This item is only available as the following downloads:
1 THE PRODUCT OF LABOR : POTTERY TECHNOLOGY IN THE UPPER XINGU, SOUTHE RN AMAZON, BRAZIL A.D. 7 00 177 0 By JOSHUA ROBERT TONEY A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2012
2 2012 Joshua Robert Toney
3 T o my parents, Robert and Marie and to the memory of Jim Petersen
4 ACKNOWLEDGMENTS This dissertation is based on data collected over the course of four successive field seasons in Brazil. This field work was supported by grants from the National Science Foundation and the University of Florida Center for Latin American Studies. I would also like to acknowledge the support of sever al Brazilian insti tutions including the Associao Indgena Kuikuru do Alto Xingu (AIKAX), the Museu Paraense Emlio Goeldi (MPEG), the National Museum of the Federal University of Rio de Janeiro (MN UFRJ) the National Council of Technological and Scienti fic Development (CNPq),the National Institute of Historical and Artistic Patrimony (IPHAN), the Nat ional Indian Foundation (FUNAI). While writing I benefited from the support of my employers and colleagues first at Garcia and Associates in Kailua, Hawaii a nd ultimately at The Joint Joint Base Pearl Harbor/Hickam All l aboratory work was conducted in Bel em do Para, Brazil at the MPEG. I would like to acknowledge all of the researchers and supp ort staff at the MPEG who provided me with a temporary work space within their very busy laboratory and graciously allowed me to photograph and reexamine collections from the Upper Xingu Specifically, I would like to thank Edithe Pereira (Dir ector of Archaeology), Nilson Gabas Jr. (Director of MPEG), Ana Vilacy Galucio (Head of Ciencias Humanas), Regina Farias Daniel Lopes, Silvio Costa Vitor Martins, Patricia Oliveira da Silva, Carlos Augusto Barbosa, and Joo Aires Fonseca All field work w as conducted while living in the Kuikuru village in the Parque Indgena do Xingu I would like to thank the entire Kuikuru c ommunity, especially Chief Afukaka and his brother Tabata for their support of the project Chief Afukaka graciously
5 supported our w ork provided our team with a home in the village, and provided daily made our research team feel at home during our successive field season s living in the Kuikuru village. The entire community was welco ming and many participated directly in our research. Among the many people who assisted in the daily archaeology Itsana, Dunga and Masinu became good friends and their assistance was an important part of my experience in the Upper Xingu Also while in t he field I met several Brazilian archaeologists and anthropologists who welcomed me and offered advice during my initial field seasons in Brazil including Carlos Fausto, Bruna Franchetto, and Eduardo Neves. At the University of Florida I benefited from di scussions with my committee members as well as my fellow graduate students. My seminars and conversations with Kenneth Sassaman provided solid ground from which to build m y overall analysis. Augusto Oyuela Caycedo provided constant support and encouragemen t as I spent many long nights and weekends struggling to write in the basement of Turlington Hall. Through field work with David Steadman I gained a deeper understanding of the scope of Amazonian archaeology and its relationship to the Caribbean My analys is would not have been possible without the diligent fieldwork of my fellow graduate students who participated in the Southern Amazon Ethnoarchaeological Project, including Christian Russell, Morgan Schmidt, David Mead, Mark Donop, Diogo Costa, Renata de G odoy and Lus Claudio Symanski I also benefited from discussions with fellow graduate students Josh Torres, Brian Tucker, Randy Crones, Isaac Shearn and Joe Hefner. My archaeological career began in Vermont where I was very lucky to work with a close gro up of archaeologists who shared both professional and personal interests and
6 continued to provide moral support throughout my graduate career At the head of this group was my mentor, friend, and colleague, Jim Petersen. M y archaeological career began with his teaching and friendship. His infectious enthusiasm for the people, culture s, and history of the Amazon prompted me to apply to graduate school to work with Michael Heckenberger. Quite simply, none of my graduate research would have been possible witho ut Jim or Michael. As part of the Southern Amazon Ethnoarchaeological Project I not only enjoyed Michaels professional and intellectual guidance but I also benefited from his two decades of previous work negotiating permits, visas, grants, and other requir ements to accomplish the very difficult task of conducting archaeological research in Brazil and in the Parque Indgena do Xingu specifically. As my gra duate advisor, both in th e field and on campus Michael was cons istently an innovative teacher and a clo se friend who provided institutional, financial, and emotional support during a long succession of shared field seasons, shared tragedies, and shared success I dedicate this dissertation to my parents, Robert and Marie, and my sister Sarah Their support and thoughtful encouragement provided a voice of reason outside of the mental chaos of graduate school and without their constant support none of this would be possible Finally I would like to thank my wife Rachel. She graciously endured my erratic work ethic while simultaneously w orking on her own dissertation. H er encouragement provided a constant source of inspiration and her positive influence cannot be overstated
7 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ .......... 10 LIST OF FIGURES ................................ ................................ ................................ ........ 11 LIST OF ABBREVIATIONS ................................ ................................ ........................... 19 ABSTRACT ................................ ................................ ................................ ................... 20 CHAPTER 1 PROBLEM ORIENTATION AND OVERVIEW ................................ ........................ 22 Introduction ................................ ................................ ................................ ............. 22 The Case Study ................................ ................................ ................................ ...... 26 Ethnoarchaeology ................................ ................................ ................................ ... 35 Technology ................................ ................................ ................................ ............. 37 Subsistence and Landscape ................................ ................................ ................... 39 Pottery, People, and Food ................................ ................................ ...................... 43 Summary ................................ ................................ ................................ ................ 45 2 AMAZONIAN ARCHAEOLOGY ................................ ................................ .............. 46 Introduction ................................ ................................ ................................ ............. 46 Archaeological Models ................................ ................................ ............................ 52 Regional Research ................................ ................................ ................................ 56 Lower Amazon Archaeology ................................ ................................ ............. 56 Upper Amazon Arc haeology ................................ ................................ ............. 60 Central Amazon Archaeology ................................ ................................ ........... 63 Summary ................................ ................................ ................................ ................ 64 3 UPPER XIN GU ARCHAEOLOGY ................................ ................................ .......... 66 Introduction ................................ ................................ ................................ ............. 66 Previous Archaeology ................................ ................................ ............................. 67 The So uthern Amazon Ethnoarchaeological Project ................................ .............. 76 The Nokugu Site (MT FX 06) ................................ ................................ ........... 80 Excavation Trenches ................................ ................................ ................. 80 Excavation Units ................................ ................................ ........................ 83 Test Units ................................ ................................ ................................ ... 92 Surface Collections ................................ ................................ .................... 96
8 The Heulugiht Site (MT FX 13) ................................ ................................ ...... 97 Excavation Trenches ................................ ................................ ................. 98 Excavation Units ................................ ................................ ........................ 98 Test Units ................................ ................................ ................................ ... 98 Surface Collections ................................ ................................ .................. 102 Other Upper Xingu Sites ................................ ................................ ................. 102 4 POTTERY METHOD AND THEORY ................................ ................................ .... 103 Introduction ................................ ................................ ................................ ........... 103 Ceramic Studies in the Amazon ................................ ................................ ............ 105 Ceramic Analytical Method ................................ ................................ ................... 106 Beyond Culture History ................................ ................................ ................... 108 Standardi zation Theory ................................ ................................ .................. 109 Determining Types ................................ ................................ ......................... 113 Labor and Society ................................ ................................ .......................... 115 Ma nufacture and Production ................................ ................................ .......... 115 Ceramic Ethnoarchaeology ................................ ................................ ............ 117 Summary ................................ ................................ ................................ .............. 122 5 POTTERY ANALYSIS ................................ ................................ .......................... 124 Introduction ................................ ................................ ................................ ........... 124 Sample Selection and Method of Analysis ................................ ............................ 124 Analysis Groups/Assemblages ................................ ................................ ............. 127 Attribute Analysis ................................ ................................ ................................ .. 129 Metric Attributes ................................ ................................ ............................. 129 Temper ................................ ................................ ................................ ........... 130 Core Color ................................ ................................ ................................ ...... 132 Decoration ................................ ................................ ................................ ...... 132 Form and Type ................................ ................................ ............................... 135 Vessel Types ................................ ................................ ................................ .. 139 Type 1 Vessels ................................ ................................ ........................ 139 Type 2 Vessels ................................ ................................ ........................ 140 Type 3 Vessels ................................ ................................ ........................ 142 Type 4 Vessels ( alato or griddles) ................................ ........................... 142 Type 5 ( undagi or pot stands) ................................ ................................ 143 The Nokugu Site (MT FX 06) ................................ ................................ ................ 143 Group 1 (Surface Collection) ................................ ................................ .......... 144 Group 2 (Plaza Berm Excavations) ................................ ................................ 167 Group 3 (Plaza Peripheral Excavations) ................................ ......................... 172 Residential Excavation ................................ ................................ ................... 177 The Heulugiht Site (MT FX 13) ................................ ................................ ........... 188 Group 1 (Surface Collection) ................................ ................................ .......... 189 Group 2 (Plaza Berm) ................................ ................................ .................... 201 Group 3 (Plaza Peripheral Excavations) ................................ ......................... 205 Other Upper Xingu Sit es ................................ ................................ ....................... 208
9 Alto Xingu ................................ ................................ ................................ ....... 208 Formadores do Xingu ................................ ................................ ..................... 208 Summary ................................ ................................ ................................ .............. 237 6 ARCHAEOLOGICAL INTERPRETATION ................................ ............................ 244 Introduction ................................ ................................ ................................ ........... 244 Upper Xingu Chron ology ................................ ................................ ....................... 245 Initial Period (Early Ipavu) ................................ ................................ .............. 251 Developmental Period (Late Ipavu) ................................ ................................ 252 Protohistoric Xingu ................................ ................................ ......................... 252 Historic Xingu ................................ ................................ ................................ 254 Upper Xingu Regional System ................................ ................................ .............. 255 Ipavu and Diauarum ................................ ................................ ....................... 257 Ipatse, Itafanunu, and Lamakuka ................................ ................................ ... 257 Amazonia and the Upper Xingu ................................ ................................ ............ 260 Summary ................................ ................................ ................................ .............. 266 7 THEORETICAL CONCLUSIONS ................................ ................................ ......... 272 Introduction ................................ ................................ ................................ ........... 272 Technological Considerations ................................ ................................ ............... 273 Social Considerations ................................ ................................ ........................... 276 Household Archaeol ogy ................................ ................................ ........................ 280 Pottery and Technology ................................ ................................ ........................ 285 Labor and Society ................................ ................................ ................................ 286 Summary ................................ ................................ ................................ .............. 290 LIST OF REFERENCES ................................ ................................ ............................. 297 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 327
10 LIST OF TABLES Table page 5 1. Metric attribute means for all vessels in all groups at MT FX 06. ........................ 151 5 2. Average temper content for all vessels in all groups at MT FX 06. ..................... 152 5 3. T test for equality of means for lip thickness measurements at MT FX 06. ........ 153 5 4. T test for equality of means for rim thickness measurements at MT FX 06. ....... 154 5 5. T test for equality of means for orifice diameter measurements at MT FX 06. ... 155 5 6. Metric attribute means for all vessels in all groups at MT FX 13. ........................ 193 5 7. Average temper content for all vessels in all groups at MT FX 13. ..................... 194 5 8. T test for equality of means for lip thickness measurements at MT FX 13. ........ 195 5 9. T test for equality of means for rim thickness measurements at MT FX 13. ....... 196 5 10. T test for equality of means for lip orifice diameter measurements at MT FX 13. ................................ ................................ ................................ .................... 197
11 LIST OF FIGURES Figure page 1 1. The Kuikuru Study Area (KSA) is located within the Parque Indgena do Xingu (PIX) on the southern periphery of the Brazilian Amazon. ................................ .. 24 1 2. Kuikuru woman processing manioc with an entirely metal suite of pots and a single ceramic pot ( ahukugu ) over the fire in the background. ........................... 28 1 3. Kuikuru woman pounding dried manioc with a w ooden pestle and a hollowed tree stump mortar. ................................ ................................ .............................. 29 1 4. Silo and sacks filled with manioc at the center of a Kuikuru house. ...................... 30 1 5. Kuikuru village adjacent to Lake Ipatse. ................................ ................................ 31 1 6. Manioc griddle ( alato ) with a rim portion of a broken manioc processing vessel leaning against it. Other sherds are used to elevate the g riddle. ........................ 33 1 7. Thematic observations, inferences, and correlations used in this study. ............... 36 2 1. Major archaeological research in the Brazilian Amazon has been historically conducted in the Central Amazon (near Manaus), the Upper Amazon (west of Manaus), the Middle Amazon (near Santar m and Oriximin ), the Lower Amazon (near Maraj Island), and in the Upper Xingu (the Parque Indg ena do Xingu or PIX). ................................ ................................ ................................ 47 2 2. Aerial view showing an example of the mosaic landscape characteristic of the Upper Xingu which combines tropical forest and savannah flood plains. This view is w ithin the Parque Indgena do Xingu near the Kuikuru Study Area and shows Lake Itafanunu in the background. ................................ .......................... 49 2 3. The Parque Indgena do Xingu (PIX) is situated at the transitional zone betw een the Amazonian Rain Forest and the Brazilian Savannah. .................... 50 3 1. Upper Xingu and Lower Culuene archaeological areas and subareas within the Parque Indgena do Xingu (PIX, green shaded area, AX designation, upper large circle, FX designation, lower large circle). ................................ ................. 68 3 2. Upper Xingu periodization based on radiocarbon dates and historic events. ........ 72 3 3. Major clusters of archaeological sites along the Culuene River and in the traditional territory of the Kuikuru near Lakes Itafanunu, Ipatse, and Lamakuka. ................................ ................................ ................................ .......... 73 3 4. Major clusters of archaeological sites within the Kuikuru Study Area (KSA) based on the traditional territory of the Kuikuru. ................................ ................. 74
12 3 5. MT FX 06 excavation and collection area location s. ................................ ............. 81 3 6. Profile drawing of ET 02 (top) and ET 04 (bottom). ................................ ............... 84 3 7. East profile view of ET04 facing northeast with chaining pins at one meter intervals. ................................ ................................ ................................ ............. 85 3 8. East wall profile view of ET 02 facing northeast with chaining pins at one meter intervals. ................................ ................................ ................................ ............. 86 3 9. Profile drawing of ET 03 and its corresponding EUs. ................................ ............ 87 3 10. Profile drawing of ET 05. EU 05 1 was excavated on the south end of this ET. 87 3 11. West wall profile view of ET03 facing northwest with chaining pins at one meter intervals. ................................ ................................ ................................ ... 88 3 12. East wall profile view of ET05 facing southeast with chain ing pins at one meter intervals. ................................ ................................ ................................ ... 89 3 13. Profile drawing of ET 08 and ET 09 and their corresponding EUs. ..................... 90 3 14. Profile vie w of the intersection of ET08 and ET09 showing the east wall profile of EU 8 1 (left) and south wall profile of EU 9 1 (right). ........................... 91 3 15. Excavation Unit 5 1 sherd count by depth. ................................ .......................... 93 3 16. Excavation Unit 5 1 sherd weight by depth. ................................ ........................ 93 3 17. Excavation Unit 8 1 sherd count by depth. ................................ .......................... 94 3 18. Excavation Unit 8 1 sherd weight by depth. ................................ ........................ 94 3 19. Excavation Unit 9 1 sherd count by depth. ................................ .......................... 95 3 20. Excavation Unit 9 1 sherd weight by depth. ................................ ........................ 95 3 21. Test unit soil profiles for transects 1 4. ................................ ................................ 96 3 22. MT FX 13 excavation and collection area locations. ................................ ........... 99 3 23. North wall profile of ET01 at MT FX 13. The single EU was excavated at the east end of the trench. ................................ ................................ ...................... 100 3 24. Northeast view of ET01 at MT FX 13 after excavation. The EU was excavated on the east end to the right. ................................ ............................. 101 4 1. Type 1 manioc processing vessels in use f rom 1884 (bottom, Steinen 1894:Tafel XV), 1950 (middle, Galvo 1953:51, Figure 9), and 2002 (top)
13 where metal vessels have replaced all but a single Type 1 ceramic vessel in the background used for boiling manioc juice into kuigiku ............................... 104 5 1. Examples of painting and modeling methods on decorated body sherds and handles from MT FX 12. ................................ ................................ ................... 134 5 2. Zoomorphic vessel designs recorded by Steinen (1894:Tafel XXIII). .................. 136 5 3. Zoomorphic vessel designs recorded by Steinen (1894:Tafel XXIV). ................. 137 5 4. Kuikuru outdoo r cooking area showing Type 1 ahukugu vessel (large blackened vessel in background) and Type 2 atange vessel (blackened vessel on wire rack) surrounded by aluminum vessels. Globular metal vessels are mainly used for water transport while the low profile ve ssels are used for manioc processing. ................................ ................................ ............. 141 5 5. Type 2 rim with chevron design from surface of MT FX 06. ................................ 146 5 6. Lip thickness and diameter on all surface collected vessels from MT FX 06. ..... 147 5 7. Distribution of Type 1 vessels in surface collection areas at MT FX 06. ............. 148 5 8. Distribution of Type 2 vessels in surface collection areas at MT FX 06. ............. 149 5 9. Distribution of Type 3 vessels in surface collection areas at MT FX 06. ............. 150 5 10. Type 1 rim profiles (700's) from the MT FX 06 surface collection assemblage. 156 5 11. Type 1 profiles (800's) from the MT FX 06 surface collection assemblage. ...... 156 5 12. Type 1 profiles (900's) from the MT FX 06 surface collection assemblage. ...... 157 5 13. T ype 1 profiles (1000's) from the MT FX 06 surface collection assemblage. .... 157 5 14. Type 1 profiles (1100's) from the MT FX 06 surface collection assemblage. .... 158 5 15. Type 1 rim profiles (1200's) from the MT FX 06 surface collection ................... 158 5 16. Exterior view of Type 1 rims from the MT FX 06 surface collection. ................. 159 5 17. Type 2 profiles from surface collection (700's) from the MT FX 06 surface collection assemblage. ................................ ................................ ..................... 160 5 18. Type 2 profiles from surface collection (800's) from the MT FX 06 surface collection assemblage. ................................ ................................ ..................... 161 5 19. Type 2 profiles from surface collection (900's) from the MT FX 06 surface collection assemblage. ................................ ................................ ..................... 162
14 5 20. Type 2 profiles from surface collection (1000's) from the MT FX 06 surface collection assemblage. ................................ ................................ ..................... 162 5 21. Type 2 and Type 3 profiles from surface collection (1100's) from the MT FX 06 surface collection assemblage. ................................ ................................ .... 163 5 22. Type 2 and Type 3 profiles from surface collection (1200's) from the MT FX 06 surfa ce collection assemblage. ................................ ................................ .... 164 5 23. Type 2 rims from the MT FX 06 surface collection assemblage. ...................... 165 5 24. Type 4 (griddle, or alato ) rim profiles from the MT FX 06 surface collection and EUs (top two examples 67 1 and 125 2). ................................ .................. 166 5 25. Lip thickness and diameter on all Group 2 vessels from MT FX 06 ................. 169 5 26. Nearly complete Type 1 flat bottom vessel from EU 3 1. ................................ .. 170 5 27. Type 2 rims from EUs at MT FX 06. ................................ ................................ 170 5 28. Type 1 rim profiles from EUs at MT FX 06. ................................ ....................... 171 5 29. Type 2 and Type 3 rim profiles from EUs at MT FX 06. ................................ .... 171 5 30. Lip thickness and diameter for all Group 3 vessels at MT FX 06. ..................... 174 5 31. Type 1 rim profiles from TUs at MT FX 06. ................................ ....................... 174 5 32. Type 2 rim profiles from TUs at MT FX 06. ................................ ....................... 175 5 33. Type 3 rim profiles from TUs at MT FX 06. ................................ ....................... 175 5 34. Type 4 rim profiles from EUs at MT FX 06. ................................ ....................... 176 5 35. Base sherd profiles from TUs at MT FX 06. ................................ ...................... 176 5 36. Block excava tion at MT FX 06 showing proposed outline of House 1(dashed line) in the peripheral/residential zone north of the main plaza. The House 1 outline is based on the location of central house posts (Feature 9, between N 314 and N 315), wall posts and associate d trench (Features 4 and 4a, between N 307 and N 309), and the interior cooking area (Features 1a 1d). A single radiocarbon date obtained from wood charcoal in Feature 9 dates the likely occupation of the house to sometime around A.D. 1450. The rear house midden contained ceramics characteristic of the Late Ipavu Period and Protohistoric Period. ................................ ................................ ......................... 178 5 37. South facing view of the block excavation of House 1 at MT FX 06. ................ 179
15 5 38. Overview of Kuikuru village in 2002 showing newly constructed house frame and older houses arranged around the central plaza. The layout and size of this house is comparable to that hypothesized in the bl ock excavation at MT FX 06. ................................ ................................ ................................ ............... 179 5 39. Distribution of Type 1 vessels in sub collection area and in relation to House 1 excavation to the north. ................................ ................................ ................. 180 5 40. Distribution of Type 2 vessels in sub collection area and in relation to House 1 excavation to the north. ................................ ................................ ................. 181 5 41. Distribution of Type 3 vessels in sub collectio n area and in relation to House 1 excavation to the north. ................................ ................................ ................. 182 5 42. Distribution of Type 4 vessels in sub collection area and in relation to House 1 excavation to the north. ................................ ................................ ................. 183 5 43. Hypothetical location of second house based on distribution of pottery in sub collection area and in relation to House 1 excavation to the north. ................... 184 5 44. Close up of sub collection area and hypothetical House 2 location in relation to road and plaza berm. ................................ ................................ .................... 185 5 45. Hypothetical House 2 location and vessel type distribu tion. ............................. 186 5 46. Distribution of Type 1 vessels in sub collection area. ................................ ........ 186 5 47. Distribution of Type 2 vessels in sub collect ion area. ................................ ........ 187 5 48. Distribution of Type 3 vessels in sub collection area. ................................ ........ 187 5 49. Distribution of Type 4 vessels in sub col lection area. ................................ ........ 188 5 50. Lip thickness and orifice diameter for Group 1 vessels from MT FX 13. ........... 192 5 51. Type 1 rim profiles (t apered variant) from surface collection at MT FX 13. ....... 198 5 52. Type 1 rim profiles (other variant) from surface collection at MT FX 13. ........... 198 5 53. Type 2 and Type 3 rim profiles from surface collection at MT FX 13. ............... 199 5 54. Vessel base profiles from surface collection at MT FX 13. ............................... 199 5 55. Type 4 rim profiles from surface collection at MT FX 13. ................................ .. 200 5 56. Type 2 adorno (Araquinoid like) from surface collection at MT FX 13. ............. 200 5 57. Type 2 rim from surface collection at MT FX 13. ................................ .............. 201
16 5 58. Lip thickness and orifice diameter for all Group 2 vessels from MT FX 13. ...... 203 5 59. Type 2 rim (Araquinoid like) from EU at MT FX 13. ................................ .......... 204 5 60. Type 1 (bottom right), Type 2 (top row), and Type 3 (bo ttom left, 004 1) rim profiles from EU at MT FX 13. ................................ ................................ .......... 204 5 61. Type 2 rim adorno (Arquinoid like) from TU at MT FX 13. ................................ 206 5 62. Lip thickness and oral diameter for all Group 3 vessels from MT FX 13. .......... 207 5 63. Temper statistics from Simes first excavation at MT AX 01. ........................... 211 5 64. Temper statistics from Simes second excavation at MT AX 01. ..................... 211 5 65. Temper statistics from Simes third excavation at MT AX 01. .......................... 212 5 66. Type 4 rim profiles (top row) and base sherd profiles (bottom rows) from Simes MT AX 01 surface collection. ................................ ............................... 212 5 67. Type 2 rim profiles (top row) and Type 1 rim profiles (bottom rows) from Simes MT AX 01 surface collection. ................................ ............................... 213 5 68. Type 5 (or pot stand, undagi ) from Simes MT AX 01 surface collection. ........ 213 5 69. Type 2 rim profiles from MT FX 01. ................................ ................................ .. 214 5 70. Rim adorno's from MT FX 01 surface collection. ................................ .............. 214 5 71. Type 2 rims from MT FX 01 surface collection with engraved decoration. ........ 215 5 72. Type 1 rims (bottom), Type 2 rims (middle), and Type 5 fragment from MT FX 01. ................................ ................................ ................................ ............... 216 5 73. Type 1 rim profiles from the MT FX 02 surface collection. ................................ 217 5 74. Type 2 rim profiles from the MT FX 02 surface collection. ................................ 218 5 75. Type 5 fragments (top two rows) base fragments (third row from top), and Type 4 rim profile (bottom row) from MT FX 02 surface collection. .................. 219 5 76. Type 2 rims from MT FX 02 surface collection with incised, engraved, and thumbnail punctate decorations. ................................ ................................ ....... 220 5 77. Type 1 rims (top row) and Type 5 fragments f rom MT FX 02 surface collection. ................................ ................................ ................................ ......... 221 5 78. Modeled and incised rim sherds from MT FX 02 surface collection. ................. 222
17 5 79. Type 1 ( bottom left) and Type 2 (top row and bottom right) rim profiles from MT FX 03 surface collection. ................................ ................................ ............ 223 5 80. Type 2 rims from MT FX 03 surface collection with engraved decoration. ........ 224 5 81. Type 2 rims with engraved decoration (top three rows) and Type1 rim (bottom row) from MT FX 03 surface collection. ................................ ............................ 2 25 5 82. Typ e 2 rims (top two rows) and Type 1 rims (bottom row) from MT FX 04 surface collection. ................................ ................................ ............................. 226 5 83. Type 1 rim from MT FX 04 with engraved decoration on the interior (top) and red slip on the ex terior (bottom). ................................ ................................ ....... 227 5 84. Type 2 rims with engraved decoration (top four rows), incised decoration (bottom right), and adorno (bottom left) from MT FX 04 surface collection. ..... 228 5 85. Type 1 rim profiles (bottom row) and Type 2 rim profiles (top row) from MT FX 05 surface collection. ................................ ................................ .................. 229 5 86. Type 2 rims with inci sed decoration from MT FX 09 surface collection. ........... 230 5 87. Type 2 rims with incised and thumbnail punctate decoration from MT FX 09 surface collection. ................................ ................................ ............................. 231 5 88. Type 2 rims with modeled and incised decoration from MT FX 09 surface collection. ................................ ................................ ................................ ......... 232 5 89. Type 1 vessel (small) from MT FX 09 surface collection. ................................ 233 5 90. Whole pot stand ( undagi ) from MT FX 11 surface collection. ........................... 233 5 91. Type 2 rims with engraved decoration from MT FX 11 sur face collection. ........ 234 5 92. Type 2 rims with engraved decoration from MT FX 12 surface collection. ........ 235 5 93. Decorated pot stan d fragment (top left), Type 2 incised rims (top right), and Type 1 rim with red slip (bottom) from MT FX 18 surface collection. ................ 236 5 94. Vessel type percentage within the three analysis grou ps at MT FX 06. ............ 240 5 95. Average orifice diameter among vessel types within the three analysis groups at MT FX 06 ................................ ................................ ................................ .... 240 5 96. Average lip thickness among vessel types within the three analysis groups at MT FX 06. ................................ ................................ ................................ ........ 241 5 97. Rim thickness by vessel type and assemblage. ................................ ................ 241
18 5 98. Temper percentage for all vessel types combined within each assemblage from MT FX 06. ................................ ................................ ................................ 242 5 99. Temper percentage for Type 1 vessels within each assemblage from MT F X 06. ................................ ................................ ................................ .................... 242 5 100. Temper percentage for Type 2 vessels from TFX06. ................................ ...... 243 7 1. A full vessel inventory for a single non chiefly Ku ikuru house. A Type 4 griddle is in the foreground, a large Type 1 ahukugu is decorated to the right, and five Type 2 vessels of various sizes sit behind the griddle, four of them blackened from cooking activity. The cooking area in the background contains examples of Type 1 and Type 4 vessels both suspended over fire. A stack of manioc filled sacks is seen in the upper right. Additional metal vessels are stored behind the house, used exclusively for processing manioc. ................................ ................................ ................................ ............. 282 7 2. This manioc processing area is shared by two households and is located in the backyard between the two houses. Two Type 1 ahukugu sit over fires cooking kuigiku and several metal vessels are scattered throughout the area 283 7 3. Community roof thatching project where all laborers will be paid with meals consisting of manioc and fish. ................................ ................................ ........... 289
19 LIST OF ABBREVIATION S CA Co llection Area, a 100 x 100 m eter gridd ed area with 121 2.0 x 2.0 meter collection units spaced every 10 m eters CU Co llection Unit, a 2.0 x 2.0 meter square divided into 1.0 x 1.0 m eter subunits ET Excavation Trench, a hand excavated trench EU Excavation Un it, a 1. 0 x 1.0 meter or 1.0 x .5 meter square excavated in 10 centimeter level s within naturally occurring soil stratigraphy KSA Kuikuru Study Area MPEG Museu Paraense Emlio Goeldi MT FX Mato Grosso Formadores do Xingu, this is the designation for archaeological sites within the headwaters of the Xingu River including those along the Culuene River in the Kuikuru Study Area MT AX Mato Grosso Alto Xingu, this is the formal designation for archaeological sites along the southern portion, or upper portion of the Xin gu River but north of the confluence of the tributaries that comprise the headwaters PIX Parque Indgena do Xingu, the Indigenous Park of the Xingu where the Kuikuru Study Area is located TU Test Unit a 0 .5 x 0 .5 meter square excavated in 10 centimeter le vels within naturally occurring soil stratigraphy
20 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy THE PRODUCT OF LABOR: POTTERY TECHNOLOGY IN THE UPPER XINGU, SOUTHERN AMAZON, BRAZIL A.D. 700 1770 By Joshua Robert Toney May 2012 Chair: Michael J. Heckenberger Major: Anthropology This study e xamines the variation in pottery technology from prehistoric sites in the headwaters region of the Xingu River in the southern Brazilian Amazon. The pottery analysis is combined with ethnographic observations regarding manioc subsistence and modern indigenous pottery use in the Upper Xingu in the traditional area of the Kui kuru indigenous tribe of the lower Culuene River, the main tributary of the Xingu River Particular attention is given to the connection between material culture and social relations among past and present societies Specifically, the analysis examines tra nsformations in the technology of pottery used specifically for the processing of manioc, which along with fish, is the main staple of the Upper Xingu diet. Three specific domestic pottery forms, or types, are analyzed from archaeological contexts that spa n the period from A.D. 700 177 0 The analysis uses a technofunctional method that focuses on ceramic attributes that specifically relate to the technology of pottery and the ir performance characteristics. These attributes include temper type and amount ve ssel shape, vessel wall thickness, and overall vessel size. The largest portion of
21 metric attribute data was collected from three separate pottery assemblages, or groups, from two separate archaeological sites These data were compared and examined to disc ern any statistically significant variations or changes through time. The most obvious changes observed are a significan t decrease in the variety of vessel size, shape, and temper contents and an increase in vessel size and uniformity, collectively classif ied as increased standardization. These changes are hypothesized to be a result of the transformation of the relations of production that compelled pottery manufact ure This essentially socioeconomic transformation is correlated with other changes document ed archaeologically including the early first millennium A.D population increase and village expansion. These archaeological observations are also correlated with ethnographic data regarding social organization and labor control among chiefly societies of the Upper Xingu. Collectively, the increase in population and the demand for more labor from chiefly elites placed stresses on manioc processing and the production of pottery used in the processing method unique to the Upper Xingu This increased the rout inization and standardization in pottery production and manufacture which narrowed its production to the utilitarian vessels used specifically for processing manioc. This is indicative of a highly controlled and centralized production and manufacturing ind ustry which likely transformed through time from a household level of production to a village level of production. This transformation is further explained by both local and regional developments including the arrival of new indigenous groups to the Upper Xingu and later the effects of the first arrival of Europeans to Brazil.
22 CHAPTER 1 PROBLEM ORIENTATION AND OVERVIEW Introduction The devaluation of the world of men is in direct proportion to the increasin g that t he object which labor produces confronts it as something alien, as a power independent of the producer. The product of labor is labor which has been embodied in an object, which has become material: it is the objectification of labor Karl Marx Economic and Philosophical Manuscripts of 1844 This study focuses on material culture and social relations within both past and present societies of the headwaters region of the Xingu River in southeastern Amazonia, Brazil ( Figure 1 1 ) Specifically it focuses on understanding two aspects of material change in Upper Xingu or Xinguano, society from circa A.D. 700 to 177 0 and how these aspects relate to each other. The first aspect is the tr ansformation of ceramic technology within the continuous tradition of Xinguano pottery spanning this entire period and into the present. The second aspect is the physical expansion and elaboration of Xinguano villages during this breadth of time. Three fur ther aspects tie these two material aspects together. First, the subsistence economy, specifically manioc horticulture and processing relying on the use of Xinguano pottery is examined Second, the political economy specifically the organization and use o f female labor by both chiefly and non chiefly heads of household in the Upper Xingu to maintain certain levels of manioc production is examined Third, the social ( and symbolic) structures that enable these head s of household to manage and organize labor are examined These five aspects of Upper Xingu society, documented ethnographically, and examined archaeologically, are viewed together to address several anthropological and
23 archaeological problems. These problems relate to issues in pottery studies and how a technofuncti onal study of pottery can shape archaeologists re construction of past societies and conversely how archaeologists reconstruction of past societies shape their understanding of pottery production and use and the mechanisms that structure that process Finally, all of this is directly related to the influence of labor and its organization on material production The case study focuses specifically on two large circular plaza villages located near the Culuene River (the main headwater of t he Xingu River) in the Parque Indgena do Xingu (PIX) and o ccupied between roughly A.D. 700 and 177 0 P ottery assemblages of the two roughly contemporaneous sites are analyzed and compared with observations on historic and ethnographic assemblages among th e Kuikuru indigenous community These data suggest that a transformation in ceramic technology through time parallels a concurrent increase in village size complexity and population, all of which put a greater demand on manioc production. The transformat ion of pottery to a more uniform and standardized group of vessels whose form and construction are more specifically suit ed for their function parallels the increase in village size, the construction of village peripheral ditches, village segmenting road s, and the likely effect these developments had on social complexity. Taken together with the size of village surrounding anthropogenic forests, the landscape alterations transformation in pottery production and increased village size, all suggest that t he intensification of manioc horticulture played a role in maintaining the increased p opulation density of the region but was not the sole facto r initiating that increase
24 Figure 1 1 The Kuikuru Study Are a (KSA) is located within the Parque Indgena do Xingu (PIX) on the southern periphery of the Brazilian Amazon.
25 T his study also contributes to the ongoing task of better understanding Upper Xingu village complexes and networks late in time and the associat ed level of ceramic produc tion with in this regional scale of social organization The stylistic differences in pottery present at these two prehistoric sites, and among other less studied site s throughout the Upper Xingu are characteristic of village level ceramic production with regional level influence on style and technology This differs from the regional level of production that exists today where one village manufactures potter y for all villages in the Upper Xingu The consolidation of manufacture in to one village is explained at least in part by the fifteenth century impact of European contact on the Xingu River and all of Amazonia The effects of contact altered the scale of regional chiefdoms and had a direct impact on craft specialization. This st udy follows in the tradition of other studies focused on pottery as technology and on pottery as the outcome of patterned behavior Jame s Deetz classic study of Arikara pottery was among the first to show that the patterning of behavior which produces sta ndardization in artifacts is largely conditioned by the culture of the makers of those objects (Deetz 1965:2). Kenneth Sassamans study of the development and adoption of pottery among hunter gatherers of southeastern North America demonstrated that social conditions not only perpetuate technological change but can also inhibited it (Sassaman 1993:218). As Randall McGuire points out, V. Gordon Childe first put forth the notion that tools reflect the social and economic conditions that produce them and that we can learn abo ut the conditions from the tools but neither the processualist view of mat erial culture as a 'fossil' record nor the postprocessualist notion of material culture as text captures the complexity implied in Childe's two axioms
26 (McGuire 199 2:102). F inally closer to the present study area, Warren DeBoer emphasized through his study of the Shipibo ceramic industry that "a fundamental resource in human life is labour. Labour fuels production by welding raw materials into cultural form and, in the case of female labour, uniquely limits the very reproduction of society (DeBoer 1986: 231). Applying this basic and complimentary theoretical framework to an examination of ceramic technological innovation encourages an explanation that accounts for bo th individual and group patterning. Following these studies and others, the present study attempts to explain the process of technological transformation in pottery within the context of its role as both a tool and a product of labor in the entirely female dominated in dustries of manioc harvesting, processing and ceramic manufacturing This is accomplished using both ethnographic data from past stu dies in the KSA and observations made during residence in the Kuikuru vil lage from 2002 2005. These ethnograph ic data support the interpretation of archaeological data from region. The Case Study T he present study focuses on the prehistory as well as the ethno historic trajectory of pottery in the Upper Xingu. Clearly analogous examples of modern Xinguano pottery used specifically for the processing of manioc are documented at t he latest prehistoric and proto historic sites in the study area. Drawing on this connection improves the following analysis by allowing certain conclusions to be drawn about the relations h ip between form and function of pottery in the p rehistoric Upper Xingu, especially as it relates to their use in processing manioc and cooking fish as is well docu mented elsewhere (Carneiro 1983 ; Dole 1978; Heckenberger 1998)
27 Much of t he field work for t his project took place during the Southern Hemisphere Winter which is the dry season between June and September. The dry season in the Upper Xingu is the harvest season and each morning before dawn women ride bicycles to the manioc gardens outside of the v illage and return with sacks and baskets of manioc tubers. While most of the village remains asleep in their hammock s the rhythmic sound of women peeling and grating manioc roots cuts through the cold morning air ( Figure 1 2 ) Ne ar the village house the rhythm is more complex as several women are busy grinding manioc into large, round, flat bottomed vessels. The processed manioc will be dried and stored for use during the feasting intensive Kuarup ceremonies of August (Ago stinho 1974) The stored manioc also provide s the main element of the Kuikuru diet throughout the rainy season This means that the sound of women g rating manioc is heard throughout the village each day to ensure that enough manioc is processed to maintain household food supplies throughout the wet season. This process of peeling and grating the manioc is followed by further processing to separate the liquid and the pulp. The grated shavings are rinsed with clean water and squeezed by rolling the mash into a tuafi a woven matt that is placed flat over staves that straddle the large vessels. The tuafi is unique to the Upper Xingu and only a few other places throughout the Amazon (Dole 1964) T he preferred method in almost the entire Amazon remains the tipiti a tube shaped plaited basket in which the grated manioc is squeezed though this technology was not adopted in the Upper Xingu. Once squeezed dry, the pulp is formed into small loaves dried in the sun, and saved as a n inferior quality reserve stock. The juice squeezed from the pulp sit s in large flat bottomed vessels until the manioc sediment settles to the bottom of the vessel
28 Figure 1 2 Kuikuru woman processing manioc with an entirely metal suite o f pots and a single ceramic pot ( ahukugu ) over the fire in the background. The water on the top is skimmed off and boiled throughout the day in a nother large ceramic vessel ( ahukugu ) suspended over fire with pot stands ( undagi ) and cooked until the juice becomes a thick starchy liquid ( kuigiku ) If left to ferment the kuigiku would transform into a type of beer but this practice is not k nown in the Upper Xingu as it is in almost the entire northern Amazon. After the water is skimmed to boil into the kuigi ku t he remaining sediment in the original processing vessel is left to dry until most of the water has evaporated. The cakes of sediment are t hen removed from the vessel to completely dry in the sun. This manioc sediment, when dried, is broken up and plac ed
29 in a hollowed out tree stump and pounded into a fine powder with a wooden pestle ( Figure 1 3 ). The powdered manioc will be mixed with water into a gruel that is consumed throughout the day. The powdered manioc will also be bake d on a large griddle into flatbread that, along with fish boiled in a separate vessel ( atange ) is the main staple of the Upper Xingu diet The flour is stored in silos and sacks within each residence ( Figure 1 4 ) A s tock of mani oc flour is essential not only because of its value within the Upper Xingu diet but also for the perform ance of a chief and his duties (Heckenberger 2003) For these purposes manioc can be used as payment for services or to host village guests, especially important if one is to host a Kuarup ceremony when visiting villages will consume manioc provided by the sponsor of the Kuarup. Figure 1 3 Kuikuru woman pounding dried mani oc with a wooden pestle and a ho llowed tree stump mortar.
30 An essential part of th e processing method just described is t he flat bottomed high rimmed vessels Today, these vessels are made of aluminum and bought in the Brazilian markets. These large metal vessels replaced the traditional ceramic vessel for all cold processing of manioc. Likewise, the traditional mussel shell has been replaced with a metal peeler for removing the skin of the manioc Today, t he Xinguano equivalent of the modern aluminum pot, the ahukugu is used only f or ho t processing at which point it is designated as a montegoho The ceramic ahukugu are over 100 cm in diameter, 3 4 c m thick on their lip and rim, almost 30 c m in height, and nearly identical in their manufacture. T he measurements and standardized manufactur e of the modern ceramic ahukugu are in contrast to the earliest ahukugu vessels found at archaeological sites Figure 1 4 Silo and sacks filled with manioc at the center of a Kuikuru house.
31 The processing and cooking activities just described, and all daily pottery related activity in the Kuikuru village take place either in the backyard of the house or in the rear domestic area within the house. S tanding in the middle of the Kuikuru village, in the cente r of the plaza, surrounded by the inward gaze of the lon g houses ringing its edge no pottery is visible ( Figure 1 5 ) Glancing between two houses some pottery is visible in the backyard cooking and processing areas but the plaza itself is clean New p ottery that is awaiting its debut or mo vement in a trading session ( uliqi ) is often stored along the interior walls of houses, or beneath hammocks. During the dry season most cooking activities take place outside the house in the bac kyard, but during the rainy season these activities are moved to the center of the house. Figure 1 5 Kuikuru village adjacent to Lake Ipatse.
32 Within these main domestic areas, and n earby the boiling kuigi ku is found another small ceramic vessel sitting over a fire. In this pot, charred black on the outside from sitting within the lapping of the flames, is a fish gruel that also cooks for many hours. In accordance with taboos against meat and blood, f ish a re not gutted or cut but placed directly in the pot with water and boiled whole. These pots are not regularly cleaned and develop a thick layer of burned residue on the inside as well as the outside. O ne additional ceramic ware is found in domestic areas ; the large flat griddle ( alato ) used for transforming the dried and pulverized manioc flour into flatbread. Nothing is mixed with the flour and the natural sugars melt and bind the flour together as it is spread evenly in a thin layer across the griddle w hich also sits low to the ground suspended by three to four undagi A small fire is kept under the griddle with small branches that are fed into the fire just enough to keep it flaming while the flat bread is being cooked. When the morning of manioc proce ssing is finished and the wet pro cessed manioc is set out to dry, a ll the large aluminum vessels used in the processing are tipped over to allow them to dry and some of them are placed on their side leaning against a tree or drying rack where they sometim es fall and crack This scene is similar around the village behind every house. What is different, and perhaps more important, is the amount of vessels that are found behind each house. Some houses are cluster ed together and share back yard domestic areas In the Kuikuru village this occurs in sections of the village where chiefly individuals live with larger extended families It is in these houses or house clusters where wealth is accumulated in manioc flour. This is an essential accumulation that is use d to sponsor feasts and serves as a reminder of the and the amount of manioc processing labor he can
33 command To produce this excess of manioc flour the chief employs the labor of his many daughters and daughters in law. To al low for this amount of processing the chief is the owner of more large ves sels than anyone in the village. The large ceramic vessels are at constant risk of breakage from accidental mishandling or less frequently from thermal stress. Disposal of the broken pots is not immediate. Because the alato are essentially the same diameter as the ahukugu when a n ahukugu is broken on the body or rim, the base can often be used as a n alato The broken sides of the vessel can also be used as wind breaks around the base of the fire under the cooking vessels. In some cases the remnant of a broken vessel, with its rim still intact will be used as a pot stand to elevate a new pot over the fire ( Figure 1 6 ) Figure 1 6 Manioc griddle ( alato ) with a rim portion of a broken manioc processing vessel leaning against it. Other sherds are used to elevate the griddle.
34 Pieces of broken pots that have no further use usually make it to a deliberate disposa l behind the backyard or off the side of a trail leading away from the village. Along these areas a steady accumulation of organic waste creates visible trash mounds. Occasional stray sherds make their way into the soil of the backyard or even into the int erior edges of the house if they are broken with in the house. This is especially true of the smallest vessels that are used for non cooking related activities such as storage of beads, salt, red pigment, or other dai ly and long term storage items. This bri ef description of ceramic use in the ethnographic context of a modern Kuikuru village provides some reference for assessing the life of a ceramic pot in the archaeological record. In assessing the overall distribution of ceramic remains at the sites in thi s study it is clear that the patterns of use and discard that took place in prehistoric times were similar to those observed in the Kuikuru village especially in the location of trash accumulation Excavations in the center of the plaza at the archaeologi cal sites resulted in the recovery of almost no ceramic remains and the small amounts that were found were not subsistence related However, t he midden ring around the edge of the village plaza of the archaeological sites revealed dense deposits of pottery Excavati ons further away from the plaza revealed the densest midden on the site in areas just behind an excavated house The accumulation of ceramics at the archaeological sites has much to do with the overall construction of the site through time. The p eriod of greatest accumulation seems to be during the period in which the plaza mounds were built up and the village periphery ditches were excavated Less accumulation is found in the various areas between the plaza mound s and the ditches. These areas see m to have densities that
35 are the result of house locations, paths, and formal roads, all progressing out from the central plaza. Subsurface densities in this area are also indicative of long term accumulation. Areas closer to the plaza including the plaza berm itself, are denser with ceramic remains and in some cases contain darker soil This correlation between ethnographic observations and archaeological realities is the premise from which this study proceeds. Ethnoarchaeology as its own field of inquiry can provide the archaeologist with many insights but must be used with caution. Directly relating ethnoarchaeological observations without accounting for differential variables between the present and the past can lead to misguided assumptions about both. A brief review of some pertinent ethnoarchaeological studies follows with special attention to those related to ceramic ethnoarchaeology. Ethnoarchaeology This study is not a ceramic ethnoarchaeological study but rather falls into what is a rare case in which the modern pottery has immediate analogs with pottery forms of the past and in e thnoarchaeological studies Allowing for such analogies are the many modern ethnographic works of the region (Agostinho 1993; Basso 1973 1977 1984, 1995; Carneiro 1960 1961 1970, 1972, 1977, 1978 a, 1978b 1983, 1987a, 1987b, 1989, 2000; Dole 1956, 1964, 1978, 1983, 1984, 1991; Gregor 1977; Heckenberger 1996, 2005) and the ethnographic sketches from the earliest period of research in greater Amazonia (Hartman 1986; Oberg 1953; Steward 1948; Steward and Faron 1959 ; Von den Steinen 1886 ). Drawing
36 on the most recent work in the KSA th e current study examines the relationship between the demand for labor and the transformation of pottery related to that labor (Heckenber ger 2005) The issue of identifying and evaluating social complexity in the Upper Xingu, indeed in the entire Amazon, has been a contentious one throughout the last half century (Carneiro 1960, 1970; Denevan 1976, 1991, 1996; Lathrap 1970; Meggers 1971, 20 00, 2007; Roosevelt 1987, 1999a, 1999b; Heckenberger et al 1999; Neves 1999b) The data presented here contributes to ongoing discussions that interweave hypotheses about landscape alteration, subsistence, and scales and rationales of social organization and complexity Correlating the modern ceramic industry and the prehistoric industry requires the use of ethnoarchaeological observation s combined with archaeological observations and inferences ( Figure 1 7 ) Figure 1 7 Thematic observations, inferences and correlations used in this study.
37 Further, t he correlation between the transformation and standardization of pottery production and the timing of developments in social comple xity in the Upper Xingu region begins with the modern Kuikuru village In terms of the ceramic industry the transformation of the in strument of manioc production pottery and its relationship to labor in an Upper Xingu village, is examined in prehistory through the archaeological record The production of pottery is intimately tied to the production of processed manioc in the modern Xinguano village Situating the transformation of both pottery and the accumulation of surplus subsistence within the broade r context of the social conditions in which it occurs in prehistory is also evaluated in terms of relations of production and the domestic and political economy of the modern Xinguano society cautions by previous researchers not withstanding ( DeBoer and L athrap 1979; Longacre1991). Technology The aim of focusing on technological change is to accentuate the anthropological aspects of archaeological research without abandoning the historical aspects. This presupposes that the reasons for technological chang e are based in social factors that may or may not be expressed in the archaeological record. Implicit among these social factors is the place and control of labor. Technological change happens through labor more specifically the intensification of labor, and the factors that influence labor thus influence technological change. More broadly, he process of technological change is a concern, implicitly or exp but it is also interesting to understand why these technological cha n ges occurred" (Skibo 1994:113). Technological change as an alternate vantage point on the processes of past societies is not without its obfuscations Perceptions of technological change may be
38 emic or etic and without the benefit of ethnographic analogy t he difference may never be known. Anna Shepard perceived some of these problems early in the development of ceramic analysis in archaeology; There are three distinct obstacles to the use of technological features as classificational criteria: (1) identific ation of materials often requires laboratory facilities; (2) evidences of some important techniques are difficult to recognize or, in some cases, they actually leave no identifiable mark in the finished vessel; (3) physical properties are often influenced by both composition and technique and hence their interpretation may be unce rtain (Shepard 1976:310) Nonetheless, observed changes in technology must be situated and explained within the context of what is known of past societies. Once the difficulties ar e recognized and incorporated into the study of ceramic technology t he advantages outweigh the faults Though weary of the faults of focusing on technology, Shepard also pointed out the advantages; The study of technological features offers at least three distinct advantages simpl e criteria for delimiting types (Shepard 1976:311) Shepard was concerned with classification on the basis of technology, which was also the focus of the first researchers in the Upper Xingu. Classification though is only one way of employing the study of technologi cal change. Technological aspects of pottery can provide simultaneous readings of change or continuity where these proces ses are otherwise not apparent; No less important is the fact that abrupt change in one feature is often rendered inconspicuous by the prominence of features that are stable and therefore exhibit only the variations that result from lack of standardization and individual differen ces in skill (Shepard 1976:311)
39 This is especially true in the Upper Xingu where continuity moving towards sta ndardization and uniformity easily masks any changes, especially technological, that are occurring simultaneously. A rchaeologists can determine through analysis what physical changes take place in the construction of pottery over a period of time but what is of interest as anthropologists is why these changes t ake place. In the Upper Xingu, a simple model of societal change is already understood based on broad landscape archaeology and ethnography (Heckenberger 1996, 2003, 2005) A model built on a timeline of Upper Xingu colonization, village expansion village peripheral ditch construction, village abandonment, and modern ethnographic occupation, presents a framework in which to situate what changes occurred in pottery assemblages and better understand why these changes occurred in conjunction with the established timeline. The correlation between village expansion and pottery transformation is of particular interest to this study. Subsistence and Landscape In addition to technology, ethnography, and relati ons of production, subsistence is at the center of this study. Indeed, subsistence has been at the center of most discussions about complexity and social organization throughout the history of Amazonian arc haeology. As Lathrap points out; Our understanding of the age and origin of Tropical Forest Culture is inextricably bound up with our understanding of the age and origin of the major cultivated plants which were basic to the agricultural system (Lathrap 1970:47) Subsistence strategy is a central issue in discussions of the spread and growth of populations and political centralization through out the prehistory of the Amazon. Our Father The Cayman, Our Mother The Gourd set the
40 tone for discussions of the spread of root crop agr iculture and the importance of early house gardens in the Amazon (Lathrap 1975) Understanding the importance of root crop agriculture in the Amazon is essential also to understanding the waves of migrations of people through into, and out of t he Amazon as well as the networks of trade and exchange as has been examined elsewhere (Anthony 1990 ; Boomert 2000 ) Again, as Lathrap noted early in the history of Amazonian archaeology ; I wish to emphasize that the use of bitter manioc as a basis for bread and f lour production is indicative not of a subsistence agriculture but instead of an intensified agricultural economy in which appreciable amounts of the food produced are being f e d into extended trade networks (Lathrap 1977:740) Whether being fed into the ex tended trade networks or fed into the local economy, there is little doubt that manioc contributed to the intensification of the political economy and was part of the economic structure of past societies as much as it is of today's Amazonian and Xinguano s ocieties It has already been noted that chiefs in the Upper Xingu "maintain well kept silos for storage of manioc flour (sometimes well over a thousand kilo s each)", far above that maintained by the average household (Heckenberger 2003:38) However, the f inal product may be less important than one's ability to command the production of that product (Heckenberger 2003) This relationship with subsistence is present throughout Amazonia and in chiefdom societies in general where "chiefs are seen to manipulate food supplies in various ways related to their position s as chiefs" (Drennan 1995:306). The main problem with addressing any issue related to major cultivated plants of the past in Amazonia is taphonomic. T he preservation of direct evidence of th e presen ce and domestication of plants in archaeological sites is very rare Ethnographic analogy, indirect evidence such as pottery remains, and landscape alteration such as
41 anthropogenic soils and forests, remain the best evidence for examining early agriculture Preservation issues th e abundance of cultivated plants, their origins and their dispersals still require much study (Pearsall 1992) Nonetheless, through meticulous research spanning the entirety of Amazonia and Mesoamerica, much has already been achie ved in the area of understanding cultivated plant origins. The first hypothesized center of the development of root crop agriculture, specifically manioc horticulture, was thought to be in the central Amazon, near the conflu ence of the Negro and Solim es r ivers. It was hypothesized by Lathrap that root crop agriculture was concomitant with the first settled village life in the Amazon basin (Lathrap 1970) New research however may place the origin closer to the Upper Xingu or at least in southwest Amazonia with the available data favoring a savannah origin (Isendahl 2011 ). According to this new research the geographical origin of manioc, domesticated as long as 10,000 years ago, was most likely in the savannas, the Brazilian Cerrado, to the south of the A mazon rainforest. Though we may be closer to understanding the origin place of the first domesticated manioc we are not much closer to understanding the choice or preference for bitter manioc over sweet manioc and the timing of its dispersal throughout the Amazon At the heart of t he problem between bitter and sweet manioc is the toxicity of prussic acid and the cyanogenic potential (CNP) in bitter manioc when eaten unprocessed unlike the prussic acid f ree sweet manioc which is simply boiled or roasted bef ore consuming But i s prussic acid a red herring in discussions about why bitter manioc was widely adopted over sweet manioc in Amazonia ? Wilson and Dufour (2002) point out that one possible explanation for the preference of bitter manioc over
42 sweet manioc is due to the higher yields it produces Wilson and Dufour (2002) find that Tukanoans of the Columbian Amazon region prefer bitter manioc in part for their ability to produce consistently higher yields. While they do not provide a conclusive explanation f or the difference in yields between bitter and sweet manioc they surmise that ce is that the high CNP plants are more likely to be disease and/or insect resistant (Wilson and Dufour 2002:49) Further, a s Nye (1991) points out, m anioc is not processed in order to drive out the prussic acid but is processed to create a storable product that can be created in surplus. Lathrap noted this as early as 1977, stating that "bitter manioc is the more evolved or ennobled cluster of cultivar s among the maniocs. The selection process leading to the bitter group of maniocs has been in terms of higher starch yield and in terms of starch of a quality more appropriate for making bread and flour" (Lathrap 1977:741). In fact, t he complex method of p rocessing bitter manioc creates a dried storable subsistence surplus and directly contributes to the creation of ceramic products that are used in the process including griddles that are used to cook the processed manioc flour Subsistence and pottery bo th enter into the domestic economy with direct linkages to the political economy. Though we may speculate about the role or non role of subsistence surplus in the formation of social hierarchy w e may not be able to ascertain whether the hierarchy of consu mption was established before the product was produced or if the presence of the product itself was the impetus for the creation of complexity (Carneiro 1987; Heckenberger 2003). What w e can examine is the correlation between the increased demand for food, an increased demand on the labor needed to produce that food, and technological changes in the pottery used by laborers to produce food
43 Pottery, People, and Food The p re ceramic age is little known in the Amazon and it appears that Pleistocene hunters an d later Holocene Hunter gatherers were concentrated on the prairies and savannahs to the east and south of Amazonia until more active gathering of mollusks and other shellfish intensified along coastal regions (Heredia1994; Roosevelt 1991, 1996, 2002). The se lithic technology based groups eventually expanded into forested areas by the ninth millennium B.C (Schmitz 1987). Quartz and chert flakes and other tools suggest that eastern South America was settled long before the appearance of the Clovis complex in North America (Gruhn 1991; Roosevelt et al 1996, 2002) but very little evidence of that is found in the Amazon ( Sanoja 1994:325). Instead, most archaeological evidence in the Amazon begins with the pottery rich cultures of Holocene horticulturalists. Com menting on the origins of pottery in South America, Sanoja succinctly summarizes the connection between pottery production, those producing it, and the subsistence economy in which this production may have taken place; T he beginning of pottery making is n ot the result of pure chance. It appears at the precise time when cultivated plants begi n to predominate over wild the introduction of instruments of production such as containers is a consequence of the development of the productive forces that began to appear in societi es with a gathering way of life ( Sanoja 1994:631) Whether or not pottery was first produced specifically for food, there can be little doubt that the productive forces associated with food production eventually had some impact on pott ery related to food production. This theoretical cont extualization of pottery fits we ll in the Upper Xingu where p ottery is almost entirely utilitarian and used specifically for processing and cooking food. Though s ome smaller vessels are used for storage of other dry goods, most vessels are used for processing staple subsistence items Vessels used for storage purpose s are not the same vessels used in the preparation
44 and cooking of food and based on the low amount of restricted mouth vessels in the Upper X ingu archaeological record pottery was not used for major food storage Subsistence related vessels consist of several varieties used for three tasks; processing manioc, cooking processed ma nioc, and cooking fish The importance of pottery to the study of subsistence in the prehistoric Amazon cannot be overemphasized and this is especially true in the Upper Xingu where little other material culture exists in the archaeological record. Though the improvement of archaeological techniques h as allowed archaeol ogists to engage other micro scale artifacts including botanical re mains and macro scale features including broad landscape alterations, it is the ceramic industries of the prehistoric Amazon that continue to lay the foundation for the construction of a cu lt ure history for the vast region T hese ceramics hold the potential to advance much more than culture history if studied properly For example, o utside of Amazonia in the nearby Orinoco river basin Anna Roosevelt's work at the site of Corozal in Venezue la demonstrated that tropical lowland sites have much more to offer than simple ceramic sequences. Roosevelt suggests that future problem oriented work will need to expand excavations horizontally across large sites...columnar excavations are not appropri ate for investigating the range of contemporary activities or the composition and function of the groups living in a settlement" (Roosevelt 1997:180). In recent research in the Central Amazon, Eduardo Neves and colleagues have demonstrated the effectivenes s of broad horizontal archaeology (Lima 2008; Neves and Petersen 2006) T his approach is effective in the Upper Xingu as well where village sites are well over 200 meters wide and ceramic
45 remains can be associated with discrete activity areas within these sites (Heckenberger et al 2009) Finally, in assessing the usefulness of domestic utilitarian pottery remains we can proceed with the premise that civilization, it can be shown that the tools. involv ed were themselves developed in response to societal demands ... t his interpretation is presented, not as a repudiation of cultural materialism, but as an elaboration on it (Carneiro 1973:179). In the Upper Xingu, thre e main lines of evidence support this b asic premise regarding material responses to societal demands First, evidence of the transformation of pottery production second, ethnographic evidence regarding the use of pottery for subsistence processing and third, landscape alteration and village e laboration suggesting increased horticulture and complex social organization that likely influenced the production and use of pottery. Summary This study combines ethno graphic observation and archaeology to examine subsistence, material culture, and politi cal and economic control over manioc processing and ceramic production in an Upper Xingu society over the past millennium The se elements all contribute to the structured relationship between labor and technological change through time. Specific to the Upp er Xingu, this study also provides a new interpretation of regional complexes late in time and situates their development within the protohistoric migrations taking place throughout the Amazon prior to European contact in the Upper Xingu during the ninetee nth century.
46 CHAPTER 2 AMAZONIA N ARCHAEOLOGY Introduction The degree to which the network of huge rivers forming the Amazon system was the major avenue for communication and travel cannot be overemphasized Donald Lathrap The Upper Amazon The social history o f the Amazon River basin or Amazonia, is inextricably tied to rivers ( Figure 2 1 ) Likewise, the history of archaeology in Amazon ia is inextricably tied to riv ers and t he trajectory of archaeology in the region (like the archaeol ogically related Orinoco River ) parallels the movement of explorers along rivers. The earliest archaeological information gathered from Amazon ia is defined by the particular river basins that became the focus of early archaeological expeditions by Brazilia ns, Germans, and North Americans. The first major archaeological investigations were conducted at the mouths of rivers such as the Negro River near Manaus, the Tapajs River near Santarm, and at the mouth of the Amazon itself near Belm and Maraj Island ( Evans and Meggers 1950; Hilbert 1952, 1955, 1968; Meggers 1945; Meggers and Evans 1957 ). Rivers also defined the movement and settlement of people in Amazonia both prior to and after the arrival of Europeans. Thus, Amerindians, explorers, and archaeolog ists all follow ed similar geographical routes Even the main dichotomy of all land in Amazonia between dry stable land, or terra firme and seasonally flooded fertile lowlands, or vrzea is defined by rivers and their effect on inhabitable land. This ce ntral difference is also what separates the savannahs of the southern Amazon, and the Upper Xingu, from the wet canopy forests of the main branch of the Amazon River.
47 Figure 2 1 Major archaeological rese arch in the Brazilian Amazon has been historically conducted in the C entral Amazon (near Manaus), the U pper Amazon ( west of Manaus), the Middle Amazon (near Santar m and Oriximin ), the Lower Amazon ( near Maraj Island), and in the Upper Xingu (the Parque Indgena do Xingu or PIX) The general ized landscape s where archaeology is conducted is fortuitous in most cases since Amazonian populations, like their ancestors, often settle in easily accessible areas along river bank s around lakes or along sea shores Without being overly deterministic we can draw conclusions about the relationship of people to the land they choose to settle. E conomic and subsistence realities associated with these landscape features likely influence these settlements although they ar e likely also influenced by social and political factors (Carneiro 1987; Gross 1983).
48 In hydrographic terms the Amazon bas in encompasses an area of about six million square kilometers or an area about the size of th e continental United States, a challengin g area to consider related developments prehist orically. Amazonia is characterized and defined by both its physical geography and its human geography where diverse groups of native peoples and extensive networks of river valleys each combine to create uni que histories (Moran 1993). However, c ombining the entire Amazon culturally just as it often is geographically or environmentally blurs these individual histories. It is this process of synthesizing the entire Amazon that creates many of the diffi culties encountered in its interpretation. As Moran (1993) points out, most anthropologists accept the terra firme/ vrzea dichotomy and in doing so place data from areas as ecologically different as the Xingu Basin, the Rio Negro Basin, and the central Brazilian s avannas into the same category of terra firme adaptations", or into the even more aggregating "lowland South America". He contends that t his process masks the evidence from ecosystems with widely different soils, plants, and climates In regard to anthrop ology, he further contends that this synthesizing is erroneously used to support radically opposing views explaining cultural development, village size, and population mobility (Moran 1993). These attempts to combine evidence from across the Amazon are a r esult of its size and the efforts of researchers working sporadically earch by applying it widely to areas that, although diverse, cannot be separated. Closer to the Upper Xingu, transitional forests an d savannas further distinguish this southern region from the northern Amazon. The transitional forests of the southern Amazon cover the plains known as the Plancie dos Parecis within the geological basin
49 for med between the southern Amazon River basin and the highland plateau of central Brazil. This geographic division is accentuated along the Xingu River by a series of rapids that separate the northern and southern portions of the river (Netto 1964). In the Upper Xingu the transitional tropical forest conn ects the southern upland scrub forests with the northern wooded savannas where distinct broadleaf forests line the Xingu River ( Figure 2 3 ) This transitional zone creates a unique environment that distinguishes the terra firme o f the Upper Xingu from the vrzea and terra firme areas of the lower, middle, and upper portions of the Amazon River to the north. Figure 2 2 Aerial view showing an example of the mosaic landscape charact eristic of the Upper Xingu which combines tropical forest and savannah flood plains. This view is within the Parque Indgena do Xingu near the Kuikuru Study Area and shows Lake Itafanunu in the background.
50 S avannas like those in the Upper Xingu cover almos t two million square kilometers throughout the Brazil ian Amazon and are found in the basins of the Middle and Upper Tocantins, Araguaia, Irir, Tapajs and Xingu rivers These savanna landscapes were mistakenly thought to have caused the limited population s first observed in the region (Steward 1939 1946). Later research demonstrated however that the limited populations were the result of the initial contact with Europeans and their diseases a nd not the soils of the savanna (Heckenberger 1996) T hough it is true that the soils of the savannas of central Brazil are acid, leached, and very deficient in nutrients they are now thought to be the origin place of the first domesticated manioc (Isendahl 2011) Figure 2 3 The Parque Indgena do Xingu ( PIX ) is situated at the transitional zone between the Amazonian Rain Forest and the Brazilian Savannah.
51 O ver fifty percent of the soils in the Brazilian savannas are oxisols with high aluminum saturation Ninety eight percent o f the soils have a pH below six and n inety two per cent of these soils have less than the critical level of phosphorous for many important cereals. Indeed, n inety six percent of the se soils have less than the critical level of calcium for many crop s Many Amerindian groups both past and present incl uding the Kuikuru, overcome these problem s by planting in areas where the soil has already been enriched through centuries of human habitation and slash and char horticulture (Schmidt and Heckenberger 20 09). This successful indigenous method is in sharp contrast to the slash and burn agriculture practiced by soy farmers today who rely on large amounts of chemical fertilizers that eventually render cleared areas unusable while contributing to the degradati on of the water quality in Upper Xingu rivers through agricultural runoff. Besides nutrients, Amazonian soils present another problem in that they only retain enough water in the top twenty centimeters to support crops for eight to ten days (Goedert 1983). When this is combined with the hydraulic restrictions imposed by the limited root depth attained by crops due to aluminum saturation, the crops are at risk during dry spells without irrigation. In contrast to modern introduced crops, the native plants of the savannas have very deep roots and are not affected by the dry spells. T he Amerindian populations of the central Brazilian savannas are hypothesized to include many groups that fled into these upland forests. A population in the savan na may be there bec ause it fled eithe r in the recent or distant past from another region in which its social and political organization developed under different environmental stresses Thus, one environmental practices and organization may be
52 satisfactory in a new e nvironment or it may be a highly adaptive strategy Another group, in contrast, may have adjusted to the upland forests and may find itself today in the savannas, but its social and cultural institutions may still reflect the constructions that came about in a forested environment rather than in grassland savannah (Moran 1993:125). All of these factors should be considered when assessing social, political, and domestic structures of indigenous groups within Amazonia and especially the Upper Xingu. Archaeolo gical Models There is no doubt that the Amazonian environment plays a large role in both the physical reality and the collective past of those living in the region but it is not the only factor in determining the course of history This is true also of the archaeological research in the region. Early a rchaeological studies in Amazonia were shaped by the environmental and ethnographic observations outlined above. The dominant paradigm of environmental determinism in the early twentieth century greatly influe nced the conclusions reached by archaeologists This was coupled with an absence of historical perspective when examining the ethnographies that were shaping the archaeological conclusions. All of this resulted in simplistic views about past Amazonian soci eties which became the basis for further work despite recent adjustments to this way of thinking (Meggers 1954 1960 ; Steward 1949) Later twentieth century studies in Amazonia have always been problem oriented rather than purely descriptive partly becaus e of the accompanying and growing spectrum of anthropological data available for the region. As Neves (1999 a ) points out, these research problems can be grouped into three categories that have shaped the overall archaeology of Amazonia These categories fo llow a linear path through the
53 history of archaeology in Amazonia beginning with a focus on environmental factors, moving to ethnic and linguistic issues, and finally addressing the impact of both European history on the Amazonia n indigenous past and the i mpact of indigenous history on the Amazonian environmental past Today, cross disciplinary, broad scale landscape archaeology often in combination with historical ecology dominates many large scale and long term research projects ( Bale and Erickson 2006; Heckenberger 2006; Heckenberger et al 2009; Neves and Petersen 2006; Schaan 2010 ; McKey et al 2010 ). Among early paradigms in Amazonian archaeology, Brazilian archaeologist Eduardo Viveiros de Castro distinguished between what he called (Viveiros de Castro 1996). The first model of indigenous development in Amazonia, the standard model was synthesized in archaeological terms in the mid 1950 s by Betty Meggers She argued that Amazonia n soils were too poo r to support the establishment of large settled villages based on root crop agriculture. This argument was based on a great deal of other work available at the time. The tendency of early twentieth century Amazonian research to focus on the environment as a contributing factor in the shaping of social and cultural processes of the past is in part attri buted to the publication of the Handbook of South American Indians and the concept of Tropical Forest Culture (Steward 1949) This characterization placed t he Amazon on the periphery of South American cultural development attributing many of its traits to outside sources in the Andes or northwest South America By the 1970s t characterization of Amaz o nia n people s was shown to be faulty and based on the ethnographies of Amerindians who
54 were the survivors of the European arrival and subsequent onslaught of diseases and violence that decimated local indigenous populations ( Heckenberger 1996 ; Hemming 1978 ). New research replaced the model of lim ited development with the revisionist model new archaeological evidence of large, permanently settled villages, in the Central Amazon, Maraj Island, and the Upper Xingu ( Deneven 1998; Lathrap 1970; Roosevelt 1999, 2000 ; Schaan 1997, 2000; Heckenberger et al 19 99; Neves et al 2003) At the center of this debate was the search for evidence of the ability to grow enough food to support large populations and cre ate some sort of surplus wealth to support social hierarchy and complexity The supporting evidence for this argument abounded in new archaeology focused on landscape alterations and the size and permanence of prehistoric villages (Heckenberger et al 1999 ) In the case of the U pper Xingu the ethnographic record combined with local indigenous oral histories provide d a direct historical link to the prehistoric record. T he mistake of early ethnographic work in the Amazon in projecting a picture of small scal e, scattered tribes, into the past, was overcome by contextualizing the basic social and political structure within an historical framework Also intimately tied to the revisionist model later to be a focus of its own is the rich anthropogenic soil asso ciated with archaeological sites known as terra preta or A nthropogenic Dark Earth (ADE) (Lehmann et al 2003; Woods et al 2009). Referred to as e gepe by the Kuikuru, ADE in the Upper Xingu, though not as rich in organics as some examples from the Central Amazon, is located at prehistoric sites
55 and well known to the Kuikuru who plant their crops in it (Heckenberger 2005; Schmidt 2010; Schmidt and Heckenberger 2009) In the Lower and Central Amazon, ADE focused studies show that variation in social organiza tion can be documented by differential ADE distribution within a site ( Schaan et al 2009:139 ; Neves et al 2004) The variability of ADE across time and space has led some to suggest that there is a difference between anth r opic (unintentional) and anthrop ogenic (intentional) soils (Neves et al 2003:35) This differentiation may help explain why ADE sites tend to be vast ly different in the ir size (Neves et al. 2003) Explanations for the creation of ADE range from those favoring population density rather t han time (Neves et al. 2004) to those favoring different use of the land especially differing cultivation practices, as evidenced by the lighter terra mulata (Denevan 2004). The quest to understand the development of ADE has lead to the proliferation of mo re microscopic analysis especially phytolyth analysis of terra preta to determine refuse, agricultural, and domestic areas (Bozarth et al 2009). This extends to pottery where micro analysis and protein residue extraction from ceramic samples may provide a new way to analyze prehistoric subsistence regimes (Barker et al. 2011). Regardless of how ADE were created or maintained, a ll are associated with dense ceramic concentration s (Costa 2004) that vary with the darkness of the soil and all are from the same b road period, speaking to "the e mergence of a new way of life" (Schaan et al. 2009:129 ). Although Amazonian archaeology generally embraces a single paradigm or trend it remains a regionally focused place of research. Amazonian archaeology consists of diver gent research and, like past societies, grows around major centers that disperse
56 their influence. Archaeologists working in the Amazon use d iverse and r egionally specific data to continually refine the broader picture of Amazonian history while maintaining locally focused archaeological projects that provide the foundation for future research (Barreto 1998) Regional Research Despite the problem oriented nature of Amazonia n archaeology, the issue of establishing culture histories persists. Archaeology is co nducted across Amazonia and in places once ignored. Archaeologists have gone from a tantalizing amount of scattered data to an overwhelmin g amount of data that covers an ever increasing area. Chronologies benefit from refinement only i n those places where archaeology has been conducted for the longest period of time These chronologies, lar g ely based on pottery and associated radiocarbon dates are still the basis for reconstructing the indigenous past of Amazonia A review of chronologies in those major ar eas of archaeological research in Amazonia is presented here with a particular emphasis on the observed developments in pottery technology, as related to the current study, which may provide clues to similar developments within the Upper Xingu specifically Lower Amazon Archaeology The earliest studies in the Lower Amazon distinguished two main centers of culture in the Amazon Valley Maraj and Santarm based on a small coll ection of ceramics (Meggers 1945 :208) Meggers also distinguished between these tw o "cultures" based on the presence of mounds and burial urns at Maraj and the lack of these features in Santarm This focus on earthworks in Amazonia continues today and is in fact some of the most reported archaeology in the region in the last decade (E ri kson 2006; Schaan 2010; McKey 2010; Heckenberger 2009 ) On Maraj Island itself
57 more recent archaeology has shown how indigenous land use has affected the landscape of the island area (Roosevelt 2000). Meggers and Evans first stratigraphic archaeological investigations led them to develop an indigenous history since challenged, of a highly developed culture arriving at Maraj Island only to find that the environment could not support them (Evans and Meggers 1950). Though this was one of the first systema tic ceramic analyses on a collection from Maraj Island it did not benefit from the extensive collection of radiometric data available today. Since the beginning of modern Amazonia n archaeology researchers have tried to tie together vast geographic areas w ith incomplete ce ramic chronologies (Howard 1947; Meggers and Evans 1961). This work began in the Lower Amazon (and on the Orinoco) especially at the mouth of the Amazon River on Maraj Island. These studies provide es throughout Amazonia and provide the basic naming systems despite their inadequacies and the inherent difficulties of comparing ceramics and cultures across such a vast region delineated by one of the largest watersheds in the world. Howards stated the distributional data pertaining to the ceramics of the lowland South America area and the formulation of historical interpretations suggested by these d ii). He expanded this lowland area to incl foothills but also the Maritime Andes of eastern and central Venezuela, the island of Trinidad, and the West Indies His predecessor, Thomas A. Joyce, penned the first broad scale analysis of lowland South American ceramics in 1912 and divided them simply into the Tupi was generalized to the southern Amazon and was characterized by round bottomed
58 vessels with inward facing shoulders. The Araw ak complex was characterized by flat bottomed vessels with red on white and modeled decoration similar to those found in the Upper Xingu Other broad studies of lowland cer amics followed through the 1920s and 1930 s (Uhle 1921; Linne 1928; Metreux 1930; No rdenskold 1930; Bennett 1936; Lothrop 1940 1942 ; Rouse 1940; Palmatary 1939; Kroeber 1942). Howards study is the first to define specific ceramic units based on style, trait, and complex, and use these units to divide lowland South America into four distin ct units including Amazonia area drained by the Amazon and its tributaries with the exception of the Lowland Bolivia Meggers and Evans (1961) study followed up on Howards Amazonia an d Orinoco geographic divisions, each of which had severa l localized series of complexes Meggers and Evans applied horizon styles to the regions identified by Howard but under the influence of the Handbook of South American India regrouped Howards subdivisions back into the Tropical Forest group and working backwards attempted to regroup those complexes that Howard analyzed, as well as newly presented complexes not available to Howard, into horizon styles following the methods of Alfred Kroeber (Meggers and Evans 1961) The result was the identificat ion of four horizon styles that, for better or worse, are still in use today. These are the Zoned Hachure, Incised Rim, Polychrome, and Incised Punctate. The Zoned Hachure horizon (5 00 B.C. to A.D. 500) includes complexes spread from the foothills of the Andes in eastern Peru to Maraj Island at the mouth of the Amazon. The diagnostic decoration on these vessels is a broad line incision bounding
59 zones of fine line incision often in a cross hatched pattern. Red bands of painting also co occur on these vessels with the zoned incision. These vessels also contain a wood ash temper ( cariap ) that Megger s and Evans conclude is a precursor to the widespread use of cariap throughout Amazonia including the Upper Xingu. The Incised Rim horizon (A.D. 100 800) includes complexes on the lower Amazon and middle Orinoco rivers. The diagnostic decoration on these vessels is a broad flat topped rim decorated with broad incisions and red paint or slip covering all of the exterior or interior of the vessels. Meggers and Evans felt that this was the most onclude that the Incised Rim horizon is identified more by lacking what other horizons have than by having a common trait. They apply one of the first radiometric dates to this horizon possibly narrowing its presence from roughly A.D. 500 800 or as Rouse and Cruxent (1963) date it along the Orinoco, from A.D. 350 1150. The Polychrome horizon (A.D. 600 1300) includes complexes from the Napo River eastward to the mouth of the Amazon. The diagnostic decoration is white slip with red and black painting and inc ised decoration combined with slipped surfaces which is absent from the other horizons. A cambered rim is also diagnostic of this horizon and the horizon occupies a similar chronological position in regard to the other defined horizon styles. The introduct ion of the urn burial into the lowlands is also associated with this h orizon style Interestingly, g riddles, an important vessel type throughout the Amazon are known from western Polychrome complexes but not from those to the east.
60 The Incised and Punctat e horizon (A.D. 1000 1500) includes complexes on the lower Amazon, th e Orinoco, and British Guiana. These include the sites of Arauquin, those near Santar m at the mouth of the Xingu River (Corra 1965), and across the Amazon River from Santarm in the Ori ximina region, especially at Konduri (Hilbert 1955) The diagnostic decoration is a combination of incision, punctates and especially modeling of anthropomorphic and zoomorphic adornos, the most elaborate examples of which come from Santar m (examples fro m the Upper Xingu are presented later) This is a uniformly late horizon and is associated with griddles and a late wave of people moving from the Orinoco into the Caribbean and south into the Amazon. S ynthesis by Cruxent and Rouse of Venezuelan archaeolog y contains two main hypotheses about connections with the south. One suggests that "traits of Arauquin may have diffused... to Amazonia to produce some, if not all, of the modeling incision of that region and Rouse 1952:37). Upper Amazon Archaeolo gy Though the lower Amazon received much of the early attention of Amazonia n archaeologists, the Upper Amazon became the next focus. extensive research on the Upper Amazon, specifically his work at Yarinacocha and the Ucayali basin formed the ba sis for his treatise on the Upper Amazon establishing it as the setting for the development of Amazonia n Tropical Forest Culture (Lathrap 1958 ) In this early work he first found the connections between Barrancoid ceramics of the Lower Orinoco and Hupa iy a even narrowing the comparison down to the Los Barrancos ceramics (Lath rap 1958:386) He also developed his ideas about the Upper or Central Amazon as an origin place excluding Maraj or the Ucayali basin "as the point of origin for this polychrome trad ition" (Lathrap 1958:386).
61 His classic work, The Upper Amazon addressed many of the issues that persist in Amazonia n arc haeology today (Lathrap 1970) Among these broad issues are the effects of the Amazonia n environment agriculture and horticulture an d their origins, language groups and their clues to past migrations, including the Barrancoid archaeological culture and the Carib and Arawak migrations landscape archaeology, including the ridged fields of Bolivia, Ecuador, and the Guiana's and finally, the archaeological cultures of the Upper Amazon, including those around Lake Yarinacocha and the Uc ayali River basin oldest clear cut evidence for human occupation in the of Lake Yarinacocha in the Peruvian Amazon (Lathrap 1970:84) Based on the midden site of Tutishcainyo, he identified the Early and Late Tutishc ainyo based on pottery remains and based on comparisons with Kotosh pottery placed the Early Tutishcainyo sett lements between 2000 B.C. and 1600 B.C. Decoration included zoned incision and exterior red paint. Of (Lathrap 1970:85) siderable variety and high standardization of vessel shape all argue for a cuisine in which vegetable foods were varied and im and he notes Late Tuti H owever, based on this pottery he still felt that there seem ed to be change in subsistence pattern or settlement size and the presence of exotic ceramic forms suggeste d trade over a considerable distance (Lathrap 1970:89) Finally, t he Shakima
62 pottery style, also a continuation of Early Tutishcainyo, is represented by large open mouthed flat bottomed vessels. Simple incision and well polished red exteriors take the plac e of the zoned incision of earlier forms. carries with it distinct attributes of the Barrancoid (or Modeled Incised) ceramic style first found in the Orinoco Basin (Lathrap 1970 :117) Designated Hupa iya in the Central Ucayali Basin, this pottery does not show continuity with Shakimu or Tutishcainyo. Decoration is present on over fifty percent of vessels and includes broad line incision and zoomorphic adornos. Lathrap surmised th pot form suggests that bitter manioc was the staple and was being processed for bread and the Hupa iya called Yarinacocha. This pottery i s thick and poorly made, according to Lathrap. Importantly, large griddles appear for the first time By A.D. 500 another poorly made form of pottery replaces the Yarinacocha. Ca lled the Pacacocha, the forms are dominated by globular vessels, poorly fired, and not well smoothed. Decoration is limited to all over red slip and inward facing rim adornos. body and pottery complex, Lathrap suggests that it evolved through three phases and d uring the middle phase, the Cashiboca o complex, the frequent use of a single row of thumb print corrugation immediately below the rim is
63 prevalent The final phase is known from a single site, Nueva Esperanza, da ted to roughly A.D. 700. This 100 m eter circular plaza is characterize d by an outer ringed midden with a clean plaza. These descriptions of pottery and village developments from the late first millennium Upper Amazon are strikingly similar to developments found in the Upper Xingu during the same period as will be presented later. The last migration into the Upper Amazon is represented by the Cumancaya pottery style. It shows no affinities with earlier styles in the area and is hypot hesized to originate in B olivia, the prehistoric correlate for later Shipibo Conibo pottery styles. Central Amazon Archaeology Although first explored archaeologically in the 1960s by Peter Paul Hilbert, t he Central Amazon began receiving more attention in the 1 990 s The Central A mazon Project ( or CAP ) has identified over 60 prehistoric Amerindian archaeological sites within a study area of 900 1000 square kilometers near the confluence of the Solimes and Negro rivers near Manaus ( Petersen et al. 2001a) This research has identifi ed two major ceramic complexes and a third possible complex intermediate between them. Major archaeological sites of the Central Amazon reveal intense terra preta in prehistoric village sites dating back more than 2,000 years ago. This extensive work in po ttery rich terra pret a deposits has produced its ow n sequence (Petersen et al. 2001a) related to that established specifically for the Central Amazon by Hilbert (1968) and that developed by Meggers and Evans in the Lower Amazon. The earliest deposits at st ratified ceramic sites in the Central Amazon date fr om roughly 360 B.C. to A.D. 850 and contain Modeled Incised pottery styles (or Incised Rim, more broadly Barrancoid) Representative of the earliest of these complexes is the Osvaldo site, dated to A.D. 4 50 650. It is a 2 4 h a site arranged in a circular village
64 pattern unlike the later ceramic complex sites which tend to be long and linear in configuration with defensive earthworks At Osvaldo there is apparently not evidence of a stratified society. Th e beginning of the pottery sequence at this site is characterized by thin, hard pottery that is grit and grog tempered and bell shaped. It consists largely of unpainted and non s lipped vessels of various forms. It is only later that high amounts of riverin e sponge temper ( cauxi ) and griddles appear along with burial urns at the very end of this sequence. Similar pottery appears in the Upper Xingu almost a century later with the exception of the burial urns. A possible intermediate period the Paredao (or I ntermediate complex) is dated from A.D. 850 to 950 .T hough this ma y be a part of the later period, Lathrap suggest that it is a short lived invasion, as is found in the Upper Amazon around this same period. Finally the latest prehistoric period in the Cent ral Amazon dates from roughly A.D. 950 to 1440 Locally known as Guarita, it is associated with the Amazon wide Polychrome or possibly Araquinoid series as found i n the Orinoco basin and possibly throughout t he Southern Amazon late in time. Guarita sites c over 30 h a or more and are long and linear stretching over three kilometers Guarita pottery shows a distant relation to the Araquinoid series with painted and slipped wares that show functional and possibly social distinctions typical of chiefdom level s ocieties. Cariap is a minority temper which appears only in fine Guarita wares. Pot stands and griddles are prevalent as are thumbnail punctate s along the lips of vessel rims, again, similar to Upper Xingu developments during this same general period. Sum mary The descriptions of each of these regional chronologies are driven by change in pottery style and construction associated with village expansion New paradigms in
65 archaeology seek to find ways to understand this change aside from simple invasions, mig ration s or exchanges. Since the 1970 s archaeologists have moved away from grand unifying theories and shifted their focus to more specific anthropological issues with a few notable exceptions that benefit from a continued focus on the bigger picture (see Hill and Santos Granero 2002 ; Fausto an d Heckenberger 2007 ). Donald Lathrap, perhaps better than anyone in the 1970 s, summarized and presented the issues central to explaining change in the form of settlement and subsequent cultural development of the ent irety of Amazonia Beginning with the idea of in situ development of what was th Tropical F or est C he proposed areas within the tropical lowlands rather than a source in the Andes based on his reading of the archaeological d a ta as well as other researchers including the geographer Carl O. Sauer who proposed areas in n orthern South America at the junction of riverine flood plains and arid savannahs (Lathrap 1970:49). While Sauer may have been right, new evidence may suggest th at the junction of flood plains and arid savannahs was to the south rather than to the north (Isendahl 2011). Although the timing of ceramic developments and population growth across the Amazon are not simultaneous, they are similar enough to be related. T he lack of contemporaneous developments is likely a result of lag from various migrations, diffusions, and other non immediate transmissions of knowledge and culture.
66 CHAPTER 3 UPPER XINGU ARCHAEOLOGY Introduction The headwaters of the Xingu River and the indigen ous tribes living in the region were entirely unknown t o the Western world until Karl v on den Steinen reached the area in 1884 (Steinen 1886 1894 ) Several other explorers navigated their way into the area over the following 60 years and although the Kuik uru recall several excavations on the Cu risevo River in the 1920s or 1930 s these were never reported (Heckenberger 1996). T he first full and detailed ethnographic description of the are a was published in 1948 with Claude Levi Han dbook of South American Indians where he described the Upper Xingu tribes collectively (Levi Strauss 1948:321 348) This was Indian Tribes of M ato Grosso, Brazil which focused on describing collectively thos e tribes living on the Batovi, Cu risevo Culuene, and Tanguro rivers (Oberg 1953) These four major rivers form the eastern headwaters of the Xingu River The combination of these two e arly publications and those of Steinen forged a collective image of Upp er Xingu society and ma terial culture. Very little in the way of oral history was recorded in these publications however and the image of Uppe r Xingu society created by these early documents became the image associated with past societies as well Levi St rauss ac knowledged as much in saying that the history of the area is not well know n ecounting the movements of the Suya, Cayabi, and Baicari in and out of the Upper Xingu he generalized that, S imilar migrations within a relatively small area are said to have been made by most of the tribes prio hus, the general trend is toward tribal intermixture and concentration of population on the river banks (Levi Strauss 1948:323).
67 This fragmentary history is the starting point f rom which archaeology commenced in the region. Further migrations and movements of groups took place over the next decade or so until the formal creation of the P arque Indgena do Xingu (PIX) in 1961. All of the archaeological research discussed in this st udy was conducted in t he Kuikuru Study Area ( KSA ) This area is located almost entirely within the PIX although the area traditionally inhabited by the Kuikuru and other Upper Xingu tribes extends to areas outside of the PIX including the site of Kamakuaka to the south Previous Archaeology Throughout the 1950s and 1960 s anthropologists continued to document various tribes of the Upper Xingu and include d limited archaeological observations in their publications ( Galvo 1953; Schultz 1961; Simes 1963; Schad en 1964; Galvo and Simes 1965). However, most of these studies remained north of the twelfth parallel and on the main branch of the Xingu River never reaching past the major confluenc e of the headwater tributaries or formadores do Xingu ( Figure 3 1 ) In 1953 and 1954, while conducting ethnographic fieldwork among the Kuikuru, Gertrude Dole and Robert Carneiro conducted the first archaeological work in the headwaters region south of the twelfth parallel along the lower Culuene R iver. The results of this work provided the first glimpse into the prehistory of the Upper Xingu limited testing provided some basic preliminary insights on the prehi story of the Upper Xingu, the pottery industry specifically, and also on the stratigraphic and taphonomic nature of the archaeological sites, anticipating future issues such as site reoccupation and reversed stratigraphy in major earthworks.
68 Figure 3 1 Upper Xingu and Lower Culuene archaeological areas and subareas within the Parque Indgena do Xingu (PIX, green shaded area, AX designation, upper large circle, FX designation, lower large circle).
69 Dole observed t hat no soil stratigraphy was present in any of the test unit s excavated and she made her chronological observations based on superposition alone (Dole 1961:400). Another important observation was made at the site of Atiki. Dole found ceramic remains almost 75 cm below the surface at this site which the Kuikuru informed her had been inhabited by their ancestors between 1870 and 1880. Dole surmised that this was not long enough to deposit ceramics to such a depth and that the deeply buried ceramics were eithe r a product of very intense erosion or more likely of a much older age. Importantly, the information collected from the Kuikuru suggested that further archaeological work would need to address the issue of recent reoccupation of many sites. Her excavations within site middens provide the first gross evidence of differences in early and late prehistoric pottery in the region. She found that pottery deposited in the upper levels was frequently painted red, rarely incised, tempered with cauxi and had gradual ly flaring rims. Pottery in the lower levels had less red paint, included cariap temper along with cauxi and tended to have angular rims. In the 1960 s Mrio Simes of the Museu Paraense Emlio Geoldi (MPEG) also conducted archaeologica l excavations at t wo locations on the Upper Xingu near the Posto Diauarum of the newly formed PIX and at Lake Ipavu on the lower Culuene River north of the KSA and the sites investigated by Dole Simes conducted fieldwork at five sites in the area of Lake Ipavu in 1966 (S imes 1967). He recorded seven sites (MT AX 01 to MT AX 07) for this area designated AX to denote the uppermost portion of the Xingu River proper (Alto Xingu) as opposed to the FX designation to denote the headwat ers region of the Xingu River (F ormadores d o Xingu). Eduardo Galvo also
70 conducted limited archaeological investigations at the site o f Yakar (MT FX 09) in the 1950 s (Galvo 1950; Oberg 1953). Simes regional survey included limited excavations which provided enough data for several preliminary co nclusions. The first broad conclusion was that the two areas, chosen because of their rel atively dense concentration of s ites, represented two distinct complexes within the Upper Xingu. This was mainly based on the fact that the sites near Diauarum contain ed pottery with a higher percentage of cariap while the sites near Ipavu contained pottery with a higher percentage of cauxi It is not clear if Simes based this observation on stratified remains, surface remai ns, or some combination of bot h, but a clos er look at his data in the following chapter may help to understand this. He also observed that pottery of the lower Culuene River sites was more likely to be decorated than th at of the Upper Xingu river sites. Despite th e differences among sites of the Xi ngu and Culuene rivers he also observed many similarities including the location of sites along river banks and lake shores the the presence of sherds characteristic of one regi region ( Simes 1967:14 2). This final observation opens the door to many questions regarding regional systems in the Protohistoric period Simes observ ation of mixing of stylistic attributes at different sites, c ombined with Levi Strauss recounting of tribes moving in and out of the Upper Xingu recording of site reoccupation in recent time s all suggest that recognizing regional systems will be somewhat difficult until solid ceramic chronologies are built so that contemporary ceramics can be identified from those of past occupations.
71 Other research has shown that Simes cultural and temporal ceramic sequence may directly relate to the ceramic seq uence from areas investigated fa rther to the south (Heckenberger 1996). Simes correctly identified differing pastes in the ceramics he identified, but his determination of whether or not this attribute was an indication of time or space remains suspect without further examination of his site stratigraphy and the context of his radiocarbon dates. Simes based very little of hi s limited analysis of the region on ceramic form and how it might have varied through time. He also did not b enefit from our current understanding of village size and regional configuration, though he did recognize their basic layout ( Simes 1967; Simes a nd Costa 1978 ). In 1973 and 1980, Pierre Becquelin conducted investigations at one of Simes sites (MT FX 07) which Becquelin named Tuatuari due to its location between Posto Leonardo the confluence of the Tuatuari River and the Culuene River (Becquelin 1993) Becquelin also conducted very limited investigations at Yakar (MT FX 09), Miararr (MT FX 08), Moren (MT AX 08), and Nokugu (MT FX 06), the only other work conducted in the KSA besides that of Dole and Heckenberger. Finally, the most comprehensive work in the Upper Xin gu is that lead by Michael Hecke nberger as pa rt of the Southern Amazon Ethno archaeological Project, beginning in the early 1990 s This work provides the basic wo rking chronology for the Upper Xingu ( Figure 3 2 ). This work focused specifically on the traditional territory of the Kuikuru tribe or the KSA ( Figure 3 3 and Figure 3 4 ) As in most regions of Amazonia archaeology in the Upper Xingu is restricted t o the ceramic age. The earliest reliable radiometric data from the Upper Xingu place the earliest ceramic using people along the Buriti ( Angahuku ) R iver by at least A.D 700 but probably much earlier (Heckenberger
72 2005:88). According to Heckenberger, the ap pearance of circular villages by ca. A.D. 7 00 is evidence of the migration of Arawak speakers into the area. To date, all prehistoric archaeological materials collected in the Upper Xingu are related to these circular village occupations which places these materials at the center of important questions related to the Arawakan Diaspora (se e Hill and Santos Granero 2002), among other issues. Ceramic remains recovered from late prehistoric circular village occupations in the KSA show similarities in form, deco ration, and technology to northern Amazonia n ceramic complexes, as observed by Dole (1961) and Heckenberger (1996, 2005) Heckenberger also observed possible similarities to the ceramics of the Uru Tradition in central Brazil south of the KSA (Wst and Bar reto 1999), accentuating the transitional location of the Upper Xingu between the Amazon Lowlands and C entral Brazil. Figure 3 2 Upper Xingu periodization based on radiocarbon dates and historic events.
73 Figure 3 3 Major clusters of archaeological sites along the Culuene River and in the traditional territory of the Kuikuru near Lakes Itafanunu, Ipatse, and Lamakuka.
74 Figure 3 4 Major clusters of archaeological sites within the Kuikuru Study Area (KSA) based on the traditional territory of the Kuikuru.
75 Simes designated the Ipavu phase to refer to all the archaeological materials he encountered in the low er Culuene River, reserving Diauarum for those along the upper Xingu River and excluding those attributed to the historic Xinguano or Waur pottery. Simes published two radiocarbon dates, however, a nalysis of additiona unpublished date s show that stratig raphic inconsistencies and poor association cast at least some doubt on Simes evaluation of the archaeological materials (Heckenberger 1996). Settlement and other site data also suggest that in fact, the Diauarum Complex and the Ipavu Complex may likely be slightly divergent complexes of the same basic cultural pattern (Heckenberger 1996:66) and related to other complexes in the lower Culuene area ( Figure 3 4 ) On the basis of new dates and more extensive stratigraphic excavati ons Heckenberger divides the prehistoric sites of the lower Culuene into two geographically distinct archaeological complexes, the Eastern and Western complexes both of which share characteristics of Simes Ipavu phase ceramics. Sites identified by Simes north of the confluence of the Xingu tributaries appear to be related to Heckenberger's Western Complex of sites while the Diauarum sites, "may represent a third distinctive complex (Heckenberger 1996:71). These complexes more closely approximate "archae ological cultures" sharing essentially identical features of material culture differentiated only by settlement patterns. The Western complex corresponds to the large fortified villages identified to the west and north of Lake Itafanunu (Heckenberger 1996 2005 ) dating back to at least A.D. 1250 while the Eastern com plex, situated closer to Lake It afanunu corresponds to the smaller unfortified villages with circular structures that do not have counterparts in the fortified villages to the west.
76 Heckenberge r surmises that the Eastern complex villages likely represent the late movement of Carib speaking peoples into the area around A.D. 1500. Since this time, these discrete complexes (Western and Eastern) have merged forming the basis for the multiethnic regi onal culture (Xinguano) which continues to the present day. Since the Simes collections have not been fully analyzed, the relationships between the Eastern complex, the Western complex, and the Diauarum phase, are not fully understood but will be explored further following the pottery analysis i n Chapter 5 The Southern Amazon Ethno archaeological Project All of the data prese nted here were collected during nine months in the KSA over the course of four years (2002 2005) and an additional four months of lab oratory work conducted at the MPEG in Belem during 2007. D uring the most recen t phases of field work in the KS A vertically stratified ceramic samples were obtained from various horizontal contexts at prehistoric sites throughout the KSA These were the sub ject of a technofunctional analysis, complimented with ethnographic observations, to provide data sufficient to document technological and stylistic variation in ceramic samples across time and space. During the first field season the primary goal of rese arch was to map the known sites in the vicinity of the KSA This mainly focused on those sites close enough to the Kuikuru village that allowed for our return to the village at the end of each day. However, our GPS survey also i ncluded several sites that i ncluded multi day trips to other smaller Kuikuru villages as well as to areas around Lake Itafanunu where no indigenous villages or houses currently exist E ach site was mapped with a sub meter accurate Trimble GPS unit with real time correction Typicall y, several Kuikuru assistants and a team of archaeologists hiked
77 through the forest following anthropogenic landscapes such as road berms plaza berms, and ditches, at prehistoric village sites. The ditches were generally known to the Kuikuru and accounted for by oral tradition and myth. E ach of the village sites are referred to using names previously given by the Kuikuru ( Dole 1961; Heckenberger 1996, 2005) Maps for each site, including all of the major landscape features were completed in the first seas on of field work. Several excavation trenches (ET) bisecting the village peripheral ditches were also completed in the first field season as well as smaller excavation units (EU) At the Nokugu site ( MT FX 06 ) a total of 13 ETs and a total of nine EUs we re hand excavated At the Heulugiht site ( MT FX 13 ) a single ET was excavated along with a single EU Each one meter wide EU was excavated to sterile non anthropogenic soil across each village peripheral ditch as well as across village plaza and roadsid e berms. After completion of each trench a 1.0 x 0.5 m EU was excavated at the end of each ET in 0 .5 x 0.5 m subunits and in 10 cm levels within the natural stratigraphy. All soil from these excavations was screened through one quarter inch mesh resulting in the recovery of pottery and charcoal remains Additional charcoal samples were hand collected from the cleaned and profiled wall s of both the main ET and the EU Many of the dated charcoal samples for m the basis for the cultural chronology of the KSA I n the following field season several 100 x 100 m surface collection areas (CA) were designated. Collection areas were place d in six locations at MT FX 06 and at three loca tions at MT FX 13. S urface CAs were placed in various portions of the site, represent ing plaza proximal and plaza marginal areas, overlapping the plaza berm in the case of the latter, and overlapping the peripheral ditches in the case of the former.
78 Each site had one CA each for the n orthwest, southwest, and s outheast quadrants of the site s, with additional CAs placed at MT FX 06 in the northcentral, n ortheast, and peripheral south of the site. These additional CAs at MT FX 06 were placed to further investigate the nature of presumed domestic areas and plaza marginal areas, respectively In side each CA a 2.0 x 2.0 m collection unit (CU) was placed at 8 m intervals (10.0 m spaced on center) along eleven transects within the CA running north to south and counted from west to east This created a total of 121 CUs per CA Additionally each CU wa s subdivided into four 1.0 x 1.0 m sub units. After each CU was designated with pin flags trained Kuikuru a ssistants carefully cleaned each CU of vegetation which in most cases was light. Each CU was then scanned for all surface artifacts and collected acc ording to sub unit, placed in a bag, and labeled for later analysis. In addition to the extensive surface collection, t ransects of sub surface test units (TU) were also excavated within each CA to supplement the surface data and characterize the soil strati graphy across the plaza marginal areas T ransects of TUs were placed a cross each CA in 20 m intervals running from west to east and from north to south. Along each transect six 0 .5 x 0 .5 m TU s were excavated in 10 cm levels and all soil was screened throug h one quarter inch mesh Profiles of all completed excavations were drawn and soil colors were recorded using the Munsell Soil Color Charts. Additionally, all excavations were photographed an d soil samples were taken in 10 cm intervals from each TU for la ter study (Schmidt 2010). In the final field season MT FX 06 was also the subject of a larger horizontal block excavation. This effort, along with a localized surface collection area south of CA01 was
79 undertaken to investigate discrete domestic and residen tial areas possibly associated with house locations, related activity areas and middens The block excavation focused on an area within CA 01 on the north side of the main plaza at MT FX 06 Previ ous surface collections and sub surface testing indicated tha t this area was characteristic of a domestic area, with house trash middens surrounding a lower concentration of artifacts in lightly colored non terra preta soils. Vegetation was caref ully cleared from a 40 x 40 m area which extended from just north of th e plaza berm to just overlapping the presumed house trash midden. The site wide grid was reestablished over the block excavation area Excavation in this area proceeded in 1 .0 x 1 .0 m units further divided into 0.5 x 0.5 m sub units E levations were taken a cross the cleared area using a transit and a localized datum The block excavation pro ceeded in 5 .0 cm levels for the first 15 cm below the relatively even and flat surface in an effort to identify domestic features related to the construction o f a house or household activities. This method identified several significant features just belo w the hard compact surface. Excavation of these features resulted in the identification of the most significant feature, a central hearth which had a definitive large central house post feature to its east T he largest concentration of household ceramics were recovered near these features including pot stand fragments ( undagi ) and rims of large manioc processing vessels ( ahukugu ) In addition to the exc avation in the domestic area, a large area to the north of the house excavation was cleared and a similar grid of 1.0 x 1.0 m units and 0.5 x 0.5 m subunits. The area was photographed, surface colle cted, and then a 1.0 x 2.0 m area w as excavated in subunits and 10 c m levels. This area was rich in ceramic deposits and
80 the soil was much darker, almost black, representing the densest portion of the midden related to the house in the larger block excavation. Excavation within the residential area specifically provided fu rther information regarding the structure of domestic areas. The Nokugu Site (MT FX 06) A total of 12 linear trenches were excavated at MT FX 06 during the 2002 and 2003 field seasons to expose the vertical stratigraphy of the anthropogenic landscape featu res ( Figure 3 5 ) A single linear trench was excavated in 1993, designated Excavation Trench 1 (ET01) which along with data from elsewhere in the KSA was the basis for the cultural chronology of the study area ( Heckenberger 1996, 2005 ). Like ET01, all ETs, EUs, and TU s were hand excavated ETs were placed across plaza and road berms ( ETs 3, 5, 6, 8, and 9) across village peripheral ditches ( ETs 2, 4, 7, and 10), a nd road berms ( ETs 11, 12, and 13). V ertical profiles of select ETs (corresponding to adjacent EUs) and all TU s are presented here as the y further resolve periodization and physical configuration of the site Excavation Trenches ET01 was excavated in 1993 and is fully described elsewhere (Heckenberger 1996, 2005). The stra tigraphic sequence combined with radiometric data from ET01 provide d the first basic chronology for the site Stratum I was devoid of cultural artifacts and dates before A.D. 700 Stratum II is a series of continuous occupation layers that date from A.D. 9 00 1400. Stratum III is the reddish overburden which represents the initial construction of Ditch 1 and dates from A.D. 1400 1450. Stratum IV represents the infilling of Ditch 1 immediately after its construction. A date of A.D. 1590 was obtained from the middle of this stratum and a date of A.D. 1770 was obtained from the top of the stratum within two distinct macro stratigraphi c units (Heckenberger 1996:44).
81 Figure 3 5 MT FX 06 excavation and collection area locations
82 Twelve more ET s were excavated during 2002 2005 fieldwork confirming and expanding on the information collecte d from ET01. The most recent ET s were placed across the main plaza berm (ET08 09 and ET05), across the secondary plaza berm (ET03) and secondary plaza ditch (ET02 and ET13) across the village outer d i tch (ET04 and ET07), across the village inner ditch (ET06 and ET10), and finally across a ro adside berm (ET11 12). These ET s pr ovided a broad view of the site wide stratigraphy related to the construction of berms and ditches. Dates obtained from these profiles further refine the chronology of the site and provide the basic framework to understand social and material de velopments. Profiles and data are presented for select ET s from this group including ET 03, ET04, ET05, ET08 ET09 and ET10 ET 02 is a 1.0 x 12.0 m trench running north to south across the primary plaza peripheral ditch north of the main plaza This ditch extends to the bank of the Buriti ( Angahuku ) River in the northwest p ortion of the site. The profile reveals a very deeply excavated trench containing later period pottery in the basin which began to infill sometime after the construction of the ditch ( Figure 3 6 and Figure 3 8 ) Radiometric data obtained from charcoal samples from the bottom of this trench date the first episodes of infilling to roughly A.D. 1250. ET03 is a 1.0 x 7.0 m trench oriented north to south across the ber m of a smaller, perhaps earlier, plaza berm to the west of the main plaza. The stratigraphy of this t rench revealed a very dark soil containing high amounts of pottery ( Figure 3 9 and Figure 3 11 ) Various episodes of filling and other disturbance are also revealed in the trench profile. Four excavation units were placed on the south side of the trench and their contents are included in the overall ceramic analysis presented here.
83 ET04 is a 1.0 x 9.0 m trench oriented north to south on the far sout h end of the site bisecting the outer ditch. The profile of this trench revealed a relatively shallow ditch compared to the same ditch on the north side of the site ( Figure 3 6 and Figure 3 7 ) Radiometri c data from charcoal obtained from the bottom layer of the excavation date the first episodes of infilling of this ditch to A.D. 1250. ET05 is a 1.0 x 9.0 m trench oriented north to south across the intersection of the main plaza berm and the south curb of the main road on the east side of the site. The pr ofile of this trench revealed a deep layer of anthropogenic earth overlaying the original substratum ( Figure 3 10 and Figure 3 12 ) The anthropogenic lay er can be roughly divided into an upper and lower stratum but the soil transition is extremely gradual and diffuse An excavation unit was placed on the south end of this ET and its contents are included in the overall ceramic analysis ET08 and ET 09 are o riented north to south and east to west respectively and meet at the south and east ends. They are located at the intersection of the main plaza berm and the road berm which leads to the smaller plaza west of the main plaza Soil str atigraphy from each of these ET s matches closely with that of ET05 ( Figure 3 13 and Figure 3 14 ) Two excavation units were placed near t he intersection of these two ET s and their contents is included in the overall ceramic ana lysis Excavation Units EU s were excavated in all major landscape features including village peripheral ditches, plaza berms, and road berms. While the ET s provided a broad overview of the general stratigraphy of the village peripheral ditches a nd the plaz a/road berms, the EU s provided controlled samples of pottery and charcoal to aid the diachronic analysis of the ceramic industry at the site.
84 Figure 3 6 Profile drawing of ET 02 (top) and ET 04 (bottom)
85 Figure 3 7 East profile view of ET04 facing northeast with chaining pins at one meter intervals.
86 Figure 3 8 East wall profile view of ET 02 facing nor theast with chaining pins at one meter intervals
87 Figure 3 9 Profile drawing of ET 03 and its corresponding EUs Figure 3 10 Profile drawing of ET 05 EU 05 1 was excavated on the south end of this ET.
88 Figure 3 11 West wall profile view of ET03 facing northwest with chaining pins at one meter intervals.
89 Figure 3 12 East wall profile view of ET05 facing southeast with chaining pins at one meter intervals.
90 Figure 3 13 Profile drawing of ET 08 and ET 09 and their corresponding EUs
91 Figure 3 14 Profile view of the intersection of ET08 and ET09 showing the east wall profile of EU 8 1 (left) and south wall profile of EU 9 1 (right).
92 The excavation units placed at the ends and centers of excavation trenches across the peripheral ditches (EU04 and EU02 ) contained very little ceramic remains compared to those placed in plaza and road berms (EU03 1, EU03 2, EU03 3, EU03 4, EU05 1 EU08 1 and EU 09 1 ) and in some cases contained mixed stratigraphy In the ditch excavation units, pott ery was likely transported there by chance and in some cases is part of a reversed stratigraphy since these features were excavated and filled in well after the first occupations at the site. Like the somewhat problematic radiocarbon dates obtained from some ditch excavations, pottery in these was not used in the overall analysis. Excavation Units in the plaza and roadside berms were useful in the ceramic analysis and provided seemingly reliable dates as well The most beneficial of th ese were EU 03 1, EU03 2, EU 03 3, EU03 4 EU05 1 EU08 1 and EU 09 1 A total of 24.05 k g (n=1,644) of pottery was collected from these 7 EU s The de pth of ceramic remains and their increasing numbers through time before dropping off are consistent with the proposed occupation sequence of the site and accumulation of the plaza berms ( Figure 3 15 to Figure 3 20 ). Test Units Test Unit s were excavated through the center of all CAs Rather than describe each t est unit individually they are presented here as linear profile s stretching across each 100 x 100 meter collection area. Soil profiles recorded for each test unit reveal much about the overall site stratigraphy which fluctuates from the center of the site (the plaza) to its periphery (the residential areas) ( Figure 3 21 ) A total of 10.26 kg (n=1,107) of pot tery was collected in the 24 TUs from CA s 1 4. Pottery from these excavations was included in the overall ceramic analysis.
93 Figure 3 15 Excavation Unit 5 1 sherd count by depth Figure 3 16 Excavation Unit 5 1 sherd weight by depth
94 Figure 3 17 Excavation Unit 8 1 sherd count by depth Figure 3 18 Excavation Unit 8 1 sherd weight by depth
95 Figure 3 19 Excavation Unit 9 1 sherd count by depth Figure 3 20 Excavation Unit 9 1 sherd weight by depth
96 Figure 3 21 Test unit soil profiles for transects 1 4 Surface Collectio ns Surface collected areas provided the largest samples of pottery from both sites. Surface collection s contained all types and styles of pottery. A total of 102.68 k g (n=2,768) o f pottery was collected from CA s 1 4.
97 The Heulugiht Site (MT FX 13) The sit e of Heulugiht (MT FX 06) is six kilometers to the south of Nokugu (MT FX 06) and is connected by a near ly straight prehistoric road identified by low laying berms which form the side s of the road. The ber ms are 25 30 cm in height and do not contain arti facts on their surface once they extend past the main plaza. MT FX 13 was examined with the same methods as MT FX 16 to test horizontally and vertically with excavations and collection areas ( Figure 3 22 ) At MT FX 13 three 100 x 100 m collection areas were situated in three distinct segments of the village divided by roads leading out from the plaza center. A single ET was excavated to examine the stratigraphy of a segment of the plaza berm where it intersects with a road side ber m. Additionally, a single EU was excavated on the end of the ET with finer vertical control. Transe cts of 0.5 x 0.5 m TUs were excavated through the center of each of the three CAs Like MT FX 06 MT FX 13 occupied continuously from late in the first mille nnium A.D. until it was abandoned around the same time as Nokugu or earlier. Based on this chronology and the connecting roads, it is hypothesized that these two villages were part of a larger complex of organized villages. The pottery from MT FX 13 is dis tinctive from the pottery at the MT FX 06 and based on its overall appearance and technology was not likely made by the same potters. Unlike the pottery at MT FX 06 the assemblages from MT FX 13 do not contain the variety of forms and styles throug h time that are present at MT FX 06 However, some forms a nd styles are present at MT FX 13 that are not present at MT FX 06 and may be evidence of the mixture of different groups moving into the Upper Xingu in the Protohistoric period.
98 Excavation Trenches ET01 i s a 1.0 x 10.0 m trench oriented east to west across the southeast portion of the main plaza berm and the west curb of the main road. The pr ofile of this trench revealed a deep layer of anthropogenic earth overlaying the original substratum ( Figure 3 23 and F igure 3 24 ). An EU was placed on the east end of this ET and its contents are included in the overall ceramic analysis. Excavation Units A single 1 .0 x 0.5 m EU was placed at the east end of the ET and excavated in 10 cm levels This EU pr ovided the most reliable stratified remains from MT FX 13. The ceramic count from this excavation was somewhat low in comparison to EU s at MT FX 06 but consistent with the overall low count of ceramics at the MT FX 13 in general. Four other 1.0 x 1.0 m EU s were placed both within the plaza and in the presumed residential areas outside the plaza and these contained almost no cultural materials and were relatively shallow, less than 30 cm in most cases to culturally ster ile soil A total of 6.23 k g (n=397) of pottery was collected from the single EU at the end of the ET Test Units Once surface collections were complet ed 0.5 x 0.5 m TU s were placed on transects across the cen ter of each CA and spaced at 20 m intervals run ning no rth to south. A total of six TU s were excavated acros s each CA an d a total of 18 TU s were excavated at MT FX 13 The clearest difference between the TUs at MT FX 13 and those at MT FX 06 is the depth of the cultural dep osits. At MT FX 06 the terra p ret a is more than one meter deep in some areas while at MT FX 13 it is difficult to see even in shallow contexts and is much lighter in color and organic material A total of 7.14 k g (n=1,087) of pottery was collected in the TU s.
9 9 Figure 3 22 MT FX 13 excavation and collection area locations
100 Figure 3 23 North wall profile of ET01 at MT FX 13. The single EU was excavated at the east end of the trench.
101 F igure 3 24 Northeast view of ET01 at MT FX 13 after excavation The EU was excavated on the east end to the right.
102 Surface Collections Surface collections were conducted in three 100 x 100 m CAs at MT FX 1 3 CA s wer e situated in the southeast (CA01), southwest (CA02), and northwest (CA 03) portions of the site. Each was placed on the same site wide grid and oriented with at least one portion extending into the plaza berm though mainly sampling the residentia l areas W ithin each CA smaller 2.0 x 2.0 m CUs were placed 10 m apart on center beginning at the edge of each CA with the first and last CU extending one meter beyon d the 100 m outline of the CA The CU s were further divided into 1.0 x 1.0 m subunits. All artifacts collected from the surface of a subunit were grouped together for la ter analysis. A total of 121 CU s were completed for each CA. This resulted in a total of 29.56 k g (n=857) of pottery collected from MT FX 13 in the surface collection effort. Ot her Upper Xingu Sites Only limited excavations were conducted at other sites during the 2002 2005 fieldwork. Test unit s were excavated at Akagaht (MT FX 18) and non systematic surface collections were made at other sites while mapping including Akagaht (MT FX 18) Ipatse (MT FX 12), and Kuhi kugu (MT FX 11) The materials from these sites provide limited but important information regarding the overall pattern of ceramics in the Upper Xingu espec ially in the KSA and are presented here for their form and decorative attributes. These samples were collected from the surface of most sites and r epresent the last occupation at these sites. Finally, limited collections from the Museu Paraense Emlio Goeldi were analyzed along with their original field notes The se include collections from the work of Mario Sim es at sites MT AX 01, MT FX 01, MT FX 02, MT FX 03, MT FX 04, MT FX 05, and MT FX 09.
103 CHAPTER 4 POTTERY METHOD AND THEORY Introduction In the Upper Xingu the correlation between pottery and subsistence is already d ocumented based on the presence of the same food producing and food consuming po ttery vessels in late prehistoric, historic, and modern times and the overwhelming ethnohistoric evidence to support these observations ( Figure 4 1 ) Pottery remains therefore provide the best tangible material connection to food production and consum ption in the Upper Xingu past since p aleobotan ical remains are few and little studied, especially in the southern Amazon. Anthropogenic and anthropic land scapes both within and a round archaeological sites also provide some clues about the scale and organization of human changes to the land related to food production and consumption. T his lack of micro scale evid ence and the abundance of macro scale evidence are balanced by the ubiquitous presence of pottery throughout all known archaeological sites of the Upper Xingu. Working with the supposition that pottery played a signifi cant role in subsistence activities its advantage as a subject of study is its abil ity to reveal how food production and consumption might have varied through time in relation to its variable employment of pottery Examined from its earliest appearance to it s modern use Upper Xingu pottery demonstrates a continued presence of certain ve ssels from late prehistory into the present It also reveals the development of a preference for certain types of vessels and a tech nology related to these vessels This trend towards uniformity and standardization is perhaps not on the level of mass produ ction where standardization is from routinization but it is standardization on the level of a determined preference for a
104 certain type of vessel and a certain technology for building that vessel. While this type of trend within a ceramic industry is tradit ionally analyzed in relationship to the urbanization or modernization of a society here it is analyzed to understand how food production relates to the expansion and elaboration of a highly organized non industrialized society. Figure 4 1 Type 1 manioc processing vessels in use from 1884 (bottom, Steinen 1894:Tafel XV), 1950 (middle, Galvo 1953:51, Figure 9 ), and 2002 (top) where metal vessels have replaced all but a single Type 1 ceramic vessel in the background used for boiling manioc juice into kuigiku
105 Ceramic Studies in the Amazon Archaeologists in the Amazon have always employed ceramics in the job of creating chronology and connecting occupations and migrations across vast area s of Amazonia Howev er, s tudies that focus specifically on the technology of ceramics through time or across space in Amazonia are absent with a few notable exceptions (DeBoer 1975, DeBoer and Lathrap 1979, DeBoer 1986 Roosevelt 1996 ). Even these studies and other s like them (Meggers and Maranca 1980) focus on how ceramic assemblages are chosen and arranged for later use in developing site or regional chronologies and not specifically focused on the technology of pottery (DeBoer et al 1996 ; Meggers 1995b ). These studies are concerned with the important archaeological processes of identifying and seriating site occupations as well as ceramic assemblages within these occupations The most detailed research in Amazonia n ceramic studies most often employed an ethnoarchaeological approach Warren DeBoer and his colleagues used ethnoarchaeological studies to a ddress important problems encountered within ceramic studies These included issues such as differential longevity of ceramic vessels ske wing frequencies (DeBoer 1974), difficu lties in identifying assemblages accurately (DeBoer et al. 1996 ; Meggers 1995b ), and the tenuous nature of using ethnographic examples to project past behavior regarding ceramic production, use, and discard (DeBoer and Lathrap 1979). These studies some of them cautionary, add stability to those studies that seek to use ethnographic analogy in a direct historical approach. Those archaeologists most keenly aware of problems with the direct historical approach are perhaps those studying hunter g atherer societ ies because they are more likely to use ethnographic analogy though these studies also are few A rchaeologists have thus
106 taken this matter into their own hands, often recording pottery specific ethnographic information for themselves (Politis 2007). The r esult has been a small proliferation of ethnoarchaeological studies since at least the 1990 s in Amazonia (though little of it ceramic) adding to the pioneering ethnoarchaeology first conducted elsewhere (Longacre 1991) Like ethnographic data that specific ally addresses the technological variation in pottery similar data associated with prehistoric pottery is even less prominent in the Amazon even though technological attributes are often recorded during routine analysis A notable exception is Anna Roose velt's study of prehistoric ceramics at Parmana (1996). Though not in Amazonia per se her study, related to the Amazon based on early chronologies, notes variation in the technological attributes of pottery including temper type. Ceramic Analytical Method In Amazonia the long tradition of focusing on decorative motifs on pottery as indicators of change through time began very early (Hartt 1871) and was in part due to the lack of technology allowing for radiometric dating Donald Lathrap acknowledged the un due focus on stylistic attributes early in the span of Amazonia n archaeology. The archaeologist has been justly criticized for his preoccupation with pottery, but in the Amazon Basin the minutiae of ceramic style must carry the full burden of our attempts to study old population movements, old trade routes, and the boundaries of now extinct political units (Lathrap 1970:63) Although ceramic style must carry some of the burden, we now have the resources to expand minutiae to include techno logical considerations in attempting to understand both "old population movements" and in situ developments in regional variation. Increasingly, archaeological ceramics are seen as the record of human
107 agency showing both the intended and unintended consequ ences of human action (McCall 1999:18). Pottery is not just used as a stylistic indicator of migrations and diffusions. Pottery viewed as material culture emphasizes "the constitutive process of artifact manufacture, use, and discard amidst the influence of structural social processes (Ch ilton 1999:1) It is the social, economic, and political processes that drive the manufacture, use, and discard within a society. In this way ceramics are a product of social interaction and as such a key to understanding the social context in which they were produced. In the Amazon material culture deepens our understanding of other important developments concerning the association of d isparate cultures and their affinity with the land. Our understanding of the social proc esses that shape the manufacture of pottery is intimately linked to our understanding of the social patterns of those who produced and used pottery. Further, by refining the characterization of production w e can situate it more accurately along a spectrum spanning from individual level to regional level production or from household level to village level production. Through a technological understanding of pottery we acquaint ourselves better with their function and the process of construction. T his informa tion allows us to develop a model of economic and social patterning in the past and in the Amazon this model extends from the past into the ethnographic present. P ottery can o f course also be used as it has been always as an indicator of change and contin uity of movement and in situ development, and of trade and exchange Especially in the Amazon where we see widespread changes concurrently across a very expansive landscape, the introduction of new style horizons or new technological innovations into a re gion provide indications
108 hinting at the extensive social transformations that took place at various times in the prehistory of Amazon ia such as the successive waves of Arawak and Carib migrations and their impact on Amazonia as a whole for example (Hecken berger 2001 ; Lathrap 1970 ) Given the intense activity on either end of the roughly 1 000 year period of occupation of the sites included in this study, one may conclude that some clear changes should be reflected in the material culture of this group of p eople who on one end are enduring the massive movements of populations and environmental pressures and on the other end of time, the encapsulation of their villages as reflected in the building of presumed defensive ditches around those villages. Given th e presumed function of many pottery vessels in production and consumption and their role in the domestic economy we can comment on their place in a changing and ap parently intensifying economy. As research already suggests this period of time in the Upper Xingu was likely witness to a great intensification of manioc agriculture as noted from the massive secondary forest growth around these prehistoric villages. As well, it is surmised that chiefdoms were a part of this intensification, and the core social dynamics of current Xinguano peoples is evidence of th at growth and later decline (Heckenberger 2005 ; Dole 1978 ). Beyond Culture History Over the course of the last thirty years ceramicists have shifted their gaze from the pot to the potter behind the vess el. Binford (1977) and others showed that asking questions about past actions and processes in different ways yielded meaningful results Pots were viewed as more than decorated pieces of a fractured culture history and became material remains of technolog ical innovation that could be exploited to
109 reveal more abou t the people who made, used, discarded, and reused them (Braun 1983). Following seminal articles by W obst (1977) and others suggesting that perhaps we could use pots as more than temporal and cult ural diagnostics, archaeologists be gan to investigate the tech nological choices of potters to see how these choices reflected on their culture (Bronitsky and Hamer 1986). In th e 1990 s material culture studies flourished resulting in more anthropological ev aluations of archaeological materials (Chilton 1999). With new questions being asked of archaeological pottery, new ways of answering these questions also began to develop spawning more micro scale levels of examination such as compositional analysis (Arno ld 1992) In the 1980s a shift from s herd based analysis to vessel lot analysis placed the focus of study on the archaeological vessel and its component attributes rather than on collectio ns of attributes and typologies. Vessel lots were not universally ac cepted however, and as some found in database focused computational archaeology, both methods may benefit archaeological research ( Buck 1993; Duff 1996). Standardization Theory The subject of standardization and specialization has been addressed extensivel y in the study of pottery by Prudence Rice (Rice 1984a, 1984b, 1984c, 1991). According to Rice standardization refers to a relative degree of homogeneity or reduction in variability in the characteristics of the pottery or to the process of achieving that relative homogeneity (Rice 1991:268). She makes also a useful distinction between "manufacture" of pottery, or the actual act of fabricating ceramics, and "production", the social and economic organizational arrangements within which pottery manufacture is carried out (1996:173). The relationship between standardization and production is
110 further developed here as the relationship between those individuals who are producing the pots and those individuals who are controlling the labor that produces the pots In 1981 Rice developed a "trial model" of the ways in which labor relationships were encoded in pottery in order to analyze the process of intensification of production using as her proxy indicator of i ntensity and scale, standardization, or reduction in variability (Rice 1981, 1996:177). One problem with testing what she calls the "standardization hypothesis" is identifying a case study that fits the proper criterion for such a study. Since standardization is seen as a process it requires comparison betw een units, or assemblages, among which that process can "legitimately be hypothesized to operate (Rice 1996:178). This suggest s that testing this hypothesis would be possible with ceramics that exhibit clear developmental relationships over a long period of time, perhaps 400 500 years or longer and are with some certainty related, traditionally and economically to the same process For this reason, as outlined in the preceding chapter, the ceramics of the Xinguano tradition specifically th ose from sites in the KSA, represent a well suited case study to test the "standardization hypothesis" outside of an industrialized setting When combined with previous ethnographic and ethnoarchaeological research in the KSA describing the control of labor through soci al capital and political surplus we can also situate the degree of standardization with in the context of the local social and political structure (Heckenberger 2003) Another concern in testing this hypothesis is related to the study of ceramics in a more technological way. As Rice points out; T hrough time, changes in relative degrees of standardization and diversity in technological, formal, and decorative variables of pottery would reflect changing patterns of organization of production, some of which co uld be interpreted as specialization (Rice 1991:257)
111 This suggest s that the measure of change in pottery should ideally focus on technological as well as decorative attributes. The abundance of utilitarian wares in the Upper Xingu creates a situation wher e this is even more important in the absence of elaborately decorated prestige vessels. Clearly dimensional attributes carry the most weight since the size and shape of a vessel are often related to the function and performance of the vessel Vessel thick ness, form, and paste relate specifically to function and performance in many cases. C riticisms of dimensional attributes notwithstanding, if problems such as analysts not recognizing size categories are taken into account, especially in the application of statistical analysis, these dimensional attributes will still prove useful (Long acre et al. 1988; Sinopoli 1988) Of course with chemical characterization of ceramic pastes in archaeology gaining increasing prominence, the addition of technological and c ompositional attributes to dimensional attributes will alleviate this problem somewhat (Neff 1992). This will only be the case though if compositional attributes are used for the same purposes that decorative or stylistic attributes have been employed. The anthropology of standardization, or understanding the processes related to standardized behavior, will also help in the analysis by guiding technological questions towards anthropological themes In other words, the factors and causes that bring about sta ndardization must be reconciled with the measured variation in the actual pottery assemblages to maintain an anthropological framework. Additionally, factors that may have stunted or prevented standardization can also be examined and factored into the stud y and juxtaposed with physical analysis of the pottery assemblage. In this effort we must distinguish between standardizat ion in manufacturing technology and the
112 associa ted factors that may cause this. R eduction in variability associated with specializatio n may be from a decrease in the number of producers and the factors that may have caused this must be accounted for In addressing these issues Rice has argued for a precise distinction between specialization and intensification ( Rice 1991, 1996) To be cl ear, intensification is an economic process that involves increasing le vels of investment of all types; it incorporates specialization but specialization does not require intensification. Specialization refers more directly to the restrictedness of manufac turing activity, and may or may not be related to the level of sociopolitical activity in which it occurs, which could be at any level. Rice suggests that we reserve the term standardization "as a process through which uniformity in ceramics increases thro ugh time" and use the term uniformity when we refer to "the qualitative st ate or result of that process" (Rice 1996) In the Upper Xingu we are concerned with both uniformity and standardization. In a more recent evaluation of the concept of standardizati on in ceramics Roux (2003) found that standardization a s measured in metric attributes falls along a continuum of production intensity where low level intensity still shows standardization but with a higher coefficient of variation. This lead to the conclu sion that o nly in a high rate production situation do we have motor habits that transcend emic conceptions of standardization (Roux 2003:781). T he apparent lack of a "high rate" production situation in the Upper Xingu or intensification, does not result in a lack of reliable data The Upper Xingu coefficient of variation falls within the low level production found in other ar eas studied by Roux implying that the Upper Xingu potters were likely well aware of the standardization and chan ge in technology th ey achieved.
113 Determining Types V essel types for the purposes of this study refer to f ormal types that correspond to basic vessel f unctions based on ethnographic analogies described by Heckenberger (1996:71 73 2005 ). Typ ing vessels w h ether by form or funct ion is not without its since the pottery type is a generalization from many fragments and since there may be no individual example including all of its features, it is not Desp ite this criticism of types, in the Upper Xingu we stand on solid ground in knowing that types are not only recognized by local indigenous groups but also observed functional types in the present day subsistence practices of these same groups. In regard to using types to form chronologies we are also warned by Shepard who far as it is selected to serve as a means of outlining relative chronologies, a purpose that ha s no relation to the conditions of production or original functions of pottery (Shepard 1976:307). Again we find ourselves on solid ground because Upper Xingu forms are selected based on the functions of pottery both observed and confirmed through ethnogr aphic examples. Despite her early warnings to archaeologists about types, Shepard does relinquish a bit in finding that ny feature that changes in time, irrespective of its possible meaning or lack of meaning to the makers and users of pottery, is accept ed as a criterion of classification ( Shepard 1976:307 ) In this regard we again find a solid basis for the use of functional forms as classificatory types that change through time. Here we break from the common use of types based on decoration, temper, or other "abstractions". By assigning types that appear meaningful
114 across a wide span of time we can address long term changes and continuities by accounting for these "same" vessel s through time (and across space). While these basically functional types hel p identify the use of many vessels through the ethnographic examples, grouping them this way archaeologically has the risk of blurring the differences between various vessel subtypes found within archaeological contexts or types that are not emic However by having explicit types, one can argue for continuity in the presence of those vessel forms while also accounting for clear technologic al differences in the types through time In other words, d iscerning types better equips us to explain change and cont inuity in the technological aspects of vessels as well. Regardless of the potters level of awareness of standardized types as Hayden (1984) points out, emic types should be noted exclusively when they overlap with archaeologi cal interests or questions S ome studies have resulted in criticism of attempts at discerning cognitive prototypes rather than archaeological types, noting the absence of local taxonomies in some areas where they were created only when asked for by anthropologists (Kaplan 1985:357). A s difficult as it may be to find meaningful types in pottery analysis either emic or etic types the general principle of remaining anthropological should guide any study or as Anna Shepard puts it ; Despite the fact that much has been written on the subje ct of the artificiality of the pottery type, the underlying cause of artificiality, the habit of viewing pottery as a physical object abstracted from the essentials of its composition and the method of its manufacture, in other words, the persistent tenden cy to ignore the role of the potter, has not received the criticism it deserves. It is indeed strange that pottery should be studied without considering its rela tions to the people who made it (Shepard 1976:310)
115 Labor and Society T he potter remains a t the center of the process of standardization the discerning of types, and the fabrication and innovation of pottery technology Their relationship to the economic and political structure of a society is determined by the structure and organization of labor. Labor provides the connection between the potters and the social structure that influences manufacture. Labor thus becomes an important element in understanding the structure between potters and those they produce for. As Sassaman puts it nly the most p owerful institutions are able to fully constrain the chaos one might expect from unbridled crafting (Sassaman1998:93). In the Upper Xingu, this "powerful institution" is the complex form of chiefdom well documented by anthropologists in the region (Carnei ro 1970; Dole 1983; Heckenberger 2005). Control of crafting is control of labor and the power to command labor. This power amounts to a surplus of symbolic capital in the fo rm of labor control, labor that can produce ceramics and subsistence. In this way, t he material surplus is handled by the chief because he is a chief (who holds control over labor), not because the material surplus makes him a chief (Heckenberger 2003) Manufacture and Production Discussions of ceramic production are common in the archae ological literature of complex societies and urbanism, where studies of craft specialization and its relation to political complexity have been abundant (Sinop oli 1999, see also; Blackman et al. 1993, Blackman and Vida le 1992; Costin 1991; Costin et al. 19 89; Feinman 1985; Stein and Blackman 1993; Vidale 1989; Wailes 1996). It is suggested that perhaps urbanism should not be seen as a binary situation, either urban or not urban, but as a continuum with development falling within various degrees of urban (He ckenberger et al. 2008 ).
116 Using the term urban in this sense we can also broaden ideas about craft specialization and stretch out the continuum from individual potter to village level production, creating space for various levels between these that may not fall in to categories already established. This does not weaken the usefulness of the term urban but requires that those using it qualify their description of urban societies without using a simple checklist. Questions about the social complexity of the sou thern Amazon in ancient times are at the heart of archaeo logical research in this region and d efining in empirical archaeological terms what constitutes specialization, complex ity, urbanism and the networks that connected them are at the core of Upper Xing u research toda y (Heckenberger 2009; Heckenberger et al 2003, 2007 2 008 ). As in other parts of the world, ceramic production is a key indicator of many economic, social, and political elements of society. Levels of ceramic complexity may or may not be co rrelated with levels of social and political complexity. In the southern Amazon we can explore both issues separately before arriving at a conclusion regarding this relationship. The importance of specialized production is further emphasized by Longacre; F or over twenty years archaeologists have been concerned with the identification of the early appearance of specialized production and the implications of such a production mode for understanding the rise of complex forms of social and political organizatio n (Longacre 1999:44) We have an established lexicon for discussing ceramic pr oduction and its various levels and Prudence Rice provides the best definitions when discussing standardization particularly. Product standardization is viewed as a by product of specialization. Routinization is responsible for decreased variability in the products produced. In the Upper Xingu, current ethnographic models of pottery production have placed the specialization or 'Producer Specialization' (Rice 1991:263; Longacre 199 9:44) in the
117 hands of one village at least and possibly one household ( Heckenberger 2005 ). If we place today's production and prehistoric production along a continuum of craft specialization that varies from small scale household production to larger scale factory production, we would surely find ancient production and modern production on two ends of this continuum though likely not at the extremes, certainly not in the case of modern production. Variation within levels of pottery production does not fit nicely into Old World models that compartmentalize and separate these levels or scales A multi scalar continuum is needed to situate and understand the variation in Upper Xingu production from A.D. 700 1500 and into the present. If this finer scale is app lied it reveal s linkages between the scale and transformation of social complexity and the scale and transformation of ceramic technology over the course of Upper Xingu history Pottery production does not simply change because the scale of society changes but rather, the change in pottery production is an unintended consequence of the changes in political and social complexity, and these are in turn are an unintended consequence of overall changes in the economic pattern of prehi storic peoples ( McCall 199 9; Mills 1999). Ceramic Ethnoarchaeology This study is strengthened by ethnographic and ethnoarchaeological observations describing the use of pottery in subsistence processing in the Upper Xingu (Carneiro 1983; Dole 1960; Heckenberger 1996, 2005) Ceramic ethnoarchaeology studies provide data that can be applied and used to assess archaeological assemblages and their comparati ve validity. These same ceramic ethnoarchaeological studies also have limitations in their ability to predict the archaeological rec ord. This is true especially in
118 helping to project past usage of all ceramic forms, by studying present usage of similar forms, where different social and functional factors may be present (DeBoer 1974:341). E thnoarchaeological studies conducted over the p ast thirty years provide a wealth of information that encourages more of these studies For example, i n regard to assessing types in the archaeological record, ethnoarchaeological research has pointed out that the frequencies of certain forms in archaeolog ical assemblages can be affected by differential ceramic longevity (DeBoer 1974). Based on ethnoarchaeological studies, we can deduce that the relative proportional frequency of a ceramic type or form in the archaeological record is not only a function of its relative usage compared to other forms but to its durability and longevity before discard (David and Hennig 1972:20; DeBoer 1974). A central problem in any ceramic analysis is the determination of types that allow for quantification of pottery through time and across space. The pros and cons of determining types have long been debated among archaeologists (Shepard 1956; Rice 1987). Like types, assemblages are equally difficult to determine, especially where clear stratigraphic separation is not clear. P roceeding from the identification of an assemblage to the identification of types and frequencies within that assemblage is the most reliable way to begin a ceramic analysis. As Barbara Mills points out ; T he relative frequencies of vessel classes in trash mounds may be the most secure method of identifying what the constant values of the relative frequencies of use classes should be...this assemblage type may then be treated as a baseline and deviations from it investigated...critical to assessing patternin g in assemblages, however, is the use of functional types that are sensitive t o differences in vessel uselife (Mills 1989:143) Ceramic ethnoarchaeology can assist archaeologists in determining functional types and assessing the validity of the types they choose to assign to ceramics within an
119 assemblage. David and H ennig found that simple typologies based on criteria obvious to the uninitiated are likely to be adequate for most archaeological purposes" (David and Hennig 1972:28). Sherd based spatial analy sis also benefits from ethnoarchaeological studies. Archaeologist s are always faced with the choice of determining whether distributional patterns based on the count or weight of ceramic remains offer the most reliable information. This choice also affects the frequencies of vessels and ethnoarchaeological studies suggest that vessel weight varies directly with vessel uselife (DeBoer 1985). This is important when assessing the relative frequencies of different vessel types within an assemblage but is less i mportant when assessing the variations and similarities of the same vessel type between assemblages. Overall, ceramic studies benefit from ethnoarchaeological studies especially in drawing out anthropological questions. As Pritchard and Van Der Leeuw sugge st, these studies necessarily place the emphasis on the "ethnography of pottery" and that through this emphasis archaeological questions that focus on when and where are replaced with questions that focus on why and how (Pritchard and Van Der Leeuw 1984:6) In some cases the answers to these questions involve other aspects of anthropological research including linguistic. For example, i n the northern Amazon, Duin shows that pottery production is intimately linked to women and their physiology both linguisti cally and through oral tradition (Duin 2000:55). Ethnoarchaeological studies often lead researchers to attempt to replicate observations from the field or from historic accounts. This le ads to experimental archaeology and the development of studies that sp ecifically address technological
120 variation in pottery (Schiffer et al. 1994; Skibo et al. 1989). Placing priorities on performance characteristics in ex plaining technological change has also been done through experimental archaeology (Schiffer and Skibo 19 87) Experimental archaeology has successfully shown that our perceptions of pottery technology are sometimes flawed as in the case of experimental archaeology in Southwest North America showing that corrugated pots substantially degrades heat transfer ins tead of improves it as was once thought (Pierce 2005) In assessing the nature of modern ethnographic pottery assemblages, archaeologists benefit from a handful of studies devoted specifically to the study of pottery produced exclusively for outsiders, or tourist pottery This has been especially fruitful in the Amazon where Lathrap concluded that the potters involved in making tourist trade pottery have a general tendency to "favor the introduction and spread of exotic and unusual forms" making this potter y largely unsuitable for studying native ceramic assemblages in terms of their context within the general social and political system especially as direct historic analogs to pre tourist pottery (Lathrap 1976:206 207). Minimally, any ethnoarchaeological st udy should prevent an archaeological conceptualization of the past that is devoid of people. To this end; Artifacts 'speak' not so much because actors created them as 'texts' but because they are marked with the gestures and habits of their production and use, they are inscribed by the social processes involved in their creati on, employment, and abandonment ( McC all 1998:18) As ceramicists are fond of pointing out, pottery production is an additive process, an d through that process are left indications abou t the actions taken by potters and perhaps about the context in which those actions were conceived and implemented Put another
121 way very action is an instance of interpretation and representation ( Sassaman 1998:93 ). Accessing the material aspects of ind ividual interpretation and rep resentation" can be difficult however, and it is often the case that group consciousness and collective action are rare in d ispersed forms of production. I n her assessment of Aztec specialists Brumfiel finds that pecialist s were arrayed along a continuum, with the part time rural producers of utilitarian goods at the lower end of the scale and full time urban producers of elite goods at the upper end (Brumfiel 1998:151). Focusing on the "lower end" of the specialist contin uum presents even further obstacles since the routinized everyday traditions of domestic and communal realms are non ideological in that they are not recognized in the fields of political action However ; D d where cultural level traditional rea lms do lie within the field of 'actual relations of power'...with local produ cers being the agents of change ( Pauk etat and Emerson 1999:302 310) E thnoarchaeology highlights th e social mechanisms that cast the products of labor as non ideological symbols of ideologies of the past. Rather than simply viewing pottery as a reflection of the past it is viewed as a projection of the past or as Pauketat and Emerson put it, ymbols project relations, not reflect them and symbolic projection involves m ore than w (Pauketat and Emerson 1991). A further elaboration of ethoarchaeology or ceramic ethoarchaeology is the closely related field of c eramic ecolo gy (Arnold 1975 ). Seeing the context in which ceramics are created, used, and discarded as a complete system provides contextual advantages to understanding both the pottery and the system. Without being deterministic the study of pottery can benefit from an understanding of the natural and
122 social environment in which it was produced. For example, f rom his work with pottery making communities throughout Latin America, Arnold finds that potters tend to obtain their ceramic resources within a five kilometer radius (Arnold 1975:192). This type of i nformation is useful in assessing boundari es either social or natural that may limit or restrict pottery production. In some cases it may be the environment that prevents the craft from developing into a full time oc cupation for example (Arnold 1975:190). Ultimately, behavior is the root of ethnoarchaeological work, specifically, ceramic ethnoarchaeological work with the objective of "improved understanding of relationships between patterned human behavior and elemen ts of material culture that may be preserved in the archaeological record (Kramer 1985:77). But as Kramer also points out "o ne problem with ceramic ethnoarchaeology is that it almost never considers change and in fact is often conducted in a small area ov er a short period of time (Kramer 1985:92). Summary There are many factors to consider when examining and evaluating pottery assemblages and ceramic technology in archaeological contexts Social, economic, political, and e nvironmental factors all have a s ymbiotic role. Alt hough this study is focused on the te chnology of pottery, that technology is influenced by social factors which can be accessed through ethnoarchaeology Standardization is one important aspect of pottery manufacture that is linked intim ately to the social and political structure in which it occurs. Standardization is visible ethnographically and for this reason ethnoarchaeological data drives much of the research and theory regarding standardization. E thnoarchaeological studies provide i nformation that can be brought to bear on the archaeological record E thnoarchaeological studies focus on archaeological
123 and anthropological questions regarding the processes of manufacture, use, and discard and how these processes affect the formation of the archaeological record. Ethnoarchaeological studies also provide cautionary data guiding archaeological methods and techniques of examining the past Combining archaeological and ceramic ethnoarchaeological data adds the dimension of time and brings cer amic studies closer to an historical ceramic ecology.
124 CHAPTER 5 POTTERY ANALYSIS Introduction The following technofunctional analysis and discussion address two aspects of Upper Xingu pottery; technology and decoration Pottery form, function, and technology are as pects that distinguish ceramics both synchronically and diachronically and are used here for those purposes locally and are the main focus of this analysis. Decoration is used by archaeologists across the Amazon to connect or separate cultural units both t emporally and spatially. Decorative attributes on Upper Xingu ceramics also allow for basic chronology building, but decorative distinction across time and space are not as well understood in the Upper Xingu as they are across Amazonia in general especial ly in relation to e xamples south of the Xingu basin Both aspects of the analysis are based on pottery from several excavations and surface collections made at the Nokugu S ite (MT FX 06) and Heulugiht S ite (MT FX 13) during fieldwork conducted between 20 02 and 2005. Pottery from surface collections at other Upper Xingu sites within the KSA are also described for their relevance in the discussion of regional stylistic variability. P ottery from the Museu Paraense Emlio Goeldi (MPEG) collected in the Upper Xingu region by Mario Simes in the 1960 s was re evaluated in 2007 along with Simes original notes and also is included to benefit the regional analysis of Upper Xingu pottery as well as the technological study within the KSA Sample Selection and Method o f Analysis As is now common practice, a vessel unit of analysis, or vessel lot analysis, was selected to emphasize the technofunctional aspect of the overall study. A vessel lot
125 oriented technofunctional analysis allowed for the maximum amount of functiona l and technological information to be collected from each "vessel" In regard to stylistic or decorative attributes, the vessel lot analysis is also preferable because it minimizes bias created by unequal coverage of decoration on certain vessels accentuat ed by simple sherd counts and sherd based analysis However, in the Upper Xingu this is not as much of a problem since decorative attributes are mostly restricted to the rim portion of vessels though highly decorated rims are often on smaller vessels, thu s giving them overall smaller numbers in a sherd based analysis Vessel lot analysis is usually conducted with small defined assemblages from small excavations where reconstruction of vessels is much easier. Given the very large area covered by the surface collections and the relative distance between each collection unit within these areas a modified approach to the vessel lot analysis was undertaken V essel lot selection was conducted within each collection unit rather than the unmanageable and unfruitful task of connecting sherds across vast collection areas The vessel lots were selected by first identifying all rim sherds in a collection unit and proceeding to match those rim sherds with non rim sherds based on macro scopic attributes such as surface fin ish, thickness, and decoration. In most cases very few positive matches were made and vessel lots usually consisted of a single rim sherd. This resulted in a total of 1,142 vessels from all of the defined assemblages combined. Vessels from each assemblag e were analyzed for formal technological ( temper and metric attributes ), and decorative attributes. Vessel form as described previously, is derived from previous studies and ethnographic analogs but simplified som ewhat based on rim form. Metric attributes included orifice diameter, rim thick ness, and lip
126 thickness. Other attributes included temper percentages, determination of complete or incomplete firing, and core color. Additionally, decorative attributes were recorded as well as non systematic observati ons regarding the application of decorative motifs. D ecorative attributes were recorded for interior, rim, and exterior portions of each vessel and were also noted for non vessel ceramic items including pot stands. Upper Xingu ceramics do not have the dist inctive and elaborate decorative motifs that are found in other parts of the Amazon such as Maraj Santarm or the Central Amazon, for e xample. D istinctive, yet conservative, rim decorations do exist through time and across space in the Upper Xingu howe ver it is the rim form and its relation to function that is much more distinct and beneficial for study in this particular area of the Amazon. The general forms that constitute Upper Xingu pottery also do not have other distinctive body attributes such as carinations or handles, in most cases. To round out the study, s elect body sherds base sherds, and other ceramic objects such as pot stands are included in this analysis Base sherds are particularly useful given their profile and ability to distinguish b etween different base styles or types such as pedestal or flat bases allowing for inferences of direct heat or indirect heat use in cooking though much of this can be inferred from direct ethnographic observation Bases also record use wear, especially o n the interior in the form of eroded and pocked surfaces from use in hot processing of manioc. Due to the factors outlined here, the analysis of ceramics recovered from the Upper Xingu sites were a lmost entirely focused on macroscopic attribute s with the e xception of temper A ll measurements were taken using a standard caliper with sub
127 centimeter accuracy However, a microscope was used to record and characterize the temper content of each vessel. Fresh breaks on rim sherds of each vessel were examined usin g a Zeiss Stemi SV6 8 50X power microscope set to 10X Each temper type observed was recorded on a sliding scale from zero to five based on the amount of temper visible in the microscopes field of view A score of zero was recorded if a temper type was no t present, a one was scored for amounts less than 10 percent of the paste, a two was scored for amounts from 10 20 percent of the paste, a three was scored for 20 30 percent of the paste, a four was scored for 30 40 percent of the paste, and a five was sco red for amounts greater than 40 percent of the paste. Analysis Groups/Assemblages The focus of this study on technological change requires clearly distinguished ceramic assemblages as discussed in previous sections The basic chronology for the Upper Xingu is already well defined without the benefit of pottery as presented previously, using radiometric and stratigraphic data from several excavations in the study area. This is represented by initial occupation of the sites from A.D. 700 1250 (essentially no n depositional, flat, village peripheral z ones, sampled with controlled test unit excavations), village elaboration, initial defensive ditch excavation and subsequent dit ch infilling from A.D. 1250 15 00 or later (depositional, plaza and road side berm depo sits, sampled with controlled excavation units ), and final site occupation and abandonment from roughly A.D. 15 00 17 7 0 (non depositional surface deposits, sampled with systematic surface collection). Because similar decorative and stylistic attributes occu r throughout most of the Upper Xingu past, they are not a good basis for well define d time periods and could not be used to separate ceramic assemblages with a typical seriation method The three analysis groups or assemblages, were chosen to
128 essentially divide the continuous tradition of pottery into early, middle, and late phases as a technique t o assess variation and change in a mostly continuous industry. That said, it was assumed that the proposed early and middle assemblages would contain a large amo unt of overlap especially in the two excavated contexts representing the two earliest group assemblages, and this is in fact reflected in the data presented below. Because of t he essentially non depositional nature of Upper Xingu sites and the continuous tradition of pottery the three unit s of analysis are a compromise between compressing and dividing the entire history of the study area This is a compromise because excavations in the plaza peripheral or residential, zones identified at least two distin ct strata above the sterile sub soil which are compressed into one assemblage representing the earliest phase of site occupation Excavations into plaza berms also identified at least two distinct strata which are compressed into one assemblage representin g the middle phase of site occupation. If left separate it is likely that the lower level in the plaza berms and the upper level in the plaza peripheral areas would be statistically similar in regard to their ceramic assemblages. While there is absolutely no evidence to s uggest that these stratigraphic separations represent different archaeological cultures or phases, they at least provide a point of departure for understanding any chang es or continuities through time and are testable at other sites At a m inimum the surface artifacts and the excavated artifacts represent the earliest and the latest occupations of these sites. E xcavations that took place across the site peripheral ditches were more complicated and represent, in some cases, reverse stratigrap hy near places that were excavated in the past and then filled
129 through natural deposition These excavations were not used in the ceramic analysis as they generally contained very small amounts of pottery that was in mixed contexts. Attribute Analysis Prev ious prelimi nary analysis of samples of these assemblage s demonstrated that attempting to reassemble once extant vessels by sorting portions did not produce positive results and that in most cases a single rim sherd represented a single vessel with only li mited fragmentary body sherds adding to its overall description. Given this factor and the desire to cover a broad horizontal area only rim sherds, base sherds, and adorno's were considered in the final attribute analysis. Like rim sherds, b ase sherds wer e counted as single vessels as they could rarely be matched to their rim counterpart. All ceramics collected from excavations and surface collection were cataloged to record count, weight and vessel portion minimally. Once the initial catalog was finished rim, base, and adorno fragments were isolated for detailed vessel attribute analysis. Each aspect of the attribute analysis is discussed below. Metric Attributes Several measurements were taken for each sherd including lip thickness, rim thickness (measu red three centimeters below the lip), and oral diameter (on rims representing at least 15 percent of the overall estimated diameter) Measurements were also taken for base sherd thickness. Additional metric attributes were recorded on particular vessel for ms, such as lip width for example, on vessels with thickened or flat everted rims. Metric attribute data was also recorded for various decorations including incision width, incision depth and punctate spacing, for example.
130 Temper T he primary method of ide ntifying and qua ntifying temper type and amount was by e xamination of sherd profiles on fresh breaks as described earlier. Four different types of temper were observed in all assemblages analyzed; cauxi (a riverine sponge processed by potters to remove o rganic matter and use the silicate content) cariap (a tree bark also processed before being added to ceramic paste) grit (s ometimes referred to as mineral or sand tempered) and grog (sometimes referred to as sherd tempered) Though all f our types of te mper were common among most vessels analyzed, their combination and amount (or recipe) varied and will be discussed later. To quantify the temper content, f or each sherd each temper type present was given a value between zero and five as described earlier The most abundant temper in all assemblages is cauxi followed by cariap and grit, with smaller amount s of grog found in just some samples. Despite the importance of cauxi in Amazonian pottery (it also appears in southeast North America) it has receiv ed very little detailed study other than measuring its presence or absence. Cauxi is a riverine sponge that cling s to tree branches during flooding When rivers drop in level during the dry s eason branches with clusters of spicules are collected. The drie d branches with cauxi covering them are then burned and the ash and burned sponge is mixed with clay as the prime temper in modern Upper Xingu ceramics. Given that high amounts of sponge spicule naturally accumulate on the river bank when water levels rec ede in the same way that marine debris, seaweed, or shell accumulates at the high tide line, in would not be unexpected to find some amount of cauxi in all pottery made from clay collected near river banks, precisely where Upper X ingu potters source their clay.
131 At least one Amazonian based study differentiates between Metania reticulata a species of sponge found in floodwater habitats, and Drulia uruguaiensis or Oncosclera navicella found on the rocky bottoms of Amazonian rivers and remarkable for its la rge spicules, high amount of silica, and low amount of spongine, or organic matter which binds the spicules (Gomes and Vega 1999:319). This distinction between at least two or three different types of c auxi could provide a better understanding of clay sou rces or possibly exchange patterns if used for sourcing clay and thus pottery. To date, these c auxi types are not well understood and difficult to distinguish. As an organic tempering agent c auxi is also a reliable source for radiometric dating but has b een little used for this as well (Evans and Meggers 1962). The thermal properties of c auxi are also significant in this study. Anna Shepard recognized early that the use of sponge spicule temper was a compromise. As an advantage it creates a light weight paste and therefore can be used in the construction of larger vessels without the risk of them collapsing under their own weight. Sponge spicule temper also reinforces against cross fracture b ut at the cost of some weakness; Temper that is platey like mica or acicular like sponge spicules affect the structure of paste because flat or elongated fragments are in part forced into parallel position in the forming and finishing processes. In low fired pottery they may then have a reinforcing effect against cross fracture but they cause weakness in their own plane (Shepard 1976:27). When used in low firing conditions fusion of the silica f rom the sponge is not possible. This means th at to enjoy the full advantages of using an organic based silica an intense firi n g strategy must be used. Even if this is achieved through an intense firing process the resultant product may not benefit from the amount of silica present. An as Shepard points out;
132 Changes in temper that have a favorable effect on the body in the low te silica of organic origin disseminated in the paste could have some fluxing effect, but chips of paste containing sponge spicules showed n o evidence of softening of the spicules when refired to 950 degrees Celsius (Shepard 1976:29). Rather than re firing Upper Xingu archaeological ceramics to ascertain the level of spicule fusion we can examine the core color and properties to understand the firing technique. Again, the Upper Xingu can benefit from ethnographic observation in knowing a priori that Upper Xingu vessels are open air fired during the dry season (Heckenberger 2005). Core Color All core colors were recorded using a Munsell Soil Co lor C hart. These colors were recorded after fresh breaks were made on each vessel In cases of incomplete firing colors were recorded for both the core and the margin. In cases of complete firing a single color was recorded for the core. Core color is an important attribute when considering firing technique. Firing technique is often reflective of intensity of production within a ny given ceramic industry (Rye 1981) It can also assist in determining level s of standardization or routine present within a pr oduction system. Decoration Decoration was recorded for each vessel when present. Several methods of decoration were observed in all assemblages analyzed. Incision was the most common form of decoration observed and metric attribute data was recorded for t hickness, depth, and spacing of incisions. So called thumbnail punctates (Oberg 1953) on the lip of rim sherds were recorded and their orientation, slanting left or right from an overhead view, was also recorded. Thumbnail punctates were first observed on prehistoric
133 pottery of the Upper Xingu in the 1940 s. Oberg provides an important observation on ceramics encountered in 1948 at a site near Jacar formerly occ upied by the Trumai in the 1880 s. Here he found pottery similar to that made by the Waur both o n the surface and in excavations. modern pottery. Some fragments showed likewise a corrugation (finger impression) whi ch we do not observe on an y modern piece" (Oberg 1953:9). Only rare instances of paint and slip were noted likely due to the eroded nature of many of the specimens ( Figure 5 1 ) Though painted decoration almost certainly existed, as it does on modern Upper Xingu pottery, the preservation of painted designs on prehistoric Upper Xingu pottery is rare Other forms of distinguishable embellishments on pottery include modeling usually in zoomorphic forms and these also are restricted to rim portions of vessels. Thus the main decorative attributes used for comparison in the Upper Xingu remain those on the rim. Though the techniques used to apply decoration are few, various symmetrical and asymmetrical line designs, thumbnail punctates, and appliqu designs are present. Decorative designs are also found on cylindrical and conical shaped pot stands ( undagi ) These are decorated on both the bottom and side of the cylindrical items in the form of incised lines and groups of single punctates In contrast to this prehistoric decorative mode both modern and historic pottery in the Upper Xingu contains no incision s, punctates, or other rim decoration. The exception is found in v arious modern vessels that have zoomorphic appen dages and in some ca ses handles as first recorded by Steinen ( Figure 5 2 and Figure 5 3 ).
134 Figure 5 1 Examples of painting and modeling methods on decorated body sherds and handles from MT FX 12
135 Form and Type A profile was drawn for each rim and base sherd representing a single vessel using a gauge Based on form, ea ch vessel was then assigned to a functional type category established during previous research in the KSA and based on modern ethnographic terminology for similar vess el forms (Heckenberger 2005:205, 209). Rim forms are the single most important attribute of this study and there are several ways to classify the rim forms found at the sites included here Dole first split the rim forms she found at several Upper Xingu sites and noted 11 different everted rim forms (Dole 1961 :405). She noted in her excavations at one Upper Xingu site that ually flari ng ones" (Dole 1961:406). The ceramic types or forms, used in the current study were firs t proposed by Heckenberger (1996:71 73) based on surface collections and limited excavations as well as extensive ethnographic research with both the Kuikuru and the Waur tribes of th e Upper Xingu. These types represent pottery forms that span the entire known history of the Upper Xingu from roughly A.D. 7 00 or earlier to the present. The addition of modern indigenous knowledge and ethnographic examples to our underst anding of pottery in the Upper Xingu also allows the addition of functional categories or forms, to the analysis that can be tracked as they change o r remain the same through time. This is especially important to this technofunctional study and its focus on the manioc processing vessel. This vessel is examined through each assemblage as it varies through time.
136 Figure 5 2 Zoomorphic vessel designs r ecorded by Steinen (1894 :Tafel XXIII )
137 Figure 5 3 Zoomorphic vessel designs recorded by Steinen (1894 :Tafel XXIV )
138 Vessels included in this study are limited to what Heckenberger referred to as Type IA Type IB Type II, Type III and Type IV vessels based on the amount of available data for each form. These represent the primary flat bottomed cooking vessels ( Type IA for manioc and Type IB for fish ), the most easily recognizable globular rim decorated vessels ( Type II and Type III) and griddles (Type IV) A l so included are pot stands ( undagi ) because they are indicative of certain cooking processes across the Amazon (such as long term boiling) are sometimes decorated, and are good indicators of domestic spaces within arc haeological sites. In the Upper Xingu equating vessel form with function is more easily done than in some regions given the modern ethnographic correlates to the prehistoric pottery in the KSA This has situated function as the primary mode of identifying pottery variation at least in the mos t recent studies within the Uppe r Xingu. T his study approaches the idea of function separately from the attribute analysis and uses form as the basis for discovering other technological transformation s through time This is particularly fruitful in the mai n vessel form s that exist throughout the archaeological record in the Upper Xingu. While f unction was likely similar, t he form of the rim changed while the form of the vessel rema ined quite similar through time Because follows nati ve classification of vessels based on function, and function is an essential part of this overall analysis, this study does not depart from the original taxonomy ; it merely narrows the focus to specific vessel forms The native classifications are useful i n placing the overall vessel forms into broad categories however, the y do omit some of the variation in rim form found within these categories. In other words, while the types
139 might be accurate in describing use, they are not necessarily accurate in descr ibing manufacture. Vessel Types The following forms, or types, were initially defined by Heckenberger as described above, based on ethnographic forms and information obtained from observations with the Kuikuru and Waur of the PIX ( Figure 5 4 ) Because the t ypes distinguished in this study are somewhat different than those established by Heckenberger they are identified by using Arabic num bers rather than Roman numerals. The Arabic numbered types are designated only for this analy sis and are not meant to replace the vessel type designations already established for the Upper Xingu. For analysis purposes Heckenberger's distinction of Type IA and Type IB is further accentuated. Here I refer to Type IA as Type 1 ( ahukugu ) Type IB and Type II as Type 2 ( atange ) Type 3 combines all vessels with straight, direct, or inverted rims. Type 4 vessels are griddles ( alato ) and Type 5 are not vessels but pot stands ( undagi ). A short description of the vessel types, their original descriptions, and general observations are presented below. Type 1 Vessels Type 1 vessels, or ahukugu are flat bottomed thick bodied vessels with gradually excurvate rims and slightly thickened lips (Heckenberger 1996:71 71) Type 1 vessels are rarely found with deco ration in archaeological contexts but modern equivalents are slipped and painted on both the interior and exterior likely affecting their porosity The largest ahukugu are those vessels associated with manioc cooking and processing ; both activities involv e the use of water Today their metal equivalent is used fo r cold processing but the ahukugu vessel remains t he preferred vessel for boiling the juice squeezed from the manioc during processing ( once cooked it is referred to as kuigiku ).
140 This is the only h ot processing that the vessel is used for and these vessels are found in every household and obtained from the Waur village. Archaeologically these vessels are also ubiquitous and though exhibiting variation across time and space as we will see below, the y are found everywhere. The replacement of these vessels by metal pots for processing manioc occurred only in the last 50 60 years. In 1950 Eduardo Galvo recorded the use of ahukugu in the Upper Xingu and his photographs from that era show the use of many ahukugu in the processing of manioc ( Galvo 1953:49 51, Figures 7 9) Karl von d en Steinen also recorded these enormous vessels in the Upper Xingu as early as 1884 (S teinen 1894, Tafel XV ). Type 2 Vessels Type 2 vessels are those vessels with folded rims. Type 2 designation includes all pots described by Heckenberger as Type IB and Type IIA. The distinction in these two types of folded rim vessels is mostly in size and this is directly related to use. They are combined into one type in this study as a blin d test based on the similarity of the rim finishes. Using only the rims of these vessels it would be impossible to distinguish them as two types of vessels other than by size which falls along a continuum Type IB are generally larger, thicker flat bottom ed vessels, used to cook fish while the Type IIA vessels are smaller, more globular in shape, with a flat pedestal bottom, and used for dry st o rage or as drinking containers. The differences between the Type IB and Type IIA do not always reveal themselves based on rim sherds alone However, these two types do distinguish themselves statistically based on other attributes as will be shown in the analysis Also, b oth of these vessel types almost always have similar decorative incision on the rim and exterior lip further blurring their distinction Of all the Type 2 vessels (IB and IIA) from
141 surface collected contexts (n=50), 64 percent (n =32) have parallel incisions on the everted rim and most are in combination with thumbnail pu nctates around the exterior edg e of the lip. Heckenberger also notes that "unlike Type IA pots, several small examples of Type IB pots were identified (Heckenberger 1996:72). Not only do small examples of Type IB pots e xist but small examples of Type IA pots also exist. This similarity in vessel size variation between Type IA and Type IB further complicates dividing these types based on size alone. Figure 5 4 Kuikuru outdoor cooking area showing Type 1 ahukugu vessel (large blackened v essel in background) and Type 2 atange vessel (blackened vessel on wire rack) surrounded by aluminum vessels. Globular metal vessels are mainly used for water transport while the low profile vessels are used for manioc processing.
142 Type 2 vessels are the mo st elaborately decorated pieces of the surface collected ceramics though the rims of modern equivalents are not decorated The combination of rim incisions on the top of the flat everted rim and thumbnail punc tates on the outer lip of these vessels is not found on any Type 1 vessels however the thumbnail punctates and incision can be found on some Type 3 vessels. The incision decorations va ry quite broadly both in their execution and their design. Incisions range from very wide, shallow strokes to very na rrow deep strokes, though they are almost always applied to wet or leather hard clay with the except ion of very late period vessels which have decoration motifs that are engraved on the rims after firing is completed The thumbnail punc tates are generally from left to right likely indicative of right handedness and the action of the potter spinning the pot with their left hand while forming the rim, and the punct ates, with their right hand Type 3 Vessels The Type 3 vessel designation combines those vessel s described by Heckenberger as Type IIIA and Type IIIB. These vessels have direct rims with little to no thickening. They are sometimes slightly incurving or constricted and have thumbnail punctates on the exterior lip and infrequently one or two parallel incised lines on the narrow top portion of the lip. These vessels have been almost completely replaced with Type 4 Vessels ( alato or griddles) Type 4 vessels are those described by Heckenberger as Type IV vessels. These are flat griddles that are not decorated and have very low upturned rims. These rims can vary in height and are more gradual or more angled where they leave the flat bottom portion of the griddle. In the present day Kuikuru village a Type 1 vessel t hat has
143 broken sides is sometimes broken further so that only the bottom portion of the Type 1 vessel remains and is used as a griddle. In this way, positive identification of griddles is reserved for those sherds found that show the intact rim and bottom portion of the griddle in profile. Flat base sherds with the proper thickness could be from either a Type 1 or Type 4 vessel, though in some cases the Type 1 base will have interior deterioration from continued use in boiling kuigiku Type 5 ( undagi or po t stands) Though not vessels, pot stands ( undagi ) are designated as a type to include them as a separately produced ceramic product that has distinct form and decoration across the Upper Xingu. The Nokugu Site (MT FX 06) A total of 8, 026 ceramic sherds we re collected from MT FX 06 between 2002 2005. This includes specimens collected from systematic surface collection, excavation units (1 x 1 m ), and test units ( 0 .5 x 0 .5 m ) placed across the site. By far the most ceramics collected were from the surface co llections with a total weight of 148.66 k g (n= 4,077). The excavation units yielded a total weight of 28.08 k g (n= 1,807 ) and test units yielded a total weight of 16.96 k g (n= 2,142 ). Based on the established chronology for the site and the dating of the accu mulation of the pl aza mounds the following data are presented as three assemblages with the surface collection (CA) data representing the most recent (Group 1) the excavation unit ( EU ) data representing the middle period (Group 2) and the test unit ( TU ) data representing the earliest data from the site (Group 3) Means for all metric attributes are presented in Table 5 1 along with average temper contents in Table 5 2 and corresponding t tests for correl ations between means presented in Table 5 3 Table 5 4 and Table 5 5 These tests show that the most
144 statistically significant differences are found in the lip thickness and ori fice diameter of Type 1 vessels between Group 1 and Group 3. Group 1 ( Surface Collection ) A total of 2,768 ceramics were collected from surface collection areas 1 4 Of this total 274 pieces were identified as rims Of the 274 rims 70 percent are Type 1 ( n=192), 26 percent are Type 2 (n=72), and 4 percent are Type 3 (n=10). Of these 274 rims only 90 were sufficient f or determining orifice diameter However, t hese 90 rims are almost identical to th e total 274 in regard to vessel t ype percentage Of the 90 r ims sufficient for measuring orifice diameter 71 percent are Type 1 (n=64), 24 percent are Type 2 (n=22), and 5 percent are Type 3 (n=4) Type 1 vessels from the surface are the most standardized vessel forms in the Upper Xingu ( Figure 5 10 Figure 5 11 Figure 5 12 Figure 5 13 Figure 5 14 Figure 5 15 and F igure 5 16 ). Type 1 vessels have an average lip thickness of 1.56 cm (n=192), an average rim thickness of 1.15 cm (n=192), and an average orifice of 54 .89 cm (n=64). However, the range for Type 1 vessel metric attributes is broad. Lip thickness ranges from 0 .6 to 3.3 cm, ri m thickness ranges from .41 to 2.68, and orifice diameter ranges from 7 to 90 cm. Despite this wide range, Type 1 vessels still cluster and mostly separate from Type 2 and Type 3 vessels ( Figure 5 6 ). The Type 2 ve ssels that cluster with the Type 1 vessels represent the fish cooking vessels ( atange ) and are what Heckenberger referred to as Type IB. Although their rim form is similar to Type 2 vessels they may be flat bottomed and are usually much larger in size. The occurrence of extremely curved body sherds typical of globular vessels is very low in all surface assemblages This fact combined with ethnographic data lead to the conclusion that many vessels are flat bottomed at least late in the prehis toric record of the Upper Xingu.
145 Type 2 v essels from the surface are somewhat more variable in their rim form ( Figure 5 17 Figure 5 18 Figure 5 19 Figure 5 20 Figure 5 21 Figure 5 22 Figure 5 23 ). Type 2 vessels have an average lip thickness of 0 .94 cm (n=72) an average rim thickness of 0 .87 cm (n=72), and an average orifice diameter of 32 cm (n= 4). Like Type 1 vessels, the range for Type 2 vessel metric attributes is somewhat broad. Lip thickness ranges from 0 .4 to 2.9 cm, rim thickness ranges from 0 .4 to 1.72 cm and orifice diameter ranges from 12 to 74 cm. Like Type 1 vessels, Type 2 vessels also cluster together when considered by their thickness to orifice ratio ( Figure 5 6 ). Type 3 vessels are few in number and are generally small ( Figure 5 21 Figure 5 22 ) Type 3 vessels from the surface have an average lip thickness of .94 cm (n=10), an average rim thickness of .80 cm (n=10), and an average orifice diameter of 25 cm (n=4). Lip thickness ranges from .46 to 2.14 cm, rim thickness ranges from .4 to 1.1 cm, and orifice diameter ranges from 12 to 40 cm. Type 3 vessels cluster together and with the smaller Type 2 vessels ( Figure 5 6 ). Temper statistics for surface collected vessels were calculated using a ll 274 rims ( Table 5 2 ) They are presented in terms of percent of total temper present. Calculated with all vessels (n=274) the average temper composition is 67 percent cauxi 15 percent cariap 13 percent grit, and 5 percent g rog. When broken down into vessel types the statistics change somewhat dramatically though this may be due to small sample size. Type 1 vessels (n=192) are 65 percent cauxi 16 percent cariap 14 percent grit, and 5 percent grog. Type 2 vessels (n=72) ar e 72 percent cauxi 11 percent cariap 13 percent grit, and 4 percent grog. Type 3 (n=10) vessels are 76 percent cauxi 17 percent cariap 7 percent grit, and they contained no grog.
146 Surface treatment for all vessels included smoothing with only occasi onal vessels showing burnishing. Type 1 vessels had the most exterior surface paint and slip (n=21) as well as interior orange slip (n=5) and only a single rim still showed signs of interior black paint. Type 2 vessels had no interior paint or slip and onl y some exterior red paint (n=4). However, Type 2 vessels have the highest frequency of decoration including rim incision with parallel line motifs, thumbnail punctates, and rare chevron motifs on very large vessels with broad, flat, handle like rims ( Figure 5 5 ). Figure 5 5 Type 2 rim with chevron design from surface of MT FX 06. Type 3 vessels showed exterior red paint on only a single vessel. Of the 274 rims analyzed 64 percent (n=176) were completely fired showing no core and the remaining 36 percent (n=98) were incompletely fired with variably colored cores and margins. Of the non vessel rims only one Type 4 griddle was identified Besides the clustering of
147 diameter to thickne ss ratios ( Figure 5 6 ), t he clearest pattern in the overall distribution of pottery in the surface collection areas is the high amount of ceramic remains found on and around the plaza berm Outside the plaza berm pottery does not generally cluster by type ( Figure 5 7 to Figure 5 9 ). Figure 5 6 Lip thickness and diameter on all surface collected vessels from MT FX 06
148 Figure 5 7 Distribution of Type 1 vessels in surface collection areas at MT FX 06.
149 Figure 5 8 Distribution of Type 2 vessels in surface collection areas at MT FX 06
150 Figure 5 9 Distribution of Type 3 vessels in surface collection areas at MT FX 06.
151 Table 5 1 Metric attribute means for all vessels in all groups at MT FX 06. Assemblage Measurement n Mean Min Max STDEV CV Type 1 Group 1 Lip Thickness 192 1.56 0.60 3.30 0.49 0.31 Rim Thickness 192 1.15 0.41 2.68 0.37 0.32 Orifice Diameter 64 54.89 7.00 90.00 17.75 0.32 Type 1 Group 2 Lip Thickness 37 1.31 0.55 2.24 0.42 0.32 Rim Thickness 37 0.95 0.30 1.81 0.30 0.32 Orifice Diameter 29 44.12 17.00 64.00 12.97 0.29 Type 1 Group 3 Lip Thickness 17 1.27 0.67 2.17 0.41 0.32 Rim Thickness 17 1.05 0.77 1.33 0.16 0.16 Orifice Diameter 6 49.66 30. 00 60.00 12.01 0.24 Type 2 Group 1 Lip Thickness 72 0.94 0.40 2.90 0.42 0.45 Rim Thickness 72 0.86 0.40 1.72 0.34 0.39 Orifice Diameter 22 32.46 12.00 74.00 20.95 0.64 Type 2 Group 2 Lip Thickness 50 0.80 0.45 1.40 0.25 0.31 Rim T hickness 50 0.78 0.36 1.22 0.19 0.25 Orifice Diameter 33 25.87 11.00 55.00 11.51 0.44 Type 2 Group 3 Lip Thickness 14 0.81 0.44 1.59 0.29 0.36 Rim Thickness 14 0.74 0.16 1.01 0.28 0.38 Orifice Diameter 5 23.40 12.00 36.00 9.04 0.38 Type 3 Group 1 Lip Thickness 10 0.94 0.46 2.14 0.54 0.58 Rim Thickness 10 0.79 0.40 1.10 0.20 0.25 Orifice Diameter 4 24.85 12.00 40.00 11.48 0.46 Type 3 Group 2 Lip Thickness 15 0.70 0.50 0.88 0.10 0.14 Rim Thickness 15 0.68 0.11 1.15 0. 23 0.34 Orifice Diameter 13 16.50 9.00 26.00 6.31 0.38 Type 3 Group 3 Lip Thickness 7 0.82 0.56 1.16 0.23 0.29 Rim Thickness 7 0.59 0.45 0.90 0.15 0.26 Orifice Diameter 5 11.60 9.00 20.00 4.77 0.41
152 Table 5 2 Average temper content for all vessels in all groups at MT FX 06. Type 1 Type 2 Type 3 All Types Assemblage Temper n Mean n Mean n Mean n Mean Group 1 Cauixi 192 65.00 72 72.00 10 76.00 274 67.00 Cariape 192 16.00 72 11.0 0 10 17.00 274 15.00 Grit 192 14.00 72 13.00 10 7.00 274 13.00 Grog 192 5.00 72 4.00 10 0.00 274 5.00 Group 2 Cauixi 37 54.00 50 69.00 15 70.00 193 61.00 Cariape 37 20.00 50 11.00 15 12.00 193 18.00 Grit 37 21.00 50 13.00 15 18.00 193 1 7.00 Grog 37 5.00 50 6.00 15 0.00 193 4.00 Group 3 Cauixi 17 55.00 14 57.00 7 62.00 38 57.00 Cariape 17 22.00 14 20.00 7 10.00 38 19.00 Grit 17 22.00 14 23.00 7 18.00 38 21.00 Grog 17 2.00 14 0.00 7 10.00 38 3.00
153 Table 5 3 T test for equality of means for lip thickness measurements at MT FX 06. 95% confidence interval of the difference t df two tailed P value mean difference standard error of difference lower upper sta tistical significance Lip Thickness Type 1 Group 1 vs. Group 2 0.5783 57 0.5654 0.09 0.156 0.2217 0.4017 not statistically significant Group 1 vs. Group 3 3.4985 63 0.0009 0.37 0.106 0.1587 0.5813 extremely statistically significant Group 2 vs Group 3 1.3902 12 0.1897 0.28 0.201 0.1588 0.7188 not statistically significant Lip Thickness Type 2 Group 1 vs. Group 2 1.3259 19 0.2006 0.25 0.189 0.1446 0.6446 not statistically significant Group 1 vs. Group 3 2.1759 20 0.0417 0 .37 0.170 0.0153 0.7247 statistically significant Group 2 vs. Group 3 0.8139 9 0.4367 0.12 0.147 0.2135 0.4535 not statistically significant Lip Thickness Type 3 Group 1 vs. Group 2 not sufficient data Group 1 vs. Group 3 not sufficient data Group 2 vs. Group 3 not sufficient data
154 Table 5 4 T test for equality of means for rim thickness measurements at MT FX 06. 95% confidence i nterval of the difference t df two tailed P value mean difference standard error of difference lower upper statistical significance Rim Thickness Type 1 Group 1 vs. Group 2 0.3925 57 0.6962 0.07 0.178 0.4271 0.2871 not statistica lly significant Group 1 vs. Group 3 1.2415 63 0.2190 0.15 0.121 0.0914 0.3914 not statistically significant Group 2 vs. Group 3 1.1248 12 0.2827 0.22 0.196 0.2062 0.6462 not statistically significant Rim Thickness Type 2 Group 1 vs. Group 2 1.9256 19 0.0692 0.29 0.151 0.0252 0.6052 not quite statistically significant Group 1 vs. Group 3 1.3198 20 0.2018 0.17 0.129 0.0987 0.4387 not statistically significant Group 2 vs. Group 3 0.9525 9 0.3657 0.12 0.126 0.4050 0.1650 not statis tically significant Rim Thickness Type 3 Group 1 vs. Group 2 not sufficient data Group 1 vs. Group 3 not sufficient data Group 2 vs. Group 3 not sufficient data
155 Table 5 5 T test for equality of means for orifice diameter measurements at MT FX 06. 95% confidence interval of the difference t df two tailed P value mean difference standard error of difference lower upper statistic al significance Orifice Diameter Type 1 Group 1 vs. Group 2 0.4068 27 0.6873 3.56 8.750 21.5142 14.3942 not statistically significant Group 1 vs. Group 3 2.2835 31 0.0294 15.44 6.762 1.6494 29.2306 statistically significant Group 2 vs. Group 3 1.7694 10 0.1073 19.00 10.738 4.9260 42.9260 not statistically significant Orifice Diameter Type 2 Group 1 vs. Group 2 0.5179 7 0.6205 9.25 17.860 32.9810 51.4810 not statistically significant Group 1 vs. Group 3 not sufficient data Group 2 vs. Group 3 not sufficient data Orifice Diameter Type 3 Group 1 vs. Group 2 not sufficient data Group 1 vs. Group 3 not sufficient data Group 2 vs. Group 3 not sufficient data
156 Figure 5 10 Type 1 rim profiles (700's) from the MT FX 06 surface collection assemblage. Figure 5 11 Type 1 profiles (800's) fr om the MT FX 06 surface collection assemblage.
157 Figure 5 12 Type 1 profiles (900's) from the MT FX 06 surface collection assemblage. Figure 5 13 Type 1 p rofiles (1000's) from the MT FX 06 surface collection assemblage.
158 Figure 5 14 Type 1 profiles (1100's) from the MT FX 06 surface collection assemblage. Figure 5 15 Type 1 rim profiles (1200's) from the MT FX 06 surface collection
159 F igure 5 16 Exterior view of Type 1 rims from the MT FX 06 surface collection
160 Figure 5 17 Type 2 profiles from surface collection (700's) from the MT FX 06 surface collection assemblage.
161 Figure 5 18 Type 2 profiles from surface collection (800's) from the MT FX 06 surface coll ection assemblage.
162 Figure 5 19 Type 2 profiles from surface collection (900's) from the MT FX 06 surface collection assemblage. Figure 5 20 Type 2 prof iles from surface collection (1000's) from the MT FX 06 surface collection assemblage.
163 Figure 5 21 Type 2 and Type 3 profiles from surface collection (1100's) from the MT FX 06 surface collection assembla ge.
164 Figure 5 22 Type 2 and Type 3 profiles from surface collection (1200's) from the MT FX 06 surface collection assemblage.
165 Figure 5 23 Type 2 rims fro m the MT FX 06 surface collection assemblage.
166 Figure 5 24 Type 4 (griddle, or alato ) rim profiles from the MT FX 06 surface collection and EUs (top two examples 67 1 and 125 2).
167 Group 2 ( Plaza Berm Excav ations ) A total of 1,807 ceramics were collected from all excavation units. Of this total 193 pieces were identified as rims ( Figure 5 17 to Figure 5 23 ) Of the 193 rims only 102 could be accurately assi gned to a vessel type. Of the 102 rims 36 percent are Type 1 (n=37), 49 percent are Type 2 (n=50), and 15 percent are Type 3 (n=15). Of these 102 rims only 75 were sufficient for determi ning orifice diameter. These 75 rims are almost identica l statisticall y to the total 102 in the vessel t ype breakdown. Of the 75 rims sufficient for measuring orifice diam eter 39 percent are Type 1 (n=29), 44 percent are Type 2 (n=33 ), and 17 percent are Type 3 (n=13 ). Type 1 vessels from the unit excavations are less standa rdized in form than those from Group 1 but more so than Group 3 ( Figure 5 28 ). Type 1 vessels have an average lip thickness of 1. 32 cm (n= 37 ), an average rim thickness of .95 cm (n= 37 ), and an average orifice of 44 cm (n= 29 ) One of the best examples of these smaller versions of the modern day vessel was excavated from ET03 ( Figure 5 26 ). T he range for Type 1 vessel metric attributes is somewhat narrower than for surface collected Type 1 vessels Lip thickn ess ranges from 55 to 2.24 cm, rim thickness ranges from 3 to 1.81 and orifice diameter ranges from 1 7 to 64 cm. Type 1 vessels generally cluster together and mostly separate from Type 2 and Type 3 vessels but not as distinctly as in the surface assembl age ( Figure 5 25 ) Type 2 vessels from the unit excavations are somewhat variable in their form ( Figure 5 29 and Figure 5 27 ). Type 2 vessels have an average lip thickness of .90 cm (n=50), an average rim thickness of .78 cm (n=50), and an average orifice diameter of 26 cm (n=33). Like Type 1 vessels, the range for Type 2 vessel metric attributes is narrower in the unit excavations. Lip thickness ranges from .45 to 1.4 cm, rim thi ckness
168 ranges from .36 to 1.22, and orifice diameter ranges from 11 to 55 cm. Like Type 1 vessels, Type 2 vessels also generally cluster together when considered by their thickness to orifice ratio, however this is much less distinctive than in the surface assemblage ( Figure 5 25 ). Type 3 vessels from the unit excavations are extremely variable in their form and usually quite small ( Figure 5 29 ). Type 3 vessels have an average lip thickness of .7 cm (n=15) an average rim thickness of .68 cm (n=15), and an average orifice diameter of 16 cm (n=13). Lip thickness ranges from .5 to .88 cm, rim thickness ranges from .11 to 1.15 cm, and orifice diameter ranges from 9 to 26 cm. Type 3 vessels cluster together and with the smaller Type 2 vessels but unlike other vessel types do not consistently increase in orifice diameter as they increase in thickness ( Figure 5 25 ). Temper statistics for unit excavation vessels were calculated using all 1 93 rims. They are presented in terms of percent of total temper present. Calculated with all vessels (n=193) the average temper composition is 61 percent cauxi 18 percent cariap 17 percent grit, and 4 percent grog. When broken down into vessel types th e statistics change somewhat dramatically though this may be due to small sample size. Type 1 vessels (n=37) are 54 percent cauxi 21 percent grit, 20 percent cariap and 5 percent grog. Type 2 vessels (n=50) are 69 percent cauxi 13 percent grit, 11 pe rcent cariap and 6 percent grog. Type 3 (n=15) vessels are 70 percent cauxi 18 percent grit, 12 percent cariap and they contained no grog. Surface treatment for all vessels included smoothing with only occasional vessels showing burnishing. All types of vessels were represented by more surface paint and slip than the surface collected vessels but his is likely a product of preservation. Type 1
169 vessels had the most exterior surface paint and slip (n=12) as well as interior black paint (n=4) and interio r orange slip (n=3). Type 2 vessels have several examples of exterior red paint (n=6), exterior black paint (n=4) and orange slip (n=4). Type 2 vessels also have several examples of interior red paint (n=4), black paint (n=2), and orange slip (n=5). Type 3 vessels have examples of exterior red paint (n=6) and orange slip (n=2) with one example each of interior red paint, black paint, and orange slip. Of the 193 rims analyzed 72 percent (n=139) were completely fired showing no core, 26 percent (n=51) were in completely fired with variably colored cores and margins, and 2 percent (n=3) were too fragmentary to determine firing. Both the completely fired and the incompletely fired pieces have variably colored profiles. Figure 5 25 Lip thickness and diameter on all Group 2 vessels from MT FX 06
170 Figure 5 26 Nearly complete Type 1 flat bottom vessel from EU 3 1. Figure 5 27 Type 2 rims from EUs at MT FX 06.
171 Figure 5 28 Type 1 rim profiles from EUs at MT FX 06. Figure 5 29 Type 2 and Type 3 rim profiles from EUs at MT FX 06.
172 Group 3 (Plaza Peripheral Excavations ) A total of 2,142 ceramics were collected from all test unit excavations. Of this total 38 pieces were identified as rims. Of the 38 rims all could be accurately assigned to a vessel type. Of the 38 rims 45 percent are Type 1 (n= 17 ), 37 percent are Type 2 (n= 14 ), and 18 percent are Type 3 (n= 7 ). Of these 38 rims only 16 were sufficient for determining orifice diameter. These 16 rims are somewhat skewed statistically compared to the total 38 in the vessel t ype breakdown likely due to the small sample size Of the 16 rims sufficient for measuring orifice diameter 38 percent are Type 1 (n= 6 ), 31 percent are Type 2 (n= 5 ), and 31 percent are Type 3 (n= 5 ). Type 1 vessels from the test unit excavations have an aver age lip thickness of 1.28 cm (n= 17), an average rim thickness of 1.05 cm (n= 17 ), and an average orifice of 50 cm (n= 6 ) ( Figure 5 31 ) The range for Type 1 vessel metric attributes is somewhat narrower than for surface collected Ty pe 1 vessels. Lip thickness ranges from 67 to 2.17 cm, rim thickness ranges from .77 to 1.33 and orifice diameter ranges from 30 to 60 cm. Type 1 vessels generally cluster together and mostly separate from Type 2 and Type 3 vessels but not as distinctly as in the surface assemblage and less than the unit assemblage ( Figure 5 30 ). Type 2 vessels from the test unit excavations have an average lip thickness of 81 cm (n= 14), an average rim thickness of .74 cm (n= 14 ), and an average orifice diameter of 23 cm (n= 5 ) ( Figure 5 32 ). Like Type 1 vessels, the range f or Type 2 vessel metric attributes is narrower in the test unit excavation s. Lip thickness ranges from .44 to 1.59 cm, rim thickness ra nges from .16 to 1.01 and orifice diameter ranges from 12 to 36 cm. Like Type 1 vessels, Type 2 vessels also generally cluster together when
173 considered by their thickness to orifice ratio, however this is much less distinctive than in the surface assembla ge and less than in the unit assemblage ( Figure 5 30 ). Type 3 vessels from the test unit excavations have an average lip thickness of 82 cm (n= 7 ), an average rim thickness of 59 cm (n= 7 ), and an average orifice d iameter of 12 cm (n= 5 ). Lip thickness ranges from .5 6 to 1 16 cm, rim thickness ranges from 45 to 90 cm, and orifice diameter ranges from 9 to 2 0 cm. Type 3 vessels cluster together and with the smaller Type 2 vessels but unlike other vessel types do not consistently increase in orifice diameter as they increase in thickness ( Figure 5 30 ) Temper statistics for unit excavation vessels were calculated using all 38 rims. They are presented in terms of percent of total temper presen t. Calculated with all vessels (n= 38 ) the average temper composition is 57 percent cauxi 21 percent grit, 19 percent cariap and 3 percent grog. When broken down into vessel types the statistics remain relatively consistent Type 1 vessels (n= 17) are 55 percent cauxi 22 percent grit, 22 percent cariap and 2 pe rcent grog. Type 2 vessels (n=14 ) are 57 percent cauxi 23 percent grit, 20 percent cariap and contain no grog. Type 3 (n= 7 ) vesse ls are 62 per cent cauxi 18 percent grit, 10 percent cariap and 10 percent grog. Surface treatment for all vessels included smoothing with only occasional vessels showing burnishing. Three Type 1 vessels have red paint present on the outside surface and only one has black paint on the inside surface. Of the 38 r ims analyzed 68 percent (n=26) were completely fired showing no core and 32 percent (n=12) were incompletely fired with variably colored cores and margins. Both the completely fired and the incompletely fired pieces have variably colored profiles.
174 Figure 5 30 Lip thickness and diameter for all Group 3 vessels at MT FX 06. Figure 5 31 Type 1 rim profiles from TUs at MT FX 06.
175 Figure 5 32 Type 2 rim profiles from TUs at MT FX 06. Figure 5 33 Type 3 rim profiles from TUs at MT FX 06.
176 Figure 5 34 Type 4 rim profiles from EUs at MT FX 06. Figure 5 35 Base sherd profiles from TUs at MT FX 06.
177 Residential Excavation Excavation within the residential zone just north of the plaza berm and east of the northern r oad was conducted in 2005 ( Figure 5 36 and Figure 5 39 ) A house outline (House 1) was identified within this excavation block based on the location of domestic features, post holes, and domestic pottery remains. Radiometric data obtained from wood charcoal (beta 272640, 510 40 BP) from a feature within the block excavation places the occupation of the house sometime between cal A.D. 1400 to 1460 or the peak of village elaboration and expansion. Pot stan ds, large Type 1 vessels, and other domestic ceramics are associated with this feature. This combination of data substantiates the assignment of the surface collected ceramics to the latest period of occupation at the site. The limited nature of the reside ntial excavation does not provide data sufficient to comment on residential population or density however, b ased on this and other subsurface features the layout of the house excavation ( Figure 5 36 and Figure 5 37 ) is consistent with the size and orientation of a modern Kuikuru house and related domestic areas ( Figure 5 38 ) To further supplement this data, an additional area was surface collected adjacent to t he plaza berm sout h of the house excavation. The surface collection confirmed another domestic area based on the presence of Type 4 vessels in association with Type 1 vessels. Voids in the surfac e collection area are likely the location of a second house (House 2). D istribu tions of pottery in the surface collected are a further characterize the use and disposal of pottery at the site ( Figure 5 39 to Figure 5 49 ). T he surface data show that the distribution of pottery across the site based on function is minimal and mainly restricted to the disposal of Type 4 ceramics.
178 Figure 5 36 B lock excavation at MT FX 06 showing proposed outline of House 1(dashed line) in the peripheral /residential zone north of the main plaza. The House 1 outline is based on the location of central house posts (Feature 9, between N 314 and N 315), wall posts and associated trench (Features 4 and 4a, between N 307 and N 309), and the interior cooking are a (Features 1a 1d). A single radiocarbon date obtained from wood charcoal in Feature 9 dates the likely occupation of the house to sometime around A.D. 1450. The rear house midden contained ceramics characteristic of the Late Ipavu Period and Protohistoric Period.
179 Figure 5 37 South facing view of the b lock excavation of House 1 at MT FX 06 Figure 5 38 Overview of Kuikuru village in 2002 showing newly co nstructed house frame and older houses arranged around the central plaza. The layout and size of this house is comparable to that hypothesized in the block excavation at MT FX 06.
180 Figure 5 39 Distributio n of Type 1 vessels in sub coll ection area and in relation to H ouse 1 excavation to the north
181 Figure 5 40 Distribution of Type 2 vessels in sub collection area and in re lation to H ouse 1 excavation to th e north
182 Figure 5 41 Distribution of Type 3 vessels in sub coll ection area and in relation to H ouse 1 excavation to the north
183 Figure 5 42 Distribution of Type 4 vessels in sub coll ection area and in relation to H ouse 1 excavation to the north
184 Figure 5 43 Hypothetical location of second house based on distribution of pottery in sub coll ection area and i n relation to H ouse 1 excavation to the north
185 Figure 5 44 Close up of sub collection area and hypothetical House 2 location in relation to road and plaza berm
186 Figure 5 45 Hypothetical H ouse 2 location and vessel type distribution Figure 5 46 Distribution of Type 1 vessels in sub collection area
187 Figure 5 47 Distribution of Type 2 vessels in sub collection area Figure 5 48 Distribution of Type 3 vessels in sub collection area
188 Figure 5 49 Distri bution of Type 4 vessels in sub collection are a The Heulugiht Site (MT FX 13) A total of 2,133 ceramic sherds were collected from MT FX 13 between 2002 2005 This includes specimens collected from systematic surface collection, excavation units (1 x 1 m ), and test units ( 0 .5 x 0 .5 m ) placed across the site. The most ceramics collected were from the surface collections with a total weight of 29.56 kg (n= 857 ). The excavation units yielded a total weight of 6.23 kg (n= 189 ) and test units yielded a total wei ght of 7.14 kg (n= 1,087 ). Based on the established chronology for the site (like MT FX 06) and the dating of the accumulation of the pl aza mounds the following data are presented as three assemblages or groups, with the surface data representing the most recent (Group 1) the excavation unit (EU) data representing the middle period (Group 2) and the test unit (TU) data representing the earliest data from the site (Group 3). Means for all metric attributes are presented in Table 5 6 along with average temper
189 contents in Table 5 7 and corresponding t tests for correlations between means presented in Table 5 8 Table 5 9 and Table 5 10 These tests show that the most statistically significant differences are found in the lip thickness and orifice diameter of Type 1 vessels between Group 1 and Group 3, the same significant difference as found at site MT FX 06. Additionally at MT FX 13, the difference between lip thickness in Type 2 vessels between Group 1 and Group 3 is also statistically significant. Just as at site MT FX 06, the low n values for most types within each group likely contributed to the lack of statistically significan t differences even though there are clearly visible trends. Group 1 ( Surface Collection ) A total of 857 ceramics weighing 29.56 k g were collected from all surface c ollection areas. Of this total 74 pieces were identified as rims. Of the 74 rims 74 percent are Type 1 (n= 55 ), 22 percent are Type 2 (n= 16 ), 3 percent are Type 3 (n= 2 ) and 1 vessel could not be typed Of these 74 rims only 30 were sufficient f or determining orifice diameter T hese 30 rims are slightly different statistically to the total 74 in the vessel t ype breakdown. Of the 30 rims sufficient for measuring orifice diam e ter 83 percent are Type 1 (n= 25 ), 14 percent are Type 2 (n= 4 ), and 3 percent are Type 3 (n= 1 ). Type 1 vessels from the surface have an average lip thickness of 1. 32 cm (n= 55 ), an average rim thickness of 1. 24 cm (n= 55 ), and an average orifice of 5 2 cm (n= 25 ) Two clearly different variations of Type 1 rims from MT FX 13 are visible. The most standardized form is the tapered variant ( Figure 5 51 ) and the less standardized form is more gradually flaring like those from MT FX 06 ( Figure 5 52 ). The range for Type 1 vessel metric attributes is somewhat wide Lip thickness ranges from .51 to 1.87 c m, rim thickness ranges from .49 to 2 .31 an d orifice diameter ranges from 12 to 8 0 cm. This
190 wide range is reflected in the lack of clustering of Type 1 vessels though they are still mostly separate from Type 2 and Type 3 vessels at least in their large size ( Figure 5 50 ). Type 2 vessels from the surface have an average lip thickness of 1.07 cm (n=16), an average rim thickness of .1.03 cm (n=16), and an average orifice diameter of 38 cm (n=4) ( Figure 5 53 Figure 5 56 and Figure 5 57 ). Like Type 1 vessels, the range for T ype 2 vessel metric attributes is somewhat broad. Lip thickness ranges from .58 to 2.1 cm, rim thickness ranges from .6 to 1.5, and orifice diameter ranges from 13 to 74 cm. Though based on a very small sample size, Type 2 vessels do not cluster together when considered by their thickness to orifice ratio ( Figure 5 50 ). Type 3 vessels from the surface have an average lip thickness of .73 cm (n=2), an average rim thickness of .74 cm (n=2), and an average orifice diameter of 17 cm (n=1) ( Figure 5 53 ) Lip thickness ranges from .71 to .76 cm, rim thickness ranges from .7 to .78 cm, and orifice di ameter on the single available vessel is 17. Type 3 vessels cluster together and with the smaller Type 2 vessels, though again, this is a very small sample size and these statistics are highly suspect ( Figure 5 50 ) Temper statistics for surface collected vessels were calculated using all 74 rims. They are presented in terms of percent of total temper present. Calculated with all vessels (n=74) the average temper composition is 67 percent cauxi 15 percent cariap 9 percent grit, and 9 percent grog. When broken down into vessel types the statistics change somewhat dramatically though this may be due to small sample size. Type 1 vessels (n=55) are 65 percent cauxi 17 percent cariap 9 percent grit, and 9 percent grog. Type 2 vessels (n=16) are 80 percent cauxi 10 percent grog, 6 percent
191 grit, and 4 percent cariap Type 3 (n=2) vessels are 57 percent cauxi 15 percent cariap 14 percent grit, and 14 percent grog. Surface treatment for all vessels included smoo thing with only four vessels showing evidence of scraping. Some Type 1 vessels have exterior surface paint or slip (n=2) but none on the interior and no evidence of interior black paint. Type 2 vessels had no interior paint or slip and only some exterior r ed paint (n=2). Type 3 vessels showed exterior red paint on only a single vessel. Of the 74 rims analyzed 65 percent (n=48) were completely fired showing no core and the remaining 35 percent (n=35) were incompletely fired with variably colored cores and ma rgins. Similarly, even the completely fired pieces have variably colored profiles. In addition to Type 4 and base sherd examples that are more variable than those found at MT FX 06 ( Figure 5 54 and Figure 5 55 ), MT FX 13 also has several examples of what may be related to the Araquinoid tradition found across the late prehistoric Amazon ( Figure 5 56 and Figure 5 57 ). These examples are decoratively differ ent, fired differently, and made of much darker clay than any pottery from MT FX 06 further accentuating the differences between these two sites and perhaps others in the Upper Xingu.
192 Figure 5 50 Lip thic kness and orifice diameter for Group 1 vessels from MT FX 13.
193 Table 5 6 Metric attribute means for all vessels in all groups at MT FX 13. Assemblage Measurement n Mean Min Max STDEV CV Type 1 Group 1 Lip Thickness 55 1.32 0.51 1.87 0.30 0.23 Rim Thickness 55 1.24 0.49 2.31 0.35 0.28 Orifice Diameter 25 52.44 12.00 80.00 16.23 0.30 Type 1 Group 2 Lip Thickness 4 1.23 0.84 1.54 0.31 0.25 Rim Thickness 4 1.31 0.98 1.46 0.22 0.17 Orifice Dia meter 4 56.00 42.00 74.00 16.40 0.29 Type 1 Group 3 Lip Thickness 10 0.95 0.67 1.80 0.35 0.36 Rim Thickness 10 1.09 0.81 1.95 0.36 0.33 Orifice Diameter 8 37.00 20.00 68.00 18.00 0.48 Type 2 Group 1 Lip Thickness 16 1.07 0.58 2.10 0 .39 0.36 Rim Thickness 16 1.03 0.60 1.50 0.30 0.29 Orifice Diameter 4 38.25 13.00 74.00 26.25 0.68 Type 2 Group 2 Lip Thickness 5 0.82 0.52 1.20 0.27 0.33 Rim Thickness 5 0.74 0.45 1.09 0.27 0.36 Orifice Diameter 5 29.00 12.00 60.00 26.90 0.93 Type 2 Group 3 Lip Thickness 6 0.70 0.45 1.14 0.22 0.32 Rim Thickness 6 0.86 0.71 1.00 0.14 0.16 Orifice Diameter 1 21.00 Type 3 Group 1 Lip Thickness 2 0.73 0.71 0.76 0.03 0.04 Rim Thickness 2 0.74 0.70 0.78 0.05 0. 07 Orifice Diameter 1 17.00 Type 3 Group 2 Lip Thickness 1 1.27 Rim Thickness 1 0.68 Orifice Diameter 1 32.00 * Type 3 Group 3 Lip Thickness 1 0.88 Rim Thickness 1 0.66 Orific e Diameter 1 20.00
194 Table 5 7 Average temper content for all vessels in all groups at MT FX 13. Type 1 Type 2 Type 3 All Types Assemblage Temper n Mean n Mean n Mean n Mean Group 1 Caui xi 55 65.00 16 80.00 2 57.00 73 67.00 Cariape 55 17.00 16 4.00 2 15.00 73 15.00 Grit 55 9.00 16 6.00 2 14.00 73 9.00 Grog 55 9.00 16 10.00 2 14.00 73 9.00 Group 2 Cauixi 4 54.00 5 100.00 1 0.00 10 64.00 Cariape 4 32.00 5 0.00 1 33.00 10 21.00 Grit 4 14.00 5 0.00 1 0.00 10 12.00 Grog 4 0.00 5 0.00 1 67.00 10 3.00 Group 3 Cauixi 10 58.00 6 62.00 1 62.00 17 60.00 Cariape 10 24.00 6 22.00 1 0.00 17 22.00 Grit 10 15.00 6 5.00 1 0.00 17 12.00 Grog 10 3.00 6 11.00 1 33.00 17 6.00
195 Table 5 8 T test for equality of means for lip thickness measurements at MT FX 13. 95% confidence interval of the difference t df two tailed P value mean difference standard erro r of difference lower upper statistical significance Lip Thickness Type 1 Group 1 vs. Group 2 0.5783 57 0.5654 0.09 0.156 0.2217 0.4017 not statistically significant Group 1 vs. Group 3 3.4985 63 0.0009 0.37 0.106 0.1587 0.5813 extremely statis tically significant Group 2 vs. Group 3 1.3902 12 0.1897 0.28 0.201 0.1588 0.7188 not statistically significant Lip Thickness Type 2 Group 1 vs. Group 2 1.3259 19 0.2006 0.25 0.189 0.1446 0.6446 not statistically significant Group 1 vs. Group 3 2.1759 20 0.0417 0.37 0.170 0.0153 0.7247 statistically significant Group 2 vs. Group 3 0.8139 9 0.4367 0.12 0.147 0.2135 0.4535 not statistically significant Lip Thickness Type 3 Group 1 vs. Group 2 not suf ficient data Group 1 vs. Group 3 not sufficient data Group 2 vs. Group 3 not sufficient data
196 Table 5 9 T test for equality of means for rim thickness measurements at MT FX 13. 95% confidence interval of the difference t df two tailed P value mean difference standard error of difference lower upper statistical significance Rim Thickness Type 1 Group 1 vs. Group 2 0.3925 57 0.6962 0.07 0.178 0.4271 0.2871 not statistically significant Group 1 vs. Group 3 1.2415 63 0.2190 0.15 0.121 0.0914 0.3914 not statistically significant Group 2 vs. Group 3 1.1248 12 0.2827 0.22 0.196 0.2062 0.6462 not statistically significant Rim Thickne ss Type 2 Group 1 vs. Group 2 1.9256 19 0.0692 0.29 0.151 0.0252 0.6052 not quite statistically significant Group 1 vs. Group 3 1.3198 20 0.2018 0.17 0.129 0.0987 0.4387 not statistically significant Group 2 vs. Group 3 0.9525 9 0.3657 0.12 0 .126 0.4050 0.1650 not statistically significant Rim Thickness Type 3 Group 1 vs. Group 2 not sufficient data Group 1 vs. Group 3 not sufficient data Group 2 vs. Group 3 not sufficient data
197 Table 5 10 T test for equality of means for lip orifice diameter measurements at MT FX 13. 95% confidence interval of the difference t df two tailed P value mean difference standard error of difference lower upper statistical significance Orifice Diameter Type 1 Group 1 vs. Group 2 0.4068 27 0.6873 3.56 8.750 21.5142 14.3942 not statistically significant Group 1 vs. Group 3 2.2835 31 0.0294 15.44 6.762 1.6494 29.2306 statistica lly significant Group 2 vs. Group 3 1.7694 10 0.1073 19.00 10.738 4.9260 42.9260 not statistically significant Orifice Diameter Type 2 Group 1 vs. Group 2 0.5179 7 0.6205 9.25 17.860 32.9810 51.4810 not statistically significant Gro up 1 vs. Group 3 not sufficient data Group 2 vs. Group 3 not sufficient data Orifice Diameter Type 3 Group 1 vs. Group 2 not sufficient data Group 1 vs. Group 3 not sufficient da ta Group 2 vs. Group 3 not sufficient data
198 Figure 5 51 Type 1 rim profiles (tapered variant) from surface collection at MT FX 13. Figure 5 52 Type 1 rim profiles (other variant) from surface collection at MT FX 13.
199 Figure 5 53 Type 2 and Type 3 rim profiles from surface collection at MT FX 13. Figure 5 54 Vessel base profiles from surface collection at MT FX 13.
200 Figure 5 55 Type 4 rim profiles from surface collection at MT FX 13. Figure 5 56 Type 2 adorno (Araquinoid like) from surface collection at MT FX 13.
201 Figure 5 57 Type 2 rim from surface collection at MT FX 13. Group 2 ( Plaza Berm ) A total of 189 ceramics weighing 6.23 k g wer e collected from all levels of the single 1.0 x 5.0 m EU placed at the end of ET01 at MT FX 13 and the five 1.0 x 1.0 m EU's placed inside and just outside the main plaza Of this total only 22 pieces were identified as rims and only 10 of those were compl ete enough to obtain comparable data a nd determine vessel type Although this is a very small number to obtain statistically significant data from, it is presented here for comparison. Of the 10 rims 40 percent are Type 1 (n= 4 ), 50 percent are Type 2 (n= 5 ) 10 percent are Type 3 (n= 1 ). Of these 10 rims all were sufficient for determining orifice diameter. Type 1 vessels from the plaza berm have an average lip thickness of 1 .23 cm (n= 4 ), an average rim thickness of 1. 31 cm (n= 4 ), and an average orifice of 5 6 cm (n= 4 ). Lip thickness ranges from 84 to 1. 54 cm, rim thickness ranges from 98 to 1.46 and orifice diameter ranges from 42 to 74 cm. This wide range is reflected in the lack of
202 clustering of Type 1 vessels though they are still mostly separate from Ty pe 2 and Type 3 vessels at least in their large size ( Figure 5 58 ). Type 2 vessels from the plaza berm have an average lip thickness of .82 cm (n=16), an average rim thickness of .75 cm (n= 5 ), and an average orifice diameter of 29 cm (n= 5 ). Lip thickness ranges from .52 to 1.2 cm, rim thick ness ranges from .45 to 1. 09 and orifice diameter ranges from 12 to 60 cm. One significantly decorated Type 2 rim exhibits Araquinoid like decoration ( Figure 5 59 ). Tho ugh based on a very small sample size Type 2 vessels do cluster together when considered by their thickness to orifice ratio with one exceptionally large vessel represented ( Figure 5 58 ). Only a single Type 3 vessel was recorded f rom the plaza berm excavation with a lip thickness of 1.27 cm a rim thickness of 68 cm, and an orifice diameter of 32 cm Temper statistics for plaza berm vessels were calculated using all 22 rims. They are presented in terms of percent of total temper present. Calculated with all vessels (n= 22 ) the average temper composition is 64 percent cauxi 21 percent cariap 12 percent grit, and 3 percent grog. When broken down into vessel types the statistics change somewhat dramatically though this may be due to small sample size. Type 1 vessels (n= 4 ) are 54 percent cauxi 32 percent cariap 14 percent grit, and no grog present Type 2 vessels (n= 5 ) are 100 percent cauxi with no cariap grit, or grog, and very high levels of cauxi The single Type 3 vessel is 33 percent cariap and 67 percent grog, a very unique mixture and likely not representative of Type 3 vessels Surface treatment for all vessels included smoothing with only three vessels showing evidence of scraping. Some Type 1 vessels have exterior surface paint or slip (n=2) but none on the interior and no evidence of interior black paint. Type 2 and Type 3
203 vessels had no interior paint or slip but all showed burnishing on the interior and exterior surface (n=6) Of the 22 rims analyzed 82 percent ( n= 18 ) were completely fired showing no core and the remaining 18 percent (n= 4 ) were incompletely fired Unlike vessels from MT FX 06 all of the core colors and margins ranged from black to very dark brown or dark gray. Figure 5 58 Lip thickness and orifice diameter for all Group 2 vessels from MT FX 13.
204 Figure 5 59 Type 2 rim (Araquinoid like) from EU at MT FX 13. Figure 5 60 Type 1 (bottom right), Type 2 (top row), and Type 3 (bottom left, 004 1) rim profiles from EU at MT FX 13.
205 Group 3 (Plaza Peripheral Excavations) A total of 1,087 ceramics weighing 7.14 kilograms, were collected from all test unit e xcavations in the plaza peripheral areas of MT FX 13 Of this total 1 8 pieces were identified as rims. Of the 1 8 rims 17 could be accurately assigned to a vessel type. Of the 17 rims 59 percent are Type 1 (n= 10 ), 35 percent are Type 2 (n= 6 ), and 6 percent are Type 3 (n= 1 ). Of these 17 rims only 1 0 were sufficient for determining orifice diameter. These 10 rims are somewhat skewed statistically compared to the total 17 in the vessel type breakdown likely due to the small sample size. Of the1 0 rims sufficien t for measuring orifice diameter 80 percent are Type 1 (n= 8 ), 10 percent are Type 2 (n= 1 ), and 10 percent are Type 3 (n= 1 ). Type 1 vessels from the test unit excavations have an average lip thickness of .95 cm (n= 17 ), an average rim thickness of 1. 10 cm (n = 17 ), and an average orifice of 37 cm (n= 10 ). The range for Type 1 vessel metric attributes is somewhat narrower than for surface collected Type 1 vessels. Lip thickness ranges from .67 to 1.80 cm, rim thickness ranges from 81 to 1.95 and orifice diamete r ranges from 20 to 68 cm. Type 1 vessels generally cluster together and mostly separate from Type 2 and Type 3 vessels but not as distinctly as in the surface assemblage and less than the unit assemblage though the sample size is very small ( Figure 5 62 ). Type 2 vessels from the test unit excavations have an average lip thickness of 70 cm (n= 6 ), an average rim thickness of 86 cm (n= 6 ), and an average orifice diameter of 21 cm (n= 1 ). Like Type 1 vessels, the range for Type 2 ves sel metric attributes is narrower in the test unit excavation s. Lip thickness ranges from .45 to 1. 14 cm, rim thickness ranges from 71 to 1.00 and orifice diameter on the only vessel measurable is 21 cm One Type 2 rim adorno, too small to measure metric attributes, does have
206 significant Araquinoid like decoration ( Figure 5 61 ). The single Type 2 vessel large enough to be measured is separate from the Type 1 and Type 3 vessels when considered by orifice to thickness ratio ( Figure 5 62 ). A single Type 3 vessel was identified from the test unit excavations and has a lip thickness of .8 8 cm a rim thickness of 66 cm and an orifice diameter of 20 cm. The single Type 3 vessel is separate from the Type 1 and Type 2 vessels when considered for orifice to thickness ration (see Figure 5 62 ). Figure 5 61 Type 2 rim adorno (Arquinoid like) from TU at MT FX 13. Temper statistics for test u nit vessels were calculated using all 17 rims. Calculated with all vessels (n= 17 ) the average temper composition is 60 percent cauxi 22 percent cariap 12 percent grit and 6 percent grog. When broken down into vessel types the
207 statistics remain relativ ely consistent. Type 1 vessels (n= 10) are 58 percent cauxi 24 percent cariap 15 percent grit, and 3 percent grog. Type 2 vessels (n= 6 ) are 62 percent cauxi 2 2 percent cariap 5 percent grit and 11 percent grog. The sole Type 3 vessel is 62 percent c auxi 33 percent grit, with no cariap or grog Surface treatment for all vessels included smoothing, occasional burnishing (n=2) and scraping (n=3) One Type 2 vessel has red paint present on the outside surface. Of the 17 rims analyzed 76 percent (n= 13 ) were completely fired showing and 24 percent (n= 4 ) were incompletely fired Core colors ranged from black to very dark brown Figure 5 62 Lip thickness and oral diameter for all Group 3 vessels from MT FX 13.
208 Other Upper Xingu Sites Alto Xingu Data for this the Alto Xingu or AX designated sites is limited to that collected by Mario Simes at MT AX 01 Posto Diauarum and reevaluated at the MPEG. Statistical n file at the MPEG and reorganized here ( Figure 5 63 Figure 5 64 Figure 5 65 ) show that similar trends were found at three excavations (Corte 1 site. This data shows a general trend with higher amounts of cauxi in the upper stratum and lower amounts of cariap. This is similar to data presented here from the sites of MT FX 06 and MT FX 13. Vessels at MT AX 01 also have thick flat bottomed bases ( Figure 5 66 ) and uniform Type 1 and Type 2 vessels, based on rim profiles ( Figure 5 67 ) as well as the presence of pot stands ( Figure 5 68 ). Formadores do Xingu Ceramic examples from other sites in the lower Culuene ( or the formadores do Xingu ) are almost entirely surface collected examples. Analysis shows that surface collected examples are clearly related to late period surface collected pottery in the KSA. Of particular inter est is the presence of various type s o f pot stands with differing decorations. Such a bulky and heavy ceramic object is durable by its nature and unlikely to be traded. In the future these objects may provide good markers for individual houses, villages, o r at least pottery producing groups given their likely stationary nature combined with their seemingly unique and localized decorative attributes ( Figure 5 68 Figure 5 72 Figure 5 77 Figure 5 90 Figure 5 93 ). Other patterns among late period pottery from across the Upper Xingu include the wide spread prese nce of Type 2 vessels with flat folded rims and engraved chevron esque designs that are found at
209 several sites Present alongside these distinctive rim forms are animal adornos, and broad line incision ( Figure 5 71 Figure 5 75 Figure 5 76 Figure 5 80 Figure 5 81 Figure 5 84 Figure 5 91 Figure 5 92 ). At MT FX 01 Type 2 rim profiles show the same uniformity as other l ate period vessels across the Upper Xingu ( Figure 5 69 ). The presence of adornos, fine line incision, red slipped Type 1 vessels, and engraved designs on folded Type 2 rims, all suggest a mixing of late period Carib pottery with e stablished local forms that are also present in the KSA ( Figure 5 69 Figure 5 70 Figure 5 71 Figure 5 72 ). Site MT FX 02 has much of the same pottery as MT FX 01 including Type 1 vessels with uniform rim profiles, Type 2 vessels with much thicker, but still uniform, rim profiles, characteristic Carib (Protohistoric period) engraving, red slipped Type 1 vessels, broad line incision, and modeling ( Figure 5 73 Figure 5 74 Figure 5 75 Figure 5 76 Figure 5 77 Figure 5 78 ). The nearby site MT FX 03 al so has late period engraved rim Type 2 vessels in association with red slipped Type 1 vessels and very uniform rim profiles for both ( Figure 5 79 Figure 5 80 Figure 5 81 ). At M T FX 04 the engraved design usually reserved for Type 2 folded ri m vessels is found on an example of a red slipped Type 1 vessel further accentuating the merging of old and new styles of potter y in the Develo pmental (Late Ipavu) and Protoh istoric period of the Upper Xingu ( Figure 5 82 Figure 5 83 Figure 5 84 ) T he sites of MT FX 05, MT FX 09, MT FX 11, MT FX 12, and MT FX 18 also show a variety of vessels with some uniformity in rim shape and a mixture of decorative attributes including modeling, incision, and late period engraving, a decorative attribute that seems to be wide spread across the Upper Xingu between the Late Ipavu and Protohistoric periods ( Figure 5 85 Figure 5 86 Figure 5 87 Figure 5 88 Figure 5 89 Figure 5 90 Figure 5 91 Figure 5 92 Figure
210 5 93 ). This cursory evidence from many sites located along the Xingu and Culuene r iver s provides the basis for establishing a hypothesis for pre European trade and exchange networks in the Up per Xingu where localized economies of village clusters (centered near lakes) were intertwined through regional networks that increasingly merged into a single Xinguano culture.
211 Figure 5 63 T emper stati stics from Simes first excavation at MT AX 01. Figure 5 64 Temper statistics from Simes second excavation at MT AX 01.
212 Figure 5 65 Temper statistics from Simes third excavation at MT AX 01. Figure 5 66 Type 4 rim profiles (top row) and base sherd profiles (bottom rows) from Simes MT AX 01 surface collection.
213 Figure 5 67 Type 2 rim profiles (top row) and Type 1 rim profiles (bottom rows) from Simes MT AX 01 surface collection. Figure 5 68 Type 5 (or pot stand undagi ) from Simes MT AX 01 surface collection
214 Figure 5 69 Type 2 rim profiles from MT FX 01. Figure 5 70 Rim adorno's from MT FX 01 surface collection.
215 Figure 5 71 Type 2 rim s from MT FX 01 surface collection with engraved decoration.
216 Figure 5 72 Type 1 rims (bottom), Type 2 rims (middle), and Type 5 fragment from MT FX 01.
217 Figure 5 73 Type 1 rim profiles from the MT FX 02 surface collection.
218 Figure 5 74 Type 2 rim profiles from the MT FX 02 surface collection.
219 Figure 5 75 Type 5 fragments (top two rows) base fragments (third row from top), and Type 4 rim profile (bottom row) from MT FX 02 surface collection.
220 Figure 5 76 Type 2 rims from MT FX 02 surface collection with i ncised engraved, and thumbnail punctate decorat ions.
221 Figure 5 77 Type 1 rims (top row) and Type 5 fragments from MT FX 02 surface collection.
222 Figure 5 78 Modeled and incised rim sherds from MT FX 02 surface collection.
223 Figure 5 79 Type 1 (bottom left) and Type 2 (top row and bottom right) rim profiles from MT FX 03 surface collection.
224 Figure 5 80 Type 2 rims from MT FX 03 surface collection with engraved decoration.
225 Figure 5 81 Type 2 rims with engraved decoration (top three rows) and Type1 rim (bottom row) from MT FX 03 surface collection.
226 Figure 5 82 Type 2 rims (top two rows) and Type 1 rims (bottom row) from MT FX 04 surface collection.
227 Figure 5 83 Type 1 rim from MT FX 04 with engraved decoration on the i nterior (top) and red slip on the exterior (bottom)
228 Figure 5 84 Type 2 rims with engraved decoration (to p four rows), incised decoration (bottom right), and adorno (bottom left) from MT FX 0 4 surface collection.
229 Figure 5 85 Type 1 rim profiles (bottom row) and Type 2 rim profiles (top row) from MT FX 05 sur face collection.
230 Figure 5 86 Type 2 rims with incised decoration from MT FX 09 surface collection.
231 Figure 5 87 Type 2 rims with i nci sed and thumb nail pu nctate decoration from MT FX 09 surface collection.
232 Figure 5 88 Type 2 rims with modeled and incised decoration from MT FX 09 surface collection.
233 Figure 5 89 Type 1 vessel (small) from MT FX 09 surface collection. Figure 5 90 Whole pot stand ( undagi ) from MT FX 11 surface collection.
234 Figure 5 91 Type 2 rims with engraved decoration from MT FX 11 surface collection.
235 Figure 5 92 Type 2 rims with engraved decoration from MT FX 12 surface collection.
236 Figure 5 93 Decorated pot stand fragment (top left) Type 2 incised rim s (top right) and Type 1 rim with red slip (bottom) from MT FX 18 surface collection.
237 Summary This discussion concerns the apparent patterns in comparing the three assigned comparative un its; Group 1 ( the surface artifacts ) Group 2 ( the plaza berm artifacts) and Group 3 ( the test unit artifacts) This discussion takes into account the almost certain overlapping of ceramics from different time periods occurring in all of these units of an a lysis especially between Group 2 and Group 3 However, it is clea r, based in the chronology outlined in previous research and in previous sections, that site deposits can be broken down into at least three contiguous periods where the deepest deposits can be assu med to be the oldest (circa A.D. 700 12 5 0) while the surface artifacts can be assumed to b e the most recent (circa A.D. 150 0 177 0 ) and at the Nokugu and Heulugiht sites, t he plaza berms associated with plaza elaboration circumferential ditch exc avation, and likely population increase fall in the middle ( circa A.D. 12 5 0 1 50 0). Given this generalized periodization at these site s we can proceed to compare pottery assemblages and draw conclusions about the development of pottery technology over this roughly 1 000 year period. The first attribute of the ceramics analyzed is the form or type. Given the importance of certain vessels, particularly the Type 1 vessel, to the processing of manioc, it is important to see how this vessel varies through time, both technologically and in its relative prefer ence compared to other vessel types Vessel types are correlated ethnographically with particular functions. Type 1 vessels are associated with manioc processing, Type 2 vessels are associated with cooking fis h, and Type 3 vessels are associated with dry goods storage and traded as tourist goods. Archaeologica lly, Type 1 vessels are clearly the dominant vessel type in the surf ace assemblage as well as the TU assemblage while they are slightly edged out
238 by Type 2 vessels in the EU assemblage ( Figure 5 94 ). Type 2 vessels are the second most dominant vessel type in each assemblage with the exception of the EU assemblage where they are the dominant type as previously mentio ned. Finally, Type 3 vessels are the minority vessel in each assemblage constituting less than 20 percent in all assemblages. A clear pattern emerges from this data showing a dramatic increase, from 45 percent to 70 percent, in Type 1 vessels from the earl iest deposits to the latest ( Figure 5 94 ) Type 2 vessels decrease overall in preference from the earliest to the latest deposits with the exception of an increase from 37 percent to 49 percent in the middle assemblage. Finally, T ype 3 vessels remain at a constant low percentage throughout time to the point of almost dropping out completely in the surface assemblage. Overall metric attributes for vessel types show a distinct trend through time ( Figure 5 95 to Figure 5 97 ) Type 1 vessels increase in average lip thickness by 22 percent from the earliest deposits to the surface assemblage while average rim thickness increases by 10 percent and average orifice diameter increases by 8 percent. Type 2 vessels increase in average lip thickness by 16 percent from .81 c m to .94 c m while average rim thickness increases by 18 percent from .74 c m to .87 c m. A verage orifice diameter increases 39 percent from 23 centimeters to 32 centimeters. T ype 3 vessels increase in average lip thickness by 15 percent from .82 c m to .94 c m while average lip thickness increased by 36 percent from .59 c m to .80 c m and average orifice diameter increase by 108 percent from 12 c m to 25 c m The increase in average size combined with a decrease in standard deviation from these
239 averages demonstrates an increased uniformity or standardization of vessels. Trends toward a decrease in temper variability add further validity to this observation. Cauxi temper remains the d ominant type throughout all assemblages; however, its dominance in the surface assemblage suggests a conscious effort to use it over other available temper types ( Figure 5 98 to Figure 5 100 ) Cauxi temp er increases by 18 percent from the earliest deposits to the surface deposits where it represents 67 percent of the total temper ( Figure 5 98 ). Cariap decreases by 27 percent constituting only 15 percent of the to tal temper in the surface assemblage. Grit temper decreases by 62 percent though this is only a decrease from 21 percent to 13 percent of the total temper present. Finally, grog temper increases by 67 percent though remains at less than 5 percent of total temper throughout all assemblages. This general pattern of increases in cauxi decrease in cariap and decrease in grit is found in the statistics for separate vessel types as well ( Figure 5 99 and Figure 5 100 ). C auxi temper has been at the center of many discussions about pottery throughout Amazonia and in the Orinoco River basin. In the Central Amazon Hilbert first identified cauxi as a possible indicator of time or m igrations (Hilbert 1955:33 37). Temper was previously used in the Upper Xingu as an indica tor of cultural difference through time and space as well (Simes 1967). As far away as the Middle Orinoco basin, sponge spicule is noted for its sudden appearance du ring the Corozal phase around 1000 B.C. making it one of the earliest lowland complexes with examples of cauxi tempered pottery (Roosevelt 1997:159). Based on these previous studies a distinction between vessels that favored cariap and vessels that favo red cauxi was anticipated thus recording the presence
240 and amount of this type of temper was a priority. The results presented here clearly show that temper is an indicator of difference and variation through time at sites in the KSA. These results are cl ear evidence for ceramic technological change at least from the Initial Period to the Protohistoric Period of the Upper Xingu. Figure 5 94 Vessel type percentage within the three analysis groups at MT FX 06. Figure 5 95 Average o rifice diameter among vessel types within the three analysis groups at MT FX 06
241 Figure 5 96 Average l ip thickness among ve ssel types within the three analysis groups at MT FX 06. Figure 5 97 Rim thickness by vessel type and assemblage
242 Figure 5 98 Temper percentage for a ll vessel types combined within each assemblage from MT FX 06 Figure 5 99 T emper percentage for Type 1 vessels within each assemblage from MT FX 06
243 Figure 5 100 T emper percentag e for Type 2 vessels from TFX 06.
244 CHAPTER 6 ARCHAEOLOGICAL INTERPRETATION Introduction Archaeological work in the Upper Xingu combined with recent work conducted in the Central and Lower Amazon, establishes the framework for the revis ionist model and a more historicized understanding of settled and complex societies of th e Amazonian past. Archaeological and ethnographic field work conducted in the KSA reveals the presence of densely settled circular plaza villages throughout the lower Culuene River in the PIX that are directly related to the present day societies of the Upper Xingu M apping of these archaeological sites and the landsca pe features that characterize them illustrates the presence of these permanent villages throughout the region in prehistory Archaeological investigation and associated radiocarbon dates show that these villages were contemporaneous and connected by roads that reached several kilometers between villages (Heckenberger 2005; Heckenberger et al. 2007, 2008) W hile this work raises serious doubts regarding the standard model it also raises several questions regarding the similarities and differences between the modern Xinguanos and their prehistoric ancestors. The most obvious question is related to how such a c omplex system of villages in the past is now represented by only a few scattered villages. Regardless of continuous doubts about the validity of ethnohistoric accounts (Meggers 1995a) the massive depopulation brought on by the spread of disease and the dis placement of people in the wake of early contact with European s is hinted at by early et hnographies (Levi Strauss 1948), recorded in later census data (Heckenberger 1996) and historical accounts (Hemming 1978; Whitehead 1993) and shown in kinship analysis where kin terms remain in use where the structure for th ere
245 origin no longer exists (Dole 1991: 396). Still other questions regarding the genesis of Xingu ano culture remain including how and when Arawakan speaking people arrived in the Xingu, how and when Carib peoples arrived and mixed with Arawakan speaking people, what village and regional social organization looked like in the past, ho w material culture was manufactured and distributed, and how subsistence regimes were formed The starting place for se veral of these questions is in the ethnographic present and each of these questions can be related to the most prevalent form of prehistoric remains, pottery. Today, all pottery in the Upper Xingu is made by Arawakan speaking people in one village and trad ed to the other villages. Almost all pottery is made in relations hip to the processing of manioc which is not clearly the case throughout the archaeological record of the KSA. Finally, besides form, function, and technology, changing styles of pottery deco ration likely tell us something about the movement of different people into the Xingu as in other major study areas of the Amazon Upper Xingu Chronology Archaeological research in Amazonia has generally focused on the culture areas of the Lower, Middle, and Upper Amazon, with relatively little attention given to the upper reaches of its major tributaries The Xingu River the focus of detailed ethnographic resea rch since the late nineteenth century, provides ample data on socio political dynamics in nativ e Amazonia, both past and present. In the last fifteen years, substantial archaeological research has complimented the ethnographic research in this area and some of the ceramic data from that research is presented here. Archaeological research in this are a during the mid twentieth century established the presence of prehistoric villages but limited excavations and almost no surface
246 mapping of the sites left only vague information about the earliest inhabitants of this area (Dole 1961) A rchaeological resea rch shows that just prior to initial contact with the first Europeans to reach Amazonia the Xingu supported vast populations that seem to be elaborate expressions of the same basic socio political system that is found among the Xinguano of today (Heckenbe rger 2005) Using a direct historical and ethnoarchaeological approach, Heckenberger successfully showed that the present day condition of Xinguano people is in fact a product of contact with Europeans (Heckenberger 1996) He also showed continuity between late prehistoric and present day Xinguano economic and social organization based in part on a comparison of late prehistoric and modern day Xinguano pottery The earliest documented occupation of the Upper Xingu, based on radiocarbon dates, places people in the area beginning circa A.D. 7 00 and probably earlier. Heckenberger favors a broad range for the initial occupation of the Xingu stating that, "the known cultural history of the region begins by circa A.D. 500 800, and perhaps much earlier", however hi s assessment concentrates on "the last 500 years", from about A.D 1500 to the present (Heckenberger 2005:67). Though other research in the southern and central Amazon is scant with a few notable exceptions (Wust 1990; Ireland 1990; Prous 1992; Kipnis 1998) Heckenberger observes that "it is clear that the known sequence represents a single evolving cultural tradition: the Xinguano regional tradition (Heckenberge r 2005:68). According to Heckenberger ceramic occupations begin in the region no later than A.D 800 and are associated with a single industry. This single industry, named the Ipavu Phase by Mario Simes based on work further to the north and downstream is likened to the several traditions from the Amazon proper
247 including the Incised rim, Incised m odeled, and Amazonia n Barrancoid traditions. According to this summary these early inhabitants united by a single ceramic industry maintained their culture unt il around A.D. 1500 when they were met by Carib and other groups. Heckenberger's prehistoric chro nology is based on "solid empirical grounds, including a well controlled sequence of radiocarbon dates, systematic mapping and distributional studies of entire sites, detailed analysis of p rehistoric technologies, and a site survey r 2003:68). Based on these solid empirical grounds he goes further than Simes and divides the cultural sequence of the region into two major cultural phases, the Ipavu Phase and the Xinguano Phase. More importantly for this study he further divides these phases into cultural periods with the Ipavu divided into Early (800 1250) and Late (1250 1650) while t he Xinguano Phase is broken into the Transitional Period (1650 1750), the Early Xinguano (1740 1884), and the Late Xinguano (1884 1950). Referring back to his assertion that "a single evolving cultural tradition" exists in the region, he breaks up these periods with a different naming sequence, where the Early Ipavu becomes the Developmental period, the Late Ipavu becomes the Galactic period, the Transition al Period is subsumed by the Early Xinguano Period, and the Late Xinguano becomes the Historical Xinguano Period. This sequence is based on single radiocarbon dates, clustering of radiocarbon dates, and a single historical event, the arrival of explorer Ka rl von den Steinen. This rather complex method of dividing Upper Xingu prehistory allows for the discussion of complex socio political developments throughout Upper Xingu history. Another way of dividing up the prehistoric period from earliest occupation t o the contact
248 period in Brazil, which begins roughly around A.D. 1500, is by the landscape alterations and other cultural formations. For this Heckenberger places the "structural elaboration", "nucleation in villages", and "establishment of 'galactic clust er' regional organization" in the period from 1250 to 1650. Radiocarbon da tes provide more detail for this At the site of Nokugu (MT FX 06) there are dates reported from several contexts. From the base of a "sub curb" on the central plaza a date of A.D. 9 80 103 0 is provided, indicating the very earliest this central plaza lands cape delineation could have begu n. From the small plaza at this site two dates were obtained from another "sub curb" though not designated as basal. These dates of A.D. 1400 1430 and A.D. 1420 1480, since not basal, suggest that this sub curb was initiated sometime before A.D. 1400. Another critical date is obtained from "ditch 2 (S), sub berm intact" ( Beta 78979 ) Given the location of this date at the base of the berm associated wit h ditch 2 it suggests that ditch 2, at least in this location, was initiated around A.D. 1260 1300. A complimentary date from ditch 1 suggests that this ditch had already filled in about 70 cm by A.D. 1270 1300 (Beta 176136). At the site of Heulugiht (MT FX 13) a date of A.D. 1260 1300 is provided from the "central plaza, sub curb, intact". Excavations beneath the plaza berm provided a date of A.D. 6 90 1030 an early occupation for this site prior to plaza berm accumulation with "artificial earthworks cons tructed some hundreds of years after initial site occupation (Heckenberger 2003:90) around A.D. 1260 1300. This suggests that the e ntire road/plaza marginal mound complex at Heulugiht is roughly contemporary with those documented at Nokugu and Kuhikugu, after circa A.D. 1300 1400 o r later (Heckenberger 2003:91).
249 The basic chronology of the Upper Xingu is shaped by the inferred changes in society based on landscape alterations detailed above, including village size and elaboration, for example (Heckenberg er 2005:68 75). Unlike much of the rest of the well studied Amazon, in the Upper Xingu there are not well established ceramic traditions, phases, or complexes nor a chronology in which to place them. As mentioned, some attempt at distinguishing complexes w as attempted by Simes. In addition to the Ipavu (also referred to as Culuene to distinguish it from the Xingu River proper) he identified the Diauarum complex based on sites far to the north of Lake Ipavu at Posto Diauarum. Heckenberger distinguished the Eastern and Western complexes which may be more aptly named the Ipatse and Tafununu complexes after the lakes around which they were found (Heckenberger 2005:102 103 ). These geographical complexes do not however address the chronological sequences for the Upper Xingu and tend to blur the earliest and latest occupations of th ese areas as single components. Simplifying this complex array of dating we may settle on a few time periods that move away from using place names to identify time periods. The suggested periods are the Initial Period (AD 7 00 1250), Developmental Period (1250 1 500 ), Protohistoric Period (1500 1884), and Historic Period (1884 1950). These periods are based on the earliest archaeological evidence for colonization of the Upper Xingu, structu ral elaboration of village configurations, the first unrecorded contact between Europeans and Xinguanos, and the first recorded contact between Europeans and Xinguanos, respectively as outlined in this section None of these periods were based on ceramic seriations or stratigraphic archaeological deposits at specific sites with the exception of
250 the beginnings the Initial and Developmental periods which are based on radiocarbon dates collected direct ly beneath village plaza mounds and site peripheral ditch es. Returning to pottery, c ertain vessel types based solely on form can be found throughout the record in all of the proposed periods Other ve ssel types based solely on form, decoration, or size, are found only in certain period s Those vessel forms that can be traced throughout the record demonstrate variation through time in size, temper, and amount. The size and temper changes, presumed to be a technological change provide s clues about the changing use of these vessels. The changing amount, both percen tage of total vessels and amount of total vessels, likely reflect the growing need for these vessels by a growing population. Because the sites examined fo r this study contain very macro scale stratigraphy, additional arbitrary stratigraphy was used to ser iate the ceramics from three broad contexts roughly correlating to the Initial, Developmental, and Protohistoric periods. These are the village wide late surface deposits (Group 1), man made plaza berms and roadside berms (Group 2) and village wi de reside ntial area excavations (Group 3) The village circumferential ditches also show distinct strata that may be correlated to those just described however, because they involved both prehistoric digging and accumulation in their construction thus reversing th eir stratigraphy in some cases, they were not considered here for analysis but can be better understood apart from the other part s of the site. In describing the pottery from these broad strata we can consider the changes and continuities visible in the su bsurface and berm zones as compare d to those from the surface artifac ts, presumably the most recent.
251 Initial Period (Early Ipavu) The Initial Xingu period is represented by pottery styles related to Incised Rim Tradition or Series known elsewhere in Amazon ia Pottery associated with the earliest radiocarbon dates in the Upper Xingu is found at sites of the Ipatse C omplex (or Western Complex) They are characterized by small, thin vessels, which exhibit a higher frequency of scraping, burnishing, and red pai nt than later vessels Their primary decorative feature is a row of parallel incised lines that appear on thickened and/or folded rim s These vessels are globular and have a pedestal base rather than the flat base found in later periods. They may be includ ed with the Incised Rim tradition of greater Amazonia Prominent on this localized (and late) version of the Incised Rim tradition are thumbnail punctuates along the edge of the lip of these vessels (which persists into later periods, though the context fo r these vessels is likely mixed). These punctates are so distinctive as to be able to discern handedness or at least rotation of the vessel during the application of this dec orative technique which may be compared with those applied at other sites in futur e studies. The frequency of these puncta tes is almost a truly Upper Xingu preference and is found rarely in such amounts in contexts along the main route of the Amazon. However, e xamples are known from Konduri sites and near Oriximina sites and also noted in early pottery of the Upper Amazon (Hilbert 1955; Lathrap 1970) Besides being generally smaller and th inner than later period vessels, Initial Period vessels in the Upper Xingu use higher proportions of cariap grit, and grog temper than later period v essels while maintaining high amounts of cauxi temper. The forms, both rim and body, are clearly part of the same tradition as later vessels though their function may have been different based on changes in overall size and technology
252 Developmental Perio d (Late Ipavu) The Developmental Xingu period is represented by the same basic assemblage of pottery from the Initial period but exhibits somewhat more uniformity. The Developmental Period represents the final local variation of pottery before Carib occupa tions in the Upper Xingu began to influence local potters perhaps towards the end of this period. Ceramics from this period are not securely separated from those of the Initial Xingu based on radiocarbon dates or depositional deposits. However, based on ce ramic chronologies elsewhere in the Amazon we would expect to find either Polychrome or Incised Punctate horizon ceramics this late. Ceramics found at MT FX 13 are dated between A.D. 1040 and 1300 and representative of the Developmental Period of the Ipatse C omplex with possible affinities to the Araquinoid pottery of far northern Amazonia ( Figure 5 56 ) Protohistoric Xingu The Protohistoric Xingu is represen ted by a mixed assemblage of Carib influenced local pottery and the continu ation and further standardization of forms from previous periods. The Protohistoric Period pottery in the Upper Xingu is a mixture of the local Arawak style developed through the Initial and Developmental periods combined with Complex (though found in limited quantities in the Ipatse Complex sites and furth er north at site of the Ipavu and Diauarum c omplex ). This pottery represents an introduced form of pottery tha t was briefly mixed with the in situ local sty les roughly between A.D. 1500 and 177 0 or later Pottery with this form, temper, and decorative application (engraved rather than incised) is documented at sites near Lake Ita fanunu (MT FX 14, MT FX 15, and MT FX 26) and Lake Ipatse (MT FX 12) (Heckenberger
253 2005:103 the flat rims. Chevron designs are engraved o nto the rim after the vessel has been fired. This is in stark contrast to potters from earlier periods who applied all incised design on wet to leather hard clay before it was fired. The extremely folded rim on large vessels is also in sharp contrast to the gradually flaring Type 1 rims developed through time in the Upper Xingu. It is not clear if the new application of the chevron motif is an imitation of an in situ local style or a poorly executed version of a style carri ed from outside the Upper Xingu al though the latter is more likely given several factors. First, there are v ery little chevron decorative motifs from the Initial through the Developmental periods of the Upper Xingu when parallel incised lines are popular. Second, as Lathrap points out, and if suppositions that these are indeed Carib invaders are true, the men ma y have brought the style without the skill of the execution with them. As a result of this kind of partial ethnic replacement, not all aspects of Carib culture would have been disseminated with uniformity and full understanding. If it is correct to assume that art style and ceramic technology were feminine domains, it would be predictable that these patterns would be transmitted in a poorly understood and garbled form, since there would be few, if any, properly trained women moving out of the old Carib hea rth land (Lathrap 1970:164 165). One other important detail that supports this Carib introduction to the Upper Xingu can be found outside of the material remains and in the presence of a regionally specific origin myth that is Carib (Carneiro 1989). The qu estion still remains whe ther or not the Carib added some imported style to an alr eady made local pot or imported the entire form and style. Radi ocarbon dates from both sites at the Itafanunu C omplex a nd a single site from the Ipatse C omplex
254 place these sit es squarely within the Protohistoric Period of the Upper Xing u, r oughly between A.D. 1500 and 177 0. In attempting to explain the mixture of local Xingu style with the newly introduced ment out of the Orinoco; The taking over of the lower Orinoco was by no means the limit of the territorial acquisitiveness of the Barrancoid peoples. Before the Barrancas style had started to evolve into Las Barrancas colonies were budding off from the set tlements on the Lower Orinoco. One wave moved east along Venezuelan and Guianan Barrancoid peoples makes it almost certain that they were the invaders who introduced Maipuran Arawak into these areas (Lathrap 1970:114 116). Describing these movements and "invasions" in terms of ceramic technology and style, specifically in the Upper Xingu as it relates to modern ethnographic ceramic styles, he states further that, The nature of the modern ceramic style of the Upper Xingu, brought into the area by Maipuran speakers, suggests an ultimate derivation from the Barrancoid tradition on the Central Amazon ... i n the two mi llennia following the Maipuran expansion many groups would have adopted the styles of their neighbors or so modified their own style as to make Barrancoid derivation no longer evident (Lathrap 1970:127). Historic Xingu The Historic Period in the Upper Xing u is the first ethnographic documentation of the ethnogenesis and cultural merging that occurred in the previous centuries. Historic period pottery, or Waur Xinguano pottery, is the final product of the Xinguano style of pottery developed from the Initial Period through the Protohistoric Period prior to the rapid destabilization of the region and the multiethnic consolidation that leads to the village specialization of several crafts including pottery. A s recorded in the nineteenth
255 and twentieth t his pottery is the closest archaeological and ethnohistoric analog to modern Upper Xingu pottery Characteristic attributes of this pottery are roughly the same as most Protohistoric pottery as recorded on the surface at Ipatse Complex sites (MT FX 12) a nd Ita fanunu C omplex sites ( MT FX 11 ) Though this pottery is found in much smaller amounts in buried contexts it appears to be primarily associated with historically known occupations of many Upper Xingu sites. These vessels were reco rded by Karl von den Steinen in 1884 ( Figure 5 3 ) and can be seen in Xinguano villages today, made exclusively by the Waur tribe (Coelho 1981; Lima 1950b). Upper Xingu Regional System Any attempt at identifying a regional system in the Upper Xingu mu st focus on the late prehistoric period when village complexes reached the zenith of regional integration. Contemporaneous clusters of villages around Upper Xingu lakes and corresponding roads linking these villages provide the clearest evidence for region al networks. This archaeological information can be combined with ethnographic data that illustrates the nature of social principles that may have ordered prehistoric village networks; Xinguano communities and regional clusters are hierarchically ordered, according to genealogy, works, gender, and age, and we see that ancestors or, more precisely, ancestral places are likewise arranged according to these social principles (Heckenberger 2007:304). Further characterization of these communities and their relat ionship to each other is discernible through ceramic remains. Pottery decorative styles and forms found among the surface assemblages of sites from Lake Itafanunu, Posto Diauarum, Lake Ipavu, and Lake Ipatse illustrate that late in time there was regional if not village level
256 distinction among pottery producers. Eventually these styles and forms developed into what is known as the Xinguano tradition made exclusively by the Waur. Earlier studies by Simes and Dole already documented differences between pott ery present at the Itafanunu, Ipatse, Ipavu, Lamakuka, and Diauarum complexes. Later work by Heckenberger further documented the differences in village and house arrangements at Itafanunu and Ipatse. In the Upper Xingu and in the study area specifically, c eramic forms and decorative motifs were previously categorized and used to document regional variation Mario Simes first divided his findings into two regional distinctions based on the presence of different types of temper at the Diauarum and Ipavu comp lexes. The only other significant regional distinction was that recorded by Heckenberger between the complexes of Itafanunu and those of Ipatse which he designated the Eastern and Western complexes. The main distinction was made between those sites with ci rcular houses and those sites with typical Upper Xingu long houses. Based on Simes dating, it appears more appropriate to describe pottery based on its locale which in most cases is near a lake or river confluence. Within these geographical locations there seems to be an apparent division of decorative and formative techniques in pottery production. Given the amount of overlapping dates we can only begin to distinguish between local groups and t heir degree of interaction when we can distinguish between their distinctive pottery styles. Within each locale there is the more arduous task of distinguishing changes and continuities through time.
257 Ipavu and Diauarum Simes first described Diauarum Compl ex pottery based on his excavations at seven sites near the confluence of the Rio Xingu, Rio Sui missu, and Rio Manitasau, though did not use the designation Diauarum until later (Simes 1967:142,1972:29,39). Simes dated the Ipavu to A.D. 1200 1300 base d on the presence of Ipavu type ceramics superimposed over Diauarum type ceramics radiocarbon dated to A.D.1120 90 at other Diauarum sites (Simes1967) He noted that Diauarum ceramics were tempered with cariap and Ipavu ceramics were tempered with cauxi the first observation that cauxi was present in later ceramics. Becquelin proposed an end date for the Ipavu around A.D. 1350 based on radiocarbon dates obtained from Lake Miararr in the vicinity of Lake Ipavu. Ipavu Complex ceramics were first described by Galvo Simes and Dole. Thus far they are the earliest dated ceramics in the upper Xingu and lower Culuene. Their rim forms vary from straight to horizontal. Horizontal rims often have notches or thumbnail punctates on the lip around the enti re vessel. These are often on a bias when looking down at the pot resulting from either left or right hand application. These everted rims also commonly have from 1 7 parallel incised lines that very in degree of precision, width, and depth. These vessels are globular commonly with pedestal bases. They are generally thin but range in oral diameter from 5 to 35 centimeters. There temper is variable and often contains a mixture of cariap cauxi grog, and grit or mineral. They never have the amount of caux i temper that is found in later Xi nguano phase vessels. Ipatse, It afanunu and Lamakuka Later period sites at Lake Itafanunu represent both Carib and Arawak ceramics and are found at sites described elsewhere (Heckenberger 2005). These are present at
258 sites across the Upper Xingu. In assessing the late Upper Xingu regional system, however, it is clear that the Itafanunu sites represent the Carib neighbors of the Arawak groups to the east. This has been demonstrated based on village configuration previously a nd by pottery style as well. pottery with this same rim form and rim decoration are known from sites at Lake Lamakuka, Lake Itafanunu, and at sites further north such as MT AX 0 1, MT AX 0 2, MT AX 0 3, and MT AX 0 4. They r epresent a widespread late form of rim and rim decoration attributable to Carib peoples just prior to the historic period based on evidence from Lake Itafanunu (Heckenberger 2005:106) This style was first referred to as Eastern Complex due to its abundant presence east of the Culuene River at presumed and historically known Carib sites along the shores of Lake Itafanunu This most distinct form of this style is a flat bottomed vessel with straight walls and a robust folded flat rim. Decoration is restricte d to the top portion of the rim and consists of engraved (as opposed to incised) chevron designs. Unlike any other line designs in the Upper Xingu, those on the Carib rims are applied after the vessels are fired giving them the appearance of being scratche d. The difficulty of engraving on hard fired pottery is apparent in the lack of precision of these designs. These vessels tend to be darker in color and are neither painted nor slipped. Based on radiometric data from the Itafanunu sites containing these ri m forms they slightly predate Von den Steinen 's visit to the Xingu and may thus be attributed to the Protohistoric Period Pottery with this form of rim and rim decoration is not present at MT FX 06, MT FX 13, or any of the other Ipatse stream sites except MT FX 12, the main archaeological site at Lake Ipatse and the most recent Carib occupation west of
259 the Culuene River, according to local oral history and radiometric data which dates the site to about A.D. 1770 (Heckenberger 2005:106). This late date for Upper Xingu Carib pottery places it as the direct contemporary of the latest pottery at MT FX 06, the direct archaeological ancestor of modern Waur, or Arawak, pottery. Also called Miarrare by Becque lin, and Group 3 by Dole, it is clear that a distinct st andardized set of rim and vessel forms with decoration limited to zoomorphic adornos on smaller pots and painted designs on larger pots, consolidated itself sometime after the Initial and Developmental periods (and perhaps concurrently with the Kuikuru or Carib ceramics, given the survival of Waur ceramics into the present and the abandonment of the Carib ceramic forms) Dole hypothesized that this was a direct result of the use of the larger vessels for manioc processing. She concluded that the Kuikuru fo rms, based on fracture lines, could not withstand the pressure of the Xingu specific technique of draining manioc pulp through mats on top of the vessels, rather than use of the tipiti for this task, as in most parts of Amazonia The absence of the tipiti in these outlaying areas among peoples who raise manioc suggests that the development of manioc cultivation was accompanied by a movement of peoples upstream, and that many of them reached the headwaters of the Amazon before having an opportunity to parti cipate in the full as a relatively recent invention (Dole 1960:246). Thus the more grad ual flare of the Waur vessels, which developed through time, replaced the Carib vessels shortly after Carib arri val in the area. T his is also illustrated in the increase d amounts of cauxi temper that facilitate these larger, thicker vessels Its light weight and thermal shock resistance made it the clear choice for very large vessels that needed to sit on the fire cooking manioc juice for many hours.
260 before having an opportunity to participate in the full fledged Topical Forest culture may have its flaws, but in regard to material cultu re, such as the tipiti and pottery, it has its merits. Like the tipiti many decorative and manufacturing styles are missing from Upper Xingu potteries that are found throughout the rest of the Amazon within similar Barrancoid like pottery assemblages Ho wever, comparable data from other southern headwaters communities is not available to verify the same process in other communities at some distance from the main Amazon River Amazonia and the Upper Xingu A persistent problem in Amazonian archaeology is th e identification of evidence to link the histories of cultural groups from vast areas of the region. Identifying migration routes and distinguishing Arawak from Carib settlements archaeologically is one element needed in solving this problem. The Upper Xin gu is particularly difficult to correlate with other parts of the Amazon because of its distance and relative isolation archaeologically. Add to that the condensed occupation history based on the earliest known radiocarbon dates from the area. Yet finding the archaeological evidence that ties the Upper Xingu into the cultural history of the rest of the Amazon remains important. Since the 1970s various hypothesis have been formulated about when, why, and how, Xinguanos arrived in the Upper Xingu. The focus o n horticulture and fertile land prompted Lathrap to suggest that the Upper Xingu was occupied as part of the search for fertile land; This competition for agricultural land has been going on for a long time, and began several millennia before Orellana's vo who have lost the battle have been many, and they have been pushed further up stream and off the major river s into the intervening expanses (Lathrap 1970:19).
261 Even this argument does not retreat completely from the standard mod el in its use of environmental factors to explain change. Because of the nature of previous models, These arguments and e xplanations for the relationship between the Upper Xingu and Amazonia are tied also into the appearance and spread of manioc horticulture. Lathrap felt that the active floodplains were much more attractive apart from the vast ancient alluvial deposits in general ( Lathrap 1970:26 48 57 ). He developed a history that suggested Amazonian a gr iculture matured on the flood plain of the Central Amazon. These Proto Arawakan groups experienced population pressure and the limited expanses of alluvial flood plain forced colonists to move out in the first Arawakan migration around 3000 BC. Later, as p opulation pressures continued to increase on the Central Amazon flood plain a second wave of migrations took place between 1000 B.C and A.D. 500 These Proto Maipuran peoples moved follo wing the same routes as their ancestors and went further even going do wn stream along the Amazon and even up the Xingu to small patches of alluvi al land in its upper watershed. This fight for the limited supply of productive farm land has been the most important single force in the culture history of the Amazon Basin, and mo re than any other factor is clearly visible in the archaeological record (Lathrap 1970:20). His entire hypothesis is based on three key factors; relatively continuous population pressure, constant rates of migration, and the search for a specific ecologic a l niche in good alluvial soils, such as the Upper Xingu (Lathrap 1970:75). Whether population pressure, environment, or social and political factors, a rchaeological research across Amazonia illustrates that some major event or sequence of events prompted c hange in the last half of the first millennium A.D. or
262 earlier he first cultures of the Tropical Forest Pattern, characterized archaeologically by settled villages and the manufacture of pottery, make the ir appearance on the Island of Maraj at a time estimated as somewhere around A.D. 700 (Meg gers and Evans 1957:598). While more recent research pushes that date back somewhat, it is st ill important to recognize that; The chronology of cultural change shows initial occupation by cultivators with a generalized economy of living in small autonomous villages. By A.D. 400, people moved to the headwaters and lakes in order to establish permanent villages and intensively exploit the abundant fish resources. In a f ew decades, cooperation and competition in such bountiful areas led to the emergence of chiefdoms (Schaan et al. 2009:130). In this description of the northern Amazon we find similarities to the Upper Xingu in the establishment of permanent villages and v illage clusters near lakes and along headwaters. Similarly, further to the south, in describing the raised fields of the It is unlikely that any of the ceramic styles at present known from the Lowlands of Bolivia date f rom earlier than A.D. 600 to 700 (Lathrap 1970:123). Moving back to the north, near the confluence of the Solimoes and Negro rivers in the Central Amazon, archaeological research places similar developments during roughly the same span of time. In additio n to the appearance of settle villages, archaeology in the Central Amazon also illustrat es the increase in village size; Early Ceramic age settlements of Amazonia n Barrancoid tradition, ca. 300 mic Age, that is, after A.D. 900 or A.D. 1000, local settlements within the CAP area ADE sediments and copious pottery sherds (Petersen et al. 2005:9).
263 Finally, in the Upper Xingu s ettled villages with ceramics appear also around the same he known cultural history of the region begins by circa A.D. 500 (Heckenberger 2005:67). Thus, the archaeological evidence seems clear; b y the late first millennium A.D. major chan ges were taking place all across Amazonia In most locations these changes were preceded by long traditions of pottery producing groups. In the other areas, such as the Upper Xingu these late first millennium changes were the first appearance of people in these regions, perhaps seeking new fertile land in an otherwise crowded Amazon Like the lowlands of Bolivia, t he Upper Xingu stand s out as unique from its Amazonia n counterparts in these developments small autonomous villages with notable ADE sediments and copious pottery sherds centuries of development prior to the appearance of larger villages. Instead the Upper Xingu shows rapid development be tween its initial settlement and the beginning of the second millennium A.D. or what can locally be called the Initial and Developmental periods (or Early and Late Ipavu p eriod s ) and what Heckenberger has called the developmental phase of the Xinguano tradi tion (Heckenberger 2005:72). In the Developmental Period large fortified villages are dwarfed in distinction by localized clusters of sites or c omplexes paralleling the developments in Maraj reported by Schaan where distinctive groups of mounds develop t he largest of which is the Camutins site ( Heckenberger 2005:73; Schaan et al. 2009:130). Another example of the disjunction among regional Amazo nian develop ments occurs around the mouth of the Tapajos River in the context of Santarem. Although
264 contemporan eous with Amazon Polychrome developments, Amazon P olychrome is virtually absent in Santarem where the Incised and Punctate horizon was dominant, much the way the Arauquinoid is present in Venezuela rather than the Amazon Polychrome. This may be true of the Upper Amazon and the Upper Xingu, where there is not a truly Amazon Polychrome tradition but rather an Incised Punctate/Araquinoid form. This lag or gap in the spread of certain styles may suggest that early suppositions by Lathrap were right. W e would ex pect to find waves of development, not simultaneous development across the Amazon. Successions of people moving out and taking new ideas and styles with them and passing these on either in wholesale replacement or in combination with previous ways of livin g. If we take what is know n of the last 2 000 years in Amazonia n prehistory we see some of these waves emerge and thus place the variability present in the Upper Xingu within a broader Amazonia n context. It is in fact the "more obvious and significant of t hese waves" that formed the basis for Lathrap's seminal work on the Upper Amazon (1970 ) Much of the changing territory, technology, and culture is hypothesized to relate to dramatic environmental changes including intense dry and wet periods and the chang ing flora and fauna as a result (Whitmore and Prance 1987; Meggers 1994a, 1994b; Meggers and Miller 2003) even in later periods the changing course of rivers can help explain the absence or presence of sites during different periods (Lathrap 1968). T he Ha ndbook of South American Indians in 1948 first identified the Tropical Forest Culture of South America by four diagnostic fea tures t he first of which is the cultivation of tropical root crops, especially bitter manioc Lathrap hypothesized that direct evi dence on the beginnings of manioc cultivation would date to around 5000 to
265 7000 B.C. based on the length of time it takes to domesticate wild cultigens (Lathrap 1970:57). He further suggested that the origins of Tropical Forest Culture were to be found in the extensive areas of riverine flood plain in the Amazon about 3000 B.C. (Lathrap 1970:67). However, not until almost 4 000 years later do we see any evidence of Tropical Forest Culture in the Upper Xingu. If Lathrap is right this would suggest that the ev idence of this culture in the Upper Xingu around A.D. 700 1000 or earlier is the result of its movement out of the Central Amazon much earlier than it appears in the Upper Xingu. This lag is accentuated by the fact that the first ceramics in the Upper Xing u represent the middle first millennium arrival of people from elsewhere in Amazonia and reflect the lack of several central aspects of Amazonian culture that are still not present in the Upper Xingu, including the brewing of beer and the burial of the dea d in urns (Mowat 1989). There is other evidence to sugge st that Amazonian peoples who moved into the Upper Xingu did not carry with them the full suite of characteristics developed later in the Central and Upper Amazon One piece of evidence is the way in which bitter manioc is processed in the Upper Xingu. In almost all of Amazonia the shredded tuber is processed through a cylindrical basket called a tipiti. In the Upper Xingu however the tipiti is substituted with the tuafi The Upper Xingu has the distin ction of being the only place in the Amazon to use the tuafi rather than the ubiquitous tipiti. ognition of the differential distribution of the tipiti and tuafi (Dole 1964) and Carneir (Carneiro 2001) are in line with oth er material evidence such as native fiber industries (Petersen et al. 2001) that show cultural distinction across the Amazon. This combined with an understanding of the relatio nship between Caribbean, Orinoco and Amazon pottery suggests that the attempts to tie together the
266 entirety of the Amazon through ceramic styles is not entirely unfounded or impossible especially considering the other evidence of successive waves of migrations throughout the prehistory of Amazonia (Petersen et al 2001a, 2004) In as sessing the movement and possible origins of Upper Xingu cultures using historic and modern ethnographic information, there has been much s peculation about native acculturation and assimilation including explanations that involve refuge from violence to th e north or south ( Gregor 1990:180; Schaden 1964) At least as far as migration from the south is concerned, t he intimate knowledge of the forest by Upper Xingu tribes, suggests that they are not purely savannah folk from the south (Carneiro 1978) They als o d istinguish between primary rain forest ( itsuni ), secondary forest ( tafuga ), terra preta ( egepe ), and savanna ( oti ), a developed lexicon that would have devel oped over a long period of integrating culture and livelihood with the tropical forest environme nt (Carneiro 1978a:203 204). Summary M uch is still unknown about the first occupants of the Upper Xingu, when they arrived and from where they came. The earliest dates are around A.D. 700 though this does not mean earlier dates are not out there to be fou nd. Clearly there are two different foci in the Upper Xingu as laid out by Heckenberger, the eastern and western complexes Other "new arrivals" are noted at this same time in other disparate portions of the Amazon. Lathrap reports evidence at Pacacocha fo r the arrival of bitter manioc using crude pot tery making peoples from "downstream" around A.D. 700. Similarly, he reports the appearance of a circular village at Nueva Esperanza around A.D. 700 much like the ones taking shape in the Upper Xingu at this time These people were later displaced by Cumancaya peoples from the south, at the same time Caribs moved into
267 the Upper Xingu presumably from the north, following a pattern of multipl e migrations and displacements along (or generally up) river valleys, as population and other pressures caused people to move into less fertil e areas where they adapted their tropical forest economies to the savannahs and gallery forests supplemented with fishing. Finally, t he Amazon basin itself, or Amazonia, is diverse in its geographical landscape, soils, topography, and ethnographic diversity. All of this lead one of the pioneers of twentieth century archaeology in the region, Donald Lathrap to the conclusion that, "It is doubtful if the culture history of the tropical f orests of South America will ever be successfully encompassed in a really simple develo (Lathrap 1970:21). Even in 1973 Meggers and Evans proclaim lthough little archaeological investigation has been conducted in the Amazon Basin, ev en along the main river, existing evidence indicates great variation in the prehistoric pottery (the principal surviving cultural remains) through space and time, implying a complicated histo ). The relationships between variat ion in prehistoric pottery technology, labor, and subsistence demands are presented throughout this study. Ethnographic and archaeological data provide a framework in which the constants and variables of Upper Xingu history can be examined simultaneously. Manioc, a constant source of subsistence throughout Amazonia and in the Upper Xingu, is the main product of labor while pottery is a secondary product of labor. Variations in pottery technology reflect both the increase in demand for manioc and an increase in the demand for labor which produces manioc. Each of these three elements of society, pottery, manioc, and labor,
268 are specific examples of structural elements of society; the control of craft production, social and political control, and subsistence rel iance respectively. Each of these is examined separately in light of the new data. The differences in pottery assemblages present at the sites of MT FX 06 and MT FX 13 and between other sites throughout the Upper Xingu region, demonstrate that each villag e or possibly each household produced its own pottery late in prehistory before shifting at some point to a village specialization system where a single village manufactured pottery consumed be each of the other villages in the regional network This archa eological evidence fits with other interpretations of the area suggesting that the impact of European c ontact on the Upper Xingu in the fifteenth century altered the scale and organization of regional chiefdoms and village organization. Sometime around A.D 700 a group of people moved into the headwaters region of the Xingu River. They brought with them a pottery tradition that is not dissimilar to the Barrancoid style of pottery, originally documented in the Orinoco River basin but also well documented in the Amazon River basin, particularly in the Upper Amazon in the Ucayali valley and the central Amazon near Santarem. There is almost no doubt that this assemblage of early Upper Xingu pottery was used in the processing of bitter manioc based on both archae ological evidence such as the presence of griddles, or alato and on ethnographic evidence from the Upper Xingu. As population increased during this initial and developmental period the people of the Upper Xingu, Xinguanos, responded with a reorganized and more centralized social and political system. The increased population demanded more food which was handled on a household basis. Village regional density increased and as this happened across the Amazon, and in the
269 Upper Xingu, conflict arose. Within the se regional stresses are situated the village level stresses. This includes a stress on those producing the increased food supplies and those producing the pottery to process those food supplies. The reflection of, or response to, this social stress is fou nd in the transformation of pottery directly linked to the processing of food. Variation between pottery assemblage s through time reflects a concurrent rise in the importance of manioc horticulture and the processing of this root crop early in the chronolo gy. The demand for processed manioc, and therefore the demand for labor to produce it, fueled the change in the pottery used specifically to process it. The transition to a more standardized and uniform suite of pottery vessels whose form and technology su ited their function, coincides with an increase in village social complexity and the construction of village peripheral ditches and village segmenting roads, all part of a complex alteration and expansion of the chiefly society of the Upper Xingu. Taken to gether, the landscape alterations and the shift in pottery production suggest that the chiefly demand for manioc production played a pivotal role in the transformation of pottery technology in the region. This study attempts to reconcile previous ceramic c hronologies from the Upper Xingu and contribute a foundation upon which further analyses in the region can rest with the hope that distinct Upper Xingu ceramic complexes can be more clearly delineated and eventually be more firmly contextualized within a b roader Amazonia n framework. Although this has been done to some degree by correlating the Upper Xingu ceramics with the Amazonian Barrancoid Incised Punctate, Incised Rim and Amazonia n Polychrome traditions, these correlations are very broad and blur loca l
2 70 distinctions both spatially and temporally. In short, upper Xingu prehistory (and proto history) is far too complex to fit well into any of the broad Amazonia n traditions. To overcome this and correlate the Upper Xingu first within its own context, a muc h more detailed attribute analysis was undertaken on ceramics that represent those found and described in each of the previous studies and also represent the full span of indigenous occupation thus far identified in the Upper Xingu. Additionally, ceramics collected by Mario Sim es during field studies in 1960s were reanalyzed at the Museu Paraense Emlio Goeldi (MPEG) and included in the overall analysis to link prehistoric remains in the lower Culuene project area with tho se little studied sites on the Xin gu River proper. This geographical distinction between the lower Culuene and upper Xingu is also important. Several other regional distinctions are made here, for example, between Lake Lamakuka and Lake Tafununu, as well as, Lake Ipavu. One of the primary goals of the ce ramic analysis was to better understand changes and continuities in ceramic technology through time and space within the KSA and throughout the Upper Xingu region and broader Amazon more generally. Through systematic sampling both horizontal ly and vertically, the samples analyzed here represent the full span of time known in the study area, roughly from A.D. 7 00 to the present. The samples also represent the full range of activity areas across the well mapped prehistoric villages of Nokugu (M T FX 06) and Heulugiht (MT FX 13). A third goal of this analysis was to identify social distinction within the prehistoric village complexes of the Upper Xingu based on hypothesis set out by previous research (Heckenberger 2005:123). According to this hy pothesis the construction of earthworks served to physically enhance social divisions that were already in place. If this is so we
271 would expect to find some difference in the material remains within each of these divisions. The second part of this analysis is also related to the development of these divisions. Previous research also suggests that the increase in population and subsequent social hierarchy was the result or helped along by the increase in manioc production, the main staple of the upper Xingu diet. This increase in production is hypothesized to be reflected in the manufacture of pottery which would have become more specialized for a single use, the processing of manioc. This technique of using the manufacture of ceramics to understand the socia l organization it produced or was a product of is layered over an already robust knowledge of the spatial organization of both the prehistoric project area and the modern d ay villages of the project area
272 CHAPTER 7 THEORETICAL CONCLUSIONS Introduction This study an d its corresponding interpretations are based on an a nalysis of pottery from archaeological sites in the KSA and limited amounts of pottery collected by Mario Simes and others north of the KSA The amount of data collected from the Nokugu site (MT FX 06) and the Heulugiht site (MT FX 13) allow for a more detailed understanding of the diachronic trends in pottery form and style from the Initial Period to the Historic Period in the Upper Xingu The data collected from various site s throughout the Upper Xin gu, mostly from surface collections, and analyzed at the MPEG, allow for some interpretation of the synchronic widespread appearance of certain pottery forms and styles sometime during the Developmental and Protohistoric period s A transformation in potter y technology is the most statistically significant change over time. This is apparent especially in the Type 1 vessel, or ahukugu used specifically for manioc processing. The most compelling evidence of this transformation is found in the increasing propo rtion of cauxi temper in comparison to cariap temper. Increased amounts of cauxi improve the thermal properties and physical composition of the vessels creating a lightweight vessel that can be used in long term hot processing and moved easily for cold processing. Thickness and overall diameter of the Type 1 vessels also increase allowing for more durability and increased productivity. In addition to these technological transformations, ahukugu vessels become more standardized in their thickened form and increased size, especially in the style of rim finish. Sometime during the Developmental Period or early in the Protohistoric Period, after Type 1 vessels were well on their way to the standardized form of their historic
273 counterparts, a regional complex o f sites is visible based on decorative variation across the Upper Xingu. Based on surface collected ceramics that date to at least 200 years prior to abandonment of these sites (and probably well into the historic period), a preliminary model of synchronic variability in the late prehistoric Upper Xingu is evident. This model suggests that a regional system existed in which village based pottery production resulted in transmission of village specific manufacturing knowledge through intervillage exchange and exogamous marriage practices and even kidnapping. Numerous reports implicate the Suya as feared burglars of the Waur and other villages of the region where they steal pots and steal women Among the Suya there was a minimum of three Waur women who make all the baked clay pots for the tribe as explained by Chief Pentoti These pots were however different from the pots that Waur wome n make in their native village Of circular form, smoothwall and tapering slightly, they measure in height almost the sam e as they do in orifice (Schultz 1961:327) [translation by author]. Contrasting this supposition other ceramic ethnoarchaeological studies suggest that "ceramic change is not simply a function of altered post marital residence patterns or of the immigratio n of new peoples" (Kramer 1985:95). Clearly, many factors influence the structure of pottery manufacture and production including social and technological circumstances. Technological Considerations As historic examples of Suya men stealing pots and women from Waur villages suggest, knowledge about pottery manufacture in the Historic Period was concentrated into the women of one tribe. This knowledge was so sought after that these women were the subject of kidnapping. Clearly, knowledge about technology in pottery is not restricted to the ceramicist, archaeologist, or scientist. As this historic example illustrates, it is reasonable to assume that prehistoric potters had a level of
274 understanding and sophistication that allowed them to understand the effects of various temper on the workability, drying shrinkage, firing behavior, and fired properties of pots (Rye 1976:109) For anthropologists this is important because the s trategies adopted to manipulate these properties are the product of available material s the transmission of product. Consideration of these strategies and influencing structures are critical to a complete understanding of the technology of pottery in the past. As Arnold points out, T he kicks that turned technological innovations in pottery production into deviation amplifying mechanisms were probably the result of feedback processes like population pressure and increased demand for ceramic vessels (Arnold 1985:220). Implicit in this statement is the increased demand for more subsistence and in the Upper Xingu this specifically relates to the hot and cold processing of manioc. This demand compelled Upper Xingu potters to develop a vessel that was tec hnologically sound for these processes. They specifically required a large vessel that was not too large to move and could withstand extended periods over a fire. One of the most important aspects of ceramic technology specifically related to performance c haracteristics in extreme firing conditions is temper. Since the vessel was flat bottomed, unusual for a pot that is used for cooking, the vessel had to be tempered to withstand direct firing during cooking. As studies in the technology of temper in prehis toric pottery show, the smaller the temper the more resistant the pot is to thermal shock (Bronitsky and Hamer 1986:96). These studies also show that an increase in the amount of small temper also increased the resistance to impact testing after thermal sh ocking. Bronitsky and Hamer concluded during one of their experimental archaeology studies on the use of small temper that
275 using small temper was in f the only instance in the entire testing program in which amount of temper affected briquette performa nce (Bronitsky and Hamer 1986:97; Steponaitis 1982). In the Upper Xingu cauxi provides a small temper that is used in vessels receiving daily long term thermal shock Although the rate at which potters increased the use of cauxi remains vague, w hat is c lear is that the technological innovations were contemporary with the increase of village size and population. As t he re was an increased need for large r manioc boiling vessels that could withstand thermal fatigue as a result of longer periods over a fire, t he increased amounts of cauxi temper provided thermal resistance in a lighter vessel that allowed for increasing vessel size Increased vessel size was likely related to increased production capacity. Although today ceramic pots are used almost exclusive ly for t he boiling process as late as the 1950s they were used for the entire process and would have been produced in much larger quantities (Galvo 1952) An alternate hypothesis for the increased preference for cauxi temper is a change in diet. If it w as not an increase in production that inspired these changes perhaps it was a shift in what was being produced at least early in Upper Xingu occupation, perhaps during the Initial and Developmental Periods (circa A.D 700 1250), when a variety of crops may have been supplanted with a more manioc focused diet The use of sweet manioc is known from some areas of the Amazon and the shift in technology may have been a shift to accommodate the introduction (and likely intensification) of bitter manioc processing among a group of people previously accustomed to the low processing requirement of sweet manioc and its variety of associated pottery.
276 Social Considerations Changes in temper constituency should not overshadow the concurrent changes in other technological and stylistic attributes. Although "distinctions of temper and firing may be useful chronological markers, refined distinction of vessel shape, rim profile, and style of decoration are more likely to be superior indicators of time" (Lathrap 1964:354). The se reflect greatly on the intended use of the vessel which has much to tell us about the sociality and economics of the people using them (Chilton 1999). The fact that Upper Xingu communities reduced the variety of vessel forms to mostly large mouthed, fla t bottomed vessels, and homogenized the technology to thicker walled, densely tempered vessels, with standardized rim forms, suggests an intensification or homogenization of subsistence processing and consumption as well as pottery production. The intensif ication of domestic ceramics in the Upper Xingu agree with even the most divergent viewpoints in the area regarding the intensive manioc agriculture that took place there in late preh istory (Heckenberger et al. 1999 ). Given the tumultuous activities that o ccurred in the final years of occupation at prehistoric Upper Xingu sites and the correlated population changes, shifts, and increases, it is likely that production was on a steady increase at least until A.D 1500 As steady increases in population created a need for more subsistence, in this case processed manioc, there was also a demand for more quantity and more durability in the core ceramic vessel used in the processing of manioc. In present day ethnographic examples it is entirely the women who proces s the mani oc, from harvest to consumption. A n increase in agricultural production would have meant a decreased amount of time for the production of pottery especially since harvest and production take place during the same short dry season
277 that is used for pottery production. This would have created an immense stress within each household and perhaps lead t o a more focused specialization where one household in each village or even one village produced all the pottery, as documented historically and as is th e case in the present day Upper Xingu This narrowing of production would surely have lead to a decrease in variety, both technologically and stylistically. Although this is somewhat in contrast to Arnold's supposition that "efficiency is not important bec ause women's time in the home does not contribute economically to agriculture (Arnold 1985:220) it does seem to fit with the Upper Xingu example where ethnographically men plant and tend manioc in the fields and women harvest and process manioc from befo re sunrise until late in the morning each day throughout the dry season In the years before A.D. 1500 (and perhaps just after), before population pressure was affected by European arrival to Brazil, innovation in technology may have resided in the individ ual household or regional precinct with different kinds of innovation being carried out in indiv idual houses or villages connected by regional networks (Arnold 1985:220). In the Upper Xingu the identificati on of social boundaries, regional networks, and id entifiable social units, is directly related to identifying the formation of chiefdoms, social hierarchy, and regional social organization. T he identification of these units also bears directly on the pattern of organization and its cultural origins. Indee d, the identification of social groups has been a perennial concern throughout the history of archaeology (Stark 1999:25). Ethnographic work in the southern Amazon provides m ost of the details on the organization of social groups, the identity of tribes, a nd the differences that are found between the two (Basso 1973; Gregor 1977; Heckenberger
278 2005 ). But as more historically oriented research has pointed out the process of history has done much to change the appearance of these organizational structures fro m what they may have looked like prehistorically ( Dole 1978; Heckenberger 1996). Besides the evidence for contemporaneous village plaza s, oriented, connected and segmented by roads, there is little other direct evidence regarding the social organization of the prehistoric Upper Xingu, and even less regarding the intravillage organization of these communities. A possible first step towards understanding how these units were organized, or even distinguished is to identify them archaeologically. Not surprisin gly, like elsewhere in the prehistoric world, archaeological ceramics offer abundant material culture from which to begin this search for identifiable, prehistoric, social units. This may be done by identifying the so called "middle range" links between so ciopolitical behavior of potters and patterns in clay composition for example (Neupert 2000). Like form and decoration, compositional variability, including temper, corresponds well with local traditional social boundaries and is a function of both natura l and cultural sources of variation such as geographic location or origin (Stark et al. 2000). However, as other research shows, material constraints are not the only factor in choosing clay or temper, social distinction also plays a role (Gosselain 1994). Focusing on other forms of pottery besides those associated with subsistence, when they are available, can also provide important clues about sociality especially since pottery is not always associated with food production (Oyuela Caycedo 1995). Besides s ocial boundaries or distinction, ceramic ethnoarchaeology reminds us that technological change and persistence can occur for a variety of reasons including
279 technofunctional performance characteristics as seems the case in Upper Xingu pottery (Stark 2003:20 7). Detecting motor habits, such as handedness when applying certain decoration or producing perishable materials, can also be useful, as shown in the thumbnail punctates of sites in the KSA and in studies elsewhere (Hill 1977; Petersen, Heckenberger, Wolf ord 2001). Bronitsky points out that "changes in ceramic technology such as shifts in types, grades or amounts of temper have ramifications beyond simple This extends to the life cycle of pottery and ho w the technological style may affect its durability (Tite 1999) E ven where social changes are reflected in material culture the correspondence may not be easily interpreted. This is especially true for pottery which "responds sluggishly to dramatic extern al or internal social and political happenings, and only then, when these make themselves felt in more basic day to are they found in ceramic changes (Rice 1984b:274). Identifying ceramic changes that correspond to day to day concerns is successful if analysis is focused on ceramics used in everyday activities. Archaeologists have long turned to domestic remains to address a wide range of social and everyday concerns. These concerns pertain to the physical process of site formation and the anthropological process of cultural, social, and economic organization ( DeBoer et al. 1996 ). The popularity of using domestic remains to address these l as the ubiquity of the domestic archaeological record, usually in the form of ceramics. Besides nuclear families, households are often described as basic units of human societies, and domestic remains are the common archaeological indicator of such units In addition,
280 households often leave a distinct archaeological signature and are thus relatively easy to recognize both through soil analysis and patterned refuse. Household Archaeology Household archaeology is typically approached as settlement pattern archaeology. This type of study generally attempts to reconstruct social and economic organization through analysis of spatial distribution and the grouping of dwellings. Differentiation in wealth, social status, and domestic activities are all gleaned fro m this type of study and aid in its conclusions. While the opportunity to have a direct historical comparison in the Kuikuru village affords the opportunity to inform and focus archaeological investigation of prehistoric households, it also risks blur ring and confusing interpretation of the archaeological data. Without embracing the Pompeii Premise we can assume a certain degree of replication between human practice in the past and patterned refuse in the archaeological record. In places like the Upper Xing u ethnographic data can help to resolve questions whose answers are not found in the archaeological record. As well, ethnographic observation can skew our reading of the past if we are not careful to account for those elements of the ethnographic present that were not a part of the archaeological past but have instead emerged from the influence of historical factors Challenging the search fo r universal categories of micro scale social units, such as the nuclear family and the house, and recognizing the va riability of the social context of domestic action facilitates a more open dialogue with the past that doesn't assume that what we see ethnographically is precisely what we should see prehistorically. That said, the overall model provided by research in t he Upper Xingu today does suggest a broad underlying principle of social organization that is replic ated on a macro scale, or within and between those units of analysis above the household level (house, village, polity,
281 regional system). The question is wh ether or not these underlying princ iples are replicated on a micro scale, or within and between those units of analysis that fall below the village level, such as the household. A s other research has demonstrated, exploring domestic processes through the household does not necessarily need to involve excavating dwelling remains (Bermann 1994). In fact, equat ing architectural remains with households could even hinder the study of prehistoric life, particularly for those societies in which co residential gro ups did not coincide with dwellings. In the Upper Xingu this example is true especially for the groups that perfo rmed domestic tasks in socially oriented sub plazas representing large extended families Domestic organization and activities in such areas ma y have little to do with the houses in which people slept and tell us more about the social unit for which their labor was directed. Investigations of domestic processes should involve excavation away from residential dwellings and information gathered fro m what are usually co nsidered non domestic contexts and house arrangements (S 1983). This is further supported by direct observations made in the Kuikuru village. These e thnographic observations include d complete inventories of the ceramic and metal vesse l assemblages for representative houses from each quadrant of the Kuikuru village ( Figure 7 1 ) Vessel inventories and relationships between houses in these quadrants revealed that certain types of vessels related to manioc proces sing were shared among related residences Although each house could establish its own vessel inventory within the ir dwelling, when it came to vessels in shared areas t here was som e difficulty in attributing vessel ownership to a particular house
282 Figure 7 1 A full vessel inventory for a single non chiefly Kuikuru house. A Type 4 griddle is in the foreground, a large Type 1 ahukugu is decorated to the right, and five Type 2 vessels of various sizes sit beh ind the griddle, four of them blackened from cooking activity. The cooking area in the background contains examples of Type 1 and Type 4 vessels both suspended over fire. A stack of manioc filled sacks is seen in the upper right. Additional metal vessels a re stored behind the house, used exclusively for processing manioc. Vessels directly related to manioc processing are exclusively located in shared areas outside the house where the occupants of a few houses are jointly in possession of the se vessels ( Figure 7 2 ). C lustering of domestic activity between houses is a key link between the ethnographic realities of the Upper Xingu and the archaeological remains How these h ouse clusters might have looked in the prehistoric Upper Xingu relates directly to identifying analysis groups. Q uestions about scale and complexity can also be addressed since the appearance of modern day villages is a scaled down version of that present prior to historic population denigration (Heckenberger 1996).
283 Figure 7 2 This manioc processing area is shared by two households and is located in the backyard between the two houses. Two Type 1 ahukugu sit over fires cooking kuigiku and several metal vessels are sca ttered throughout the area. In addition to these bridging ethnographic observations regarding human practice, domesticity and social organization there are taphonomic and sampling issues that must be reconcile d to validate any direct historical compariso n. Distribution of specific vessel types throughout the archaeological sites suggest that broad horizontal sampling is not biased towards specific activity areas. Additionally, the identification excavatio n, and analysis of ceramic remains from House 1 an d House 2 show that with the exception of Type 4 vessels, there is very little clustering of vessel types in relation to specific activity areas. This archaeological data conversely fits well with further ethnoarchaeological observations made between 2004 and 2005. Generalized m apping of modern Xinguano houses took place both during occupation in 2004 and immediately
284 after several houses within the village burned and were subsequently abandoned in 2005 Mapping of the burned houses also took place one year after their destruction, noting the process of reuse of materials, abandonment of materials, and the fate of those abandoned materials. Abandoned house s were mostly free of ceramic and other domestic remains. Even the burned remains of house posts were rem oved and reused. The most characteristic remnants were the in filled holes from the posts and the trenches around the outside of the house where several rows of various sized wall posts had been located. This area also contained the most accumulated debris from daily house sweeping and the partial burying of the walls on the outside from windblown soil and intentional burying to keep water out during the rainy season. As outlined here, direct ethnographic observations bring much to bear on the archaeologica l record regarding the identification of dwellings, social groups and their organization at the house and village level. The house and its related domestic space used as a unit of analysis can be applied to the archaeological record to denote the possible space occupied and used by a household in areas where archaeological houses are not immediately visible by the distribution of archaeological features. For example, dividing the village into pie shape wedges, as is indicative of the modern Xinguano villag e, include s b ackyard trash middens and plaza side activity areas, for which very little refuse is found. Given the presumed larger populations of the prehistoric Upper Xingu this method may need to be altered slightly to account for clusters of houses that do not adhere to the historically scaled down modern village's pie shaped scenario.
285 Still, this division, or some relative form of it, can be overlai n on the archaeological village using the observations from the Kuikuru village and the single excavated H ouse 1 at MT FX 06 as a guide. The extremely low coun ts of non domestic pottery may situate these vessels as key indic ators of non domestic areas or the location of specialized places or persons. Despite these possibilities, domestic wares are the main foc us of this study and provide the context needed to divide ceramics collected archaeologically into meaningful groupings that can be analyzed for the detection of social bodies contained within houses leading then to the further division of activities and s ocial distinction within households. Pottery and Technology The Upper Xingu is a perfect case study for examining change in ceramic technology precisely because it is a long, continuous tradition, with recognizable potte ry forms and decorative styles that can be traced and compared through a more than 1,000 year tradition which includes modern ethnographic examples. The Upper Xingu also offers a well documented ethnographic and ethnohistoric record that compliments archaeological research. Upper Xingu ceram ics are sparsely decorated which situates variation in ceramic technology as a better indicator of diachronic societal development. This leaves difference and similarity in pottery decoration as a better indicator of synchronic societal variation especiall y late in time when technological developments seem to plateau and village and regional complexes exhibit distinct decorative styles and techniques. While both technological and decorative attributes can be used to address chronology and identify regional systems of interaction, t he main focus of this study is on techn ological variation, or change, through time. Technological c hange is tracked
286 here as change in vessel attributes that affect performance, acknowledging that new insights can be provided about the causes and consequences of ceramic change if one focuses on vessel performance characteristics (Skibo 1994:113). Performance characteristics are accounted for in vessel construction including rim form, overall size, and paste recipes Because nothing is abandoned or replaced along the trajectory of Upper Xingu pottery development, technological attributes can be followed through time and at some point conjoined with societal changes as understood through traditional archaeological research such as reg ional survey and mapping as well as village level household archaeology. This is all accomplished with the understanding that "t he best designed study of even the smallest technological c hange must eventually take the 'leap of faith' when trying to explain the proces s of interest" (Skibo 1994:113). The process es of interest in this study are two; the process of technological change in pottery and the processes operating within the social structure in which the technological change occurred. These are examin ed under the premise that t he variation in ceramic technology found in the archaeological ceramics of the Upper Xingu is inextricably tied to the social and political developments reflected in village plaza organization and the contemporaneous centralizati on of political and social control over labor. Labor and Society One of the benefit s of working in Amazonia is the broad scope of ethnographic research, conducted in various forms for well over 100 years, available in almost all areas of the region includi ng the Upper Xingu Recent and historical ethnographies of the people of the Upper Xin gu provide important elements for this study and allow for a direct historical approach to be combined with the technofunctional analysis of pottery presented here. A t th e core of this study is an explanation for the observed variation in
287 pottery technology through time. Part of that explanation is found in the specific tasks that pottery was implemented in achieving; the processing of manioc and the cooking of fish. The t wo main vessel forms, the ahukugu and the atange follow these functions. Pottery in the Upper Xingu was produced as a tool for those processing manioc and those cooking fish. More importantly, the demand for increased amounts of manioc resulted in the inc reased demand for labor to process manioc and thus to produce more pottery to be used in production of manioc. In exploring the role of labor, the role of technology must remain neutral, as one aspect, or indicator, and not as the reason for other factors. Jona than Friedman suggests navigating this dilemma by enforcing that the "relations of production are not The process of historical development depends on the relation between 1974:450). The relations between those who make and use pottery, and those who provoke or compel these actions, are the relations of production in this study. Further explained; We must always distinguish the technological from the social process of reprod uction. Relations of production of those social relations which dominate (i.e. determine the economic rationality of) the material process of production in given techno ecological conditions at a given stage of development of the forces of production (Frie dman 1974:446). Following this structure of reasoning, the social relations between the physical processors of manioc and those who demand the processing are the relations that directly affect the tools of production, especially in a society where those wh o process manioc and those who produce pottery are the same people, the women, as is the case in the Upper Xingu ( Carneiro 1983; Dole 1978; Oberg 1948).
288 Those who demand processing in the Upper Xingu, or those who control labor, are These individuals are imbued with a social and symbolic capital that allows them to demand proce ssing. As Heckenberger surmises; Some individuals have a capacity, due to their accumulated symbolic capital, to transform symbolic power into political or ec onomic power that far outstrips that of others. This is particularly evident in terms of labor control, the true measure of chiefly power ( Heckenberger 2003:39) Though it has long been argued that a surplus of food is required for increased social complex ity across the globe, in the Upper Xingu the model suggests that a surplus of social capital is perhaps more critical and required first. Specifically, this social capital, usually a chiefly lineage or ownership of ritual privelage is converted to the abi lity to command labor (Heckenberger 2003) Commanding non subsistence related labor, such as thatching a roof, requires that food be supplied to the laborers ( Figure 7 3 ). This food is produced by the wives and daughters of the ma n commanding the labor. In this way the resource of labor, both to produce food, and bought with food, is the ultimate form of wealth. Ag ain, as Heckenberger points out; In Amazonia, there is little evidence of agricultural surplus or hoarding of wealth by elite families. If there is something controlled it is labor, not material goods (Heckenberger 2003:11). The relationship between those in power and those whose labor produces material wealth is found between the existence of potential l abo r and the autho rity to demand use of that labor. Part of the ability to command labor begins with the ability to accumulate laborers within your house. Negotiating residence patterns are part of the ocess manioc.
289 Figure 7 3 Community roof thatching project where all laborers will be paid with meals consisting of manioc and fish. Pottery use to process manioc is situated as a n object, a subj ect, and a product of labor. In these roles it is well placed to reveal the nature of those relationships between labor and its use in production. Pottery is an object of labor because it is used by laborers in the processing and production of manioc. It is a subjec t of labor because it reflects the demands of those who use it and those who manufacture it. Finally, pottery is a product of labor both as the object produced by potters and the subject transformed by the influence of labors necessity; it is the final pro duct of several processes. Since women are both the laborers who produce the pottery and the laborers who use the pottery, sending the message to the producers that the users need a larger more durable vessel is a simple transmission. Ethnographic work in the Upper Xingu suggests that overall as symbolic
290 capital (Heckenberger 2003; Menget 1993). Among the Kuikuru this is mostly demonstrated by his ability to compel women to produce surplus manioc. In the p ast this power likely extended to major community projects such as the excavation of the ditches surrounding MT FX 06 but through time that power has somewhat waned. The question remains, however, whether chiefs, even though pregnant with symbolic capital can transform this into economic capital, in the form of goods and labor, to ac hieve greater political power" (Heckenberger 2003:36). In 1973 even where technology gave an impetus to the rise of civilization, it can be shown involved were themselves developed in response to societal demands In the Upper Xingu, three main lines of evidence suggest this is true. First, an examination of the variation in pottery through time second, ethnograp hic evidence regarding the use of pottery, and third, landscape and village configuration suggesting a rise in population and an elaboration of social organization. Car neiro quotes Robert Adams in a statement that sums up the situ ation in the Upper Xingu n icely, by saying that changes in social institutions that precipitated changes in technology, subsistence and e versa" (in Carneiro 1973:179). Summary This study is driven by two br oad research questions First, in what ways can we study archaeological ceramics to learn more about the society that produced and used them and second, what can archaeological ceramics tell us about prehistoric social dynamics in Amazonia and the Upper Xi ngu? To answer these questions this study focused on two prehistoric villages in the Upper Xingu of the Southern Amazon occupied between ca. A.D. 700 1770 and ethnographic data from the historic period
291 from roughly 1884 to the present. These dates are part icularly important to this study for two reasons First, it is around A.D. 700 or earlier that widespread movements of prehistoric people have been documented archaeologically in the Amazon basin and second, it is document ed that between A.D. 1250 and 1500 these villages were fortified with newly constructed defensive ditches both in the Upper Xingu and the Central Amazon. We may now perhaps refine the research question to this; does the archaeological ceramic assemblage from two prehistoric sites in the U pper Xingu reflect the dramatic activities that occurred at the beginning and end of the occu pation of these sites? At what point does pottery production in the Upper Xingu move beyond simple household production and become a household industry ( Rice 19 87: 184; Feinman 1999). This basic connection between pottery development and complex societies remains a central theme for many archaeologists; For archaeologists this subject has been important because, despite the operational problems of identifying special ists in prehistory, economic specialization in production and distribution is generally acknowledged to be a concomitant of large, complex, highly differentiated societies and to depend on other intensive production arrangements, for example, in agricultur e (Dow 1985; in Rice 1987). Given this generally accepted principle, the Upper Xingu provides the perfect ca se study to test this hypothesis by examining changes between the Initial Period and the Developmental Period It is in the Developmental Period tha t Heckenberger has documented the expansion of large fortified villages, connected by roads, and the sudden appearance of defensive ditches around these villages. As well, satellite imagery shows marked forest alteration around prehistoric villages dated t o this period.
292 All of this suggests that in the Developmental Period populations were large, horti culture was intense, and socio economic complexity was reaching new heights. Some of this change in production level can be attributed to the waves of Europea n contact in more recent times beginning at A.D. 1500 on the coasts and rapidly spreading along river routes However, much of this change occurs well before European contact and in the context of broader environmental and social changes happening througho ut Amazonia These changes are well documented in other areas around the end of the first mil lennium A.D In the U pper Xingu we see these changes with the initial sett lement of the area around A.D. 700 and with the growth and fortification of these settlem ents from about A.D. 1250 to 1 500 Caught in the maelstrom of change, or in fact, reflective of this change, is the Upper Xingu ceramic industry. At first this production can be classified as one of household or village level production and later perhaps a s a regional level of production (Feinman 1999; Rice 1987; Longacre 1991). Though the period between household and village production is not well understood apart from other periods something clearly caused the specialization of ceramic production to be f ocused among one group of people It is possible that each village may have been responsible for producing different ceramics until the intense focus on manioc production caused the decline of production variation in ceramics among other groups and a monop oly on production was created by the group responsible for producing the vessels needed for the processing of manioc. Indeed, this group, the Waur is one of only three Arawak groups in the area, the language group hypothesized to have arrived in the regi on first and in this case likely to have reached a
293 local level of sophistication in manioc production before other invading groups, such as the Carib, Paraci, and others arrived. Further evidence of this relationship may be found in how man ioc is processed in this region which as others have pointed out, is entirely unique to the Upper Xingu. The process of using tuafi's rather than relies on the use of the large low collared ahukugu vessels ; in fact this type of pro cessing would be impossible or v ery difficult and inefficient with a smaller diameter, pedestal base vessels. In other parts of the Amazon a technique using the tipiti a long woven tube that is hand squeezed and left to drain does not require such specific vessels and there are no corr elates to this vessel in other parts of the Amazon to the north (Carneiro 2000) Observable changes in village organization also illustrate the changes of the early prehistoric Upper Xingu. Heckenberger presents evidence for the construction of defensive d itches during the Developmental Period at both MT FX 06 and MT FX 11 beginning around A.D. 1250 He notes the construction of defensive ditches and corresponding roads and berms, suggesting that the internal constructions and the resultant village configur ation were the "developmental apogee" of the concentric circular village pattern seen today in the region (Heckenberger 1996:97). Hecken berger goes further to say that "the pie shaped partitioning of the site created discrete domestic precincts or 'neighbo rhoods' situated between roads and also delimited by the plaza and excavated ditches (Heckenberger 1996:97). He further suggests that within each "were undoubtedly much la rger, more clearly defined and more internally complex (Hec kenberger 1996:97). Given the undoubtedly more complex social divisions and their clearly defined hierarchy we should expect to see some reflection of this in the
294 material culture of these groups or factions across the village unit depending on the l evel of manufacture of pottery. In characterizing the development of Xinguano culture during prehistory, three aspects of the ceramic technology remain important. The first aspect of this industry is re lated to its function. The survival into modern times of a large vessel that is used in the processing of manioc suggests that the development of this technology was an important process for the subsistence economy of the Xinguano. The second aspect of the ceramic industry is the structure of social relations involved in its production. Today, a single tribe is responsible for the manufacture of pottery for several separate tribes. Understanding the scope and timing of this development from presumed househo ld production to village production is a key factor in understanding the development of the Xinguano society. Finally, contextualizing ceramic technology and its development in the Upper Xingu with developments elsewhere in the Amazon situate s the Xingu in to a wider context of regional and local influencing factors In summarizing the results of this analysis three clear transitions can be documented from the Initial Period through the Protohistoric Period First, a clear preference for cauxi or sponge spi cule temper appears to originate in the later part of the Initial P eriod and continues to increase through the Developmental Period while other tempering agents decrease or remain the same. Second, the overall variety in vesse l forms diminishes in the late Initial P eriod and continues into the Developmental Period where Type 1 manioc cooking and processing vessels become more common. Third, as the preference for Type I vessels in the Developmental Period continues their rims become more standardized and inc rease in thickness. Vessel size measured by
295 oral diameter also increases creating an overall more robust vessel where the average rim thickness has almost doubled. A discussion of these transitions in ceramic technology can be placed within the general out line of Upper Xingu culture history specifically in relation to three major developments. First, while earlier dates may exist in the Upper Xingu, research thus far suggests that the earliest settlement of this area by pottery produci ng groups occurred aro und A.D. 7 00. Second, between A.D. 700 and A.D. 1500 a transition in the manufacture of pottery, significantly the transition to flat bottomed vessels, large amounts of cauxi temper, and gradually out flaring rims, a vessel essential to manioc processing, suggests a concurrent and dramatic transition in subsistence which is taken to be more intensive processing and accumulation of bitter manioc. Third and finally between A.D. 1250 and A.D. 177 0, Upper Xingu pottery demonstrates a marked process of standar dization that coalesced concurrently with the first expression of social complexity found in the archaeological visibility of a connected network of circular plaza villages. Given the connection between manioc processing and its specialized pottery, these developments suggest that manioc production played a key role in the expansion and development of the social complexity of the region as expressed in the landscape transformations and standardization of ceramics. In relating this to the rest of the Amazon we can first look to the Central Amazon, where many suppose a settled Amazonia n agricultural way of life first developed. Here a continuous complex of pottery is documented uninterrupted from 350 B.C. until A.D. 750 when a new complex of pottery appears. T his noticeable pattern of change concurrently appearing across the Amazon still lacks a definitive explanation although environmental pressure, population
296 pressure, subsistence changes, and social transformations have all been suggested, often in relation to each other. Clearly, m uch has changed over the last 1 200 years in the Upper Xingu. Settlement patterns grew from small single plaza villages to larger village complexes connected by networks of roads. The ceramic industry changed from one of greater va riability to one of great homogeneity with increasing vessel sizes likely related to chiefly demand on labor. U nfortified circular villages became larger fortified villages and again in recent times singular, isolated, non fortified villages. Finally, disc ussions of ceramic production are common in the archaeological literature of complex societies and urbanism, where st udies of craft specialization, and standardization are considered in relation to political complexity T he Upper Xingu case fall s within t hese discussions of standardization, specialization, and routinization In the Upper Xingu there is a direct correlation between the development of ceramic production and the development of complex social organization. This is most clearly illustrated in t he transformation of pottery which parallels the elaboration of regional and village level social complexity The role that ceramics played in the intensified producti on of pottery squarely in the center of any discussion about the development of complex social and political organization in the Upper Xingu The timing of these events and their seeming ubiquity across the Amazon raises still more questions for Amazonian archaeologists.
297 LIST OF REFERENCES Agostinho da Silva, Pedro 1974 Kwarp : mito e ritual no Alto Xingu Editora Pedaggica e Universitria, So Paulo. 1993 Testemunhos da ocupao pr xinguana na bacia dos formadores do Xingu In Karl von den Steinen: Um Sculo de Antropologia no Xingu edited by V. P. Coelho, pp. 233 287. Edusp, So Paulo. Anthony, David W. 1990 Migration in Archaeology: The Baby and the Bathwater. American Anthropologist 92(4):895 914. Arnold, Dean E. 1975 Cera mic Ecology of the Ayacucho Basin, Peru: Implications for Prehistory. Current Anthropology 16(2):183 194. 1985 Ceramic Theory and Cultural Process Cambridge University Press, Cambridge. 1992 Comments on Section II In Chemical Characterization o f Ceramic Pastes in Archaeology edited by H. Neff, pp. 159 166. vol. Monographs in World Archaeology. Prehistory Press, Madison. Bale, William 1995 Historical Ecology of Amazonia In Indigenous Peoples and the Future of Amazonia : An Ecological Ant hropology of an Endangered World edited by L. E. Sponsel, pp. 97 106. University of Arizona Press, Tucson and London. 2006 The Research Program of Historical Ecology. Annual Review of Anthropology 35:75 98. Bale, William and Clark L. Erickson 2006 Time, Complexity, and Historical Ecology In Time and Complexity in Historical Ecology: Studies in Neotropical Lowlands edited by W. Bale and C. L. Erickson, pp. 1 17. Columbia University Press, New York. Barker, Andrew, Barney Venables, Stanley M. Stevens Jr., Kent W. Seeley, Peggy Wang and Steve Wolverton 2011 An Optimized Approach for Protein Residue Extraction and Identification from Ceramics After Cooking. Journal of Archaeological Method and Theory (Published Online 8 September). Bar reto, Cristiana 1998 Brazilian archaeology from a Brazilian perspective. Antiquity 72:573 581.
298 Basso, Ellen B. 1973 The Kalapalo Indians of Central Brazil Holt, Rinehart and Winston, New York. 1977 The Carib Speaking Indians: Culture, Society and Language 28. University of Arizona Press, Tuscon. 1984 A Husband for His Daughter, a Wife for Her Son: Strategies for Selecting a Set of In laws among the Kalapalo In Marriage Practices in Lowland South America edited by K. M. Kens igner, pp. 33 44. University of Illinois Press, Urbana. 1995 The Last Cannibals: A South American Oral History University of Texas Press, Austin. Becquelin, Pierre. 1993 Arqueologia xinguana In Karl von den Steinen: um sculo de antropologi a no Xingu edited by V. P. Coelho, pp. 224 232. Edusp, So Paulo. Blackman M. James, G. J. Stein, P. B. Vandiver 1993 The Standardization Hypothesis and Ceramic Mass Production: Technological, Compositional, and Metric Indexes of Craft Specializat ion at Tell Leilan, Syria American Antiquity 58(1):60 80 Blitz, John H. 1993 Big Pots for Big Shots: Feasting and Storage in a Mississippian Community. American Antiquity 58(1):80 96. Boomert, Arie 2000 Trinidad, Tobago, and the Lower Orin oco Interaction Sphere: An Archaeological/Ethnohistorical Study Bozarth, S. R., K. Price, W. I. Woods, E. G. Neves and R. Rebellato 2009 Phytoliths and Terra Preta : The Hatahara Site Example In Amazonia n Dark Earths: Wim Sombroek's Vision edited by W. I. Woods, W. G. Teixeira, J. Lehmann, C. Steiner, A. M. G. A. WinklerPrins and L. Rebellato. Springer Science and Business Media, New York. Braun, David P. 1983 Pots as Tools In Archaeological Hammers and Theories edited by J. A. Moore and A S. Keene, pp. 108 134. Academic Press, New York. Bronitsky, Gordon 1989 Introduction In Pottery Technology: Ideas and Approaches edited by G. Bronitsky, pp. 1 11. Westview Press, Boulder.
299 Bronitsky, Gordon, and R. Hamer 1986 Experiments in Ceramic Technology: The Effects of Various Tempering Materials on Impact and Thermal Shock Resistance. American Antiquity 51:89 101 Brumfiel, Elizabeth M. 2000 On the Archaeology of Choice: Agency as a Research Stratagem In Agency in Arch aeology edited by M. A. Dobres and J. E. Robb, pp. 249 255. Routledge, London. Buck, C. E. 1993 The provenancing of archaeological ceramics: a Bayesian approach In Computing the Past: Computer Applications and Quantitative Methods in Archaeology edited by J. Andresen, T. Madsen and I. Scollar, pp. 293 301. vol. CAA92. Aarhus University Press, Aarhus. Carneiro da Cunha, Manuela (editor) 1992 Histria dos ndios no Brasil Companhia das Letras: Secretaria Municipal de Cultura: FAPESP, So Pau lo. Carneiro, Robert L. 1960 Slash and Burn Agriculture: A Closer Look at Its Implications for Settlement Patterns In Men and Cultures edited by A. F. C. Wallace, pp. 229 234. University of Pennsylvania Press, Philadelphia. 1961 Slash and b urn cultivation among the Kuikuru and its implications for cultural development in the Amazon Basin In Antropologica, Supplemental Publication 2, The evolution of horticultural systems in native South America, causes and consequences: a symposium edited by J. Wilbert, pp. 47 67. 1970 A Theory of the Origin of the State. Science 169:733 738. 1972 From Autonomous Villages to the State, a Numerical Estimation In Population Growth: Anthropological Implications edited by B. Spooner, pp. 64 77. Massachusetts Institute of Technology, Cambridge. 1977 Recent Observations of Shamanism and Witchcraft among the Kuikuru Indians of Central Brazil. Annals of the New York Acadamy of Science 293:215 228. 1978a The Knowledge and Use of Rain Fo rest Trees by the Kuikuru Indians of Central Brazil In The Nature and Status of Ethnobotany edited by R. I. Ford, pp. 201 216. Anthropological Papers No. 67. Museum of Anthropology, University of Michigan, Ann Arbor. 1978b Some Speculations on t he Culture History of the Upper Xingu Proceedings of the Annual Meeting of the American Anthropological Association
300 1983 The Cultivation of Manioc Among the Kuikuru Indians of the Upper Xingu In Adaptive Responses in Native Amazonia ns edited by R B. Hames and W. T. Vickers, pp. 65 111. Academic Press, New York. 1987a Further Reflections on Resource Concentration and Its Role in the Rise of the State. British Archaeological Reports (BAR) International Series No. 349 :245 260. 1987b Vi llage Splitting as a Function of Population Size In Themes in Ethnology and Culture History, Essays in Honor of David F. Aberle edited by L. Donald. Folklore Institute, Meerut, India: Archana Publications. 1989 To the village of the jaguars: the M aster Myth of the Upper Xing. Antropologica 72:3 40. 2000 The Evolution of the Tipiti: A Study in the Process of Invention In Cultural Evolution: Contemporary Viewpoints edited by G. M. Feinman and L. Manzanilla, pp. 61 93. Klewer Academic/Plenum Publishers, New York. Carsten, Janet 1995 The substance of kinship and the heat of the hearth: feeding, personhood, and relatedness among Malays in Pulau Langkawi. American Ethnologist 22(2):223 241. Carsten, Janet and Philip Hugh Jones (editors) 1995 About the House: Lvi Strauss and Beyond Cambridge University Press, Cambridge. Carsten, Janet and Stephen Hugh Jones 1995 Introduction In About the House: Lvi Strauss and Beyond edited by J. Carsten and S. Hugh Jones, pp. 1 46. Camb ridge University Press, Cambridge. Carvalho, Jos Cndido M. 1951 Relaes Entre os Indios do Alto Xingu e a Fauna Regional Publicaes Avulsas do Museu Nacional. Oficina Grfica da Universidade do Brasil, Rio de Janeiro. Chilton, Elizabeth S. (ed itor) 1999 Material Meanings: Critical Approaches to the Interpretation of Material Culture The University of Utah Press, Salt Lake City. Cleary, David 2001 Towards an Environmental History of the Amazon: From Prehistory to the Nineteenth Ce ntury. Latin American Research Review 36(2):65 96.
301 Coelho, Vera Penteado 1981 Alguns Aspectos Da Ceramica Dos Indios Waur In Contribuicoes A Antropologia em Homenagem Ao Professor Egon Schaden edited by T. Hartmann and V. P. Coelho, pp. 55 83. U niversidade de Sao Paulo, Sao Paulo. Cohen, G. A. 1978 Karl Marx's theory of history : a defence Princeton University Press, Princeton. Correa, Conceicao Gentil 1965 Estatuetas de Ceramica na Cultura Santarem. ) 4:3 25. Costa, Maria Helosa Fnelon and Hamilton Botelho Malhano 1987 Habitao Indgena Brasileira In Suma: Etnolgica Brasileira 2: Tecnologia Indgena edited by B. G. Ribeiro. Financiado ra De Estudos E Projetos, Petrpolis. Costa, Marcondes Lima da, Dirse Clara Kern, Alice Helena Eleoterio Pinto and Jorge Raimundo da Trindade Souza 2004 The ceramic artifacts in archaeological black earth (terra preta) from lower Amazon region, Braz il: Mineralogy. Acta Amazonica 34(2):165 178. Costin, Cathy Lynne 1998 Introduction: Craft and Social Identity In Craft and Social Identity edited by C. L. Costin and R. P. Wright, pp. 3 16. Archaeological Papers of the American Anthropological As sociation Number 8, Arlington. Curet, Luis Antonio 1992 House Structure and Cultural Change in the Caribbean: Three Case Studies from Puerto Rico. Latin American Antiquity 3(2):160 174. David, Nicholas and Hilke Hennig 1972 The Ethnography o f Pottery: A Fulani Case Seen in Archaeological Perspective Addison Wesley Modular Publications Module 21. Addison Wesley Publishing Company, Reading, Mass. DeBoer, Warren R. 1974 Ceramic Longevity and Archaeological Interpretation: An Example from the Upper Ucayali, Peru. American Antiquity 39(2):335 343. 1984 The Last Pottery Show: System and Sense in Ceramic Studies In The Many Dimensions of Pottery edited by S. E. Van Der Leeuw and A. C. Pritchard, pp. 529 571. Universiteit van Amsterd am, Amsterdam.
302 1985 Pots and pans do not speak, nor do they lie: the case for occasional reductionism In Decoding Prehistoric Ceramics edited by B Nelson, pp. 347 357 Southern Illinois University Press Carbondale 1986 Pillage and production in the Amazon: a view through the Conibo of the Ucayali Basin, eastern Peru. World Archaeology 18(2):231 246. 1990 Interaction, imitation, and communication as expressed in style: the Ucayali experience In The Uses of Style in Archaeology edited by M. Conkey and C. Hastorf, pp. 82 104. Cambridge Univesity Press, Cambridge. DeBoer, Warren R., Keith Kintigh and Arthur G. Rostoker 1996 Ceramic Seriation and Site Reoccupation in Lowland South America. Latin American Antiquity 7(3):263 278. DeB oer, Warren R. and Donald W. Lathrap 1979 The Making and Breaking of Shipibo Conibo Ceramics In Ethnoarchaeology: Implications of Ethnography for Archaeology edited by C. Kramer, pp. 102 138. Columbia University Press, New York. Deetz, James 1 965 The dynamics of stylistic change in Arikara ceramics Illinois studies in anthropology, no. 4. University of Illinois Press, Urbana,. Denevan, William M. 1970 Aboriginal Drained Field Cultivation in the Americas. Science 169(14 August):647 654 1976 The Aboriginal Population of Amazonia In The Native Population of the Americas in 1492 edited by W. Denevan, pp. 205 234. University of Wisconsin, Madison. 1991 Prehistoric Roads and Causeways of Lowland Tropical America In Ancient Road Networks and Settlement Hierarchies in the New World edited by C. D. Trombold. Cambridge University Press, Cambridge. 1992 Stone vs. Metal Axes: The Ambiguity of Shifting Cultivation in Prehistoric Amazonia Journal of the Steward Anthropologi cal Society 20(1+2):153 165. 1996 A Bluff Model of Riverine Settlement in Prehistoric Amazonia Annals of the Association of American Geographers 82:369 385. 1998 Comments on Prehistoric Agriculture in Amazonia Culture & Agriculture 20(2/3): 54 59.
303 2004 Semi intensive Pre European Cultivation and the Origins of Anthropogenic Dark Earths in Amazonia In Explorations in Amazonia n Dark Earths edited by W. I. Woods and B. Glaser, pp. 135 143. Springer, Verlag Berlin Heidelberg. 2006 Pre European Cultivation in Amazonia In Time and Complexity in Historical Ecology: Studies in the Neotropical Lowlands edited by W. Bale and C. L. Erickson, pp. 153 163. Columbia University Press, New York. Dole, Gertrude E. 1956 Ownership and E xchanges among the Kuikuru Indians of Mato Grosso. Revista do Museu Paulista 10:125 133. 1960 Techniques of Preparing Manioc Flour as a Key to Culture History in Tropical America In P roceedings of the Fifth International Congress of Anthropological and Ethnological Sciences edited by A. F. C. Wallace University of Pennsylvania Press, Philadelphia 1961 A Preliminary Consideration of the Prehistory of the Upper Xingu Basin. Revista do Museu Paulista XIII:399 423. 1964 Shamanism and Poli tical Control Among the Kuikuru In Peoples and Cultures of Native South America edited by D. R. Gross, pp. 294 307. Natural History Press, New York. 1978 The Use of Manioc Among the Kuikuru: Some Interpretations In The Nature and Status of Ethnob otany edited by R. I. Ford, pp. 217 247. 1983 Some Aspects of Structure in Kuikuru Society. Antropologica 59 62:309 329. 1984 The Structure of Kuikuru Marriage In Marriage Practices in Lowland South America edited by K. M. Kensinger, pp. 45 62. University of Illinois Press, Urbana and Chicago. 1991 The Development of Kinship in Tropical South America In Profiles in Cultural Evolution edited by A. T. Rambo and K. Gillogly, pp. 373 403. vol. Anthropological Papers No. 85. Museum of Anthropology, University of Michigan, Ann Arbor. Drennan, Robert D. 1995 Chiefdoms in Northern South America. Journal of World Prehistory 9(3):301 340. Duff, Andrew I. 1996 Ceramic Micro Seriation: Types or Attributes? American Antiquity 61(1 ):89 101.
304 Duin, Renzo 2000 A Wayana Potter in the Tropical Rain Forest of Suriname/French Guyana. Newsletter of the Department of Pottery Technology (Leiden University) 18/19:45 58. Ember, Melvin 1973 An Archaeological Indicator of Matriloca l versus Patrilocal Residence. American Antiquity 38(2):177 182. Erickson, Clark L. 1995 Archaeological Perspectives on Ancient Landscapes of the Llanos de Mojos in the Bolivian Amazon In Archaeology in the American Tropics: Current Analytical Meth ods and Applications edited by P. Stahl, pp. 66 95. Cambridge University Press, Cambridge. 2000 An Artificial Landscape Scale Fishery in the Bolivian Amazon. Nature 408(9 November):190 193. 2006 The Domesticated Landscape of the Bolivian Am azon In Time and Complexity in Historical Ecology: Studies in the Neotropical Lowlands edited by W. Bale and C. L. Erickson, pp. 235 278. Columbia University Press, New York. Erickson, Clark L. and William Bale 2006 The Historical Ecology of a C omplex Landscape in Bolivia In Time and Complexity in Historical Ecology: Studies in the Neotropical Lowlands edited by W. Bale and C. L. Erickson, pp. 187 234. Columbia University Press, New York. Evans Jr., Clifford 1967 Amazon Archaeology: A C entennial Appraisal Proceedings of the Atas do Simpsio sbre a Biota Amaznica 2:1 12. Evans Jr., Clifford and Betty J. Meggers 1950 Preliminary Results of Archaeological Investigations at the Mouth of the Amazon. American Antiquity 16(1):1 9. 1962 Use of Organic Temper for Carbon 14 Dating in Lowland South America. American Antiquity 28(2):243 244. Evans Jr., Clifford, Betty J. Meggers and Jos M. Cruxent 1959 Preliminary Results of Archaeological Investigations Along the Orinoco and Ventuari Rivers, Venezuela Proceedings of the Actas del XXIII Congresso Internacional de Americanistas II :359 369. San Jose. Fausto, Carlos and Michael Heckenberger 2007 Time and memory in indigenous Amazonia : anthropological perspectives Un iversity Press of Florida, Gainesville.
305 Feinman, Gary M., Steadman Upham and Kent G. Lightfoot 1981 The Production Step Measure: An Ordinal Index of Labor Input in Ceramic Manufacture. American Antiquity 46(4):871 884. Franchetto, Bruna and Michael Heckenberger 2000 Os povos do Alto Xingu : histria e cultura Editora UFRJ, Rio de Janeiro. Friedman, Jonathan 1974 Marxism, Structuralism and Vulgar Materialism. Man 9(3):444 469. Galvo, Eduardo 1953 Cultura e Sistema de Parentesco d as Tribos do Alto Rio Xingu. Boletim de Museu Nacional (Antropologia) 14:1 56. Galvo, Eduardo and Pedro Agostinho 1973 Indios do Brasil, reas culturais e reas de subsistncia Ncleo de recursos didticos 67. Centro Editorial e Didtico, Univ ersidade Federal da Bahia, Salvador, Bahia. Gaspar, Maria Dulce 1998 Considerations of the sambaquis of the Brazilian coast. Antiquity 72:592 615. Gassn, Rafael A. 2002 Orinoquia: The Archaeology of the Orinoco River Basin. Journal of World Prehistory 16(3):237 311. Godelier, Maurice 1977 Perspectives in Marxist anthropology Cambridge studies in social anthropology. Cambridge University Press, Cambridge ; New York. Gomes, Denise C. and Oscar Vega 1999 Dating Organic Temper of C eramics by AMS: Sample Preparation and Carbon Evaluation. Radiocarbon 41(3):315 320. Gonzlez, Erika Marion Robrahn 1998 Regional pottery making groups in southern Brazil. Antiquity 72:616 624. Gosselain, Olivier 1994 Skimming Through Potter' s Agendas: An Ethnoarchaeological Study of Clay Selection Strategies in Cameroon In Society, Culture, and Technology in Africa edited by S. T. Childs, pp. 99 107. MASCA Research Papers in Science and Archaeology, Supplement to Volume 11, University of P ennsylvania Museum of Archaeology and Anthropology, Philadelphia.
306 Gregor, Thomas A. 1977 Mehinaku: The Drama of Daily Life in a Brazilian Indian Village The University of Chicago Press, Chicago. 1990 Uneasy Peace: Intertribal Relations in Brazil's Upper Xingu In The Anthropology of War edited by J. Haas. Cambridge University Press, Cambridge. Gross, D. 1975 Protein Capture and Cultural Development in the Amazon Basin. American Anthropologist 77(3):526 554. 1983 Village Move ment in Relation to Resources in Amazonia In Adaptive Responses of Native Amazonia ns edited by R. B. Hames and W. T. Vickers, pp. 429 449. Academic Press, New York. Harris, Edward C., Marley R. Brown, III and Gregory J. Brown (editors) 1993 Practi ces of Archaeological Stratigraphy Academic Press, San Diego. Hartmann, Gnther 1986a Keramik des Alto Xing Zentral Brasilien Museum Fr Vlkerkunde Berlin, Berlin. 1986b Xing: Unter Indianern in Zentral Brasilien Dietrich Reimer Verlag, Berlin. Hartt, C.F. 1871 The Ancient Indian Pottery of Maraj Brazil. The American Naturalist V(5):259 271. Hayden, Brian 1984 Are emic types relevant to archaeology? Ethnohistory 31(2):79 92. Heckenberger, Michael J. 1996 War and Pe ace in the Shadow of Empire: Sociopolitical Change in the Upper Xingu of Southeastern Amazonia A.D. 1400 2000 Unpublished Ph.D. Dissertation, University of Pittsburgh, UNI, Ann Arbor. 1998 Manioc Agriculture and Sedentism in Amazonia : The Upper Xi ngu Example. Antiquity 72(277):633 648. 1999 Xinguano History and Hierarchy: Toward an Archaeology of Cultural Meanings Proceedings of the 98th Annual Meeting of the American Anthropological Association Chicago.
307 2001 Rethinking the Arawak an Diaspora: Hierarchy, Regionality, and the Amazonia n Formative In Comparative Arawak Histories: Rethinking Culture Area and Language Group in Amazonia edited by J. Hill and F. Santos Granero, pp. 99 122. University of Illinois Press, Urbana. 2003a The Enigma of the Great Cities: Body and State in Amazonia Tipiti: Journal of the Society for the Anthropology of Lowland South America 1(1):27 58. 2003b The Wars Within: Xinguano Witchcraft and Balance of Power In Assault Sorcery in Amazonia edited by N. L. Whitehead and R. Wright. Duke University Press, Durham (in press). 2005 The ecology of power : culture, place, and personhood in the southern Amazon, A.D. 1000 2000 Critical perspectives in identity, memory, and the built environmen t. Routledge, New York. 2006 History, Ecology, and Alterity In Time and Complexity in Historical Ecology edited by W. Bale and C. L. Erickson, pp. 311 340. Columbia University Press, New York. 2007 Xinguano Heroes, Ancestors, and Others: M aterializing the Past in Chiefly Bodies, Ritual Space, and Landscape In Time and Memory in Indigenous Amazonia : Anthropological Perspectives edited by C. Fausto and M. J. Heckenberger, pp. 284 311. University Press of Florida, Gainesville. 2009 Lost Cities of the Amazon: The Amazon tropical forest is not as wild as it looks. Scientific American 301(4):44 51. Heckenberger, Michael J., A. Kuikuru, U.T. Kuikuru, J. C. Russell, M. Schmidt, Carlos Fausto and Bruna Franchetto 2003 Amazonia 149 2: Pristine Forest or Cultural Parkland? Science 301:1710 1714. Heckenberger, Michael J. and Eduardo Ges Neves 2009 Amazonia n Archaeology. Annual Review of Anthropology 38:252 266. Heckenberger, Michael J. and James B. Petersen 1996 Circular Village Patterns in the Caribbean: Comparisons from Amazonia In Proceedings of the Seventeenth International Congress for Caribbean Archaeology Guadeloupe. Heckenberger, Michael J., James B. Petersen and Eduardo G. Neves 1999 Village Size and Per manence in Amazonia : Two Archaeological Examples from Brazil. Latin American Antiquity 10(4):353 376.
308 Heckenberger, Michael J., J. Christian Russell, Carlos Fausto, Joshua R. Toney, Morgan J. Schmidt, Edithe Pereira and Afukaka Kuikuru 2008 Pre Col umbian urbanism, anthropogenic landscapes, and the future of the Amazon. Science 321(5893):1214 1217. Heckenberger, M. J., J. C. Russell, J. R. Toney and M. J. Schmidt 2007 The Legacy of Cultural Landscapes in the Brazilian Amazon: Implications for Biodiversity. Philosophical Transactions of the Royal Society London B: Biological Sciences, Special Edition "Biodiversity Hotspots Through Time" 362(1478):197 208. Hemming, John 1978 Red gold : the conquest of the Brazilian Indians Harvard Univer sity Press, Cambridge, Mass. Heredia, Osvaldo R. 1994 Prehistory of the Non Andean Region of South America (Brazil, Paraguay, Uraguay and Argentina, 31,000 5,000 years ago) In History of Humanity edited by S. J. De Laet, pp. 328 341. Routledge, Ne w York. Hilbert, Peter Paul 1952 Contribuio arqueologia da Ilha de Maraj : os "tesos" Maraj aras do alto Camutins e a atual situao da ilha do Pacoval, no Arar Publicao (Instituto de Antropologia e Etnologia do Par) n. 5. Instituto de Ant ropologia e Etnologia do Par, Belm, Par, Brasil. 1955 A cermica arqueolgica da regio de Oriximin : (com 1 mapa e 49 estampas) Publicao (Instituto de Antropologia e Etnologia do Par) n. 9. Instituto de Antropologia e Ethnologia do Par, Be lm, Par, Brasil. 1957 Contribuio a arqueologia do Amap : Fase Arist Boletim do Museu Paraense Emlio Goeldi. Nova srie. Antropologia no. 1. Conselho Nacional de Pesquisas, Instituto Nacional de Pesquisas da Amazonia Museu Paraense Emlio Go eldi, Belm, Brasil. 1962a New Stratigraphic Evidence of Culture Change on the Middle Amazon (Solimes) Proceedings of the Akten des 34 Internationalen :471 476. Wien. 1962b Preliminary Results of Archaeological research on the Japur River, Middle Amazon Proceedings of the Akten des 34 Internationalen :465 470. Wien. 1968 Archologische Untersuchungen am mittleren Amazonas Reimer, Berlin,. Hill, Jonathan D. and Fernando Santos Granero (editors) 2002 Comparative Arawakan Histo ries: Rethinking Language Family and Culture Area In Amazonia University of Illinois Press, Urbana and Chicago.
309 Hill, James N. 1977 Individual Variability in Ceramics and the Study of Prehistoric Social Organization In The Individual in Prehistory : Studies of Variability in Style in Prehistoric Technologies edited by J. N. Hill and J. D. Gunn, pp. 55 108. Academic Press, New York. Howard, George Delvigne and Jay I. Kislak Reference Collection (Library of Congress) 1947 Prehistoric ceramic s tyles of lowland South America, their distribution and history Yale university publications in anthropology,. Pub. for the Dept. of Anthropology, Yale Univ., by the Yale Univ. Press, New Haven Hugh Jones, C. and S. Hugh Jones 1993 The storage of m anioc products and its symbolic importance among the Tukanoans In Tropical forests: People and food biocultural interactions and applications to development edited by C. M. Hladik, A. Hladik, H. Pagezy, O. Linares, G. F. A. Koppert, and A. Froment, pp. 5 89 594 UNESCO, Paris Iriarte, Jos, J. Christopher Gillam and Oscar Marozzi 2008 Monumental burials and memorial feasting: an example from the southern Brazilian Highlands. Antiquity 82(318):947 961. Isendahl, Christian 2011 The Domesticati on and Early Spread of Manioc ( Manihot Esculenta Crantz ): A Brief Synthesis Latin American A ntiquity 22(4):452 468 Joyce, Thomas Athol and Jay I. Kislak Reference Collection (Library of Congress) 1912 South American archaeology; an introduction to the archaeology of the South American continent with special reference to the early history of Peru Macmillan and co., London,. Junqueira, C. 1975 Os ndios de Ipavu tica, So Paulo. Kaplan, Flora S. 1985 The measuring, mapping, and meani ng of pots. American Anthropologist 87(2):357 364. Kingery, W. David 1981 Plausible Inferences from Ceramic Artifacts. Journal of Field Archaeology 8(4):457 467. 1984 Interactions of Ceramic Technology with Society In Pots and Potters edite d by P. M. Rice, pp. 171 178. UCLA Insitute of Archaeology, Los Angeles.
310 Kipnis, Renato 1998 Early hunter gatherers in the Americas: perspectives from central Brazil. Antiquity 72:581 592. Kramer, Carol 1985 Ceramic Ethnoarchaeology. Annual Review of Anthropology 14:77 102. 1997 Pottery in Rajasthan Smithsonian Institution Press, Washington. Kvamme, Kenneth L., Miriam T. Stark and William A. Longacre 1996 Alternative Procedures for Assessing Standardization in Ceramic Assembla ges. American Antiquity 61(1):116 126. Lathrap, Donald W. 1958 The Cultural Sequence at Yarinacocha, Eastern Peru. American Antiquity 23(4):379 388. 1964 An Alternative Seriation of the Mabaruma Phase, Northwestern British Guiana. American An tiquity 29(3):353 359. 1968 Aboriginal Occupation and Changes in River Channel on the Central Ucayali, Peru. American Antiquity 33(1):62 79. 1970 The upper Amazon Ancient peoples and places,. Praeger Publishers, New York,. 1972 Al ternative Models of population movements in the tropical lowlands of South America. Actas y Memoias, XXXIX Congreso Internacional de Americanistas 4:12 23. 1973 The antiquity and importance of long distance trade relationships in the moist tropics o f Pre Columbian South America. World Archaeology 5(2):170 186. 1976 Shipibo Tourist Art In Ethnic and Tourist Arts: Cultural Expressions from the Fourth World edited by N. H. Graburn, pp. 197 207. University of California Press, Berkeley. 197 7 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. A. Reed, pp. 713 751. 1983 Recent Shipibo Conibo ceramics and their im plications for archeological interpretation In Structure and Cognition in Art edited by D. W. Washburn, pp. 25 39. Cambridge University Press, Cambridge.
311 Lea, Vanessa 1992 Mebengokre (Kayapo) Onomastics: A Facet of Houses as Total Social Facts in Central Brazil. Man 27(1):129 153. Lvi Strauss, Claude 1948 The Tribes of the Upper Xingu River In The Handbook of South American Indians edited by J. H. Steward, pp. 321 348. Smithsonian Institution Bureau of American Ethnolology Bulletin 143 vol. 3. 7 vols. United States Government Printing Office, Washington. Lightfoot, Kent, Antoinette Martinez and Ann M. Schiff 1998 Daily Practice and Material Culture in Pluralistic Social Settings: An Archaeological Study of Culture Change and Pe rsistence from Fort Ross, California. American Antiquity 63:199 222. Lima, H., Eduardo G. Neves and James B. Petersen 2006 A fase Acutuba: Um novo complexo ceramica na Amazonia Central. Arqueologia Suramericana 2(1):26 52. Lima, Pedro 1950 Os Indios Waur : Observacoes Gerais A Ceramica. Boletim do Museu Nacional 9(8 de maio):1 25. Linn, S. 1965 The Ethnologist and the American Indian Potter In Ceramics and Man edited by F. R. Matson, pp. 20 42. vol. Viking Fund Publications in Anth ropology Number Forty One. Wenner Gren Foundation for Anthropological Research Incorporated, New York. Longacre, William A. 1970 Archaeology as anthropology; a case study Anthropological papers of the University of Arizona, no. 17. University of Ar izona Press, Tucson,. 1991 Ceramic ethnoarchaeology University of Arizona Press, Tucson. Longacre, William A., Kenneth L. Kvamme and Masashi Kobayashi 1988 Southwestern Pottery Standardization: An Ethnoarchaeological View from the Philippine s. The Kiva 53(2):101 112. Longacre, William A. and James M. Skibo 1994 Kalinga ethnoarchaeology : expanding archaeological method and theory Smithsonian Institution Press, Washington.
312 Matson, Frederick R. 1984 Ceramics and Man Reconsider ed with Some Thoughts for the Future In The Many Dimensions of Pottery: Ceramics in archaeology and anthropology edited by S. E. Van Der Leeuw and A. C. Pritchard, pp. 27 54. Universiteit van Amsterdam, Amsterdam. Maybury Lewis, David 1974 Akwe S havante Society Oxford University Press, New York. McCall, John C. 1999 Structure, Agency, and the Locus of the Social: Why Poststructural Theory is Good for Archaeology In Material Symbols: Culture and Economy in Prehistory edited by J. E. Rob b, pp. 16 20. Center for Archaeological Investigations Occasional Paper No. 26. Southern Illinois University, Carbondale. McEwan, Colin, Cristiana Barreto and Eduardo Neves (editors) 2001 Unknown Amazon. Culture in nature in ancient Brazil British Museum Press, London. McGuire, Randall H. 2002 A Marxist archaeology Percheron Press, Clinton Corners, N.Y. McKey, Doyle, Stphen Rostain, Jos Iriarte, Bruno Glaser, Jago Jonathan Birk, Irene Holst and Delphine Renard 2010 Pre Columbian agr icultural landscapes, ecosystem engineers, and self organized patchiness in Amazonia PNAS 107(17):7823 7828. Meggers, Betty J. 1945 The Beal Steere Collection of Pottery from Maraj Island, Brazil. Papers of the Michigan Acadamy of Science, Arts, a nd Letters XXXI. 1954 Environmental Limitation on the Development of Culture. American Anthropologist 56:801 823. 1960 Environment and Culture in the Amazon Basin: An Appraisal of the Theory of Environmental Determinism In Studies in Human E cology pp. 71 89. Pan American Union Social Science Monographs 3, Washington, D.C. 1971 Amazonia : Man and Culture in a Counterfeit Paradise Worlds of Man: Studies in Cultural Ecology. Aldine, Chicago. 1977 Vegetational fluctuation and prehi storic cultural adaptation in Amazonia : some tentative correlations. World Archaeology 8(3):287 303 1994a Archeological Evidence for the Impact of Mega Nio Events on Amazonia During the Past Two Millennia. Climatic Change 28:321 338.
313 1994b Biogeographical Approaches to Recontructing the Prehistory of Amazonia Biogeographica 70(3):97 110. 1995a Amazonia on the Eve of European Contact: Ethnohistorical, Ecological, and Anthropological Perspectives. Revista de Arqueologa Americana 8:91 115. 1995b Archaeological perspectives on the potential of Amazonia for intensive exploitation In The fragil tropics of Latin America: Sustainable management of changing environments edited by T. Nishizawa and J. I. Uitto, pp. 68 93. United Nation s University Press, New York. 2000 Natural Versus Anthropogenic Sources of Amazonia n Biodiversity: the Continuing Quest for El Dorado In How Landscapes Change edited by G. A. Bradshaw, pp. 89 107. Springer, New York. 2007 Sustainable Intens ive Exploitation of Amazonia : Cultural, Environmental, and Geopolitical Perspectives In The World System adn teh Earth System edited by A. Hornborg and C. L. Crumley, pp. 105 209. Left Coast Press, Walnut Creek. Meggers, Betty J. and Clifford Evans Jr. 1957 Archaeological Investigations at the Mouth of the Amazon Bureau of American Ethnology Bulletin 167. Smithsonian Institution, Washington, D.C. 1961 An Experimental Formulation of Horizon Styles in the Tropical Forest Area of South America In Essays in Pre Columbian Art and Archaeology edited by S. K. Lothrop, pp. 372 388. Harvard University Press, Cambridge. 1973 A reconstituicao da pre historia amazonica: algumas consideracoes teoricas. io Goeldi) 20:51 69. Meggers, Betty J. and Eurico Th. Miller 2003 Hunter Gatherers in Amazonia during the Pleistocene Holocene Transition In Under the Canopy edited by J. Mercader, pp. 291 316. Rutgers University Press, New Brunswick. Menget, Pat rick 1993 Les Frontieres de la chefferie: Remarques sur le systeme politique du haut Xingu (Bresil). L'Homme 33(2 4):59 76. Mtraux, Alfred 1942 The Native Tribes of Eastern Bolivia and Western Matto Grosso Smithsonian Institution Bureau of A merican Ethnology Bulletin 134. United States Government Printing Office, Washington, D.C.
314 Mills, Barbara J. 1989 Integrating Functional Analyses of Vessels and Sherds through Models of Ceramic Assemblage Formation. World Archaeology 21(1):133 147 Montgomery, Barbara K. 1990 An Instance of Rapid Ceramic Change in the American Southwest. American Antiquity 55(1):88 97. Mowat, Linda 1989 Cassava and Chicha: Bread and Beer of the Amazonia n Indians Shire Publications, Bucks, UK. Neff, Hector (editor) 1992 Chemical Characterization of Ceramic Pastes in Archaeology Prehistory Press, Madison. Netto, Annibal Brasil 1964 A Bacia dos Formadores do Xingu. Museu Paraense Emilio Goeldi Publicacoes Avulsa 1:1 8. Neupert, Mark A. 2000 Clays of Contention: An Ethnoarchaeological Study of Factionalism and Clay Composition. Journal of Archaeological Method and Theory 7(3):249 272. Neves, Eduardo Gos 1998 Twenty Years of Amazonia n Archaeology in Brazil (1977 1997). Antiquit y 72(277):625 632. 1999a Changing perspectives in Amazonia n Archaeology In Archaeology in Latin America edited by G. G. Politis and B. Alberti, pp. 216 243. Routledge, New York. 1999b Complexity or Not in the Amazon? Proceedings of the Four th World Archaeological Congress Cape Town. 2001 Indigenous Historical Trajectories in the Upper Rio Negro Basin In Unknown Amazon: Culture in Nature in Ancient Brazil edited by C. McEwan, C. Barreto and E. G. Neves. The British Museum Press, Lon don. 2006 Political Economy and Pre Columbian Landscape Transformations in Central Amazonia In Time and Complexity in Historical Ecology: Studies in the Neotropical Lowlands edited by W. Bale and C. L. Erickson, pp. 279 309. Columbia University P ress, New York. 2008 Ecology, ceramic chronology and distribution, long term history and political change in the Amazonia n floodplain In Handbook of South American Archaeology edited by H. Silverman and W. H. Isbell, pp. 359 379. Springer, New Yor k.
315 Neves, Eduardo Gos, James B. Petersen, R. N. Bartone and M. J. Heckenberger 2004 The Timing of Terra Preta Formation in the Central Amazon: Archaeological Data from Three Sites In Explorations in Amazonia n Dark Earths edited by W. I. Woods and B. Glaser, pp. 125 134. Springer, Verlag Berlin Heidelberg. Neves, Eduardo Gos, James B. Petersen, R. N. Bartone and C. A. Silva 2003 Historical and socio cultural origins of Amazonia n dark earths In Amazonia n Dark Earths: Origins, Properties, and Management edited by J. Lehmann, D. Kern, B. Glaser and W. Woods, pp. 29 50. Kluwer Academic Publishers, Dordrecht. Nimuendaj, Curt 1948a Little Known Tribes of the Lower Tocantins River Region In Handbook of South American Indians Volume 3: The Tropical Forest Tribes edited by J. H. Steward, pp. 203 211. Bureau of American Ethnology, Smithsonian Institution, Washington, D.C. 1948b Tribes of the Middle and Lower Xingu In Handbook of South American Indians Volume 3: The Tropical Forest Tr ibes edited by J. H. Steward, pp. 213 244. Bureau of American Ethnology, Smithsonian Institution, Washington, D.C. Nimuendaj, Curt and Per Stenborg 2004 In pursuit of a past Amazon : archaeological researches in the Brazilian Guyana and in the Ama zon region Etnologiska studier,. Vrldskulturmuseet, Gteborg. Noelli, Francisco Silva 1998 The Tupi: explaining origin and expansions in terms of archaeology and of historical linguistics. Antiquity 72:648 663. Nye, Margaret M. 1991 The Mis measure of Manioc ( Manihot esculenta Euphorbiaceae). Economic Botany 45(1):47 57. Oberg, Kalervo 1948 The Bacairi of Northern Matto Grosso. Southwestern Journal of Anthropology 4(3):305 319. 1953 Indian Tribes of Matto Grosso, Brazil Insti tute of Social Anthropology Publication No. 15, Smithsonian Institution, Washington, D.C. 1955 Types of Social Structure among the Lowland Tribes of South and Central Amrica. American Anthropologist 57:472 487.
316 Oliver, J. R. 1989 The Arch aeological, linguistic and ethnohistorical evidence for the expansion of Arawakan into Northwestern Venezuela and Northeastern Colombia Ph.D. Dissertation, University of Illinois at Urbana Champaign, Illinois. Oyuela Caycedo, Augusto 1995 Rocks ver sus Clay: The Evolution of Pottery Technology in the Case of San Jacinto 1, Colombia In The Emergence of Pottery: Technology and Innovation in Ancient Societies edited by W. K. Barnett and J. W. Hoopes, pp. 133 144. Smithsonian Institution Press, Washing ton. Patterson, Thomas Carl 2003 Marx's Ghost: Conversations with Archaeologists Berg, New York. Pauketat, Timothy R. 2001 Practice and History in Archaeology: An Emerging Paradigm. Anthropological Theory 1:73 98. Pauketat, Timothy R. and Thomas E. Emerson 1991 The Ideology of Authority and the Power of the Pot. American Anthropologist 93(4):919 941. Pearsall, Deborah M. 1992 The Origins of Plant Cultivation in South America In The Origins of Agriculture edited by C. W. Cowan and P. J. Watson, pp. 173 205. Smithsonian Institution Press, Washington and London. Petersen, James B. 1997 Taino, Island Carib, and Prehistoric Amerindian Economies in the West Indies: Tropical Forest Adaptations to Island Environments In The In digenous People of the Caribbean edited by S. M. Wilson. University Press of Florida, Gainesville. Petersen, James B., Michael J. Heckenberger and Eduardo G. Neves 2001 a A Prehistoric Ceramic Sequence From The Central Amazon And Its Relationship To The Caribbean Proceedings of the XIX International Congress for Caribbean Archaeology Aruba. Petersen, James B., Michael J. Heckenberger and Jack A. Wolford 2001 b Spin, Twist, and Twine: An Ethnoarchaeological Examination of Group Identity in Nat ive Fiber Industries from Greater Amazonia In Fleeting Identities: Perishable Material Culture in Archaeological Research edited by P. B. Drooker, pp. 226 253. Center for Archaeological Investigations, Southern Illinois University, Carbondale.
317 Peterse n, James B., Corinne L. Hofman and L. Antonio Curet 2004 Time and Culture: Chronology and Taxonomy in the Eastern Caribbean and the Guianas In Late Ceramic Age Societies in the Eastern Caribbean edited by A. Delpuech and C. L. Hofman. BAR Internati onal Series. Archaeopress, Oxford. Petersen, James B., Eduardo G. Neves and Michael J. Heckenberger 2001 c Gift from the Past: Terra Preta and Prehistoric Amerindian Occupation in Amazonia In Unknown Amazon: Culture in Nature in Ancient Brazil edit ed by C. McEwan, C. Barreto and E. G. Neves, pp. 86 105. British Museum Press, London. Pierce, Christopher 2005 Reverse Engineering the Ceramic Cooking Pot: Cost and Performance Properties of Plain and Textured Vessels. Journal of Archaeological Met hod and Theory 12(2):117 157. Piperno, Dolores R., Anthony J. Ranere, Irene Holst and Patricia Hansell 2000 Starch Grains Reveal Early Root Crop Horticulture in the Panamanian Tropical Forest. Nature 407:89 897. Pritchard, Alison C. and Sander E. V an Der Leeuw 1984 Introduction: The Many Dimensions of Pottery In The Many Dimensions of Pottery: Ceramics in archaeology and anthropology edited by S. E. Van Der Leeuw and A. C. Pritchard, pp. 3 23. Universiteit van Amsterdam, Amsterdam. Prous, Andr 1991 Arqueologia Brasileira Universidade de Braslia, Braslia. Rice, Prudence M. 1984a The Archaeological Study of Specialized Pottery Production: Some Aspects of Method and Theory In Pots and Potters edited by P. M. Rice, pp. 45 54 UCLA Institute of Archaeology, Los Angeles. 1984b Change and Conservatism in Pottery Producing Systems In The Many Dimensions of Pottery: Ceramics in archaeology and anthropology edited by S. E. Van Der Leeuw and A. C. Pritchard, pp. 233 293. U niversiteit van Amsterdam, Amsterdam. 1984c Overview and Prospect In Pots and Potters edited by P. M. Rice, pp. 245 255. UCLA Institute of Archaeology, Los Angeles. 1984d Pots and Potters UCLA Institute of Archaeology, Los Angeles. 19 87 Pottery analysis: a sourcebook University of Chicago Press, Chicago.
318 1991 Specialization, Standardization, and Diversity: A Retrospective In The Ceramic Legacy of Anna O. Shepard edited by R. L. Bishop and F. W. Lange, pp. 257 279. Universit y Press of Colorado, Niwot. 1996a Recent Ceramic Analysis: 1. Function, Style, and Origins. Journal of Archaeological Research 4(2):133 163. 1996b Recent Ceramic Analysis: 2. Composition, Production, and Theory. Journal of Archaeological Re search 4(3):165 201. 1999 On the Origins of Pottery. Journal of Archaeological Method and Theory 6(1):1 54. Riley, John A. 1984 Pottery Analysis and the Reconstruction of Ancient Exchange Systems In The Many Dimensions of Pottery: Ceramics in archaeology and anthropology edited by S. E. Van Der Leeuw and A. C. Pritchard, pp. 57 78. Universiteit van Amsterdam, Amsterdam. Roe, P.G. 1987 Village Spatial Organization in the South Amerindian Lowlands: Evidence from Ethnoarchaeology Proce edings of the 52nd Annual Meeting of the Society for American Archaeology Toronto, Canada. Roosevelt, Anna Curtenius 1980 Parmana : prehistoric maize and manioc subsistence along the Amazon and Orinoco Studies in archaeology. Academic Press, New Y ork. 1987 Chiefdoms of the Amazon and Orinoco In Chiefdoms in the Americas edited by R. D. Drennan and C. Uribe, pp. 153 185. University Presses of America, Lanham. 1989 Resource Management in Amazonia before the Conquest: Beyond Ethnograph ic Projection. Advances in Economic Botany 7:30 62. 1991 Moundbuilders of the Amazon : geophysical archaeology on Maraj Island, Brazil Academic Press, San Diego. 1994 Amazonia n Indians from prehistory to the present : anthropological perspe ctives University of Arizona Press, Tucson. 1995 Early Pottery in the Amazon In The Emergence of Pottery: Technology and Innovation in Ancient Societies edited by W. K. Barnett and J. W. Hoopes, pp. 115 131. Smithsonian Institution Press, Washing tone.
319 1997 The Excavations at Corozal, Venezuela: Stratigraphy and Ceramic Seriation Yale University Publications in Anthropology Number Eighty Three. Department of Anthropology and the Peabody Museum Yale University, New Haven. 1999a The De velopment of Prehistoric Complex Societies: Amazonia A Tropical Forest In Complex Polities in the Ancient Tropical World edited by E. A. Bacus and L. J. Lucero. Archaeological Papers of the American Anthropological Association 9, Washington, D.C. 1 999b The Maritime, Highland, Forest Dynamic and the Origins of Complex Culture In The Cambridge History of the Native Peoples of the Americas edited by F. Salomon and S. Schwartz, pp. 265 349. vol. 3 South America. Cambridge University Press, New York. 2000 The Lower Amazon: A Dynamic Human Habitat In Imperfect Balance: Landscape Transformations in the Pre Columbian Americas edited by D. L. Lentz, pp. 455 491. Columbia University Press, New York. Roosevelt, Anna Curtenius, John Douglas and Lin da Brown 2002 The Migrations and Adaptations of the First Americans Clovis and Pre Clovis Viewed From South America In The First Americans edited by N. G. Jablonski. Wattis Symposium Series in Anthropology Memoirs of the California Academy of Scien ces, San Francisco. Roosevelt, Anna Curtenius, R. A. Housley, M. Imazio, S. Maranca and R. Johnson 1991 Eighth Millennium Pottery from a Prehistoric Shell Midden in the Brazilian Amazon. Science 254:1621 1624. Roosevelt, Anna Curtenius, M. Lima da Costa, C. Lopes Machado, M. Michab, N. Mercier, H. Valladas, J. Feathers, W. Barnett, M. Imazio da Silveira, A. Henderson, J. Sliva, B. Chernoff, D. S. Reese, J.A. Holman, N. Toth and K. Schick 1996 Paleoindian Cave Dwellers in the Amazon: The Peopli ng of the Americas. Science 272(19 April):373 384. Roscoe, Paul B. 1993 Practice and Political Centralization. Current Anthropology 34:111 140. Rouse, Irving 1965 Caribbean Ceramics: A Study in Method and in Theory In Ceramics and Man edit ed by F. R. Matson, pp. 88 103. vol. Viking Fund Publications in Anthropology Number Forty One. Wenner Gren Foundation for Anthropological Research Incorporated, New York. Rouse, Irving and Jos Mara Cruxent 1963 Venezuelan archaeology Caribbean s eries 6. Yale University Press, New Haven.
320 Roux, Valentine 2003 Ceramic Standardization and Intensity of Production: Quantifying Degrees of Specialization. American Antiquity 68(4):768 782. Rye, Owen S. 1976 Keeping Your Temper Under Control. Archaeology and Physical Anthropology in Oceania 11(2):106 137. 1981 Pottery Technology Manuals On Archaeology 4. Taraxacum, Washington. S, Cristina 1983 Observaes Sobre A Habitao em Trs Grupos Indgenas Brasileiros In Habitaes Ind genas edited by S. C. Novaes, pp. 103 145. Livraria Nobel S. A. Editora da Universidade de So Paulo, So Paulo. Sanoja Obediente, Mario 1994a Central America, the Caribbean, Northern South America and the Amazon: the beginnings of food production In History of Humanity edited by S. J. De Laet, pp. 630 633. Routledge, New York. 1994b Central America, the Caribbean, Northern South America and the Amazon: the way of life of ancient hunters In The History of Humanity edited by S. J. De Laet, pp. 317 327. Routledge, New York. Sanoja Obediente Mario and Iraida Vargas Arenas 1978 The Formative Cultures of the Venezuelan Oriente In Advances in Andean Archaeology edited by D. Browman, pp. 259 276. Mouton Publishers, The Hague. Santos Gr anero, Fernando 2001 The Arawakan Matrix: Ethos, Language, and History in Native South America In Comparative Arawakan Histories: Rethinking Language Family and Culture Area in Amazonia edited by J. D. Hill and F. Santos Granero. University of Illi nois Press, Urbana and Chicago. Sassaman, Kenneth E. 1993 Early pottery in the Southeast : tradition and innovation in cooking technology University of Alabama Press, Tuscaloosa. Saunders, Rebecca 2000 Stability and Change in Guale Indian Po ttery, A.D. 1300 1702 The University of Alabama Press, Tuscaloosa. Schaan, Denise Pahl 1997 Evidncia arqueolgica e organizao social na Fase Maraj ara. Estudos Ibero Americanos XXIII(1):97 114.
321 2000 Recent investigations on Maraj ara Cult ure, Maraj Island, Brazil. Antiquity 74:469 470. 2001 Estatuetas Maraj ara: o simbolismo de identidades de gnero em uma sociedade complexa Amaznica. Boletim do Museu Paraense Emlio Goeldi. Srie Antropologia 17(2):437 477. 2010 Long Term Human Induced Impacts on Maraj Island Landscapes, Amazon Estuary. Diversity 2:182 206. Schaan, Denise Pahl, Dirse Clara Kern and F. J. L. Frazao 2009 An Assessment of the Cultural Practices Behind the Formation (or Not) of Amazonia n Dark Earths in Maraj Island Archaeological Sites In Amazonia n Dark Earths: Wim Sombroek's Vision edited by W. I. Woods, W. G. Teixeira, J. Lehmann, C. Steiner, A. M. G. A. WinklerPrins and L. Rebellato, pp. 127 141. Springer Science and Business Media, New York. Scha den, Egon 1964 Aspectos e Problemas Etnologicos de uma Area de Aculturacao Intertribal: O Alto Xingu In Aculturacao Indigena Universidade de So Paulo, So Paulo. Schiffer, Michael B. and James M. Skibo 1987 Theory and Experiment in the Stud y of Technological Change. Current Anthropology 28(5):595 622. Schiffer, Michael B., James M. Skibo, Tamara C. Boelke, Mark A. Neupert and Meredith Aronson 1994 New Perspectives on Experimental Archaeology: Surface Treatments and Thermal Response o f the Clay Cooking Pot. American Antiquity 59(2):197 217. Schmidt, Max 1902 Ausden Ergebnissen Meiner Expedition in das Schingu Quellgebiet. Globus 82, 86(2):29 31, 119 125. 1917 Die Aruaken: Ein Beitrag zum Problem de Kulturverbrietung Leip zig. Schmidt Morgan J 2010 Reconstructing Tropical Nature: Prehistoric and Modern Anthrosols (Terra Preta) in the Amazon Rainforest, Upper Xingu River, Brazil Unpublished Ph.D. Dissertation, University of Florida UNI, Ann Arbor Schmidt, Morgan J. and Michael Heckenberger 2009 Amerindian Anthrosols: Amazonia n Dark Earth Formation in the Upper Xingu In Amazonia n Dark Earths: Wim Sombroek's Vision edited by W. I. Woods, W. G. Teixeira, J. Lehmann, C. Steiner, A. M. G. A. WinklerPrins and L Rebellato, pp. 163 191. Springer Science and Business Media, New York.
322 Schmitz, Pedro Ignacio 1987 Prehistoric Hunters and Gatherers of Brazil. Journal of World Prehistory 1(1):53 126. Schultz, Harald 1961 Informaes etnogrficas sobre os ndios Suy. Revista do Museu Paulista 13:315 332. Scott, Andy 1993 A parametric approach to seriation In Computing the Past: Computer Applications and Quantitative Methods in Archaeology edited by J. Andresen, T. Madsen and I. Scollar, pp. 317 324 vol. CAA92. Aarhus University Press, Aarhus. Seeger, A. 1976 Fixed Points on Arcs in Circles: The Temporal, Processual Aspect of Suy Space and Society. Proceedings of the International Congress of Americanists 2:341 359. 1981 Nature and So ciety in Central Brazil: The Suy Indians of Mato Grosso, Brazil Harvard University Press, Cambridge. Sewell, William H., Jr. 1992 A Theory of Structure: Duality, Agency, and Transformation. American Journal of Sociology 98:1 29. Shepard, Anna O. 1956 Ceramics for the Archaeologist Publication 609. Carnegie Institution of Washington, Washington, D.C. Siegel, Peter E. and Peter G. Roe 1986 Shipibo Archaeo Ethnography: Site Formation Processes and Archaeological Interpretation. World Ar chaeology 18(1):96 115. 1963 Os "Txikao" e Outras Tribos Marginais do Alto Xingu. Revista do Museu Paulista 14:76 104. 1967 Consideraes preliminares sobre a arqueologia do Alto Xingu. In Programa Nacional de Pesquisas Arq ueolgicas: resultados preliminares do primeiro ano: 1965 1966 pp. 129 144. vol. 6. Museu Paraense Emlio Goeldi (Publicaes avulsas), Belm. and Fernanda Araujo Costa ) 30.
323 Skibo, James M. and Gary M. Feinman (editors) 1999 Pottery and People: A Dynamic Interaction The University of Utah Press, Salt Lake City. Skibo, James M., Michael B. Schiffer and Kenneth C. Reid 1989 Organic Tempered Pottery: An Experimental Study. American Antiquity 54(1):122 146. Stahl, Peter W. 2002 Paradigms in Paradise: Revising Standard Amazonia n Prehistory. The Review of Archaeology 23( 2):39 51. Stahl, Peter W. and James A. Zeidler 1990 Differential Bone Refuse Accumulation in Food Preparation and Traffic Area on an Early Ecuadorian House Floor. Latin American Antiquity 1(2):150 169. Stanislawski, Michael B. 1973 Ethnoarcha eology and Settlement Archaeology. Ethnohistory 20(4):375 392. Stark, Miriam T. 1991 Ceramic production and community specialization: a Kalinga ethnoarchaeological study. World Archaeology 23(1):64 78. 2003 Current Issues in Ceramic Ethnoarch aeology. Journal of Archaeological Research 11(3):193 242. Stark, Miriam T., Ronald L. Bishop and Elizabeth Miksa 2000 Ceramic Technology and Social Boundaries: Cultural Practices in Kalinga Clay Selection and Use. Journal of Archaeological Method and Theory 7(4):295 331. Steinen, Karl von den 1886 Durch Zentral Brasilien F.A. Brockhaus, Leipzig. 1894 Unter den Naturvolkern Zentral Braziliens Dietrich Reimer, Berlin. 1966 Among the Primitive Peoples of Central Brazil: A Travel Account and the Results of the Second Xingu Expedition 1887 1888 Translated by F. Schutze. Human Relations Area Files. Steward, Julian Haynes (editor) 1946 Handbook of South American Indians 6 vols. Washington, D.C., Bureau of American Ethnology, Smithsonian Institution. Steward, Julian Haynes and Louis C. Faron 1959 Native peoples of South America McGraw Hill, New York,.
324 Sullivan, Alan P., III 1989 The technology of ceramic reuse: formation processes and archaeological evidence. Wo rld Archaeology 21(1):101 114. 2008 Ethnoarchaeological and archaeological perspectives on ceramic vessels and annual accumulation rates of sherds. American Antiquity 73(1):121 135. Tite, M.S. 1999 Pottery Production, Distribution, and Consum ption The Contribution of the Physical Sciences. Journal of Archaeological Method and Theory 6(3):181 233. Tucker, Robert C. 1978 The Marx Engels Reader W.W. Norton and Company, New York Turner, Terrence S. 1979 Kinship, Household, and Comm unity Structure among the Kayap In Dialectical Societies: The G and Bororo of Central Brazil edited by D. Maybury Lewis, pp. 179 217. Harvard University Press, Cambridge. Upham, Steadman, Kent G. Lightfoot and Gary M. Feinman 1981 Explaining Soc ially Determined Ceramic Distributions in the Prehistoric Plateau Southwest. American Antiquity 46(4):822 833. Van Der Leeuw, Sander E. and Alison C. Pritchard (editors) 1984 The Many Dimensions of Pottery: Ceramics in archaeology and anthropology Universiteit van Amsterdam, Amsterdam. Velde, Bruce and Isabelle C. Druc 1999 Archaeological Ceramic Materials: Origin and Utilization Natural Science in Archaeology. Springer, New York. Verswijver, G. 1982 Intertribal relations between the Juruna and the Kayapo indians (1850 1920) In Jahrbuch des Museums fur Volkerkun Leipzig. Villas Boas, Orlando and Cladio Villas Boas 1973 Xingu: The Indians, Their Myths Far rar Straus and Giroux, New York Wagley, Charles 1940 The Effect s of Depopulations upon Social Organization as Illustrated by the Tapirape Indians. Transactions of the New York Academy of Science 3(2):12 16.
325 Waterson, Roxana 1995 Houses and hierarchies in island Southeast Asia In About the House: Lvi Strau ss and Beyond edited by J. Carsten and S. Hugh Jones, pp. 47 68. Cambridge University Press, Cambridge. Weber, Thomas 1993 Mathematical models for the reconstruction of prehistoric settlement patterns: Central German examples In Computing the Past : Computer Applications and Quantitative Methods in Archaeology edited by J. Andresen, T. Madsen and I. Scollar, pp. 377 388. vol. CAA92. Aarhus University Press, Aarhus. Whitehead, Neil L. 1993 Ethnic Transformation and Historical Discontinuity in Native Amazonia and Guayana, 1500 1900. L'Homme 33(2 4):285 305. 1994 The Ancient Amerindian Polities of the Amazon, the Orinoco, and the Atlantic Coast: Preliminary Analysis of Their Passage from Antiquity to Extinction In Amazonia n Indians from Prehistory to the Present edited by A. C. Roosevelt, pp. 33 53. University of Arizona Press, Tuscon. 1996 Amazonia n Archaeology; Searching for Paradise? A Review of Recent Literature and Fieldwork. Journal of Archaeological Research 4(3):241 264. Whitmore, T.C. and G.T. Prance 1987 The Early History of Man in Amazonia In Biogeography and Quaternary History in Tropical America edited by T. C. Whitmore and G. T. Prance, pp. 151 174. Clarendon Press, Oxford. Wilson, David John 1999 Indi genous South Americans of the Past and Present: An Ecological Perspective Westview Press, Boulder. Wilson, Warren M. and D. L. Dufour 2002 Why "Bitter" Cassava? Productivity of "Bitter" and "Sweet" Cassava in a Tukanoan Indian Settlement in the Nor thwest Amazon. Economic Botany 56(1):49 57. Wst, Irmhild 1994 The Eastern Bororo from an Archaeological Perspective In Amazonia n Indians from Prehistory to the Present: Anthropological Perspectives edited by A. C. Roosevelt, pp. 315 342. Universi ty of Arizona Press, Tuscon. 1998 Continuities and discontinuities: archaeology and ethnoarchaeology in the heart of the Eastern Bororo territory, Mato Grosso, Brazil. Antiquity 72:663 675.
326 Wst, Irmhild and Cristiana Barreto 1999 The Ring V illages of Central Brazil: A Challange for Amazonia n Archaeology. Latin American Antiquity 10(1):1 23. Zeidler, James A. 1984 Social Space In Valdivia Society: Community Patterning and Domestic Structure at Real Alto, 3000 2000 B.C. Ph.D. Dissertati on, University of Illinois at Urbana Champaign, Urbana. 1998 Cosmology and Community Plan in Early Formative Ecuador: Some Lessons From Tropical Ethnoastronomy. Journal of the Steward Anthropological Society 26(1/2):37 68. Zucchi, Alberta 2001 A New Model of the Northern Arawakan Expansion In Comparative Arawakan Histories: Rethinking Language Family and Culture Area In Amazonia edited by J. D. Hill and F. Santos Granero, pp. 199 222. University of Illinois Press, Urbana and Chicago.
327 BIOG RAPHICAL SKETCH Joshua Robert Toney was born in Brattleboro, Vermont in 1974 and graduated from Brattleboro Union High School in 1992 After one year at the University of Rhode Island he moved back to Vermont to study history and a nthropology at the University of Vermont. He began his archaeological career in 1995 when he enrolled in a summer field school taught by Dr. James B. Petersen He received a BA in history and p hilosophy in 1997 and continued travel ling the lower 48 states in search of archaeological fieldwork before moving back to Vermont in 1999 where he worked as At the encouragement of Dr. Petersen he began graduate studies at the University of Florida under Dr. Michael J. Heckenberger, travelling to the Upper Xingu in 2002 to begin his fieldwork. He received his MA from the University of Florida in 2005, moved to Hawaii in 2009 to finish writing his dissertation and received his PhD from the University of Florida in 2012.