protocol is responsible for the high degree of stylistic homogeneity among kaolin-ware pottery traditions across the Early Intermediate Period northCentral Andean highlands (T. Topic 1987:9). The so-called Recuay "positive-painted" kaolin wares in particular appear to be functionally specialized for exchange protocol as their spatial distributions extend well beyond the upper 602 Santa Valley into coastal and montane forest areas. These vessels bear relatively innocuous geometric decorative motifs to the exclusion of militaristic and mythological imagery common in the upper Santa Valley Recuay homeland. While Paste A style displacement at the onset of the Empedrada Phase inaugurates southern Chachapoyas' entry into Central Andean interaction spheres, a similar transition seems to have occurred earlier in northern Chachapoyas where Central Andean neighbors in Cajamarca had adopted llama transport technologies during the Layz6n Phase. Not only are resemblances between Huayabamba and Empedrada shapes clearly evident, but stylistic correspondences are startlingly precise between the Huayabamba Complex and Early Intermediate Period Cajamarca non-kaolin vessel shapes, especially given the small Huayabamba sample. Huayabamba Form A is comparable to Cajamarca Light-colored Form 1 jars (Onuki and Terada 1982: Pl. 92:1-6) and Cajamarca Coarse Red Form 2 jars (Ibid.: Pl. 94:2-10). Ravines' Forms c and D have close highland counterparts in Cajamarca Light-colored Forms 3 and 4 jars (Ibid.: Pls. 92:14-17 and 93:1-13). The unusual modeled profiles of the Cajamarca "upturned" and "irregularly thickened" rims are identical to the Huayabamba vessel rims. Surface decorations based upon applique and incision are also common to the two assemblages. Applique and incised decorations are likewise typical of Cancharin Phase styles documented in the upper Utcubamba 603 Valley of northern Chachapoyas (Ruiz 1972). The Huayabamba Complex style amalgamation can be interpreted as a reflection of systematic, linear interaction between Cajamarca and the Huallabarnba waterway by way of Chachapoyas intermediaries that controlled the mountain passes between Cajamarquilla (now Bolivar) and Leimebarnba. Reichlen and Reichlen (1950:241-242) who worked in both areas first drew attention to the stylistic correspondences between Cajamarca and northern Chachapoyas ceramic assemblages. This east- west exchange relationship probably began by the beginning of the Early Cajamarca Period around a.d. 200 (Matsumoto 1993:188), and lasted with varying degrees of intensity through the Inca-dominated Late Horizon. Such east-west exchange systems ultimately left fine kaolin-ware bowls distributed widely across northern and southern Chachapoyas. Evidence for the Cajamarca-northern Chachapoyas relationship is particularly contrary to the expectations of verticality frameworks because it demonstrates specific highland interest in Amazonian resources, despite the close proximity of tropical forest resources west and north of the Cajamarca Basin. In fact, the archaeological data show that Early Intermediate Period highland polities already controlled lowland ecological zones on the western slopes (J. Topic and T. Topic 1985; Proulx 1982). In this scenario, the eastern premontane and subtropical forest style amalgams like Huayabarnba, Camonal and Simariba 604 conceivably represent societies with strategic access to lowland riverine exchange networks as border intermediaries. Archaeological data from the Early Intermediate Period Huallaga Basin attest to similar processes of technoeconomic change at the edge of the montane forests farcher south. The Kotosh-Higueras Phase faunal assemblage (Wing 1972: Table 5) exhibits a sudden rise in the relative abundance of camelids from 43 percent during the prior Sajara-patac Phase to 72 percent. Despite Wing's potentially conflicting interpretation of camelid utilization, the coincidence of this shift with the dramatic changes in ceramic technology and settlement patterns described by Lathrap and Isbell suggest a close parallel to events in the Pataz-Abiseo area. The Higueras "Coarse Brown" ware vessels are morphologically analogous to Empedrada Phase jars and bowls, while the sudden shift to hilltop settlement locations above the upper Huallaga likely corresponds to the increased importance of securing improved, secure access to high-altitude pasturage for cargo animals. From the Huanuco Basin, the head of Huallaga River canoe navigation near Tinge Maria can be accessed (see Chapter 3). In the far north-Central Andes, Sechura E (Lanning 1963:209-210) and Tamarindo C (Kaulicke 1991:417) assemblages may likewise represent the same technoeconomic transition to llama caravan transport seen in the Empedrada Phase Paste A assemblage. However, Lanning and Hocquenghem 605 et al. {1993:455) implicate expansion of the adjacent Mochica empire. In Cajamarca, the Layz6n Phase shift to llama caravan transport apparently predates the widespread upward settlement relocation that occurred during the Initial Cajamarca Phase {Sechtin 1986). Browman {1989:263) asserts that "the shift from a meat-orientation to a cargo/wool orientation does not occur until the Middle Cajamarca Phase c. A.D. 500." He cites M. Shimada's report on Cajamarca faunal remains, but Shimada's findings are inconclusive {M. Shimada 1985:289). Poor chronometric control and other problems frustrate attempts to determine whether comparable shifts in ceramic technologies, settlement patterns and animal utilization documented elsewhere are approximately synchronous and indicative of similar local and regional processes. The evidence from Huamachuco suggests that such changes were complex and uneven across the north-Central Andes. Huamachuco's prehistoric occupations had been situated on promontories since the Early Horizon {J. Topic and T. Topic 1986: Fig. 2, Cuadra 9; Zaki 1983) but, as previously noted, archaeological data from the area indicate that camelid pastoralism was unimportant until much later {McGreevey 1989) . Eastern montane forest modes of exchange during the Early Intermediate Period can be effectively characterized utilizing Hirth's gateway model. The interpretation of 606 Andean yunga settlements as uniquely permeable boundary sites functioning in the context of interregional interaction is not unprecedented. At the lower Meche Valley Early Intermediate Period site of Cruz Blanca, J. Topic and T. Topic (1983a} documented highland kaolin pottery amounting to 20 to 50 percent of assemblage totals, cooccurring with pottery from the upper western slopes near Otuzco. · Local pottery shows attributes diagnostic of the coastal Gallinazo polity, but the site should not necessarily be interpreted as a Gallinazo-controlled node for that reason. Apparently Cruz Blanca served as a western slope Andean gateway community, and provides an analogy for the interpretation of eastern montane forest sites and settlement systems. Comprising an estimated 16 percent of the total assemblage, Manachaqui Cave's Empedrada Phase large sample of Paste C pottery attests that the rockshelter served to channel interregional interaction as part of a gateway system. Deep within the adjacent montane forest, the Early Intermediate Period settlement at Gran Pajaten can be characterized as a gateway community functionally analogous to Cruz Blanca. Foreign kaolin ware pottery from the adjacent highlands comprises seven percent of the pre-Abiseo Phase assemblage from Building No. 1's fill deposits (Church 1994}. The gateway model may serve as well to approach an understanding of late prehistoric northern and southern 607 Chachapoyas settlement patterns, but only continued research can evaluate the effectiveness and weaknesses of such models. In the central subdivision of the eastern montane forest, the Tarma Valley Tranca Phase settlement system conforms to some expectations of the gateway model. Apparently, reciprocal exchange and redistribution of local produce took place at the paramount settlement of Tranca (cf. Renfrew's 1975:48 central place mode of exchange). However, Tranca also probably functioned as an autonomous gateway community coordinating interregional exchange between Andean, local montane forest and Amazonian societies. The greater diversity of wares relative to most other coeval sites supports this interpretation. The Tranca settlement system, especially at Tranca and Pomohuain (Site H15), shows clustered associations between Huacrapuquio, Malambo and Camonal assemblages despite the otherwise discrete regional distributions of the three wares. Some sites like Site CH41 in the lower montane forest additionally show mixing of Andean and Amazonian architectural techniques (Hastings 1985:639). As exotic elements like the Huacrapuquio pottery do not cluster particularly strongly at Tranca, it may be that the system hierarchy reflects mostly local redistribution while mediation of interregional exchange was carried out at several nodes within a gateway system. On the other hand, 608 small surface and subsurface pottery samples from the Tarma Valley sites may not adequately reflect real distributions. While present evidence for the gateway analogy is not particularly strong, neither do the limited data support other interpretations any more convincingly. Ostensibly "weak" patterning may be a function of interaction's comparatively low levels of integration and intensity in the Tarma Valley and adjacent regions. Most likely such interpretations will remain debatable until more evidence is available. Future investigations in the eastern montane forest should also target ethnohistorically described trade nodes at strategic lowland locations to explore other models of interaction that may or may not resemble the early Historic Period Urubamba River trade fairs below Cuzco described by Gade (1972), Camino (1977) and Lyon (1981) Conclusion Through the course of this thesis, we have seen how scholars have presented a variety of perspectives regarding prehistoric settlement in the Central Andean montane forests. Their viewpoints are usually consistent with the either the "unitary origins" or the "verticality" ruling theoretical frameworks. Few, if any of these perspectives are truly free of all of the assumptions listed by Raymond and included in my Chapter 1. Most scholars presume that insurmountable difficulties posed by this purportedly inhospitable ecological zone inhibited independent cultural 609 development, and permitted only subsidized permanent settlement. Consequently, they have viewed population movement and colonization as the salient montane forest settlement process. In this study I have attempted to evaluate the diverse migration hypotheses bearing on the problem of montane forest settlement and cultural development. Detailed examination of these hypotheses in Chapter 2 has revealed that they were formulated post-hoc to account for anomalous settlement distributions (Adams et al. 1978), and to further larger theoretical frameworks. Rather than ignoring these migration hypotheses, or categorically rejecting them on theoretical grounds, I have weighed each one separately, first with the new data from Manachaqui Cave, and second with extant data from the Central Andes and adjacent regions. Virtually all of the hypotheses can be rejected on purely empirical grounds utilizing the new data from Manachaqui Cave. That these postulated migrations were judged improbable does not mean that prehistoric migrations did not occur in the Central Andes or elsewhere. It does suggest however, that advocates of migration explanations still must demonstrate that some other interpretation does not better suit the data (Rouse 1958). Chapter 3's examination of documentary and archaeological evidence offered a glimpse of late prehistoric northeastern montane forest societies 610 participating in interregional interaction connecting the highland Andes and Amazonia. This evidence suggests Inca incorporation of a pre-existing exchange system linking northeastern montane forest and Amazonian Central Huallaga societies within the imperial interaction network. Subsequent chapters presented a long, and possibly unbroken prehistoric sequence of human occupation in the Pataz-Abiseo study area. Manachaqui Cave's sequence stands as a record of autochthonous northeastern montane forest cultural development, and all of the evidence for prehistoric population movements and colonization can be accounted for within an alternative framework of interregional interaction. Terminal Preceramic and Initial Period material and contextual evidence indicates that Manachaqui Cave served as a wayside station, and perhaps alternately as a temporary habitation. Initial Period north-Central Andean population nucleation on the upper slopes of the cordillera runs counter to evolutionary expectations of population clustering within ecological zones most favorable for early Andean mid-elevation and high altitude crop cultivation. Yet rather than accepting Onuki's (1985) ecological model, or replacing it with another ecologically-based explanation, it may be more appropriate to examine the archaeological evidence for other factors potentially governing regional demography. 611 Clear evidence for interregional interaction in the northern montane forests appears with north-Central Andean ceramic technology displaying style attributes that have temporal precedents in distant lowland regions to the north. However, the Alca source obsidian from northern Arequipa recovered from Manachaqui Cave's Preceramic Period deposits offers a rare, imperishable bit of evidence suggesting that such long-distance exchange was an ancient feature of prehistoric montane forest economies. Populations had already begun clustering around the northern rim of the Central Andes prior to the onset of the Initial Period. With new patterns of interregional interaction introduced during the late Initial Period flash horizon, the rockshelter became functionally specialized as a wayside station where travelers rested, ate and warmed themselves. While interregional interaction began to fuel burgeoning Initial Period and Early Horizon political economies, Manachaqui Cave became a node in a gateway system channeling long-distance interchange with the Northern AI1des into the heart of the Central Andes. Early Horizon pedestrian traffic probably included religious specialists, pilgrims and other travelers involved in ritualized gift exchange between population centers, and perhaps in less systematic trading and questing for knowledge and experience (Helms 1991, 1993), much as Amazonian shamans and curanderos undertake today (Oberem 1980; Langdon 1981; Bastien 1987). 612 Archaeological data from Kotosh, La Galgada and Chavin de Huantar strongly suggest a correlation between the amalgamation of local Andean styles with "foreign" attributes, symbols and imagery, and the early emergence of economically privileged, socio-political leaders in the Central Andes. Hirth's gateway model emphasizes the catalytic role of long-distance interregional exchange in stimulating increased social stratification, and the evidence for emerging sociopolitical complexity at Central Andean sites strategically situated to participate in montane forest interaction spheres (e.g. Kotosh, La Galgada, Poro-Poro and Cerro Blanco} supports his position. Although Manachaqui Cave is situated at the edge of the eastern montane forest, almost all of the archaeological patterning prior to the mid-Early Intermediate Period can be understood in terms of north-south, rather than east-west interaction. At present, the strength and importance of the connection between the Chorrera and Early Horizon interaction spheres may be under-appreciated. In fact, future research may demonstrate that interregional interaction with northern regions had a far more substantial role in the emergence of Central Andean civilization than did interchange with the Amazonian lowlands to the east. The ethnohistorically documented modes of east-west exchange between Andean and Amazonian societies may have been infrequent prior to the Early Intermediate Period, or highly 613 localized like the Kotosh-Tutishcainyo linkage. Between a.d. 200 and 400, interregional interaction involving study area societies underwent a techno-economic revolution with the introduction of llama-caravan transport technology. Interaction now connected centralized highland Central Andean polities and lowland Amazonian riverine exchange networks via montane forest intermediaries. Manachaqui Cave again seems to have functioned in the context of a gateway system of exchange. If it is determined that portions of the Empedrada Phase Paste A assemblage are intrusive from the Callej6n de Huaylas, then the archaeological evidence may reflect Recuay access to "wholesale" (Hirth 1978:38) commodities (i.e. bypassing additional intermediaries) at one or more autonomous gateway communities within the montane or premontane forests. During both the Suitacocha and Ernpedrada Phases, long- distance exchange would in theory have been tightly directed between sources and destinations. Continued regional survey, subsurface sampling and comparative ceramic analyses in the Pataz-Abiseo area, and especially in the Montane Rain Forest Zone, are all crucial to confirm or reject such a hypothesis. Mcst frustrating is t:he paucity of direct evidence for the commodities excha~ged, although virtually all goods from the middle and lower Amazon basin were perishable. Material evidence for these interactions abounds in the form of style 614 manipulation. During the Early Intermediate Period, highland amalgamation of lowland design attributes is no longer seen at Manachaqui Cave as Andean-Amazonian boundaries moved down-slope. At this time the most common pattern of style amalgamation seems to be the grafting of lowland style attributes like surface decorations onto montane forest vessel shapes. This pattern reflects the upslope movement of information rather than subsistence produce. Thus, commodities with the greatest impact on Central Andean's developmental trajectory may have been ideas, symbols and experiences from distant regions. Pan-regional changes in interaction during the Chavin and Huari horizons resulted in Manachaqui Cave's relative disuse. Chavin horizon interaction spheres emphasized intra-regional over interregional interaction. Middle Horizon events north of the Huari-controlled upper Santa drainage are still poorly understood. Lack of evidence for the rockshelter's Middle Horizon use is therefore not unexpected. There is no reason to believe that the montane forests were depopulated at these times, although populations previously clustered at principle exchange nodes would be expected to disperse. The Late Horizon prosperity indicated by Gran Pajaten's monumental architecture and elaborate lithic sculpture can be understood in terms of the lucrative material advantages acquired by "middlemen" (Lathrap 1973a:l72-173; Oberem 615 1974:350). Montane forest intermediaries at gateway communities are situated at the interface of two distinct regional economies. Hirth (1978:37) describes such interfaces as "economic shear lines where cost factors change." Gran Pajaten's extraordinary architectural achievements, the associated ornate funerary complex of Los Pinchudos and other evidence suggestive of a ranked social organization can be accounted for in terms of the PatazAbiseo area's privileged geographic position and gateway interaction's common association with centralized sociopolitical authority. This Andean-Amazonian exchange network was apparently so important to the maintenance of Andean social and political structures that it was left intact, albeit overseen, by the Inca conquerors. Historical documents reviewed in Chapter 3 offer scant, but unequivocal evidence for autonomous late prehistoric montane forest societies and Andean-Amazonian exchange activities in the Pataz-Abiseo area. Thus, Sucos, Puymal and the Late Horizon monumental architectural complexes encountered in the Pataz-Abiseo Montane Rain Forest Zone represent intermediary components of a gateway system, while the elaborate late prehistoric northern Chachapoyas settlements arrayed along the forest edges probably represent Hirth's "multiple gateway case" (1978: Fig. 2c). With the arrival of Europeans, unfamiliar strains of Old World diseases coursed through interregional exchange 616 network vectors, and montane forest gateway communities like Gran Pajaten would have been disproportionately buffeted. Exchange networks disintegrated as the indigenous Central Andean regional economy shrank, and an ancient tradition of montane forest trade and trade mediation dwindled. Ultimately, the eastern montane forests were abandoned, not for ecological reasons as archaeologists usually suppose, but for historical reasons. Montane forest families not wiped out by pestilence were hauled out under Viceroy Toledo's mid-16th century mandate. Those families escaping Toledo's torch either dispersed to attempt a living on their own, or departed the montane forest in search of community elsewhere. The agricultural potential of tropical montane forest ecological zones warrants future investigation. However, the chief limiting factor in all but the high Tropical Montane Rain Forest Life Zone (see Chapter 3) was probably the capacity of societies to mobilize communal labor for the necessary forest clearing and soil conservation. Archaeological evidence from the Central Andes demonstrates that native societies have possessed that capacity since the Preceramic Period. Montane forest agriculture may have had advantages only dimly appreciated by archaeologists. Parsons and Hastings (1988:214) observe that "a greater variety of crops can be grown than in the highlands and usually in a significantly shorter period of time." 617 Kauffmann's (1987:6} conclusion tha~ intensive (non- shifting) agriculture cannot be sustained in the montane forest due to soil erosion and rapid invasion of secondary regrowth is based upon his observation of modern, rather than pre-Hispanic, methods of cultivation, and may therefore be unwarranted. Reasons why the eastern montane forests are only lightly populated today are numerous and complex, but the current lack of infrastructure joining them to the rest of Peru contrasts sharply with the abundant evidence for prehistoric economic integration. In response to gloomy commentaries regarding the inability of prehistoric montane forest societies to "evolve" independently, it might be appropriate to consider Clarke and Blake's observation quoted on the first page of this chapter. In short, prehistoric communities in the Pataz-Abiseo area were both instigators and outgrowths of regionally evolving change. Within this thesis, I have attempted to show that the Pataz-Abiseo area is more profitably investigated as a cultural interface, than as a frontier. The term "frontier" has been avoided because of its core-periphery connotations (e.g. Lerner 1984:67; Green and Perlman 1985:4). At this juncture we might conclude that the prehistoric montane forests constitute one or more archaeological culture areas as Lumbreras and Morales have suggested. However, it might be argued on the basis of academic conceptions of "culture 618 areas" (e.g. Kroeber 1939) that the montane forests lacked a center or "culture core." Also, our archaeological knowledge of the Central Andean montane forests is still so sketchy that utilization of the culture area concept seems premature. On the other hand, referring to the montane forests as one or more sub-regions of the Central Andes analogous to the "north coast" or "south highlands" (cf. Kauffmann 1989:6) is fully justifiable. Decisions to refer to the Central Andean montane forests as a frontier, culture area or interaction sphere have been based upon scholars' theoretical perspectives, and terms will probably continue to reflect disparate research agendas. I have attempted to interpret evidence for what appear to be shifting complementary and nested ceramic interaction spheres that have alternately coincided with and cross-cut montane forest ecological boundaries. The montane forest interaction spheres circumscribing the Central Andes from north to south were loci of constant lateral cultural interaction contributing cohesion to the Peruvian cotradition throughout prehistory. An approach highlighting the montane forests' connectedness, rather than their marginality, provides an alternative perspective on dynamic prehistoric societies situated at montane forest cultural crossroads, and a sense of interregional interaction's catalytic role in the development of Central Andean civilization. REFERENCES CITED Acosta, Jose de 1958 El P. Jose de Acosta, Prov. al P. Everardo Mercuriano [1577]. In Monurnenta Peruana. Edited by A. de Egafia, Vol. 2, pp. 210-286. Monurnenta Historica Societatis IESU. Rome. Adams, William Y. 1979 On the Argument from Ceramics to History: a Challenge Based on Evidence fro Medieval Nubia. Current Anthropology 20 (4) :727-744. Adams, William Y., Dennis P. Van Gerven and RichardS. Levy 1978 The Retreat from Migrationism. Annual Review of Anthropology 7:483-532. Adovasio, J.M. and Robert F. Maslowski 1980 Cordage, Basketry and Textiles. In Guitarrero Cave: Early Man in the Andes. Edited by T. Lynch, pp.253292. Academic Press. New York. Aldenderfer, Mark S. and Charles Stanish 1993 Domestic Architecture, Household Archaeology, and the Past in the South-Central Andes. In Domestic Architecture. Ethnicity. and Complementarity in the South-Central Andes. Edited by M. Aldenderfer, pp. 1-12. University of Iowa Press. Iowa City. Allen, William L. 1968 A Ceramic Seqyence from the Alto Pachitea. Peru: Some Implications for the Development of Tropical Forest Culture in South America. Ph.D. dissertation, University of Illinois, Urbana. University Microfilms International. Ann Arbor. Alva, Walter 1986 FrUhe Keramik aus dem Jequetepeque-Tal, Nordperu. Materialen zur Allgemeinen und Vergleichenden Archaologie, Band 32. Verlag c. H. Beck. Mtinchen. 1988a Excavaciones en el Santuario del Tiempo Formative Udima-Poro Poro en la Sierra del Norte del Peru. Beitrage zur A1lgemeinen und Vergleichenden Archaologie 8:301-352. Mainz. 619 620 1988b Investigaciones en el Complejo Formative con Arquitectura Monumental de Purulen, Costa Norte del Perli. Beitrage zur Allgemeinen und Vergleichenden Archaologie 8:283-299. Mainz. Arnich, Jose 1854 Compendia Hist6rico (1768). Libreria de Rosa y Bouret. Paris. Ammerman, A. J. and L. L. Cavalli-Sforza 1973 A Population Model for the Diffusion of Early Farming in Europe. In The Explanation of Culture Change. Edited by C. Renfrew, pp. 343-357. University of Pittsburgh Press. Pittsburgh. Anthony, David 1986 The "Kurgan Culture," Indo-European Origins, and the Domestication of the Horse: A Reconsideration. Current Anthropology 27:291-313. 1990 Migration in Archaeology: the Baby and the Bathwater. American Anthropologist 92 (4) :895-914. APECO (Asociaci6n Peruana para la Conservaci6n de la Naturaleza) 1988a Results of 1987 Fieldwork, Faunal Inventory, Rio Abiseo National Park. Peru. Report submitted to the David and Lucile Packard Foundation. Lima, Peru and Boulder, Colorado. 1988b Results of 1988 Fieldwork, Faunal Inventory, Rio Abiseo National Park. Peru. Report submitted to the David and Lucile Packard Foundation. Lima, Peru and Boulder, Colorado. 1991 Results of 1989 Fieldwork, Faunal Inventory, Rio Abiseo National Park. Peru. Report submitted to the David and Lucile Packard Foundation. Lima, Peru and Boulder, Colorado. Arriaga, Pablo Joseph de 1968 The Extirpation of Idolatry in Peru [1621] Translated and edited by L. Clark Keating. University of Kentucky Press. Lexington. ARSI (Archivurn Romanurn Societatis Iesu, Novi Regni et Quitensis) 1636~37 Carta Annua de la Provincia del Peru; Missiones de Infieles. 621 Athens, Steven 1986 The Site of Pumpuentsa and the Pastaza Phase in Southeastern Lowland Ecuador. Nawpa Pacha 24:111124. Institute of Andean Studies. Berkeley. Bandelier, Adolf 1907 The Indians and Aboriginal Ruins Near Chachapoyas in Northern Peru. Historical Records and Studies. New York. Bankmann, Ulf 1979 Meche und Recuay. Baessler-Archiv. Neue Folge Band 27:253-271. Berlin. Bard, E. M. Arnold, R. G. Fairbanks and B. Hamelin 1993 230 Th-, 234 U and 14 C Ages Obtained by Mass Spectrometry on Corals. Radiocarbon 35:191-199. Bastien, Joseph 1987 Healers of the Andes: Kallawaya Herbalists and their Medicinal Plants. University of Utah Press. Salt Lake City. Bautista de Escobar, Juan 1899 Memorial de Juan Bautista de Escobar, en Nombre de las Ciudades de Chachapoyas y Moyobamba sabre que No Haga Don Martin de la Riba por Dichas Ciudades la Entrada que Tiene Capitulada. In Autos, Fechos y Actuados por el General D. Martin de la Riva Herrera [1653]. Revista de Archives y Bibliotecas Nacionales Afio II, Vol. III, entregas 3 y 4, pp.81-83. Emprenta de El Tiempo. Lima. II II Benevides C., Mario 1971 Analisis de la Ceramica Huarpa. Revista del Museo Nacional 37:63-88. Lima. Bennett, Wendell C. 1944 The North Highlands of Peru: Excavations in the Callej6n de Huaylas and at Chavin de Huantar. Anthropological Papers of the American Museum of Natural History Vol. 39, Part 1. New York. 1946 Excavations in the Cuenca Region, Ecuador. Yale University Publications in Anthropology No. 35. New Haven. 1948 The Central Andean Co-tradition. In b Reappraisal of Peruvian Archaeology. Memoirs of the Society for American Archaeology No.4, pp. 1-7. Salt Lake City. 622 Bennett, Wendell C. and Junius B. Bird 1949 Andean Culture History. American Museum of Natural History. Handbook Series No. 15. New York. Binford, Lewis R. 1983 The Archaeology of Place. Journal of Anthropological Archaeology 1:5-31. Binford, Lewis R. and Jeremy Sabloff 1982 Paradigms, Systematics and Archaeology. Journal of Anthropological Research 38:137-153. Bingham, Hiram 1930 Machu Picchu: A Citidel of the Incas. Books. New York. Hacker Art Bird, Junius B. 1963 Pre-ceramic Art from Huaca Prieta, Chicarna Valley. Nawoa Pacha 1:29-34. Institute of Andean Studies. Berkeley. Bird, Robert McK., David L. Browrnan and Marshall E. Durbin 1983-84 Quechua and Maize: Mirrors of Central Andean Culture History. Journal of the Steward Anthropological Society 15 (1-2) :187-240. Urbana. Birkeland, Peter W., Donald T. Rodbell, Daniel C. Miller and Susan K. Short 1989a Investigaciones Geologicas en el Parque Nacional Rio Abiseo, San Martin. Boletin de Lima 64:11:55-64. Lima. Birkeland, Peter W., Donald T. Rodbell and Susan K. Short 1989b Radiocarbon Dates on Deglaciation, Cordillera Central, Northern Peruvian Andes. Quaternary Research 32:111-113. Seattle. Bischof, Henning 1982 La Fase Engoroy: Periodos, Cronologia y Relaciones. In Primer Simposio de Correlaciones Antropol6gicas Andino-Mesoamericano. Edited by J. Marcos and P. Norton, pp.135-176. Escuela Superior Politecnica del Literal, Guayaquil. Bonavia, Duccio 1964 Investigaciones en la Ceja de Selva de Ayacucho. Arqueol6gicas No. 6. Institute de Investigaciones Antropologicas, Lima. 1968a Nucleos de Poblaci6n en la Ceja de Selva de Ayacucho (Peru). Aetas y Memorias de XXXVII Congreso de Americanistas Vol. 1, pp.75-83. Mar de Plata. 623 1968b Las Ruinas del Abiseo. Universidad Peruana de Ciencias y Tecnologia, Lima. 1967-68 Investigaciones Arqueol6gicas en el Mantaro Media. Revista del Museo Nacional 35:211-294. Lima. 1972a La Ceja de Selva: Colonizadores y Avanzadas. In Pueblos y Culturas de la Sierra Central del PerU, pp. 91-99. Cerro de Pasco Corporation. Lima. 1972b Reconocimiento Arqueol6gico en el Area del Mantaro. Argyeol6gicas 14, pp. 13-38. Institute de Investigaciones Antropologicas, Lima. 1978 Ecological Factors Affecting the Urban Transformation in the Last Centuries of the PreColumbian Era. In Advances in Andean Archaeology, ed. D. L. Browman, pp. 393-411. Mouton Publishers. The Hague. 1981 Tello y la Arqueologia de la Ceja de Selva. Hist6rica 5 (2) :149-158. Lima. 1990 Les Ruines du Abiseo. Inca-Peru: 3000 Ans d'Histoire, pp. 248-261. Musees Royaux d'Art et d'Histoire. Imschoot, uitgevers. Bruxelles. 1991 Peru Hombre e Historia: de los Origenes al Sigle XV. Ediciones Edubanco. Bonavia, Duccio and Rogger Ravines 1967 Las Fronteras Ecol6gicas de la Civilizaci6n Andina. Amaru 2:61-69. Lima. 1968 Villas del Horizonte Tardio en la Ceja de Selva del Peru: Algunas Consideraciones. Aetas y Memorias del XXXVII Congreso Internacional de Americanistas, Vol. 1, pp.153-160. Mar de Plata. Bonnier, Elisabeth 1983 Piruru: Nuevas Evidencias de Ocupaci6n Temprana en Tantamayo, Peru. Gaceta Argyeol6gica Andina 2 (8) :810. Lima. 1988 Aquitectura Preceramica en la Cordillera de los Andes: Piruru Frente a la Diversidad de los Datos. Antropol6gica 6 (6) :336-361. Lima. 624 Bonnier, Elisabeth, Julio Zegarra and Juan Carlos Tello 1985 Un Ejernplo de Crono-estratigrafia en un Sitio con Superposicion Arquitectonica: Piruru-Unidad I/II. Bulletin de L'Institut Francais D'Etudes Andines 14 (3-4) :80-101. Lima. Bray, Tamara L. 1995 The Panzaleo Puzzle: Non-local Pottery in Northern Highland Ecuador. Journal of Field Archaeology 22:137-156. Braun, Robert 1982 The Formative as Seen from the Southern Ecuadorian Highlands. In Primer Simposio de Correlaciones Antropologicas Andino-Mesoamericano, Edited by J. Marcos and P. Norton, pp. 41-100. Escuela Superior Politecnica del Litoral, Guayaquil. Brennan, Curtis T. 1978 Investigations at Cerro Arena. Peru: Incipient Urbanism on the Peruvian North Coast, Vols. 1 and 2. Ph.D. dissertation. University of Arizona. University Microfilms International. Ann Arbor. Brockington, Donald L., David M. Pereira Herrera, Ramon Sanzetenea Rocha and Maria de los Angeles Mufioz C. 1995 Estudios Arqueologicos del Periodo Formative en el Sur-este de Cochabamba 1988-1989. Cuadernos de Investigacion. Serie Argueologia No.8. Institute Antropologico y Museo de la Universidad Mayor de San Simon. Cochabamba, Bolivia. Brockington, Donald L. and Ramon Sanzetenea Rocha 1989 Decorated Formative Period Pottery from Cochabarnba, Bolivia: Ryden's Observation. Ethnos 54 (1-2) :63-68. Folkens Museum Etnografiska. Stockholm. Browman, David 1970 Early Peruvian Peasants: The Culture History of a Highland Valley. Ph.D. Dissertation. Department of Anthropology, Harvard University. Cambridge, MA. 1975 Trade Patterns in the Central Highlands of Peru in the First Millennium B.C. World Archaeology 6 (3) :322-329. 1989 Origins and Development of Andean Pastoralism: An Overview of the Past 6000 Years. In The Walking Larder: Patterns of Domestication. Pastoralism and Predation. Edited by J. Clutton-Brock, pp. 256-268. Unwin Hyman. London. 625 Bruhns, Karen 0. 1989 Intercambio entre la Costa y la Sierra en el Forrnativo Tardio: Nuevas Evidencias del Azuay. Proceedings of the 46th International Congress of Americanists. Edited by J. F. Bouchard and M. Guinea, pp. 57-74. BAR International Series 503. Brush, Stephen B. 1974 El Lugar del Hombre en el Ecosistema Andino. Revista del Museo Nacional 40:277-299. Lima. 1977 Mountain Field and Family: the Economy and Ecology of an Andean Valley. University of Pennsylvania Press. Philadelphia. Burger, Richard L. 1981 The Radiocarbon Evidence for the Temporal Priority of Chavin de Huantar. American Antiquity 46:592-602. 1984a Archaeological Areas and Prehistoric Frontiers: the Case of Formative Peru and Ecuador. In Social and Economic Organization in the Prehispanic Andes. Edited by D. L. Browman, R. L. Burger and M. A. Rivera, pp.33-71. BAR International Series No. 194. Manchester. 1984b The Prehistoric Occuoation of Chavin de Huantar, Peru. University of California Press. Berkeley. 1985a Concluding Remarks: Early Peruvian Civilization and its Relation to the Chavin Horizon. In Early Ceremonial Architecture in the Andes. Edited by C. B. Donnan, pp. 269-289. Durnbarton Oaks Research Library and Collection. Washington, D. C. 1985b Prehistoric Stylistic Change and Cultural Development at Huaricoto, Peru. National Geographic Research 1 (4) :505-534. 1988 Unity and Heterogeneity within the Chavin Horizon. In Peruvian Prehistory, edited by R. Keatinge, pp. 99-144, Cambridge University Press. Cambridge. 1992 Chavin and the Development of Andean Civilization. Thames and Hudson. London. 1993 The Chavin Horizon: Stylistic Chimera or Socioeconomic Metamorphosis. In Latin American Horizons. Edited by D. Rice, pp. 41-82. Dumbarton Oaks Research Library and Collection. Washington, D.C. 626 Burger, Richard L. and Lucy Salazar-Burger 1980 Ritual and Religion at Huaricoto. Archaeology 33:26-32. 1986 Early Organizational Diversity in the Peruvian Highlands: Huaricoto and Kotosh. In Andean Archaeology: Papers in Memory of Clifford Evans. Monograph XXVII, Institute of Archaeology, University of California, Los Angeles. Edited by R. Matos, S. A. Turpin and H. H. Eling Jr., pp.65-82. Los Angeles. Burger, Richard L., Michael Glascock and Hector Heff 1995 The Sources of the Obsidian Artifacts from Manachaqui Cave: A Preliminary Consideration. Unpublished Manuscript in possession of the authors. Bushnell, George H. S. 1946 An Archaeological Collection from Macas, on the Eastern Slopes of the Ecuadorian Andes. Man 46 (2) :2-6. London. 1951 The Archaeology of the Santa Elena Peninsula in Southwest Ecuador. Cambridge University Press. 1966 Some Archaeological Discoveries from the Frontier Region of Peru and Ecuador, Near Jaen. Aetas y Memorias del XXXVI Congreso de Americanistas, 1964 Vol. 1, pp. 501-507. Sevilla. Bush, M. B. and P. A. Colinvaux 1988 A 7000-year Pollen Record from the Amazon Lowlands, Ecuador. Vegetation 76:141-154. Caldwell, Joseph R. 1964 Interaction Spheres in Prehistory. In Hopewellian Studies. Scientific Papers Vol. XII, edited by J. R. Caldwell and R. L. Hall. Springfield. Cameron, Catherine M. 1995 Migration and the Movement of Southwestern Peoples. Journal of Anthropological Archaeology 14:104-124. Campana M., Roosevelt 1992 Litoescultura en Nunamarca. Yunga No. 3 (1988), pp. 13-20. Universidad Nacional de Trujillo. Trujillo, Peru. Camino, Alejandro 1977 Trueque, Correrias e intercambios entre los Quechuas Andinos y los Piro y Machiguenga de la Montana Peruana. Amazonia Peruana 1 (2) :123-140. Lima. 627 Cardich, Augusto 1973 Excavaciones en la Caverna de Huargo, Peru. Revista del Museo Nacional 39:11-29. Lima. 1985 The Fluctuating Upper Limits of Cultivation in the Central Andes and their Impact on Peruvian Prehistory. Advances in World Archaeology 4:293-333. Academic Press. New York. de las Casas, Enrique 1930a Provincia de Huallaga. Boletin de la Sociedad Geografica de Lima Torno XLVII, Trimestre cuarto, pp. 162-210. Lima. 1930b Monografia de la Provincia de Huallaga. Boletin de la Sociedad Geografica de Lima Torno XLVII, Trimestre cuarto, pp. 368-390. Lima. 1932 Provincia del Huallaga: Apuntes Monograficos (continuaci6n) . Boletin de la Sociedad Geografica de Lima Tomo XLIX, Trimestre tercero, pp. 167-215. Lima. 1933 Provincia del Huallaga: Apuntes Monograficos (continuaci6n) . Boletin de la Sociedad Geografica de Lima Torno L, Trimestre primero, pp. 55-71. Lima. 1934 La Mejor Ruta a Iquitos por el Rio Huallaga. Boletin de la Sociedad Geografica de Lima Tomo LI, Trimestre primero, pp. 51-69. Lima. 1935a Monografia de la Provincia de Huallaga (Continuaci6n) . Boletin de la Sociedad Geografica de Lima Tomo LII, Trimestre tercero, pp. 234-256. Lima. 1935b Monografia de la Provincia de Huallaga (Continuaci6n) . Boletin de la Sociedad Geografica de Lima Tomo LII, Trimestre cuarto, pp. 335-354. Lima. Cedr6n Goicochea, Elke 1989 Cronologia e Identificaci6n de Funci6n en Tres Edificios Prehispanicos del Sitio La Playa. Departamento de San Martin. Bachelor's thesis, Universidad Nacional de Trujillo. Trujillo, Peru. Chapman, John and Pavel M. Dolukhanov 1992 The Baby and the Bathwater: Pulling the Plug on Migrations. American Anthropologist 94 (1) :169-174. 628 Church, Warren B. 1988 Test Excavations and Ceramic Artifacts from Building No.1 at Gran Pajaten. Department of San Martin, Peru. Unpublished Master's thesis, Department of Anthroplogy, University of Colorado, Boulder. 1994 Early Occupations at Gran Pajaten, Peru. Andean Past 4:281-318. Cornell Latin American Studies Program. Ithaca. Cieza de Leon, Pedro de 1976 The Incas (1553). Translated by H. de Onis and edited by V. von Hagen. University of Oklahoma Press. Norman. Clark G. A. 1994 Migration as an Explanatory Concept in Paleolithic Archaeology. Journal of Archaeological Method and Theory 1 (4) :305-343. Clark G. A. and J. M. Lindly 1991 On Paradigmatic Biases and Paleolithic Research Traditions. Current Anthropology 32 (5) :577-587. Clarke, John E. and Michael Blake 1994 The Power of Prestige: Competitive Generosity and the Emergence of Rank Societies in Lowland Mesoamerica. In Factional Competition and Political Development in the New World. Edited by E. M. Brumfiel and J. W. Fox, pp. 17-30. Cambridge University Press. Cambridge. Clarke, John E. and Dennis Gosser 1995 Reinventing Mesoamerica's First Pottery. In The Emergence of Pottery: Technology and Innovation in Ancient Societies. Edited by W. K. Barnett and J. W. Hoopes, pp.209-222. Smithsonian Institution Press. Washington, D. C. Cobo, Padre Bernabe 1890 Historia del Nuevo Mundo [1653], Torno I. Edited by D. M. Jimenez de la Espada. Sociedad de Bibliofilos Andaluces. Sevilla. Coe, Sophie D. 1994 America's First Cuisines. University of Texas Press. Austin. Collier, Donald and John V. Murra 1943 Survey and Excavation in Southern Ecuador. Anthropological Series, Vol. 35, Field Museum of Natural History. Chicago. 629 Cook, Noble David 1981 Demographic Collapse. Indian Peru 1520-1620. Cambridge University Press. Cambridge. 1992 Impact of Disease in the Sixteenth-century Andean World. In Disease and Demography in the Americas. Edited by J. W. Verano and D. H. Ubelaker, pp. 20713. Smithsonian Institution Press. Washington, D.C. Cooper, John M. 1946 The Araucanians. In Handbook of South American Indians, Vol. 2: Andean Civilizations. Edited by J. H. Steward, pp.687-761. Bureau of American Ethnology Bulletin 143. Smithsonian Institution. Washington, D.C. Cordell, Linda S. 1984 Prehistory of the Southwest. Academic Press. San Diego. 1995 Tracing Migration Pathways from the Receiving End. Journal of Anthropological Archaeology 14:104-124. Correal Urrego, Gonzalo 1979 Investigaciones Arqueol6gicas en Abrigos Rocosos de Nemoc6n y Sueva. Fundaci6n de Investigaciones Arqueol6gicas Nacionales, Banco de la RepUblica. Bogota. Curtin, Philip D. and Gordon R. Willey 1951 A Survey of New Archaeological Sites in Central Pataz, Peru. Journal of the Washington Academy of Sciences 41 (2) :49-63. Cussler, Clive 1994 Inca Gold. Simon and Schuster. New York. DeBoer, Warren R. 1972-74 Bin6 Style Ceramics from Iparia. Nawpa Pacha 1012:91-107. Institute of Andean Studies. Berkeley. 1981 The Machete and the Cross: Conibo Trade in the Late Seventeenth Century. In Networks of the Past, Proceedings of the XII Chacmool Conference, edited by P. Francis, F. Kense and P. Duke, pp. 31-48. University of Calgary Archaeological Association. Calgary. 1984 Archaeological Reconnaissance in the Central Huallaga, Department of San Martin, Northeastern Peru. Unpublished report on 1983 field season in possession of the author. 630 n.d. Ceramic Assemblage Variability in Ecuador and Peru. Paper presented Archaeology of Formative Ecuador. Raymond and R. L. Burger, October Durnbarton Oaks. Washington, D.C. the Formative of to the session Organized by J. s. 7-8, 1995, at DeBoer, Warren R. and J. Scott Raymond 1987 Roots Revisited: The Origins of the Shipibo Art Style. Journal of Latin American Lore 13:1:115-132. Denevan, William M. 1992 The Aboriginal Population of Amazonia. In The Native Population of the Americas in 1492, edited by W. Denevan, 2nd edition, pp. 205-234. University of Wisconsin Press. Madison. Deza Rivasplata, Jaime 1975-76 La Playa: Un Complejo Arqueol6gico de la Cuenca del Abiseo. Boletin del Seminario de Arqueologia, No.17-18, pp.43-50. Institute Riva Aguero, Pontificia Universidad Cat6lica del Perli. Lima. Diez de San Miguel, Garci 1964 Visita Hecha a la Provincia de Chucuito por Garci Diez de San Miguel en el Afio 1567. Casa de la Cultura Peruana. Lima. Dillehay, Tom D. and Patricia J. Netherly 1983 Exploring the Upper Zafia Valley of Peru. Archaeology 37 (4) :23-30. Dillehay, Tom D., Patricia J. Netherly and Jack Rossen 1989 Middle Preceramic Public and Residential Sites on the Forested Slope of the Western Andes, Northern Peru. American Antiquity 54 (4) :733-759. Dobyns, Henry F. 1992 Native American Trade Centers as Contagious Disease Foci. In Disease and Demography in the Americas. Edited by J. W. Verano and D. H. Ubelaker, pp. 215222. Smithsonian Institution Press. Washington, D.C. 1993 Disease Transfer at Contact. Annual Review of Anthropology 22:273-291. Drewett, Peter 1993 Review of Migrations in Prehistory, by Irving Rouse. Antiquity 67 (255) :456-457. 631 Earle, Timothy, Terence D'Altroy and C. LeBlanc 1978 Arqueologia Regional de los Periodos Prehispanicos Tardios en el Mantaro. In III Congreso Peruano El Hombre y la Cultura Andina: Aetas y Trabajos. Edited by R. Matos, Torno II, pp. 641-672. Editora Lasontay. Lima. Earle, Timothy, Terence D'Altroy, Christine Hastorf, Catherine Scott, Cathy Costin, Glenn Russell and Elsie Sandefur 1987 Archaeological Field Research in the Upper Mantaro, Peru, 1982-1983: Investigations of Inka Expansion and Exchange. Institute of Archaeology. Monograph 28 University of California, Los Angeles. Los Angeles. Eddy, J. 1977 Climate and the Changing Sun. Climatic Change 1:173190. Espinoza Soriano, Waldemar 1967 Los Sefiorios Etnicos de Chachapoyas y la Alianza Hispano-Chacha. Revista Hist6rica 30:224-332. Lima. Erickson, Jonathon E. and Clement W. Meighan 1984 Boundaries, Alliance and Exchange in California. In Exploring the Limits: Frontiers and Boundaries in Prehistok¥. Edited by S. P. De Atley and F. J. Findlow, pp. 143-152. BAR Series 223. Estrada, Emilio 1958 Las Culturas Pre-Clasicas. Formativas o Arcaicas del Ecuador. Publicaci6n del Museo Victor Emilio Estrada No. 5. Guayaquil. Estrada, Emilio, Betty J. Meggers and Clifford Evans 1964 The Jambeli Culture of South Coastal Ecuador. Proceedings of the U.S. National Museum Vol. 115, pp. 483-558. Washington, D.C. Evans, Clifford and Betty J. Meggers 1968 Archaeological Investigations on the Rio Napo, Eastern Ecuador. Smithsonian Institution Press. Washington, D.C. 1982 Tecnicas Decorativas Diagn6sticas y Variantes Regionales Chorrera, un Analisis Preliminar. In Primer Simposio de Correlaciones Antropol6gicas Andino-Mesoamericano, Edited by J. Marcos and P. Norton, pp. 121-134. Escuela Superior Politecnica del Literal, Guayaquil. 632 Farrand, William R. 1985 Rockshelter and Cave Sediments. In Archaeological Sediments in Context. Edited by J. K. Stein and W. R. Farrand. Peopling of the Americas Edited Volume Series : Vol. 1. Center for the Study of Early Man. Institute for Quaternary Studies, University of Maine at Orono. Orono. Fejes, Paul 1944 Archaeological Explorations in the Cordillera Vileabamba. Southeastern Peru. Viking Fund Publication. Flores-Ochoa, Jorge A. 1979 Pastoralists of the Andes: The Alpaca Herders of Paratia. Institute for the Study of Human Issues. Philadelphia. Fogel, Reidy 1993 Settlements in Time: A Study of Social and Political Development during the Gallinazo Occupation of the North Coast of Peru. Unpublished Ph.D. dissertation. Department of Anthropology, Yale University. New Haven. Ford, James 1969 A Comparison of the Formative Cultures in the Americas: Diffusion or the Psychic Unity of Man. Smithsonian Institution Press. Washington, D.C. Fung, Rosa 1972 Las Aldas: Su Ubicaci6n dentro del Proceso Hist6rico del Peru Antigua Dedalo 5:9-10. Sao Paolo. 1975 Excavaciones in Pacopampa, Cajamarca. Revista del Museo Nacional. Torno XLI, pp. 129-207. Lima. 1982 El Temprano Surgimiento en el Peru de los Sistemas Socio-politicos Complejos: Planteamiento de una Hip6tesis de Desarrollo Original. In Primer Simposio de Correlaciones Antropol6gicas AndinoMesoamericano, Edited by J. Marcos and P. Norton, pp. 457-495. Escuela Superior Politecnica del Literal, Guayaquil. Gade, Daniel W. 1967 The Guinea Pig in Andean Folk Culture. The Geographical Review 57 (2) :213-224. 633 1972 Comercio y Colonizaci6n en la Zona de Contacto Entre la Sierra y las Tierras Bajas del Valle del Urubamba, Perli. Aetas y Memorias del XXXIX Congreso Internacional de Americanistas Vol. 4, pp.207-221. 1975 Plants. Man and the Land in the Vilcanota Valley of Peru. Dr. W. Junk B. V., Publishers. The Hague. 1979 Inca and Colonial Settlement, Coca Cultivation and Endemic Disease in the Tropical Forest. Journal of Historical Geography 5 (3) :263-279. Gambini E., Wilfreda 1984 Santa y Nepefta: Dos Valles. Dos Culturas. Imprenta M. Castillo R. Lima. Garcilaso de la Vega, Inca 1966 Royal Commentaries of the Incas and General History of Peru. Part One (1609). Translated by Harold V. Livermore. University of Texas Press. Austin. Gentry, Alwyn 1992 Diversity and Floristic Composition of Andean Forests of Peru and Adjacent Countries: Implications for their Conservation. In Memorias del Museo de Historia Natural. Edited by K. R. Young and N. Valencia, No. 21, pp. 11-30. Universidad Nacional Mayor de San Marcos. Lima. Goldstein, Paul 1993 House, Community, and State in the Earliest Tiwanaku Colony: Domestic Patterns and State Integration at Omo M12, Moquegua. In Domestic Architecture. Ethnicity and Complementarity in the South-Central Andes. Edited by M. S. Aldenderfer, pp. 25-41. University of Iowa Press. Iowa City. Green, Stanton W. and Stephen M. Perlman 1985 Frontiers, Boundaries, and Open Social Sysytems. In The Archaeology of Frontiers and Boundaries. Edited by S. W. Green and S. M. Perlman, pp.3-13. Academic Press. New York. Greenberg, Joseph H. 1987 Language in the Americas. Stanford University Press. Stanford. Grieder, Terence 1978 The Art and Archaeology of Pashash. University of Texas Press, Austin. 634 1988 Ceramics. In La Galgada. Peru: A Preceramic Culture in Transition. By T. Grieder, A. Bueno M., C. E. Smith, Jr. and R. M. Malina, pp. 185-191. University of Texas Press. Austin. Grieder, Terrance and Alberto Bueno M. 1985 Ceremonial Architecture at La Galgada. In Early Ceremonial Architecture in the Andes. Edited by C. B. Donnan, pp. 93-109. Dumbarton Oaks Research Library and Collection. Washington, D. C. Grieder, Terrence, Alberto Bueno M., C. Earle Smith, Jr. and Robert M. Malina 1988 La Galgada. Peru: A Preceramic Culture in Transition. University of Texas Press. Austin. Grossman, Joel W. 1972 Early Ceramic Cultures of Andahuaylas. Apurimac. Peru. Unpublished Ph.D. dissertation, University of California, Berkeley. 1985 Demographic Change and Economic Transformation in the South Central Highlands of Pre-Huari Peru. Nawga Pacha 21:45-126. Institute of Andean Studies. Berkeley. Guaman Poma de Alaya, Felipe 1936 Nueva Cor6nica y Buen Gobierno <Codex Peruvien Illustrel . Travaux et Memoires de l'Institut D'Ethnologie. Vol. XXIII. Universite de Paris. Paris. Guffroy, Jean 1989 Un Centro Ceremonial Formative en el Alto Piura. Bulletin de l'Institut Francais d'Etudes Andines 18 (2) :161-208. Lima. 1992 Las Tradiciones Culturales Formativas en el Alto Piura. In Estudios de Argueologia Peruana. Edited by D. Bonavia, pp. 99-121. Fomciencias. Lima. Guffroy, Jean, N. Almeida, P. Lecoq, C. Caillavet, F. Duverneuil, L. Emperaire and B. Arnaud 1987 Loja Prehispanique: Recherches Archeologigues dans les Andes Meridionales de L'Equateur. Institut FranGais D'Etudes Andines. Paris. Guillet, David 1985 Hacia una Historia Socio-Econ6mica de los Basques en los Andes Centrales del Peru. Boletin de Lima 38:7:79-84. Lima. 635 Hampe M., Teodoro 1978 Relaci6n de los Encomenderos y Repartimientos del PerU en 1561. Historia y Cultura No. 12, pp. 76-116. Lima. Hantman, Jeffrey L. and Stephen Plog 1982 The Relationship of Stylistic Similarity to Patterns of Material Exchange. In Contexts for Prehistoric Exchange. Edited by J. Ericson and T. Earle, pp. 237-263. Academic Press. New York. Harlan, Jack 1975 Crops and Man. American Society of Agronomy. Madison. von Hagen, Victor W. 1955 Highway of the Sun. Little, Brown and Co. Boston. Harner, Michael J. 1972 The Jivaro: People of the Sacred Waterfall. Doubleday/Natural History Press. Garden City, NY. Hartshorn, Gary 1983 Introduction to Chapter 7, Plants. In Costa Rica Natural History. Edited by D. Janzen, pp. 118-157. University of Chicago Press. Chicago. Hastings, Charles M. 1985 The Eastern Frontier: Settlement and Subsistence in the Andean Margins of Central Peru, Vols. 1 and 2. Ph.D. dissertation, University of Michigan. University Microfilms International. Ann Arbor. 1987 Implications of Andean Verticality in the Evolution of Political Complexity: a View from the Margins. In The Origins and Development of the Andean State, edited by J. Hass, S. Pozorski and T. Pozorski, pp. 145-157. Cambridge University Press. Cambridge. Hegmon, Michelle 1992 Archaeological Research on Style. Annual Review of Anthropology 21:517-536. Helms, Mary W. 1991 Esoteric Knowledge, Geographic Distance, and the Elaboration of Leadership Status: Dynamics of Resource Control. In Profiles in Cultural Evolution Papers from a Conference in Honor of Elman R. Service. Edited by A T. Rambo and K. Gillogly, pp. 333-349. Anthropological Papers, Museum of Anthropology, University of Michigan No. 85. Ann Arbor. 636 1993 Craft and the Kingly Ideal: Art. Trade and Power. University of Texas Press. Austin. Henrickson, Elizabeth F. and Mary M. A. McDonald 1983 Ceramic Form and Function: An Ethnographic Search and an Archeological Application. American Anthropologist 82:630-643. Herndon, William L. 1952 EXPloration of the Valley of the Amazon. Edited by H. Basso. McGraw-Hill Book Co. New York. Hilbert, Peter Paul 1968 Archaologische Untersuchungen am Mittleren Amazonas. Dietrich Reimer/Andrews and Steiner. Berlin. Hill, Betsy 1972-74 A New Chronology of the Valdivia Ceramic Complex from the Coastal Zone of Guayas Province, Ecuador. NawPa Pacha 10-12:1-39. Institute for Andean Studies. Berkeley. Hirth, Kenneth G. 1978 Interregional Trade and the Formation of Prehistoric Gateway Communities. American Antiquity 43 (1) :3546. Hocquenghem, Anne Marie 1993 Rutas de Entrada del Mullu en el Extreme Norte del Perli. Bulletin de l'Institut Francais d'Etudes Andines 22 (3) :701-719. Lima. Hocquenghem, Anne Marie, Jaime Idrovo, Peter Kaulicke and Dominique Gomis 1993 Bases del Intercambio entre las Sociedades Norperuanas y Surecuatorianas: Una Zona de Transici6n entre 1500 A.C. y 600 D.C. Bulletin de l'Institut Francais d'Etudes Andines 22 (2) :443-466. Lima. Hodder, Ian 1982 Symbols in Action. Cambridge University Press. Cambridge. Holdridge, L. R. 1967 Life Zone Ecology. Revised edition. Tropical Science Center. San Jose, Costa Rica. 637 Hoopes, John 1994 Ford Revisited: A Critical Review of the Chronology and Relationships of the Earliest Ceramic Complexes in the New World, 6000-1500 B.C. Journal of World Prehistory 8 (1) :1-49. Horkheimer, Hans 1959 Algunas Consideraciones Acerca de la Arqueologia en el Valle del Utcubamba. Aetas del II Congeso Nacional de Historia del PerU: Epoca Pre-hispanica Vol. 1, pp. 71-90. Lima. Hyslop, John 1984 The Inca Road System. Academic Press. New York. Ibarra Grasso, Dick Edgar and Roy Querejazu Lewis 1986 30.000 Afios de Prehistoria en Bolivia. Editorial Los Amigos del Libra. La Paz. INC (Institute Nacional de Cultura, PerU) 1983 Inventario de Monumentos Argyeologicos del Peru: Zona Norte (Primera Aproximacion). Compiled by R. Ravines and R. Matos. Departamento de Investigacion del Patrimonio Monumental. Lima. Institute Geografico Nacional 1984 Mapa Fisico Politico del PerU (1:1'000,000). Ministerio de Guerra. Lima Institute de Geologia y Mineria 1975 Mapa Geologico del Perti. Institute de Geologia y Mineria, Ministerio de Energia y Minas. Lima. Isbell, William H. 1968 New Discoveries in the Montana of Southeastern Peru. Archeology 21 (2) :108-114. 1970 Review of Las Ruinas del Abiseo, by Duccio Bonavia. American Antigyity 35 (3) :237. 1971 Un Pueblo Rural Ayacuchano durante el Imperio Huari. Aetas y Memorias del 39 Congreso Internacional de Americanistas, Vol. 3, pp.89-105. Lima. 1974 Ecologia de la Expansion de los Quechua-hablantes. Revista del Museo Nacional 40:139-155. 1977 The Rural Foundation for Urbanism. Illinois Studies in Anthropology No. 10. University of Illinois Press. Urbana. 638 1983-84 Andean Linguistics and Culture History: An Examination of Competing Interpretations. Journal of the Steward Anthropological Society 15 (1-2) :241258. Isbell, William H. and Katherine Schreiber 1978 Was Huari a State? American Antiquity 43 (3) :372389. Izaguirre, Bernardino 1923 Historia de las Misiones Franciscanas y Narraci6n de la Geografia en el Oriente del Peril. Vol. 2. Talleres Tipograficos de la Penitenciaria. Lima. Izumi, Seiichi 1971 Development of the Formative Culture in the Ceja de Montana: A View Point Based on the Materials from the Kotosh Site. In Dumbarton Oaks Conference on Chavin edited by Elizabeth Benson, pp.49-72. Washington D.C. Izumi, Seiichi and Toshihiko Sono 1963 Andes 2: Excavations at Kotosh. Kadokawa Publishing. Tokyo. Peru. 1960. Izumi, Seiichi and Kazuo Terada 1966 Andes 3: Garbanzal and Pechiche. Tumbes Valley, Peru. 1960. Kadokawa Publishing. Tokyo. 1972 Andes 4: Excavations at Kotosh. Peru. 1963 and 1966. University of Tokyo Press. Tokyo. Izumi, Seiichi, Pedro Cuculiza and Chiaki Kana 1972 Excavations at Shillacoto. Huanuco. Peru. University Museum, University of Tokyo, Bulletin No. 3. Tokyo. Jakobsen, Jan, J. Balslev Jorgensen, L. Kempfner Jorgensen and Inge Schjellerup 1986-87 "Cazadores de Cabezas" en Sitios Pre-Inca de Chachapoyas. Revista del Museo Nacional 48:139-186. Lima. Jij6n y Caamafto, Jacinto 1927 Puruha: Contribuci6n al Conocimiento de los Aborigenes de la Provincia de Chimborazo de la Republica de Ecuador. Tipografia y Encuadernaci6n Salesianas. Quito. 1943 El Ecuador Interandino y Occidental antes de la Conguista Castellana, Vol. 3. Editorial Ecuatoriana. Quito. 639 Jouanen, Jose 1941 Historia del Campania de Jesus en 1a Antigua Provincia de Quito 1570-1774, Vol. 1. Editorial Ecuatoriana. Quito. Julien, Catherine J. 1985 The Historical Geography of Chachapoyas. Ethnohistory of Chachapoyas, Part 1. Unpublished manuscript in possession of the author. Julien, Daniel 1993 Late Pre-Inkaic Ethnic Group in Highland Peru: An Archaeological-Ethnohistorical Model of the Political Geography of the Cajamarca Region. Latin American Antiquity 4 (3) :246-273. Kana, Chiaki 1971 Excavaciones en Shillacoto, Huanuco. Revista del Museo Nacional 37:52-62. Lima. 1972 Pre-Chavin Cultures in the Central Highlands of Peru: New Evidence from Shillacoto, Huanuco. In The Cult of the Feline. Edited by E. P. Benson, pp. 139-152. Dumbarton Oaks Research Library and Collection. Washington, D.C. 1979 The Origins of the Chavin Culture. Dumbarton Oaks Studies in Pre-Columbian Art and Archaeology No. 22. Washington, D.C. Kauffmann Doig, Federico 1976 El Perli Argueol6gico. Kompaktos. Lima. 1980 "Los Pinchudos": Exploraci6n de Ruinas Intactas en la Selva. Boletin de Lima 7:26-31. Lima. 1983 Manual de Argueologia Peruana. Octava edicion, Peisa. Lima. 1986 Sarcofagos Pre-Incas en los Andes Amaz6nicos. Kuntur 1:4-9. Lima 1987 Andes Amaz6nicos: Investigaciones Argueol6gicas 1980-1986 (Expediciones Antisuyol . Publication of the Banco Continental, Lima. 1991 Sobrepoblaci6n en los Andes: Una Explicaci6n del Origen y Proceso de la Cultura Andina. l'imaginaire, 1 (3) :45-48. Publicaci6n Trimestral de la Alianza Francesa. Lima. 640 1992a Historia del Perli AntiqUo: Una Nueva Perspectiva. Vol. 1. Editorial Monterrico. S.A. 1992b Historia del Perli AntiqUo: Una Nueva Perspectiva. Vol. 2. Editorial Monterrico. S.A. Kauffmann Doig, Federico and Giancarlo Ligabue 1990 Les Andes Amazoniennes. In Inca-Peru: 3000 Ans d'Histoire, pp. 262-337. Musees Royaux d'Art et d'Histoire. Imschoot, uitgevers. Bruxelles. Kauffmann Doig, Federico, Miriam Salazar, Daniel Morales, Iain Mackay and Oscar Sacay 1989 Andes Amaz6nicas: Sitios Intervenidas por la Expedici6n Antisuyo/86. Arqyeol6gicas No. 20. Publicaci6n del Institute de Investigaciones Antropologicas. Museo Nacional de Antropologia y Arqueologia. Lima. Kaufman, Terrence 1990 Language History in South America: What We Know and How to Know More. In Amazonian Linguistics. Edited by D. Payne, pp. 13-74. University of Texas Press. Austin Kaulicke, Peter 1975 Pandanche: Un Caso Formative en los Andes de Cajamarca Seminario de Historia Rural Andina. Universidad Nacional Mayor de San Marcos. Lima. 1981 Keramic der Fruhen Initial-periode aus Pandanche, Dept. de Cajamarca, Peru. Beitrage zur Allgemeinen und Vergleichende Archaeologie. 3:363-389. Munchen. 1991 El Periodo Intermedio Temprano en el Alto Piura: Avances del Proyecto Arqueol6gico "Alto Piura (1987990). Bulletin de l'Institut Francais d'Etudes Andines 20 (2) :381-422. Lima. Kendall, Ann 1976 Preliminary Report on Ceramic Data and the Pre-Inca Architectural Remains of the (Lower) Urubamba Valley, Cuzco. Baessler Archiv Neue Folge, Band XXIV, pp. 41-160. Berlin. Kent, Jonathan D. 1993 A New Research Facility in Andean Zooarchaeology and Recent Investigations in Ecuador and Peru. Paper presented at the 21st Annual Midwest Conference on Andean and Amazonian Archaeology and Ethnohistory. St. Louis. 641 1994 The Faunal Remains of Manachaqui Cave, North Highlands, Peru. Unpublished manuscript in Possession of the author. Kalata, Alan 1983 The South Andes. In Ancient South Americans. Edited by J. D. Jennings, pp. 241-285. W. H. Freeman and Company. San Francisco. 1993 The Tiwanaku: Portrait of an Andean Civilization. Blackwell, Cambridge, Massachusetts and Oxford. Kroeber, Alfred L. 1939 Cultural and Natural Areas of Native North America. University of California Publications in American Archaeology and Ethnology Vol. 38. University of California Press. Berkeley. Krzanowski, Andrzej 1986 The Cultural Chronology of Northern Andes of Peru (The Huamachuco-Quiruvilca-Otuzco Region) . Acta Archaeologica Carpathica Torno XXV, pp. 231-263. Landar, Herbert 1977 South and Central American Indian Languages. In Native Languages of the Americas. Edited by T. A. Sebeok, pp. 401-527. Plenum Press. New York. Langdon, E. Jean 1981 Cultural Bases for Trading of Visions and Spiritual Knowledge in the Colombian and Ecuadorian Montana. In Networks of the Past, Proceedings of the XII Chacmool Conference, edited by P. Francis, F. Kense and P. Duke, pp. 101-116. University of Calgary Archaeological Association. Calgary. Langlois, Louis 1939 Utcubamba: Investigaciones Arqueologicas en Este Valle del Departmento de Amazonas. Revista del Museo Nacional 8 (2) :224-249. Lima. 1940a Utcubamba: Investigaciones Arqueol6gicas en Este Valle del Departmento de Amazonas. Revista del Museo Nacional 9 (1) :34-71. Lima. 194Gb Utcubamba: Investigaciones Arqueol6gicas en Este Valle del Departmento de Amazonas. Revista del Museo Nacional, 9 (2) :191-228. Lima. 642 Lanning, Edward P. 1963 A Ceramic Sequence for the Piura and Chira Coast, North Peru. University of California Publications in American Archaeology and Ethnology, Vol. 48, No. 2, pp. 135-284. Berkeley. 1967 Peru Before the Incas. Englewood Cliffs, NJ. Larco H. , Rafael 1941 Los Cupisniques. Casa Editora "La Cr6nica y Variedades" S .A., Ltda. Lima. 1962 La Cultura Santa. Valverde, S.A. Lima. Lathrap, Donald W. 1958 The Cultural Sequence at Yarinacocha, Eastern Peru. American Antiquity 23 (4) :379-388. 1962 Yarinacocha: Stratigraphic Excavations in the Peruvian Montana. Unpublished Ph.D. dissertation. Department of Anthropology. Harvard University. Cambridge. 1963 Possible affiliations of the Machalilla complex of coastal Ecuador. American Antiquity, 29:239-241. 1965 Origins of Central Andean Civilization: New Evidence. Science Vol. 148, pp. 796-798. 1970 The Upper Amazon. Thames and Hudson. London. 1971 The Tropical Forest and the Cultural Context of Chavin. In Dumbarton Oaks Conference on Chavin. Edited by E. P. Benson, pp. 73-100. Dumbarton Oaks Research Library and Collection. Washington, D.C. 1973a The Antiquity and Importance of Long-distance Trade Relationships in the Moist Tropics of Pre-Columbian South America. World Archaeology 5 (2) :170-186. 1973b Gifts of the Cayman: Some Thoughts on the Subsistence Basis of Chavin. Variation in Anthropology: Essays in Honor of John C. McGregor. Edited by D. W. Lathrap and J. Douglas, pp. 91-103. Illinois Archaeological Survey. Urbana. 1974 The Moist Tropics, the Arid Lands and the Appearance of Great Art Styles in the New World. In Art and Environment in the Native America. Edited by I. R. Traylor and M. E. King. Texas Technological University Press. Lubbock. 643 1977 Our Father the Cayman, Our Mother the Gourd: Spinden Revisited, or a Unitary Model for the Emergence of Agriculture in the New World. In Origins of Agriculture. Edited by C. Reed, pp. 713-752. Mouton Publishers. The Hague. 1982 Complex Iconographic Features by Olmec and Chavin and Some Speculations on their Possible Significance. In Primer Simposio de Correlaciones Antropol6gicas Andino-Mesoamericano, Edited by J. Marcos and P. Norton, pp. 301-327. Escuela Superior Politecnica del Literal, Guayaquil. 1985 Jaws: The Control of Power in the Early Nuclear American Ceremonial Center. In Early Ceremonial Architecture in the Andes. Edited by C. B. Donnan, pp. 93-109. Dumbarton Oaks Research Library and Collection. Washington, D. C. Lathrap, Donald (editor) 1956 An Archaeological Classification of Culture Contact Situations. Memoirs of the Society for American Archaeology No. 11. Salt Lake City. Lathrap, Donald W., Donald Collier and Helen Chandra 1975 Ancient Ecuador: Culture, Clay and Creativity 3000300 B.C. Field Museum of Natural History. Chicago. Lathrap, Donald W., Angelika Gebhart-Sayer and Ann Mester 1985 The Roots of the Shipibo Art Style: Three Waves on Imiriacocha or There Were "Incas" Before the Incas. Journal of Latin American Lore 11 (1) :31-119. Lathrap, Donald W., Angelika Gebhart-Sayer, Ann Mester and Thomas P. Myers 1987 Further Discussion of the Roots of the Shipibo Art Style: A Rejoinder to DeBoer and Raymond. Journal of Latin American Lore 13 (2) :225-271. Lathrap, Donald W, Jorge G. Marcos, James A. Zeidler 1977 Real Alto: An Ancient Ceremonial Center. Archaeology 3 0 ( 1) : 2-13 . Lathrap, Donald W. and Lawrence Rays 1963 The Archaeology of the Cave of the Owls in the Upper Montana of Peru. American Antiquity 29 (1) :27-38. Lavallee, Daniele 1977 Telarmachay: Campamento de Pastores en la Puna de Junin del Periodo Formative. Revista del Museo Torno 43, pp. 61-96. Lima. 644 Lavallee, Daniele, Michele Julien and Jane Wheeler 1982 Telarmachay: Niveles Preceramicos de Ocupaci6n. Revista del Museo Nacional Torno 46, pp. 55-127. Lima. Lavallee, Daniele, Michele Julien, Jane Wheeler and Claudine Karlin 1985 Telarmachay: Chasseurs et Pasteurs Prehistorigues des Andes Torno 1. Institut Fran9ais d'Etudes Andines. Editions Recherche sur las Civilisations No. 20. Paris. Lecoq, Patrice 1987 La Periode des developpements Regionaux. In Loja Prehispanigue: Recherches Archeologigues dans les Andes Meridionales de L'Eguateur by J. Guffroy, N. Almeida, P. Lecoq, C. Caillavet, F. Duverneuil, L. Emperaire and B. Arnaud, pp. 217-258. Institut Fran9ais D'Etudes Andines. Paris. Ledergerber Crespo, Paulina 1995 Factores Geograficos en la Localizaci6n de Sitios Arqueol6gicos: El Caso de Morena-Santiago, Ecuador: Un Informe Preliminar. In Cultura y Medio Ambiente en el Area Andina Septentrional. Co-ordinated by M. G. Guinea, J. Bouchard and J. Marcos, pp. 343-375. Colecci6n Biblioteca Abya-Yala 21. Ediciones Abyaala. Quito. LeMoine, Genevieve and J. Scott Raymond 1987 Leishmaniasis and Inca Settlement in the Peruvian Jungle. Journal of Historical Geography 13 (2) :113129. Lennon Thomas J., Warren B. Church and Miguel Cornejo G. 1989 Investigaciones Arqueol6gicas en el Parque Nacional Rio Abiseo, San Martin. Boletin de Lima 62:11:43-56. Leo, Mariella 1992 Problematica del Parque Nacional Rio Abiseo. Amazonia Peruana 11 (21) :109-144. Lima. Leon, Blanca, Kenneth R. Young and Lois Brako 1992 Analisis de la Composici6n Floristica del Bosque Montano Oriental del Peru. Memorias del Museo de Historia Natural 21:141-154. Universidad Nacional Mayor de San Marcos. Lima. Lerche, Peter 1995 Los Chachapoya y los Simbolos de su Historia. Servicios Editoriales Cesar Gayoso. Lima. 645 Lerner, Shereen 1984 Defining Prehistoric Frontiers: A Methodological Approach. In Exploring the Limits: Frontiers and Boundaries in Prehistory. Edited by S. P. De Atley and F. J. Findlow, pp. 67-80. BAR International Series 223. Oxford. Lightfoot, Kent G. and Antoinette Martinez 1995 Frontiers and Boundaries in Archaeological Perspective. AnnUal Review of Anthropology 24:471-492. Lippi, Ronald 1983 La Ponga and the Machalilla Phase of Coastal Ecuador. Doctoral dissertation. Department of Anthropology, University of Wisconsin-Madison. University Microfilms International. Ann Arbor. Lira, Jorge A. 1970 Breve Diccionario Kechwa-Espafiol. Edici6n Popular. Cusco. Lopez de Alvarado, Domingo 1907 Carta a S. M. del Cabildo de Moyobamba sobre las Conquistas de la Riva Herrera [1656] . In Juicio de Limites entre el Peru and Bolivia: Prueba Peruana. Presented by v. M. Maurtua. Vol. 3, pp.361-365. Imprenta de Henrich y Comp. Barcelona. Loredo, Rafael 1958 Los Repartimientos: Bocetos para la Nueva Historia del PerU. Lima. Lumbreras, Luis G. 1974 The Peoples and Cultures of Ancient Peru. Smithsonian Institution Press. Washington. 1981 Arqueologia de la America Andina. Editorial Milla Batres. Lima. 1993 Chavin de Huantar: Excavaciones en la Galeria de las Ofrendas. Materialen zur Allgemeinen und Vergleichenden Archaologie, Band 51. Verlag Philipp von Zabern. Mainz am Rhein. Lynch, Thomas F. 1981 Zonal Complementarity in the Andes: A History of the Concept. In Networks of the Past, Proceedings of the XII Chacmool Conference, edited by P. Francis, F. Kense and P. Duke, pp. 31-48. University of Calgary Archaeological Association. Calgary. 646 Lynch, Thomas F. (editor) 1980 Guitarrero Cave: Early Man in the Andes. Academic Press. New York. 1992 Summary of Investigations in Bolivar submitted by Warren Church. Current Research: Andean South America. American Antiquity 57 (1) :156. Lynch, Thomas F. and Susan Pollock 1981 La Arqueologia de la Cueva Negra de Chobshi. Miscelanea Antropol6gica Ecuatoriana No. 1, pp.92119. Guayaquil. Lyon, Patricia J. 1981 An Imaginary Frontier: Prehistoric Highland Lowland Interchange in the Southern Peruvian Andes. In Networks of the Past, Proceedings of the XII Chacmool Conference, edited by P. Francis, F. Kense and P. Duke, pp. 3-18. University of Calgary Archaeological Association. Calgary. MacNeish, RichardS., Thomas C. Patterson and David L. Browman 1975 The Central Peruvian Interaction Sphere. Papers of the Robert S. Peabody Foundation for Archaeology Vol. 7. Phillips Academy. Andover. MacNeish, RichardS., Angel Garcia Cook, Luis G. Lumbreras, Robert K. Vierra and Antoinette Nelken-Turner 1981 Prehistory of the Ayacucho Basin. Peru, Vol. II: Excavations and Chronology. University of Michigan Press. Ann Arbor. Maguifia c., Teofilo 1988 Chavin: La Epopeya Jamas Contada. Editorial "Impulse." Lima. Mariategui, Jose Carlos 1971 Seven Interpretive Essays on Peruvian Reality Translated by M. Urquidi. University of Texas Press. Austin. Mason, J. Alden 1950 The Languages of South American Indians. In Handbook of South American Indians, Vol. 6: Physical Anthropology, Linguistics and Cultural Geography of South American Indians. Edited by J. H. Steward, pp. 157-318. Bureau of American Ethnology Bulletin 143. Smithsonian Institution. Washington, D.C. 647 Masucci, Maria 1992 Ceramic Change in the Guanga1a Phase. Southwest Ecuador: A Typology and Chronology. Doctoral dissertation. Department of Anthropology, Southern Methodist University. University Microfilms International. Ann Arbor. Masuda, Shozo, Izumi Shimada and Craig Morris (editors} 1985 Andean Ecology and Civilization. University of Tokyo Press. Tokyo. de la Mata, Fr. Pedro 1923 Arte de la Lengua Cholona. Inca Vol. I, pp. 690-700. Lima. Matos, Ramiro and John W. Rick 1981 Los Recursos Naturales y el Poblamiento Preceramico de la Puna de Junin. Revista del Museo Nacional Torno 44, pp. 23-68. Lima. Matsumoto, Ryozo 1993 Dos Modos de Proceso Socio-Cultural: El Horizonte Temprano y el Periodo Intermedio Temprano en el Valle de Cajamarca. In El Mundo Ceremonial Andino. Edited by L. Millones andY. Onuki, pp. 169-202. Senri Ethnological Studies No. 37. National Museum of Ethnology. Osaka Mayer, Enrique 1985 Production Zones. In Andean Ecology and Civilization. Edited by S. Masuda, I. Shimada and C. Morris, pp. 45-84. University of Tokyo Press. Tokyo. McCown, Theodore D. 1944 Pre-incaic Huamachuco: Survey and Excavations in the Region of Huamachuco and Cajabamba. University of California Publications in American Archaeology and Ethnology Vol. 39, No. 4, pp. 223-400. University of California Press. Berkeley. McGreevy, Tom 1989 Prehispanic Pastoralism in Northern Peru. In The Walking Larder: Patterns of Domestication, Pastoralism and Predation. Edited by J. CluttonBrock, pp. 231-239. Unwin Hyman. London. McGreevy, Tom and Roxanne Shaughnessy 1983 High Altitude Land Use in the Huamachuco Area. In Investigations of the Andean Past. Edited by D. Sandweiss, pp. 226-242. Latin American Studies Program. Cornell University. Ithaca. 648 Means, Philip A. 1931 Ancient Civilizations of the Andes. Charles Scribner's Sons. New York. Meggers, Betty J. 1954 Environmental Limitation on the Development of Culture. American Anthropologist 56:801-824. 1966 Ecuador. Frederick A. Praeger, Inc. New York. 1971 Amazonia: Man and Culture in a Counterfeit Paradise. Aldine-Atherton. Chicago. 1975 Application of the Biological Model of Diversification to Cultural Distributions in Tropical Lowland South America. Biotropica 7 ( 3) : 141-61. 1977 Vegetational Fluctuation and Prehistoric Cultural Adaptation in Amazonia: Some Tentative Correlations. World Archaeology 8 (3) :287-303. 1979 Climatic Oscillation as a Factor in the Prehistory of Amazonia. American Antiquity 44 (2) :252-266. Meggers, Betty and Clifford Evans 1961 An Experimental Formulation of Horizon Styles in the Tropical Forest Area of South America. In Essays in Pre-Columbian Art and Archaeology. By S. K. Lothrop and others, pp. 372-388. Harvard University Press. Cambridge. 1983 Lowland South America and the Antilles. In Ancient South Americans. Edited by J. D. Jennings, pp. 287335. W. H. Freeman. San Francisco. Meggers, Betty J., Clifford Evans and Emilio Estrada 1965 Early Formative Period of coastal Ecuador: The Valdivia and Machalilla phases. Smithsonian Contributions to Anthropology, Vol. 1. Washington, D.C. Mejia X., M. Toribio 1956 Appendice. Historia de la Expedici6n Arqueologica al Marafion de 1937. In Arqueologia del Valle de Casrna. By J. C. Tello, pp. 321-337. Publicaci6n Antropol6gica del Archive "Julio C. Tello" de la Universidad Nacional Mayor de San Marcos, Vol. I. Lima. 649 Metraux, Alfred 1963 Jesuit Missions in South America. In Handbook of South American Indians. Vol. 5: The Comparative Ethnology of South American Indians [1949], edited by J. H. Steward, pp. 645-655. Bureau of American Ethnology Bulletin 143. Cooper Square Publishers. New York. Miasta G., Jaime 1979 El Alto Amazonas: Arqueologia de Jaen y San Ignacio, Perli. Seminario de Historia Rural Andina, Universidad Nacional Mayor de San Marcos. Lima Middendorf, Ernst 1895 Peru. Beobachtungen und Studien uber das Land und seine Bewohner wahrend eines 25 jahrigen Aufenhalts, III, Das Hochland von Peru. Robert Oppenheim. Berlin. Miller, George R. 1979 An Introduction to the Ethnoarchaeology of Andean Camelids. Ph.D. dissertation. Department of Anthropology, University of California at Berkeley. University Microfilms International. Ann Arbor. Miller, George and Richard L. Burger 1995 Our Father the Cayman, Our Dinner the Llama: Animal Utilization at Chavin de Huantar. American Antiquity 60 (3) :421-458. Miranda, Cristobal de 1906 Relaci6n de los corregimientos y otros oficios que se proveen en los reynos e provincias del Piru, en el distrito e gobernaci6n del vissorrey dellos [1583]. Juicio de Limites entre el Peru y Bolivia. Prueba Peruana. Torno I, pp. 168-280. Edited by Victor M. Maurtua. Imprenta de Henrich y Comp. Barcelona. 1925 Relaci6n hecha por el Virrey D. Martin Enriquez de los oficios que se provee en la gobernacion de los reinos y provincias de Peru, 1583. Gobernantes del Peru. Cartas y Papeles. Siglo XVI. Edited by Roberto Levillier, Torno IX, pp. 114-230. Imprenta de Juan Pueyo. Madrid. 650 Mogrovejo, Toribio Alfonso de 1921 Libra de Visitas 1593. Diario de la Segunda Visita Pastoral que Hizo de su Arquidiocesis el Ilustrisimo Senor Don Toribio Alfonso de Mogrovejo, Arzobispo de los Reyes (1593). Revista del Archive Nacional del PerU Torno II, Entrega I, pp. 37-77. San Martiny Ca. Lima. Mohr-Chavez, Karen L. 1980 The Archaeology of Marcavalle, An Early Horizon Site in the Valley of Cuzco, Peru. Part I. BaesslerArchiv, Neue Folge, Band XXVIII, pp. 203-329. Berlin. 1981a The Archaeology of Marcavalle, An Early Horizon Site in the Valley of Cuzco, Peru. Part II. BaesslerArchiv, Neue Folge, Band XXVIII, pp. 107-205. Berlin. 1981b The Archaeology of Marcavalle, An Early Horizon Site in the Valley of Cuzco, Peru. Part III. BaesslerArchiv, Neue Folge, Band XXVIII, pp. 241-386. Berlin. Morales, Edmundo 1995 The Guinea Pig: Healing. Food and Ritual in the Andes. University of Arizona Press. Tucson. Morales Ch., Daniel 1978 Excavaciones en In III Congreso Andina: Aetas y I, pp. 325-344. las Salinas de San Blas (Junin) . Peruano El Hombre y la Cultura Trabajos. Edited by R. Matos, Torno Editora Lasontay. Lima. 1992 Chambira: Alfareros Tempranos de la Amazonia Peruana. In Estudios de Argueologia Peruana. Edited by D. Bonavia, pp. 149-176. Fomciencias. Lima. 1993 Compendia Hist6rico del Perli Torno I: Historia Arqueol6gica del PerU. Editorial Milla Batres. Lima. Morris, Craig 1978 The Archaeological Study of Andean Exchange Systems. In Social Archaeology: Beyond Subsistence and Dating. Edited by C. Redman et al., pp. 315-327. Academic Press. New York. Moseley, Michael E. 1983 Central Andean Civilization. In Ancient South Americans. Edited by J. D. Jennings, pp. 179-240. W.H. Freeman and Co. San Francisco. 651 1992 The Incas and Their Ancestors: The Archaeology of Peru. Thames and Hudson. New York. Mujica, Elias 1984 Cerro Arena-Layzon: Relaciones Costa-Sierra en el Norte del PerU. Gaceta Arqueologica Andina 10:12-15. Lima. 1985 Altiplano-Coast Relationships in the South-Central Andes: From Indirect to Direct Complementarity. In In Andean Ecology and Civilization. Edited by S. Masuda, I. Shimada and C. Morris, pp.l03-140. University of Tokyo Press. Tokyo. Murra, John V. 1964 Una Apreciacion Etnologica de la Visita. In Visita Hacha a la Provincia de Chucuito por Garci Diez de San Miguel en el Afio 1567. Casa de la Cultura del Peru. Lima 1967 La Visita de los Chupachu como Fuente Etnol6gica. In Visita de la Provincia de Leon de Huanuco en 1562 Torno I. Edited by J. Murra. Universidad Nacional Hermilio Valdizan. Huanuco. 1972 El "Control Vertical" de un Maximo de Pisos Ecologicos en la Economia de las Sociedades Andinas. In Ifiigo Ortiz de Zuniga, Visita de la Provincia de Leon de Huanuco en 1562, 2:429-476. Universidad Nacional Hermilio Valdizan. Huanuco. 1975 Formaciones Economicas y Politicas del Mundo Andino. Institute de Estudios Peruanos. Lima. 1980 The Economic Organization of the Inka State. Research in Economic Anthropology Supplement 1. JAI Press. Greenwich, CT. Muscutt, Keith, Vincent R. Lee and Douglas Sharon 1994 Vira Vira: A Chachapoyas Site. Sixpac Manco Publications. Wilson, WY. Myers, Thomas P. 1976 Formative Period Occupations in the Highlands of Northern Ecuador. American Antiquity 41 (3) :353-360. 198la Aboriginal Trade Networks in Amazonia. In Networks of the Past, Proceedings of the XII Chacmool Conference, edited by P. Francis, F. Kense and P. Duke, pp. 19-30. University of Calgary Archaeological Association. Calgary. 652 1981b Ceramics from the Hacienda Tarapoto, Department of San Martin, Peru. Nawpa Pacha 19:155-166. Institute of Andean Studies. Berkeley. 1988 El Efecto de las Pestes Sobre las Poblaciones de la Amazonia Alta. Amazonia Peruana 8 (15) :61-82. Lima. Narvaez Vargas, Luis Alfredo 1988 Kuelap: Una Ciudad Fortificada en los Andes Nororientales de Amazonas, Peru. Argyitectura y Argueologia, compiled and edited by Victor Rangel Flores, pp.115-142. Consejo Nacional de Ciencia y Tecnologia. Chiclayo. NASA (National Aeronautics and Space Administration, USA) 1976 Landsat image ERTS E-2518-14285-7, June 23, 1976. Nieves, z., Maria n.d. Analisis de Ceramica Chorrera del Sitio Penon del Rio. Tesis titular. Escuela de Arqueologia, Escuela Superior Politecnica del Literal. Guayaquil. Niles, Susan A. 1987 Callachaca:Style and Status in an Inca Community. University of Iowa Press. Iowa City. Norton, Presley, Richard Lunnis and Nigel Nayling 1983 Excavaciones en Salango, Provincia de Manabi, Ecuador. Miscelanea Antropol6gica Ecuatoriana Vol. 3, pp. 9-72. Banco Central del Ecuador. Guayaquil. Nufiez del Prado, Oscar 1958 El Hombre y la Familia: su Matrimonio y su Organizaci6n Politico-social en Q'ero. Revista Universitaria No. 114, primer semestre. Cuzco. Oberem, Udo 1974 Trade and Trade Goods in the Ecuadorian Montana. In Native South Americans: Ethnology of the Least Known Continent. Edited by P. Lyon, pp. 346-358. Little, Brown and Co. Boston. 1980 Los Ouijos: Historia de la Transculturaci6n de un Grupe Indigena en el Oriente Ecuatoriano. Institute Otavaleno de Antropologia. Otavalo, Ecuador. Olivier, Jose 1992 Donald W. Lathrap: Approaches and Contributions in New World Archaeology. Journal of the Steward Anthropological Society 20 (1-2) :283-344. Urbana. 653 ONERN (Oficina Nacional de Evavulaci6n de Recursos Naturales) 1976 Mapa Ecol6qico del PerU. ONERN. Lima. Onuki, Yoshio 1985 The Yunga Zone in the Prehistory of the Central Andes: Vertical and Horizontal Dimensions in Andean Ecological and Cultural Processes. In Andean Ecology and Civilization. Edited by S. Masuda, I. Shimada and C. Morris, pp. 339-356. University of Tokyo Press. Tokyo. 1993 Las Actividades Ceremoniales Tempranas en la Cuenca del Alto Huallaga y Algunos Problemas Generales. In El Mundo Ceremonial Andino. Edited by L. Millones andY. Onuki, pp. 69-96. Senri Ethnological Studies No. 37. National Museum of Ethnology. Osaka. Osborn, Ann 1989 Multiculturalism in the Eastern Andes. In Archaeological Approaches to Cultural Identity. Edited by S. Shennan, pp. 141-156. Unwin Hyman. London. Otte, Marcel and Lawrence H. Keeley 1990 The Impact of Regionalism on Paleolithic Studies. Current Anthropology 31 (5) :577-582. Pardo, Luis A. 1957 Clasificaci6n de la Ceramica Cuzquefia del Antigua Imperio de los Incas, Parts 1 and 2. Historia y Arqueoloqia del Cuzco, (II:541-561, 565-596). Cuzco. Parker, Gary J. 1963 La Clasificaci6n Genetica de los Dialectos Quechuas. Revista del Museo Nacional Torno 32, pp. 242-252. Lima. Parsons, Jeffrey and Charles M. Hastings 1988 The Late Intermediate Period. In Peruvian Prehistory. Edited by R. W. Keatinge, pp. 190-229. Cambridge University Press. Cambridge. Patterson, Thomas C. 1966 Pattern and Process in the Early Intermediate Period Pottery of the Central Coast of Peru. University of California Publications in Anthropology Vol. 3. Berkeley. 654 1971 Chavin: An Interpretation of its Spread and Influence. In Dumbarton Oaks Conference on Chavin. Edited by Elizabeth Benson, pp.29-48. Washington D.C. 1988 Merchant Capital and the Formation of the Inca State. Dialectical Anthropology 12:217-227. 1991 Early Colonial Encounters and Identities in the Caribbean: A Review of Some Recent Works and their Implications. Dialectical Anthropology 16:1-13. Paulsen, Allison C. 1970 A Chronology of Guangala and Libertad Ceramics of the Santa Elena Peninsula in South Coastal Ecuador. Ph.D. dissertation. Faculty of Political Science, Columbia Universtiy. University Microfilms. Ann Arbor. 1977 Patterns of Maritime Trade Between South Coastal Ecuador and Western Mesoamerica, 1500 B.C.- A.D.600. In The Sea in the Pre-Columbian World. Edited by E. P. Benson, pp.141-160. Dumbarton Oaks Research Library and Collection. Washington. Pearsall, Deborah 1980 Pachamachay Ethnobotanical Report: Plant Utilization at a Hunting Base Camp. In Prehistoric Hunters of the High Andes. Edited by J. W. Rick, pp. 191-232. Academic Press. 1989 Paleoethnobotany. A Handbook of Procedures. Academic Press. San Diego. 1992 The Origins of Plant Cultivation in South America. In The Origins of Agriculture: An International Perspective. Edited by C. W. Cowan and P. J. Watson, pp. 173-205. Smithsonian Institution Press. Washington D.C. Pearson, G. W. and M. Stuiver 1993 High-precision Bidecadal Calibration of the Radiocarbon Time Scale 500-2500 B.C. Radiocarbon 35: 25-33. Pefia Mesa, E. 1935 Pajaten. In "Monografia de la Provincia de Huallaga (Continuaci6n) . " By Enrique de las Casas. Boletin de la Sociedad Geografica de Lima Torno LII, Trimestre cuarto, pp. 343-344. Lima. 655 Perez c., Ismael 1988 Monumentos Arqueologicos de Santiago de Chuco, La Libertad. Boletin de Lima No. 60, pp.33-44. Lima. Perez v., Jose Luis 1969 Petroglifos en Pataz. Boletin del Seminario de Argueologia Afio I, No. 1, pp. 97-100. Instituto Riva Agliero, Pontificia Universidad Catolica del Perli. Lima. Peterson, Emil 1984 Morteros Ceremoniales: the Early Development and Distribution of a Decorated Stone Bowl Tradition in Northwest South America. In Social and Economic Organization in the Prehistoric Andes. Edited by D. L. Browman, Burger, R. L. and M. A. Rivera, pp. 2131. BAR International Series No. 194. Manchester. Pimentel G., Victor 1967 "Pajaten". Fenix, Revista de la Biblioteca Nacional 17:34-38. Lima. Plowman, T. 1984 The Origin, Evolution and Diffusion of Coca, Erythroxylum spp., in South and Central America. In Pre-Columbian Plant Migrations. Edited by D. Stone, pp. 125-163. Papers of the Peabody Museum of Archaeology and Ethnology No. 76. Cambridge. Porras G., Pedro I. 1974 Historia y Argueologia de la Ciudad Espanola Baeza de los Ouijos. Siglo XVI. Centro de Publicaciones de la Pontificia Universidad Catolica del Ecuador. Quito. 1975a Fase Cosanga. Estudios Cientificos sobre el Oriente Ecuatoriano, Torno II. Centro de Publicaciones de la Pontificia Universidad Catolica del Ecuador. Quito. 1975b Fase Pastaza: El Formativo en el Oriente Ecuatoriano. Separata de la Revista de la Universidad Catolica Afio III, No. 10, pp. 75-134. Quito. 1982 Argueologia de Quito. Fase Cotocollao. Centro de Investigaciones Arqueologicas, Universidad Catolica del Ecuador. Quito. 1987 Investigaciones Arqueologicas a las Faldas del Sangay: Tradicion Upano. Quito. 656 1989 Investigations at the Sangay Mound Complex. National Geographic Research 5 (3) :374-381. Pozorski, Thomas 1983 The Caballo Muerto Complex and its Place in the Andean Chronological Sequence. Annals of Carnegie Museum Vol. 52, Article 1, pp. 1-40. Proulx, Donald 1982 Territoriality in the Early Intermediate Period: the Case of Moche and Recuay. Nawpa Pacha No. 20, pp. 83-96. Institute of Andean Studies. Berkeley. 1985 An Analysis of the Early Cultural Sequence in the Nepena Valley. Research Report Number 25. Department of Anthropology. University of Massachusetts, Amherst. Amherst. Pulgar Vidal, Javier 1987 Geografia del Peru. Novena edici6n. PEISA. Lima. Raimondi, Antonio 1876 El PerU. Torno II. Imprenta del Estado. Lima. 1900a Itinerario de los Viajes de Raimondi en el Peru: De Cajamarca a ... Bambamarca (1860). Boletin de la Sociedad Geoqrafica de Lima Torno X, Nos. 1, 2 y 3, pp. 1-40. Lima. 1900b Itinerario de los Viajes de Raimondi en el Peru: De Bambamarca a ... Huallaga y Regreso por Pizana (1860). Boletin de la Sociedad Geoqrafica de Lima Torno X, Nos. 4, 5 y 6, pp. 123-171. Lima. 1942 Notas de Viaj e para su Obra El PerU. Imprenta Torres Aguirre. Lima. II II Torno I. Ramp6n S., Lino M. 1959 Sitio Argueol6qico F.P. Cuadernos de Investigaciones Cientificas No. 1. Misiones Cat6licas de la Amazonia. Quito. Ravines, Rogger 1967-68 Review of Las Ruinas del Abiseo, by Duccio Bonavia. Revista del Museo Nacional. Torno 35, pp. 335-336. Lima. 1971 Grupos de Tradici6n Cazadora en las Tierras Altas de Huancavelica, Peru. Revista del Museo Nacional Torno 37, pp. 17-27. Lima 657 1972 Los Caciques de Pausamarca: Alga Mas sabre las Etnias de Chachapoyas. Historia y Cultura No. 6, pp. 217-248. 1978 Antiguos Sitios de Ocupacion en el Rio Huayabamba, Perli. Historia. Problema y Promesa. Homenaje a Jorge Basadre Vol. 1, pp. 521-532. Pontificia Universidad Cat6lica del Perli. Lima. 1981a Los Reines y Sefiorios Locales de los Andes Centrales: 800-1470 DC. In Historia del Peru. Torno II Perli Antiquo, pp. 93-184. Editorial Juan Mejia Baca. Lima. 1981b Yacimientos Arqueologicos de la Region Nororiental del Peru. Amazonia Peruana 4 (7) :139-175. Lima. 1982 Panorama de la Argyeologia Andina. Institute de Estudios Andinos. Lima. 1986-87 Colan: Evidencias Arqueologicas. Revista del Museo Nacional Torno 48, pp. 55-118. Lima. Raymond, J. Scott 1972 The Cultural Remains from the Granja de Sivia, Peru: An Archaeological Study of Tropical Forest Culture in the Montana. Ph.D. dissertation, Department of Anthropology, University of Illinois-Urbana. University Microfilms International. Ann Arbor. 1976 Late Prehistoric and Historic Settlements in the Upper Montana of Peru. In Canadian Archaeology Abroad, edited by P. L. Shinnie, J. H. Robertson and F. Kense, pp. 205-213. University of Calgary Archaeological Association. Calgary. 1985 Quechuas y Chunchos: Ethnic Boundaries in Eastern Peru. In Status. Structure and Stratification: Current Archaeological Reconstructions, Proceedings of the Sixteenth Annual Chacmool Conference, edited by M. Thompson, M. T. Garcia and F. J. Kense, pp. 39-50. The Archaeological Association of the University of Calgary. Calgary. 1988 A View from the Tropical Forest. In Peruvian Prehistory. Edited by R. W. Keatinge, pp. 279-300. Cambridge University Press. 658 1992 Highland Colonization of the Peruvian Montana in Relation to the Political Economy of the Huari Empire. Journal of the Steward Anthropological Society 20 (1-2) :17-36. Urbana. Raymond, J. Scott, Warren R. DeBoer and Peter G. Roe 1975 Cumancaya: A Peruvian Ceramic Tradition. Occasional Paper No. 2, Department of Archaeology, University of Calgary. Calgary. Raymond, J. Scott, Augusto Oyuela-Caycedo and Patrick Carmichael 1994 Una Comparaci6n de las Tecnologias de la Ceramica Temprana de Ecuador y Colombia. In Tecnologia y Organizaci6n de la Producci6n Ceramica Prehispanica en los Andes. Edited by I. Shimada, pp. 33-52. Pontificia Universidad Cat6lica del Perti, Fonda Editorial. Lima. Reeve, Mary-Elizabeth 1994 Regional Interaction in the Western Amazon: the Early Colonial Encounter and the Jesuit Years. Ethnohistory 41 (1) :106-138. Reichlen, Henry and Paule 1950 Recherches Archeologiques dans les Andes du Haut Utcubamba, Journal de la Societe des Arnericanistes, 39:219-246. Paris. Renard-Casevitz, F. M., Th. Saignes and A. C. Taylor 1988 Al Este de los Andes: Relaciones entre las Sociedades Amaz6nicas y Andinas entre los Siglos XV y XVII. 2 Volumes. Institute Frances de Estudios Andinos. Ediciones ABYA-YALA. Quito. Renfrew, Colin 1975 Trade as Action at a Distance: Questions of Integration and Communication. In Ancient Civilization and Trade. Edited by J. A. Sabloff and C. C. Lamberg-Karlovski, pp. 3-59. University of New Mexico Press. 1987 Archaeology and Language: The Puzzle of IndoEuropean Origins. Jonathan Cape. London. Rice, DonS., Charles Stanish and Phillip R. Scarr (editors) 1989 Ecology. Settlement and History in the Os~ore Drainage. Peru. Parts 1 and 2. BAR International Series 545. London. 659 Rice, Prudence M. 1987 Pottery Analysis: A Sourcebook. University of Chicago Press. Chicago. Richardson, III, James B. 1973 The Preceramic Sequence and the Pleistocene and Post-Pleistocene Climate of Northwest Peru. In Variation in Anthropology: Essays in Honor of John C. McGregor. Edited by D. W. Lathrap and J. Douglas, pp. 199-211. Illinois Archaeological Survey. Urbana. Rick, John W. 1980 Prehistoric Hunters of the High Andes. Academic Press. New York. 1988 The Character and Context of Highland Precerarnic Society. In Peruvian Prehistory. Edited by R. W. Keatinge, pp. 3-40. Cambridge University Press. Rios, Manuel A., C. F. Ponce, P. Vasquez and A. Tovar 1982 Planificaci6n Para el Establecimiento de Unidades de Conservaci6n en el Bosque Nublado del Noreste del Perli. Informe Final del Proyecto WWF-IUCN/1792. Departamento Manejo Forestal, Universidad Nacional Agraria La Molina. Lima. de la Riva Herrera, Martin 1907 Testimonio de lo que Ha Obrado el Gobernador Don Martin de la Riva Herrera en la Reducci6n de Differentes Provincias de Indios [1655]. In Juicio de Limites del Perli: Prueba Peruana. Presented by V. M. Maurtua. Vol. 3, pp. 259-346. Imprenta de Henrich y Comp. Barcelona. Rivera, Mario A. 1975 Una Hip6tesis sobre Movimientos Poblacionales Altiplanicos y Transaltiplanicos a las Costas del Norte de Chile. Chungara No. 5, pp. 7-32. Universidad del Norte, Depto. de Antropologia. Arica. 1991 The Prehistory of Northern Chile: A Synthesis. Journal of World Prehistory 5:1:1-47. Rivera, Mario A. and Francisco Rothhammer 1986 Evaluaci6n Biol6gica y Cultural de Poblaciones Chinchorro: Nuevas Elementos para la Hip6tesis de Contactos Transaltiplanicos, Cuenca Amazonas-Costa Pacifico. Chungara No. 16-17, pp. 295-306. Universidad de Tarapaca. Arica. 660 Rivera S. , Raul 1957 Libro Primero de Cabildos de San Juan de la Frontera de Chachapoyas. Penix. Revista de la Biblioteca Nacional 12:280-336. Lima. Rivet, Paul 1968 Les Elements Constitutifs des Civilisations du NordOuest et de L'Ouest Sud-americain. International Congress of Americanists, pp. 1-20. Gothenburg 1924. Kraus Reprint. Nendeln/Liechtenstein. Robinson, Roger W. 1994 Recent Excavations at Hacha in the Acari Valley, Peru. Andean Past 4:9-37. Latin American Studies Program. Cornell University. Ithaca. Rodbell, Donald 1991 Late Quaternary Glaciation and Climatic Change in the Northern Peruvian Andes. Ph.D. dissertation. Department of Geology, University of Colorado, Boulder. Rodbell, Donald T. and Barbara C. S. Hansen n.d. A Late Pleistocene-Holocene Palynologic and Sedimentologic Record from the Eastern Andes of Northern Peru. Unpublished manuscript in possession of the authors. Rojas Ponce, Pedro 1966 Un Informe sobre las Ruinas del Pajaten. Cuadernos Americanos 148 (5) :119-127. Mexico City. 1967 The Ruins of Pajaten, Archaeology 20 (1) :9-17. 1969 La "Huaca" Huayurco, Jaen Cajamarca. Boletin del Seminario de Arqueologia 4:48-56. Institute Riva Agliero, Pontificia Universidad Cat6lica del Peru. Lima. Roosevelt, Anna C. 1980 Parmana: Prehistoric Maize and Manioc Subsistence along the Amazon and Orinoco. Academic Press. New York. 1995 Early Pottery in the Amazon: Twenty Years of Scholarly Obsurity. In The Emergence of Pottery: Technology and Innovation in Ancient Societies. Edited by W. K. Barnett and J. W. Hoopes, pp. 115132. Smithsonian Institution Press. Washington, D.C. 661 Roosevelt, Anna C., R. A. Housley, M. Imazio da Silviera, S. Maranca and R. Johnson 1991 Eighth Millennium Pottery from a Prehistoric Shell Midden in the Brazilian Amazon. Science 254:16211624. Rosas, Hermilio 1976 Investigaciones Arqueol6gicas en la Cuenca del Chotano. Aetas del XLI Congreso de Americanistas, 1974, Vol. 3, pp. 564-578. Mexico. Rosas, Hermilio and Ruth Shady 1970 Pacopampa: Un Complejo Temprano del Periodo Formative Peruano. Argueologia y Sociedad. No. 3, pp. 3-16. Museo de Arqueologia y Etnologia de la Universidad Nacional Mayor de San Marcos. Lima. 1974 Sobre el Periodo Formative en la Sierra del Extreme Norte del Perli. Argueol6gicas 15:6-35. Institute Nacional de Cultura. Lima Rostoker, Arthur G. 1988 An Archaeological Collection from Eastern Ecuador. Unpublished Master's thesis. Department of Anthropology, San Francisco State University. Rostworowski de Diez Canseco, Maria 1977 Coastal Fishermen, Merchants, and Artisans in PreHispanic Peru. In The Sea in the Pre-Columbian World. Edited by E. P. Benson, pp. 167-183. Dumbarton Oaks Research Library and Collection. Washington. Rowe, John H. 1946 Inca Culture at the Time of Spanish Conquest. In Handbook of South American Indians. Bureau of American Ethnology Bulletin, 143, Vol. 2, Part 2, edited by J. H. Steward, pp.182-330. Washington, D.C. Rowe, John H. and Dorothy Menzel 1967 Introduction. Peruvian Archaeology. Selected Readings, edited by J. Rowe and D. Menzel, pp.v-x. Peek Publications. Palo Alto. Rozenberg, Catherine 1982 Le Materiel Archeologique de Piruru II: La Collection Louis Girault. Bulletin de l'Institut Francais d'Etudes Andines 11 (3-4) :115-141. Lima. 662 Rozenberg, Catherine and Maurice Picon 1990 Circulation, Echange et Production de Poteries dans les Andes Centrales au Deuxieme Millenaire Av. J.C. Bulletin de l'Institut Francais d'Etudes Andines 19 (1) :1-14. Lima. Rouse, Irving 1958 The Inference of Migrations from Anthropological Evidence. In Migrations in New World Culture History Edited by R. H. Thompson, pp. 63-68. University of Arizona Press. Tucson. 1986 Migrations in Prehistory: Inferring Population Movement from Cultural Remains. Yale University Press. New Haven. Rouse, Irving and Jose Cruxent 1963 Venezuelan Archaeology Yale University Press. New Haven. Ruiz Estrada, Arturo 1969 Alfareria del Estilo Huari en Cuelap. Boletin del Seminario de Arqueologia, No. 4, pp. 60-65. Pontificia Universidad Cat6lica del Peru. Lima. 1972 La Alfareria de Cuelap: Tradici6n y Cambio, Bachelor's thesis, Universidad Nacional Mayor de San Marcos, Lima. Salazar, Ernesto 1983 El Hombre Temprano en el Ecuador. In Nueva Historia del Ecuador. Vol. 1: Epoca Aborigen. Edited by E. Ayala Mora, pp. 73-128. Corporaci6n Editora Nacional Grijalbo. Salomon, Frank 1985 The Dynamic Potential of the Complementarity Concept. In Andean Ecology and Civilization. Edited by S. Masuda, I. Shimada and C. Morris, pp. 511-532. University of Tokyo Press. Tokyo. 1986 Native Lords of Quito in the Age of the Incas: The Political Economy of North Andean Chiefdoms. Cambridge University Press. Cambridge. Sahlins, Marshall D. 1961 The Segmentary Lineage: An Organization of Preditory Expansion. American Anthropologist 53:322-345. Sauer, Carl 1952 Agricultural Origins and Dispersals. American Geographical Society. New York. 663 Savoy, Gene 1965 El Gran Pajaten Expedition: a Lost Pre-Inca Civilization in the Eastern Andes. Andean Air Mail and Peruvian Times Vol.25, No. 1294. Lima. 1970 Antisuyo: The Search for the Lost Cities of the Amazon, Simon and Schuster, New York. Schaedel, Richard P. 1948 Stone Sculpture in the Callejon de Huaylas. In b Reappraisal of Peruvian Archaeology. Edited by W. C. Bennett, pp. 66-79. Memoirs of the Society for American Archaeology, No. 4. Menasha. Schiffer, Michael B. 1983 Toward the Identification of Formation Processes. American Antiquity 48 (4) :675-706. Schortman, Edward M. and Patricia A. Urban 1987 Modeling Interregional Interaction in Prehistory. In Advances in Archaeological Method and Theory Vol. 11, pp. 37-95. Academic Press. New York. 1992 Current Trends in Interaction Research. In Resources, Power. and Interregional Interaction. Edited by E. M. Schortman and P. Urban, pp. 235-255. Plenum Press. New York. Schjellerup, Inge 1979-80 Documents on Paper and in Stone: A Preliminary Report on the Inca Ruins in Cochabamba, Province of Chachapoyas, Peru. Folk 21-22:299-311. 1984 Cochabamba: An Incaic Administrative Centre in the Rebellious Province of Chachapoyas. Current Research in the Andes. Edited by A. Kendall. BAR International Series No. 210, pp. 161-187. Oxford. 1985 Observations on Ridged Fields and Terracing Systems in the Northern Highlands of Peru. Tools and Tillage Vol. V, No. 2, pp. 100-121. Copenhagen. 1990 Recherches Archeologiques et Historiques au Chachapoyas - Perou. In Inca-Peru: 3000 Ans d'Histoire, pp. 236-247. Musees Royaux d'Art et d'Histoire. Imschoot, uitgevers. Bruxelles. 1992 Patrones de Asentamiento en las Faldas Orientales de los Andes de la Region de Chachapoyas. Estudios de Argueologia Peruana. Edited by D. Bonavia, pp. 355373. Fomciencias. Lima. 664 Science News 1985 A 'Lost and Found' City in Peru Gets New Perspective. Science News 127 (8) :117. Washington, D.C. Sechlin C., Jorge 1986 Patrones de Asentamiento en e1 Proceso Cultural Prehispanico del Valle de Cajamarca. Serie: Materiales para la Arqueo1ogia de Cajamarca-3. Editorial Sudamericana. Trujillo, Perli Seltzer, G. and C. Hastorf 1990 Climatic Change and its Affect on Prehispanic Agriculture in the Central Andes. Journal of Field Archaeology. 17:397-414. Shady, Ruth 1976 Investigaciones Arqueologicas en la Cuenca del Utcubamba. Aetas del XLI Congreso de Americanistas, 1974. Vol. 3, pp.579-589. 1983 Una Aproximaci6n al Mundo de las Creencias Andinas: la Cultura Pacopampa. Patrones de Enterramientos durante el Formative en la Sierra Norte del Peru. Museo Nacional de Antropologia y Arqueologia Boletin No. 8, pp. 17-24. Lima. 1987a La Arqueologia del Norte Peruano. In II Seminario de Investigaciones Sociales en la Region Norte, pp.7191. Concytec. Lima. 1987b Tradici6n y Cambia en las Sociedades Formativas de Bagua, Amazonas, Perli. Revista Andina 5 (2) :457-488. 1993 Del Arcaico al Formative en los Andes Centrales. Revista Andina 11 (1) :103-132. Shady, Ruth and Hermilio Rosas 1979 El Complejo Bagua y el Sistema de Establecimientos durante el Formative en la Sierra Norte del Peru. Nawpa Pacha No. 17, pp. 109-142. Institute of Andean Studies. Berkeley. 1982 Algunas Consideraciones sobre las Relaciones de Bagua con otros Centres Formativos del Area Andina. In Primer Simposio de Correlaciones Antropol6gicas Andino-Mesoamericano, Edited by J. Marcos and P. Norton, pp. 383-398. Escuela Superior Politecnica del Literal, Guayaquil. 665 Shimada, Izumi 1987 Horizontal and vertical Dimensions of Prehistoric States in North Peru. In The Origin and Development of the Andean State. Edited by J. Haas, S. Pozorski and T. Pozorski, pp. 130-144. Cambridge University Press. Cambridge. Shimada, Izumi and Adriana Maguifia 1994 Nueva Vision sobre la Cultura Gallinazo y su Relaci6n con la Cultura Moche. In Moche: Propuestas y Perspectivas. Edited by S. Uceda and E Mujica. Travaux de l'institut FranGais d'Etudes Andines No. 79, pp. 31-58. Lima. Shimada, Izumi, Carlos Elera and Victor Chang J. 1990 Excavaciones en Hornos de Ceramica de la Epoca Formativa en Batan Grande, Costa Norte del Perli. Gaceta Argyeol6gica Andina 5 (20) :19-44. Lima. Shimada, Izumi, Carlos Elera y Melody Shimada 1982 Excavaciones Efectuadas en el Centro de la Huaca Lucia-Ch6lope, del Horizonte Temprano, Batan Grande, Costa Norte del Peru: 1979-1981. Argyeol6gicas 19:109-208. Lima. Shimada, Melody 1982 Zooarchaeology of Huacaloma: Behavioral and Cultural Implications. In Excavations at Huacaloma in the Cajarnarca Valley. Peru. 1979: Report of the Japanese Scientific Expedition to Nuclear America. Edited by K. Terada andY. Onuki, pp. 303-336. University of Tokyo Press. Tokyo. 1985 Continuities and Changes in Patterns of Faunal Resource Utilization: Formative through Cajarnarca Periods. In The Formative Period in the Cajarnarca Basin. Peru: Excavations at Huacaloma and Layzon, 1982. Edited by K. Terada andY. Onuki, pp. 289-310. University of Tokyo Press. Tokyo. Silva S. Jorge 1988 La Alfareria de Pacharnachay, Junin. Boletin de Lima 57:21-30. Lima. Silverman, Helaine 1994 Paracas in Nazca: New Data on the Early Horizon Occupation of the Rio Grande Drainage, Peru. Latin American Antigyity 5 (4) :359-382. Snow, Dean 1995 Migration in Prehistory: The Northern Iroquoian Case. American Antiquity 60 (1) :59-79. 666 Solecki, Ralph S. 1979 Contemporary Kurdish Winter-time Inhabitants of Shanidar Cave, Iraq. World Archaeology 10 (3) :318330. Sonin, Robert 1977 Appendix: The Processes Involved in the Production of Iridescent-Painted Pottery found Archaeologically on the South Coast of Ecuador. In The Sea in the Pre-Columbian World. Edited by E. P. Benson, pp. 141-160. Dumbarton Oaks Research Library and Collection. Washington, D.C. Spinden, Herbert J. 1917 The origin and Distribution of Agriculture in America. Proceedings of the XIX International Congress of Americanists, pp. 269-276. Washington, D.C. Stanish, Charles 1992 Ancient Andean Political Economy. University of Texas Press. Austin. Stark, Louisa R. 1985 Ecuadorian Highland Quechua: History and Current Status. In South American Indian Languages. Edited by H. E. M. Klein and L. R. Stark, pp. 443-481. University of Texas Press. Austin. Steward, Julian H. 1948 Tribes of the Montana: An Introduction. In Handbook of South American Indians, Vol. 3: The Tropical Forest Tribes. Edited by J. H. Steward, pp.507-533. Bureau of American Ethnology Bulletin 143. Smithsonian Institution. Washington, D.C. 1949 South American Cultures: An Interpretive Essay. In Handbook of South American Indians. Vol. 5: The Comparative Ethnology of South American Indians, edited by J. H. Steward, pp. 669-772. Bureau of American Ethnology Bulletin 143. Washington, D.C. Steward, Julian H. and Louis C. Faron 1959 Native Peoples of South America. McGraw Hill. New York. Steward, Julian H. and Alfred Metraux 1948 Tribes of the Peruvian and Ecuadorian Montana. In Handbook of South American Indians, Vol. 3: The Tropical Forest Tribes. Edited by J. H. Steward, pp.535-656. Bureau of American Ethnology Bulletin 143. Smithsonian Institution. Washington, D.C. 667 Stewart, T. D. and Marshall T. Newman 1950 Anthropometry of South American Indian Skeletal Remains. In Handbook of South American Indians, Vol. 6: Physical Anthropology, Linguistics and Cultural Geography of South American Indians. Edited by J. H. Steward, pp.19-42. Bureau of American Ethnology Bulletin 143. Smithsonian Institution. Washington, D.C. Stothert, Karen E. 1993 Un Sitio de Guanga1a Temprano en el Suroeste del Ecuador. Smithsonian Institution and Banco Central del Ecuador. Washington, D.C. and Guayaquil. Straus, Lawrence Guy 1979 Caves: A Paleoanthropological Perspective. World Archaeology 10 (3):331-339. 1990 Underground Archaeology: Perspectives on Caves and Rockshe1ters. In Archaeological Method and Theory Vol. 2. Edited by M. B. Schi~ier, pp. 255-304. University of Arizona Press. Strong, William Duncan and Clifford Evans, Jr. 1952 Cultural Stratigraphy in the Viru Valley, Northern Peru: The Formative and Florescent Epochs. Columbia University Press, New York. Stuiver, M. and G. W. Pearson 1993 High-precision Bidecadal Calibration of the Radiocarbon Time Scale, A.D. 1950-500 B.C. and 2500000 B.C. Radiocarbon 35:1-23. Sued-Badillo, Jalil 1992 Facing Up to Caribbean History. American Antiquity 57 (4) :599-607 Tarnawiecki, M. C. 1927 Monografia Minera de la Provincia de Patas. Boletin de la Sociedad Geograf;ca de Lima. Torno XLIII, Trimestre segundo de 1926, pp. 165-182. Lima. Tellenbach, Michael 1981 Volbericht tiber die erste Kampagne der Ausgrabung bei Montegrande im Jequetepeque-Tal, Nordperu. Beitrage zur Allgemeinen und Vergleichende Archaeologie. 3:415-435. Munchen. Tello, Julio C. 1921 Introducci6n a la Historia Antigua del Peru. Editorial Euforion. Lima. 668 1923 Wira Kocha. Inca Vol. I, No. 1, pp. 93-320; No. 3, pp. 583-606. Organa del Museo de Arqueologia de la Universidad Mayor de San Marcos. Lima. 1929 Antiquo PerU: Primera Epoca. Editado para la Comisi6n Organizadora del Segundo Congreso Sudamericano de Turismo. Lima. 1938 Objeto y Prop6sito de la Expedici6n Arqueol6gica al Marafion. Boletin de la Sociedad Geografica de Lima. Torno LV, Trimestre segundo y tercero, pp. 123-131. Lima. 1942 Origen y Desarrollo de las Civilizaciones Prehist6ricas Andinas. Aetas y Trabajos Cientificos del XVII Congreso Internacional de Americanistas, Lima 1939. Torno I, pp. 589-720. Lima. 1943 Discovery of the Chavin Culture in Peru. American Antigyity 9:1:135-160. 1960 Chavin: Cultura Matriz de la Civilizaci6n Andina. Publicaci6n Antropol6gica del Archive "Julio C. Tello" de la Universidad Nacional Mayor de San Marcos, Volumen 2. Lima. 1967 Paginas Escoaidas. Edited by T. Mejia X. Universidad Nacional Mayor de San Marcos. Lima. Terada, Kazuo 1979 Excavations at La Pampa in the North Highlands of Peru. 1975, Report No. 1 of the Japanese Scientific Expedition to Nuclear America. University of Tokyo Press, Tokyo. Terada, Kazuo and Yoshio Onuki (Editors) 1982 Excavations at Huacaloma in the Cajamarca Valley, Peru. 1979, Report No. 2 of the Japanese Scientific Expedition to Nuclear America. University of Tokyo Press, Tokyo. 1985 The Formative Period in the Cajamarca Basin. Peru: Excavations at Huacaloma and Layzon. 1982. Report 3 of the Japanese Scientific Expedition to Nuclear America, University of Tokyo Press. Tokyo. 1988 Las Excavaciones en Cerro Blanco y Huacaloma. Cajamarca. Peru. 1985. Andes Chosashitsu. Tokyo. Tessmann, Gililter 1930 Die Indianer Nordost-Perus. Friederichsen, de Gruyter & Co. M.B.H. Hamburg. 669 Thatcher, John P. 1972 Continuity and Change in the Ceramics of Huamachuco, North Highlands. Peru. Ph.D. dissertation, University of Pennsylvania. University Microfilms International. Ann Arbor. 1972-74 Early Intermediate Period and Middle Horizon 1B Ceramic Assemblages of Huamachuco, North Highlands, Peru. Nawpa Pacha Nos. 10-12, pp. 109-127, Institute of Andean Studies. Berkeley. 1979 Early Ceramic Assemblages from Huamachuco, North Highlands, Peru. Nawpa Pacha, No. 17, pp. 91-106. Institute of Andean Studies. Berkeley. Thomas, David J. 1981 Cultural Homogeneity, Resource Distribution and Exchange Patterns in the Guiana Highlands of South America. In Networks of the Past, Proceedings of the XII Chacmool Conference, edited by P. Francis, F. Kense and P. Duke, pp. 85-100. University of Calgary Archaeological Association. Calgary. Thompson, Donald E. 1973 Archaeological Investigations in the Eastern Andes of Northern Peru, XL Congresso Internazionale Degli Americanisti 1:363-369. Roma-Genova. 1976 Prehistory of the Uchucmarca Valley in the North Highlands of Peru. Aetas del XLI Congreso Internacional de Americanistas Vol. 2, pp. 99-106. Mexico City. 1980 The Precolumbian and Colonial Heritage of Rapayan. Archaeology 32 (2} :45-51. 1984 Ancient Highland Connections with Selva and Coast: Evidence from Uchucmarca, Peru. Current Research in the Andes. Edited by A. Kendall, pp. 153-160. BAR International Series, 194:153-160. Oxford. Thompson L., E. Mosley-Thompsen, M. Davis, P. Lin, K. Henderson, J. Cole-Dai, J. Bolzan and K. Liu 1995 Late Glacial Stage and Holocene Tropical Ice Core Records from Huascaran, Peru. Science 269:46-50. Thompson, Raymond (Editor} 1958 Migrations in New World Culture History. University of Arizona Press. Tucson. 670 Topic, John R. and Theresa Lange Topic 1982 Huamachuco Archaeological Project Preliminary Report on the First Season. July-August 1981. Unpublished report on file at the Department of Anthropology, Trent University. Peterborough. 1983a Coast-Highland Relations in Northern Peru: Some Observations on Routes, Networks and Scales of Interaction. Civilization in the Ancient Americas, Essays in Honor of Gordon R. Willey, edited by R. M. Leventhal and A. L. Kalata, pp. 237-260. University of New Mexico Press and Peabody Museum. 1983b Huamachuco Archaeological Project Preliminary Report on the Second Season. June-August 1982. Unpublished report on file at the Department of Anthropology, Trent University. Peterborough. 1985 Coast-Highland Relations in Northern Peru: the Structure and Strategy of Interaction. Status, Structure and Stratification. Current Archaeological Reconstructions, Proceedings of the Sixteenth Annual Chacmool Conference, edited by M. Thompson, M. T. Garcia and F. J. Kense, pp. 55-65. The Archaeological Association of the University of Calgary. Calgary. Topic, Theresa Lange 1987 The Kaolin Ceramic Tradition in Northern Peru. Paper presented at the Fourth Annual Northeast Conference on Andean Archaeology and Ethnohistory, SUNY-Albany, November 2-3, 1985. Revised November, 1987. Topic, Theresa Lange and John R. Topic 1987 Huamachuco Archaeological Project: Preliminary Report on the 1986 Field Season. Trent University Occasional Papers in Anthropology No.4. Peterborough. Torero, Alfredo 1964 Los Dialectos Quechuas. Annales Cientificos de la Universidad Agraria Vol. II, No. 4, pp. 447-478. Lima. 1974 El Ouechua y la Historia Social Andina. Universidad Ricardo Palma. Lima. Torres c., German 1903 La Via Fluvial del Huayabamba. Boletin del la Sociedad Geografica de Lima. Torno XIII, Trirnestre tercero, pp. 295-309. Lima. 671 Tosi, Joseph A. Jr. 1960 Zonas de Vida Natural en el Perli. Institute Interamericano de Ciencias Agricolas, Boletin Tecnico No. 5, Lima. Trigger, Bruce 1989 A History of Archaeological Thought. Cambridge University Press. Cambridge. Troll, Carl 1958 Las Culturas Superiores Andinas y el Medio Geografico [1931]. Monografias y Ensayos Geograficos No. 1. Facultad de Letras, Universidad Nacional Mayor de San Marcos. Lima. Turner, Christy G. and Laurel Lofgren 1966 Household Size of Prehistoric Western Pueblo Indians. Southwestern Journal of Anthropology 22:117-132. UC and FPCN (University of Colorado and the Fundaci6n Peruana para la Conservaci6n de la Naturaleza) 1991 Plan Operative 1990-1992. Parque Nacional Rio Abiseo. Volumen I. Parques Nacionales del Peru. Lima. Ulbert, Cornelius 1994 Die Keramik der forrnativzeitlichen Siedlung Montegrande, Jequetepequetal, Nord-Peru. Materialen zur Allgemeinen und Vergleichenden Archaologie, Band 52. Verlag Philipp von Zabern. Mainz am Rhein. Uhle, Max 1922a Influencias Mayas en el Alto Ecuador. Boletin de la Academia Nacional de Historia, Vol. IV, Nos. 10-11, pp. 202-240. Quito. 1922b Origenes Centroamericanos. Boletin de la Academia Nacional de Historia, Vol. IV, No. 9, pp. 1-6. Quito. Unanue, Hipolito and Manuel Sobreviela 1963 Historia de las Missiones de Cajarnarguilla v Reducci6n de la de Manoa. Biblioteca Tenanitla: Libros Espafioles e Hispanoarnericanos 6. Ediciones Jose Porrua Turanzas. Madrid. Valcarcel, Luis E. 1964 Etnohistoria del Peru AntiqUo. Universidad Nacional Mayor de San Marcos. Departamento de Publicaciones. Lima. 672 Valencia, Niels 1992 Los Bosques Nublados Secos de la Vertiente Occidental de los Andes del PerU. In Memorias del Museo de Historia Natural. Edited by K. R. Young and N. Valencia, No. 21, pp. 155-170. Universidad Nacional Mayor de San Marcos. Lima. Van Buren, Mary 1996 Rethinking the Vertical Archipelago: Ethnicity, Exchange, and History in the South Central Andes. American Anthropologist 98 (2) :338-351. Vargas Ugarte, Ruben 1941 Los Jesuitas del Peru (1568-1767}. A. Castefieda. Lima. 1959 Historia de la Iglesia en el Peru. Torno II (15701640). Burgos. Spain. Vasquez de Espinoza, Antonio 1969 Description of the Indies (c.1620}. Translated by C. U. Clark. Smithsonian Miscellaneous Collections Vol. 102. Smithsonian Institution Press. Washington, D.C. Vega Ocampo, Abel 1978 Complejo Arqueol6gico de Uchucmarca: Descripci6n del Elemento Ceramica de Pirca-Pirca. Investigaci6n Argueol6gica, No. 2, pp. 8-19. Universidad Nacional de Trujillo. Trujillo, Peru. 1979 Irnportancia Argueol6gica de Bolivar. Unpublished "Trabajo de Habilitaci6n." Departamento Academico de Ciencias Sociales. Universidad Nacional de Trujillo. Trujillo, Peru. Villabla 0., Marcelo 1988 Cotocollao: Una Aldea Formativa del Valle de Quito. Miscelanea Antropol6gica Ecuatoriana. Serie Monografica No. 2. Museos del Banco Central del Peru. Quito. Weberbauer, Augusto 1920 La Salida de Patas al Huallaga Estudiada en la Ruta de Pajaten. Boletin de la Sociedad Geografica de Lima Torno XXXVI, Trimestre prirnero, pp. 5-13. Lima. 1945 El Mundo Vegetal de los Andes Peruanos. New edition. Ministerio de Agricultura. Lima. 673 Wegner, Steven A. 1981 Identifying Recuaydom: An Expanded Ceramic Inventory for Identifying Recuay Settlement Sites. Revised version of a paper given at the 46th Annual Meeting of the Society for American Archaeology, San Diego. Weissner, Polly 1990 Is there a Unity to Style? In The Uses of Stvle in Archaeology. Edited by M. Conkey and C. Hastorf, pp. 105-112. Cambridge University Press. Cambridge. Wheeler, Jane 1985 De la Chasse a l'Elevage. In Telarmachay: Chasseurs et Pasteurs Prehistorigues des Andes. Torno 1. By D. Lavallee, M. Julien, J. Wheeler and C. Karlin. Institut Fran~ais d'Etudes Andines. Editions Recherche sur las Civilisations No. 20, pp. 61-80. Paris. Wiener, Charles 1884 Amazone et Cordilleres. Le Tour du Mende. 48 (2) :385-416. Paris Willey, Gordon R. 1945 Horizon Styles and Pottery Traditions in Peruvian Archaeology. American Antiquity 11:49-56. 1971 An Introduction to American Archaeology, Vol. 2 South America. Prentice Hall Inc. Englewood Cliffs, New Jersey. Willey, Gordon R. and Jeremy A. Sabloff 1974 A History of American Archaeology. W. H. Freeman and Co. San Francisco. Wilson, David J. 1988 Prehispanic Settlement Patterns in the Lower Santa Valley. Peru. Smithsonian Institution Press. Washington, D.C. Wilson and Reyes 1964 Geologia del Cuadrangulo de Pataz. Carta Geologica Nacional Serie A, Boletin No. 9. Comision de la Carta Geologica Nacional. Lima. Wing, Elizabeth 1972 Utilization of Animal Resources in the Peruvian Andes. Appendix IV in Andes 4: Excavations at Kotosh. Peru. 1963 and 1966. By S. Izumi and K. Terada. University of Tokyo Press. Tokyo. 674 1980 Faunal Remains. In Guitarrero Cave: Early Man in the Andes. Edited by T. F. Lynch, pp. 149-172. Academic Press. New York. Wise, Karen Ruth 1976 Apuntes sabre la Influencia Inca entre los Arnuesha. Factor que Oscurece la Clasificaci6n de su Idioma. Revista del Museo Naciona1 Torno 42, pp. 355-366. Lima. 1985 Indigenous Languages of Lowland Peru: History and Current Status. In South American Indian Languages: Retrospect and Prospect. Edited by H. E. Manelis Klein and L. R. Stark, pp. 194-223. University of Texas Press. Austin. Wobst, Martin 1977 Stylistic Behavior and Information Exchange. In For the Director: Research Essays in Honor of James B. Griffin. Edited by C. E. Cleland, pp. 317-342. Museum of Anthropology Anthropological Paper 61. Ann Arbor. Wright Jr., H. E., G. 0. Seltzer and B. C. S. Hansen 1989 Glacial and Climatic History of the Central Peruvian Andes. National Geographic Research 5 (4) :439-445. Young, Kenneth R. 1990 Biogeography and Ecology of a Timberline Forest in North-central Peru. Unpublished Doctoral thesis, University of Colorado, Boulder. 1991 Floristic Diversity on the Eastern Slopes of the Peruvian Andes. Candollea 46:125-143. 1992a Biogeografia y Conservaci6n de los Basques Montanos Tropicales. In Memoria del Decima Congreso Nacional de Biologia. Edited by E. Castillo de Maruenda, pp. 57-62. Consejo Nacional, Colegio de Bi6logos del Peru. 1992b Biogeography of the Montane Forest Zone of the Eastern Slopes of Peru. In Memorias del Museo de Historia Natural. Edited by K. R. Young and N. Valencia, No. 21, pp. 119-140. Universidad Nacional Mayor de San Marcos. Lima. 1993 National Park Protection in Relation to the Ecological Zonation of a Neighboring Human Community: An Example from Northern Peru. Mountain Research and Development 13 (3) :267-280. 675 n.d. Una Perspectiva Dinamica de la Vegetaci6n del Parque Nacional Rio Abiseo. In Memorias del Simposio Biodiversidad. Historia Cultural y Futuro del Pargue Nacional Rio Abiseo. In Press. Asociaci6n Peruana para la Conservaci6n de la Naturaleza. Lima. Young, Kenneth R. and Blanca Leon 1988 Vegetaci6n de la Zona Alta del Parque Nacional Rio Abiseo, San Martin, Peru. Revista Forestal del Perli, 16:1. Lima. 1990 Catalogo de las Plantas de la Zona Alta del Parque Nacional Rio Abiseo, Peril. Publicaciones del Museo de Historia Natural (B) 34:1-37. Universidad Nacional Mayor de San Marcos. Lima. 1993 Distribution and Conservation of Peru's Montane Forests: Interactions between the Biota and Human Society. In Tropical Montane Cloud Forests. Compiled and edited by L. Hamilton, J. Juvik and F. Scatena. East-West Center. Honolulu. Young, K. R., W. B. Church, M. Leo and P. F. Moore 1994 Threats to Rio Abiseo National Park, Northern Peru. Ambio 23 (4-5) :312-314. Stockholm. Young, Kenneth R. and Neils Valencia 1992 Introducci6n: Los Bosques Montanos del Peril. In Memorias del Museo de Historia Natural. Edited by K. R. Young and N. Valencia, No. 21, pp. 5-10. Universidad Nacional Mayor de San Marcos. Lima. Zaki, Andrzej 1983 Cultura Pelon. Una Desconocida Cultura en la Sierra Norte. Boletin de Lima 5 (29) :13-19. Lima. Zevallos Q., Jorge 1982 Omnastica Prehispanica de Chachapoyas. Investigaci6n Argueol6gica, No. 4, pp. 3-18. Universidad Nacional de Trujillo. Trujillo, Peru. 1987 Introducci6n al Estudio Etnohist6rico de Chachapoyas. Unpublished manuscript in possession of the author. Zeidler, James A. 1988 Feline Imagery, Stone Mortars, and Formative Period Interaction Spheres in the Northern Andean Area. Journal of Latin American Lore 14 (2) :243-283. APPENDIX A ILLUSTRATIONS AND ARTIFACT PROVENIENCE 676 677 SOUTH AMERICA PACIFIC OCEAN Scale 500km Fig. 1. Map of the Central Andes, western Amazonia and the Northern Andes (after Institute Geografico Nacional 1984). 678 ! ECUADOR COLOMBIA SOUTH r'/ AMERICA I 10 o PACIFIC OCEAN Legend (jjJ Western Slope • Northern 0 Eastern Slope Scale 500km Fig. 2. Map of the Central Andean Montane Forests in Peru (after Young· and Leon 1993: Fig. 1). The highland montane forest is not shown. 679 SOUTH AMERICA PACIFIC OCEAN Scale 500km Fig. 3. Map of some sites mentioned in the text. 680 Legend • 0 Towns Archaeological ~ Rivers Scale 50km f Fig. 4. Map of the northeastern Peruvian Andes. Shaded area is above 3,000 m {after Institute Geografico Nacional 1984 and NASA 1976). 00' Legend - ......... Study Area Boundaries ffiTI Dry Forest Zone 0 Moist Montane Zone D Tropical Alpine Zone ~ Montane Rain Forest Zone Ill Premontane Forest Zone Town o Arcpaeolog!cal !;lite Study Area Macro-ecological Scale 25km Fig. 5. Map of the Pataz-Abiseo study area and macroecological zones described in the text (after Young 1993: _Fig. 3 and Young et al. 1994: Fig. 1). 0'1 00 f-' Manachaqul Cave (3625ml 4500 m I Cran Pojalen 12850ml I Tropical Alpine Zone I 3500 m , Montane Rain Forest Upper Montane Forest 2500m Lower Montane Forest Maraflon-H uallaga Divide 1500m 4000m I 3000m Montane Rain Forest Zone 2000m Premontane Forest lOOOm 500m Okm 50km IOOkm Fig. 6. Profile of the Maranon-Huallaga divide showing distribution of macro-ecological zones. 0"1 00 1:\.) ~ N I - • .,.___..__ Approx. :J km 1\taral'lon·Huallag;t river dl\,dc 1\toncane Rain forest Zone lakes and rivers I ~ Fig. 7. Map of the Manachaqui and Montecristo River valleys and the pre-Hispanic roads. "'w 00 684 D \ , \ \ \ I I \ \ \. I \ SECTOR B 2m \ ......... - Shelter interior Fig. 8. Plan map of the Manachaqui Cave site complex (Site M-1). 685 .--.29 _-·-·-·30 26 ·-·-. 5 32 31 6 ........ .. -.... 27. 28 ........ ....... 18 20 21 4 17 3 22 ./ / \ 24 ,... 23 N I 1 m J \ \ \ \ \ \ \ '\ '' - · - · - Area "cl.eaned" Edge of roof -··· ··· .... Drip line Fig. 9. Plan map with detail of Manachaqui Cave (Site M-lA) interior. The principal interior space is shaded. 686 2m · ____, __ ',9..., ~--·-. .. ,...... ..' \ ' It '\,. Contour interval =0.50 rn \ .....' t, '·\ ~ ~... ~ l '\, . . _:.:= _....---.. , Shelter interior Cl Test pit No. 1 Fig. 10. Map of Manachaqui Cave excavation units 1988-1990. 687 Legend for Stratigraphy Illustrations Charcoal or carbon-rich lenses Lenses of yellow-brown fill Gravel ROCK ~- Ul U2 U3 UNEXCAVATED ~~- ~~!!'~'&'- ~~ Fig. 11. west profile of Units 1, 2 and 3. m (X) (X) us U4 ~a <:::::> <J to.~ ' - - - - - R-3 ~ (]:!---~----- U35 U6 ~ .... 0 ao o ~ ----~~~ ~ ---~~-c::> ""' 2A/B c. 0 (/ c::::?\:7 c:;J Doao <10 o b () Q cJ Oa::::?o"'c::::::J 6 ~ ° 1::::> Fig. 12. West profile of Units 4, 5, 6 and 35. Features R-2 and R-3 are shown. 0"1 (X) lO U38 U39 U7 us c. D () 0 () ROCK Fig. 13. West profile of Units 38, 39, 7 and 8. 0'\ ~ 0 ( ROCK U3 Ul U2 (] 0 2A 2A/B Fig. 14. East profile of Units 3, 2 and 1. ~ -0\ ~ I-' us U6 U35 oJ) U4 c:;. 0'\:> ~ Q c:J =o -c::::>o () c()<)o_g-_1,_ 0... Q.... "" [J ... -- 0 r:::::.o ~ ~~ C:P~ co 0[] c:J tV ~ 2A/B o <» 2A ~ODcv a 2s6 Q (l CJ Fig. 15. East profile of Units 35, 6, 5 and 4. Feature R-4 is shown. 3A/B 0"1 1.0 1'0 U8 U7 U39 PEDISTAL Fig. 16. East profile of Units 7, 8, 39 and 38. "'w \0 694 U37 U28 U32 2A/B F'lg. 17. East profile of Units 37, 32 and 28. Ull Ul Ul3 ROCK Fig. 18. South profile of Units 11, 1, 13 and 12. Feature R-5 is shown. 0"1 \.0 Ul 696 \ U12 \ U13 -;;;~~~~~c=~--~--~-­ - \ A ~~....--~15~;;;;;;;::;::::===~~zg--=2:A~~[: w X y z AA Fig. 19. North profile of units 12 and 13. Feature R-5 is shown and sector A floors are indicated in right margin. Ul9 U20 U21 ~ Ul8 ~:;!_ Qc::J CJ Q c:::J r::::,.(;> ~ ~ <::r eo cO c:::> 2A c:=o () 0 "V -·o lA/B ______ ~ -<::> ~ t;> ~ p 0 ()~ 0 2A/B 0 0 '\:::> -=- 0 Q 0 0 () 3A ~ ~~~- £2 Fig. 20. South profile of Units 21, 20, 19 and 18. Feature R-3 and Floor Z's Hearths 1 and 2 are shown. 0'\ lO ..,_] U29 ~l I ( c::> ~, ' ' I FALLEN ROCK ·, c:::::2 Pre-1986 Surfia , _____(ji' ______. :.,~~Cl CJ I ~ ---- 4 --~ --=-==-- <:::::.. 0 C7 ,;:> ~ 0 !R t;> 0 qC),() \)Dg'<l~ 0 d Cl (;P Qb' = U32 ce :::::::::::::g2;:=:::. • R 8:; U31 U6 U30 c 2A/B C7 Fig. 21. North profile of Units 29, 30, 35, 31 and 32. 0"1 \.0 00 699 U5 U6 LEVEL 16 18 20 22 c:::::::::::::. 24 26 0 0 0 0 2A/B 0 CJ Cl ~ D CJ 28 32 34 36 38 40 Fig. 22. Schematic diagram illustrating relationships between 1988 excavation levels and strata as seen in Unit 5 and Unit 6 east profiles. 700 U32 \ \ 2A/B 17 19 30 ~..---::...--~ ~-=~~~~~~~32 \ \ \ Fig. 23. schematic diagram illustrating relationships between 1990 excavation levels and strata as seen in Unit 31 and unit 32 north profiles. 701 od 3 I ·-·-·-·-·-·-,·-· i I i I ~~ i 14 2 --·-·-·~-·- --- ----1 12 I I 0.50m ------- WALLS OF DRAIN ~DIRECTION OF FLOW R-7 10 Fig. 24. Plan view of Feature R-5, Sector A. 702 Ul6 Ul7 ......._ ' ..; , ' fl.. / I , I '\ I ' ROCK '' ,,' , I ROCK Ul4 Ul5 Ul6 Ul7 A <JeD D(} C) f3:J ()0 ROCK ROCK Ul4 Ul5 B Fig. 25. Sector A floors in Units 14-17. a: Floor FF and associated rock-filled hearth; b: Floor EE and associated rock-filled hearth. 703 Ul6 Ul7 ROCK ROCK Ul4 Ul5 A Ul7 Ul6 ROCK ROCK Ul4 Ul5 B Fig. 26. Sector A floors in Units 14-17. a: Floor CC and associated rock-filled hearth; b: Floor BB and associated hearth. 704 A UlO U9 Fig. 27. Feature R-7 in rear of Sector A, Units 9 and 10. a: north face of stone wall in Unit 10. b: east profile of Feature R-7. Cave wall protrusion is shaded. 705 SECTORB ,,-..:;.-:.":.. .:::::-.::\ ;I' ~......• k t!'' ••• N SECTOR A ., • ...,_~ • • ~::-: + .. r:·· .......~ ...... •. "•• ~ ······ ) lrn Fig. 28. Plan view of Manachaqui Cave interior showing rock wall features R-6, R-7 and R-8. 706 U16 U17 f ' , ,, ,,- ...... '' \ ,I \.._ ,,-'HEARTH ROCK U14 t:J16 \) U15 A c=:>O U17 I I I t? I I I \ , HEARTH 1''----------' 0(} 0 0 / ,' t \ HEARTH2- tJ ROCK U14 U15 B Fig. 29. Sector A floors in Units 14-17. a: Floor AA and associated hearth; b: Floor Z and associated Hearths 1 and 2. 707 U16 Ul7 ,..,..,..,---- ....... ,, / 'q HEARTH\ { I I \ \ U14 \ QI'-.J I I I r I\., ' ..... _.,, 'I './ U15 Fig. 30. Sector A Floor P and associated hearth in Units 1417. 708 UNIT 17 A ------.. 0 ' ___ r -" I I I I .................. _,, ,. / I 1 I I I UNIT 16 8 Fig. 31. Sector A floors containing hearths with embedded stones. a: FloorS in Unit 17; b: Floor M? in Unit 16. 709 U28 U19 8 Fig. 32. Plan views of Sector Brock wall features. a: Feature R-6 in Units 23 and 28 (Rockshelter exterior is shaded). b: Feature R-8 in Units 18 and 19. 710 U31 A A---.J:)_....... - ...- .- .-. .- ;:r--F=>"<--)- A ' 0 8 Fig. 33. Sector B Feature R-4 in Unit 31, Level 16. a: Plan view of hearth. b: Profile of hearth. 711 Provenience of Illustrated Artifacts: Unit#-Leve1# (rim diameters) Fig.34a: b: c: d: e: f: g: h: i: j: k: 1: m: n: o: p: q: r: s: t: Fig.35a: b: c: d: e: f: g: h: i: j: k: 1: m: n: o: p: Fig.38a: b: c: d: e: 24-20 24-15 2-21 5-32 31-14 25-8 2-20 2-21 6-33 6-33 1-20 4-29 30-14 28-19 5-30 37-9 4-25 27-16 35-17 19-12 f: g: h: i: j: K: 1: m: n: o: p: 34-10 (10) 31-20 (12) 1-24 (11) 28-16 (12) 34-11 (11) 27-17 (11) 30-20 (10) 2-21 (9) 1-24 (11) 32-19 (18) 32-18 (18) 35-16 (17) 39-7 (14) 28-17 (14) 29-12 (13) 36-14 (12) 26-18 (14) 29-14 (19) 36-12 (10) 5-33 (26) 34-11 (26) f: 28-18 (24) g: ? (23) h: 39-7 (14) Fig.36a: b: c: d: e: Fig.37a: b: c: d: e: f: g: h: i: 30-16 37-8 36-12 24-19 26-19 21-6 28-13 28-16 28-14 (14) (12) (14) (11) (13) (14) (25) (24) (16) 4-29 (24) 6-29 (30) 6-34 (25) 5-33 (23) 6-30 (22) 22-18 (14) 38-10 (13) 22-17 (12) 34-11 (12) 25-8 (14) 30-16 (13) 20-10 (14) 22-19 (14) 22-17 (14) 34-12 (13) 32-17 (12) Fig.39a: b: c: d: e: 21-13 (14) 20-12 (13) 20-11 (13) 19-15 (12) 37-11 (11) f: 24-15 (11) g: 2-23 (13) h: 13-9 (16) i: 31-13 (16) j: 30-18 (16) k: 22-14 (14) 1: 31-16 (14) m: 28-15 (13) n: 31-14 (12) o: 31-14 (12) p: 22-16 (11) q: 31-13 (16) r: 22-16 (15) s: 5-24 (14) t: 3-21 (14) u: 31-14 (14) v: 30-15 (13) w: 31-15 (12) x: 16-04 (10) y: 6-30 (14) z: 1-22 (12) aa: 13-9 (?) Fig.40a: b: c: d: e: f: g: h: i: j: k: 1: 18-10 (12) 31-15 (12) 28-21 (19) 30-19 (16) 4-28 (14) 31-15 (14) 30-19 (13) 31-20 (12) 28-22 (12) 14-34 (17) 5-31 (14) 5-32 (13) m: n: o: p: 5-32 13-13 1-21 5-28 (12) (11) (11) (10) Fig.41a: 5-26 (12) b: same c: 5-32 (18) d: 4-30 (16) e: 30-17 (15) f: 25-6 (14) g: 5-32 (13) h: 5-30 (13) i: 5-33 (11) j: 30-15 (16) k: 30-16 (15) 1: 30-20 (14) m: 12-9 (13) n: 28-23 (11) o: 31-14 (11) p: 30-19 (11) q: 5-31 (18) r: 5-31 (18) s: 31-17 (15) t : 5-32 (14) u: 16-45 (11) v: 25-6 (ll) w: 30-19 (10) Fig.42a: 32-16 (19) b: 2-18 (15) c: 31-16 (15) d: 28-18 (14) e: 31-16 (14) f: 31-12 (12) g: 30-18 (10) h: 22-17 (12) i: 4-21 (20) j: 6-30 (16) k: 4-26 (13) 1: 4-27 (12) m: 5-28 (11) n: 31-16 (14) o: 22-17 (11) p: 30-16 (10) Fig.43a: 29-17 (15) b: 26-17 (16) c: 29-17 (15) Fig.44a: b: c: d: 29-11 (12) 28-19 (13) 32-17 (17) 36-14 (17) Fig.45a: 17-43 (11) b: 28-20 (15) 712 c: d: e: f: g: h: i: j: k: 1: m: n: o: 30-17 (14) 30-14 (12) 31-14 (11) 31-18 (11) 30-12 (?) 30-16 (13) 31-14 (13) 22-16 (13) 5-31 (13) 30-16 (12) 1-18 (14) 24-18 (12) 6-30 (12) Fig.46a: 28-15 (18) b: 31-17 (18) c: 22-19 (14) Fig.47a: b: c: d: e: f: 30-15 (13) 28-19 (13) 1-23 (14) 28-21 (15) 5-29 (19) 28-14 (24) Fig.48a: b: c: d: e: 31-14 30-17 37-12 28-17 30-15 k: 35-18 (13) 1: 30-20 (13) m: 6-35 (12) n: 20-13 (12) o: 18-18 (13) Fig.52a: b: c: d: e: f: g: 32-15 (11) 20-12 (11) 30-16 (12) 31-17 (12) 30-16 (14) 4-27 (28) 2-23 (12) Fig.53a: b: c: d: e. f: g: h: i: 2-23 (?) 6-27 (26) 13-12 (23) 32-14 (18) 6-27 (15) 1-21 (13) 1-20 (13) 11-7 (11) 24-21 (18) 37-9 (18) 28-16 (14) 20-10 (13) 1-20 (15) j: k: (16) (14) (14) (16) (15) Fig.49a: 17-44 (16) b: 5-31 (16) c: 28-17 (17) Fig.50a: b: c: d: e: f: g: h: i: j: k: 1: m: n: o: p: q: 21-16 (18) 11-6 (17) 31-17 (17) 13-11 (15) 30-18 (14) 26-19 (12) 30-19 (12) 32-20 (11) 13-E2 (18) 28-23 (13) 19-13 (13) 19-14 (?) 25-9 (18) 29-15 (17) 25-9 (16) 1-20 (13) 31-17 (12) 4-28 (18) 26-19 (16) 12-10 (16) 32-17 (14) 18-18 (13) 5-33 (12) 36-15 (12) 5-30 (11) 22-15 (12) j : 4-26 (18) Fig.51a: b: c: d: e: f: g: h: i: 1: m: 1-20 (15) 5-28 (14) 5-27 (13) 36-12 (12) 6-35 (19) 31-20 (18) 19-16 (14) 24-15 (18) 27-18 (17) j: 36-9 (17) k: 5-31 (10) 1: 34-12 (9) Fig.54a: b: c: d: e: f: g: h: i: Fig.55a: b: c: d: e: f: g: h: i: 26-21 (16) 16-45 (14) 3-26 (13) 28-14 (13) 31-17 (12) 28-22 (14) 22-14 (18) 32-14 (14) 14-42 (16) j: 31-17 (14) k: 38-10 (16) Fig.56a: 25-L (18) b: 16-44 (22) c: 30-17 (18) b: 26-15 C: 6-33 d: 6-34 e: 1-E f: 2-23 g: 22-17 h: 6-33 i: 30-14 j: 4-30 k: 20-10 1: 31-12 m: 36-14 Fig.59a: b: c: d: 29-18 19-12 15-26 5-30 and 27-20 e: 5-30 f: 22-15 g: 6-32 Fig.60a: b: c: d: e: f: g: h: i: 14-41 34-7 36-15 1-19 29-19 14-33 29-11 36-14 34-11. j: 15-30 k: 24-13 Fig.61a: b: c: d: e: f: g: 22-17 37-9 35-15 22-16 1-23 23-17 4-29 Fig.62a: 29-15 and 35-16 b: 22-11 c: 27-16 d: 26-21 e: 30-18 f: 5-28 Fig.63a: b: c: d: 22-16 24-19 37-11 30-16 Fig.57a: 27-16 and 32-17 b: 27-15 c: 24-14 Fig.64a: 21-12 b: 26-9 C: 6-28 d: 36-11 e: 31-17 f: 30-15 Fig.58a: 28-19 Fig.65a: 27-17 and 713 24-17 (13) b: 36-12 (14) c: 38-9, 36-13 and 6-30 (13) Fig.66a: 28-23 (?) b: 27-14, 22-16 and 22-14 c: 6-30 (9) Fig.67a: 6-32, 34-10 and 30-16 (13) b: 23-17 c: 6-24 f: g: h: i: j: k: 1: m: n: o: p: q: r: s: t: U: v: W: Fig.68a: 1-19 (11) b: 6-25 (10) C: 5-23 (10) d: 12-4 (13) e: 30-10 (12) f: 11-8 (10) g: 36-8 (12) h: 22-9 (12) i: 25-5 (17) j: 12-7 (15) k: 31-11 (13) 1: 30-11 (12) m: 32-8 (11) n: 31-13 (10) o: 16-2 (8) p: 12-10 (7) q: 28-14 (14) r: 31-13 (10) s: 4-22 (14) Fig.69a: 6-25 and 31-13 (12) b: 4-22 (13) c: 21-8 (10) d: 31-9 (14) e: 31-9 (12) f: 31-7 (12) g: 4-15 (11) h: 30-9 (10) i: 24-6 (?) j: 6-23 (8) k: 28-6 (14) 1: 4-20 (15) m: 4-21 (14) n: 25-5 (13) o: 4-23 (12) p: 28-5 (11) q: 28-6 (10) r: 16-10 (7) Fig.70a: b: c: d: e: 30-11 5-24 5-23 6-25 6-23 (14) (13) (12) (11) (10) 5-23 (10) 6-25 (10) 6-26 (8) 3-21 (11) 14-21 (12) 4-19 (11) 36-11 (12) 6-27 (16) 30-11 {14) 4-23 (13) 4-22 (13) 31-11 (12) 5-24 (10) 6-24 (10) 6-25 (14) 6-24 (12) 4-23 (11) 4-20 (10) c: d: e: f: g: h: i: j: k: 1: m: n: 0: p: q: r: s: t: u: Fig.71a: b: c: d: e: f: g: h: i: j: k: 1: m: n: 0: p: q: r: s: t: u: 31-11 (?) 30-9 (18) 30-10 (15) 6-26 (12) 30-10 (12) 22-8 (10) 6-26 (15) 31-12 (14) 31-13 (12) 31-12 (11) 31-12 (10) 13-9 (10) 24-11 (16) 28-7 (14) 4-17 (13) 4-17 (13) 31-11 (12) 28-11 (12) 4-23 (11) 22-13 (16) 4-22 (9) Fig.72a: 22-11 (14) b: 20-9 (14) C: 28-14 (20) d: 21-9 (15) Fig.73a: b: c: d: e: f: g: h: i: 17-11 (16) 28-10 (16) 27-10 (14) 34-7 (16) 28-10 (18) 28-9 (10) 3-20 (?) 14-33 (13) 30-11 (12) j: 22-11 (27) k: 20-4 (12) 1: 11-8 Fig.74: 21-8 Fig.75a: 30-12 (14) b: 4-20 (13) v: Fig.76a: b: c: d: e: f: g: h: i: j: k: 1: m: n: Fig.77a: b: c: d: e: f: 13-4 (12) 4-21 (12) 30-11 (12) 25-1 (11) 5-24 (7) 1-18 (11) 30-10 (16) 30-10 (?) 30-10 {12) 5-23 {10) 28-14 { 8) 26-8 (10) 6-26 (?) 30-10 (17) 24-12 (14) 6-24 (12) 15-32 (12) 1-16 (11) 31-12 (11) 26-12 (15) 13-8 (14) 1-18 (13) 36-7 (12) 31-11 (11) 36-8 (11) 24-8 (10) 36-7 ( 9) 6-25 (10) 22-13 (10) 28-12 (15) 6-24 24-11 (?) 22-11 (?) 4-20 (13} 24-9 22-15 24-12 36-12 28-12 11-10 (18) (10) (12) (18) (14) (14) 5-23 (14) 30-10 (14) 13-9 (10) 4-22 (9) 31-14 ( 10) 35-9 (12) 29-9 (15) 30-11 (10) 29-11 (15) j: 6-26 (13) k: 27-11 (21) Fig.78a: b: c: d: e: f: g: h: i: Fig.79a: b: c: d: e: f: g: h: 16-3 6-29 31-13 28-10 3-15 28-11 17-10 31-9 (15) (15) (12) (14) (13} (12) (12) (12) 714 i: 20-7 (11) j: 4-17 (9) k: 30-11 and 31-12 (18) 1: 5-24 (18) m: 30-10 (12) n: 35-8 (?) 0: 28-12 (15) p: 14-42 (14) q: 4-21 (12) r: 3-21 (11) Fig.80a: 6-26 ( 2 0) b: 31-9 (12) c: 13-1 (16) d: 5-26 e: 25-7 f: 32-10 g: 2-19 h: 22-11 and 28-12 Fig.81a: 37-7 b: 32-9 C: 39-6 d: 31-10 e: 29-9 f: 27-10 g: 6-23 h: 28-9 i: 27-12 Fig.82a: b: c: d: e: f: g: h: i: j: 38-6 32-9 5-24 1-19 4-22 and 6-24 30-10 24-12 22-10 22-14 21-7 Fig.83a: 4-23 b: 6-25 C: 30-11 d: 35-9 e: 4-19 f: 4-24 g: 31-7 h: 28-11 i: 22-11 j: 20-8 k: 35-10 1: 30-13 Fig.84a: b: c: d: e: f: 4-20 22-9 6-27 5-24 22-15 36-10 g: 22-12 h: 22-13 i: 28-12 j: 34-6 k: 4-20 1: 19-9 Fig.85a: b: c: d: e: f: g: 4-21 3-22 4-23 17-11 20-9 11-5 22-11 h: 24-12 30-12 11-6 5-24 28-12 6-27 27-11 6-17 6-25 4-22 j: 19-11 k: 27-12 1: 2-24 Fig.86a: b: c: d: e: f: g: h: i: Fig.87a: b: c: d: e: f. g: h: i: j: 30-11 26-9 24-13 26-10 6-25, 27-11, 35-10 and 3-17 4-22 28-13 22-11 4-22 21-8 Fig.88a: b: c: d: e: 22-12 25-8 17-11 23-13 21-7 and 6-25 f: 6-25 g: 25-5 h: 1-22 Fig.89a: 30-10 b: 19-13 C: 20-10 d: 36-9 e: 30-10 Fig.90a: 35-10 b: 30-10 C: 30-12 d: 30-10 e: f: g: h: 26-10 6-26 30-12 11-8 Fig.91a: b: c: d: e: f: g: h: i: 21-7 5-20 6-22 12-10 32-6 32-8 6-23 4-19 6-21 32-3 3-20 31-9 28-9 13-8 3-15 26-5 19-10 2-20 32-8 20-4 13-4 5-20 j: k: 1: m: n: o: p: q: r: s: t: u: V': Fig.92a: b: c: d: e: f: g: h: i: j: k: 1: m: n: o: p: 31-9 30-9 30-9 1-19 28-9 28-9 17-12 29-6 28-10 29-9 20-7 13-9 14-17 18-7 37-5 6-22 Fig.93a: 36-8 b: 36-8 c: 2-20 d: 28-7 e: 5-22 f: 2-21 g: 1-17, and 2-19 Fig.94a: b: c: d: e: f: g: 1-14 12-7 2-19 2-19 13-9 13-10 3-22 (?) (16) (12) (10) (10) (12) (10) (?) (16) (?) (19) (14) (13) (12) (?) (14) (16) (13) (12) (?) (18?) (12) (16) (14) (14) (17) (16) (15) (15) (14) (14) (16) (17) (17) (11) (?) (19) (15) (?) (20) 2-18 (20) (?) (18) (13) (13) (13) (16) (17) Fig.95a: 29-2 (16) 715 b: 15-32 (15) c: 13-5 (13) d: 24-3 (13) e: 3-19 (14) f: 3-16 (10) g: 30-7 (20) h: 27-4 (18) i: 6-22 (16) j: 28-9 (14) k: 4-20 (14) 1: 13-8 (16) m: 2-18 (13) n: 23-8 (14) o: 3-13 and 3-14 (14) p: 2-22 (12) q: 24-3 (10) r: 25-1 (9) s: 14-30 ( 6) t:2-16 (?) u: 1-14 (17) V: 3-15 (14) w: 5-20 (12) Fig.96a: b: c: d: e: f: g: h: i: j: k: 1: m: n: o: p: q: r: S: 14-21 (18) 32-6 (16) 5-20 (16) 16-45 (15) 13-12 (23) 1-19 (17) 31-8 (13) 25-6 (11) 23-3 (20) 17-10 (12) 4-18 ( 8) 28-8 (14) 28-7 (14) 32-6 (10) 17-5 (3 0) 6-22 (20) 12-5 (13) 6-21 (?) 19-4 (?) Fig.97a: 5-18 b: 30-3 c: 17-6 d: 35-6 e: 32-13 f: 35-5 g: 30-6 h: 30-4 i: 23-3 j: 31-3 k: 24-6 1: 18-5 m: 32-7 (22) (20) (18) (15) (14) (13) (10) (10) (?) (11) (?) (?) (10) Fig.98a: b: c: d: e: (19) (19) (17) (16) (15) 16-39 1-20 16-4 27-8 3-20 f: 37-5 and 24-6 (14) g: 15-3 (14) h: 31-6 (11) i: 19-4 and 23-7 (24) j: 30-7 (16) k: 12-4 (13) 1: 2-21 (14) m: 28-7 (13) n: 5-21 (11) o: 6-20 (15) p: 21-4 (20) q: 3-14 (14) r: 3-21 (11) s: 15-10 (19) t: 34-4 (17?) u: 11-5 (12) v: 13-4 (?) Fig.99a: 34-4 and 18-6 (16) b: 31-8 (15) c: 28-9 (14) d: 28-7 (13) e: 1-13 (25) f: 5-19 (11) g: 1-14 (10) Fig.100a: 27-6 b: 38-3 c: 5-22 d: 1-14 e: 14-18 f: 29-6 (30) g: 28-6 (15) Fig.101a: 2-16 (14) b: 21-6 c: 15-20 d: 12-7 (16) e: 1-18 (?) f: 13-5 Fig.102a: 11-8 (20) b: 29-7 (13) c: 2-18 (14) d: 4-17 e: 1-23 f: 12-13 (18) Fig.103a: 18-4 (?) b: 11-8 (18) c: 17-7 (14) d: 6-20 (?) e: 6-20 ( ?) f: 5-21 (14) g: 20-3 (14) h: 17-7 (14) i: 22-5 (20) j: 26-4 (?) k: 18-L (?) 1: 5-19 (14) m: n: o: p: q: r: S: t: u: v: w: X: 5-19 (14) 15-6 (10) 35-7 (8) 17-7 (14) 21-6 (20) 15-11 (18) 16-1 (16) n/p (14) 1-19 (14) 3-17 (14) 1-19 (12) 2-20 (15) Fig.104a: 2-21 (14) b: 2-18 (16) c: 16-10 (14) d: 3-22 (12) e: 3-19 (18) Fig.105a: 21-7 b: 2-18 C: 12-5 d: 5-20 e: 5-20 (10) (14) (22) (14) (14) Fig.106a: 16-M3 b: 2-20 C: 4-18 d: 21-5 e: 1-20 f: 2-16 (6) g: 20-3 h: 6-21 (24) i: 20-3 (14) j: 6-21 (14) k: 39-3 (10) 1: 18-6 ( 7) m: 17-6 (14) n: 11-7 (12) Fig.107a: 6-20 (18) b: 2-21 (10) C: 15-29 (16) d: 3-18 (18) e: 15-15 Fig.108a: 11-2 (14) b: 1-18 c: 20-3 (16) d: 17-7 (14) e: 2-16 (12) f: 20-3 (14) g: 15-12 (18) Fig.109a: 12-13 b: 15-8 c: 2-18 d: 4-17 and 4-18 (12) e: 31-9 (16) f: 21-6 (12) g: 26-7 (14) h: 30-7 (14) 716 Fig.l.l.Oa: l.l.-l. b: 12-7 c: 34-7 (14) d: 10-2 717 Legend for Ceramic Illustrations* Red paint or slip D Orange paint or slip Brown paint or slip * except for Figs. 103x-106e and 109d, e; see captions. 718 e a c b i f \\g h d ~) 1 j ) q m f T r s Fig. 34. Manachaqui Phase Paste A body profiles. a-h: globular, semi-carinated and carinated profiles. i-p: same, with applique medial ribs. q-t: rare "stepped" shoulder profiles with shoulder ribs. 719 , ," '' /, / ' / '' / '\ / / '\ I \ \ I \ I \ I I I \ \ I I I I I I I I I I I \ \ \ I I I \ ' \ I I \ / \ \ I ' ' ' ' ...... / / ' ' .... , ....... / / / ;' - .... ;' ... ___ ..... __ ----- --- -- ..... ...-"' / i ,, 0 m p n Fig. 35. Manachaqui Phase Paste A, Shape A rim profiles. ah: Rim 1. i: hypothetical Rim 1 jar. j-p: Rim 2. 720 /~I ···· . f'·: ·. . . ~·. . .. .. .. . . ' a ) b -, -, d g e c f , , 'II I I I I I I I I I I I I I I I I I I \ \ I \ I \ I \ \ '' / '' I I / ' ' ''", '~ ... '-, ............ ____ .... --- _.. . ,,/ ,, .... "" / -==--==5 em h Fig. 36. Manachaqui Phase Paste A, Shape A rim profiles. a, b, g: Rim 2 decorated rims. c-f: Rim 2 vessels with large orifices. h: hypothetical Rim 2 jar. 721 \ ) ' ) a b ' c ~--, , ._J • I ~ f I e f 5 em g i Fig. 37. Manachaqui Phase Paste A, Shape A rim profiles. ac: Rim 3. d-f: Rim 4. g-i: Rim 5. 722 e b f g h i . • \ I I I \ I \ • / J '\ \ / '' / / '' / / ' ' ..... / / ' ..... .... .... .... ........ .... _ -- --- ------- ------- --- ---- ..,"" ~~,... / -I k 5 em Fig. 38. Manachaqui Phase Paste A, Shape A rim profiles. ae: Rim 6. f-i: Rim 7. j: hypothetical Rim 7 jar. k-m: Rim 8. n-p: Rim 9. 723 a g h p X u t v q 5 em y z w aa Fig. 39. Manachaqui Phase Paste A and Brim profiles. a-e: Shape A Rim 10. f: Shape A Rim 11. g: Shape A Rim 12. h: Shape A Rim 13. i-p: Shape BRim la. q-r: Shape BRim lb. yaa: Shape B Rim lc. 724 a I \ b I I \ ( \ ( \ I ' I I I ( I \ I I \ \ I I \ '' I I I '' ' / / '' / ' ..... ..... ................................. -- --- -- --- _..... ..... / / / // /' i 1 5 em Fig. 40. Manachaqui Phase Paste A, Shape B rim profiles. a: Rim la with handle. b: hypothetical Rim lb jar. c-p: Rim 2a. 725 b (top view) 5 em Fig. 41. Manachaqui Phase Paste A, Shape B rim profiles. a, b: Rim 2a profile and top view. c-i: Rim 2b. j-p: Rim 3a. qw: Rim 3b. 726 '~''''r a c d e f g h r i m j ~ p 0 n 5 em Fig. 42. Manachaqui Phase Paste A, Shape Brim profiles. ah: Rim 4. i-p: Rim 5. 727 a b / , / / I I ~ 5 em ---- --- -- Fig. 43. Manachaqui Phase Paste A, Shape Brim profiles. a, b: Rim 5 with applique decoration. c: hypothetical Rim 5 jar. 728 ' a c d 5 em Fig. 44. Manachaqui Phase Paste A, Shape B rim profiles. a: Rim 5 with rare notched decoration. b-d: Rim 5 with incision and applique decoration. 729 l (. ~ a b ~~ lf ' , , h i g r j 1 k m n o Fig. 45. Manachaqui Phase Paste A, Shape B rim profiles. ag: Rim 6. h-1: Rim 7. m-o: Rim 8. 730 ) ' \. \. ' / \. / a / '' / '' / / ./ ' .... ... ... .... ......... .... __ .... .... _ ---- --- ------ -- _- .,.......--"' ,./ / ./ / ,./ ...... 5 em Fig. 46. Manachaqui Phase Paste A, Shape Brim profiles. a: Rim 8 partially reconstructed jar. b, c: Rim 9. 731 1" a I b c { d Ej I .,;:.,-.·'/;· .. ···. e ( VI .., ...:-. · . f 5 em Fig. 47. Manachaqui Phase Paste A, Shape BRim 9 profiles. 732 , \0 f a b I I I c / \ ' / '' ' / '' ' ' ,, / / / / / ,,-....... ............. __ _ / .,/ / __, ~~ ~---- 5 em Fig. 48. Manachaqui Phase Paste A, Shape Brim profiles. a, b: Rim 9. c: Rim 9 partially reconstructed vessel. d, e: Rim 10. 733 ""', " ' ... ' ... ./a ,., "'"' ' ...... ...... /' ,. ,.-"" .......... --- --- --- ------ --------------., .... ,'"' b 5 em Fig. 49. Manachaqui Phase Paste A, Shape Brim profiles. a: Rim 10 partially reconstructed jar. b, c: Rim 10 profiles with top views. -,- -,c-, -,~\X\h a d 734 g b i ! ([I I ILJI ..: .. ·.; ~ ' . ' ' l .. j k n q I I I I I I r I (i I \ I \ f I / \ '' I ' '- ' .... _ ' •-==--==-· . . ----5 em / ------ / / .... ·-""' Fig. 50. Manachaqui Phase Paste A, Shape C rim profiles. ak: Rim la. 1-p: Rim lb. q: hypothetical Rim la bowl. 735 h c i -, -\k_\_, . J IQ 1 I m \ n \ 0 / / \ I '\ I '\ \ I \ I \ I \ I I I I q p \ I \ I \ / '' I / ' ' ', / / ' .......... ............ ---- ----- ----~--r----- ---- --- ...-...,..,..,, ...... // 5 em Fig. 51. Manachaqui Phase Paste A, Shape C rim profiles. ai: Rim 2a. j-n: Rim 2b. o: hypothetical Rim 2a bowl. 736 1. ( ( c / 5 em ' d -, f Fig. 52. Manachaqui Phase Paste A, Shape C rim profiles. a, b: Rim 3. c-e: Rim 4. f: Rim 5. g: Rim 6. 737 c k 1 i j I \ m f I I I I I I ( \ I '' I ) / '' / __ __ _ '' '-.. ..... _ ....... ----- ------- --- -- / / / .// 5 em Fig. 53. Manachaqui Phase Paste A, ShapeD rim profiles. ah: Rim 1. i-1: Rim 2. m: hypothetical Rim 3 bowl. 738 I c b d e f g a h i k 1 5 em Fig. 54. Manachaqui Phase Paste A, Shape D rim profiles. ad: Rim 3. e-g: Rim 4. h-j: Rim 5. k, 1: Rim 6. 739 c b a -, e ~I 1 h g j 5 em Fig. 55. Manachaqui Phase Paste A, Shape E rim profiles. a, b: Rim 1. c, d: Rim 2. e-h: Rim 3. i, j: Rim 4. k: Rim 5. 740 a 5 em Fig. 56. Manachaqui Phase Paste A, Shape E rim profiles. a: Rim 5. b: Rim 6. c: Rim 7. 741 I I • I I C. c 5 em Fig. 57. Manachaqui Phase Paste A miscellaneous shapes. a: spoon fragment. b, c: fragments of unidentified artifacts. 742 a c b '\?~ e f g /\-\ ~ ). ~ h ~ . . \ I> k j -) <:;;;-\ . ! m 1 5 em Fig. 58. Manachaqui Phase Paste A applique rib decorations. a-d: unembellished medial ribs. e-g: unembellished shoulder ribs. h-m: Notched A medial ribs. 743 ~ d ·, f g 5 em Fig. 59. Manachaqui Phase Paste A applique rib decorations. a, b: Notched A medial ribs. c: Notched B shoulder rib. d: Notched B medial rib. e, g: Incised A medial ribs. f: Incised A shoulder rib. 744 c ~~/ ~ Ye d h \ i j 5 em Fig. 60. Manachaqui Phase Paste A applique rib and band decorations. a: Incised A medial rib. b-e: Incised B medial and shoulder flanges. f-k: notched bands. 745 ~ c e 5 em Fig. 61. Manachaqui Phase Paste A applique band decorations. a, b: notched bands with Incised A ribs. c, d: bands with circular punctations. e, f: bands with ovoid incisions. g: applique serpent head. 746 a c I ~=~ ~}) d e J 5 em r f Fig. 62. Manachaqui Phase Paste A assorted applique decorations. a: Incised A rib/band. b: lug with notched medial rib and notched band. c: incised button on unembellished rib. d: incised button on Incised A medial rib. e, f: snake adornos on unembellished ribs. 747 a .· c -13 d 5 em Fig. 63. Manachaqui Phase Paste A adornos. a: head of amphibean or fish. b: bird wing or fish fin. c: zoomorphic heads. 748 a d)-' b . . -\ c d ~ f 5 em e Fig. 64. Manachaqui Phase Paste A incised decoration. a-c: Incised Vessel 1 with zoned punctation. d, e: Incised Vessel 2 with zoned punctation. f: incised sherd. 749 b c 5 em Fig. 65. Manachaqui Phase Paste B1 vessels. a: Vessel 1, b: Vessel 2, c: Vessel 3. 750 -,. ,.....----" \ I I II I 'I I I I b \. \. 5 em Fig. 66. Manachaqui Phase Paste Group B. a: Paste B1 , Vessel 4. b: Paste B1 Vessel 5. c: Paste B2 vessel. 751 I c 5 em Fig. 67. Manachaqui and Suitacocha Phase artifacts. a: Manachaqui Phase Paste B3 vessel. b: Manachaqui Phase incised slate disk. c: Suitacocha Phase sherd with net impressions on interior surface. 752 -,a-,~,c -, -, h i g f 1 j k f -, -( -( m 1 n r p 0 q r ~:c -· . . :.--::·--:·.> -~- ~ ~·.: ~~·~-, r._.'.::.._. .:· ·..;. .:./·:_:;·.-..-.. ~~~ . :_~·;~.~ \ '' ,' s 5 em Fig. 68. Suitacocha Phase Paste A, Shape A and B rim profiles. a-c: Shape A Rim 14. d-f: Shape A rim 15. g, h: Shape A Rim 16. i: Shape A Rim 17. j-s: Shape BRim 11a. 753 Jr:::JI ~b ) tid t -l -, -, -, -, 1. ··.· ...... . . ;·. ~- ... :· .·. I c l d f e h g i J ?~ ~1 -\ -t p n r 0 q m 5 em Fig. 69. Suitacocha Phase Paste A, Shape B rim profiles. ac: Rim lla. d-k: Rim llb. 1-r: Rim llc. 754 -, -, -t -~ a b c -, d ~--::-~~. g :~ ---: .,_ '] . :' ·-·- .-_ ·._:! -, -, h j k i 1 0 -, -, q v s w 5 em Fig. 70. Suitacocha Phase Paste A, Shape B rim profiles. ah: Rim lld. i-1: Rim lle. rn-s: Rim llf. t-w: Rim llg. 755 -( l r -~ a b d e c -, f k -,m -t -, 1 -~ 0 -,q -r -~ n t u 5 em Fig. 71. Suitacocha Phase Paste A, Shapes B and C rim profiles. a-f: Shape BRim 12. g-1: Shape BRim 13. m-u: Shape C Rim 7. 756 a b c d 5 em Fig. 72. Suitacocha Phase Paste A, Shape C rim profiles. a: Rim 7. b-d: Rim 8. 757 -, a b c d -, -, -, -, -, i f g j h e ( 81 1 5 em Fig. 73. Suitacocha Phase Paste A, Shape C and E rim profiles. a-c: Shape C Rim 9. d: Shape C Rim 10. e-h: Shape E Rim 8. i: Shape E Rim 9. j: Shape E Rim 10. k, 1: Shape E Rim 11. 758 Fig. 74. Suitacocha Phase Paste A partially reconstructed Shape F Rim 1 vessel. 759 -, -, a -, -, -, -, b d f e g c h -, 7 j m i -, -, -, 1 n r s t u q v 5 em Fig. 75. Suitacocha Phase Paste A, Shape F rim profiles. an: Rim la. o-v: Rim lb. 760 -1 d c f e g h J ltiT] \ tj;j;j I l. ~ r I j J e~?J ;;:;.~----~- ;.·~~~·.::·.:-. lk .· r : .. --- -:~.;;._~ 1 ---------- 5 em ••• • •• . 1 m r .n Fig. 76. Suitacocha Phase Paste A, Shape F rim profiles. ak: Rim 2. 1-n: Rim 3. 761 a r b r ! WI c r \~]1 ) .....·. .. I d r f 5 em Fig. 77. Suitacocha Phase Paste A, Shape F Rim 3 profiles. 762 a b e f r r 1 i h J 5 em Fig. 78. Suitacocha Phase Paste A, Shape F rim profiles. a: Rim 3. b-d: Rim 4a. e, f: Rim 4b. g, h: Rim 5. i, j: Rim 6. k: Rim 7. 763 -, c b -, -, a e d j k -, m n 1 5 em -, -, 11 r 0 p q Fig. 79. Suitacocha Phase Paste A, Shape F and X rim profiles. a-c: Shape F Rim 8. d-j: Shape X Rim 1. k-n: Shape X Rim 2. o-r: Shape X Rim 3. 764 a b c 5 em Fig. 80. Suitacocha Phase Paste A rim profiles and basal sherds. a: Shape X Rim 4. b: Shape X Rim 5. c: Shape X Rim 6. d: jar or bowl base. e-h: basal angles from carinated bowls. 765 - L).. . ~ c b a ~. I f e h g 5 em Fig. 81. Suitacocha Phase Paste A miscellaneous shapes. a-c: fragments of mammiform vessel legs. d-h: handle fragments? i: unidentified fragment. 766 b a c d I Q~ I f g e I i h 5 em Fig. 82. punctate notched. j: round Suitacocha Phase Paste A sherds with notched and rib or band decoration. a-d: notched. e: parallel f, g: high relief notched. h, i: ovoid punctation. punctation. j 767 I I @L 1 ~ d I I . . . __ :. I c ~ t!l~ I fifijW ~e f I I @I g\ . . . . ~--.~-:- b a e( ' ·. g I h i ' 1 5 em Fig. 83. Suitacocha Phase Paste A sherds with applique and incised decoration. a: round punctation. b, c: unembellished applique. d-f: flanges. g-i: incised lines. j: divergent arrays of parallel lines. k, 1: cross-hatching. 768 I - . <fJ···~. J l I o ... I . r::;;;:::;;;;; c b ~~.) ~1··1 I I I a ~ . . . ·.·: .. •.· ''.0:::<~:.-•• d I 'd!. e · . . . ·<:. <) .... I f I ~, ~ I g i 5 em Fig. 84. Suitacocha Phase Paste A decorated sherds with stamped circles. a: in row. b, h: with paint. c-f, i-1: with incision. j, k: with punctation. 769 a, I ,..,. Q' . - ~ I . I a c I ~l ~ I f e d h 5 em Fig. 85. Suitacocha Phase Paste A decorated sherds. a-f: stamped circles and applique with round punctations. g, h: punctation and incision. 770 a b ~ ~e c d I I Af-2l €91 ~I I At ~ . I g i h f I Q\ j I I k 5 em Fig. 86. Suitacocha Phase Paste A decorated sherds with punctation and incision. 1 771 ~1. c \ . , . ' . a\, I I . ·. ··. .. ' I f I e . . g . ' h I ~~~ ~ I 5 em i j Fig. 87. Suitacocha Phase Paste A sherds with punctation. ah: with punctation and incision. i: punctation and applique with round punctation. j: punctation, incision, notched applique and red paint. 772 a b I 1)DJ c d I g •I I h 5 em Fig. 88. Suitacocha Phase Paste A assorted decorated sherds. a, b: punctation, incision and notched applique. c, d: combing. e: brushing. f: rouletting. g: incised boss. h: fine-line scratched. 773 a c e 5 em Fig. 89. Suitacocha Phase Paste A assorted sherds. a, b: mat and/or fabric impressed. c: zoomorphic adorno depicting parrot head. d: zoomorphic adorno (opposurn head?). e: zoomorphic adorno (probably a bat's head and face). 774 a b c e \~ ·I . .. . . .. ~ I h 5 em Fig. 90. Suitacocha Phase Paste A adorno and Group B sherds. a: anthropomorphic head adorno. b: Paste B4 • c, d: Paste B4 jar rims. e-g: Paste B4 sherds with punctation and incision. h: Paste B5 sherd from bowl. 776 7 7 7 a 1 \ c b ~ 1~ ~- !19.·.·.···.. ·::i:·. . . ....·:.·. 7~ h I j 7 1 1 m P· n 5 em Fig. 92. Colpar Phase Paste A, Shape E rim profiles and decorated sherds. a: Rim 3, b, c: Rim 9. d-j: Rim 12. k, 1: Rim 13. m: Rim 14. n-p: notched applique. 777 e 5 em Fig. 93. Colpar Phase Paste Group B. a, b: Paste B6 rim and basal angle profiles. c-e: Paste B, rim profiles. f: Paste B, rim with iridescent red paint. g: Paste B8 rim with whiteon-red decoration. 778 UTUI d e f ( g 5 em Fig. 94. Colpar Phase Paste Groups B and C. a, b: Paste B8 rim profiles. c-e: Paste B9 rims with red paint. f: Paste B10 rim with orange-red paint. g: Paste C1 rim with red paint. 779 -, 7 a -J b 7. f d c j 1 n s q 0 5 em J p -I 1 t v w u Fig. 95. Empedrada Phase Paste A Shape E rim profiles. a-c: Rim 3. d-k: Rim 9. 1-s: Rim 15. t-w: Rim 16. 780 7 e 1/ f g ]) m }1 p r q n 1 s 0 5 em Fig. 96: Empedrada Phase Paste A Shape E rim profiles. a-d: Rim 17. e-h: Rim 18. i-k: Rim 19. 1-n: Rim 20. o: Rim 21. p, q: Rim 22. r: Rim 23. s: Rim 24. 781 ,,,, e f g h j • •• • • • • 7 k r 1 5 em Fig. 97. Empedrada Phase Paste A Shape GRim profiles. a-m: Rim 1. 782 ~, l/1 k 1 m n i I!' 17 s t q v Fig. 98. Empedrada Phase Paste A Shape G rim profiles. a-h: Rim 2. i-k: Rim 3. 1-n: Rim 4. o: Rim 5. p, q: Rim 6. r: Rim 7. s, t: Rim 8. u: Rim 9. v: Rim 11. 783 a f I g 5 em Fig. 99. Empedrada Phase Paste A Shape H rim profiles. a-d: Rim 1. e: Rim 2. f: Rim 3. g: Rim 4. 784 e~ • a I ' c b d l g e 5 em Fig. 100. Empedrada Phase artifacts. a, b: Paste A spoon fragments. c: unidentified artifact. d-e: notched applique. f: Paste B11 jar rim profile. g: Paste B12 rim profile. 785 c b 5 em Fig. 101: Empedrada Phase Paste Group B. a-c: Paste B13 bowl with red paint. d-f: Paste Bu sherds from bowls with red and brown paint. 786 c d e 5 em Fig. 102. Empedrada Phase Paste Group B. a, b: Paste B15 bowl with dark red paint, pale orange slip on exterior. c: Paste B15 : bowl with iridescent red paint. d, e: Paste B17 sherds with orange and white paint on brown slip. f: Paste B18 bowl with orange slip and negative-resist decoration. 787 -, ~I f e i g n 7 w s X 5 em Fig. 103. Empedrada Phase Paste Group C2 • a-p: unpainted. qw: traces of red paint. x: Paste C2 a with red paint. 788 a J b I c } \lJ d \~ J ----- 1 e 5 em Fig. 104. Empedrada Phase Paste Group C2 • a-e: Paste C2 a with red paint. 789 I \ i a I [)) b d 5 em Fig. 105. Empedrada Phase Paste Group C2 • a-e: Paste C2 a with red paint. 790 Lli[J_ b c 0 ~ - -, j ~-< 2 ) 1 -, g f e d ~ k 1 m n 5 em Fig. 106. Empedrada Phase Paste Group Cl. a-e: Paste Cla with red paint. f-n: Paste Cla with orange slip. 791 J b ( ~I ) c \~I J d 5 em e Fig. 107. Empedrada Phase Paste C2 • a-e: Paste C2b with black paint. 792 1. --b l < --- ( \Ji I ( Fti7 I d J. J f g 5 em Fig. 108. Empedrada Phase Paste C2 • a-d: Paste C2b with black paint. e-g: Paste C2 c with red and black paint. 793 I~~ a I b r~1 I VI I c )f I \ ( g IS j h 5 em Fig. 109. Ernpedrada Phase Paste C2 • a-c: Paste C2 C with red and black paint. d-f: Paste C2 d with tan paint. g, h: Paste C2 e with black paint on orange slip. 794 a c 5 em d Fig. 110. Empedrada Phase Pastes C2 • a: Paste C2 f with brown slip (bottom), negative-resist smudging (top} and white painted circles on natural tan paste (middle) . b: Paste C2g with red paint on light brown paste. c: Paste C2h unpainted jar rim. d: notched ground slate disk from Unit 10 Level 2. 795 PRE-LAVASEN PHASE (N=102) Q. (/) z 50 45 40 35 30 25 20 15 10 5 0 SMALL SMALL· MED MEDIUM MEDLARGE LARGE Animal Size Fig. 111. Pre-Lavasen Phase animal size distribution (NISP=Number of Individual Specimens). LAVAS EN PHASE (N=21} 6 5 4 Q. ~ z 3 2 SMALL SMALL· MED MEDIUM MEDLARGE LARGE Animal Size Fig. 112. Lavasen Phase animal size distribution (NISP=Number of Individual Specimens). 796 MANACHAQUIPHASE(N=~) 16 c. en z 10 8 SMALL SMALLMED MEDIUM MED- LARGE LARGE Animal Size Fig. 113. Manachaqui Phase animal size distribution (NISP=Number of Individual Specimens). SUITACOCHA PHASE (N=29) c. en z 18 16 14 12 10 8 6 4 2 0 SMALL SMALLMED MEDIUM MED- LARGE LARGE Animal Size Fig. 114. Suitacocha Phase animal size distribution (NISP=Number of Individual Specimens). 797 COLPAR PHASE (N::36) 25 20 a. en z 15 10 5 0 SMALL SMALLMED MEDIUM MEDLARGE LARGE Animal Size Fig. 115. Colpar Phase animal size distribution (NISP=Number of Individual Specimens). EMPEDRADA PHASE (N=211) 100 90 eo a. en z 70 60 50 40 30 20 10 0 SMALL SMALLMED MEDIUM MEDLARGE LARGE Animal Size Fig. 116. Empedrada Phase animal size distribution (NISP=Number of Individual Specimens) . 798 • • MOS CJPfTVLODgLQSMAJORDD MAlO 2)0MOSMI[I1 E ' 'e#.S 'YI'raj""'-f~tq,k&.&.i)14.fdoStu~;JoUJ1 .. - - - ,~~syJcy·ytfpMc.lJfL!y•tuf~or.-o-3 ...::=--- Fig. 117. Guaman Poma's rendering of a camelid with a cloth or net alforja containing jars for transport (from Guaman Poma 1936:524). 799 SOUTH AMERICA Scale 500km Fig. 118. Map showing relative degrees of stylistic similarity to the Manachaqui Phase Paste A style, and the probable sources of Paste Group B sherds. Darkest shading indicates closest similarity to Manachaqui Paste A. 800 SOUTH AMERICA PACIFIC OCEAN Scale 500 krn 7 Fig. 119. Map showing relative degrees of stylistic similarity to the Suitacocha Phase Paste A style, and the probable sources of Paste Group B sherds. Darkest shading indicates closest similarity to Suitacocha Paste A. 801 SOUTH AMERICA PACIFIC OCEAN Scale 500km Fig. 120. Map showing relative degrees of stylistic similarity to the Colpar Phase Paste A style, and the probable sources of Paste Group B and C sherds. Darkest shading indicates closest similarity to Colpar Paste A. 802 SOUTH AMERICA PACIFIC OCEAN Scale 500 km Fig. 121. Map showing relative degrees of stylistic similarity to the Empedrada Phase Paste A style, and the probable sources of Paste Group Band C sherds. Darkest shading indicates closest similarity to Ernpedrada Paste A. APPENDIX B PLATES 803 804 Plate I. View of the Manachaqui Cave site complex from the north side of Manachaqui Valley facing south. 805 Plate II. Manachaqui Cave and 1988 trench looking east. Plate III. Manachaqui Cave 1990 work in progress. 806 Plate IV. Manachaqui Cave Unit 6 north profile 1988. 807 Plate V. Manachaqui Cave Unit 13 north profile 1990. Plate VI. Manachaqui Cave Units 14-17 Floor CC, 1990. 808 7 Plate 9 10 II 12 13 Manachaqui Phase Plate VIII. Manachaqui Phase Paste B2 (see Fig. 66c). 809 Plate IX. Ground slate points. m: Manachaqui Phase. K, 1: Suitacocha/Colpar Phase. f, g: Colpar Phase. a-e, h-j: Empedrada Phase. 810 Plate X. Suitacocha Phase $hape F jar (see Fig. 74). 811 Plate XII. Ernpedrada Phase Paste C2 a kaolin sherds . ... ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,1,,,, 111f111111111f11 13 14 15 16 17 18 Plate XIII. Empedrada Phase projectile points. APPENDIX C TABLES 812 813 UNIT 1.5 Stratum Color pH Organic % Gravel Silt Sand % % % Clay % 2A 10YR2/1 6.60 5.80 18.5 68.25 0.95 12.30 2B 7.5YR3/2 6.95 3.72 23.3 62.42 0.91 9.65 3A 10YR3/2 6.30 4.41 22.7 61.66 0.89 10.34 3B 10YR3/3 6.80 3.60 28.5 57.34 1.03 9.53 Fill Brown 8.20 0.0 29.6 62.70 1.50 7.20 Table 1. UNIT Soil sediments from Unit 15, Sector A. 36 Stratum Color pH Organic Gravel % % Silt % Sand % Clay % 2A 10YR2/1 5.6 1.77 18.3 69.47 1.40 9.06 2A 10YR2/1 5.6 1. 79 18.9 67.69 1.42 10.2 2B 10YR2/1 6.4 1. 78 19.2 68.22 1.45 9.35 2B 10YR2/1 6.4 1. 80 23.1 65.55 1.56 7.99 3A 10YR3/2 6.8 1.28 39.0 50.17 1. 75 7.80 3B 10YR3/3 6.85 1.30 39.4 50.24 1. 76 7.30 3B 10YR3/3 6.85 1.30 41.5 48.96 1. 74 6.50 3B 7.5YR3/2 5.80 1.32 42.7 47.60 1. 78 6.60 Bedrock Brown 7.80 0.0 84.2 9.20 1.10 5.50 Table 2. Soil sediments from Unit 36, Sector B. 814 Floor/Unit Class Diam. Depth Phase B I 14 s 38 em 5 em Empedrada J I 15 B 12 em 1.5 em Empedrada L I 14 B 23 em 6 em Empedrada M? I 16 s 28 em 4.5 em Empedrada I 17 s 19 em 5 em Empedrada 15 B 35 em 5 em Empedrada p I 14 s 45 em 13 em Col par z z I 16 R? 48 em 11 em Manaehaqui I 17 .S? 30 em 5 em Manaehaqui AA I 17 B 26 em 2 em Manaehaqui BB I 15 B 38 em 2 em Lavas en cc I 17 R 50 em 5 em Lavas en DO I 15 R 26 em 3 em Lavas en EE I 16 R 50 em 13 em Lavas en FF I 15 R so em 7 em Lavas en s T I Table 3. Hearth classification, measurements and chronology. Hearth classes are: R = rock-filled; s = single embedded stone; and B = Simple basin (no stones). 815 Lab I I Fie1d I (Floor) I Context I Sector A: C-14 h.p. I calih.* I Phase Shelter Interior I-16,980 I-16,975 I-16,973 M1A-14-ll (H) M1A-14-15 (L) M1A-15-25 (T) Hearth Hearth Hearth 1500 + 80 1380 + 80 1460 !: 80 AD 619 AD 671 AD 646 Empedrada Empedrada Empedrada I-16,978 I-16,979 M1A-14-21 (P) M1A-15-30 (W) Hearth Floor 1840 + 80 2110 + 80 AD 239 50 BC Col par Col par Beta-75,230 I-17,756 M1A-15-35 (X) M1A-15-38 (Y) Floor Floor 2560 :!:. 100 2450 :!:. 90 765 BC 409 BC Suitacocha Suitacocha I-17,428 I-16,976 M1A-16-41 (Z) M1A-17-43 (AA) Hearth Hearth 2800 + 90 2850 + 90 902 BC 927 BC Manachaqui Manachaqui I-17,318 I-17,487 I-16,974 M1A-15-54 (BB) M1A-16-63 (££) M1A-15-66 (FF) Hearth Hearth Hearth 3670 + 100 3830 + 100 3520 + 100 1973 BC 2197 BC 1758 BC Lavas en Lavas en Lavas en Sector B: - Shelter Exterior I-17,321 I-17,319 M1A-22-ll M1A-6-25 Lens Lens 2630 :!:. 100 2740 :!:. 90 J95 BC 827 BC suitacocha Suitacocha I-17,320 M1A-31-16 Hearth 2810 :!:. 100 906 BC Manachaqui I-17,322 I-17 ,429 M1A-31-25/26 M1A-31-30 Strat 3A Strat 3B 4120 + 130 4280 :!:. 110 2586 BC 2881 BC Unnamed Unnamed CAMS-13,151 M1A-31-35 strat 3C 10270 :!:. 60 10152 BC Unnamed • Pearson and Stuiver 1993; Stuiver and Pearson 1993; Bard et al. 1993. The laboratory measurements were adjusted for the southern hemisphere prior to calibration. Table 4. Radiocarbon dates from Manachaqui Cave. 816 Rim Form I Sherds I 1 2 3 4 5 6 7 8 9 10 11 12 13 48 38 14 12 12 10 8 8 7 6 5 5 2 Table 5. Rim Form la lb lc 2a 2b 3a 3b 4 5 6 Range 6-16 10-26 9-15 9-14 13-25 22-29 12-14 13-14 12-14 11-14 11 13 12,16 I em em em em em em em em em em em em em Mode I R 12 em 25 Bi-modal 29 13 11 N/A 9 Bi-modal? 5 N/A 8 N/A 8 N/A 3 N/A 3 N/A 6 N/A 1 N/A 2 N/A 2 I Sherds I Range l Mode I l H 8 8-16 9-18 10-14 9-19 10-18 9-17 10-18 10-16 10-20 8-17 11-13 12-18 9 10 189 109 Decorated Groups 54 10-24 em 15 em 16 em 43 10-22 em Table 6. I Average 11.64 em 15.97 em 12.55 em N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A S.D. 2.45 em 4.20 em 1.44 em N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A Manachaqui Phase Shape A Rims 250 245 8 454 41 224 67 51 37 25 19 9 7 I em em em em em em em em em em em em 12 em 88 12 em 65 N/A 3 12 em 144 20 13 em 12 em 57 12,13 em 25 13,15 em 18 15,16 em 22 11 em 15 N/A 9 N/A 8 Average 12.98 12.98 N/A 13.08 13.70 13.25 13.72 13.89 14.45 11.20 N/A N/A J em em em em em em em em em 14.80 em 15.42 em S.D. 1.65 em 1.91 em N/A 1.87 em 1. 79 em 1.91 em 2.74 em 1.45 em 2.33 em 2.45 em N/A N/A 2.70 em 2.18 em Manachaqui Phase Shape B Rims. * Calculations were not made for sample sizes of less than 10 sherds. 817 Rim Form I Shards I 1a 1b 2a 2b 3 4 5 6 57 11 18 9 8 7 2 1 Table 7. Rim Form 10-18 7-18 11-19 12-18 11 12-14 12 28 JN Mode 12 em N/A 12 em N/A N/A N/A N/A N/A em em em em em em em em I Sherds I 32 14 13 13 13 2 Table 8. Range 11-26 12-18 12-17 14-20 15-18 9.5 I em em em em em em Mode l 40 8 13 6 4 5 1 1 Manachaqui Phase Shape 1 2 3 4 5 6 Rim Form I Range c Average 13.95 em N/A 14.46 em N/A N/A N/A N/A N/A S.D. 2.18 em N/A 2.73 em N/A N/A N/A N/A N/A Rims. I I R Average 18 6 7 9 5 2 17.5 em N/A N/A N/A N/A N/A 15 em N/A N/A N/A N/A N/A l I S.D. 3.99 em N/A N/A N/A N/A N/A Manachaqui Phase Shape D Rims. I Shards _I 1 2 3 4 5 6 7 6 6 5 3 3 2 1 Range 10-18 11-14 13-18 14-16 16-18 22 18 em em em em em em em Table 9. Manachaqui Phase Shape E Rims. Technique Total Interior Surface Treatment Finished Unfinished Position Medial Rib ends Shoulder Unembellished 122 20 (16\) 102 (84\) 115 (94\) 7 (6\) 10 (8\) Notched A 292 231 (79\) 61 (21\) 292 (100\) 0 (0\) 14 (5\) Notched B 38 36 (95\) 2 (5\) 14 (37\) 24 (63\) Incised A 502 Incised B 75 24 (5\) 17 (23\) 6 ( 16\) 478 (95\) 485 (97\) 17 (3\) 18 (4\) 58 (77\) 55 (73\) 20 (27\) N/A 961 (93\) 68 (7\) 48 ( Table 10. Totals pertaining to rib and flange embellishment techniques. Severely eroded sherds were not included for tabulations (N=l,029). 818 Rim Form j 14 15 16 17 I 14 8 7 1 Table 11. Rim Form Sherds 10-14 10-14 11-12 17 252 110 86 52 35 28 22 16 11 llf llg 12 13 Table 12. B 6 5 4 1 Average I N/A N/A N/A N/A S.D. N/A N/A N/A N/A I Mode I 12 em 12 em 10 em 12 em 12 em 13 em N/A N/A N/A em em em em em em em em em B I 64 44 41 26 11 15 7 9 7 Average 11.58 11.57 11.44 11.54 11.18 12.0 N/A N/A N/A em em em em em em I S.D. 1. 73 em 1. 68 em 2.02 em 1. 95 em 1. 80 em 1. 86 em N/A N/A N/A Suitacocha Phase Shape B rims. I Sherds I 7 8 9 10 37 9 7 1 Range 9-16 14-20 14-17 16 em em em em I Mode I I Average 13/14 em N/A N/A N/A 14 6 5 1 13.07 em N/A N/A N/A H suitacocha Phase Shape Rim Form I Sherds I 8 9 10 11 7 6 1 1 Table 14. J l N/A N/A N/A N/A em em em em Range 7-15 8-15 7-15 8-16 8-15 9-16 9-12 10-18 10-15 Rim Form Table 13. Mode Suitacocha Phase Shape A rims. I Sherds I lla llb lle lld lle I Range Range 11-18 em 8-12 em 27 em 12 em I Mode N/A N/A N/A N/A I S.D. 1. 91 em N/A N/A N/A c rims. I I H Average 4 2 1 1 N/A N/A N/A N/A suitacocha Phase Shape E rims. I S.D. N/A N/A N/A N/A 819 Rim Form I Sherds I 109 29 79 26 7 3 6 6 4 3 1a 1b 2 3 4a 4b 5 6 7 8 Table 15. Rim Form 7-16 10-17 5-15 10-18 7-14 10-12 10-15 13-15 21 12-15 11 71 27 14 10 9 • 6 4 4 3 2 2 Table 16. Mode 11 em 11 11 14 N/A N/A N/A N/A N/A N/A em em em em em em em em em em I I Average 31 15 19 17 5 2 2 4 1 3 11.52 em 12.33 em 10.84 em 13.94 N/A N/A N/A N/A N/A N/A R I S.D. 2.09 1. 92 2.54 2.55 N/A N/A N/A N/A N/A N/A em em em em Range I Mode 12-17 em N/A 9-20 em 13/14 em 6-20 em 14 em N/A 12-20 em N/A 15-18 em N/A 11-23 em N/A 8-13 em N/A 10-14 em 30 em N/A 13-20 em N/A N/A N/A N/A N/A I I Average 5 29 19 7 4 5 4 4 1 2 0 0 N/A 14.06 em 13.11 em 't:T/A N/A N/A N/A N/A N/A N/A N/A N/A R I S.D. N/A 3.02 em 3.58 em N/A N/A N/A N/A N/A N/A N/A N/A N/A Empedrada Phase Shape E rims. Rim Form 1Sherds I 1 2 3 4 5 6 7 8 9 10 56 43 21 12 8 4 3 2 1 1 1 11 I Suitacocha Phase Shape F rims. I Sherds I 3 9 15 16 17 18 19 20 21 22 23 24 Range Range 9-24 em 11-23 em 12-21 em 11-20 em 15-20 em 12 em 11 em 19 em 7 em 12 em N/A I Mode 10 em 17 em N/A N/A N/A N/A N/A N/A N/A N/A N/A I I Average 21 25 7 8 3 1 1 1 1 1 0 15.48 em 15.96 em N/A N/A N/A N/A N/A N/A N/A N/A N/A R Table 17. Empedrada Phase Shape G rims. I S.D. 4.79 em 2.88 em N/A N/A N/A N/A N/A N/A N/A N/A N/A 820 Phase Lavas ell Mallachaqui suitacocha C:olpar C:ol/E=p Floors PF - BB AA- z Y, Z: 'tf v, u l4 4 4 2 2 14 Sample Weight 565.01 139.91 98.98 50.76 13.25 527.2 Wooc! Weight 430.32 54.58 20.985 19.745 3.14 291.84 \ of Sample Weight 85.2 39 21.2 38.9 23.7 55.4 Maize Kel:'llals 37 46 1 236 Maize Cob Frags 42 15 4 555 Phaseolus 2 No. of Samples E=pec!rac!a or- 5 cf. Sapot. Fruit 76 8 6 Fruit Rille!& 57 20 5 1 'l!uber/Root Frags 40 5 1 2 17 1 1 LupillUS 1 6 cf. Ribes Rubus Festuca 9 1 1 1 3555 C:hellopoc!/amerallt 3 Polygollum/Rumex 5 2 12 32 3 4 11 6 123 1 1 2 2 Table 18. Botanical remains by phase (see Appendix F). 3 G 821 Class Order Family Genus Species Mammalia •••••••• • ••••.••.•. . 19 Rodentia . ...................... . Caviidae ••••••••••• Cavia •..•...•..•. Artiodactyla •••••••••••• cf. Artiodactyla •••••••• Aves. RISP • ........•.•. 2 . ............. 1 . ............. l . ........•••. 2 • ••..•...•.•. 1 .•••.......... 1 . •••.......... 1 Strigiformes ••••• Table 19. Fauna of the Lavasen Phase. Class Order Family Genus Species RISP 15 Marmnalia • .................•. Rodentia •••••••••••••• S igmodon tinae •• . .....•. 1 . •........•... 1 Table 20. Fauna in LavasenjManachaqui Strata 822 Class Order Family Genus Species NISP ............................ 51 . .............•............... 1 Mammalia •••••• cf. Mammalia •• Rodentia ••••••••••••• Caviidae ..•••.• Cavia .• Agoutidae ...•.•.......... Agouti ... ....... . cf. taczano. ..5 • •2 • •2 . ........... 1 . ....... 1 . ...... . 1 Artiodactyla •..••••••••••••.•••• Cervidae . ............... . • .•...•• 3 • •••••••• 2 Odocoileus ••••••• . ......... 1 virginianus ••••••••• . ...•.. . 1 cf. Odocoileus ••••••••••••••• . ....... 1 cf. virginianus ••••••• . . . . . . . . . . 1 Xenarthra •••••.•••••••.• Dasypodidae •••••• Dasypus .•••••.••. •••••••••• 5 • •••.••••• 5 .......... 1 Table 21. Fauna of the Manachaqui Phase Class Order Family Genus Species Mammalia •••• NISP . ••.•••. 6 Rodentia. • •2 Artiodactyla . ................ . Cerv-idae . ............. . . . . . Odocoileus • •..• virginian us •• .1 .1 .1 .1 . .1 Osteichthyes .•••. Table 22. Fauna in ManachaquijSuitacocha Strata. Class Order Family Genus Species Mammalia •.•• NISP • •••••. 28 . •••.•...••• 8 . .1 Rodentia ••••..••• cf. Rodentia ...•. Caviidae .• Cavia. cf. porcellus • •• Sigmodontinae. . ......... 3 • •3 .• 1 . .1 Osteichthyes. .2 Aves ...••• .1 Reptilia . ................................................... 1 Table 23. Fauna of the Suitacocha Phase. 823 Class Order Family Genus Species HISP ....................... Mammalia •••• 6 Artiodactyla . ......................................... 2 Cervidae . ..................................... . 1 cf. Odocoileus ••••.•• .1 c f . Camelidae . ....................•............ 1 Table 24. Fauna in SuitacochajColpar Strata. Cl.ass Order Family Genus Species Mammal.ia •••• HISP •••.•....... 33 Rodentia •• cf. Rodentia ••••••••• Caviidae ••••• • .11 ... 2 . •• 6 Cavia •••••••• cf. Cavia. 3 . ..•.......... 2 Artiodactyla ••.•.•••••••• . .....•. 3 Cervidae . . . . . . . . . . cf. Mazama •••• cf. Camelidae ••••••••• . ........ 2 ...... ~ . 2 . ....... 1 Xenarthra. Dasypodidae ••••••••••• cf. Dasypus •• cf. Dasypodidae •••••• . .......... 1 . ........ 1 . ..... 1 Osteichthyes •••• ........ 2 . ..... 1 Table 25. Fauna of the Colpar Phase. Class Order Famil.y Genus Species Mammalia •••• HISP • • 28 Rodentia ••••••••.•••• Caviidae ••••••• . .11 Cavia. porcellus ••• cf. porcellus. Aves •• ••••••• Osteichthyes ••••••••••••••• cf. Eleotridae •• Table 26. Fauna in ColparjEmpedrada Strata. . •. 6 . •. 6 .•• 1 .1 .1 .1 . .. 1 824 Class Order Family Species Genus Mammalia ••••••• cf. Mammalia ••• RISP . •••••.•.•..•••.••• • 19 9 . •••.••....••••.•••• 3 Rodentia. cf. Rodentia ••••• Caviidae •• cf. Caviidae •••• Cavia ••. cf. Cavia. porcellus. cf. porcellus •• Sciuridae •• Sigmodontinae ••• • .•••••••••••• 4 3 • •••.•••••••••• 2 • •• 16 • .••.•.•.•••••. 6 • •• 12 . ••. 2 .3 .6 .1 .1 Ar-tiodactyla. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 cf. Ar-tiodactyla. . . ......................... 2 Cervidae. . . ................. . ... 7 cf. Cervidae. . . . . ........................... 2 Odocoileus. . . . . . . . . . .............. 4 virginianus • • . •• 4 cf. Pudu •. ..•• 1 Camelidae ••••••••• • ••• 7 cf. Camelidae ••••• • ••• 2 cf. Lama. .1 Xenarthra ••••••••••••• Dasypodidae •••• •••• 6 . .•• 6 Carnivora •••••••• cf. Carnivora •••• Canidae.. . ...... . Dusicyon ••• culpaeus andinus. .3 .3 .1 .1 .1 . .• 1 . ..... 1 Primates . .......... . Hominidae ••.•• Homo •• sapiens •••• Aves. Strigiforrnes .•••• Strigidae. cf. Reptilia •••••••• cf. Amphibia .• cf. Osteichthyes ••• Table 27. Fauna of the Empedrada Phase. ... 1 . .1 • •• 11 . .. 1 . ... 1 •••• 2 . ... 1 .1 825 I i!HA.SE~.SIZE I .SHALL I .SHALL-MED I MEDIJ.lM I HED-LABGE I LAB!:ZE I :t~AL I EMPEDRADA 98 43 25 12 33 211 COLPAR 21 5 4 2 4 36 SUITACOCHA 17 3 2 3 4 29 MANACHAQUI 11 16 10 5 6 48 LAVASEH 4 2 4 6 5 21 PRE-LAVASEN 7 10 17 48 20 102 Table 28. Animal size by phase. Sums are NISP. APPENDIX D MACROCHRONOLOGY 826 827 Unit 1 Unit 2 Level Shards Carin Incised Kaolin Lithics Level Shards Carin Incised Kaolin Lithics 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 10 12 30 84 172 136 141 128 183 201 180 150 83 193 80 14 6 Total 1803 2 2 9 1 6 1 3 2 3 1 5 2 6 9 20 13 1 26 1 2 10 28 31 37 36 20 28 22 4 8 3 9 10 11 2 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 8 54 66 73 46 43 63 85 142 117 154 58 170 54 38 11 3 1 1 1 2 7 1 4 2 2 2 1 3 1 2 3 4 4 1 2 1 10 41 41 45 21 18 7 8 10 14 4 7 5 7 8 286 Total 1185 20 3 22 246 Unit 3 Unit 4 Level Shards Carin Incised Kaolin Lithics Level Shards Carin Incised Kaolin Lithics 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 3 25 140 111 30 35 16 75 3 99 98 76 83 82 47 12 4 Total 1004 3 3 5 2 3 1 1 1 27 45 12 2 4 8 8 5 7 4 5 6 19 141 1 1 1 2 2 2 6 4 1 1 1 1 2 2 1 4 18 9 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 so 161 79 112 88 155 96 123 4 10 2 1 1 1 3 4 5 1 lOS 100 80 72 36 80 65 152 205 74 18 8 4 3 2 2 4 4 2 Total 1860 23 1 53 62 17 18 11 17 4 2 5 2 1 1 3 5 6 13 24 12 23 13 18 279 Note: "Carin" refers to sherds from medial and shoulder angles, and sherds bearing applique ribs. It excludes the 16 Suitacocha Phase sherds from basal angles. "Incised" refers only to line-incised, but excludes the 31 Manachaqui Phase sherds with line-incision. 828 Unit 5 Unit 6 Level Shards Carin Incised Kaolin Lithics Level Shards Carin Incised Kaolin Lithics 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 67 140 105 118 160 137 118 255 172 138 85 eo 120 95 149 172 120 135 110 15 1 Total 2492 1 1 3 1 5 11 13 11 6 4 1 55 1 1 1 5 4 4 2 2 2 17 7 6 1 2 51 69 41 25 23 16 14 32 4 2 1 3 6 2 19 37 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 54 87 101 11 75 66 1 lOB 95 147 190 157 134 150 290 104 98 87 130 150 142 162 88 58 7 4 1 1 2 2 3 8 12 3 4 10 23 14 5 10 5 3 2 2 l l ll l l l 5 4 2 2 289 Total 2543 37 58 68 38 23 8 4 48 26 56 439 Unit 7 Unit 8 Level Shards Carin Incised Kaolin Lithics Level Shards Carin Incised Kaolin Lithics 1 2 3 4 5 6 7 8 9 10 12 13 22 7 20 90 142 97 126 85 130 60 36 15 5 Total 842 ll 5 l 2 10 3 8 32 26 2 3 l l 5 4 3 1 2 12 14 5 2 3 4 5 6 7 8 9 10 11 12 5 6 19 28 31 82 8 105 45 34 11 30 Total 404 l 3 2 4 5 5 l l 2 1 1 l 1 l 4 5 27 102 Unit 11 Unit 12 Level Shards Carin Incised Kaolin Lithics Level Shards Carin Incised Kaolin Lithics l 2 3 4 5 6 7 8 9 10 11 12 13 70 65 93 120 102 118 115 188 104 130 18 1 3 Total 1148 l l 1 3 2 2 3 l l 2 2 3 1 2 4 1 l l l 2 13 9 14 3 2 12 8 7 13 17 18 12 15 4 1 6 120 12 13 15 21 27 62 38 25 95 23 118 56 14 18 3 Total 515 l 2 3 4 5 6 7 8 9 10 ll l l l 1 l 6 2 2 1 3 1 1 3 3 8 12 17 6 l 2 1 l 12 4 9 so 829 Unit 13 Unit 15 Level Sherds Carin Incised Kaolin Lithics Level Shards Carin Incised Kaolin Lithics 14 10 30 17 77 27 39 12 43 105 53 43 48 5 14 Total 523 1 2 3 4 5 6 7 8 9 10 11 12 13 5 1 2/A 3/B 4/C 5 6/0 7 8/E 9/F 10/G 11 10 26 2 10 2 1 1 1 4 1 1 6 4 2 1 1 11/Ei 15 78 2 1 Unit 14 Level Sherds Carin Incised Kaolin Lithics 1 2/A 3/B 4/C 5 6/0 7 8/E 9/F 10/G 11/Ei 12/I 13/J 14/K 15/L 16/LL 17/M 18/N 19/N 20/0 21/P 22/Q 23/R 24/S 25/T 26/U 27 28/V 29 30/W 31 32 33 34 35/X 36 37 38/Y 39 40 41/Z 42 43/AA 44 45 46 47 48 49 50 Total 2 10 20 18 20 23 3 5 4 7 2 1 5 3 5 4 2 3 2 7 5 7 2 2 4 3 4 31 4 52 5 27 32 53 24 1 10 11 4 8 16 3 15 I 1 1 1 6 3 3 1 1 1 Total 1 1 1 2 11 2 8 16 4 2 1 1 1 486 12/I 13/J 14/K 15/L 16/LL 17/M 18/N 19/N 20/0 21/P 22/Q 23/R 24/S 25/T 26/U 27 28/V 29 30/W 31 32 33 34 35/X 36 37 38/Y 39 40 41/Z 42 43/AA 44 45 46 47 48 49 50 51 52 53 54/BB 55 14 5 1 32 4 12 2 11 27 10 3 14 2 8 5 4 1 1 4 3 1 4 4 2 2 4 5 3 2 4 3 4 5 3 12 4 29 1 12 20 23 39 7 6 15 2 18 8 1 1 1 1 1 1 1 1 2 1 2 1 1 1 1 1 1 10 11 2 7 1 1 2 2 1 2 2 1 2 370 6 1 11 30 830 Unit 16 Unit 17 Level Shards Carin Incised Kaolin Lithics Level Shards Carin Incised Kaolin Lithics 1 2/A 3/B 4/C 5 6/D 7 6/E 9/F 10/G 4* 5* 6* 7• B* 9* 10* 11* 36/Y 39 40 41/Z 42 43/AA 44 45 46 47 46 49 50 53 46 93 25 64 6 12 5 6 11 262 63 67 133 lOB 96 46 46 7 4 73 39 33 1 35 33 45 11 9 7 3 Total 1444 1 1 19 14 25 6 1 2 3 1 9 3 17 6 6 5 7 1 2 1 2 1 7 1 3 1 1 2/A 3* 4* 5• 6• 7* B* 9* 10* 11* 12* 36/Y 39 40 41/Z 42 43/AA 44 45 46 47 46 49 50 51 52 53 4 66 233 142 40 53 63 104 49 34 94 31 3 26 25 11 1 19 25 37 10 6 2 2 1 1 1 1 3 5 1 1 1 2 4 1 1 4 3 4 1 1 2 1 1 1 1 1 1 1 2 5 2 Total 1104 4 18 41 15 3 10 10 14 101 2 16 3 13 119 Note: Totals do not include artifacts contained in soil samples. * Denotes arbitrary levels Unit 16 Unit 19 Level Shards Carin Incised Kaolin Lithics Level Shards Carin Incised Kaolin Lithics 4 110 96 325 136 276 256 279 139 132 102 47 161 71 93 96 26 33 16 4 3 1 1 1 3 1 3 2 2 7 1 5 2 Total 2410 29 1 2 3 4 5 6 7 6 9 10 11 12 13 14 15 16 17 16 19 20 21 22 2 1 1 2 4 1 1 3 6 6 6 4 2 23 31 45 17 46 26 19 6 4 6 5 3 3 11 2 5 1 10 29 1 2 3 4 5 6 7 6 9 10 11 12 13 14 15 16 17 16 19 20 21 22 23 5 59 190 506 121 165 247 326 164 150 236 166 295 205 119 194 53 12 2 3 1 3 1 5 25 13 5 6 2 1 1 2 3 6 1 2 1 1 19 5 6 1 62 119 16 22 25 32 4 6 17 1 10 13 12 11 7 1 2 4 4 3 8 1 256 Total 3245 62 16 35 401 831 Unit 20 Unit 21 Level Shards Carin Incised Kaolin Lithics Level Shards Carin Incised Kaolin Lithics 1 2 3 4 5 6 7 8 9 10 11 12 13 14 78 211 188 151 239 268 320 255 170 235 283 185 88 19 Total 2690 1 1 2 2 3 2 5 6 4 1 4 6 8 1 4 11 28 6 3 2 1 1 2 1 27 31 28 33 19 21 21 20 9 7 1 6 1 20 56 199 1 2 3 4 5 6 7 8 9 10 ll 12 13 14 15 16 17 18 50 194 132 175 134 170 248 314 167 82 93 77 87 51 51 25 1 3 1 2 5 6 1 1 2 5 15 35 7 1 10 2 3 3 13 3 1 3 2 5 1 2 4 3 Total 2053 20 13 56 9 25 19 10 25 3 4 4 67 35 183 Unit 22 Unit 23 Level Shards Carin Incised Kaolin Lithics Level Shards Carin Incised Kaolin Lithics 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 51 85 84 78 70 101 105 166 86 182 496 224 248 166 236 257 280 161 119 71 Total 3266 1 1 2 7 9 16 12 13 3 6 5 10 5 6 2 1 1 13 6 2 1 37 51 26 19 24 24 16 14 12 8 1 2 3 4 5 6 7 8 9 10 11 ll 1 2 12 13 14 15 16 17 18 19 20 1 3 5 1 1 6 38 77 40 ll 24 292 6 3 42 92 56 86 107 78 62 82 43 70 145 125 43 40 65 70 59 72 1 2 4 4 2 Total 1346 16 1 1 2 l 3 2 1 4 6 22 66 71 40 34 32 27 22 13 17 34 9 1 1 7 1 399 832 Unit 24 Unit 25 Level shards Carin Incised Kaolin Lithics Level Shards Carin Incised Kaolin Lithics l 2 3 4 5 6 7 8 9 10 ll 12 13 14 15 16 17 18 19 20 21 22 23 24 8 39 128 94 152 200 360 321 200 171 160 265 100 124 172 166 193 123 252 135 47 36 17 1 l 2 3 4 5 7 11 7 20 7 2 3 2 4 7 5 3 18 5 13 8 4 ll 6 12 l 2 3 4 5 6 7 8 9 10 Total l 1 82 59 11 28 125 159 136 100 46 7 2 2 l l l l 3 l l 3 3 3 2 l 4 2 5 12 3 8 12 17 3 ll 9 9 7l l l l 754 4 3 1 1 3 4 6 12 ll Total 3474 56 71 154 87 85 84 125 2 5 5 5 73 46 . 25 710 Unit 26 Unit 27 Level Shards Carin Incised Kaolin Lithics Level Shards Carin Incised Kaolin Lithics 1 2 3 4 5 6 7 8 9 10 ll 12 13 14 15 16 17 18 19 20 21 100 104 124 140 182 176 146 98 172 135 25 43 94 133 128 151 220 145 73 24 9 Total 2422 3 3 4 2 2 4 3 4 2 4 2 3 2 7 3 3 44 46 35 17 19 17 9 4 10 6 7 8 9 10 ll 1 ll 10 1 ll l 3 4 l l 63 l 2 3 4 5 l 23 16 12 13 14 15 16 17 18 19 20 21 22 55 138 181 93 109 145 162 164 95 192 218 179 153 103 130 156 121 200 172 l l l 2 11 8 15 10 6 3 2 5 44 53 69 31 29 33 14 15 14 4 6 3 l 4 2 4 10 18 12 9 66 6 57 10 2 Total 2899 67 21 193 24 51 327 833 Unit 2B Unit 29 Level shards Carin Incised Kaolin Lithics Level Shards Carin Incised Kaolin Lithics l 2 3 4 5 6 7 B 9 lO ll 12 13 14 15 16 17 lB 19 20 21 22 23 24 25 eo ll4 150 104 109 173 155 lOB 66 66 10 21l 172 91 103 67 llB 51 162 1l 61 29 36 13 l Total 2503 l 7 19 13 6 l l l l 6 6 3 lO 5 3 l 2 36 4 14 2 2 43 40 76 47 35 37 54 40 21 lO 2 l 6 2 l l ll 33 23 46 l 2 3 4 5 6 7 B 9 lO ll 12 13 14 15 16 17 lB 19 20 17 66 240 69 127 159 62 169 122 270 301 192 324 242 225 166 lOB 21 19 4 Total 2949 l 5 l 5 6 15 B ll 6 B l l 6 2 3 1 2 2 3 3 l 3 l l 2 1 l 66 11 75 76 24 65 62 42 93 44 27 lB l 2 9 21 576 460 Unit 30 Unit 31 Level Shards Carin Incised Kaolin Lithics Level Shards Carin Incised Kaolin Lithics l 2 3 4 5 6 1 B 9 lO ll 12 13 14 15 16 17 lB 19 20 21 22 40 74 209 166 69 ll3 137 206 273 497 492 276 206 322 342 322 252 297 236 40 lO 6 Total 46ll l l 2 16 9 2 l l 5 16 lO ll 6 12 B 2 72 6 19 ll 4 40 31 lB 46 100 49 22 34 50 44 64 65 20 17 15 l 2 ll 6 B 15 5 14 19 1 2 3 4 5 6 7 B 9 lO ll 12 13 14 15 16 17 lB 19 20 21 22 23 65 66 42 69 l2B 90 120 162 161 104 276 206 227 212 64 383 221 l2l 106 54 39 12 B l l 3 l 7 B 15 6 7 l l 12 4 5 5 2 6 7 2 32 16 24 34 ll2 27 46. 47 41 17 6 3 4 2 2 l l 4 2 4 16 ll5 2 l 667 Total 3022 50 19 29 557 834 Unit 32 Unit 34 Level Sherds Carin Incised Kaolin Lithics Level Sherds Carin Incised Kaolin Lithics l 2 3 4 5 6 7 8 9 10 ll 12 13 14 15 16 17 18 19 20 21 22 23 46 61 26 58 79 164 l 3 9 2 2 us 129 267 365 264 234 226 237 218 201 254 136 96 64 l l l l 5 16 14 3 12 4 3 2 6 3 2 4 2 l l 2 6 7 2 10 5 68 13 ll 4 4 Total 3259 63 15 27 19 42 49 66 48 47 34 21 4 4 5 6 17 19 l 2 3 4 5 6 7 8 9 10 ll 12 13 14 15 193 7 97 247 96 98 520 182 135 176 145 51 18 8 5 Total 1977 1 7 3 4 5 2 14 5 1 7 3 l 3 6 6 15 l 34 86 14 130 79 25 20 54 8 11 439 500 Unit 35 Unit 36 Level Sherds Carin Incised Kaolin Lithics Level Sherds Carin Incised Kaolin Lithics l 2 3 4 5 6 7 8 9 10 ll 12 13 14 15 16 17 18 19 20 52 89 91 6 24 138 161 165 206 298 250 178 34 91 107 136 92 27 20 10 12 l l 5 4 10 6 6 4 15 12 34 4 10 36 55 38 13 l 2 2 l l 4 5 7 8 l 6 10 2 2 17 l 2 3 4 5 6 7 8 9 10 ll 12 13 14 15 16 17 18 19 8 28 115 35 78 240 195 205 447 218 236 303 308 366 210 75 ll 46 17 12 260 4 8 8 l 3 4 5 12 12 8 10 34 19 73 7l 44 29 4 6 2 l l 4 l Total 3082 Total 2175 2 l 49 13 22 281 835 Unit 37 Unit 38 Level Shards Carin Incised Kaolin Lithics Level Shards Carin Incised Kaolin Lithics l 2 3 4 5 6 7 8 9 10 11 12 13 14 15 133 26 120 237 164 220 201 123 182 157 374 188 110 53 6 Total 2295 2 57 34 67 lOS l? l 35 29 2 l 4 7 13 21 2 6 l 1 l 5 56 3 3 333 l 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 40 33 3 11 57 191 135 328 386 144 60 Laval Shards Carin Incised Kaolin Lithics 1 2 3 4 5 6 7 8 9 10 81 16 201 75 270 180 216 128 35 1 Total 1203 1 ll 1 l 2 7 10 2 1 21 1 14 40 7 118 69 25 2 8 1 3 17 290 5 1 3 11 5 1 9 1 16 12 9 6 2 Total 1443 Unit 39 1 10 5 4 1 2 24 10 12 5 1 7 2 3 15 31 35 21 11 7 167 APPENDIX E VESSEL SHAPES AND RIM FORMS BY PHASE 836 837 Shs:b!~ A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A Bim 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 ~risnt I2ism :U:nit 6 6 9 9 10 10 10 11 11 11 12 12 12 12 12 12 12 13 13 14 14 14 15 15 16 2 2 2 2 2 2 2 10 10 11 2 7 7 7 7 7 7 8 19 20 25 26 27 29 29 29 33 33 34 34 35 37 37 2 28 2 24 1 30 30 1 27 34 6 7 24 28 31 32 34 17 30 4 7 11 6 22 7 2 21 26 28 36 36 38 38 38 30 36 20 MANACBAQUJ: PHASE R:IMS I.!~v~l 24 7 8 9 9 9 9 7 7 13 7 16 17 15 15 16 19 22 11 11 11 11 12 18 12 21 20 19 16 20 24 17 11 28 9 20 16 20 9 10 48 18 30 9 8 32 20 9 19 13 19 21 8 11 9 10 12 17 12 10 Shsg~ Bim A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A 2 2 2 2 2 2 12 12 12 13 13 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 4 4 4 4 4 4 4 4 4 4 4 4 5 5 5 5 5 5 5 13 14 14 14 14 15 15 16 16 16 17 18 18 18 18 19 23 24 26 26 ~~igDt ~ism 13 9 12 12 12 12 13 13 13 13 13 14 15 9 9 9 10 10 11 11 13 14 Unit 29 31 36 1 5 29 29 26 28 34 39 13 39 8 18 24 35 4 25 32 32 29 0 28 5 34 21 24 26 15 30 30 31 37 5 26 30 31 31 36 30 29 34 36 25 26 34 29 35 24 36 26 21 18 19 35 36 36 36 36 l!~v~l 16 20 14 12 33 12 16 18 17 :..1 7 11 7 10 18 17 16 28 0 18 19 14 0 18 33 11 4 14 13 39 15 16 16 8 27 17 15 14 14 12 14 13 11 14 5 12 11 16 20 19 13 19 6 7 6 10 11 12 13 13 838 MANACBAQUI: PHASE RDIS .Sh'l:r2~ Bim :ls!.I:iii!.Dt 1:2ism! Ynit A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A B B B B 5 5 5 5 5 6 6 6 6 6 6 6 6 6 6 7 7 7 7 7 7 7 7 8 8 8 8 8 8 8 8 9 9 9 9 9 9 9 10 10 10 10 10 10 12 13 14 11 12 13 13 14 14 11 11 11 11 11 11 13 15 16 24 25 19 22 22 22 23 24 25 27 29 12 12 12 12 13 13 14 14 13 14 14 12 12 12 12 12 13 13 12 16 13 13 1 1 1 1 a a a a 19 24 28 28 28 20 4 4 6 27 5 4 6 20 6 22 25 34 34 38 38 22 25 5 20 26 26 27 30 20 22 1 18 20 38 32 34 22 37 19 20 20 21 30 11 19 24 24 24 8 26 35 2 2 31 13 2 2 8 13 L~:!l~l 13 19 14 16 13 10 27 28 30 15 33 29 34 12 29 17 8 11 12 10 10 18 8 33 12 16 18 17 16., 10 19 21 11 11 10 17 12 17 11 15 11 12 13 18 5 7 9 13 15 8 12 11 23 23 16 9 20 23 8 10 Sh'l:b!~ Rim B B B B B 3 B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ~'l.:iii!.nt a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a 1:2i'lm ~nit L~~l 15 15 15 17 18 18 18 18 18 18 19 19 19 19 19 19 19 19 20 20 20 21 21 21 21 22 22 22 22 22 22 22 22 22 22 22 22 22 22 22 22 22 22 22 23 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 28 30 43 41 4 7 7 10 14 16 11 12 12 12 13 13 13 13 5 9 11 6 6 7 7 12 13 13 13 14 14 14 15 15 15 15 15 16 16 16 16 17 17 18 18 14 14 14 14 15 15 15 15 16 16 17 17 17 17 17 839 MANACBAQOJ: PHASE RIMS Sha:Q~ B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B Bim Vg:t::is.nt l:2is.m unit 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a 24 24 26 26 26 26 26 26 26 26 26 27 27 27 27 27 27 27 27 27 28 28 29 29 29 29 29 29 29 29 30 30 32 32 32 32 32 32 32 32 32 32 32 32 32 33 33 34 34 34 34 34 35 35 35 35 35 35 35 35 J:.~~l 17 19 5 6 6 10 12 14 14 14 16 14 15 15 15 16 16 17 17 17 15 19 11 11 12 12 13 13 14 15 14 14 9 12 12 12 13 13 14 15 15 15 15 16 17 19 20 6 8 8 8 9 6 12 12 13 14 15 16 16 Sha:Q~ B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B Bim 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Ys.~is.nt ~is.m a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a 8 9 9 10 10 10 11 11 11 11 11 12 12 12 12 12 12 12 12 12 12 12 unit 35 35 36 36 36 36 36 36 36 36 36 36 36 36 36 36 36 36 36 37 37 37 37 37 37 37 37 37 37 37 38 38 38 38 38 38 39 39 5 12 30 1 5 30 1 1 6 17 28 5 5 5 6 14 18 20 28 28 28 28 I.!~v~l 17 18 9 11 11 11 12 12 12 12 12 12 12 12 12 12 12 13 14 8 9 9 9 10 10 10 11 11 11 12 8 8 8 8 8 9 6 7 26 9 15 22 30 15 21 21 30 8 17 24 29 33 27 13 10 8 13 15 15 16 840 MANACBAQUI PHASE RIMS Shsu:2~ Bim Ysu::is:mt l:li~m B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 13 a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 15 15 15 15 15 15 15 15 15 llnit 28 30 30 30 30 30 30 30 30 31 31 31 31 31 31 1 4 4 4 5 5 5 27 28 28 28 28 30 31 31 31 3 5 5 6 6 13 17 17 17 28 28 28 30 30 30 30 31 31 31 31 5 5 6 12 16 17 17 30 30 l!~~~l 16 13 14 15 15 15 15 15 18 14 14 14 15 16 17 20 19 20 26 27 27 28 20 16 17 17 17 15 13 15 17 1 27 33 27 27 8 6 39 45 13 14 14 12 13 14 15 16 16 16 16 26 33 27 11 40 7 42 14 18 Sh~:12~ B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B a a a a B B a a Rim 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 v~~i~nt Di~m a a a a a a b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b 16 16 16 16 16 16 Qnit 6 14 28 30 31 36 2 2 2 2 3 3 3 3 3 3 3 3 3 6 6 6 7 11 11 11 11 15 18 18 18 19 19 19 19 20 20 20 20 20 20 20 20 21 21 21 21 22 22 22 22 22 22 22 22 22 22 23 23 23 ~gv~l 29 26 17 18 13 13 20 20 22 22 20 20 20 21 23 23 23 24 24 19 26 33 7 5 7 9 10 35 6 8 14 4 11 14 16 5 7 9 9 10 10 11 11 10 :o 11 12 13 14 14 14 14 15 ,15 ... -0 l7 17 17 17 17 841 MANACBAQUZ PHASE RIMS Shg:b!~ Bim B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ~.:isnt b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b ;QiS!J!l :Unit 23 23 23 23 23 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 26 26 26 26 26 26 26 27 27 27 27 27 27 27 27 27 27 27 27 27 29 29 29 29 29 29 29 29 29 29 30 31 L~Y:~l 17 18 19 20 20 12 12 13 13 13 13 14 14 14 16 16 16 16 16 16 16 17 17 18 18 19 19 19 5 13 14 14 15 16 16 13 14 15 15 15 16 16 16 16 16 18 19 19 9 10 10 12 12 13 13 13 13 14 14 12 SlJg:Q~ Bim B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ~~ignt b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b Qism :Unit 31 32 32 32 32 32 32 32 32 32 32 32 34 34 34 34 34 34 34 34 34 34 34 34 34 34 35 35 35 35 35 35 35 35 35 36 36 36 36 36 36 36 36 36 36 36 37 37 37 37 37 38 38 38 38 38 38 38 38 38 ~~Y:~l 16 10 12 13 13 15 15 15 15 15 16 16 8 8 9 9 9 9 9 9 9 10 10 10 10 11 10 11 11 11 11 11 12 12 14 10 11 11 11 12 12 12 12 13 13 13 7 9 10 10 11 9 9 9 9 9 9 9 9 9 842 Sh5li2~ Rim B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 :sle.::iS~.Dt b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b Q;i.SU!l Unit Ls:vs:l 9 9 10 10 10 10 11 11 11 11 11 11 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 13 13 13 13 13 13 13 13 13 13 13 13 13 14 14 14 14 14 14 14 14 14 14 14 14 15 38 38 39 39 39 39 12 30 4 5 30 31 1 5 5 6 28 31 4 5 6 6 6 6 6 12 15 28 28 28 30 30 31 31 5 5 5 5 5 6 6 17 28 28 30 30 31 5 5 5 6 14 17 28 30 30 31 31 31 16 9 10 5 6 6 7 9 15 24 26 15 13 22 28 30 25 15 16 26 27 29 30 30 31 31 4 54 16 17 18 13 15 14 15 25 26 28 30 30 28 30 43 14 17 15 15 14 24 29 31 30 40 45 15 15 18 10 14 17 42 MANACBAQU:I PHASE RDIS :2he!2~ Eim B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 :sla~iSlnt b b b b b b b b b b b c c c c c c c c a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a Qiam Qnit 15 15 15 15 15 16 16 16 17 18 18 10 12 14 17 22 22 31 31 6 28 31 17 1 6 13 27 30 32 32 5 1 6 2 2 2 2 2 3 3 4 5 6 6 7 7 7 7 7 7 7 8 8 8 8 8 8 8 8 8 8 8 8 11 11 11 11 16 18 18 18 18 18 18 :bl~v~l 12 15 18 13 14 32 13 13 44 22 30 9 18 13 13 14 30 22 30 19 19 20 22 23 20 23 22 33 25 30 7 7 8 9 9 9 10 8 8 8 8 9 9 9 10 10 10 10 :.2 2 5 7 9 49 7 7 8 8 9 9 843 MAHACHAQUX PHASE RDIS Shial2~ Bim B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 :ila..::i~an:t a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a Dia.m Qnit 18 18 18 18 18 18 18 18 18 18 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 20 20 20 20 20 21 21 21 21 21 22 22 22 22 22 22 22 22 22 22 22 22 l:i~!l~l 10 10 12 13 13 14 15 15 16 16 4 8 9 9 10 11 11 11 13 13 13 13 13 14 14 14 15 16 16 7 7 10 10 11 11 11 12 12 12 12 12 13 13 6 8 9 10 12 15 15 15 16 16 17 17 18 18 18 18 18 Sha.t2~ B B B B B B B B B B B B B B B B B B B B E B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B Bim 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 ~~i~ant a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a Dia.m Qnit 22 22 23 23 23 23 23 23 23 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 25 25 25 26 26 26 26 26 26 26 26 26 26 26 26 26 26 26 26 26 26 26 26 26 26 L~:sz:~l 18 19 13 18 18 19 20 20 20 10 12 14 15 16 16 17 17 18 18 18 18 18 19 19 19 19 19 19 19 19 19 20 20 20 20 6 7 8 6 6 8 10 12 14 14 15 16 16 16 16 16 16 16 16 16 16 17 17 17 17 844 MANACHAQUI: PHASE RIMS Sbaa~ Bim B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B 2 2 2 2 2 2 2 2 B B B B B B B B B B B B B B B B B B B B B B B B 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 ~;z;:iant a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a I:! ism Ynit 26 26 26 26 26 26 26 27 27 27 27 27 27 27 27 27 27 27 27 27 27 27 27 27 27 27 27 27 27 28 28 28 28 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 30 30 31 31 31 31 31 32 L~vd 17 17 17 17 18 18 19 13 13 14 14 15 15 15 15 15 16 17 17 17 18 18 18 18 18 19 19 20 22 18 19 19 22 11 12 13 13 14 14 14 15 15 15 15 16 16 16 16 17 17 17 18 17 20 3 14 16 16 18 14 Shaa~ Bim B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 Va~iant ~ism a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a Ynit 32 32 32 32 32 32 32 32 32 32 32 32 32 32 32 33 33 33 33 33 34 34 34 34 34 34 34 34 34 34 34 34 34 35 35 35 35 35 35 35 35 35 35 35 35 35 35 35 36 36 36 36 36 36 36 36 36 36 36 36 :b!~:l~l 14 14 14 14 16 16 17 17 17 18 19 19 20 20 20 16 17 18 18 19 8 8 9 9 10 10 10 10 11 11 11 11 11 3 3 11 11 12 12 14 15 15 15 16 16 16 16 17 9 9 10 11 12 13 13 13 13 13 14 14 845 MANACHAQO::t PHASE R:tMS Shsi2~ Bim va:risnt DiSlm Ilnit B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a 9 9 10 10 10 10 10 10 10 11 11 11 11 11 11 11 11 11 11 11 11 11 11 12 12 12 12 12 12 12 12 36 36 36 36 36 36 37 37 37 37 37 37 37 38 38 38 38 38 38 38 38 39 39 39 39 39 39 39 39 5 25 1 4 5 6 6 28 30 1 1 1 4 4 5 5 13 18 25 27 28 30 30 1 1 1 5 5 5 6 6 I.!~:ll~J. 14 14 14 15 16 16 10 10 11 11 11 13 14 9 9 9 9 10 10 11 11 7 7 7 7 7 8 9 9 30 5 20 25 28 25 32 22 16 18 20 21 19 27 28 29 13 9 21 19 15 18 17 19 20 26 32 33 24 28 Sllsi2~ Bim YSl:tisnt Qigm Unit B B B B B B B B B B B B B B 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 6 6 13 13 14 14 14 24 24 24 28 28 28 28 28 30 30 30 30 30 30 31 31 31 31 31 31 32 1 1 1 3 4 4 4 4 5 5 5 5 5 6 6 6 8 16 17 19 24 24 25 25 28 30 30 30 31 31 31 31 ~~:ll~J. 30 30 10 11 33 35 40 19 22 23 17 19 19 21 22 15 15 16 16 17 19 14 15 15 16 18 20 20 17 20 23 25 22 24 29 29 28 29 29 31 32 25 32 34 10 46 11 16 10 20 1 6 19 19 19 19 15 19 20 20 846 MANACHAQUI PHASE RDIS ~hai2~ Bim 1!a:r::isnt Dism :unit B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a b b b b b b b 13 13 13 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 15 15 15 15 15 15 15 15 15 15 15 16 16 16 16 16 17 17 17 17 17 18 18 19 19 31 36 36 4 5 5 5 6 6 6 13 13 16 17 25 25 25 27 28 28 30 30 30 30 31 31 31 37 1 1 5 6 11 12 15 16 21 25 30 4 11 11 30 35 1 14 24 30 31 6 30 5 28 5 5 7 7 7 14 19 ~!;i1j!!ii1l 21 14 15 28 31 31 31 29 30 33 13 14 39 11 6 6 7 11 19 19 16 19 19 19 15 17 18 12 19 24 31 33 10 10 34 41 15 7 17 26 8 10 19 3 23 34 23 19 18 30 15 27 21 32 34 7 7 10 34 9 ~hal2!ii1 Bim B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 Vs~isnt b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b a a a a a a a a a a a a a a a a a a a a a a a a a a J:lism Ynit 10 11 12 12 13 13 13 13 13 13 14 14 14 14 15 15 15 16 16 18 19 19 20 24 24 26 26 29 32 35 37 37 37 38 30 5 1 4 5 5 5 5 6 21 5 24 25 28 1 30 30 4 5 5 2 2 3 3 3 3 5 5 6 6 7 7 7 7 7 7 7 ~!;;1V!;i1:!, 11 11 7 16 19 6 18 14 16 18 11 11 11 10 18 33 23 24 30 30 32 32 29 15 33 22 6 21 24 17 17 30 31 33 25 26 22 22 22 23 33 34 29 33 7 7 7 7 7 8 8 7 11 8 8 8 8 8 8 8 8 8 8 8 9 9 9 9 10 847 MAHACHAQU:I PHASE R:IMS Sba:bl~ B B B B B B B B B B B B B B B B B B B B B B B B B B B 8 8 B B 8 B B B 8 B B B B B B B B B B 8 8 B 8 8 B B 8 B B B B B B Bim Jla:z:iant 1:2iam IIDit 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a 8 10 11 14 15 16 18 18 18 18 18 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 20 21 21 21 22 22 22 22 22 22 22 22 23 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 26 26 26 26 tt~3l:~l 12 10 9 29 29 16 4 11 14 16 16 8 12 13 13 13 14 15 6 6 7 9 9 10 11 12 12 13 10 12 14 15 16 18 18 18 18 19 19 20 7 12 12 15 16 17 17 18 19 19 20 20 20 20 21 22 6 7 8 16 Sba:bl~ B B B B B B B B B B B B B B B B B B B B B B 8 8 8 B B 8 8 B B 8 8 8 8 B B 8 B 8 8 8 B B B B 8 8 8 8 B B B 8 B B B B B B Bim Jla:z:iant Qiam Yoit 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a 26 26 26 26 26 26 26 26 26 27 27 27 27 28 28 29 29 29 29 29 29 29 29 29 31 32 32 32 32 32 32 32 33 34 34 34 34 34 34 34 34 35 35 35 35 35 35 36 36 36 36 36 36 36 36 36 36 36 36 36 ~~3ld 16 16 17 17 17 17 17 18 19 18 18 19 21 18 19 14 15 15 15 15 15 16 16 19 16 12 15 16 16 17 17 20 14 9 9 10 10 10 11 11 11 8 8 10 13 15 19 4 10 11 12 12 13 13 13 14 14 14 14 15 848 MANACHAQUI: PHASE RI:MS Shii!.l2~ Bim :slsu::iii!.nt Diii!.m lhlit B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a 9 10 10 10 11 11 11 11 11 12 12 12 12 12 12 12 12 12 12 12 12 12 12 13 13 13 13 13 13 13 13 13 13 13 14 14 14 14 14 36 36 37 37 37 37 37 37 37 37 37 37 37 38 38 38 38 39 39 39 39 4 13 28 31 5 28 28 30 31 4 5 6 6 6 13 17 22 28 30 30 30 30 31 1 4 4 5 5 5 6 12 25 28 30 5 6 28 30 30 L~v~l 15 16 10 11 11 11 11 11 11 11 12 12 13 9 10 10 10 7 7 9 9 20 9 18 11 32 21 22 19 14 29 32 31 32 32 10 43 11 19 16 16 18 18 14 20 27 29 30 32 34 32 9 6 24 15 29 32 13 14 20 Shii!.l2~ Bim B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 :slii!.~iii!.nt a a a a a a a a a a a a a a a a a a b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b Diii!.m 14 14 15 15 15 15 15 15 16 16 16 16 16 16 16 16 17 17 ~nit L~:2:~l 36 36 4 4 5 22 30 31 1 5 6 6 15 25 30 31 5 6 2 2 2 3 3 5 6 7 7 7 7 7 8 15 18 18 19 19 19 19 19 20 20 21 22 22 22 24 24 26 26 29 29 32 34 34 35 35 35 36 37 39 14 17 27 30 28 19 16 15 19 30 32 33 29 7 15 13 30 32 20 21 21 24 25 31 30 7 9 9 9 9 12 3 14 15 9 11 13 13 14 7 11 12 12 16 17 18 20 10 15 13 16 14 8 12 12 16 16 13 8 7 849 :;!hgg~ B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B Rim VSl.::iSlnt D;i.sm~ Ynit; 3 b 10 1 10 3 b 25 3 b 10 30 3 b 11 16 3 11 b 25 3 12 b 5 3 12 b 14 3 b 12 17 3 b 12 17 3 12 b 22 3 b 13 1 3 b 13 5 3 b 13 6 3 13 b 6 3 b 13 19 14 3 b 5 3 14 b 30 3 b 15 31 3 16 b 5 3 16 b 6 3 18 b 4 3 18 b 5 3 18 b 5 3 b 18 6 3 b 19 22 4 2 4 3 4 7 4 19 4 20 4 20 4 20 4 22 4 24 4 24 4 24 4 25 4 26 4 26 4 26 4 27 4 27 4 27 4 28 4 29 4 32 4 32 4 32 4 32 4 33 4 34 4 36 4 36 4 36 4 36 4 38 4 39 4 39 4 10 30 4 12 31 ~~:sz:~l 17 5 19 45 6 29 33 11 40 17 23 27 29 34 16 32 14 17 30 34 15 31 31 34 15 18 27 9 12 8 9 11 17 15 15 15 7 6 14 18 15 15 16 18 11 12 16 16 17 20 12 9 12 12 14 9 6 7 18 12 :;!bSla~ B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B MANACHAQOI PHASE RIMS Bim V51ri51nt QiSlm Ynit ~~:sz:~l 13 29 5 4 13 22 11 4 13 22 11 4 13 18 28 4 13 18 28 4 14 5 32 4 14 5 32 4 14 30 19 4 14 16 31 4 15 1 21 4 15 28 6 4 15 12 21 4 17 15 28 4 15 16 31 4 16 10 13 4 16 12 21 4 22 3 5 11 19 5 5 20 5 20 12 5 17 22 5 20 23 5 4 26 5 4 26 5 26 15 5 27 17 5 27 19 5 14 30 5 34 11 5 14 36 5 37 10 5 10 30 16 5 28 11 5 5 27 12 4 5 12 20 22 5 12 29 11 5 26 13 4 5 13 28 19 5 20 13 28 5 13 12 35 5 14 16 31 5 15 18 1 5 15 17 29 5 15 31 13 5 14 15 36 5 16 26 5 5 16 32 6 5 17 16 24 5 17 16 26 5 17 23 1 5 17 17 31 5 17 17 32 5 21 20 4 5 26 2 6 6 20 6 9 20 6 11 21 6 16 22 6 17 23 6 16 26 6 850 MANACBAQOJ: PHASE RIMS Shs:!.:l;!~ Bim B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B 6 6 6 6 6 6 6 B B B B B B B B B B B B B B 6 6 6 6 6 6 6 6 6 6 6 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 8 8 8 8 8 8 8 8 8 9 9 9 9 9 9 9 9 9 9 9 9 9 9 ~su:::is:!.nt l2is:!.m Ynit 8 8 9 9 10 10 11 11 11 11 12 12 14 15 17 11 12 12 12 13 13 13 13 13 12 12 12 12 13 14 18 18 29 30 35 30 31 11 32 28 28 17 21 31 31 20 30 30 28 34 18 20 20 22 27 27 27 35 38 39 5 19 20 30 5 21 22 30 31 29 6 7 24 26 1 1 28 31 2 2 3 3 3 4 6 6 6 7 7 7 7 7 t.~~~l 11 12 11 15 12 6 10 17 19 43 11 14 18 11 14 17 20 9 14 6 13 11 18 18 20 14 11 6 27 16 12 16 31 13 16 16 14 10 30 11 18 13 18 18 15 15 19 25 17 25 25 25 24 32 33 7 8 8 9 9 Slls:!.:l;!~ Bim B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 B B B B B B 9 9 9 9 9 9 9 9 9 9 9 9 9 B 9 Vs:!.~is:!.nt ~is:!.m Ynit 7 7 7 7 8 8 8 8 8 10 13 14 14 18 19 19 19 19 20 20 20 20 20 20 21 21 21 22 22 23 24 24 24 24 24 24 24 24 24 24 24 24 24 26 26 26 26 26 26 26 26 26 27 27 27 27 27 28 28 28 L~~d 9 9 12 13 9 9 10 10 12 10 12 29 33 13 13 14 15 17 8 10 11 11 11 19 12 12 13 15 18 18 14 14 14 15 15 18 18 19 19 19 20 20 21 6 8 10 13 15 15 16 17 19 18 19 20 20 21 19 20 21 851 MANACRAQtn PRASE RXMS Sh~:&!~ Bim B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 Y~;r;:ignt I;!i~m Ilnit 29 29 29 29 29 29 29 30 30 31 32 32 32 32 32 32 32 32 32 32 33 33 34 34 34 34 34 34 34 34 34 34 34 35 35 35 36 36 36 36 36 36 36 36 36 36 36 36 36 37 37 37 37 37 37 38 38 38 38 39 ~~::ll:~l 10 11 13 15 15 16 17 12 12 20 9 9 13 15 16 17 17 18 19 19 17 18 9 9 10 10 11 11 11 11 11 12 12 14 16 16 1.1 13 13 13 13 14 14 14 14 14 14 14 15 11 11 11 12 12 15 4 9 10 10 7 SbgJ;l~ Bim B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 10 10 10 10 10 yg;r;:i~nt Qi~m 10 12 12 12 12 12 12 12 12 12 12 12 13 13 13 13 13 13 13 14 14 14 14 14 14 14 14 14 15 15 15 15 15 15 15 15 15 16 16 16 16 16 17 17 17 17 17 18 18 18 19 21 22 24 Unit ~~v~l 39 30 6 24 28 28 30 30 30 31 31 31 36 4 4 6 28 28 30 31 1 5 6 30 30 30 31 31 31 1 4 6 6 12 24 24 25 31 4 14 22 25 31 1 1 7 15 31 22 20 22 17 17 19 18 19 20 14 25 29 32 19 21 15 17 23 28 24 17 17 18 18 22 23 24 27 13 11 14 28 1 22 31 5 6 6 28 2 3 3 11 11 16 19 20 20 17 29 32 33 14 22 23 24 3 8 31 32 10 19 19 9 18 28 44 19 0 14 20 24 852 MAHACHAQtn PHASE RXMS Sbsl2~ B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B Ri.m Veriant Qimn IDlit 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 11 14 15 17 18 18 19 19 19 19 19 20 20 21 21 21 22 22 23 23 24 24 24 24 24 24 26 26 27 27 27 27 27 27 27 27 28 28 29 29 29 29 29 29 29 30 34 34 35 35 35 35 36 36 36 36 36 36 37 37 w~:2:~l 9 21 3 44 13 15 13 13 13 14 14 10 10 5 9 10 15 15 16 18 12 13 13 16 17 17 6 8 14 15 15 16 16 16 17 18 15 17 10 12 12 14 14 14 14 13 8 9 9 13 14 16 9 11 12 12 12 13 9 12 Sh212~ Rim B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 c c c c c c c c c c c c c c c c Va~i.ant Qism Ynit 10 12 12 12 13 13 13 14 14 14 14 14 14 14 15 15 15 15 15 15 15 16 16 16 16 16 16 16 16 16 16 16 16 17 17 17 18 18 18 18 19 19 22 a a a a a a a a a a a a a a a a 38 1 6 22 28 4 5 22 6 6 13 16 22 22 30 5 6 15 28 28 30 30 5 5 5 5 6 15 22 22 28 30 30 37 16 28 32 5 5 6 12 1 5 5 1 4 6 7 13 18 19 19 22 24 29 29 30 36 36 36 L~:ll~1 9 18 22 16 15 25 30 16 29 31 11 44 13 14 14 29 31 38 12 17 15 17 28 28 31 31 27 44 15 16 17 13 16 12 40 17 17 29 31 27 9 19 29 31 23 14 32 9 11 6 16 17 13 10 7..7 7..7 19 12 l3 14 853 MANACHAQO:I PHASE RIMS SbsU;l~ Bim Y5!.:ti"'nt I:lism IInit c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a b b b b b b b b b b b a a a a a a a a 10 39 37 11 11 11 12 12 12 12 12 12 12 12 12 12 12 32 6 6 13 19 19 21 24 26 29 30 30 13 13 11 13 13 13 14 14 14 14 14 14 14 15 15 15 15 16 16 16 17 17 17 18 18 18 18 7 12 13 14 16 17 18 18 19 19 28 29 16 21 22 23 25 30 34 6 6 13 29 6 19 19 11 25 31 6 13 19 21 7 19 25 13 11 1 21 25 29 4 25 11 11 12 12 19 19 22 34 37 5 22 l!~:ll~l 8 12 9 20 32 34 12 15 16 14 19 19 11 18 19 10 13 14 23 17 46 14 14 20 6 18 10 30 30 11 17 33 15 15 6 10 17 28 7 16 11 14 5 4 10 20 13 9 15 27 9 9 13 16 17 12 11 33 15 Sbsl2~ Bim c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 4 4 4 4 4 4 4 5 5 6 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 D D D D D D D D D D D D D D D D D D D D D D D Y5!.~isnt a a a a a a a a a a b b b b b b b b b I:lism Unit 12 12 13 14 14 16 16 18 19 19 12 12 13 13 14 18 11 11 11 11 12 12 12 14 14 12 28 26 36 18 19 32 12 26 4 4 13 6 32 35 6 20 30 35 35 4 7 36 39 39 20 22 32 36 30 30 18 30 31 17 30 1 2 4 8 18 20 20 21 22 23 24 26 27 27 32 32 35 11 11 13 13 14 15 15 15 16 16 1 1 25 3 6 12 6 14 l:i~:ll~l 17 15 18 17 17 10 19 29 30 33 20 14 35 13 20 18 14 26 9 14 7 10 12 18 15 14 16 16 4 16 17 46 16 21 23 27 10 8 12 12 15 13 17 17 15 17 18 l3 16 17 7 20 21 10 23 27 9 34 36 854 MANACHAQU:I PHASE R:IMS Sbs:!.:g~ D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D Bim :llii!..::is:!.Dt I:liam Ilnit 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 3 3 3 3 17 18 18 19 21 22 23 23 26 12 13 13 14 18 18 3 3 3 3 3 3 3 3 3 4 4 4 4 4 4 4 4 4 4 4 4 4 5 5 5 5 5 5 5 5 5 5 5 12 13 14 14 15 17 17 14 14 15 16 18 19 20 20 20 15 17 18 10 14 32 4 31 5 5 13 6 22 27 29 32 32 32 38 39 28 20 29 28 24 37 26 26 26 27 34 39 36 5 5 24 1 27 39 7 8 11 39 19 31 1 25 31 6 6 8 8 18 24 24 24 27 34 38 38 5 36 24 L~:ll:~l 11 31 14 26 18 30 30 12 27 15 19 10 15 16 17 8 8 7 10 11 16 21 9 6 10 13 15 9 8 12 27 28 17 20 19 8 9 9 6 7 16 17 21 6 20 35 27 8 8 15 15 15 19 18 12 9 9 28 9 15 Shag~ Bim :llii!..::is:!.nt I:liam Unit D D D D E E E E E E E E E E E E E E E E E E E E E E E E E E 5 5 6 6 1 1 1 1 1 1 2 2 2 2 2 2 3 3 3 3 3 4 4 4 5 5 5 6 6 7 18 18 10 10 10 14 14 18 11 11 12 14 13 14 16 18 14 16 16 16 16 18 22 18 24 24 5 34 34 34 5 32 35 22 6 23 26 31 31 28 11 3 16 26 17 31 14 25 31 38 25 27 16 30 I.i~:ll:~l 15 15 31 12 11 11 31 14 14 14 34 20 17 18 17 22 7 26 45 21 41 17 42 6 19 10 0 20 44 17 855 stJITACOCHA PHASE RIMS Sluu;~~ Rim A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B 14 14 14 14 14 14 14 14 14 14 14 14 14 14 15 15 15 15 15 15 15 15 16 16 16 16 16 16 16 17 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 :..1 11 11 11 ~~ism.t t!iam IInit 10 10 10 11 11 14 10 13 13 14 14 11 12 12 12 17 a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a 4 4 18 19 22 24 26 36 5 6 27 1 1 24 5 30 30 11 12 27 30 30 7 24 36 35 22 31 36 25 1 1 2 2 3 3 4 5 5 5 6 6 6 6 7 7 7 7 8 10 11 11 11 11 11 11 11 11 15 15 I.!~v~l 20 20 8 6 12 12 8 9 23 25 14 19 19 12 21 10 10 8 4 10 10 10 6 8 7 11 9 10 8 5 20 21 21 25 21 22 19 21 22 22 25 26 26 26 4 4 7 7 7 7 2 4 4 6 8 8 8 9 18 30 Sha:r2~ Bim B B B B B B B B B B B B B B B B B B B 3 3 B B B B B B 3 3 3 B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 :sla~i.Sl.nt a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a t!iam IInit 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 18 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 21 21 21 21 21 21 21 21 21 21 21 21 22 22 22 22 22 22 22 22 22 22 23 23 23 L~v~l 6 6 6 7 7 7 7 8 8 8 8 9 10 10 11 11 12 8 8 8 8 9 9 10 10 11 12 6 6 7 7 8 8 10 12 6 7 7 7 8 8 8 8 9 11 11 11 8 9 10 11 11 11 12 12 13 13 14 16 17 856 SUXTACOCHA PHASE RD!S Sh5!.!2~ Bim YS!.;(iS!.nt I:2iS~.m Qnit ~~v~l B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B E B B B B B B B B B B B B B B B B B B B B B B B B B B B 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a 23 24 24 24 24 24 24 24 24 24 24 24 24 24 24 25 26 26 26 26 26 26 26 26 26 27 27 27 27 27 27 29 29 29 29 29 29 29 29 29 29 30 31 31 32 32 32 32 32 32 32 32 33 33 34 34 34 34 34 35 20 1 9 9 9 10 10 11 11 12 12 12 12 12 13 5 5 6 7 7 8 9 9 9 10 11 11 12 12 12 14 7 10 10 10 11 11 11 11 11 13 10 13 14 8 10 11 12 12 12 13 14 14 15 7 7 7 7 7 8 Sh5!.!2~ R;i.m B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 Va;riS~.nt D;i.gm Qnit I.!~v~l a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a 7 8 9 9 9 9 9 10 10 10 10 10 10 10 10 10 10 10 11 11 11 11 35 35 35 35 35 35 35 35 35 35 36 36 36 36 36 36 36 36 36 37 37 37 37 37 38 38 38 38 38 38 38 38 38 38 38 38 39 39 12 16 1 5 5 30 31 28 28 28 30 30 30 30 31 31 31 31 4 4 5 5 9 9 9 10 10 11 11 11 11 12 8 9 9 10 10 10 10 12 12 6 7 7 7 8 6 8 8 8 8 8 8 8 8 8 9 9 6 6 10 2 15 24 25 13 11 13 13 13 11 12 13 14 11 12 13 13 22 22 22 22 857 ~hga~ B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B Rim yarignt Diam ];lnit 5 11 a 11 6 11 11 a 6 11 11 a 6 11 11 a 14 11 a 11 28 11 a 11 30 11 a 11 31 11 a 11 31 11 a 11 12 1 11 a 6 11 a 12 6 11 12 a 25 11 12 a 28 11 a 12 28 11 a 12 30 a 12 11 31 12 11 a 31 12 11 a 31 12 a 11 31 a 12 11 12 31 a 11 31 12 a 11 31 12 11 a 4 13 11 a 4 13 11 a 4 11 a 13 12 a 13 11 30 13 a 11 31 13 a 11 31 a 13 11 4 14 a 11 4 14 11 a 14 4 a 11 14 4 a 11 4 14 11 a 14 5 11 a 24 14 11 a 28 14 11 a 30 14 11 a 30 11 a 14 14 31 a 11 12 15 11 a 2 11 b 2 b 11 4 b 11 6 b 11 6 b 11 10 b 11 10 11 b 18 b 11 b 18 11 18 11 b 18 11 b 18 b 11 19 b 11 19 b 11 19 b 11 20 11 b 20 b 11 21 b 11 L~v~l 25 26 27 27 36 12 11 11 13 22 25 25 6 11 12 11 10 11 13 13 13 13 14 22 23 23 11 11 11 14 19 22 22 22 22 23 16 14 10 11 11 7 19 23 15 17 25 5 5 7 8 12 13 13 8 9 10 7 7 3 ~hga~ B 3 3 3 B B 3 B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B StJ:t'l'ACOCBA PHASE R:tMS Rim Vsrignt Digm J;lnit L~v~l b 22 9 11 22 10 b 11 22 11 b 11 22 11 b 11 12 b 22 11 14 22 b 11 b 23 13 11 23 14 b 11 24 3 b 11 5 b 24 11 7 b 24 11 24 9 b 11 24 9 b 11 3 b 26 11 3 b 26 11 26 9 b 11 10 26 b 11 27 9 b 11 10 27 b 11 27 13 b 11 7 29 b 11 8 b 29 11 8 b 29 11 10 b 29 11 29 11 b 11 11 29 b 11 11 29 b 11 13 29 b 11 2 30 b 11 2 30 b 11 1 32 b 11 7 32 b 11 9 b 32 11 9 32 11 b 9 32 b 11 10 32 b 11 10 32 b 11 11 32 b 11 12 32 b 11 13 32 b 11 12 b 33 11 8 b 35 11 10 b 35 11 10 35 b 11 7 37 b 11 37 8 11 b 37 8 11 b 37 11 b 11 23 8 6 b 11 25 6 b 9 11 7 28 b 9 11 7 28 b 9 11 24 10 6 b 11 8 21 b 10 11 15 23 b 1.0 11 24 3 b 10 11 4 24 b 10 11 9 10 30 b 11 11 10 31 b 11 11 10 31 b 11 858 StnTACOCHA PHASE RDIS Sh5ll2~ B B B B 8 8 B B B B B B 3 3 B B 3 3 3 3 B 3 B 3 3 B 8 B B B B B B 3 B B B B B B 3 3 3 B B B 3 3 3 3 3 3 3 3 3 B B B B 3 Bim 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 ~;r.:i5lnt b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b c c c c c c c c c c c c c c c c c c c c c c c c c c c c ;Qis;!,m Ynit 10 10 11 11 11 11 11 11 11 12 12 12 12 12 12 12 12 12 12 12 12 13 13 14 14 14 14 14 14 14 14 15 31 32 3 4 5 6 24 31 35 6 6 6 18 30 30 30 31 31 31 31 31 3 37 3 4 6 28 30 31 32 32 24 2 4 6 6 6 6 6 6 7 8 12 17 18 19 19 19 20 20 20 20 20 21 21 22 23 24 24 26 l:!~:!.!:~l 12 8 22 15 24 25 9 11 10 24 27 27 12 10 12 16 7 8 9 10 11 22 6 22 20 25 6 6 9 10 10 11 18 18 2 18 23 24 25 25 3 6 9 10 7 4 8 8 4 4 5 5 8 4 9 11 13 7 7 4 Sh2l2~ Bim B B B 11 8 8 8 8 8 8 8 8 8 B 8 8 8 B 8 8 8 B B B 8 8 8 B B B 8 B 8 B 8 B B B B B B B B B B B B B B B B B B B B B B 8 B B B 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 ~~i2nt c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c d d Oi2m :unit 7 8 8 9 9 9 10 10 10 10 10 10 10 10 10 11 11 11 11 11 11 11 12 12 12 12 12 12 13 13 13 13 13 13 14 14 14 15 15 15 15 27 29 29 29 30 30 30 32 32 32 34 35 37 37 37 38 38 22 24 28 5 17 31 4 6 14 19 21 28 28 30 30 4 4 6 28 28 28 35 3 4 4 28 30 30 4 4 4 4 4 25 4 16 20 4 5 6 15 3 6 J:t~v~l 10 6 10 10 8 10 11 9 9 9 6 8 7 8 11 6 7 12 8 13 23 5 11 19 25 33 7 5 6 10 10 11 15 23 27 5 10 11 12 14 19 23 12 9 10 19 20 20 22 23 5 21 39 9 20 26 25 17 21 23 859 Sh~:b!!i: Bim B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B 11 11 11. 11 11 11. 11. 11. 11 11 11 11. 11 11. 11. 11. B B B B B B B B B B B B B B 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 :L~;r;:i~nt d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d e e e e e e e e e e Qigm IDlit 8 8 8 10 10 10 10 10 11 11 11 12 12 12 12 12 12 12 12 12 13 14 14 14 14 16 l:!!i:~!i:l 10 19 20 21 23 23 24 24 24 24 24 24 27 29 30 31 32 32 34 35 36 38 38 39 6 30 8 8 4 6 13 14 6 7 7 7 9 9 12 8 8 16 10 13 7 7 8 7 8 6 26 12 31 11 5 6 6 31 37 6 23 23 25 11 31 5 6 6 6 24 24 30 31 31 5 5 6 6 30 1 2 3 6 7 11 18 19 20 22 22 11 7 25 4 13 23 23 24 26 10 10 11 11 14 24 21 25 25 11 18 18 21 25 7 2 5 8 7 8 13 Sha:12~ B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B SUITACOCHA PHASE RIMS Bim Va;r;:ignt Qigm :unit l::.!!i:~!i:l 1.1 1.1. 1.1. 1.1. 11 1.1. 1.1. 1.1 1.1. 1.1. 1.1. 1.1. 1.1 1.1 11 11 1.1. 1.1 1.1 1.1 1.1 1.1 1.1 1.1 1.1 1.1 1.1 1.1 11 1.1 1.1 1.1 1.1 1.1. 1.1 11 11 11 11 1.1 1.1 11 1.1 1.1 1.1 11 11 11 11 1.1 11 11 11 11 11 1.1 11 11 11 11 e e e e e e e e e e e e e e e e e e e e e e e e e f f f f f f f f f f f f f f f f f f f f f f f f f f f f g g g g g g g 8 9 10 10 23 32 32 34 36 36 36 36 36 36 37 38 38 39 31 4 30 31 11 4 12 12 12 12 12 15 3 14 31 36 36 6 4 12 18 19 20 20 21 21 21 21 27 35 36 6 31 5 9 9 10 10 11 12 12 12 13 13 13 13 13 14 16 11 31 6 28 31 4 4 28 31 37 30 6 3 5 6 19 19 22 22 13 9 10 7 8 9 11. 11. 11. 11. 6 7 8 4 12 19 9 13 19 21 21 13 11 11 29 24 9 8 7 5 5 5 6 11 11 11 10 10 24 10 24 2 11 27 10 11 22 23 12 13 9 11 27 16 25 24 6 8 8 9 860 stnTACOCBA PHASE R:tMS Sh~:b!Sl B B B B B 3 B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B c c c c c c c c c c c c c c c c c c Bim 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 13 13 13 13 13 13 13 13 13 13 13 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 ~.:i~nt g g g g g g g g g g g g g g g !:!ism unit 9 9 10 11 11 12 12 10 12 12 14 14 15 15 15 18 10 10 11 12 12 14 15 22 22 22 22 27 27 39 39 4 11 4 4 6 11 31 2 22 22 23 23 23 31 22 6 30 19 30 5 22 30 30 21 23 36 36 13 31 31 20 31 31 6 4 13 15 18 18 19 19 19 22 22 22 24 24 27 27 27 32 32 LS1~~l 10 10 10 15 8 10 6 6 20 6 20 23 24 5 9 18 8 8 9 9 12 11 8 26 10 7 11 25 9 10 9 9 20 9 11 9 12 12 9 13 12 26 22 9 33 4 10 9 9 9 3 8 10 9 12 9 11 12 8 10 Sh~:b!~ Bim c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 8 8 8 8 8 8 8 8 8 9 9 9 9 9 9 9 10 8 8 8 8 8 8 8 9 9 9 9 9 9 10 E E E E E E E E E E E E E E E F F F F F F F F F :llsu:i~nt Di~m 9 11 11 12 12 13 13 13 14 14 14 15 16 16 14 14 14 15 15 20 14 14 16 16 17 16 11 13 18 18 8 11 12 27 12 11 1 1 1 1 1 1 1 1 1 a a a a a a a a a Ynit 32 33 34 35 35 4 4 24 18 28 4 4 22 18 22 28 24 22 24 21 21 27 6 20 21 6 21 28 23 24 24 27 17 28 27 34 3 24 28 18 14 24 28 18 22 22 5 30 30 22 20 2 3 3 3 3 3 5 11 11 :b!~VS1l 11 13 7 10 10 22 23 8 8 11 17 17 14 6 11 7 12 13 11 8 8 13 18 9 9 27 9 14 10 6 11 10 11 10 8 7 20 9 9 4 33 9 10 8 12 13 23 19 11 11 4 21 18 18 20 20 22 23 3 3 861 SUZTACOCBA PHASE RIMS Sba:g~ F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F Bim Va;tiant Qis.m JJ:nit a 11 1 11 1 a a 11 1 11 1 a 14 1 a a 14 1 15 a 1 a 18 1 a 19 1 19 1 a 19 1 a 19 1 a 1 a 19 19 1 a 19 1 a 20 1 a 1 a 20 20 1 a 20 1 a 20 1 a a 20 1 a 21 1 21 1 a 21 1 a 21 1 a 21 1 a 21 1 a 21 1 a 21 1 a 21 1 a a 21 1 22 a 1 22 a 1 a 22 1 1 a 22 24 1 a 26 1 a 1 a 26 a 26 1 27 1 a 27 1 a 27 1 a 27 1 a 27 1 a 29 1 a 1 a 29 1 30 a 1 32 a 1 32 a 32 1 a 1 32 a 33 1 a 1 a 33 33 1 a 1 a 34 1 a 34 1 a 34 1 34 a 1 a 34 34 1 a l:!~:lld 4 5 8 10 12 34 26 9 3 8 8 9 9 9 10 4 5 5 7 9 10 4 5 5 7 8 8 9 9 9 9 11 11 12 12 12 3 7 8 6 11 12 12 13 10 12 10 10 10 10 11 15 15 17 4 7 7 7 7 9 Sbal2~ F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F Birn Va.:iant I2iam J;.Init 35 a 1 35 a 1 35 a 1 36 a 1 36 a 1 36 a 1 37 a 1 a 38 1 a 39 1 7 5 a 1 28 a 8 1 1 a 9 1 16 a 9 1 a 10 5 1 11 a 10 1 26 a 10 1 10 30 a 1 10 31 a 1 32 a 10 1 1 a 11 1 3 a 11 1 4 11 a 1 11 13 a 1 17 11 a 1 25 a 11 1 25 a 11 1 4 a 12 1 12 a 12 1 13 a 12 1 25 a 1.2 1 12 30 a 1 30 12 a 1 4 13 a 1 28 a 13 1 24 14 a 1 30 a 14 1 21 a 15 1 21 15 a 1 34 15 a 1 30 a 16 1 2 b 1 2 b 1 6 b 1 14 b 1 18 b 1 19 b 1 21 b 1 22 b 1 23 b 1 26 b 1 27 b 1 32 b 1 35 b 1 36 b 1 26 b 10 1 28 b 10 1 1 11 b 1 17 11 b 1 22 b 11 1 28 b 11 1 l:!~v~l 11 11 12 8 8 9 3 8 7 24 14 18 40 23 7 8 11 12 13 18 20 19 6 6 1 5 21 8 4 10 11 20 13 12 12 8 8 7 10 21 24 26 34 7 5 8 7 13 7 12 10 10 8 8 12 16 11 8 12 862 SUXTACOCBA PHASE RZMS Sh5!.9~ Bim Vs;t:iS!.nt Qism Ynit F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 "E' b b b b b b b b b 11 12 12 13 13 14 14 15 17 31 6 15 27 27 5 6 26 30 2 3 3 3 3 6 6 11 11 11 15 15 15 15 18 18 19 19 19 19 19 19 19 22 23 24 24 24 26 27 27 27 27 27 27 27 27 29 29 29 29 32 33 36 36 36 36 36 36 36 36 L~v~l 12 24 32 10 10 24 24 12 10 12 15 18 18 18 24 24 3 6 8 4 6 6 9 4 7 7 7 8 8 8 8 11 9 10 7 8 8 9 9 10 10 10 11 11 12 12 9 9 10 11 8 12 7 7 7 7 8 8 8 9 Sh5!.!2~ F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F Bim ~.-isnt Dism Unit 2 2 2 2 2 2 2 2 2 2 2 2 5 7 7 9 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 4 4 4 4 4 4 9 10 10 11 11 11 11 12 12 12 13 13 14 14 15 10 10 12 12 12 13 13 13 14 14 14 14 14 18 18 18 18 a a a a a a 36 37 37 37 37 38 39 39 39 3 15 32 11 36 6 24 4 20 31 36 6 30 36 1 29 13 15 28 2 4 22 22 24 24 24 28 38 22 31 24 24 35 4 4 25 5 5 11 11 28 24 24 36 36 19 L~Y:~l 9 4 4 7 10 8 5 6 6 18 29 10 6 8 25 8 20 5 11 8 24 11 7 18 4 8 30 12 2 23 11 11 5 11 11 12 6 15 14 12 12 9 20 20 23 23 10 10 12 9 9 12 12 9 7 9 9 24 4 4 12 8 22 22 10 13 9 22 863 SUITACOCHA PHASE Rms Shiil:bl~ Bim JlS!.;r;:iiilnt !:liiilm Ynit F F F F F F F F F F F F F F F F F F F F F F F X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X 4 4 4 4 5 5 5 5 5 5 6 6 6 6 6 6 7 7 7 7 8 8 8 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 3 3 3 3 a 14 b b b 10 12 10 15 13 13 13 15 21 12 15 15 9 10 10 11 11 12 12 12 12 12 12 13 14 11 12 18 18 18 20 11 12 14 30 2 31 35 3 29 34 37 30 29 24 35 6 19 19 29 19 27 37 27 31 6 16 19 21 22 22 22 22 22 22 28 28 31 36 4 4 4 4 20 17 28 31 31 31 32 3 28 32 35 3 30 5 30 31 25 17 3 4 14 l!~vd 10 2 14 9 13 9 12 6 11 9 8 11 26 10 10 11 8 11 5 11 13 29 3 4 6 7 8 8 8 10 11 4 8 8 6 17 20 21 17 7 10 11 9 11 11 8 15 10 9 8 20 10 24 11 2 1 7 21 21 42 Sbiil:bl~ Bim V5l;r;:i5lnt Q;i.gm Unit X X X X X X X X X 3 3 4 4 4 5 5 6 7 15 15 20 20 12 16 19 28 28 6 6 6 35 31 13 13 ~:t~v~l 12 12 25 26 27 10 9 1 9 864 COLPAR PHASE R:IMS &haoe BUl 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 B 8 B 8 8 8 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 c c c c c c c c c c c c c c c c c c c c c c c c c c E .:. E E .:. .:. E E E E E E 12 12 12 12 12 12 13 13 14 15 3 3 3 3 9 9 9 12 12 12 12 12 VI.;J;:i.Ut h h h h h h h h h h h h h Dig JZnit LeV!!l 10 12 12 12 16 i i i i i 10 10 12 j j j j 16 16 12 13 14 16 18 19 12 13 14 16 18 12 14 14 16 8 20 21 29 29 32 32 32 12 6 27 27 5 4 37 4 6 32 32 38 6 6 3 24 24 27 29 36 38 38 38 38 13 28 31 7 36 3 22 27 32 2 26 19 19 20 13 5 19 20 22 28 30 30 31 24 29 36 36 36 6 6 7 6 10 6 10 10 10 22 7 7 20 19 6 22 23 8 3 8 21 22 15 10 11 9 8 6 6 6 6 6 8 9 9 4 8 20 8 4 8 20 5 10 7 4 4 20 6 3 8 12 9 9 9 8 7 5 7 8 &ha:ge Rim varigt Diam E E E E E E E E E E E 12 12 12 12 12 12 12 12 14 14 15 15 16 16 17 20 13 13 14 17 11 Jlnit 28 29 17 28 28 29 1 21 20 13 14 ~§vel 10 6 12 9 9 9 19 3 7 9 :..7 865 EMPEDRADA PHASE R:IMS ShS!.I2Sl Bim ~ii!.;~;:iS~.nt I2iS!.rn Ynit l:iSl:ll:Sll 3 19 3 E 3 18 1 E 3 2 21 E 2 3 21 E 3 5 21 E 5 3 38 E 16 3 12 4 E 3 5 13 13 E 3 32 15 15 E 17 3 8 16 E 3 17 2 29 E 14 9 E 1 16 9 1 E 17 9 E 1 18 9 1 E 9 18 1 E 22 9 2 E 15 9 3 E 19 9 3 E 18 9 4 E 20 9 5 E 18 9 6 E 20 9 6 E 21 9 6 E 12 9 15 E 15 9 15 E 20 9 E 15 23 9 15 E 26 15 E 9 8 18 E 9 5 19 E 9 5 E 9 19 2 9 20 E 20 3 E 9 4 9 E 20 6 9 21 E 7 9 21 E 9 1 E 22 9 22 1 E 9 23 5 E 9 5 24 E 1 9 E 26 1 9 26 E 27 3 E 9 27 3 E 9 8 29 E 9 8 E 9 29 9 4 E 34 6 9 34 E 9 8 E 35 9 6 E 36 7 9 E 36 E 9 39 5 7 17 E 9 9 7 9 E 28 9 E 9 10 2 15 9 10 16 E 3 E 9 11 20 5 9 16 E 12 2 9 12 21 5 E Sbii!.I2Sl Bim ~ii!.:dS~.nt I2ism! :unit l:iSlV§l1 5 12 21 E 9 5 32 12 E 9 23 13 6 E 9 3 13 11 E 9 3 13 24 E 9 5 13 26 E 9 7 13 28 E 9 14 19 3 E 9 20 14 4 E 9 10 14 23 E 9 5 14 28 E 9 9 14 28 E 9 22 16 6 E 9 3 16 19 E 9 3 16 19 E 9 5 16 24 E 9 15 4 18 E 9 27 4 18 E 9 27 4 18 E 9 4 27 18 E 9 17 8 20 E 9 7 20 30 E 9 22 6 E 15 8 15 E 15 2 20 E 15 2 20 E 15 2 20 E 15 2 20 E 15 5 35 E 15 7 35 E 15 30 14 6 E 15 23 8 5 E 15 1 25 9 E 15 3 24 10 E 15 18 11 2 E 15 20 11 2 E 15 22 12 2 E 15 17 12 31 E 15 18 13 2 E 15 5 15 13 26 E 14 14 3 E 15 9 14 11 E 15 3 14 18 E 15 8 14 23 E 15 8 14 28 E 15 8 13 E 15 16 15 19 6 E 15 9 19 13 E 15 18 20 4 E 15 20 2 E 16 23 2 E 16 15 3 E 16 10 8 E 16 1 20 E 16 12 22 E 16 7 24 E 16 20 E 16 12 5 21 5 16 12 E 3 13 21 E 16 28 14 E 13 16 866 Shs:b!~ Bim E E E E E E E E E E E E E E E E E E E E E E E E E E E 16 16 16 17 17 17 17 17 17 17 17 17 17 18 18 18 18 18 18 18 18 18 19 19 19 19 19 19 20 20 20 20 21 21 21 21 22 22 22 23 23 24 24 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 E E E E E E E E E E E E E E E E G G G G G G G G G G G G G G G G G ~:r::iant I:lism Ynit 15 17 20 8 12 35 1 20 19 19 19 20 20 34 16 5 32 14 6 15 21 34 25 31 18 1 13 5 23 36 4 17 13 13 10 14 14 14 32 28 28 28 21 21 21 17 6 12 6 6 6 19 19 4 7 13 15 18 18 22 22 22 23 23 23 24 24 24 24 26 15 16 16 18 11 13 14 17 23 30 13 20 I.!~vd 2 14 1 5 7 16 5 5 4 45 20 6 21 22 22 4 6 6 8 4 19 12 17 3 3 18 10 5 6 7 8 8 2 2 5 5 22 5 22 21 21 4 4 18 4 4 25 4 5 3 3 3 4 4 7 3 6 6 6 3 ~hii!::b!~ BMPEDRADA PHASE RIMS Rim Vs;r;:iii!:nt Digm Ynit L~v~l G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 27 27 31 9 10 10 10 10 11 12 12 13 14 15 16 17 18 18 19 20 21 22 24 24 11 12 12 31 32 32 33 34 35 36 36 36 37 38 38 38 39 39 6 30 30 31 32 31 28 31 35 31 22 24 28 17 34 18 36 4 5 18 18 3 6 6 20 22 22 23 24 26 27 29 32 35 35 36 37 38 39 4 27 31 3 6 7 9 7 13 12 3 6 6 7 7 5 5 5 5 3 4 17 4 6 6 7 4 6 3 5 2 4 4 7 6 3 6 8 18 18 6 8 20 22 29 4 1 7 10 6 3 B 9 6 7 7 3 5 5 3 17 7 6 867 EMPEDRADA PHASE RDIS Shiil:b!~ G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G Bim :sliil:tiiilDt l:liiilm Unit 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 4 4 4 4 4 4 4 4 4 4 4 4 5 5 5 5 5 13 13 14 14 14 15 15 15 15 16 16 17 17 17 17 17 18 18 19 19 22 23 12 13 14 15 16 20 21 11 11 13 13 13 14 16 20 22 24 15 26 29 17 27 27 35 27 27 2 16 24 27 36 20 30 1 16 7 14 19 19 19 19 19 19 19 22 27 27 27 29 29 34 23 12 19 19 30 23 22 19 29 29 38 5 24 28 28 31 35 25 18 2 8 8 8 20 ~~:ll~l 6 6 3 2 2 12 8 10 8 7 7 14 4 4 10 7 7 5 20 39 4 14 4 4 4 4 4 4 4 5 5 6 6 3 6 4 9 4 8 7 7 7 6 4 6 6 1 21 4 4 7 8 6 5 16 3 4 5 7 Shiil:b!~ Bim Va:tiiilDt I2iiilm Ilnit G G G G G G G G G G G G G G G 5 5 5 6 6 6 6 7 7 7 8 8 9 10 11 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 3 4 H H H H H H H H H H H H H H H H H H H H H H 15 15 20 12 11 19 7 12 10 13 14 14 15 16 25 11 10 6 11 7 3 21 21 39 18 34 3 34 15 4 11 13 1 18 22 23 24 29 29 34 34 34 34 37 37 36 28 27 28 31 28 1 5 1 I.!~v~l 20 4 4 14 4 4 3 4 4 21 4 10 20 5 4 16 6 9 10 6 6 18 4 4 4 6 4 4 7 7 7 9 8 6 13 19 14 APPENDIX F REPORT ON BOTANICAL REMAINS 868 869 ANALYSIS OF CHARRED BOTANICAL REMAINS FROM MANACHAQUI CAVE, PERU by Deborah M. Pearsall Universicy of Missouri 10-95 INTRODUCTION This reporc presencs che results of analysis of charred botanical materials recovered from Manachaqui Cave by Warren Church during excavations in 1988 and 1990. of three chambers. The cave is a boulder cluster comprised It is located at 3,625 m elevation near the eastern edge of the subalpine grasslands of the northeastern Peruvian Andes between the Maranon and Huallaga rivers, on the crest of the divide (Church 1993). The study area corresponds to the districts of Pataz (La Libertad Department) and Huicungo (San Martin Department), and is located jusc to che west of che Rio Abiseo National Park. The cave is in the paramo (wee Andean grassland) vegetation formation; montane forest is located nearby and may have been closer to the cave in the past, although the area is thought to have been paramo during the period the cave was used. Manachaqui Cave was utilized at various times from the Late Preceramic (2200 (A.D. 1532). b.c., uncalibrated) to the end of the Late Horizon Table 1 summarizes the C-14-based chronology and assignment of excavation levels to cultural phases and periods. hiatuses may exist in the occupation, between 500 (beginning of Early Intermediate), a.d. 200 a.d. 700 and Three b.c. and 200 b.c. and a.d. 400, and between A.D. 1470 (beginning of Late Horizon). Given the location of the shelter on the Maranon-Huallaga divide, the fact that a prehistoric road passes by it, and the modern function of the cave as a shelter for travelers, Church has proposed that Manachaqui Cave functioned as a wayside station during prehistory, rather than exclusively as a ritual site or hunting or herding camp. I 870 will therefore examine the charred botanical materials recovered from the site in light of this hypothesis, to determine if data support this function for some or all of the cultural periods represented at the site. other Further, the date of the appearance of corn, and the presence of ~exotic" (i.e., non-paramo) resources, may shed light on regional interactions during the periods the cave was used. I am aware of no ethnoarchaeological or experimental data that would help in framing botanical correlates of the hypothesis or alternative explanations of cave function (most work of this type has focused on crop processing, e.g., Hillman (1984). However, discussing the ways in which botanical materials become deposited in a site such as Manachaqui may help in setting up a framework to test the hypothesis. The first point co emphasize is that only those botanical materials that became charred survived in this environment. This included plants used as fuels for cooking or heating fires, edible plant parts lost during cooking, inedible plant parts burned as refuse, and any plant material, food or nonfood, accidentally burned (bedding catching fire, example) . for Any food items not cooked in the shelter are very unlikely to become part of the deposit. This would include "trail foods• prepared elsewhere and consumed in the cave, as well as any foods consumed raw. This may limit the utility of the macroremain data to address the hypothesis of cave function. Phytolith analysis, which can identify plants that decay without leaving charred remains (Pearsall 1989; Pearsall and Piperno 1993), would be an interesting addition to the present study. As detailed in Pearsall (1988), there are a variety of avenues through which botanical materials become charred and deposited in sites: (1) wood may be collected and burned as fuel or used in construction or tool manufacture and later burned; (2) seeds may be gathered for food and accidentally charred during cooking, brought in dung burned as fuel, brought in incidentally with root or stem foods and discarded as the waste portion, or brought in with plants for nonfood purposes and 871 discarded as the waste portion; and (3) roots may be roasted as foods or included in sod burned as fuel. Peru, At Panaulauca cave in the Junin puna of for example. the source of many of the small seeds recovered in the deposits was camellid dung fuel. (< 2.0 mm) Small. medium, and large grass seeds. Cyperaceae, Sisyrinchium, and Calandrinia seeds were observed in dung and therefore interpreted as nonfood items at this site (Pearsall 1988). It is thus important to look not only at what is likely to become charred. but at the source(s) of those materials, since not all "edible" remains may be foods. It is not always possible to determine whether charred plant remains are foods, or should be considered to represent some other function. Such is the case with Festuca, or ichu grass. This useful plant may have been used for bedding or for fire starter at the cave. However, ichu also produces large grains that may have been an attractive local food source. Both interpretations will be presented below. Finally, it is important to realize that not all charred materials deposited in a site will survive or be recovered during excavation. Fragile root remains, for example, are likely to be underrepresented relative to more robust corn cob fragments or woody fruit rinds. Any food with inedible parts likely to be disposed of in the fire is likely to be better represented than foods wholly consumed. Smaller materials, like the seeds of many annual plants, may be underrepresented due to the size of screen or flotation technique used in recovering remains. Given these considerations. the environment of the site, and the time periods in question, the following ethnobotanical indicators might be helpful for distinguishing use of the cave as a wayside station from use as a residential camp for hunters/gatherers/herders: Use of camelid dung fuel source): (and presence of small seeds from this more abundant at a camp; less at a wayside shelter (sheltered pack animals would produce little dung; a concentrated quantities). ~anaged herd much larger, This indicator cannot be used in the 872 Manachaqui case, since few small seeds were recovered (see below) . local food items: more use at a permanent or seasonal camp; less use at a shelter. "exotic" food items: might be present in either case, but would be more abundant at a wayside shelter, less so at a habitation site. abundance of food remains: more abundant relative to amount of fuel burned at a habitation site; less abundant at a wayside shelter--i.e., difference in frequency of cooking fires (more accidental charring of foods) versus fires for warmth (less accidental food charring) . In the case of corn, evidence for cooking would be indicated by presence of kernels {lost during parching, for example), rather than just the presence of the inedible cob, which could become deposited either after in-situ food preparation or consumption of a "traveler's meal" of corn still on the cob. . diversity of botanical remains: low diversity at a wayside shelter, higher diversity at an occupation site (i.e., following Yellen 1977, specialized versus generalized activities; illustrated in Pearsall 1983 for Pachamachay Cave). This measure cannot be used in the Manachaqui case due to the small quantities of smaller seeds recovered. Richness (number of taxa present) will be discussed briefly. One would expect local wood species to dominate in both situations, since it is unlikely travelers would carry fuel. EXCAVATION AND FIELD BOTANICAL PROCESSING PROCEDURES Manachaqui cave is actually three small "shelters• created by a cluster of boulders. Shelter M-lA, the largest and most intensely used, is the source of the samples discussed here. was excavated, Approximately 35% of M-lA to a maximum depth of 2 m (Church 1993). Excavations proceeded in 5 em levels, in 39 1 x 1 m units placed in the interior (Sector A) and berm (Sector 8) of the shelter. Soil for botanical recovery was taken from Sector A's floor deposits in Unit 15 , and from Sector A hearths in units 14, 15, 16 and 17 , and Sector B's Unit 31, 873 for a total of 83 samples. To reduce sample weight and volume for transport from the site, larger soil samples were wet-screened in the field using 1/16" mesh, and sherds, lithics, and so on were removed in the Trujillo laboratory. were recorded. Weight and volume of the "reducedw samples Residues were later chemically floated using zinc chloride to recover botanical rnacerials, which were sent to che University of Missouri Paleoechnobotany laboratory for analysis. LABORATORY PROCEDURES Because budget constraints made it impossible to analyze all 83 samples, the following sampling strategy was employed (refer to Table 1): all samples (floors and hearths) from the preceramic Lavasen and Initial period Manachaqui levels were sorted (14 precerarnic, 4 Manachaqui), all samples from the Early Horizon Suitacocha period were sorted (4), one sample from each of the Early Intermediate Colpar and Empedrada floors was sorted, with the exception that 2 from W, 2 from N, and none from 0 were done (18 total), and two samples were selected from among the Late Horizon Poblano floors (F, B). The sample total was 42. The large quantity of charred wood in some of the individual samples also necessitated subsampling: 13 samples were split using a riffle sampler (Pearsall 1989), and 25% or 50% of the sample analyzed. Sample sorting followed standard paleoethnobotanical procedures (Pearsall 1989). Briefly, large samples were subsampled using the riffle sampler, and the fraction to be analyzed passed through a 2.0 mm screen. The > 2.0 rnm fraction was sorted in its entirety, typically into wood, root material, unidentifiable fragments. fruit rind, large seeds, corn, and All materials were counted and weighed. < 2.0 rnm fraction was then examined for seeds and corn fragments. The Since larger samples were wet screened using 1/16" mesh prior to chemical flotation, relatively few small seeds were recovered. Identifications of seeds, fruic fragments, and roots were made by comparing unknowns to materials in che MU Paleoethnobotany comparative collection, standard 874 reference books, and herbarium specimens curated at the Missouri Botanical Garden, St. Louis. Some remains are still unidentified, however, some because the materials are highly fragmented; a few, tallied as unknown types in the raw data table because they did not match available comparative collections or published sources. Corn remains were callied as kernels, cupules (single), or rachis segments (groups of joined cupules) . The maize remains will be studied at a later date to determine the race or races present. Wood was not identified. Raw data were selected and combined in various ways to address the main hypothesis of the study, and to observe any patterning in occurrence of food plants over time. For floors in which multiple samples were analyzed, I summed the data over all samples. Any sample not sorted in its encirety was extrapolated to 100% before being summed with other samples from that floor. A series of ratios designed to address the issue of cave function were then calculated (Figs. 1-6; see bottom of Table 2 for values): local:exotic+local foods;'corn kernel:corn cob fragments (corn processing question); and total food:wood (abundance of food items) . As mentioned above, there was some question as to how Festuca, or ~ grass, should be used in the ratios: as food or not. As detailed in Table 2, "total food, 2• represents the total omitting Festuca; "total food• contains it. Further, it was unclear whether unidentified fruit hull fragments should be considered as "localw or "exoticw foods (they are considered food remains). Ratios using "local foodsw include these remains; those using "local foods, 2" do not include them. \ DISCUSSION OF RESULTS Food Plants Recovered Food plants recovered from the excavation are summarized in Table 2. There are probably other foods represented among the unknown seeds, but none of the unknowns are common (most are single occurrences), and therefore not likely co change the results significantly. 875 Maize, Phasolus beans, and thin epidermal fragments of the pit of a fruit in the Sapotaceae are considered "exotic• food resources. cave is located above the cultivation zone for maize and beans. The Trees in the Sapotaceae {genera such as Pouteria, formerly Lucyma, Manilkara; Bumelia} would occur wild {or tended, in the case of cultivated taxa} in the moist. lower elevation forest to the east of the study region. The few fragrnencs of the Sapotaceae fruit pic type thac had the hilum (atcachmenc areal presenc permicced idencificacion to family, but are too fragmencary co idencify chese conclusively as lucuma (Pouteria spp.}. A fruit similar co lucuma is indicated, however. Local foods, those present in the paramo grassland formation, are Lupinus, Ribes, Rubus, Festuca, Cheno/Am, Polyqonum/Rumex, fruit rinds {omitted in "local foods, 2•), and tuber/root fragments. Most of these occurred at Panaulauca and Pachamachay caves in the Junin pyng of central Peru (refer to Pearsall 1980, 1988 for descriptions of habitat and use of these taxa} . Briefly, ~ grass, Festuca, produces large seeds that are a potential food resource. This grass is not a preferred forage of camellids {smaller grass seeds recovered, likely Calamaqrostis and ~. are not considered foods in this analysis, since camellids do favor these} and thus less likely to be in dung. The presence of Festuca seeds in the cave could be the result of gathering for food, or inclusion by che burning of ichu grass as tinder or through accident. Lupinus seeds (lupine). though not common, are potential food resources. These seeds are not large enough to be identified as the cultivated species. Small, clover-sized legume seeds are not tallied as food plants {I have observed these in camellid dung samples from Junin}. ~ and~ seeds may represent the eating of these small berries. Both planes are small shrubs. Seeds tallied as Cheno/Am were too distorted to identify precisely to genus, but are either Chenopodium or Affiaranthus. These are annual plants that produce large quantities of edible seeds (and also greens}, and often favor disturbed habitats. Seeds are not of che domescicated forms. Polygonaceae seeds {either 876 ~ or Polygonum) are also included in the local foods group, since representatives of these genera occur in high elevation grasslands, and are considered edible by some cultures. The categories fruic rinds and tuber/root fragments require some additional explanation. Although many of the plant remains recovered from the cave were highly fragmented, it was possible to distinguish fragments that appeared to be root or tuber tissue. Such tissue is very porous in texture, and lacks the internal morphology associated with wood or other larger plant remains like bean or corn fragments. few rare cases, an "eye" (node) was preserved. In a None of these had the clear shape of potato eyes, however; a few looked like sedge (Cyperus or Scirpus) roots. Given the abundant wild and domesticated root and tuber foods found in the higher elevations of the Andes, it seemed best to include these fragmented remains in the list of local foods. case of fruit fragments, fragments this is a residual category: (thickness variedl that were~ In the all curved tissue Sapotaceae were tallied here. None of these are squash; none have diagnostic features. I included these in local foods because they did not look like any of the tropical forest tree fruits in my collection; ratios calculated with "local foods, 2" omit these remains, for comparison. Patterning through Time The three "exotic" taxa, corn, bean, and Sapotaceae fruit, appear in the record at Manachaqui cave at different points in the sequence. Sapotaceae fruit pit fragments appear earlier, in a hearth sample (#80) and a floor sample (#64) from floor FF, the lowest level analyzed, dated to 1500 b.c. Thus one food from outside the resource area of the site was present from the earliest occupation. throughout the sequence. The fruits continue to occur The largest quantity (29 fragments), occurs in sample #58, one of the samples from the CC floor, preceramic. in the later Corn and bean appear together in floor Y (sample #50), the earliest level of the Suitacocha phase, 800 b.c .. There are no 877 Preceramic or Initial period macroremains of corn or bean at the site. Corn also occurs in floor X, the upper Suitacocha floor; bean does not. Corn is consistently and often abundantly present throughout the rest of the sequence; bean occurs only in floors Y, M, J, I, and B (1-3 fragments per occurrence) . There are few foods. striki~g patterns among the occurrence of local Festuca seeds are present in high concentrations in several samples: 172 in #78 !hearth in CC); 55 in #59 (floor CCI; 1566 in #77 (hearth in 88); 56 in <floor 81. ~41 !floor Tl; 37 in #39 (floor Sl; 67 in #1 Otherwise, 1-10 ichu seeds occur in most samples. The other seven local plants/plant categories are present in low numbers sporadically throughout the sequence. in numbers of taxa present: Samples are not especially rich one sample has 5 local taxa, 8 samples have 4 taxa; the rest have 0-3 taxa present (Table 2). The Late Preceramic levels appear to be less rich than the rest of the sequence. Looking at abundance of food remains in comparison to quantity of wood charcoal is one way to determine how commonly food was cooked at the site. If fires were used primarily for warmth, and little food was cooked, one would expect fewer "cooking accidents• and disposal of inedible food residues than if fires were used for cooking on a daily basis. Fig. 1 illustrates the ratio of total food remains to wood charcoal. as fuel. and S. I am assuming that wood was brought to the site only for use Four levels stand out with high food:wood ratios: BB, Z, T, In the case of 88, a 1566-seed concentration of Festuca seeds pushes the food:wood ratio up; in floor Z a very small amount of wood is present (0.836 g), which has the same effect. The high ratios in floors T and S are caused in part by low wood abundance (1.23, 2.7 g), but food remains are also more abundant in these samples than in many examined. Floors T and S, and perhaps 8B, represent a depositional pattern consistent with increased cooking (relative to the other levels). Since it is not possible to examine these data in terms of amount of charred material in a standard volume of matrix, differences among levels in the 878 amount of burning activity are not clearly delineated. Higher amounts of charcoal can also be an indicator of more intensive use of a site. Looking at abundance of foods using "Total food,2:wood" (i.e., omitting Festuca as a food) food ratios: (Fig. 21. three levels stand out as having higher Z, T, and S. The only change by omitting Festuca is thus the omission of 88 as a level with a high food ratio. Examining the data in terms of the relative abundance of local and "exotic" foods highlights the impact of the introduction of corn into the regional economic system (Fig. 3). Corn appears at the site for the first time in floor Y, and this results in a dramatic drop in the ratio of local to exotic foods. Although Sapotaceae fruit fragments are present in the earlier levels, the Preceramic and Initial period assemblages are dominated by local food items. The pattern of local use reverses with the introduction of corn, and reappears at only two points in the later part of the sequence: floors T and S (early part of the Early Intermediate) and floors F and B (the two Incaic levels analyzed). These two periods may represent a change in use of the cave relative to the rest of the "post-maize" sequence. There are a number of changes in the ratio of local to exotic foods when "Local foods, 2" (i.e., omitting Festuca and fruit rinds) is used in the calculation (Fig. 4). The earlier levels of the cave are much less even (ie, a less consistent pattern of high local foods) when Festuca and fruit rinds are removed. Sample size falls considerably, which may contribute to the unevenness. of exotics--a new pattern. Levels EE and CC show dominance After corn comes in, exotic foods dominate, as before, and there are no pronounced reversals of this pattern, as appeared using the other sum. This weakens the argument for periodic changes in cave use (see below). Looking for the relative abundance of maize cob (inedible) and kernel (edible) remains (Fig. Sl may shed additional light on whether corn use changed during the period Manachaqui Cave was in use. Cob fragments are generally more abundant than kernel fragments at the site 879 (i.e., in most levels the ratio is >1). This shows that corn was not brought to the cave exclusively as shelled grain, and that cobs, perhaps with a few kernels still attached, were disposed of in the fire. Floor S stands out with a very high cob:kernel ratio; only 2 kernel fragments were recovered in comparison to 26 cob fragments. Levels W and F are the exceptions to the dominance of cob material; in these levels kernels are three times as common as cob fragments (Fig. 6). Little corn is present in level F, but W has abundant remains, dominated by kernels. Since it is unlikely kernels were burned deliberately, on site cooking or parching of kernels (and accidental burr.ing) seems indicated. CONCLUSXONS Considering all the charred botanical data, and remembering that results are impacted by loss of some smaller remains from the assemblage, the results of this study generally support the hypothesis that Manachaqui Cave was a camp for travelers for much of its history. Food remains are not very abundant relative to charred wood (evidence that fires were used more for warmth than for cooking); uexotic" food stuffs are present throughout the sequence (evidence that foods were carried in); once corn was introduced (Early Horizon), use of local foods drops off (evidence that local foods were relatively unimportant); corn kernels are generally less abundant than cob remains (evidence that parching/cooking of corn was uncommon). Sincd some cob remains are always present, it is clear that travelers were not just carrying shelled corn. Within this overall pattern, however, there are some interesting exceptions that suggest the cave may have changed functions for brief periods. The floors in question are BB and Z (Manachaqui phase); W, T, and S (early part of Ernpedrada phase), and F and B (Poblano phase). The Manachaqui floors show one indicator of more intensive use of the cave: higher food:wood ratios, in one case (Z) based on a small sample. are multiple indicators that ~he There cave was more intensely used during the 880 early part of the Empedrada phase: a high ratio of kernels to cobs (floor W); high food co wood ratios (floors T and S); high local to exotic food ratios !T and Sl. These indicators suggest the cave functioned as a habitation site during this period. continued to be available. however. "Exotic" foods The two Poblano phase samples analyzed (F and Bl, presenc a slightly different pattern: there is increased use of local foods, but no overall increase in foods relative to charred wood. Perhaps the cave had a more specialized function during the Late Horizon, different from a traveler's stop, but not as intensive as a habitation site. As mentioned above, the pattern involving local:exotic indicators weakens if Festuca and fruit remains are removed as local foods. This study of charred botanical remains from Manachaqui Cave illustrates how ethnobotanical data can be used to test hypotheses of site function. The shifts in cave function suggested by these data can be tested through analysis of other macroremain samples, comparisons to artifact and faunal data, detailed study of corn remains, and future research, perhaps including phytolith sampling. LITERATURE CITED Church, W. 1993 Evidence for prehistoric highland-tropical forest interaction from Manachaqui Cave in the northeastern highlands of Peru: Preliminary report. Paper presented at the Society for American Archaeology meetings, St. Louis. Hillman, G. C. 1984 Interpretation of archaeological plant remains: The application of ethnographic models from Turkey. In Plants and Ancient Man: Studies in Palaeoethnobotany, edited by W. V. Zeist and Ttl. Casparie, pp. 1-41. A.A. Balkerna'Hubbard, R. N. L. B., Rotterdam. Pearsall. D. M. 1980 Pachamachay ethnobotanical report: Plant utilization at a hunting base camp. In Prehistoric Hunters of the High Andes, J. W. Rick, pp. 191-231. Academic Press, New York. 881 1983 Evaluating the stability of subsistence strategies by use of paleoethnobotanical data. Journal of Ethnobiology 3(2):121-137. 1988 Interpreting the meaning of macroremain abundance: The impact of source and context. In Current Paleoethnobotany. Analytical Methods and Cultural Interpretations of Archaeological Plant Remains, edited by C. A. Hastorf and V. s. Popper, pp. 97-118. University of Chicago Press, Chicago. 1989 Paleoethnobotany. San Diego. A Handbook of Procedures. Academic Press, and D. R. Piperno (editors) 1993 Current Research in Phytolith Analysis: Applications in Archaeology and Paleoecology. MASCA, University of Pennsylvania Museum., Philadelphia. Yellen, J. E. 1977 Archaeological Approaches to the Present: Models for Reconstructing the Past. Academic Press, New York. 882 Floors/Samples Period ~ Date Cuncalibrated l Late Horizon Poblano AD 1532 A B C-E F 1 AD 1470 1 Hiatus Middle Horizon Early Inter. G H I J K L M N 0 p T 1 1 1 1 1 u v w 1 1 2 Q R s Empedrada AD 700 1 1 1 1 1 1 1 2 AD 200/400 Col par AD 200/400 200 BC Chavin horizon Hiatus? Early Horizon X y Suitacocha 2 2 BOO BC Initial Period Manachaqui z AA 1 3 cc DD EE FF 3 3 2 3 3 BOO BC 1500 BC Late Preceram. Lavas en BB 400 BC 1500 BC 2200 BC Table 1. Manachaqui Cave chronology and samples analyzed. 79 63 62 80 65 64 LAB NUMBER MIA.15.66 M1A.15.66 M1A.15.66 M1A.16/17.63 MIA.15.63 MIA.15.63 PROVENIENCE FF-65 Total FF EE-79-H EE-63 EE-62 FF-64 FLOOR FF·80-H Total EE 3 1 2 1 4 2 FS# 0.625 0.25 1.25 2.65 LITERS 0.3 0.475 222.75 17.25 TOTAL SAMPLE WGHT 60.22 214.56 11.68 18.66 %EXAMINED 25 100 100 100 25 100 100 100 SUB-SAMPLE WEIGHT 51.71 53.33 WOOD WEIGHT 26.56 0.775 24.91 131.93 19.26 4.14 6.64 87.82 Maize kernels, total count Maize cob fra_gs, total count Phaseolus, total count cf. Sapotaceae fruit 1 2 6 7 2 1 13 3 2 2 14 Fruit rinds 3 11 Tuber/root frags 3 12 3 3 ~lnus (Legumlnosae} cf. Ribas (Saxifragaceae) Rubus (Rosaceae) Festuca (Poaceae) 1 2 12 1 3 1 Cheno/Am Polygonum/Rumex (Polygonac.) 1 4 4 Total local foods, 1 1 2 26 31 6 7 4 0 0 Total local foods, 2 4 0 0 16 3 3 0 Total exotic foods 1 2 6 7 13 2 1 Fruit rinds 3 14 2 3 2 11 8 37 Total foods, 1 1 4 11 5 39 8 0 8 4 36 10 Total foods 2 3 27 5 Locai:Exotlc+Local, 1 0.84 0.67 Total food:wood, 1 0.28 0.44 0.73 Locai:Exotlc+local, 2 0.19 Total food:wood 2 0.27 0.31 kernel:cob cob:kernel Table 2. Summary of botanical data. (X) 00 w 1LABNUMBER IPROVENIENCE I IA.OOR FS# UTERS TOTAL SAMPLE WGHT %EXAMINED SUB-SAMPLE WEIGHT WOOD WEIGHT Maize kernels, total count Maize cob frags, total count Phaseolus, total count cf. Sapotaceae fruit Fruit rinds Tuber/root frags Luoinus (Leguminosae) cf. Ribas (Saxlfragaceae) Rubus (Rosaceae) Festuca (Poaceae) Cheno/Am 58 60 78 59 77 57 61 M1A.17.57 MIA.15.57 MIA.15.57 M1A.15.54 MIA.15.54 MIA.15.60 MIA.15.60 CC-58 DD·60 Total DD CC-78 CC-59 Total CC BB·77·H 88·57 DD·61 2 1 2 5 2 1 0.65 2 0.19 0.45 2.225 0.25 0.225 206.18 7.64 42.5 6.3 61.77 199.15 13.13 50 100 50 100 100 100 50 100 100 100.59 20 98.58 8.472 56.633 1.568 128 22.85 2.007 18.951 1.3 13.17 2 2 8 7 3 3 16 8 14 4 29 5 2 Table 2 (continued) . 1 4 3 1 4 1 6 172 1 55 14 8 2 2 1G 12 7 3 7 3 14 13 28 14 16 8 44 38 0.64 2.32 0.47 2.01 175 3 4 60 0 0 5 60 5 Pol~gonum/Rumex {Pol~gonac.) Total local foods, 1 ]'otal local foods, 2 Total exotic foods -------· Fruit rinds Total foods, 1 Total foods, 2 !:Q.c::.ai:Exotic+'=.Q.c__al, 1 Total food:wood, 1 Locai:Exotlc+local, 2 Total food:wood, 2 kernel:cob cob:kernel 37 5 4 ---- 179 7 5 1 6 1 29 35 30 404 2 1 416 7 37 5 453 49 0.92 3.54 0.16 0.38 1566 1573 3 1 4 1574 8 1 1 0 1 - - -1 - - (X) (X) tJ:>. LAB NUMBER PROVENIENCE FlOOR FS If UTEAS TOTAL SAMPLE WGHT %EXAMINED SUB-SAMPLE WEIGHT WOOD WEIGHT Maize kernels, total count Maize cob frags, total count Phaseolus, total count cl. Sapotaceae fruit Bull rinds Tuber/root frags Lupinus (Leguminosae) cl. Ribas (Saxllragacealll_ Rubus_ (Rosaceae) Festuca (Poaceae) Cheno/Am Polygonum/Rumex (Poly_gonac.' Total local foods, 1 Total local foods, 2 Tot~xotlc food_s Fruit rinds Total foods, 1 Total foods 2 Locai:Exolic+Local, 1 Total food:wood, 1 Locai:Exolic+local, 2 Total food:wood, 2 kernel:cob cob:kernel 76 56 55 54 53 51 MIA.15.43 MIA.17.43 M1A.15.43 MIA.15.54 M1A.15.41 M1A.15.38 AA-76-H AA-55 AA-54 88-56 Total 88 Y-51 Total AA 2 2 3 1 1 2 0.525 0.25 0.085 0.85 0.2 0.12 44.15 5.35 87.43 40 7.13 12.69 100 100 100 100 100 100 100 100 z 16.62 63.62 2 11 1 4 19 7 1 1 1 3132 2 2 15 4 2 11 17 17 - Table 2 (continued). 1.16 3162 11 4 1s 3166 34 1.00 49.76 0.73 0.53 - ---- - L___. _ _ _ 41.64 10.95 53.75 0.836 2 8 10 5 10 4 8 10 1 1 1 10 1 11 1 2 1 13 2 8 10 21 20 2 28 7 8 10 36 25 0.78 0.67 0.47 0.47 11 0 0 10 11 10 1.00 13.16 0.00 11.96 4 1 0 1 4 ___2_ _ 1 1 1 0 14 4 0 1 1 14 4 -- ---- - -- 00 00 U1 50 49 48 47 46 45 LAB NUMBER M1A.15.30 M1A.15.30 M1A.15.35 M1A.15.35 PROVENIENCE M1A.15.38 M1A.15.28 Total X W-47 W-46 X-49 X-48 Total W Y-50 Total Y v FlOOR 3 2 3 3 4 2 FS# 0.37 0.65 0.175 0.16 0.425 0.2 UTERS 58.66 5.22 22.41 10.35 40.41 TOTAL SAMPLE WGHT 8.31 50 100 100 100 100 100 100 %EXAMINED 100 100 SUB-SAMPLE WEIGHT 30.45 11 .635 13.64 0.83 6.515 7.345 16.413 19.745 WOOD WEIGHT 3.332 2.12 26 Maize kernels, total count 13 11 11 46 46 1 24 18 18 15 Maize cob !rags, total count 12 15 4 1 2 Phaseolus, total count 6 3 6 ct. Sapotaceae fruit 6 5 1 2 1 1 Fruit rinds 1 2 2 Tuber/root fraos 1 1 Lupinus Cleouminosae) cf. Ribes (Saxlfragaceae) 1 2 Rubus (Rosaceae) 1 6 3 4 8 2 1 Festuca (Poaceae) 3 2 1 1 1 1 Cheno/Am !:Q!Y.gonum/Rumex {Polygonac.) 4 16 2 9 Total local foods, 1 6 2 2 11 2 3 1 1 0 1 0 4 4 !_o_!a.!_l~cal food~,_2 1 _ 2_ _ Total exotic foods 29 58 0 29 29 0 67 67 0 2 5 Fruit rinds 1 1 1 35 74 2 31 33 2 76 78 Total foods, 1 7 32 _6p_ 0 30 0 72 72 I.l?tal fooqs, 2 7 i--30 Locai:Exotic+Local, 1 0.22 0.12 0.14 Q~~ 4.49 Total food:wood 1 5.43 3.95 3.30 0.05 0.03 0.06 Locai:Exolic+local, 2 0.17 4.08 4.84 Total food:wood, 2 3.65 3.30 1.08 0.61 3.07 0.25 kernel:cob 0.92 1.64 0.33 cob:kernel 4.00 i Table 2 (continued) . 00 00 0'1 43 41 39 36 34 32 31 26 LA.BNUMBER M1A.15.26 M1A.15.25 M1A.15.24 M1A.15.23 M1A.15.22 M1A.15.21 MIA.15.18 MIA.15.18 PROVENIENCE p N-26 N-31 u T R FLOOR 0 2 3 2 3 3 2 3 3 FS# I 0.2 0.15 0.2 0.37 0.43 0.35 0.35 0.75 LITERS 5.19 8.91 4.94 33.19 39 38.7 18.3 90 TOTAL SAMPLE WGHT I 100 100 100 100 100 100 100 100 %EXAMINED I SUB-SAMPLE WEIGHT 1.23 2.7 13.52 17.06 18.46 1.02 2.25 42.6 WOOD WEIGHT I 7 2 5 19 16 Maize kernels, total count 41 26 14 19 35 18 Maize cob frags, total count 165 Phaseolus, total count 4 1 1 cf. Sapotaceae fruit 8 2 1 3 7 Fruit rinds 3 2 Tuber/root frags 1 Luplnus (Legumlnosae) 1 cf. Ribes (Saxlfragaceae) Rubus (Rosaceae) 3 2 56 1 5 3 37 Festuca (Poaceae) 1 1 Cheno/Am 1 1 ~gonum/Rumex (Poly_gonac.) 3 60 15 0 37 6 7 Total local foods, 1 11 0 3 0 2 2 2 0 2 IotaiJgE-!!1 foods, 2 0 21 28 28 54 35 206 Total exotic foods 1 Fruit rinds 1 3 8 2 7 81 4 0 65 34 69 42 217 Total foods, 1 0 25 28 64 215 Total foods, 2 33 39 1 1-·--0.00 0.74 0.57 0.22 0.17 Locai:Exotic+Local, 1 __9.18 Total food:wood, 1 0.00 65.85 24.07 2.51 4.04 2.28 0.00 0.13 0.00 0.07 0.04 0.05 Locai:Exotlc+local, 2 0.00 20.33 Total food:wood, 2 10.37 2.44 3.75 2.11 kernel:cob 0.00 0.50 0.08 0.26 0.54 0.89 0.00 2.00 13.00 3.80 1.84 cob:kJ~mel -- 1.13 s I Table 2 (continued) . (X) 00 -1 20 18 16 14 12 9 24 11 LA.BNUMBER MIA.15.17 MIA.15.15 MIA.15.14 MIA.15.13 MIA.15.12 MIA.15.11 MIA.15.10 MIA.15.9 PROVENIB'JCE K J I Total N M L H G F FLOOR 1 2 3 3 3 3 3 3 FS# 0.575 0.5 0.5 0.5 0.4 1 0.75 1.025 LITERS 76.74 73.9 170.4 35.43 24.35 132.25 74 158.7 TOTAL SAMPLE WGHT 50 50 50 50 25 25 100 100 100 %EXAMINED 35.7 39 37.86 41.4 39.52 65 SUB-SAMPLE WEIGHT 20.728 17.719 44.85 32.43 18.55 24.385 22.79 11.545 13.07 WOOD WEIGHT 19 19 34 27 41 27 4 16 6 Maize kernels 1 total count 62 2 165 76 16 33 36 40 15 Maize cob rraos, total count 3 1 1 Phaseolus, total count 1 1 1 1 cr. Sapotaceae fruit 2 2 1 3 2 7 1 Fruit rinds 7 3 1 1 1 Tuber/root rrags 1 Luplnus (Leoumlnosae) 1 1 cr. Rlbes (Saxlrraoaceae) 2 1 1 Rubus (Rosaceae) 6 1 4 1 4 4 Festuca (Poaceae) 5 Cheno/Am 1 1 1 fQ!Y.gonum/Rumex (Polygonac.' 6 8 5 12 3 9 8 1 Total local foods, 1 14 I 2 1 2 2 10 2 3 1 0 Total local roods, 2 105 82 55 8 207 21 71 67 31 Total exotic foods 2 1 2 1 3 2 Fruit rinds 7 90 60 221 117 24 79 75 32 14 I Total foods, 1 11 7 24 87 59 75 69 31 _10_ _1 216 ToJ.?~_ods, 2 ---0.10 0.13 0.10 0.08 0.10 0.11 0.03 0.43 0.06 Locai:Exolic+Local, 1 ---·-1 3.39 3.24 3.29 4.93 3.61 1.29 4.34 2.77 Total food:wood, 1 1.07 I 0.09 0.04 0.02 0.03 0.01 0.00 0.20 0.01 0.09 Locai:Exolic+local, 2 I 1.29 4.20 4.82 3.33 3.08 3.03 2.69 0.77 Total food:wood, 2 3.61 0.36 0.25 0.31 0.58 0.94 0.68 1.07 3.00 0.25 kernel:cob 4.00 3.26_ 0.94 4.02 2.81 cob:kernel -0.33 _I 1_.7_~- __1_.06 ----- .. _1.48_ I Table 2 (continued). 00 00 (X) LAB NUMBER PROVENIB'JCE 1 MIA.15.3 FLOOR FS# UTERS TOTAL SAMPLE WGHT %EXAMINED SUB-SAMPLE WEIGHT WOOD WEIGHT Maize kernels, total count Maize cob frags, total count Phaseolus, total count cf. Sapotaceae fruit Fruit rinds Tuber/root frags Lupinus (leguminosae) cf. Ribes (Saxifragaceae) Rubus (Rosaceae) Festuca CPoaceae) Cheno/Am f2_!Y.gonum/Rumex (Polygonac.l Total local foods, 1 Total local foods, 2 Total exotic foods Fruit rinds Total foods, 1 Total foods, 2 Locai:Exolic+Local, 1 Total food:wood, 1 Locai:Exotic+local, 2 Total food:wood, 2 kernel:cob cob:kernel 8 3 1.25 274.675 50 138.35 78.86 36 41 2 4 7 1 I I 1 67 5 81 7 83 7 164 97 0.49 2.08 0.08 1.23 0.88 1.14 ' Table 2 (continued) . (X) (X) \0 890 8 :f E) H I r }I 1 L"J N "'0 0 0 ~ - (ij 0 I- r-1 d ~ 0 ..Q ...0en s 0 u: .l '0 ~ "'0 0 0 H 0 0 >, 0 0 ~ 0 .jJ n '0 1\ ~ M ~ X U) A ·.-! z vv 88 a:> (](] 33 .:H 0 0 0 0 .jJ co l-1 ~ 0 E co l-1 Ol 0 .jJ U) ·.-! :r: .-i 0 0 0 0 0 0 r-... co 0 0 0 lO 0 0 0 0 0 0 -.:t one~:~ . C") 0 0 0 0 0 C\1 0 ,... 0 0 0 Ol ·.-! r:.. 891 8 :f E) H I r >I 1 1-1 0 0 ~ ~ ~ N "C 0 0 ~ "!. d >., ..0 0 H '0 0 0 s "C 0 .E l. ~ n 0 I- A M ...en 0 0 u: ~ 0 .w N ' '0 0 0 ~ ~ 0 X (/} A ·.-i z 1-1 vv 88 0) 00 33 ±I 0 .w CIS ~ 0 E CIS 1-1 Ol 0 .w (/} ·.-i ::r: N 0 0 0 0 0 0 0 0 0 0 1.0 0 10 0 C\1 0 ,.... 1.0 C\J 0 0 . ,.... one~:~ Ol ·.-i rr.. Local :Exotic+ Local 1.00 0.90 0.80 0.70 0.60 0 iii 0.50 a: 0.40 0.30 0.20 0.10 0.00 lf: I:H 8 8 fB ~ N >- X 3: > :::::> t- (J) a: 0 a. z ;::E ...J ~ -, - J: (!) U. m Floors Fig. 3. Histogram of ratios of local to exotic + local by floor. 00 1.0 1:\..) Local ,2 :exotic+ local, 2 0.80 0.70 0.60 0.50 0 ~ 0.40 a: 0.30 0.20 0.10 0.00 l:f: l:f:J 8 8 ffi ~ N >- X ~ > :::> r Cl) 0: 0 a. z ~ __. ~ -, - :I: CJ u. m Floors Fig. 4. Histogram of ratios of local,2 to exotic + local,2 by floor. OJ \.0 w cob:kernel 14.00 12.00 10.00 0 8.00 ':;::: co a: 6.00 4.00 2.00 0.00 ~ ill 8 8 m ~ N ~ X ~ > ~ r 00 ~ 0 ~ Z ~ ~ ~ ~ - I ~ ~ m Aoo~ Fig. 5. Histogram of ratios of maize cobs to kernels by floor. CXl \0 ~ kernel:cob 3.50 3.00 2.50 2.00 0 :;:: m a: 1.50 1.00 0.50 0.00 l:f: HJ 8 8 83 ~ N >- X 3: > :::> 1- cn a: 0 a. z :::2: ...J ~ , - :X: (!) u_ m Floors Fig. 6. Histogram of ratios of maize kernels to cobs by floor. 00 1.0 lTI