intrasite spatial organization in chavín de huántar during the andean formative
DESCRIPTION
The ceremonial center of Chavín de Huántar is located in the north central Peruvian highlands at 3,200 m.a.s.l. I have investigated the Wacheqsa sector, located immediately to the north of the monumental core with the objective of understanding the archaeological deposits that formed this area, and to analyze their chronological and spatial relationships.This research presents a new methodology for the investigation of intrasite space organization of stratigraphic components. I demonstrate that careful sampling programs can be extremely advantageous in investigating intrasite variation, in particular when all stratigraphic records are modeled using computer aided design (CAD). I identify five prehistoric spatial analytical units in the Wacheqsa Sector: Early Platforms, Water Flood, Late Platforms, Stone Rooms and Midden. I use bivariate kernel density estimations in order to investigate ceramic modalities and comprehend the nature of the activities developed in each unit, cross- referencing this line of evidence with the distribution of archaeological materials. I also use the Boone index as a measure of diversity in order to quantitatively segregate the analytical units identified.The Wacheqsa Sector was occupied from 1200 BC to 500 BC, during the Middle and Late Formative Periods. I have divided the prehistoric occupation into two phases. The oldest one (Urabarriu 1200-800 BC) encompasses the Early Platforms and Water Flood analytical units. The Early Platforms unit represents the oldest domestic settlement located in this sector. The Water Flood analytical unit provides evidence regarding canalization of the Wacheqsa River during this phase. The latest occupation phase (Janabarriu 800-500 BC), encompasses the Late Platforms, Stone Rooms and Midden analytical units. The Stone Room analytical unit represents a late settlement in the Wacheqsa Sector, the Midden provides evidence for suprahousehold food and beverage consumption and the Late Platforms unit seems to be a buffer area between these two units.In addition, the dating of these units shows that Chavín was contemporary with ceremonial centers of the Andean area during the Middle and Late Formative. It also demonstrates that the Janabarriu ceramic phase is 400 years earlier than previously suggested, being contemporary with the largest architectural phase at monumental core.TRANSCRIPT
INTRASITE SPATIAL ORGANIZATION AT CHAVÍN DE HUANTAR DURING
THE ANDEAN FORMATIVE: THREE DIMENSIONAL MODELING,
STRATIGRAPHY AND CERAMICS
A DISSERTATION
SUBMITTED TO THE DEPARTMENT OF ANTHROPOLOGICAL SCIENCES
AND THE COMITEE ON GRADUATE STUDIES
OF STANFORD UNIVERSITY
IN PARTIAL FULLFILLMENT OF THE REQUIREMENTS
FOR THE DEGREE OF DOCTOR OF PHILOSOPHY
Christian Mesía
September 2007
ii
© Copyright by Christian Mesía 2007
All Rights Reserved
iii
I certify that I have read this dissertation and that, in my opinion, it is fully adequate in
scope and quality as a dissertation for the degree of Doctor of Philosophy.
Signed in original
(John Rick) Principal Adviser
I certify that I have read this dissertation and that, in my opinion, it is fully adequate in
scope and quality as a dissertation for the degree of Doctor of Philosophy.
Signed in original
(Ian Robertson)
I certify that I have read this dissertation and that, in my opinion, it is fully adequate in
scope and quality as a dissertation for the degree of Doctor of Philosophy.
Signed in original
(William Durham)
I certify that I have read this dissertation and that, in my opinion, it is fully adequate in
scope and quality as a dissertation for the degree of Doctor of Philosophy.
Signed in original
(Luis G. Lumbreras)
Approved for the University Committee on Graduate Studies.
iv
ABSTRACT
The ceremonial center of Chavín de Huántar is located in the north central Peruvian
highlands at 3,200 m.a.s.l. I have investigated the Wacheqsa sector, located immediately to the
north of the monumental core with the objective of understanding the archaeological deposits
that formed this area, and to analyze their chronological and spatial relationships.
This research presents a new methodology for the investigation of intrasite space
organization of stratigraphic components. I demonstrate that careful sampling programs can be
extremely advantageous in investigating intrasite variation, in particular when all stratigraphic
records are modeled using computer aided design (CAD). I identify five prehistoric spatial
analytical units in the Wacheqsa Sector: Early Platforms, Water Flood, Late Platforms, Stone
Rooms and Midden. I use bivariate kernel density estimations in order to investigate ceramic
modalities and comprehend the nature of the activities developed in each unit, cross-
referencing this line of evidence with the distribution of archaeological materials. I also use
the Boone index as a measure of diversity in order to quantitatively segregate the analytical
units identified.
The Wacheqsa Sector was occupied from 1200 BC to 500 BC, during the Middle and
Late Formative Periods. I have divided the prehistoric occupation into two phases. The oldest
one (Urabarriu 1200-800 BC) encompasses the Early Platforms and Water Flood analytical
units. The Early Platforms unit represents the oldest domestic settlement located in this sector.
The Water Flood analytical unit provides evidence regarding canalization of the Wacheqsa
River during this phase. The latest occupation phase (Janabarriu 800-500 BC), encompasses
the Late Platforms, Stone Rooms and Midden analytical units. The Stone Room analytical unit
represents a late settlement in the Wacheqsa Sector, the Midden provides evidence for
suprahousehold food and beverage consumption and the Late Platforms unit seems to be a
buffer area between these two units.
In addition, the dating of these units shows that Chavín was contemporary with
ceremonial centers of the Andean area during the Middle and Late Formative. It also
demonstrates that the Janabarriu ceramic phase is 400 years earlier than previously suggested,
being contemporary with the largest architectural phase at monumental core.
v
ACKNOWLEDGMENTS
My doctoral research at Chavín de Huántar was supported by a Stanford University
School of Humanities and Social Sciences’ Graduate Research Opportunity Grant, the
Department of Anthropological Sciences, the Center for Latin American Studies and the
Stanford Archaeology Center. This dissertation could not have been completed without the
help of several people. I thank the members of my dissertation reading committee, especially
my main advisor, John Rick, for his valuable insight and for sharing his data and views about
Chavín de Huántar with me. His support and advice during my years at Stanford have been
extremely helpful and I value our extensive discussions regarding Chavín, statistics, digital
models and Andean prehistory. Ian Robertson was instrumental in developing the statistical
approach used in this dissertation, and I thank him for our long conversations regarding the
value of quantitative methods in archaeology. William Durham offered valuable comments
regarding the anthropological nature of the stratigraphic record; his support has been unfailing
since I arrived at Stanford. Luis Lumbreras spent several hours discussing with me the
significance of Chavín in Andean prehistory, the site’s ceramic chronology, and the caution
that must be exercised when interpreting absolute radiocarbon dates. His guidance and
insights were tremendously helpful. At Chavín the Huántar, the town’s residents welcomed
me with warm hospitality. The Arana and Rosemberg families kept my team and I very well
rested and fed. I thank my field crew, Maritza Pérez, Ivan Falconí, Victor Hugo Rojas, Lizbeth
Tepo, Adriana Aguayo, Diana Sandoval, Laura Driscoll, Diana Galindo, Lise Mesz, Zozimo
Meljarejo, Fortunato Garay, Roger Garay, Andres Caurino, Benigno Dionisio, Gaspar Cruz
Virgilio Dionisio, Nando Dionisio, Florencio Melgarejo, and Andrés Díaz for their support
and professionalism. Rosa Mendoza and Maria Mendoza were crucial in organizing field and
laboratory logistics and their friendship is invaluable.
In Lima, the Museo Nacional de Arqueología, Antropología e Historia del Perú
(MNAAH) provided me with space for analyzing the Chavin ceramic assemblage. The
MNAAH also granted me access to its archives. Special thanks go to its former Director,
Carlos Del Águila, for granting me permission to conduct research at the Museum as well as
to Fedora Martinez and its former director Enrique Gonzáles Carré for their generous help
with this phase of the project. The Museo de Arqueología y Antropología at the Universidad
Nacional Mayor de San Marcos allowed me to review the Chavín collection excavated by
Rosa Fung. Its former Directors Ruth Shady and Javier Alcalde granted the permits necessary
vi
to conduct such research and established a space within which to work. The Tello Archive at
the Universidad Nacional Mayor de San Marcos approved access to review Tello’s fieldnotes.
I thank Rafael Vega Centeno and Victor Paredes for making all the arrangements necessary
for my research in the archives.
Rosa Fung kindly shared her insights regarding the Wacheqsa sector, providing
helpful information of her unpublished excavations. Likewise, Silvia Kembel shared her
insights regarding the architectural sequence at Chavín. Their willingness to share their data
without reservations has been truly inspiring.
Reimman Ramirez illustrated part of the ceramic collection at the lab in Chavín and
Ivan Falconí, Adriana Aguayo, Karla Alarcón, Natalí Ramírez, and Gabriela Ferrando drew
ceramics at the Museo Nacional. Karla Alarcón and Patricia Quintana drew those ceramics
that were reviewed at the San Marcos Museum. Ceramic drawings were redrawn in CAD by
Iván Falconí. Several excavation profiles and plans were drawn by MaFe Córdova. César
Trigoso, Jose Luis Cruzado and María Mendoza were instrumental when laptops and desktops
refused to work 10 days before filing this dissertation; to all of them I extend my greatest
appreciation.
Matthew Velasco and Megan Kane read drafts of this dissertation and provided
valuable assistance with English grammar and vocabulary. They shared with me the
frustration of completing our respective projects in the midst of power outages at Chavín; to
them I offer my endless thanks.
I am grateful to my fellow graduate students, Daniel Contreras, John Wolf, Nikki
Slovak, Ignacio Cancino, Fernando Amstrong and Silvana Rosenfield for their friendship and
support over the years; their camaraderie and sense of humor have been unparalleled; graduate
school would not have been the same without them. I also thank Matt Sayre for his friendship
during these years of graduate studies and fieldwork.
Jennifer Kidwell, Mary Cahill and Nancy Lonhart provided endless amounts of advice
throughout graduate school; their warmth, kindness and advice are greatly appreciated. I am
also grateful to The Global Heritage Fund (GHF), who provided me with registered licenses of
Autodesk Land and Autodesk Map software. I also thank Herbert Hass and Greg Hodgins for
their assistance with the radiocarbon dates.
vii
Finally, my mother has been a never-ending source of support and encouragement;
she taught me that the only way to achieve a goal in life is through hard work. To her, my
deepest thanks in every imaginable way.
viii
1 INTRODUCTION, RESEARCH HISTORY AND RESEARCH DESIGN 1
1.1.1 Introduction 1
1.2 Research History at Chavín de Huántar 2
1.2.1 Early Accounts 2
1.2.2 Travelers 4
1.2.3 The Archaeologists arrive 6
1.2.4 After the Aluvión more archaeologists arrive 9
1.2.5 A New Turning Point 12
1.2.6 The Stanford Archaeological Project 13
1.2.7 Where do we stand now? 14
1.3 Research Design 14
1.3.1 Research Process 14
1.3.2 Research Questions 16
2 THE ANDEAN FORMATIVE 18
2.1 The Concept of Formative in Andean Prehistory 17
2.1.1 Early Formative (1800-1200 BC) 20
2.1.2 Middle Formative (1200-800 BC) 23
2.1.3 Late Formative (800-500 BC) 28
2.1.4 Final Formative (500-50 BC) 32
3 THE WACHEQSA SECTOR AT CHAVÍN DE HUÁNTAR 34
3.1 Wendell Bennett’s excavations 36
3.2 Julio C. Tello’s excavations 36
3.3 Rosa Fung’s excavations 38
4 THEORY AND METHODS 41
4.1 Theory of the Archaeological Record 41
4.2 Data Processing 44
TABLE OF CONTENTS
ABSTRACT iv
ACKNOWLEDGMENTS v
TABLE OF CONTENTS viii
LIST OF TABLES xiii
LIST OF ILLUSTRATIONS xviii
ix
4.2.1 Stratigraphic analysis 44
4.2.2 Digital Models and Archaeology 45
4.2.2.1 GIS and CAD 50
4.2.2.1.1 Geographic Information System and Stratigraphic Modeling 50
4.2.2.1.2 CAD Modeling 51
4.2.2.2 Modeling the Wacheqsa Sector 54
4.2.3 Quantitative analysis 56
4.2.3.1 Analysis of Deposits 56
4.2.3.2 Boone Index 56
4.2.3.3 Kernel density estimates (KDE): univariate and bivariate 59
5 ARCHAEOLOGICAL EXCAVATIONS 63
5.1 Units Excavated 65
5.1.1 WQ-01 (N 849, E 484.5) 65
5.1.2 WQ-01 (west extension) (N 847, E 483) 65
5.1.3 WQ – 02 (N 852, E 483) 66
5.1.4 WQ-03 (N 830, E 481) 67
5.1.5 WQ-4 (N 821, E 466) 68
5.1.6 WQ-5 (N746, E432) 70
5.1.7 WQ-6 (N786, E430) 70
5.1.8 WQ – 07 (N771, E434) 69
5.1.8.1 Sector I 71
5.1.8.1.1 Unit 1 (N777, E435) 71
5.1.8.1.2 Unit 4 (N763, E441) 73
5.1.8.2 Sector II 73
5.1.8.2.1 Unit 1 (N770, E444) 73
5.1.8.2.2 Unit 4 (N764, E 451) 75
5.1.8.3 Sector III 75
5.1.8.3.1 Unit 1 (N760, E 435) 75
5.1.8.3.2 Unit 2 (N758, E441) 76
5.1.8.3.3 Unit 4 (N753, E451) 77
5.1.8.3.4 Unit 4A (N756, E 440) 78
5.1.8.4 Sector IV 79
5.1.8.4.1 Unit 3 (N754, E445) 79
x
5.1.8.4.2 Unit 4 (N754, E441) 81
5.1.9 Unit WQ-8 83
5.1.10 Unit WQ-9 83
5.1.11 Unit WQ-10 85
6 INTRASITE COMPLEXITY 88
6.1 Spatial Intrasite Complexity 88
6.1.1 Prehistoric Occupation 89
6.1.1.1 Urabarriu Phase 89
6.1.1.1.1 Water Flood 89
6.1.1.1.2 Early Platforms 91
6.1.1.2 Janabarriu Phase 92
6.1.1.2.1 Midden 92
6.1.1.2.2 Late Platforms 95
6.1.1.2.3 Stone Rooms 96
6.1.2 Modern Phase 98
6.1.2.1 Aluvion 98
6.1.2.2 Modern Canal 99
6.1.2.3 Agricultural Land 99
6.2 Boone Index Measurement 100
6.3 Ceramic intrasite complexity 103
6.3.1 OSC 104
6.3.1.1 Midden 105
6.3.1.2 Stone Rooms 106
6.3.1.3 Early Platforms 106
6.3.1.4 Water Flood 107
6.3.1.5 Late Platforms 108
6.3.2 Bowls 109
6.3.2.1 Midden 110
6.3.2.2 Stone Rooms 110
6.3.2.3 Early Platforms 111
6.3.2.4 Water Flood 112
6.3.2.5 Late Platforms 111
6.3.3 Jars 113
xi
6.3.3.1 Midden 114
6.3.3.2 Stone Rooms 114
6.3.3.3 Early Platforms 115
6.3.3.4 Water Flood 115
6.3.3.5 Late Platforms 116
6.3.4 Bottles 117
6.3.4.1 Midden 118
6.3.4.2 Stone Rooms 119
6.3.4.3 Early Platforms 119
6.3.4.4 Water Flood 120
6.3.4.5 Late Platforms 120
6.3.5 Cups 121
6.3.5.1 Midden 122
6.3.5.2 Stone Rooms 122
6.3.5.3 Early Platforms 122
6.3.5.4 Water Flood 122
6.3.5.5 Late Platforms 123
6.3.6 Plates 123
6.3.6.1 Midden 124
6.3.6.2 Stone Rooms 124
6.3.6.3 Early Platforms 125
6.3.6.4 Water Flood 125
6.3.6.5 Late Platforms 125
7 THE WACHEQSA SECTOR AS A MULTICOMPONENT AREA 129
7.1 Inferred Activities 129
7.1.1 Feasting Activities 129
7.1.1.1 Ceramics 130
7.1.1.2 Faunal Remains 132
7.1.1.3 Narcotic Paraphernalia 133
7.1.1.4 Exotic Items 134
7.1.2 Domestic Activities 136
7.1.2.1 Early Platforms 136
7.1.2.2 Stone Rooms 137
xii
7.1.3 Water Flood 139
7.1.4 Intermediate Area 139
7.2 Implications and Relevance 140
7.2.1 Feasting and Power 140
7.2.2 The Wacheqsa Sector as a domestic area 145
7.2.3 Radiocarbon Dates 150
7.2.3.1 Early Platforms 150
7.2.3.2 Water Flood 152
7.2.3.3 Midden 152
7.2.3.4 Stone Rooms 153
7.2.3.5 Late Platforms 153
7.2.4 Regional Chronological Implications 155
8 CONCLUSIONS 158
APPENDIX A. ILLUSTRATIONS 164
BIBLIOGRAPHY 276
xiii
LIST OF TABLES
Table 01: Chronological chart according to Guaman Poma and Buenaventura Salinas 18
Table 02: Proposed chronological chart of the Andean Formative 20
Table 03: Historical development of digital technologies in archaeology 46
Table 04: Cultural History of digital technologies 46
Table 05: Stratigraphic summary of unit WQ-01 65
Table 06: Stratigraphic summary of unit WQ-0, West Extension 66
Table 07: Stratigraphic summary of WQ2 67
Table 08: Stratigraphic summary of WQ3 67
Table 09: Stratigraphic summary of WQ4 68
Table 10: Stratigraphic summary of WQ5 70
Table 11: Stratigraphic summary of WQ6 71
Table 12 Stratigraphic summary of WQ7-SI-U1 72
Table 13: Stratigraphic summary of WQ7-SI-U4 73
Table 14: Stratigraphic summary of WQ7-SII-U1 74
Table 15: Stratigraphic summary of WQ7-SII-U2 75
Table 16: Stratigraphic summary of WQ7-SIII-U1 76
Table 17: Stratigraphic summary of WQ7-SIII-U2 76
Table 18: Stratigraphic summary of WQ7-SIII-U4 77
Table 19: Stratigraphic summary of WQ7-SIII-U4A 78
Table 20: Stratigraphic summary of WQ7-SIV-U3 79
Table 21: Stratigraphic summary of WQ7-SIV-U4 82
Table 22: Stratigraphic summary of WQ8 83
Table 23: Stratigraphic summary of WQ9 85
Table 24: Stratigraphic summary of WQ9 85
Table 25: Chronological chart of the Wacheqsa Sector 88
Table 26: Water Flood analytical unit strata 89
Table 27: Densities of archaeological materials from the Water Flood analytical unit 90
Table 28: Early Platform analytical unit strata 91
Table 29: Densities of archaeological materials from Early Platforms analytical unit 92
Table 30: Midden analytical unit strata 93
Table 31: Densities of archaeological materials from Midden analytical unit 94
xiv
Table 32: Late Platforms analytic unit strata 95
Table 33: Densities of archaeological materials from Late Platforms analytical unit 96
Table 34: Stone Rooms analytic unit strata 97
Table 35: Densities of archaeological materials from Stone Rooms analytical unit 98
Table 36: Aluvión analytical unit strata 98
Table 37: Layers that are part of the Modern Canal analytical unit 99
Table 38: Strata that are part of the Modern Canal analytical unit 99
Table 39: Densities of archaeological classes per analytical unit 100
Table 40: Student´s t test Least Significant Difference threshold matrix. 102
Table 41: Student´s T test connecting lines report 102
Table 42: Student´s T test ordered differences report 102
Table 43: Distribution of sampled sherds 103
Table 44: Distribution of ceramic types sampled per analytical unit 103
Table 45: Percentages of ceramic types sampled per analytical unit 103
Table 46: Results of univariate KDE for diameter and thickness measurements OSC 104
Table 47: Modal clustering table of OSC’s 104
Table 48: Measurements of total population of OCS’s types 104
Table 49: Types of OCS’s 104
Table 50: Midden’s OSC. Results of univariate KDE for diameter and thickness 105
Table 51: Modal clustering table of Midden’s OSC 105
Table 52: Measurements of OCS’s types from the Midden Analytical Unit 105
Table 53: Types of OCS’s from the Midden Analytical Unit 105
Table 54: Results of Stone Rooms univariate KDE for diameter and thickness 106
Table 55: Modal clustering table of Stone Rooms OSC’s 106
Table 56: Measurements of OCS’s types from the Stone Rooms Analytical Unit 106
Table 57: Results of Early Platfforms OSC univariate KDE for diameter and thickness 106
Table 58: Modal clustering table of Early Platfroms OSC’s 107
Table 59: Measurements of OCS’s types from the Early Platforms Analytical Unit 107
Table 60: Types of OCS’s from the Early Platforms Analytical Unit 107
Table 61: Results of Water Flood OSC univariate KDE for diameter and thickness 107
Table 62: Modal clustering table of Water Flood OSC’s 107
Table 63: Measurements of OCS’s types from the Water Flood Analytical Unit 108
Table 64: Types of OCS’s from the Water Flood Analytical Unit 108
xv
Table 65: Results of Late Platforms univariate KDE for diameter and thickness 108
Table 66: Modal clustering table of Early Platforms OSC’s 108
Table 67: Measurements of OCS’s types from the Late Platforms Analytical Unit 108
Table 68: Types of OCS’s from the Late Platforms Analytical Unit 108
Table 69: Measurements of different types of ollas sin cuello per analytical unit 109
Table 70: Sizes of ollas sin cuello per analytical unit 109
Table 71: Results of univariate KDE for diameter and thickness measurements 109
Table 72: Modal clustering of total bowl sample 109
Table 73: Results of Midden’s bowls univariate KDE for diameter and thickness 110
Table 74: Modal clustering of bowl sample from Midden 110
Table 75: Results of Stone Rooms’s bowls univariate KDE for diameter and thickness 110
Table 76: Modal clustering table of bowl sample from Stone Rooms 111
Table 77: Results of Stone Rooms’ bowl univariate KDE for diameter and thickness 111
Table 78: Modal clustering of bowl sample from Early Platforms 111
Table 79: Measurements of bowl types from Early Platforms 111
Table 80: Types of bowls from Early Platforms 111
Table 81: Results of Water Flood’s bowl univariate KDE for diameter and thickness 112
Table 82: Modal clustering table of bowls from Water Flood 112
Table 83: Measurements of bowl types from Water Flood 112
Table 84: Types of bowls from Water Flood 112
Table 85: Results of Late Platforms bowl’s univariate KDE for diameter and thickness 112
Table 86: Modal clustering table of bowls from Late Platforms 113
Table 87: Measurements of bowls types from Late Platforms 113
Table 88: Types of bowls from Late Platforms 113
Table 89: Measurements of different types of bowls per analytical unit 113
Table 90: Sizes of bowls per analytical unit 113
Table 91: Results of Jars univariate KDE for diameter and thickness 113
Table 92: Modal clustering table of jars 114
Table 93: Results of Midden jars univariate KDE for diameter and thickness 114
Table 94: Modal clustering table of jars from Midden 114
Table 95: Results of Stone Rooms jars univariate KDE for diameter and thickness 114
Table 96: Modal clustering table of jars from Stone Rooms 115
Table 97: Results of Early Platforms jars univariate KDE for diameter and thickness 115
xvi
Table 98: Modal clustering table of jars from Early Platforms 115
Table 99: Measurements of jars types from Early Platforms 115
Table 100: Types of jars from Early Platforms 115
Table 101: Results of Water Flood jars univariate KDE for diameter and thickness 116
Table 102: Modal clustering table of jars from Water Flood 116
Table 103: Measurements of jars types from Water Flood 116
Table 104: Measurements of jars types from Water Flood 116
Table 105: Measurements of jar sample from Late Platforms 116
Table 106: Measurements of jars types from Late Platforms 117
Table 107: Measurements of jars types from Late Platforms 117
Table 108: Measurements of overall jars types per analytical unit 117
Table 109: Jar types per analytical unit 117
Table 110: Results of Bottles univariate KDE for diameter and thickness 118
Table 111: Modal clustering table of bottles 118
Table 112: Results of Midden’s bottle univariate KDE for diameter and thickness 118
Table 113: Modal clustering table of bottles from Midden 119
Table 114: Measurements of bottle’s spouts types from Midden 119
Table 115: Types of bottle’s spouts types from Midden 119
Table 116: Measurements of bottle sample from Stone Rooms 119
Table 117: Measurements of bottle sample from Early Platforms 120
Table 118: Measurements of bottle sample from Water Flood 120
Table 119: Measurements of bottle spouts per analytical unit 121
Table 120: Bottle spouts types per analytical unit 121
Table 121: Results of cup’s univariate KDE for diameter and thickness 121
Table 122: Modal clustering table of cups 121
Table 123: Modal clustering table of cups from Midden 122
Table 124: Measurements of cup types from Midden 122
Table 125: Types of cups from Midden 122
Table 126: Measurements of types of cups per analytical unit 123
Table 127: Types of cups per analytical unit 123
Table 128: Results of plate’s univariate KDE for diameter and thickness measurements 123
Table 129: Modal clustering table of plates 124
Table 130: Modal cluster table of plates from Midden 124
xvii
Table 131: Measurements of type of plates from Midden 124
Table 132: Measurements of plate rims from Stone Rooms 124
Table 133: Measurements of types of plates from Stone Rooms 125
Table 134: Types of plates from Stone Rooms 125
Table 135: Measurements of plate rims from Early Platforms 125
Table 136: Measurements of plate rims from Water Flood 125
Table 137: Measurements of type of plates per analytical unit 126
Table 138: Types of plates per analytical unit 126
Table 139: Summary of vessels type from Midden 126
Table 140: Summary of vessels type from Stone Rooms 126
Table 141: Summary of vessels type from Early Platforms 127
Table 142: Summary of vessels type from Water Flood 127
Table 143: Summary of vessels type from Late Platforms 127
Table 144: Student’s t test of faunal remains. LSD threshold matrix 132
Table 145: Student’s t test of faunal remains. Ordered differences report 132
xviii
LIST OF ILLUSTRATIONS
Figure 01 Satellite photograph of Chavín de Huántar 165
Figure 02: Map of the monumental core of Chavín de Huántar 166
Figure 03: The Wacheqsa sector viewed from the site of Shallapa, 167
Figure 04: The Wacheqsa sector viewed from the top of Mound D at Chavín de Huántar 167
Figure 05: Garagay date associated to Janabarriu-like ceramics 168
Figure 06: Garagay dates 168
Figure 07: Examples of Janabarriu-like ceramics from the Wacheqsa sector 169
Figure 08: Chavín dates recovered by Burger 169
Figure 09: Huarás date GIF-1079 170
Figure 10: Dates from the site of Kotosh 170
Figure 11: Dates from the site of La Pampa 171
Figure 12: Kunturwasi dates from the Kunturwasi phase 171.
Figure 13: Huarás dates published by Lau (2002) 171
Figure 14: Map of the Wacheqsa Sector 172
Figure 15: The modern town Chavín and the Wacheqsa sector after the 1945 landslide 173
Figure 16: Tello’s excavations at the Wacheqsa Sector 173
Figure 17: Terracing of the Wacheqsa Sector 174
Figure 18: Agricultural fields’ delimeted by parkas 175
Figure 19: Agricultural terraces 175
Figure 20: Retention wall along the Wacheqsa Sector 176
Figure 21: Retention wall along the Wacheqsa Sector 176
Figure 22: Map of Bennett’s, Tello’s and Fung’s inferred excavations 177
Figure 23: Tello’s excavation profile and associated ceramics recovered 178
Figure 24: Ceramics recovered by Rosa Fung. Unit H2, Layer 5, level 2 178
Figure 25: Ceramics recovered by Rosa Fung. Test pit 3, layer 2 and level 2 179
Figure 26: Screenshot of the process of modeling stratigraphy from the Wacheqsa Sector 180
Figure 27: Texture subsurface strata modeled with Autodesk Land 180
Figure 28: Wireframe model of strata from the Wacheqsa Sector 181
Figure 29: Same strata after textures are applied 181
Figure 30: Excavations at the Wacheqsa Sector 182
Figure 31: Wacheqsa Sector before excavations started in 2003 183
xix
Figure 32: Systematic sampling strategy used in year 2005 183
Figure 33: Location of units WQ1 and WQ2 on the north edge of the Wacheqsa sector 184
Figure 34: Stone platform located on WQ1 184
Figure 35: Excavation of Feature 1 in WQ1-AW 185
Figure 36: South profile of WQ1 and WQ1-WE 186
Figure 37: Visible walls at the northern edge of the Wacheqsa sector 187
Figure 38: Exposed deposits in WQ2 before excavations started 188
Figure 39: Feature 02 in WQ2, Layer 03 and exposed section of Floor 2. 188
Figure 40: Excavation of WQ3 and WQ4 189
Figure 41: Excavation of aluvión layer in WQ3 189
Figure 42: Stone platform (layer 2a) associated to wall (Feature 01) 190
Figure 43: Stone platform, wall (Feature 01) and floor associated 191
Figure 44: Plan of exposed architecture in WQ4 192
Figure 45: Stone room, associated floor and wall (Feature 2) that delimits an alley 191
Figure 46: Layer 8 and associated Features 03 and 04 192
Figure 47: WQ4, east profile 193
Figure 48: Location of WQ5 194
Figure 49: Hearth excavated in WQ6 194
Figure 50: Panoramic view of WQ7 195
Figure 51: Profile of WQ7, SI, U1 195
Figure 52: Plant drawing of stone platform 196
Figure 53: Stratigraphic section of WQ7, SI, U1 196
Figure 54: Profile of WQ7, SII, U1 197
Figure 55: Plan of architecture exposed in WQ7, SII, U1, 197
Figure 56: Stone platform in WQ7, SIII, U2 198
Figure 57: Profile of WQ7, SIII, U4 198
Figure 58: East profile of WQ7, SIII, U4A 199
Figure 59: Excavation unit WQ7, SIII, U4A 199
Figure 60: Excavation WQ7, SIII 2, 4 and 4A 200
Figure 61: South profile of WQ7, SIV, U3 201
Figure 62: Stratigraphic detail of WQ7, SIV, U4 202
Figure 63: WQ8, plan of architecture exposed 203
Figure 64: Spatial distributions of analytical units in the Wacheqsa Sector 204
xx
Figure 65: Water Flood analytical unit 205
Figure 66: Water Flood analytical unit. Common depositional events in units excavated 206
Figure 67: Detail of WQ4. Stone Rooms analytical unit on top of Early Platforms 207
Figure 68: Ceramics recovered in unit WQ-1 WE, layer 8 208
Figure 69: Stone miniature found in WQ4, layer 6 208
Figure 70: Ceramics recovered in WQ4, layer 7 209
Figure 71: Ceramics recovered in WQ4, layer 6 209
Figure 72: Idem 210
Figure 73: Stratigraphic relationship Midden and Water Flood analytical units 211
Figure 74: Stratigraphic modelling of Midden analytical unit 212
Figure 75: Slate projectile points recovered in Midden analytical unit 213
Figure 76: Bone artefacts recovered in the Midden Analytical Unit 213
Figure 77: Slate projectile point recovered in WQ7, SIII, U4A, L10 214
Figure 78: Molded Frieze recovered in WQ7, SIII, U4A, L16 214
Figure 79: Ceramics found in the Midden Analytical unit 215
Figure 80: Ceramics found in the Midden Analytical unit 216
Figure 81: Ceramics found in the Midden Analytical unit 217
Figure 82: Fragments of columns 217
Figure 83: Fragments of burnt architectural features (floors) 218
Figure 84: Ceramics retrieved from WQ4, L2 218
Figure 85: Ceramics retrieved from the Stone Rooms analytical unit 219
Figure 86: Fragment of unworked chrysocolla. 220
Figure 87: Fragment of cooper ore 220
Figure 88: Spatial distribution of Stone Rooms analytical unit in relation with
Late Platforms and Midden units 221
Figure 89: WQ8, stratigraphic relationship between Early Platforms
and Stone Room analytical units 222
Figure 90: Aluvion and Agricultural Land sections 223
Figure 91: Stratrigraphic Harris Matrix of strata recorded 224
Figure 92: Volume excavated per Analytical Unit 225
Figure 93: Density of archaeological materials per Analytical Unit 225
Figure 95: Distribution of analytical units according to sample size (log) 227
Figure 96: Distribution of analytical units according to Hi values 227
xxi
Figure 97: Confidence interval (90%) of Hi values 228
Figure 98: Probabilities of Hi [p(Hi)] values after 10000 228
Figure 99: Distribution (log) of p(Hi) values per analytical unit 229
Figure 100: Percentage of ceramic types sampled per analytical unit 229
Figure 101: KDE plot of rim diameters of ollas sin cuello 230
Figure 102: KDE plot of rim thicknesses of ollas sin cuello 230
Figure 103: Bivariate KDE plot of ollas sin cuello. 231
Figure 104: Rim diameters of ollas sin cuello from Midden 231
Figure 105: Rim thicknesses of ollas sin cuello from Midden 232
Figure 106: Bivariate KDE of OSC’s from Stone Rooms 232
Figure 107: Rim diameters of ollas sin cuello from Stone Rooms 233
Figure 108: Rim thicknesses of ollas sin cuello from Stone Rooms 233
Figure 109: Bivariate KDE of OSC’s from Stone Rooms 234
Figure 110: Rim diameters of ollas sin cuello from Early Platforms 234
Figure 111: Rim thicknesses of ollas sin cuello from Early Platforms 235
Figure 112: Bivariate KDE of OSC’s from Early Platforms 235
Figure 113: Rim diameters of ollas sin cuello from Water Flood 236
Figure 114: Rim thicknesses of ollas sin cuello from Water Flood 236
Figure 115: Bivariate KDE of OSC’s from Water Flood 237
Figure 116: Rim diameters of ollas sin cuello from Late Platforms 237
Figure 117: Rim thicknesses of ollas sin cuello from Late Platforms 238
Figure 118: Bivariate KDE of OSC’s from Late Platforms 238
Figure 119: Rim diameters of bowls 239
Figure 120: Rim thicknesses of bowls 239
Figure 121: Bivariate KDE of bowls 240
Figure 122: Rim diameters of bowls from Midden 240
Figure 123: Rim thicknesses of bowls from Midden 241
Figure 124: Bivariate KDE of bowls from Midden 241
Figure 125: Rim diameters of bowls from Stone Rooms 242
Figure 126: Rim thicknesses of bowls from Stone Rooms 242
Figure 127: Bivariate KDE of bowls from Stone Rooms 243
Figure 128: Rim diameters of bowls from Early Platforms 243
Figure 129: Rim thicknesses of bowls from Early Platforms 244
xxii
Figure 130: Bivariate KDE of bowls from Early Platforms 244
Figure 131: Rim diameters of bowls from Water Flood 245
Figure 132: Rim thicknesses of bowls from Water Flood 245
Figure 133: Bivariate KDE of bowls from Water Flood 246
Figure 134: Rim diameters of bowls from Late Platforms 246
Figure 135: Rim thicknesses of bowls from Late Platforms 247
Figure 136: Bivariate KDE of bowls from Late Platforms 247
Figure 137: Rim diameters of jars 248
Figure 138: Rim thicknesses of jars 248
Figure 139: Bivariate KDE of jars 249
Figure 140: Rim diameters of jars from Midden 249
Figure 141: Rim thicknesses of jars from Midden 250
Figure 142: Bivariate KDE of jars. Arrow indicates the mode identified 250
Figure 143: Rim diameters of jars from Stone Rooms 251
Figure 144: Rim thicknesses of jars from Stone Rooms 251
Figure 145: Bivariate KDE of jars from Stone Room. 252
Figure 146: Rim diameters of jars from Early Platforms 252
Figure 147: Rim thicknesses of jars from Early Platforms 253
Figure 148: Bivariate KDE of jars from Early Platforms 253
Figure 149: Rim diameters of jars from Water Flood 254
Figure 150: Rim thicknesses of jars from Water Flood 254
Figure 151: Bivariate KDE of jars from Water Flood 255
Figure 152: Rim diameters from bottles 255
Figure 153: Rim thicknesses from bottles 256
Figure 154: Bivariate KDE of bottles 256
Figure 156: Rim thicknesses of bottles from Midden 257
Figure 157: Bivariate KDE of bottles from Midden. 258
Figure 158: Rim diameters of cups 258
Figure 159: Rim thicknesses from cups 259
Figure 160: Bivariate KDE of cups. Arrows indicate modes identified 259
Figure 161: Rim diameters of cups from Midden 260
Figure 162: Rim thicknesses of cups from Midden 260
Figure 163: Bivariate KDE of cups from Midden. 261
xxiii
Figure 164: Rim diameters of plates 261
Figure 165: Rim thicknesses of plates 262
Figure 166: Bivariate KDE of plates. Arrows indicate the modes identified 262
Figure 167: Rim Diameters of plates from Midden 263
Figure 168: Rim Thickness of plates from Midden 263
Figure 169: Bivariate KDE of plates. Arrow indicate the mode identified 264
Figure 170: Density of faunal remains per analytical unit (log) 264
Figure 171: Fragment of a small spoon retrieved in WQ7,SIV, U3, L11 265
Figure 171: Fragment of a small spoon retrieved in WQ7,S III, U4A, L16 265
Figure 172: Polished bone tube retrieved in WQ7,SIII, U4A, L19 266
Figure 173: Polished bone tube retrieved in WQ7, SIII, U4A, L18 266
Figure 174: Polished bone tube retrieved in WQ7, SIII, U4A, L11 267
Figure 174: Polished bone tube retrieved in WQ7, SIII, U4A, L12 267
Figure 175: Bottle fragment found associated to hearth in WQ-6 268
Figure 176: Ceramics associated to hearth in WQ-6 268
Figure 177: Ceramic associated to hearth in WQ-6 269
Figure 178: Ceramics from WQ1, layer 5 269
Figure 179: Ceramic sherd from WQ1, Floor 3 270
Figure 180: Ceramic sherds from WQ4, layer 6 270
Figure 181: Ceramic sherds from WQ5, layer 6 271
Figure 182: Ceramics sherds from WQ7, SIV, U4, layer 20 271
Figure 183: Ceramics retrieved from WQ-7, SIII, U4, layer 8 272
Figure 184: Ceramics retrieved from WQ-7, SIV, U4, layer 14 272
Figure 185: Ceramics retrieved from WQ-7, SIV, U4, layer 10 273
Figure 186: Ceramics retrieved from WQ8, layer 3 273
Figure 187: Ceramics retrieved from WQ-7,SIII, U1, layer 9 274
Figures 188: Urabarriu dates retrieved by Burger 274
Figure 189: Wacheqsa Sector radiocarbon dates 275
101
CHAPTER 1
INTRODUCTION, RESEARCH HISTORY AND RESEARCH DESIGN
1.1 Introduction
The ceremonial center of Chavín de Huántar is located in the province of Huari,
department of Ancash, in the north central Peruvian highlands at 3,200 meters above sea level
at the junction of the Wacheqsa and Mosna Rivers (figure 01). Since the work of Peruvian
archaeologist Julio C. Tello in 1919, Chavín has been a pivotal element in the discussion of
the origins of social complexity in the Andes, as the Andean “mother culture” (Tello 1942,
1943, 1960), as a derivative of social development started in Mesoamerica (Uhle 1902), as a
synthesis of previous Peruvian coastal and highland developments (Burger 1988; Burger
1992), or as a place where authority was being crafted and transmitted (Kembel and Rick
2004; Rick 2005, 2006).
Either as a “mother culture” or as a “synthetic culture”, Chavín de Huántar1
has been
constantly present in any discourse regarding early social complexity in the Peruvian Andes.
One cannot help but ask the question, why is Chavín so preeminent in Andean archaeological
studies? The reason for this extensive interest in Chavin lies in its engraved sculpture, and
monumental architecture characterized by superimposed platforms, impressive buildings,
underground galleries, and open plazas (figure 02). As a result of its impressive architecture,
most of the archaeological work in Chavin has focused upon the monumental area (Bennett
1944; Kembel 2001; Lumbreras 1989, 1993; Lumbreras and Amat 1965; Rick, et al. 1998),
trying to understand the construction sequence of the site, the meaning of the iconography,
and the relative chronological sequences.
However, a limited amount of research has been carried out in the areas surrounding
the site of Chavín de Huántar (Burger 1984; Burger 1998; Espejo 1941). The nature of the
social activities believed to have been carried out in these areas varies according to the
location and proximity to the ceremonial area. For instance, the contexts identified at La
Banda – located at the east bank of the Mosna river – are primarily domestic but with a
ceremonial imprint (Rick 2005), while the contexts identified at the West Field mostly have
ceremonial components (Contreras 2007). Along the same lines, the location of an extensive
Chavín domestic area was suggested by Rivero (Rivero de Ustariz 1851) and confirmed by
Richard Burger who excavated samples of an extensive domestic settlement to the north of the
1 From now on, Chavín de Huántar will be referred to as the archaeological site unless otherwise stated.
2
Wacheqsa river where the modern town of Chavín lies (Burger 1984; Burger 1998). The
investigation of the immediate periphery provides important information regarding the
different activities that have surrounded the monumental core, varying from sacred to
domestic, that are no less complex than those carried out in the architectural center.
Following this line of research, I investigated the Wacheqsa sector, located
immediately to the north of the ceremonial core, and enclosed by the Mosna River, the
Wacheqsa River and the northern platform of the monumental area (figures 03 and 04). The
main objective of my investigation has been the understanding of the archaeological deposits
that have formed this area, and analyzing their chronological and spatial relationships. The
goal of my investigation is the development of contextual information previously unknown in
Chavín de Huántar and a clear understanding of this sector’s role during the Middle and Late
Formative periods. This information is relevant for the comprehension of power strategies
developed by the authorities of Chavín de Huántar for enticing foreign elites and local
population into their religious system and also provides relevant data about the chronology of
Chavín de Huántar and the relationships of the ceremonial center with others of the Middle
and Late Formative (1200 – 500 BC).
1.2 Research History at Chavín de Huántar
In this section I intend to explore the historical development of knowledge constructed
about Chavín de Huántar and how specific historic moments have marked qualitative and
quantitative changes in the comprehension of the site. I consider this information relevant in
order to have a clear understanding of how interpretations about the site have changed from
the early Spanish accounts to the present. Also I would like to provide an updated research
history that can be used in conjunction with the excellent overview published by Lumbreras in
1989 (Lumbreras 1989)
1.2.1 Early Accounts
The first account of the archaeological site comes from Pedro Cieza de León who in
1553 wrote Crónica del Perú (Cieza de León 1553 [1984]). As part of his duties as a soldier
of the Spanish army, he extensively traveled across the Inca Empire, writing in his free time
the impressions of his traveling experiences. He describes an “old fortress” located eight
leguas from the province of Piscobamba in the Callejón de Conchucos,
3
“Among the old rooms, a big fortress can be seen, which is like a block and had 140 steps and was even wider, and in many places of them are faces and human sculptures, everything beautifully made, and some natives say that the Incas, as a sign of victory, built this
memorial 2” (Cieza De León, 1553: 271. Emphasis added)
The references to walls covered by sculpted faces as well as the location that Cieza provides
are compelling enough to state that the site described by him is actually Chavín de Huántar.
He not only described the site but assigned a rough chronological placement to the site. Cieza
believed that Chavín de Huántar was constructed long before the time of the Incas,
“Others tell and have it more certain, that this is not true, that in very ancient times, long before the Incas ruled there were in those lands men like giants, so tall that they show it in the stone sculptures but with time and the big war they had with those who are rulers now,
the giant lords weakened and disappeared3” (Cieza De León, 1553:
271)
After Cieza’s account, in 1593, Toribio Alfonso de Mogrovejo, head of the
Archdioceses of this region arrived at the town of San Pedro de Chavín, visiting the
archaeological site (Mogrovejo 1593 [1920]). Unlike Cieza, Mogrovejo assigns Chavín de
Huántar the category of huaca, a sacred place,
“Near this town, there is an ancient huaca, which is in a fortress and inside this huaca, there are several passages under it, and there are stories that the huaca has had a lot of richness; it [the huaca] has not
been completely discovered and some parts are contaminated4”
(Toribio de Mogrovejo, 1593: 412. Emphasis added)
Adding more information regarding the sacred nature of Chavín de Huántar, Father
Antonio Vásquez de Espinoza wrote in 1630 that Chavín de Huántar was a pilgrimage center,
similar to Rome or Jerusalem (Vázquez de Espinoza 1616 [1948]). In 1631 the Catholic Jesuit
order took control of San Pedro de Chavín until 1650 when Lima’s Archbishop took over.
During this time, Jesuits actively worked in the extirpation of idolatries mandated by Viceroy
Toledo, and the effects of their work can be seen in a large trench dug in the middle of
Building A with the attempt to discover “false idols” venerated by native populations. Stone
pieces were even used as construction materials for the local church (Mejía 1945). The reason
for the active presence of clerics in this section of the Callejón de Conchucos is that this valley
2 My translation. 3 My translation. 4
My translation.
4
occupied a strategic position as an alternative route to the Marañón valley, in the eastern
slopes of the Cordillera de los Andes (Polo 1900).
Peru freed itself from the Spanish in 1821 a year afterwards the National Museum
was created, with Mariano Rivero Ustaríz as its first director. In 1840 he visited San Pedro de
Chavín5
and Chavín de Huántar, attracted by chronicler’s accounts. His observations were
published in 1851, in the book called “Antigüedades Peruanas” (Rivero de Ustariz 1851)6.
There are three elements that must be emphasized in Rivero’s account: a) a description of the
inner rooms of the main building (Building A), b) the identification of the town of San Pedro
as an extension of the civic ceremonial center, and c) the linking of the puna settlement of
Pojoc with Chavín de Huántar (Rivero 1851: 284). He is the first scholar who attempted to
link Chavín’s surroundings with the site, he mentioning that the town of San Pedro was
constructed above a web of aqueducts and that Pojoc probably had the same function as
Chavín. However, he does not state which function this was.
1.2.2 Travelers
Polo de Ondegardo’s presence in 1871 marks the beginning of a time characterized
by several travelers’ accounts of Chavín de Huántar. While he was in town as a public officer,
he found a rectangular stone slab that depicted a supernatural being in frontal position,
wearing a headdress and holding one staff in each hand. This stela was found in the house of
Lazaro Palacios, an inhabitant of Chavín who used it as a dinner table and it was taken to Lima
in 1873. Currently this slab is known as the “Raimondi Stone”. Polo de Ondegardo also visited
Chavín de Huántar and provided a description of it.
“Here lie the remains of an ancient building called ‘El Castillo’. Judging from the terrain, the rubble and the underground galleries, it looks like the site was almost rectangular, 150 m long and 60 m – 70 m wide, with the main façade probably looking towards the Puccha and towards Posoc […] The entrance must have been guarded by two wings oriented to the river as secondary constructions, giving the whole building the shape of an E. There are several galleries and rooms running in different directions creating a labyrinth, it has to be
noted that these galleries intersect at a central point7” (Polo 1900:
220).
5
When referring to the modern town of Chavín de Huántar I will use its original name, San Pedro de
Chavín. 6 This is the first book ever written about Peruvian archaeology. 7
My translation.
5
The central point where the four galleries join is the Lanzón gallery, a gallery where a
stone sculpture (the Lanzón) more than fours meter high sits at the conjunction. Polo offers
the first description of the Lanzón.
“There is monolith that looks like a big lance [lanzón], 2.20 m high. Its lower part has three sides, stretching towards its upper portion; it fits the ceiling leaning towards a rounded stone that supports it. It resembles an ox’s head, and engravings of fangs, staffs, snakes, lizards and even condor heads and monkeys can be noticed on its
sides8” (Polo 1900:222)
One of the main characteristics of Chavín art is the recurrent use of zoomorphic images as
metaphors for anthropomorphic parts (Rowe 1962a); the snakes, fangs, condors and monkeys
that Polo refers to are metaphors of body sections of an anthropomorphic deity sculpted in the
Lanzón. In 1873, the Italian traveler Antonio Raimondi arrived at San Pedro de Chavín,
attracted by the fame of Chavín de Huántar. He stated that the site had at least a military
component as part of it and that,
“Crossing the smaller river [Wacheqsa] there is a big wall towards the side of the road, built with quarried stones, joint together without clay. This wall is just a small detail of the splendid construction. Judging from the remains existing, it looks like this castle had a rectangular form with two wings oriented towards the river, giving the whole structure the shape of an open parallelogram. In front of the wings, and almost at the shore of the river, there are two
platforms which are the remains of two fortresses9” (Raimondi,
1873: 211. Emphasis added).
In 1875, French traveler Charles Wiener arrived at San Pedro de Chavín, attracted by
Rivero’s descriptions. Wiener indicated that Chavín de Huántar was composed of two
terraces, the lower one located 11 m above the Wacheqsa River, and that the upper one was
crossed by several ducts. Wiener paid attention to the iconography depicted in different slabs
dispersed across the site. He thought that the images represented on these architectural
elements were depictions of a Chavín god, leading Wiener to conclude that Chavín was not a
fortress but a temple (Wiener 1880 [1993])
At this point most of the accounts of Chavín de Huántar were merely descriptions of
its architecture – with the possible exception of Rivero – making few references to its
iconography. The situation dramatically changed with the arrival of Ernst Middendorf who
8 My translation 9
My translation.
6
tried to link Chavín de Huántar with other archaeological sites he explored during his travels
across Peru. Middendorf paid special attention to sites on the coast, hypothesizing the
existence of two kingdoms, one on the coast, and one in the Callejón de Conchucos that
maintained close contacts in a time before the Incas. He based his interpretations on the
iconography depicted on the Lanzón and in other lithic art scattered across Chavín de Huántar
and in its architecture. He recognized that the construction technique and art styles from
Chavín were very different from those used by the Incas and consequently Chavín de Huántar
must have been constructed in a pre-Inca time. Middendorf believed that Chavín de Huántar
functioned as a temple and as a palace but not as the capital of the Kingdom of the Callejón
de Conchucos (Middendorf 1893 [1974]).
1.2.3 The Archaeologists arrive
As might be expected, the arrival of trained professionals triggered a quantitative and
qualitative jump in the knowledge of Chavín de Huántar. Archaeologists formally entered
Chavín de Huántar in 1919. Julio C Tello, then- Director of the San Marcos University
Museum of Archaeology and Anthropology stopped in Chavín on his first archaeological
expedition to the Marañón River. He did not carry out excavation work, but rather cleaned
one of the staircases of the main building, collected dispersed stone sculptures from the site
and the town of San Pedro de Chavín, and excavated the foundations of the Lanzón, exposing
2.23 additional meters to the already exposed 2.20 m. of the sculpture (Tello 1940). This was
in a true sense an exploratory field season, but also productive enough to allow Tello to make
some crucial interpretations about the nature of Chavín de Huántar, interpretations that would
frame the debate about the origins of social complexity in the Andes for several years. Based
on his experience in Chavín and his knowledge of Andean ethnography, Tello wrote an essay
named “Wiracocha” meant to explain the unique character of the Andean religious system
that was grounded in the worshiping of a divinity that lasted from the origins of Andean
civilization until Inca times. Tello argued this divinity was represented on the Lanzón and
Tello obelisk, the latter being an elongate carved stone he found in one of the plazas of
Chavin that depicted two zoomorphic deities incorporating different crop plants and animals
(Tello 1923)
Chavín de Huántar was, therefore, conceptualized by Tello as the place where
Andean civilization originated under the cult of the Jaguar or Dragon, entities later
reinterpreted by the Incas as Wiracocha. Chavín acted as the core of civilization from where
7
the Jaguar/Dragon religion spreads to the civilizing Andean population. In 1929 Tello defined
Chavín as part of the “Megalithic Culture” described in the XVII century by Guaman Poma
de Ayala (Poma de Ayala 1992 [1613])and Father Buenaventura Salinas (Salinas y Córdoba
1957 [1630]). Inca oral tradition recorded by these two chroniclers mentions a first age of
Andean history characterized by monumental stone architecture, which led Tello to identify
Chavín as part of this early age, named as Auca Runa by Poma and Buenaventura. Tello used
Chavín as evidence against the exogamic interpretations of Andean social complexity of Max
Uhle. In 1902 the German archaeologist Uhle wrote that civilization in Peru was a product of
migrating populations coming from Mesoamerica, an explanation that Tello opposed using
Chavín as evidence not only of the antiquity of Andean civilization but also as the center from
where Andean culture spread to the rest of the Andes (Tello 1929)
In 1925 a landslide from the east bank of the Mosna River modified the course of the
river and in 1930 a similar event destroyed the east portion of Building E, exposing an
architectural cross-section of the structure (Tello 1940). This event was crucial in Tello’s
understanding of Chavín de Huántar. When Tello arrived for a second visit in 1934 he studied
the stratigraphy of the cut left in the building by the river, noticing the presence of ceramics
under the exposed architecture. Before this event Tello was willing to concede that ceramics
from the coast may have been earlier than Chavín, because in his first visit in 1919 he did not
find in situ Chavín ceramics. He was concerned about the viability of the model that he had
constructed for Chavín de Huántar of Chavín as a mother culture. The discovery of these
sherds in stratigraphic layers gave him the elements needed not only to solidify his model of
Chavín as the ‘core’ and the coast as the ‘periphery’ but also to elaborate on the long
construction tradition that had occurred at Chavín (Tello 1943).
In 1938, Wendell Bennett arrived at Chavín de Huántar and excavated 16 test pits in
28 days, two of them located in the Wacheqsa sector which I will treat separately. Bennett’s
excavations were based on coarse arbitrary levels of 50 cm. each. The materials he recovered
were presented as a whole with no differentiation of context or excavation unit in which they
were found (Bennett 1944). In a 1943 paper he stated that the archaeological site of Chavín de
Huántar was the manifestation of a distinctive lithic and ceramic style present in different
areas of the Andean region, but in contrast to Tello, it was not necessarily the origin of these
styles (Bennett 1943). Bennett established the chronological position of Chavín de Huántar
based mostly on correlations with materials coming from different areas rather than from
Chavín itself, arguing, “Excavations in Chavín itself have so far shed little light on the
8
ultimate chronological position of this period” (Bennett 1943: 323). Bennett found Chavín to
be the oldest pan-Andean style, older than Tiahuanaco and the ‘white on red’ style identified
by Willey on the central coast.
In late 1940 Tello returned to Chavín for the third and last time with the main task of
repairing the Mosna retaining wall that was destroyed by landslides in 1925 and 1930. Tello
uses this project as an excuse to carry out excavations in different sectors of Chavín. Tello
benefited from this visit by excavating in different areas and creating the first museum at
Chavín which reflected the importance that Chavín had for him and how museums fit his
political agenda (Mesia 2006). Tello inaugurates the Museo de Sitio de Chavín on November
14, 1940, declaring,
“At first we thought in building a small fence in the main square;
later we received the suggestion of cleaning one of the galleries to
store the stone sculptures; at last we decide to take over the chapel on
top of Mound B” (Tello 1940:9)
Tello cleaned the western façade, the south façade and north section of the eastern
façade of Building A. He also conducted research in the Wacheqsa sector10
. Like Wiener,
Tello noted the presence of a complex web of canals extending within almost all the
monumental core, even to the west, in the West Field:
“If there are canals along the Wacheqsa edge, whose upper part is 15 m above the river, it is to be expected there will be canals crossing these lands [West Field lands] in order to drain water to the Puchka River [Mosna] and not destroy the lands and buildings built in this
area11
” (Tello 1940:26).
Upon finishing his fieldwork at Chavín, Tello proposed the first architectural growth
sequence for the site, composed of three phases:
“What appears in this façade [Building A], are the remains of three
buildings joint together. It is possible that these buildings not only
occupy the eastern section of the monument but even the western
portion. One of my excavations exposed the complete eastern façade
and recorded its foundations and lower walls, allowing the
uncovering of these three joint buildings, which nowadays form a
quadrangular platform” (Tello 1940:63).
Tello offered a well-thought-out summary of his interpretations in the paper “Origen y
Desarrollo de las Civilizaciones Prehistóricas Andinas”, published in 1942 (Tello 1942). In
10 The nature of this research will be discussed in chapter 3. 11
My translation.
9
this work Tello amply elaborates several lines of diffusion derived from Chavín de Huántar.
Chavín became the foundation from which two early civilizations develop: the Paracas-Cusco
and the Pucara-Tiahuanaco. According to Tello, these cultures emerge between 1000 BC and
the beginnings of our era. The importance of these statements is that for the first time there is
an attempt to place Chavín in calendar years. In spite of having been developed from a
combination of relative chronologies, his temporal construction is reasonably accurate.
1.2.4 After the aluvión, more archaeologists arrive
In 1945, a landslide coming from the mountains west of Chavín nearly completely
covered the ceremonial center, destroying the site museum and taking with it 155
archaeological sculptures, including tenon heads, engraved slabs, stone mortars and other
artifacts, not to mention the loss of life due to this unfortunate natural disaster. The
archaeological site was completely covered by mud and rocks and had to be uncovered all
over again. Tello, as Director of the National Museum, gave local disciple Marino Gonzales
the task of recovering Chavín from the landslide sediments.
Gonzales was very active between the years of 1947 and 1968, carrying on diverse
excavations in the monumental core, some of them under the label of “cleaning operations”.
Regrettably, there are no major sources of information about this work, but some salient
points are known (Lumbreras 1989). In 1958 Gonzales discovered the Rocas Canal, the
major drain of the monumental core. In 1959 he and Manuel Chavez Ballón excavated the
north façade of Building A with the purpose of finding a doorway similar to the one located in
the eastern façade of Building A. Gonzales and Ballón found an entrance without columns,
and the Stairway Gallery. This gallery was excavated by Gonzales in 1961, revealing its stairs
and walls covered by yellow plaster (Lumbreras 1989). Unfortunately, the total scope of
Gonzales’s work is yet unknown and awaits careful scrutiny of Gonzales’ field notes and
photographic record.
In 1961 and 1963 John Rowe made short but important visits to Chavín. He
excavated near the Mosna River, but unfortunately, there are no published accounts about the
outcome of these excavations (Lumbreras 1989). However, Rowe’s relevance to Chavín
studies does not derive from excavation but from his observations on Chavín architecture and
lithic art. Like Tello, Rowe proposed that the architectural seams in the east façade of
Building A defined chronological differences and, as Tello before him, Rowe suggested the
presence of three architectural phases. However, unlike Tello, Rowe included the rest of the
10
buildings that are part of the monumental core, establishing the pivotal architectural sequence
for Chavín: Old temple, Old Temple’s south extension, and New Temple (Rowe 1962a).
Rowe also continued with what Tello had left unfinished in Wiracocha and established a
diachronic dimension of the lithic art proposing four phases of Chavín stone art.
“The anchor point in this sequence is phase D, for which sculptures of
the Black and White portal serve as a standard comparison […] For
phase C at Chavín the standard of comparison is the most elaborated
single monument at Chavín art, the so called ‘Tello Obelisk’ […]
Phase AB includes all examples of Chavín art which are earlier than
the ‘Tello Obelisk’, while EF includes all those which are later than
the Black and White portal. Among the monuments assigned to AB
by these criteria are the cornice blocks from the main structure of the
new temple […] The Great Image in the old temple can be assigned to
Phase AB […] Phase EF includes the famous Raimondi Stone”
(Rowe 1962a:14)
Even though his statements were mainly supported by superficial observations and
objects with no contexts, this sequence, as well as the architectural one, would be highly
influential in years to come.
In 1966, Luis Lumbreras and Hernán Amat began a long-term project in the civic
ceremonial center. Lumbreras carried out excavation in the Old Temple’s atrium and the
Ofrendas Gallery located immediately at the north of the same atrium. While excavating in
the atrium, Lumbreras and his team identified a circular sunken court or circular plaza
covered by two extensive post Chavín domestic settlements, one settled above Chavín
abandonment layers and the other above this last one (Lumbreras 1989, 1993). The
abandonment layers to which Lumbreras refers are composed by the collapse of the walls of
the Old Temple (layer G) and by the collapse of the plaster and molded friezes that may have
covered the walls (layer H), (Lumbreras 1989). He only excavated 15% of layer H in the
Circular Plaza, the rest of this layer was excavated by John Rick between 1998 and 2004
(Kembel and Rick 2004; Rick 2005).
On the north arc of the Circular Plaza Lumbreras found five vertical engraved slabs
and nine horizontal engraved slabs below the vertical ones. On the vertical slabs, there are
representations of a crowned personage, an individual grasping a strombus, and another
human-like entity holding a hallucinogenic cactus likely to be that known as San Pedro. These
individuals seem to represent a hybridization of felines and human beings following the
conventions of Chavín art. On the horizontal slabs, a set of felines are represented with
strictly zoomorphic features. Given the iconographic nature of the Circular Plaza slabs,
11
Lumbreras thought that they could be dated to phase C in Rowe’s sequence, even suggesting
that the Tello Obelisk was placed in the middle of the circular plaza, surrounded by the
individuals represented on the slabs (Lumbreras 1993).
As mentioned before, Lumbreras also excavated the Ofrendas Gallery, where he
found, in a sealed context, a group of 681 broken, but mostly complete ceramic vessels, 191
human bones, and 171 non-ceramic artifacts (Lumbreras 1993). He identified seven ceramic
styles among the vessels recorded: Dragoniano, Ofrendas, Qotopuquio, Floral, Pucaorqo,
Mosna, Wacheqsa, Raku y Puchka. The first five are labeled as local styles while the
remaining four were identified as foreign. The foreign pottery makes up 27% of the total
sample and three of the styles have been assigned to the North Coast while the other one
(Mosna) to the northern highlands (Lumbreras 1989, 1993).
At the same time Hernán Amat worked in the New Temple excavating the Rocas
Canal (Lumbreras and Amat 1965). Regrettably there are no major publications resulting
from the work of Hernán Amat in Chavín12
. In front of the Black and White Portal he
excavated what Rick et. al. (Rick, et al. 1998) have interpreted to be a small rectangular plaza
built with cut granite and limestone. Under this plaza there is a set of rectangular stone-walled
rooms, though disappointingly there is no information about their contexts. Amat also
excavated at the south of Building E, again with no published references.
Lumbreras’ excavations terminated abruptly at the moment he began excavating the
Caracolas Gallery, located at the opposite side of Circular Plaza from the Ofrendas Gallery.
He found engraved fragments of strombii scattered in the gallery in his very brief excavations
there. The gallery was later fully excavated in 2001 by John Rick with startling results (Rick
2005).
In 1973, Rosa Fung began her work at Chavín de Huántar, excavating in the Loco
galley in Building C, where she found three archaeological layers laying above the floor of its
passages; no other contexts were found. She also excavated at the north edge of the Wacheqsa
Sector. The results of her excavations in this sector will be treated separately. She worked in
Chavín again in 1974, 1975 and 1976, and the results of her excavations in the ceremonial
center and the Wacheqsa sector are still unpublished.
12 He has only published some accounts regarding post-Chavín contexts at Chavín (Amat 2004)
12
1.2.5 A New Turning Point
Since the death of Julio C. Tello, knowledge about Chavín de Huántar had notably
increased. Chavín had an improved architectural sequence, and new ceramic and lithic
chronologies tied to carbon dates were obtained from Lumbreras’ excavations.In 1975
Richard Burger began excavations in the town of San Pedro de Chavín with the purpose of
obtaining absolute and relative chronologies, defining the extent of the ancient settlement and
finding information relating to Chavín’s economy and subsistence (Burger 1998). With these
objectives in mind, Burger excavated 13 units scattered mainly in the modern town of Chavín
de Huántar (A1/2/3, A/4/5/6, B1/2/3/5, B5/6/7, B/8 y B/9), southwest of the monumental core
(D1, D2 y 20-A1), La Banda (4-A1 y 4-B1), west of the monumental core (19-A1) and
northeast of it (E1).
These excavations were supplemented by surface collections and observations of the
profiles exposed by civic works in town. Burger confirmed what Rivero had stated in 1851,
also identifying one area of craft production located in the northern section of town,
“Another activity of the residents of this locus was the production of
bone tools and ornaments. This is amply demonstrated by the cut
bone cylinders and articulated ends of mammal bones as byproducts
of the workshops” (Burger 1984: 227).
He also recorded platforms to the southeast of the monumental core associated with
offerings (Burger 1984, 1992, 1998). Regrettably there is no information about the units he
excavated in La Banda.
A result of his fieldwork is the documentation of the complex and continuous
archaeological occupation around the monumental core within a range of one kilometer,
which not only confirmed what Rivero had already suggested but also added more
information. However, not only are Rivero’s interpretations expanded, but also those of Rowe
and Lumbreras. Burger identified three prehistoric moments, associating every one of these
phases with one of Rowe’s architectural periods. He also established the chronological limits
for Chavín de Huántar between 900 – 200 BC, based on ten radiocarbon dates.
In 1985, Burger proposed the concept of Chavín as synthesis rather than as mother
culture, based on the observations Carlos Williams made in 1980. Williams noticed that in the
monumental core of Chavín de Huántar there were architectural forms that originated between
3000 and 1000 BC (Williams 1980); among these forms are the circular plaza, quadrangular
plaza and stepped platforms. Burger took a step further and indicated that the presence of
these architectural forms was due to the collapse of social formations in the coast that obliged
13
coastal inhabitants to find more suitable lands in the highlands (Burger 1985). In a way what
he proposed was closely related to Larco regarding the nature of the Chavín phenomenon
(Larco 1945, 1948)
Until this time the knowledge about Chavín could be summarized in the following
points:
• Chavín chronology falls within the period of 900 – 200 BC.
• The architectural phases developed by John Rowe correlate one-on-one with the
ceramic phases defined by Richard Burger.
• Chavín is the result of the migration of coastal groups that abandoned their original
settlements due to an environmental stress.
• Chavín thus is the synthesis of older coastal archaeological traditions.
• Chavín reached its peak during 400 – 200 BC, exerting a tremendous influence in the
central Andes at that time.
1.2.6 The Stanford Archaeological Project
After a very long hiatus, research at Chavín resumed with the work of the Stanford
Archaeological Project, directed by John Rick. This work started with the production of a
three-dimensional map that allowed Rick to recognize that the architectural complexity was
more intricate than Rowe and others imagined. In order to add more information to the three-
dimensional model, Rick excavated 10 test pits in the architectural seams that are known in
the facades of Buildings A, B, and C in order to evaluate Rowe’s growth hypothesis. Along
the same lines, Silvia Kembel researched the growth of the monumental core from the inside,
examining the patterns of distribution, and growth of the galleries located inside the different
buildings. The outcome of this architectural research is as follows:
• The monumental core was built in 5 stages, 15 phases and 51 construction events.
• Rowe’s architectural and lithic art sequences, although hinting in some right
directions, are too incomplete and erroneous to remain the underpinning of Chavín
chronology.
• Construction began perhaps as early as 1500 BC, with the ceremonial center ceasing
to function at such around 500-600 BC.
• Thus, Burger’s hypothetical peak of Chavín influence at 400-200 BC does not
correspond to the ritual use of the civic ceremonial center of Chavín.
14
With this, research at Chavín de Huántar came to a new turning point. Excavations
not only in the monumental core but also in the near-periphery added more information
regarding the nature of the activities carried on in Chavín de Huántar. The West Field, La
Banda, the South Area and the Wacheqsa sectors continue to be explored, while the work in
the monumental core focuses on the Rocas canal, Circular Plaza and East Atrium.
Recent work has led Rick to propose the existence of a very well crafted convincing
system developed by the authorities of Chavín de Huántar. This system was materialized in
the external and internal spaces of the civic ceremonial center, which was a kind of stage or
theater meant to impress initiates who sought in Chavín the mechanisms to exert power and
authority, in many cases over remote polities. In order to accomplish success, the authorities
of Chavín perverted shamanistic traditions, using them as a façade in order to give a sense of
continuity from older traditions to Chavín times (Rick 2005, 2006).
1.2.7 Where do we stand now?
Research at Chavín has moved forward a long way since the first accounts written by Spanish
chroniclers, travelers and even archaeologists. It has reached a new level basically due to the
extensive work in the surrounding areas, complemented by ongoing research in the
monumental core. Long-standing Chavín paradigms have been challenged, and this
dissertation only hopes to add more data to the current debate.
1.3 Research Design
1.3.1 Research Process
My research at Chavín de Huántar is focused on the Wacheqsa sector, located
immediately to the north of the monumental core. This area is enclosed by the Mosna and
Wacheqsa rivers which in turn separate the Wacheqsa sector from the ancient domestic
settlement described first by Rivero and then by Burger (Burger 1984; Rivero de Ustariz
1851). This space was actually part of the ceremonial center but lacked any sort of surface
monumental architecture, eluding any major investigation through the research history of
Chavín de Huántar. Nevertheless it has been hypothesized that this area was either the
settlement of workers, craftspeople or even priests that lived there while the ceremonial center
was functioning (Bennett 1944; Burger 1984; Fung 1975, 2006; Lumbreras 1993; Tello 1940,
1960). My research has tried to define the nature of the occupations in this specific area in
relation with the ceremonial center, while also trying to identify the specific roles or roles that
15
this sector had in the prehistory of Chavín de Huántar during the Middle and Late Formative
periods. In this regard, I see the Wacheqsa sector as an active social space defined by activity
areas in which humans socially reproduced themselves, developing actions that would allow
the origination, altering or maintaining of their roles within the social order(s) in which they
participated. In general, social spaces can be places of social change depending on different
elements such as the current material conditions of existence, the economic and/or religious
effectiveness of the social system in which the social agent lives, and the interests of those
who socially reproduce themselves at a given time.
I have investigated the Wacheqsa sector with the premise that it was a social space
that was within the sphere of those who held authority in Chavín de Huántar and who created
different mechanisms of control of the people that lived there. Conversely, the archaeological
record recovered in the Wacheqsa sector equally shed light about luring mechanisms out by
those who held power and authority at Chavín de Huántar developed towards foreign elites
that wanted to participate in the Chavín religious system.
The religious component has been understood as the main character of Andean
societies in general and of Formative societies in particular (Burger 1992; Kaulicke 1994;
Lumbreras 1989; Tello 1923). During the Andean Formative, specifically during the Middle
Formative (1200 – 800 BC), there is a major profusion of ceremonial centers on the Peruvian
North Coast, Central Coast, North Highlands and Central Highlands, where religious systems
regulated the social life of the Andean people. In most cases these centers satisfied a religious
need in exchange for allegiance represented in labor– in the form of goods or manpower-
needed by the religious system to survive and prevail among others. As stated by Burger and
Burger (Burger and Salazar-Burger 1991), ceremonial centers are generators and receivers of
an intense social life that regulate diverse aspects of human nature such as those related to
spiritual satisfaction and the productive apparatus necessary for an adequate functioning of the
social system From a more skeptical view, this basic need for social cohesion (Durkheim
1947) can be the subject of a conscious manipulation by those who want to benefit from the
other segments of the social system (Marx 1973).
Through the use of three dimensional stratigraphic modeling, I have been able to
identify five prehistoric spatial analytical units in the Wacheqsa sector. In order to infer the
activities carried on in these units I have examined the intrasite variation of the archaeological
record, focusing primarily on ceramic distinctions using bivariate kernel density estimations of
diameter and thickness from ceramic rims. Then I associate these results with the distribution
16
of different classes of archaeological materials in each analytical unit and I assign specific
activities and/or functions to each unit. Subsequently, I discuss the nature of the activities
inferred in relation with the ceremonial center. I frame the discussions from two perspectives:
chronological and political. Using a set of ten dates coming from radiocarbon samples
collected from deposits excavated at the Wacheqsa sector, I examine the relationships
between the Wacheqsa sector and the monumental core, putting emphasis on the existing
ceramic and architectural sequences (Burger 1984; Burger 1998; Kembel 2001). Next I
discuss the political implications of the inferred activities in the Wacheqsa sector in relation
with the power strategies exerted by the authorities of Chavín de Huántar. Lastly I elaborate
on the relationships among the ceremonial center of Chavín de Huántar and other ceremonial
centers during the Middle and Late Formative.
The results of these analyses shed light on the ways the Wacheqsa sector was used by
the authorities of Chavín de Huántar, the chronological positioning of the ceremonial center,
the political strategies used by its authorities of Chavín de Huántar for enticing followers into
their system, and the relationships of the ceremonial center with other similar centers during
the Middle and Late Formative periods.
1.3.2 Research Questions
This dissertation addresses the following questions related to the Wacheqsa sector during
the time the ceremonial center of Chavín de Huántar was in function:
• What are the archaeological contexts created in the Wacheqsa sector during the
Formative period?
• Are these contexts isolated or can they be grouped into spatial units based on the
variation of its archaeological components? Do they reflect intrasite variation? And if
so, how do they relate to each other in terms of space and time?
• Based on the intrasite variation of the archaeological record, what were the activities
carried out in the Wacheqsa sector? How did the Wacheqsa sector evolve through
time? Does this change relate to changes in the ceremonial area? How do these
activities inform us about social organization and political strategies within Chavín de
Huántar?
• Do absolute dates from the Wacheqsa sector support the chronological parameter of
the existing ceramic sequence? How do they relate to the architectural sequence of
the ceremonial center?
17
• How do the materials recovered inform us about chronology and regional interaction
between Chavín de Huántar and other ceremonial centers of the Middle and late
Formative periods?
Finally, I evaluate the answers to these questions and address their implications for the
understanding of the social processes that occurred during the Middle and Late Formative
period in the Andes.
18
CHAPTER 2
THE ANDEAN FORMATIVE
In this chapter I intend to summarize of the Andean Formative in order to
contextualize my research in relation to the social processes that characterize this period. I
provide a brief historical overview of the evolution of the Formative concept in the study of
the Andean past, and then give a general updated account of the Andean Formative and its
chronological subdivisions. This chapter itself provides the background for answering the
research questions laid out in the previous chapter but also addresses the following aspects
with a strong emphasis:
• The architectural and ceramics antecedents of Chavín de Huántar during the Early
Formative.
• The contemporaniety of Chavín de Huántar with ceremonial centers from the Middle
and Late Formative.
• The chronological location of the Janabarriu phenomenon.
At the end of this chapter I provide a summary of the state of the art regarding the
aforementioned topics, setting the tone of the discussion for the upcoming chapters.
2.1 The Concept of Formative in Andean Prehistory
In 1919 archaeologist Julio C. Tello first visited the site of Chavín de Huántar and
reported being impressed by its monumentality and the complexity of the lithic art associated
with the ceremonial center (Tello 1929). Tello argued that the site dated from a pre-Inca
“Megalithic Age” and was related to the first of the four ages proposed by chroniclers Guaman
Poma de Ayala and Buenaventura Salinas in 1613 (Poma de Ayala 1992 [1613]; Salinas y
Córdoba 1957 [1630]). According to these chroniclers, Andean history had four ages, as
shown in the following chart:
Epochs Guaman Poma Buenaventura Salinas Auca Runa 1000 BC 0 Purun Runa 2100 BC 1000 BC Wari Runa 3400 BC 1500 BC Wari Wiracocha Runa 4200 BC 2500 BC
Table 1: Chronological chart according to Guaman Poma and Buenaventura Salinas
19
Tello calculated the years of duration of each age in western calendar years based on
the information provided by Guaman Poma and Buenaventura (Tello 1929:19). Inspired by the
aforementioned chroniclers Tello equally proposed 4 epochs of Andean prehistory that he
named first, second, third and fourth, equating the first epoch with Chavín de Huántar (Tello
1929, 1942). Since the work of Tello the study of the chronological aspects of Andean
prehistory have greatly improved either by refining the stylistic study of archaeological
materials associated with solid stratigraphic columns (relative chronology) or by the use of
carbon dates (absolute chronology).
The use of the term Formative began in the decade of the 1940’s (Steward 1948;
Strong 1948), the typical society from this period being one “oriented towards a priest-temple
complex, as evidenced by mounds of somewhat dispersed settlements” (Steward 1948:103).
Evidently at the time the term was defined there was no information regarding the complex
architectural developments of the Late Archaic (Hass and Creamer 2006; Shady, et al. 2001).
An alternative for the use of the term Formative was discussed by John Rowe, who in 1962
suggested the use of the terms Initial Period and Early Horizon (Rowe 1962b). Rowe proposed
a new chronological framework of Andean archaeology based on the chronological sequence
of the Ica Valley. According to him, the Initial Period is formed by the space of time between
the introduction of ceramics in the Ica valley until the beginning of the Chavín influence in
there. This Chavín influence comprises the Early Horizon until the abandonment of post fire
decoration in favor of polychrome slip decoration in the same valley. There are three different
criteria implied in this classification: material (ceramics), stylistic (Chavín style) and
technological (post fire and polychrome slip decoration) (Kaulicke 1994:259). Lately the use
of this chronological framework has been complicated by the addition of the term “Chavín
Horizon”,
“The Chavín horizon style is presumed to have begun during the
final epochs of the Initial Period and continued during the first five
or six epochs of the Early Horizon if we follow the Ocucaje
sequence” (Burger 1993:54).
The use of terms likes Early Horizon and Chavín Horizon together is complicated and
cumbersome when considering that the Early Horizon is defined on the grounds of a not very
well-known master sequence and that the Chavín Horizon has been defined on the grounds of
incorrect readings of carbon dates and conflicting stratigraphy, as I will explain later in this
chapter. Nevertheless, the term Formative is still widely accepted and extensively used in
Andean archaeology. It seems to be the most familiar for archaeologists investigating the time
20
framework of 1800 – 200 BC (Kaulicke 1994; Lumbreras 1989), hence that’s the term that
will be used in this dissertation. The Formative period in the Central Andes encompasses
approximately 1600 years and given the variability of the archaeological record and the way it
changes through time, it is necessary to subdivide such a lengthy period of time. In this sense I
use the following chronological subdivisions inspired by the works of Kaulicke (Kaulicke
1994) and Lumbreras (Lumbreras 1989),
Years (B.C.) Period Definition Ceramic Styles 500-0 Final Formative Post Chavín ceramic
styles Late Janabarriu?, White-on-Red
900-500 Late Formative Black and White Stage at Chavín de Huántar,
particularly the styles
related to the Janabarriu
phase.
Janabarriu, Late Cupisnique
1200-900 Middle Formative Pre Black and White stages of Chavín de Huántar
Kotosh-Kotosh,
Urabarriu, Idolo,
Early and Middle
Cupisnique 1800-1200 Early Formative Pre Chavín Pandanche A,
Kotosh Wairajirka,
Chira, Haldas,
Sechín
Table 2: Proposed chronological chart of the Andean Formative
This chart uses the subdivisions proposed by Lumbreras and Kaulicke but centers the
chronological sections on Chavín de Huántar. Even though this emphasis on Chavín can be
criticized. I argue that centering the chronological sections on Chavín will help to understand
the processes and regional interactions to be discussed in the following pages. Although the
Formative period is a time of great complexity and it is difficult to simply summarize it in a
few pages, centering the discussion on Chavín will improve the comprehension of the role of
Chavín in the Andean Formative and will provide a good model for the role of ceremonial
centers during the Formative, especially during the Middle and Late Formative periods.
2.1.1 Early Formative (1800-1200 B.C.)
The chronological marker for the beginning of the Formative period has been the
introduction or/and invention of ceramic and metal technologies, as well as the appearance of
textiles made with looms (Bonavia 1991; Burger 1992; Burger and Gordon 1998; Morales
1993; Rosas 1970). Early ceramics appear in the site of Pandanche (Kaulicke 1975, 1994)
21
showing resemblances with the ceramics from Ecuador, specially with the Early Machalilla 8
and Valdivia 8 phases (Burger 1992), Toril (Burger and Salazar 1985), Kotosh (Izumi and
Sono 1963), Yarinacocha (Lathrap 1960b), Chira Villa (Lanning 1953), Ancón (Rosas 1970),
La Pampa (Terada 1979) and La Florida (Patterson 1985). `
These early ceramic developments can be segregated into two trends: elaborate and
rudimentary. Early ceramics in the sites of Pandanche, Tutishcainyo, and Kotosh present
complex forms and elaborate decoration while ceramics from the sites of Guañape, Ancón,
Toril, La Pampa and Chira Villa are characterized by their simple and limited array of forms
and plain decoration. Early complex ceramics in areas where there are no signs of previous
ceramic technological experimentation may indicate that this technology was brought from
other areas or at least give insights into regional contacts. In this vein, Donald Lathrap
(Lathrap 1960b) has argued that at around 2000 B.C. people from the central Ucayali drainage
had contacts with the Huánuco area, with the same decoration being found on Early
Tutishcainyo at Yarinacocha and Kotosh-Waira-Jirca pottery at the site of Kotosh. He has
traced the changes in these pottery assemblages and found that they occur at the same time in
both areas, suggesting a continuing cultural contact. Along the same lines he has also argued
that the Early Tutishcainyo ceramics resemble ceramics from the Colombian tropical forest
regions (Lathrap 1960a, b) which would explain the form and iconographic assemblage of this
phase. In the case of the North Highlands, ceramics from Pacopampa and Huacaloma show
similarities to Early Pandanche ceramics that are related to Ecuadorian traits. On the other
hand, there seems to be evidence of localized initial ceramic development in the sites of
Ancón, Chira Villa, Toril, La Pampa and Guañape (Burger and Salazar 1980; Morales 1993;
Rosas 1970; Strong and Evans 1952; Terada 1979).
Unlike ceramics, monumental architecture is present in the Central Andes before the
Formative Period in the Late Archaic period (3000 B.C. – 1800 B.C) in which large
ceremonial centers were part of the cultural landscape of coastal valleys, especially in the
region between the north central coastal valleys of Supe and Fortaleza (Haas, et al. 2004; Hass
and Creamer 2004, 2006; Shady 1997, 2004; Shady, et al. 2001; Shady and Leyva 2003;
Vega-Centeno 2007; Williams 1980) and at a minor scale in the highlands (Bonnier 1983,
1997; Burger and Salazar 1980, 1985; Izumi and Sono 1963). On the coast the following
architectural features were the basic units of any sort of construction: platforms, mounds, and
circular plazas during the Late Archaic (Williams 1980) while in the highlands, rectangular
buildings with central hearths were the most recurrent unit of monumental architecture. These
22
rectangular buildings were relatively small but the constant closure of old structures and
construction of new ones created large mounds that in some cases reached 12 m height (Izumi
and Sono 1963). This architectural feature has also been described as the Mito Architectural
Tradition (MAT) (Bonnier 1997) which was part of what Richard Burger has called the
Kotosh Religious Tradition (KRT) (Burger and Salazar 1985). These rectangular buildings are
also present in the central coast and have been reported in the valley of Supe and the coast of
Ancash (Pozorski and Pozorski 1987; Shady and Leyva 2003).
During the Initial Formative, the basic architectural units of the Late Archaic were
rearranged creating new architectural patterns of monumental architecture. For example, in the
central coast the U-shaped building architectural tradition dominated the landscape of the
valleys of Lurín, Rímac, Chillón and Chancay (Burger and Salazar-Burger 1991; Ravines and
Isbell 1975; Silva and García 1997; Williams 1980). This tradition was characterized by a
terraced central mound flanked by two long platforms enclosing a large quadrangular plaza, in
some cases with a small vestibule in front of the central mound. Additionally, circular sunken
plazas were located either in the quadrangular plaza or to the sides of the flanking platforms
(Scheele 1970; Williams 1980). The scale of the these buildings was massive, for example the
Huaca La Florida,
“represents an investment of at least 6.7 million persons-days, and
this excludes the labor needed to level the area and to plaster and/or
decorate the outer surfaces of the buildings. La Florida is by no
means the largest of the U-shaped pyramidal mounds. San Jacinto in
the Chancay valley, is four times larger, and it would have required
almost 2 million cubic meters of materials just to level the 30 ha
plaza” (Burger 1992:61)
In the Lurín valley, U-shaped complexes tended to be present in closely spaced pairs:
La Candela, Buenavista; Mina Perdida, Parka; Cardal, Manchay Bajo; Piedra Liza, Anchucaya
(Mesia 2000). In the Rímac valley ceremonial centers are mostly located at the north of side of
the river and not as closed as the ones in the Lurín Valley (Silva and García 1997). This
architectural tradition persists during the Middle Formative in which iconographic designs
related to the iconography represented in ceramics found at Chavín de Huántar adorn the
facades of the central mound.
In the Casma valley, a set of ceremonial centers was constructed following a different
pattern than the one presented in the U-shaped ceremonial buildings of the central Peruvian
coast. The sites of Sechín Alto, Taukachi-Konkan and Pampa de las Llamas share an
architectural pattern of one longitudinal axis with a high mound along the ends of the axis and
23
a set of superimposed platforms with large rectangular plazas in front of the main mound with
an array of symmetrically positioned rooms interpreted as domestic units positioned along the
superimposed platforms (Pozorski and Pozorski 1998). There are examples of mural painting
in the Casma valley, for example, the site of Moxeque, displayed on its façade molded friezes
2.5 m high. The friezes flanked the front staircase that gave access to the summit of the
mound. They represented human images wearing tunics, small skirts and unfastened mantles,
and one of the images is holding in both hands bicephallic snakes with forked tongues (Tello
1956). The Casma valley during the Early Formative must have been similar to what it was in
the Supe valley during the Late Archaic, with a large number of massive ceremonial centers
scattered in a relatively small area.
On the north coast in the middle Jequetepeque Valley, the site of Montegrande is one
of the most important ceremonial centers of the Early Formative in the northern Andes. It is
composed of platforms interconnected by stairways, large buildings, a sunken quadrangular
plaza, and rectangular domestic units surrounding the complex. The ceramics found are
similar to the ones from Pandanche and Early Huacaloma (Kaulicke 1975; Tellenbach 1986;
Terada and Onuki 1982).
The Early Formative is characterized by the continuation of construction practices
developed during the Late Archaic. There is a strong emphasis on monumental architecture
with large plazas and decorated facades. Public architecture seems to be the most accurate
definition as it provided spaces necessary for public ceremonies of religious nature (given the
type of iconography depicted). The religious nature of the ceremonies could have been just a
filter that encompassed the economic and political aspects of these societies. This trend is
maintained during the Middle and Late Formative, and as we will see later, it practically
defined the ceremonial aspects of Chavín de Huántar.
2.1.2 Middle Formative (1200-800 BC)
The Middle Formative is characterized by the construction of the ceremonial center of
Chavín de Huántar and its associated iconography as well as the surfacing of a set of coastal
sites that have been previously identified as Cupisnique. The chronological start of this period
should be readjusted according to new dates available. Cupisnique was recognized for many
years as a coastal manifestation of Chavín de Huántar. Julio C. Tello believed that early social
development on the north coast was based on a Chavin expansion. He called all the material
elements that showed some resemblance with the expressions found in Chavín de Huántar
24
“Chavin costeño13
” (Tello 1960). These ideas were challenged by Rafael Larco who, based on
his excavations in the Cupisnique and Chicama Valleys, proposed that the Cupisnique term
should be applied to early developments in Chicama and the immediately surrounding area
(Larco 1945). Unlike Julio C. Tello, Larco argued for the coastal development of what had
been understood as coastal Chavín derived from the sierra.
“if we were to analyze carefully the different cultures that have been claimed to be included within the so- called Chavín Civilization, we would reach the conclusion that, although they have cultural elements in common, they have others in greater quality that allows
us to distinguish one culture from another” (Larco 1948:16)14
Instead of looking at Cupisnique as a consequence of highland migration, Larco saw
the opposite, the emergence of Chavín as a consequence of Cupisnique influence. Critical
review of Cupisnique’s absolute dates (Bischof 1998), especially the ones coming from the
Cupisnique site of Huaca de los Reyes (Pozorski 1975) argues for a contemporary
development of Cupisnique sites and Chavín de Huántar. Further, the analysis of Huaca de los
Reyes architecture made by William Conkin indicates that instead of two architectural
sequences, the site encompassed eight phases, with the iconography depicted in all of them
related to the Middle Formative (Conklin 1985). Huaca de los Reyes is part of the Caballo
Muerto complex that encompassed seven more mounds (Pozorski 1975), and it appears to be
the pre-eminent Cupisnique center, as it was,
“one of the main inter-regional centers or the main center in the
Cupisnique heartland. It seems probable that this represented an
initial centralization of Classic Cupisnique culture under a religious
hierarchy or authority. Maybe this pattern was also common in the
neighboring valleys.” (Elera 1998:276)
Other impressive examples of Cupisnique architecture can be found in the sites of
Poro Poro, Purulén, Limoncarro, Huaca Lucia and Huaca el Gallo (Alva 1988a, b; Barreto
1984; Shimada, et al. 1983; Zoubek 1997). In general, these ceremonial centers were built on
low-tiered platforms, with massive central inset stairways leading to rectangular forecourts,
and with the presence of colonnades, which are distinctive elements of early North Coast
monumental architecture (Burger 1992).
This local tradition of material culture may have been restricted to the north coast,
where centers were sharing material and ideological elements but maintaining their political
13 Coastal Chavín. 14
My translation
25
independence. This may have been started during the Middle Formative, when the first
material elements of the Cupisnique tradition were found. Elera (Elera 1998) suggests that
during the early development of Cupisnique, populations surrounding ceremonial centers
probably consisted of autonomous political units. As complexity emerged on the north coast
by the end of the Middle Formative, the area was full of small, powerful, and rich polities. The
end of this social tradition is not clear yet. According to Elera (Bird 1987; Elera 1998;
Inokuchi 1998; Onuki 2001), Cupisnique centers were abandoned as a result of a mega El
Niño event:
“[In the Cupisnique sequence] there is a clear cultural continuity
from the late Preceramic to the Middle Formative periods, which
ended abruptly as a result of a natural catastrophe that forced the
abandonment of coastal settlements”. (Elera 1998:257)
In the north coast sites of Huaca Negra, Huaca Prieta, Temple of the Llamas, Puemape
and others from the Moche Valley, there is evidence of abandonment, associated with the lack
of mussel shell in the archaeological record. From the Late Archaic through the Middle
Formative, coastal populations sought species from the cold Peruvian current, which
according to Carlos Elera and Robert Mc Bird is reflected in the archaeological record (Bird
1987; Elera 1998). These species include fish, birds, and sea mammals and the disappearance
of some of these in the record can be interpreted as evidence of an El Niño event: “One of the
consequences of this natural disaster was the almost complete elimination of mollusks adapted
to the typical cold Peruvian waters” (Elera 1998: 274). Bird elaborated on the possible date of
this event, suggesting that it happened around 800 B.C; he based his assumptions on
radiocarbon dates from Las Haldas, Huaca Prieta, and Huaca de los Reyes which correlate
with evidence of sudden abandonment and haphazard resettlement on the North Coast (Bird
1987). Bird recognizes the problems that radiocarbon dating has in this particular time “it must
be remembered that all radiocarbon dates relating to the Early Horizon should be treated with
caution because there are two anomalies in the long term shift in 14C levels” (Bird 1987:286).
The presence of an El Niño on the north coast seems probable but the dating of this
phenomenon has to be considered with extreme caution. According to Nials et al, this
phenomenon occurred at around 500 B.C., reflecting tensions between this and Bird’s
interpretations (Neils, et al. 1979a, b).
Currently it is common to refer to nearly all the materials of the Middle Formative
from the Virú Valley to the Lambayeque drainage as “Cupisnique”. The populations from
these valleys shared a host of material elements whose traits are limited to the north coast such
26
as stirrup bottles which appear for the first time in the north coast at around 1500 B.C (Elera
1998). Even though the understanding of the social processes that occurred on the north coast
is still in an initial stage because the lack of data, it can be said that around 1200 B.C. the
north coast shared a common set of material manifestations, known as Cupisnique, that were
not present in other sites of the central Andes.
Parallel to this north coast development, on the central coast the Early Formative U-
shaped building tradition persisted and new U-Shaped buildings were being constructed.
Among this new set of sites are Cardal and Manchay Bajo in the Lurín Valley (Burger and
Gordon 1998). The site of Cardal had at least three superimposed atriums in the central
mound; the earliest one could not be excavated and the later one was damaged by the time
archaeological work began at the site. Excavations in the middle atrium presented impressive
molded friezes, the iconography adorning its façade represented
“A mouth band of interlocking triangular teeth and massive upper
fangs. A lower horizontal band, painted red and probably
representing the lower lip, runs below the teeth, and a parallel upper
lip once existed above them, judging from a few poorly preserved
fragments in the western part of the landing” (Burger and Salazar-
Burger 1991:283).
The two mouths depicted flanked the entrance of the atrium. The iconography
described above suggests resemblance to the mouths engraved in the central sections of the
two principal mythic animals depicted in the Tello Obelisk found at Chavín de Huántar.
Richard Burger indicates that the Middle Atrium, and consequently, its mural decoration may
have been completed around 970 B.C. (Burger and Salazar-Burger 1991). This date is roughly
contemporary to the beginnings of the Black and White phase at Chavín de Huántar (Kembel
2001; Rick 2005). Cardal is not the only U-shaped building that shows Chavín related
iconography on its facades; another example is found in the archaeological record of Garagay.
Garagay is located in the lower Rímac valley and was first excavated by Manuel de Ontaneda
and Aquiles Ralli from Peru’s National Museum of Archaeology in 1959 and then by William
Isbell and Roger Ravines in 1974 (Ravines and Isbell 1975). Central Mound B had a height of
23 m while the mounds that flank it (mounds A and C) had 6 and 9 m heights respectively.
Two superimposed atriums were present in the central mound, the latest one partially
excavated by Ontaneda and Ralli and the middle Atrium excavated by Isbell and Ravines.
Regrettably the latest atrium was destroyed before it could be completely excavated.
27
Excavations at the Middle Atrium revealed that a set of colored friezes molded in low
relief decorated its walls, the images represented showing strong resemblances to the
supernatural beings depicted on ceramics belonging to the Dragoniano style identified in
Chavín de Huántar (Lumbreras 1989, 1993). These images are anthropomorphic faces with
feline mouths with three fangs separated by geometric panels. Interestingly enough,
Lumbreras has argued that the supernatural being of three fangs represented in Dragoniano
ceramics is actually the female version of what Tello called the Chavín Dragon (Lumbreras
1993; Tello 1942). Hence the relation with Chavín might be more evident than hypothesized
by Burger or Ravines who have recognized Garagay as a pre-Chavín site (Burger 1981;
Burger 1992; Ravines, et al. 1982; Ravines and Isbell 1975).
Ceramic analysis from Garagay has recognized six ceramic wares, two of them
according to Ravines related to the diffusion of the Chavín style (Janabarriu), characterized by
the presence of rocker stamping, impressed circles with and without dots (Ravines, et al.
1982). Ravines also identified four ceramic stages at Garagay but he only provides
chronological control for the latest one – which is composed by two Janabarriu like wares –
indicating that “the chronological position of this phase is uncertain; however, a carbon date
associated with fills that contained ceramics associated with this phase gave a date of 780 BC”
(Ravines, et al. 1982:227).
Interestingly enough, Janabarriu related ceramics at Garagay are dated before the 400-
200 BC mark stated by Burger as the dispersion of Janabarriu-like ceramics in the Central
Andes. When calibrated, the two sigma range of this date is 1132 – 761 BC (figure 05).
Ravines does provide three other carbon dates but there is no indication of their proveniences
or associated ceramics, rendering them of little utility other than suggesting that Garagay had
architectural stages that go back to the Early Formative (figure 06).
The first two ceramic phases of Garagay are related to central coast Early Formative
ceramic styles such as Curayacu A (Lanning 1953; Ravines, et al. 1982) and the Florida and
Hacha phases from Ancón (Rosas 1970). The ceramic assemblage at Garagay is consistent
with what happened at Ancon, where local ceramics are replaced by Janabarriu-like ceramics,
or as Rosas calls them, Chavinoid ceramics (Rosas 1970). Unfortunately Ravines has not
indicated which ceramic assemblage was associated with the Dragoniano friezes uncovered in
the Middle Atrium but its close iconographic similarities with Dragoniano ceramics at Chavín
suggests at least its contemporaniety with the Black and White stage at Chavín de Huántar, if
not earlier. Lumbreras has also pointed out the contemporary relationship between Garagay
28
and Chavín, indicating that Garagay is not that pre-Chavín and that at least it was
contemporary with the deposition of the offerings at the Ofrendas Gallery (Lumbreras
1989:107).
Nevertheless, the U-shaped building tradition continues in the Central Coast during
the Middle Formative but it is associated with Chavín-related iconography and in some cases
like Garagay, going further into the Late Formative, associated with Janabarriu-like ceramics.
Also, excavations in Ancon have shown the replacement of Early Formative ceramic
assemblages with Chavín related ceramics (Lumbreras 1989)
In the north highlands the situation is different. There is also monumental architecture
in the sites of Pacopampa, Kuntur Wasi and Huaca Loma. It is not as massive in volume but
certainly involved major energy expenditure, especially in working with clean or cut stone.
For example, in the site of Kuntur Wasi, during the Middle Formative, the ceremonial center
consisted of two rectangular platforms and a sunken rectangular plaza constructed on top of
sterile soil (Onuki 1995). The ceramics associated with these structures are very similar to the
ones in the Late Huaca Loma phase at Huaca Loma and in the Pacopampa I phase at
Pacopampa (Inokuchi 1998; Seki 1998; Seki, et al. 2006), which are derived from the north
highlands ceramic styles of the Early Formative. At the site of Pacopampa the situation is very
similar, with only a platform identified as a Middle Formative architectural component (Seki,
et al. 2006). There is a sense of stability in the ceramics and architecture of the north highlands
during the Middle Formative. Ceramics maintain their close relationship with the Ecuadorian
area while monumental architecture remains stable, with no major change within the Middle
Formative. This situation will drastically change during the Late Formative where there are
major changes in architectural design associated with the introduction of Cupisnique-related
ceramics and Chavín-related ceramics in the area.
2.1.3 Late Formative (800-500 BC)
The Late Formative period is the time of the Black and White stage at Chavín de
Huántar and the appearance of Janabarriu-related ceramics in the Central Andes (figure 07).
According to Richard Burger, the Janabarriu phase at Chavín de Huántar reflected the time
when the ceremonial center reached its maximum development, being surrounded by a “proto
city” that had 42 ha at that time (Burger 1984; Burger 1998). Following Burger, the
ceremonial center attained such prestige during these years that people from different areas
were going there, leaving offerings and transporting materials (or ideas) from Chavín to their
29
own places. Burger’s perspective would thus explain the presence of Janabarriu diagnostic
elements in areas like Kuntur Wasi, Paracas, Ayacucho, Lima, etc. (Burger 1993). As we have
seen, at Chavin de Huantar construction of the ceremonial center may have begun during the
Middle Formative (Kembel 2001; Rick 2005) but its apogee and subsequent decline happened
during the Late Formative (Kembel 2001; Rick, et al. 1998). As I will show in the following
pages, the Black and White phase can be equated with Burger’s Janabarriu phase.
Richard Burger defined Janabarriu as the last Formative ceramic phase at Chavín de
Huántar indicating that the
“Janabarriu materials make up the majority of our excavated and
collected sample from the valley floor. The richness of the sample
provides a comparatively complete glimpse of a large inventory of forms and decoration, which dwarfs the two previous phases
[Urabarriu and Chakinani] in sheer variety” (Burger 1984).
This phase was recorded in five excavated units, but only one of them was dated –
unit D1, located 100 meters south of the ceremonial center. This date, ISGS-506, comes from
a layer on top of a floor associated with a wall, which is covered by a platform that contains
Janabarriu ceramics as part of its refuse and in the mortar of its walls. Above this platform an
offering of guinea pigs and Spondylus was placed. There were other two dates, UCR-748 and
UCR-747 taken from units D2 and E1 which Burger considered too young to be acceptable.
This date was also rejected by Burger as,
“The resulting measurement of 2520+100 radiocarbon years: 570
BC (ISGS-506) for the Janabarriu carbon sample conflicts with the
internal stratigraphy of the excavation from which it is taken, as well as being at odds with the estimate for the Janabarriu phase” (Burger
1981:596)
The date of ISGS-506 resulted earlier than the samples taken from lower levels
assigned to Chakinani ceramics in the same unit. Burger not only rejected dates UCR-748 and
UCR-747 but also rejected date ISGS-506, as it was conflicting with the stratigraphy of the
unit (figure 08). The question at hand is how was this phase dated? He used the Chakinani
dates and a date from the immediate post Chavín occupation (Huarás) in order to create early
and late ends of the Janabarriu phase. The early end would be the late end for the Chakinani
phase and the late end of the Janabarriu phase would be the early end of the Huarás
occupation. The Huarás date he used was GIF-1079 which has one sigma of 383 BC-70 AD,
but two sigmas of 209 BC – 3 AD). When looking at figure 09, one can not help but to think
that the one sigmas value was the one Burger used to squeeze in the Janabarriu date.
30
In observing the absolute dated sequences of Kotosh, La Pampa, Kuntur Wasi,
Garagay and the relative dated sequences of Ancon, it can be noted that Janabarriu equivalent
materials do not date as late as Richard Burger has suggested, and in this situation Bischof’s
statement is compelling: “Late Chavín, meaning Janabarriu, could have ended towards 500
and even 600 cal BC, way before the III century as proposed by Burger” (Bischof 1998:68)
The implications of a relocation of Janabarriu as an early phenomenon does not only
have to do with Chavín de Huántar itself but also with regional processes of social complexity
for the Andean Formative. For example, let’s now look at the site of Kotosh, specifically the
Kotosh-Chavín phase, which has two carbon dates, GaK-263 and N-65-2. In terms of ceramic
elements, the Kotosh-Chavín phase is characterized by the same elements that define the
Janabarriu phase at Chavín de Huántar,
“Along with plain rocker stamping, dentate rocker stamping appears.
Stamping designs in circle-and-dot, double circles, S-shapes, hook
shapes, etc, are very popular” (Izumi and Sono 1963).
The dating of this phase is problematic as it falls within the same time framework
established for the Kotosh-Kotosh phase which unfortunately has sigmas that are separated in
some cases by 600 years, GaK-150, N-66-a and N-67-2 (figure 10).
The ceramics from the Kotosh-Kotosh phase have been identified as related or
predecessor to the ceramic style known as Urabarriu at Chavín de Huántar, located
stratigraphically under Janabarriu deposits (Burger 1984; Burger 1998; Lumbreras 1993). It
can be argued that the superimposition shown in Chavín is also represented in the Kotosh site
with Urabarriu-like materials under Janabarriu-like materials. Following this line of thought it
can be argued that the ceramics from the Kotosh Chavín phase are in fact Janabarriu materials,
as the ceramics from this phase,
“… claim very close kinship with the pottery from Chavín de
Huántar and coastal Chavín sites. It is true that the preceding
Kotosh-Kotosh period shows a marked Chavinoid impact. However
when viewed in a wider context, we are inclined to think that the
Kotosh Well Polished [Kotosh-Chavín] type stands alone in the
Kotosh sequence; it shows little relationship with other Kotosh
pottery types but rather more affinities with other Chavín sites”
(Izumi and Sono 1963:156)
The dates for both phases have the problem of the broadness of their sigmas, but still
it is quite suggestive how the dates for the Kotosh-Chavín phase – even with that broadness
31
stated above – are not in the chronological framework stated by Richard Burger of 390-200
BC (Burger 1984:277) but rather in a much earlier one.
Another example that needs to be considered is the site of La Pampa, located at the
department of Ancash, province of Corongo at 1800 m.a.s.l. It is composed of nine semi-
artificial mounds with an estimated size of 100 ha (Terada 1979:1). The cultural history of the
site is divided in four phases, Yesopampa, La Pampa, Tornopampa and Caserones (Terada
1979). Janabarriu like ceramics appear in La Pampa during the La Pampa phase. There are
two radiocarbon dates from this phase TK 176 and TK 195; these dates fall within the Late
Formative (figure 11) and “coincide perfectly with the estimated age of this period which is
related to the Chavín culture as was evidenced by pottery types”. (Terada 1979: 177.
Emphasis added). During the La Pampa phase there were substantial changes in architecture,
including changes in construction techniques such as the alteration of the orientation of
platforms and the types of stones used in walls, which has led Terada to argue that there was a
strong impact coming from outside (Terada 1979).
At Kuntur Wasi, ceramics that can be identified as Janabarriu appear for the first time
during the Kuntur Wasi phase. The presence of Janabarriu ceramics together with a major
investment in architecture defines the Kuntur Wasi phase. There are five radiocarbon dates
coming from the Kuntur Wasi phase, TK 913, TK 908, TK 912, TK 909 and TK 910.
According to Onuki the Kuntur Wasi phase dates from 950 BC to 540 BC (Onuki
1995), but if we look at fig 12, we will realize that it would make more sense to locate it at the
range between 800-400 BC. Either way the presence of Janabarriu-like materials in Kuntur
Wasi, with the data available, started at least 600 years earlier than what Burger stated for the
Janabarriu phase at Chavín de Huántar.
“If the Janabarriu phase was situated between 390 and 200 BC, it
would be contemporary with the Copa phase at Kuntur Wasi, but, the
characteristics that are representative of the Janabarriu phase are
present in the Kuntur Wasi phase and the Sangal complex, which are
chronologically earlier than the Janabarriu phase at Chavín de
Huántar (Onuki 1995:270)
This introduction of Janabarriu-related ceramics in the north highlands is associated
with a major transformation of architectural design in the area. The Kuntur Wasi phase at
Kuntur Wasi is not only characterized by the presence of this ceramic complex but by the
construction of an impressive ceremonial center composed of imposing platforms and plazas.
32
This architectural project marks a radical transformation compared with the two-platform
structure of the previous Idolo phase. The same situation happens at Pacopampa where
“most of the visible architecture is related with the Pacopampa II
phase: the quadrangular plaza, the retaining wall of the third platform
and most likely, the rest of the retaining walls of all platforms that
surround the upper platform and the rest of the minor structures
(rooms, courtyards)” (Seki, et al. 2006:17)
Seki et al. place the Pacopampa II phase within the range of 850 – 585 BC, which is
consistent with the argument stated in the previous pages. The ceramics associated with this
phase are Janabarriu-related with the presence of impressed circles with and without dots. The
Late Formative can be equated to a distribution of Janabarriu related ceramics in the Central
Andes.
2.1.4 Final Formative (500 – 50 BC)
The final Formative is a post-Chavín/Cupisnique world in the Central Andes. At
Chavín de Huántar, the ceremonial center ceases to function as such and is inhabited by
squatters living in the circular plaza associated with a post-Chavín ceramic style locally
defined as Huarás that appears extensively in the Ancash region (Amat 2004; Bennett 1943;
Lau 2002; Lumbreras 1993). However, layer H in the circular plaza excavated by Lumbreras
may suggest the possibility of a Late Janabarriu phase associated with the collapse of the
ceremonial center.
“We suppose that layer H represent a period in which the site was
abandoned […] period in which the plaster of the walls started to fall
down and people threw away food waste and broken ceramics. We
propose the hypothesis that this trash is contemporary with the latest
occupation of the Chavín” (Lumbreras 1989:147).
The ceramics associated to this layer are decorated bowls with stamped horizontal
“S”, stamped circles, stamped concentric circles as well as vessels with red slip and broad
incised lines with graphite in the incisions (Lumbreras 1989). Further studies are needed in
order to differentiate the Janabarriu components contemporary with the use of Chavín as a
ceremonial center and those components that are contemporary with the collapse and
abandonment of the ceremonial use of the site. Nonetheless, after this layer H is formed in the
circular plaza, this area is occupied by the squatters mentioned above, associated with the
Huarás ceramic assemblage which is very different from Chavín-related ceramics,
33
“As a horizontal influence it pertains to ceramics and is characterized
by the use of white-painted decorations on a natural red or red-
slipped ground color. Another decorative technique, the use of thin
incised lines to outline the painted areas, is often associated. The
simplicity of the geometric design elements and certain vessel forms
are also a means of identifying the White-on-Red horizon and its
various contexts” (Wiley 1948:10)
At a regional scale, this distinctive style has been defined as White-on-Red, having
been found in the sites of Kotosh, San Blas, Puerto Moorin, Baños de Boza, Chonta Ranra
Punta, Cerro Trinidad, Pashash, La Pampa, Huaricoto, Kuntur Wasi and Salinar (Burger and
Salazar 1985; Inokuchi 1998; Izumi and Sono 1963; Morales 1993; Strong and Evans 1952;
Terada 1979; Wiley 1948). Unlike the Late Formative in which Janabarriu related ceramics
can be associated to the florescence of Chavín de Huántar, the distribution of white-on-red
ceramics over the Andes does not seem to be related to a particular ceremonial center.
According to Wiley the presence of this style could have been linked to a technological shift
related to the open kiln firing of ceramics rather than keeping the reduction-firing tradition
that was pervasive over the Middle and Late Formative (Wiley 1948).
In assessing the absolute dating of the red-on-white style in the Ancash region, Lau
uses four dates: Gif-1079 (Chavín de Huántar), AA-32484 (Chonta Ranra), Beta-31354
(Queyash Alto) and Beta-31357 (Queyash Alto) (Lau 2002:183). Interestingly enough, these
dates better support an early date for the white-on-red style in the Ancash region (figure 13).
In Kuntur Wasi after the Kuntur Wasi phase, the Copa phase is dominated by a
ceramic assemblage that is predominantly white-on-red that co-exists with ceramics that are
reminiscent of the Kuntur Wasi phase and consequently to the Janabarriu-related styles, being
characterized by a great variety of designs that combine straight horizontal, vertical and
curvilinear lines, triangles, stairs, rectangles and concentric circles and circles with dots. This
phase is dated from 450 to 250 BC (Inokuchi 1998; Onuki 1995).
When considered along with the previous review of the Janabarriu phase, an early
date of the white-over-red style seems logical, and is well supported by the dates available.
Thus the red-over-white style at the Ancash region may have started around 500 BC and
continued until the appearance of the styles associated with Recuay towards the beginnings of
the modern era. Janabarriu-related styles may have not been completely replaced by white-
over-red ceramics as shown in Kuntur Wasi and probably Chavin, but were drastically
reduced in the archaeological record, favoring the new white-on-red style.
34
The social landscape of the Central Andes drastically changed during the Final
Formative, “the pattern of large civic-ceremonial centers in the North Highlands appears to be
replaced by a more fragmented social landscape based on small communities and territories”
(Lau 2002). This transformation set the pace for the development of regional social processes
characterized by state level societies during the first 600 years AD.
This summary was intended to provide a road map towards the following points:
• Chavín de Huántar is earlier than previously suggested.
• The Janabarriu phase at Chavín de Huántar is at least 600 years earlier than suggested.
• Cupisnique and central coast U-shaped buildings from the Middle Formative are not
Pre-Chavín but contemporary to the ceremonial center.
• There was a complex network of ceremonial centers interacting among themselves
during the Middle Formative and especially during the Late Formative that included
Coastal sites as well as Highland sites
These points are implicit just by doing a critical revision of the literature published
and examining carbon dates and their respective ceramic associations. In the upcoming
chapters I intend to make these points rather explicit using the data retrieved from my research
at the Wacheqsa sector of Chavín de Huántar.
Nevertheless, what I have presented in the previous pages is intended as a summary of
the Formative period. Certainly, many details could not have been covered given the space
constraints of the present dissertation but at least I hope to have covered the basic elements
that gave form and shape to the Formative, mixed with my personal observations on the issues
discussed above.
35
CHAPTER 3
THE WACHEQSA SECTOR AT CHAVÍN DE HUÁNTAR
In this chapter I intend to give a specific background regarding the Wacheqsa sector
and the archaeological interventions previous my own research. The Wacheqsa sector is
located immediately to the north of the monumental core, enclosed by the Wacheqsa River to
the north and the Mosna River to the east (figure 14). Tello describes it as “a trapezoidal field,
60 m long and 50 m wide looking towards the Wacheqsa river and 30 m from Mound D15
”
(Tello 1960:317). Spatially, this sector is located between the monumental core and the
domestic settlement that stretched north into the land occupied now by the modern town of
San Pedro de Chavín. It encompasses an area of 1.4 ha; it has an overall slope of 15.85°
downwards towards north and a difference of eight meters between its south and north ends.
The modern topography of this sector is the outcome of the 1945 landslide which practically
changed the topography of the entire sector (figure 15). This statement is supported by Tello’s
field notes, pre-landslide pictures and my own excavations. In his unpublished field notes
Tello stated that there were at least three major platforms in 1940, covered by small
agricultural fields where farmers were raising corn, “passing the north side building [he may
mean either Building C or D], there is a land divided among two or three platforms, that are
used as farming lands16
” (Tello 1940:25).
He went further and stated that “this land is only flat in some portions, in others it has
a marked step towards the river” (Tello 1940:27). In the map that is provided in his
posthumous 1960 publication “Chavín Cultura Matríz”, Tello showed two terraces and a
probable one extending to the north (Tello 1960:49) while today no terrace is observed in this
sector unless we count the one that serves as the south boundary of it. The land was heavily
used and occupied in 1940 as can be observed in figure 16 in which a pirka serves as a
delimiting element between two agricultural fields. Also, there are two eucalyptus trees in the
figure. Today eucalyptus trees are only located on the shores of the Wacheqsa and Mosna
rivers and not in the middle of the sector. Figure 17 gives a broader portrait of the pre–
landslide terracing in the Wacheqsa Sector as well as the presence of eucalyptus trees. Figures
18 and 19 equally show the terracing as well as the houses of the owners of the agricultural
land or chacras, and figures 20 and 21 show the prehistoric retention walls constructed along
15 My translation. 16
My translation.
36
the Wacheqsa river in order to protect the land from river flooding. Looking at the figures and
at Tello’s map, it looks like the Wacheqsa sector had at least three platforms before the 1945
landslide, which is totally different than what can be seen at present. Also, there was major
canalization at work along the Wacheqsa River, as evidenced by the retaining walls shown in
figures 18 and 19. The evaluation of Tello’s descriptions and photographs, taken before 1945,
support the assertion that what is observed now is not how the Wacheqsa sector was before
1945.
This modification was not the only one that the sector witnessed over its history as
intense land use and occupation occurred in this sector during the periods known as Middle
and Late Formative. As I will elaborate in this dissertation, topographic modification was a
constant cultural endeavor over time in the Wacheqsa sector.
3.1 Wendell Bennett’s excavations
The site was first investigated by Wendell Bennett in 1938. He excavated one unit,
Ch-15 that was located approximately 100 m north of Building D with an extension of 4 x 1.5
m (figure 22). He indicated that great numbers of sherds were recovered in the first 1.5 m
depth, decreasing in amount below that depth until totally disappearing at more than 2.00 m
depth. Interestingly enough, Bennett indicates that “the cross-section shows no obvious
layers” (Bennett 1944:80) and that “the materials seem to be house refuse” (Bennett 1944:80).
Bennett dissected this unit using horizontal arbitrary levels of 50 cm each, recovering 1465
ceramic sherds in total mixed with “animal bones, charcoal and sections of small and large
stones” (Bennett 1944:80). The use of arbitrary levels is not ideal in sites with complex
stratigraphies, as the stratified contexts and archaeological materials are not properly
identified and mixed with no relation to natural deposits. Nevertheless, judging from the
materials he recovered such as charcoal, animal bones and fragments of large stones it is
possible that these materials were produced by the discharge of food consumption activities
either at the household or suprahousehold levels; regrettably there were no major indicators of
the nature of the contexts excavated or the locations of these elements within the layers.
3.2 Julio C. Tello’s excavations
In 1940 Julio C. Tello excavated one unit sized 4 x 3 m, four meters south of where
Bennett excavated unit Ch-15. Tello had two hypotheses regarding the nature of the Wacheqsa
37
Sector. He thought, as Bennett, that it probably served as the location of the domestic
settlement related to the Ceremonial Center:
“[The Wacheqsa sector] must correspond to the places where hamlets and houses were. The kind of trash found there can be followed around the contours of the ruins and domestic wares can be
observed. It does not show up in the plaza or next to the buildings17
” (Tello 1940: 25. Emphasis added).
His second hypothesis was,
“These extensive cultivated lands have a thick layer of agricultural dirt and abundant domestic ceramic sherds on surface, and for those reasons these lands could be considered as trash areas. This brownish or chocolate land only tends to appear at the edge of main
buildings18
” (Tello 1940: 25)
It is necessary to note that these hypotheses are not mutually exclusive, “as this
brownish matrix is only present in the edges of the main buildings, it probably was an ancient
trash area and consequently the area where the domestic settlement was located19
” (Tello
1960: 317). Tello also thought that the Wacheqsa sector was contemporary with the deposits
exposed by the Mosna River under Building E, mainly because of the similarities between
stratigraphies and associated materials, assuming that it was also an area were domestic thrash
was deposited.
“The land’s aspect is similar to the trash layer with incised ceramics under the foundations of Building E, and to the land on which Building A was built. Thus, this type of land with black, grey and red incised ceramic sherds must belong to the same period, whose age is
difficult to establish20
” (Tello 1960: 146)
Marino Gonzales was the person in charge of supervising the unit Ch-1 in the
Wacheqsa sector (Tello 1940), identifying four archaeological strata. The first layer is
basically agricultural land in which post Chavín material (Recuay) was recovered. The second
layer produced Chavín ceramics, specifically the ones identified as Janabarriu by Richard
Burger (Burger 1984, 1988; Tello 1960: figs 152, 164, 167 and 169), in this layer Tello also
finds faunal remains of camelid and deer mixed with fragments of stone clubs and stone
mortars. He writes that the ceramics are highly polished with fine shallow engravings on the
17 My translation. 18 My translation. 19 My translation. 20
My translation.
38
surface that belong to the “classic Chavín period” (Tello 1960:317). According to Tello, this
layer could be “considered as the remains of a large midden” (Tello 1940:27) . The third layer
also contained Chavín ceramics, but in this case I note the presence of a mixture of decoration
and forms later identified by Burger as Urabarriu and Janabarriu (Burger 1984, 1998; Tello
1960: figs 153, 155, 156 and 160). These materials were found in a green matrix that
according to Tello seemed to be the product of an alluvial flood that destroyed some domestic
settlement located on the upper part of the civic ceremonial center (Tello 1960). Below the
third layer there was nothing but sterile soil according to Tello (Tello 1960) (figure 23).
The excavations in the Wacheqsa sector reached a depth of 1.00 m below surface,
recording 4 stratigraphic layers. During my own research, I excavated near the area Bennett
and Tello excavated, finding for example in a 2x2 m unit, a total of 24 stratigraphic layers in
average depth of 3.50 m. The stratigraphic complexity in the Wacheqsa sector is greater than
the one Tello and Bennett recorded. I believe this has to do with the excavation procedures at
that time, which employed arbitrary levels and were focused on obtaining of cultural change in
the archaeological materials. Nevertheless, Tello’s profile is indicative of a patterning
observed during my own excavations: the presence of a midden deposit on top of what seems
to be an area where water either was running or was deposited. While Tello recorded the
midden deposit and the water flood deposit as single stratigraphic layers, I have been able to
reconstruct the depositional history of these two analytical units through careful excavations,
confirming Tello’s rough stratigraphical estimations but adding a new level of complexity to
what was already known. In general, 44 stratigraphic deposits have been identified as part of
the Midden and 17 stratigraphic deposits as part of the Water Flood area.
3.3 Rosa Fung’s excavations
After Tello’s intervention the entire site of Chavín de Huántar was covered by the
1945 landslide, and the Wacheqsa Sector was entirely eroded and filled by mud to its current
condition. Research in this sector occurred again in 1973, 1974 and 1975, conducted by
archaeologist Rosa Fung. None of her data has been published yet, but the materials she
recovered are accessible in the storage rooms of the Archaeology and Anthropology Museum
of the San Marcos National University. Rosa Fung excavated 39 units (regrettably there is no
information regarding the size of the units), located at the north edge of the Wacheqsa sector.
Following Burger’s sequence, the majority of the levels excavated yielded ceramics similar to
the shapes and decorations identified as Urabarriu while a few upper layers contained
39
ceramics related to those recognized by Burger as Janabarriu. Rosa Fung believes she found a
Kotosh-Kotosh component in her excavations, below a domestic floor that covered the
Kotosh-Kotosh contexts; “during our last excavations at Chavín de Huantar, in domestic areas,
we have found Kotosh-Kotosh ceramics in deep strata, but it cannot be said that the
superimposed Chavín ceramics descends from them” (Fung 1975:199. Emphasis added).
There are indeed ceramics resembling the styles associated with the Kotosh phase in
Huánuco, specifically the Kotosh Grooved style but these items are present in a mixed deposit
with ceramics showing Janabarriu components (concentric circles with central dot) in layer 2,
test pit 3. The ceramics Fung recovered (figures 24 and 25) suggest a strong Urabarriu
component along the north edge of the Wacheqsa sector, a fact that was confirmed by my own
excavations in the same area during the beginning of my research.
Rosa Fung also identified a cremation area (Fung 2006) characterized by the presence
of pale gray ashes that according to her might have been derived from the burning of human
remains. Lumbreras takes this idea and suggests that some of the carbonized human remains
found in the Ofrendas Gallery might have been processed in the Wacheqsa sector (Lumbreras
1993).
Before my research started, the existing information regarding the Wacheqsa sector
could be summarized in the following way:
• A midden associated with Janabarriu like ceramics is located at the southern section of
the Wacheqsa sector. This midden is characterized by the presence of finely decorated
Janabarriu wares and abundant faunal remains (camelid and deer).
• Under the midden, an area with evidence of water flooding is associated with
Urabarriu like materials.
• A domestic component is located on the northern edge of the Wacheqsa sector. This
component is mostly associated with Urabarriu like ceramics and Kotosh-Kotosh
ceramics.
The information presented in this chapter has been used as a basis for my own
research in the Wacheqsa sector, in which I have been able to confirm and expand Bennett,
Tello and Fung’s assertions regarding the Wacheqsa sector. The Wacheqsa sector was truly a
multicomponent area that was consistently used by the authorities of the ceremonial center.
Activities occurring in the Wacheqsa sector could be easily monitored from the tops of
Buildings C and D in the monumental area. Its geographical location allowed it to be a buffer
40
zone, keeping the ceremonial center and the prehistoric domestic settlement apart from each
other. These ideas are further elaborated in Chapter 7, but for now, allow me to switch gears
towards the theory and methods that have guided the present research.
41
CHAPTER 4
THEORY AND METHODS
Archaeology as a social science seeks to comprehend the social processes responsible
for the existence of the archaeological record. In the pursuit of this endeavor archaeologists
have developed different sets of theoretical and methodological tools in order to accurately
and efficiently investigate the nature of this archaeological record. In this chapter I seek to
elaborate the theoretical grounds upon which my investigation has been developed, as well as
to elaborate on the methodological frameworks used to systematize the data retrieved in the
field and lab.
4.1 Theory of the Archaeological Record
An archaeological site is a spatial cluster of material culture and contexts that are
spatially and stratigraphically distributed. It is the product of a systemic context, that through
social practices give origin to the elements that together form the archaeological record. The
systemic context is an aggregation of social actions that is part of a behavioral chain, and
which
“[…] labels the condition of an element which is participating in a
behavioral system. Archaeological context describes materials which
have passed through a cultural system, and which are now the objects
of investigation of archaeologists” (Schiffer 1972:159)
But the systemic context by itself is more than a behavioral chain; it is generated by a
temporal chain of social practices that in turn emerges from a set of specific historical and
material conditions that induce men and woman to organize themselves and to create social
institutions. As part of a social environment, the amount of energy invested and the way it is
invested in producing the elements archaeologists find in the archaeological record is,
“conditioned by the circumstances in which men [and women] find themselves, by the
productive forces already won, by the social form which exists before they do” (Marx 1973:4)
The social nature of the archaeological record is not self evident and needs to be extracted from
the record itself. In doing this the degree to which the record is a manifestation of the activities
that created them needs to be considered. In ideal cases, Binford’s dictum may be appropriate
42
“The loss, breakage and abandonment of implements and facilities at
different locations, where groups of variable structure performed
different tasks leaves a fossil record of the actual operation of an
extinct society” (Binford 1964:425)
Regrettably there are several conditions that need to be taken into consideration regarding
the exact nature of the archaeological record and how these conditions may have altered the
original disposition of the record and how they can bias our interpretations. This problem was
briefly addressed by Binford (Binford 1968) and extensively discussed by Schiffer (Schiffer
1972, 1983, 1988) who labeled the study of the affecting conditions of the archaeological record
as Formation Processes. Formation process theory is probably one of the most important
contributions to the understanding of the archaeological record, being crucial in the elucidation
of behavioral and social practices. According to Schiffer, (Schiffer 1983:676-678) there are
three areas, in which Formation Processes can be grouped: entropy, statistical sampling and
transformation.
Entropy refers to the process of deterioration that the archaeological record goes through
since its original formation and the moment it is excavated by the archaeologist. In 1970
Cowgill drew a distinction between 1) the events that occurred in the past, 2) the artifacts
created and deposited and 3) the artifacts that are uncovered by the archaeologist. Regarding the
latter he stated that “physical finds populations depend on ancient human activities, but also on
subsequent events, human and non-human, and on the techniques, concepts, and equipment of
investigators” (Cowgill 1970:163). According to him, it is the task of the archaeologist to
distinguish the discontinuities between these three populations and to think about those
discontinuities in terms of sampling problems (Cowgill 1970). The discontinuities are due to a
sampling bias that progressively acted towards the reduction of the quantity and quality of the
archaeological record in relation to its original form. Entropy and statistical sampling are both
related to the transformation of the archaeological record because of time, preservation or
human and non human events. Entropy and statistical sampling can be subsumed into a
transformational approach towards formation processes (Schiffer 1983). Cultural and non-
cultural post depositional events create new patterns of artifact organization, new associations
and may even alter the soil composition of the deposits transforming entirely the original fossil
record deposited. But also it has to be clear that Formation Processes are not entirely related to
post depositional events but also to the original events that created the archaeological record.
43
The nature of the archaeological record sometimes does not allow the perfect
identification of the activities that originated it, but it is equally important to identify the
patterning of the record and its intrinsic spatial variation in horizontal and vertical planes.
“We may not be able to specify or determine what specific activities
resulted in observed differential distributions, but we can recognize
that activities were differentiated and determine the formal nature of
the observed variability” (Binford 1964:425)
In the case of my research, I have chosen to study the intrasite variability of the
Wacheqsa sector, taking into account two dimensions that are intrinsically connected: spatial
and vertical distribution of deposits,
“Beyond the representation of excavation units and their usual
components (such as soil, color, inclusions and texture), the spatial
relationships of the excavation units, be either horizontal or vertical is
crucial to field archaeologists” (Losier, et al. 2007:273)
This approach is useful for the understanding of the deposits of the Wacheqsa sector
and how they vary in their depositional context, in their depositinal history and functionally.
As well as how the deposits that are part of the site differ in: artifact quantity or the variability
within any class of features, artifact diversity or the formal content of the population of
features, artifact density of deposits and non-cultural composition of the deposit (i.e. type of
dirt, types of inclusions, etc).
Ultimately, the basic analytical unit in my research will be the archaeological deposit.
For the purposes of my research, a deposit will be interpreted as a context that is “a bounded
and qualitatively isolated unit that exhibits a structural association between two or more
cultural items and types of nonrecoverable or composite matrices” (Binford 1964:431)
For example, the relationships existing among waste materials in a midden deposit
need to be analyzed in relationship with their context and with the specific quantities of each
element. Artifacts themselves are products of specific actions like extraction of raw materials,
the preparation and transformation of those materials into final products, and use and discard
processes; these actions require a social organization capable of mobilizing labor. Each artifact
is a product of a social process but many analyses of individual artifacts yield only sterile
sequences of typological and stylistic change. With a contextual approach, hierarchical levels
of analysis can aid in the comprehension of the social processes that generated –at a minor
scale—the archaeological deposits, and—at a major scale—the archaeological site. In this
44
respect the analysis of deposits provides a better understanding of social processes as it poses
the following questions: why are these artifacts in the same deposit; what produced the
deposit; and what is the contextual significance of this deposit? The analysis of associations
has to do with the association of materials making up the artifacts within a discrete context, it
also refers to the relationships that artifacts have within a context rather than disconnected
delimited elements; if the context changes, the relationships do too. Thus, “if archaeology is
the study of the material remains of a social activity, by definition, the unit [of analysis]
should be any association that reveals the result of the social event that originated it21
”
(Lumbreras 2005:84. Emphasis added)
4.2 Data Processing
4.2.1 Stratigraphic Analysis
Even though I place the “Stratigraphic Analysis” section under the label of “Data
Processing”, the scrutiny of archaeological strata begins during the process of excavation; the
outcome of the analysis will rely not only on the post-field data analysis, but also “upon the
analysis of the stratigraphic deposits and their excavation as such. This depends upon the units
having been recognized and separated in the field” (Warburton 2003:57). The purpose of
stratigraphic analysis is to establish the spatial and chronological changes in the site before,
during and after its use. The understanding of those processes are, in fact, what almost every
archaeologist attempts to accomplish during an excavation, “breaking the stratigraphic
sequence down into contexts, and representing their relationship in a matrix” (Clark
2000:157). In order to do that, it is necessary to reassemble and synthesize the units recorded,
structuring the constituent parts into groupings. Elaborating on Pearson and Williams
(Pearson and Williams 1993), and Herzog (Herzog 2004) the nomenclature and structure
imposed on the stratigraphic sequence that will be used in the present work is as follows: a)
feature, b) deposit, c) analytical units. A feature is a set of associated artifacts that were
deposited together as part of the same social event; a deposit is a stratigraphic layer that “can
be visually distinguished from the sediments above, below and beside it, differing from these
other sediments by definition” (Warburton 2003:6); an analytical unit is an assemblage of
deposits that can be segregated in groups or clusters based upon their physical characteristics,
artifact content and chronology. Analytical units can be also grouped into chronological
blocks that will determine the phases of the site.
21 My translation.
45
In order to elaborate part of this stratigraphic analysis I have used a computer program
called Stratify (Herzog 2004) that generates a Harris Matrix using the data entered by the user.
The following data is processed by the software per feature or deposit:
• Minimum and maximum x, y, z coordinates.
• Logical relationships among deposits (earlier than, later than, equal to, part of).
The procedure I have for understanding the stratigraphic relationships of deposits in
the Wacheqsa sector is as followed is:
• A stratigraphic matrix was constructed for each excavation unit using the laws of
superposition, original horizontality and original continuity (Harris 1992)
• Using Stratify I have merged individual matrices in order to create a single matrix for
the Wacheqsa sector.
• Using a 3D Computer Aided Design (CAD) model and quantitative methods I have
grouped deposits into Analytical Units and Analytical Units into chronological blocks
(phases) in order to reconstruct the site formation of the Wacheqsa and the
stratigraphic relationships of the deposits.
Each deposit has a correlative number within each analytical unit. For example deposit 101,
102, 103…n belong to the same analytical and deposits 201, 202, 203 …n belong to a
different one.
4.2.2 Digital Models and Archaeology
The understanding of past social processes can be partly an outcome of reconstructing the
depositional history of a site. This has also been elaborated by Warburton, who indicates that
“Although [it is] generally assumed that stratigraphy is about time, stratigraphy can also
provide substantial information about architectural space” (Warburton 2003:29) and I would
go even further to argue that stratigraphy not only reflects time and architectural space but also
strategies of terrain modification. In this sense I have chosen to explore the stratigraphy of the
Wacheqsa sector at Chavín de Huántar through the integration of a stratigraphic matrix into a
CAD digital model that will allow the accurate reconstruction of the excavated areas with the
potential to reveal intrasite correlations of the different strata excavated and recorded, taking
into consideration that
“the visualizing process resulting from solid modeling can sometimes
reveal relationships within an archaeological ‘reconstruction’ more
clearly than any other current methods of display” (Malinverni, et al.
2002:411)
46
The use of digital models in archaeology allows the representation of the real world in
a compact an efficient package, the efficient modeling and simulation of real world processes
in order to understand complex interacting processes of humans and their environments.
Additionally, they allow one to count, do statistics, manipulate and evaluate measurements in
a variety of summary and analytical forms.
Zubrow (Zubrow 2006) makes an interesting culture history of the development of
digital technologies in archaeology and its correlation with trends of archaeological thought
through time.
Date Archaeological
Thought Types of
Problems Hardware-
Software Subjects of Use
Pre-1930 Natural observation
Descriptive Calculating machines
Statistical analysis
1930-65 Culture history Temporal and
geographic
gapsmanship as
well as
reconstructive
Mainframes, Fortran, Cobol
Statistical analysis, data
storage and
manipulation
1965-80 Processual Systemic,
hypothetical,
nomethetic,
behavioral group
oriented
Mini’s Vaxs,
PC, Pascal, C,
Basic
Causation, modeling, simulation GIS
1980-95 Post-processual Individual, interpretative
PC’s, C++, Prolog
Expert systems, non-
causative, AI, field use,
GIS
Table 3: Historical development of digital technologies in archaeology according to Zubrow.
In the same vein Gary Lock also tries to trace the relationship between computerized
technologies and archaeology schools of thought (Lock 2003). The following chart is a
modified version of the one Lock presents in the referenced publication,
Year Theoretical School Theoretical Tools Technological Tools 1960’s Cultural history Classification
Typology Multivariate statistics Mainframe computers
1970’s Processual Quantification Confirmatory
Theory led
Microprocessors Increased software
1980’s -90’s Postmodern Exploratory Non confirmatory
Graphics Visualization
Multimedia
Virtual reality
World wide web
Table 4: Cultural History of digital technologies, adapted and modified from (Lock 2003:8)
47
One can argue about the accuracy of this chronological chart, but what is evident is
that the use of technologies in archaeology has not been theoretically agnostic, at least at its
beginning. Theoretical orientations define the type of problems to be investigated but the
development of a particular theoretical trend exists in a particular time that has a set of
available technologies existing for the resolution of those problems.
On a very general level, a theory can be understood as a proposed explanation of why
things are the way they are; in this sense archaeology offers different brands of explanations
that are translated into competing visions of social systems and human behavior.
Technological methods profusely populated the realm of archaeology during the onset of
processualism as a predominant theoretical school; the search of new, accurate ways of
exploring archaeological data was an endeavor necessary for acquiring an appropriate
understanding of the social processes responsible for the archaeological record, related with a
scientific/objective approach to data. In this regard, the use of the scientific method formalized
the testing of hypotheses invoking the objective nature of archaeology as a scientific
discipline. With the formalization of archaeology a new set of tools borrowed from other
disciplines started to be part of the methodological toolkit necessary for the construction of
scientific knowledge.
“Principles, methods, and techniques from fields as diverse as
systems theory, biological ecology, information theory, and locational
geography now frequently punctuate the archaeological literature.
Although the ultimate utility of many of these ideas remains to be
demonstrated, such borrowings are inevitable and necessary” (Reid,
et al. 1975:865)
Nowadays the use of technologies in archaeology is still as theoretically oriented as it
was at its onset. Postprocessualism has emphasized the need of experiencing the data (Shanks
and Tilley 1992) in order to engage the archaeologist/public with the archaeological record
and investigate the nature of the relationship within individual perception and data; in the
same vein technology has been used to build interactive reconstructions of complete or partial
archaeological sites in order to communicate the past to the public (Campo 2004; Guipert, et
al. 2004; Lock 2003; Meister and Bastian 2004; Meister 2004). The same technologies can be
used also for analytical purposes as I have done in this dissertation with Computer Aided
Design (CAD)
48
Before going into the details of the modeling method I have used with the data I
recorded from my excavations, I believe it is worthwhile to explore the different modeling
methods available for the archaeologists in order to explain why I have chosen Computer
Aided Design (CAD) technology for my work.
In 1989 Paul Reilly introduced the concept of “virtual archaeology” referring to the
use of three dimensional computer assisted models of archaeological data, arguing for their
intrinsic potential not only for research but also for conservation and management of cultural
heritage given the fact that “it is an unfortunate irony that in order to reveal what lies below,
the archaeological excavator must remove, and thereby destroy, what lies above” (Reilly
1989:569)
Given this destructive nature, archaeologists have developed a set of guidelines
regarding archaeological excavations, which warrants not only an adequate extraction of
information but also the exchange of information within the academic community. When put
in practice, these guidelines result in a vast set of information that can be segregated into
graphic and written records that need to be dissected into different categories for further
storage, questioning, elucidation, and presentation. It is in these aspects (format, storage
capacity, availability and manipulation) that digital technologies have increasingly become
attractive for archaeologists from different theoretical fences.
The archaeologist never sees the entire site during the excavation process but only
portions of it, and these portions are conditioned by sampling choices made in determining
which site or which portion of a site to excavate (Binford 1964; Cowgill 1970; Orton 2000;
Schiffer 1983). Part of the site is removed, out of view or buried beneath the exposed surface.
Computer technologies offer an analytical tool that allows a virtual reconstruction of the
features excavated, simulating their original deposition and superposition of strata according
to ways the data was retrieved.
Three-dimensional reconstruction of archaeological stratigraphy has been an
undertaking since the mid 80’s, as shown in a seminal paper published in Nature (Ottaway, et
al. 1986) in which computer graphics designed for protein crystallography were used for three
dimensional visualization of archaeological stratigraphy. This approach was used with data
coming from a Bronze Age pit from Altheim in Bavaria, Germany. The basic principle was
the displaying of color-coded three dimensional contour maps of archaeological layers, the
storage of the information in digital format and the modification of the model when new
information was available (Ottaway, et al. 1986).
49
The next attempt of three dimensional stratigraphic reconstructions was the
Winchester Geographic System (WGS) relational database management system linked to
graphic facilities. This system was previously used in chemistry and medicine fields (Colley
1988) and was developed by the IBM United Kingdom Scientific Centre (Reilly 1989). The
system was composed on three main elements 1) a relational database, 2) a 3D graphics
system, and 3) a bridge software that links the relational database with the graphics system.
Every record in this database holds a three dimensional position coordinate, an orientation,
and a potentially large set of additional attributes including length, width, weight and fabric.
The data can be scrutinized from various locations and viewpoints, many of which are
inaccessible in the physical world. WGS was tested in the middle Saxon site of Hamwic,
specifically in a single midden pit of 1.5 x 2.0 m wide and 1.2 m deep that had 17 strata
recorded. Descriptions and positions of materials recorded in a trash pit were entered into
WGS’s database; the basic units of analyses were the shapes of stratigraphic layers and the
location of selected items. Each query into the database would generate a graphic image of the
spatial relationship of the layers/items queried (Colley 1988:101). The WGS system operated
under an extensive hardware that included a mainframe computer and terminals controlled by
mini computers.
The early days of different technological experimentation ceased with the availability
of CAD and GIS tools. With the sophistication and extensive distribution of software available
to archaeologists, there has been a massive increase in experimentation with visual models
that allow the investigation of the ways human have manipulated the space and constructed it
according to their needs.
The trigger for visualization has been the search for better ways to communicate the
existing knowledge to the public. The model needs to be a representation of real data. The
more reliable the data, the more useful the resulting model. The user not only moves within
the model but also extracts information about different aspects of the model. In the case of
stratigraphic modeling, the virtualization of the deposits is significant because,
“Displaying three-dimensional excavation units could be an important
aid in understanding stratigraphical relationships and identifying
potential patterning. These kinds of systems would also allow us to
perform metric analysis, to validate interpretations or to formulate
new ones since archaeologists would be able to revisit their site in
immersive reality” (Losier, et al. 2007:273)
50
4.2.2.1 GIS and CAD
In the next pages I will focus in three technologies that to a certain degree are similar
in the creation of virtual models: GIS and CAD which share the following advantages in spite
of their particularities: capture of data, modeling of data, storing of data, sharing of data,
analyzing data, displaying geo-referenced data.
4.2.2.1.1 Geographic Information System and Stratigraphic Modeling
Geographic Information Systems (GIS) is a broad term that covers a wide variety of
software that works with geo-referenced digitized coverages. GIS works mainly with raster
images but lately there is an increased integration of raster and vector data. GIS is designed to
“integrate the spatial data with an attribute database so that spatial data elements can have
large amounts of text (and image) data associated with them” (Lock 2003:54)
One example of GIS use for stratigraphic reconstruction is the case of the site of Tell Leilan
(Potzolu, et al. 2004). An excavation of a trench of 9 x 1 m was modeled using the following
assumption
“Assuming that the surfaces between the western and eastern sections
(separated by just one meter) had a regular slope, the section’s upper
and lower profile of each stratigraphic unit [layer] was enough for the
interpolation of the final surfaces” (Potzolu, et al. 2004:435)
The modeling process went trough a series of complicated steps that can be summarized as
follows:
• Digital photography of profiles
• Recording of control points on each profile using a EDM
• Raster rectification and mosaic creation using Rolleimetric
• Vectorization raster mosaic in GIS environment
• Geo-referencing of vector images in CAD environment
• Attribute data base creation in GIS
• Conversion of profile shapefiles into 3D polylines in CAD
• Creation of Triangular Irregular Networks in GIS environment.
Another alternative GIS approach to stratigraphic recording (no less complicated) has
been proposed by Doneus and Neubauer (Doneus and Neubauer 2004) who recommended the
following procedures:
51
• Single surface mapping of every surface excavated
• Digital photography of each surface.
• Digital rectification of each image.
• EDM recording of surface boundaries, topography and positioning of features.
• Geo-referencing of images in GIS.
Stratigraphical reconstruction with GIS technology is cumbersome and time expensive
when compared with reconstructions made using CAD technologies,
“[..] CAD tools and GIS systems perform differently in 3D spatial modeling context. GIS software is usually employed with larger data
sets whereas CAD tools are exploited in local applications.” (Losier,
et al. 2007:274)
GIS technologies seem to be more appropriated in analyzing large geographical areas
with ample data sets while CAD is more suited for site specific data sets. This quality of CAD
will be discussed in the following pages.
4.2.2.1.2 CAD Modeling
Computer Aided Designs (CAD) is a vector based system in which primary data is
composed by points and lines. It works with a “referenced coordinate system22
” (Lock
2003:53) that can either be global UTM values or arbitrary site grids. Initially it was merely a
two dimensional tool until 3D exploration was developed allowing the creation of complex
surface and three dimensional models. The computer models created are composed by
different drawings that are stored into databases; they are geometric models in which the level
of accuracy of the model depends on the number of points recorded per entity. EDM’s allow
enough precision if the surveyor records as many points as possible if surfaces are highly
curved and decrease the number of points once the surface flattens. CAD models are entirely
flexible, as entities can be grouped into different drawing layers which can be selectively
hidden from the user’s view, allowing the creation of different views that can be segregated
either contextually or chronologically or using any criteria that the archaeologist thinks is
useful. A more detailed description of the methodology used for modeling stratigraphy at
22 X and Y values for north-south and east-west positions and z values for heights in reference to any
assigned datum.
52
Chavín de Huántar will be developed later in this chapter but for the moment I believe it is
pertinent to cite Eiteljorg in order to have a big picture of what CAD can do visually,
“As the model is created, the underlying base of a 3D coordinate
system means that the complete geometry of the item being modeled
is always retained. Furthermore, CAD programs have the capacity to
provide a view of the model from any point in space, using the
geometric information and the rules of geometry and perspective to
generate any desired view on command. The view can be a plan,
elevation, axonometric, or perspective view, at the pleasure of the
user” (Eiteljorg II 2007:154)
The current sophistication of CAD models comes from a long experience in adjusting
this technology to be efficient in the fields of architecture and engineering. Architecture is one
of the types of data that archaeologists deal with and it is to be expected that archaeologists
would embrace this technology and exploit its potential. It was not too long before
archaeologists started to experiment with other types of data using CAD technology,
stratigraphy being one of those types of data. An early example of stratigraphic analysis with
the aid of Computer Assisted Design (CAD) programs was published in 1989 and according to
his developer “combines the philosophy of the Harris Matrix with single context plan elements
to reconstruct composite plants and three dimensional models for investigation and analysis”
(Alvey 1993). This system was named Hindsight and tested in the site of Stakis in York (UK).
Hindsight worked under a CAD platform allowing the digitizing of composite plans or single
context plans, each plan has a unique height that is contrasted with the height of the different
plans drawn in order to create an isometric view of the strata surfaces superimposed
displaying the sequence of deposition on the site. It is important to note that Hindsight is
essentially a visualization of a Harris Matrix, in which strata surfaces retain their scale but the
vertical or superimposed representation does not represent depth but the position of each
surface in the matrix. Hindsight allows the management of strata according to their position in
the Matrix which is generated when the depths of each single deposit are entered. Each deposit
immediately is recorded into a database and when a stratum is called from the database
Hindsight checks its stratigraphic position (confronted basically by heights) in order to look
for inconsistencies. As with any CAD application, the view on the screen can be rotated and
viewed from an assortment of angles and directions that can be plotted. Hindsight has the
ability to texture strata using different colors and hatching grids which act as a useful tool for
segregation of phases or groups in the three dimensional Harris Matrix. The limitations on
53
Hindsight lie in the impossibility to carry out spatial analysis of deposit distribution; it does
not allow for the possibility of intrasite analysis as deposits are visualized in a vertical
relationship that does not display horizontal complexity. Deposits from different sectors of a
site are shown in a chronological relationship but not in a spatial relationship. Hindsight is
basically a three dimensional Harris Matrix that worked very well in the field of
archaeological visualization.
In 1994 William Beex stated that the main tasks of using CAD tools in archaeology
needed to be related with the creation of a rapid general view, linked with attribute databases
and capable of reconstructing the recorded archaeological remains (Beex 1994). The
advantage of digitizing excavation plans lies in: the production of an unlimited amount of hard
copies, the possibility of dealing with multiple coordinate systems and the opportunity of
transforming maps of different scales into one single scale since the input will follow any
chosen basic unit. The uniformity of a single scale will allow the identification of measuring
mistakes.
The novelty of this approach lies in the possibility of organizing strata into surface models
with specific colors but most importantly it allows the comparison of surface models from
different excavation units. It was suddenly possible to carry out intrasite analysis on extensive
areas.
In 2000, Mikhail Zhukovsli tested the use of CAD stratigraphic modeling in the site of
Gnzdovo that has an extension of 16 ha divided in four groups of mounds built during 800 and
1100 AD (Zhukovsky 2001). The decision of using CAD over GIS technology was made on
CAD’s coordinate geometry (COGO) engine that “satisfies strict requirements for accurate
and fast direct recording of excavation records” (Zhukovsky 2001:433. Emphasis added). The
software used was AutoCAD 2000 which has the ability to support vector and raster data and
also has better three dimensional modeling and analytical capacities than GIS.
The area excavated was a trench of 40 m² and the strata was recorded following
traditional hand made drawings that were digitized into CAD. Each strata surface was
measured using a grid of 40 x 40 cm in order to take as many points as possible for an
accurate modeling of it. Data was collected manually and with the help of an EDM. Surfaces
were modeled following these steps: 1) surface points and surface boundaries were integrated
into a triangulated irregular network (TIN), 2) an extrapolation grid was created with 20 x 20
cm cell density and 3) these two elements were merged together giving shape to three
dimensional surface deposits (Zhukovsky 2001) 433. The application of CAD technology in
54
the site of Gnzdovo was restricted to the 40 m² excavated, additional areas were not excavated
in order to test the intrasite possibilities of the application of CAD modeling for stratigraphic
analysis, but this approach has set the grounds for this kind of analytical approach.
A more sophisticated analysis was recently published by Losier at al (Losier, et al.
2007). The authors used a non commercial software named GoCAD with data from the site of
Tell ‘Acharneh, Syri. They chose to model adjacent excavation units of 5 x 5 m that were
recorded using a Trimble 5800 Real-Time Kinematic (RTK) Surveying GPS . GoCAD allow
the modeling of stratigraphic units as solid models using either voxels or equilateral
tetrahedrons23
. Voxels models are usually extremely heavy and also show holes and empty
spaces in the solids when the voxel size is too small. Tetrahedron models on the other side are
not as heavy as voxel models and they are “faithful to the limits of the surface model because
nodes used to create the tetrahedron pass exactly by the control points since tetrahedrons
honor complex geometry of objects” (Cattani, et al. 2004:282).
Stratigraphic analyses and visualizations have been greatly improved with the use of digital
technologies: they allow the storage of records in a more durable format, strata can be visually
reconstructed the way they were before archaeological excavations started, layers can be
observed from different perspectives usually inaccessible in the real world and they can be
segregated into spatial analytical units of horizontal or vertical distribution allowing the
examination of intrasite variation.
4.2.2.2 Modeling the Wacheqsa Sector
The model I have constructed is an empirical model based on direct information
recorded through archaeological excavation. In order to construct this model I have chosen to
do it using Autodesk Land software which works under a CAD environment. The reasons for
using this particular technology are it’s availability, flexibility and its excellent resources and
wide range of technical manuals. But most importantly it allowed me to build a three
dimensional model of the strata recorded, to perform spatial and stratigraphic quantifications
and to identify analytical units.
Archaeological deposits are composed of layers, features and interfaces that have
surfaces and volume information. The task of the archaeologist is at first to accurately record
these elements in order to initiate the interpretive process of the archaeological record. The
23 A tetrahedron is a polyhedron formed by four triangular faces, three of which converge at each
vertex.
55
procedure of model building is useful for articulating ideas about the original structure of
archaeological features. Geometric reconstructions necessitate that the archaeologist identify
explicitly each and every element in the model and their spatial relationship with one another.
The definition of the model forces archaeologists to reconsider the original data, which can
focus attention on problematic areas and gaps, causing them to observe, or record in a
different manner, specific categories of data in future investigations (Eiteljorg II 2007).
All stratigraphic deposits in the Wacheqsa sector were modeled as a three dimensional
geometric representation of its upper and lower surfaces (figures 26, 27, 28 and 29). In
Autodesk Land surface models are made of triangles created when the program connects the
points that structure the surface data, these triangles form a triangulated irregular network
(TIN) surface, the lines that form these triangles are known as TIN lines (Autodesk 2006). The
points used for the creation of a stratigraphic surface model were taken in the field during the
excavation process. In general points should be taken segregating two types, the boundary and
the surface itself, a boundary is the polygon selected as surface boundary which is read as a
two dimensional item, the contour is the three dimensional information that the surface
polygon carries. When created, the program records the number of lines that make up the
polygon, the elevation range of the surface contour and its maximum and minimum
coordinates.
Points can be either entered manually into CAD at the time of creating the surface
model or can be organized into an ASCII text file or an Access database. The boundary points
are identified as contour vertices and integrated into the surface model as breaklines that
prevent triangulation lines from crossing the contours. It is very important to define the
boundary of a stratigraphic surface model as it controls how the surface extends to its external
limits.
Contour vertices are joined with three dimensional polylines, each of these polylines
have an x, y, z value in the plan that when joined create a three dimensional polygon (3DP).
Once the 3DP is produced a surface model can be created. Autodesk Land has a feature called
“Terrain Model Explorer” (TME) that manages the creation of surface models. Before the
creation of a surface model, data must be entered into the TME. In adding surface data to the
TME the user determines which points, breaklines, contours or boundaries are parts of the
surface model. These items are entered separately per surface model into the TME.
After choosing and entering the information the user wants to include in the surface, the
program is ready to start building the surface model. CAD-LAND process the data entered and
56
calculates the surface triangulation combining the boundary and contour information
interpolating the results. In addition to the information already provided, in the process of
building the contour of the surface model when a surface is built, the following information is
provided: 1) Number of triangles, the number of triangles in the surface TIN; 2) two
dimensional surface areas, the apparent surface area if one looks at the surface from plan view.
It is obtained by projecting the visible triangles onto the XY plane along the Z axis and
summing the areas of the triangles. If areas on the surface are hidden within boundaries, then
these areas are not included in the surface area; 3) three dimensional surface areas, it is the
true area of the surface and accounts for variations in the surface elevation. The 3D area is the
sum of the areas of each of the visible triangles in the surface without projecting the triangles.
The greater the variation in elevations, the more the 3D area differs from the 2D area.; 4)
Mean Elevation, the mean elevation of the surface; 5) minimum triangle area, the area of the
smallest triangle in the surface TIN; 6) maximum triangle area, the area of the largest triangle
in the surface TIN; 7) minimum grade, the minimum grade of the surface, in Autodesk-Land a
grade is always expressed as a percentage, a value of 5 would imply a 5% grade which is
translated into 5 vertical units every 100 horizontal units; 8) maximum grade, the maximum
grade of the surface model; and 9) the average grade of the surface model.
The model allows the visualization of the totality of the surface strata spatially and
stratigraphically located as they were prior the excavation. Measurements can be performed
and deposits can be segregated according to their spatial organization, geometry and grade.
This is important to understand the complexity of the entire Wacheqsa sector and also for
planning strategies of future excavations.
4.2.3 Quantitative Analyses
4.2.3.1 Analysis of Deposits
Frequency densities of class artifacts were tabulated in order to identify patterns of
distributions per deposits and analytical units. Frequency densities per deposit were entered
into JMP Statistical Package and plotted against the analytical units identified in order to see
the variation in artifact density per analytical unit.
4.2.3.2 Boone index
The Boone index was first conceived for the analysis of midden deposits; however, I
have adapted it here in order to use it with stratigraphic units. In Boone’s original estimation,
57
each midden will constitute the unit of analysis, while in this dissertation the unit of analysis
will be each stratigraphic deposit. Boone measured the number of artifacts and the number of
classes of artifacts per midden while I have measured the same items but for each stratigraphic
deposit.
The purpose of this index (Hi) is to compare the individual provenience units
(deposits) of artifacts with the cumulative distribution of all deposits combined. Larger values
of Hi indicate a high measure of homogeneity (the prevalence of one class over the rest) and
lower Hi values will in turn indicate a low measure of it (Boone 1987). Spatial variation in
occupational density within a settlement would be an obvious reason for non uniform
distribution and size of classes within deposits in a given area. Deposit size refers to the
number of artifacts present in the deposit.
Elaborating on Boone (Boone 1987), perfect heterogeneity (Hi value of 0) is reached
when all defined categories or entities in a population are present in equal quantities. Perfect
homogeneity (Hi value of 1) exists when the population consists almost entirely of only one
category. Consequently, in analyzing a series of stratigraphic deposits at the Wacheqsa sector,
a measure of heterogeneity was followed in which the site-wide artifact mix, that is, the
relative proportion of artifacts over the whole site is considered heterogeneous in that it
reflects the relative proportion of deposit-producing activities over a given area (in this case,
the Wacheqsa sector). Having this expression of site-wide artifact mix, it is possible to
measure the differences of Hi values among deposits and even among analytical units
grouping all Hi values of each analytical unit in order to see how they behave in comparison
with each other.
In calculating this index, the following procedure is followed:
- Primary data consists of frequency counts (denoted yij) from a range of artifact classes
(j) which were retrieved from a number of distinct deposits (i). Site-wide totals of
each artifact are denoted Yj
- An expression of site-wide relative frequencies is obtained by calculating the ratio of
one class total to each remaining class. This ratio becomes the weighting factor (Wj).
Wj is calculated for every class from every deposit.
- Weighed percentages (pij) of each class in each deposit are obtained by dividing each
individual weighed value by the sum of all weighed values from a given deposit.
- The difference between 1 and the number of classes recorded (j) is calculated.
58
- Finally, the index (Hi) is calculated upon the sum of the squared deviations of pij
minus Pj.
Wj= yij
Yj
pij= Wj
∑Wj
Pj= 1
j
Hi= [∑ (pij-Pj) ²]
yij= class artifact frequency counts per deposit (number of certain class artifact per deposit)
Yj= site wide totals of each artifact class (total number of certain class artifact in all deposits)
Wj= ratio of one class total to each class total
pij= weighed percentage of each class in each deposit.
j=number of classes
The first two steps were followed according to Boone (Boone 1987), while the last
three are suggestions recommended by Ian Robertson (Robertson 2007).
Ten classes of artifacts have been considered: decorated ceramic sherds, diagnostic
ceramic sherds, obsidian, burnt clay, anthracite mirrors, mollusks, lithics, projectile points,
bone artifacts and chrysocolla.
The null hypothesis to test while using this index is that all spatial analytical units are
composed of deposits that have the same Hi value, meaning that all deposits of each analytical
unit have equal proportions of classes of archaeological materials.
Before trying to evaluate the richness of the deposits of the Wacheqsa sector, it is
necessary to follow Cruz-Uribe when he states that “the relationship between sample size and
diversity and richness should be investigated prior to any interpretation” (Cruz-Uribe
1988:194). Sample size can seriously affect measures of diversity as large deposits may
contain a larger number of classes (Boone 1987; Kintigh 1989; Orton 2000); moreover, given
two populations with an equal number of classes, one of which has equal frequencies for all
classes, and the other for which high frequencies are concentrated in a small subset of the
classes, for small sample sizes the former case will give raise to smaller values of Hi than the
latter.
59
Once Hi values have been calculated, I generated a 90% confidence interval in order
to identify those deposits that are outside the confidence interval expected by sample size.
Normally this would be enough for testing sample size bias but I have gone a step further.
Once the expected richness (Hi) and associated confidence intervals are generated, I
repeatedly sampled from the observed population using a Monte Carlo routine in order to
determine whether Hi values calculated could reasonable be due to sample size bias. Monte
Carlo routines are particularly useful for testing the significance of a test (in this case the
Boone index) as “with a Monte Carlo test the significance of an observed test statistic is
assessed by comparing it with a sample of test statistics obtained by generating random
samples using some assuming model” (Manly 1991:21. Cited by Shennan 2006:64). Using a
Monte Carlo routine I was able to elucidate if the Hi values observed in the Boone Index and
even those observed outside the 90% confidence interval, are representing a real behavior of
the archaeological materials within deposits and analytical units or if it is just a reflection of a
simple size bias.
Either way I consider it important to test this measure of diversity in order to clarify
its validity and application in the archaeological assemblage of the Wacheqsa sector.
All calculations (Boone index and Monte Carlo routines) were made using R
statistical software package.
4.2.3.3 Kernel density estimates (KDE): univariate, bivariate
A sub set (n=3020) of the total population of diagnostic ceramic sherds (n=12017)
was tabulated using the following parameters: strata, analytical unit, shape, diameter,
thickness and phase in order to find patterns of association between the categories of diameter
and thickness per type of ceramic vessel and to identify possible differences of those
patterning’s among analytical units. In this regard the following ceramic shapes have been
identified: neckless jar, bowl, jar, cup and plate.
I have used Kernel Density Estimates (KDE) for the purpose of identify modalities in
diameters and thickness per type in each analytical unit. KDE can be separated into univariate
and bivariate. A univariate KDE can be understood as a smoothed histogram that avoids the
constraints of a histogram (Baxter 2003; Baxter, et al. 1997; Shennan 2006; Wand and Jones
1995). Given n points X1, X2, . . . , Xn situated on a line a KDE can be obtained by placing a
‘‘bump’’ (essentially a uni-modal density function) at each point and then summing the height
of each bump at each point on the X-axis. The kernel is usually a symmetric probability
60
density function (Baxter, et al. 1997). The spread of the bump is determined by a window- or
band-width, that is analogous to the bin-width of a histogram. The determination of the kernel
band-width is very important as the size of the band width will determine the output of the
estimation. There is no uniform theory on how to regulate the band-width size but it is
appropriate to do it intuitively trying to find the balance between undersmoothed and
oversmoothed results (Baxter 2003; Baxter, et al. 1997; Shennan 2006). In order to do that it is
necessary “to begin with a large band-width and to decrease the amount of smoothing until
fluctuations that are more random than structural start to appear” (Wand and Jones 1995:58)
If univariate analysis can be regarded as an alternative to the histogram,
“It might be argued that, with univariate data, the advantages of using
a KDE as opposed to a histogram for data representation are not so
great as to cause them to be preferred on a routine basis. For bivariate
data the case for using KDEs is much stronger” (Baxter, et al.
1997:347)
The potential of KDE is stronger in bivariate analysis; KDE can be very effective
when applied to scatter plots, showing concentrations of points or modality in the data,
especially for large data sets as it is difficult to make sense of scatter plots. Bivariate KDE are
straightforward to contour in terms of inclusion of specific percentages of the most densely
clustered points, KDE can be used as an informal kind of clustering method that does not
impose structure on the data in the way that more formal methods often do (Wand and Jones
1995).
Univariate KDE plots of diameter and thickness were built per type in each analytical unit
using R statistical software package and bivariate KDE were build using JMP statistical
software package. Bivariate KDE plots a smooth surface that describes how dense the data
points are at each point in that surface; these plots can be used for producing contour plots
which leads to graphical representations of the data examined” (Baxter, et al. 1997:349). In
creating this graphical representation JMP adds a set of contour lines showing densities around
the modes identified. Optionally the contour lines are quantile contours in 5% intervals with
thicker lines at the 10% quantile intervals. This means that about 5% of the points are below
the lowest contour, 10% are below the next contour and so forth. The highest contour has
about 95% of the points below it (Sall, et al. 2005). With this information a mesh plot can be
created, which is basically the three dimensional representation of the bivariate KDE. This
nonparametric density method is relatively costly for a small number of points, being this
61
reason why a selective approach has to be used with this method in regards to sample size,
having a low number of points scattered will only fabricate modes.
The steps in the method are as follows: JMP divides each axis into 50 binning
intervals, for a total of 2,500 bins over the whole surface, counts the points in each bin,
decides the smoothing kernel standard deviation using the recommendations of Bowman and
Foster (1992),runs a bivariate normal kernel smoother using a Fast Fourier Transformation
algorithm (FFT) and inverse FFT to do the convolution, and creates a contour map on the
2,500 bins using a bilinear surface patch model (Sall, et al. 2005). Based on the KDE
estimates JMP performs a modal clustering analysis in which clusters or density
concentrations are translated into a table where the definitive modes are identified. Ultimately,
these modes are the ones that represent the ceramic behavior regarding diameter and thickness
relationships among the ceramic types identified in the Wacheqsa sector.
It is important to mention the reason why I decided to work with both univariate and
bivariate KDE. I wanted to test if the modalities observed in univariate KDE were replicated
using bivariate KDE using thickness in addition to diameter. I also wanted to test whether
diameter was a good indicator of vessel size modality as it has been used consistently in the
literature (Blitz 1993; Drennan 1996; Longacre 1999; Mills 1999; Potter 2000; Rosenswig
2007). I expected that the modalities reflected by using diameter would resist when adding
thickness as a new dimension, so both measurements combined would replicate the patterns
observed when only diameter was considered.
In closing this chapter I would like to enumerate the methodological contributions of the
present dissertation:
• Three-dimensional stratigraphic modeling: allows the visual reconstruction of the
strata recorded as well as the observation of strata distribution either horizontally or
vertically.
• Harris Matrix: allows the chronological organization of the deposits excavated. In a
stratigraphically complex area as the Wacheqsa sector, the use of a Harris Matrix is
almost imperative. When combined with three-dimensional stratigraphic modeling,
the understanding of the archaeological deposits is spatially and chronologically
clearer than using traditional methods such as bi-dimensional trench profiles or bi-
dimensional composite plans; it allows the segregation of spatial analytical units based
on strata spatial organization.
62
• Density distribution of the archaeological artifacts. The measurement of density
indexes per each class of artifact within each deposit allows the characterization of
each analytical unit in terms of artifact distribution.
• Boone Index of diversity: allows the measurement of diversity indexes per each
artifact class within each deposit. When these indexes are grouped into analytical
units, it should be possible to identify artifact behavior in each analytical unit. Higher
Hi values would indicate homogeneity within the archaeological record while lower
Hi values would indicate heterogeneity. Homogeneity can be interpreted as indicative
of tasks that involved only certain archaeological materials while heterogeneity would
indicate activities that involved a diverse array of archaeological artifacts. Sample size
bias should be addressed before interpreting the results of the Boone Index.
• Kernel Density Estimations: the use of bivariate KDE in a large subset (n=3020) of
diagnostic ceramic sherds (rims) accurately identify the predominant ceramic modes
within each analytical unit, characterizing the particularities of their ceramic
assemblages and shedding light regarding the activities that characterized each unit.
The outcome of the use of these methods is described, explained and discussed in chapters 6
and 7. For now, let me turn to the description of the excavations and strata recorded in the
Wacheqsa sector.
63
CHAPTER 5
ARCHAEOLOGICAL EXCAVATIONS
The present research at the Wacheqsa sector started in year 2003 with three
excavation units of 4 x 1.5, 2 x 2 m and 4x4 m (WQ-1, WQ-1 WE and WQ-2) located at the
north edge of the sector. These excavations had the goal of tracing early deposits reported by
Rosa Fung (Fung 1975, 2006) and cited by Lumbreras (Lumbreras 1993) and Burger (Burger
1984; Burger 1998). In year 2004 four units were excavated (WQ-3, WQ-4, WQ-5 and WQ6),
two of them located 15 meters south of the north edge, another located over the terrace that
serves as the south edge of the sector, and a last one placed over the west end of the sector
(figures 30 and 31).
These six units excavated between years 2003 and 2004 provided a rough idea of the
nature of the archaeological deposits present in this sector but were insufficient for
characterizing either the degree of variability inherent in them, or the social processes
responsible for their depositions. The sampling strategy used in this first stage of research can
be considered informal purposive sampling, as “the choice [of the areas to excavate] may be
based on archaeological criteria or those of time, cost and convenience, or a combination of
them” (Orton 2000:2). Orton defined two types of informal sampling: purposive and
haphazard. Purposive sampling is related to the excavation of areas carefully selected given
topographical features while haphazard sampling is related with the excavation of areas where
there are concentrations of archaeological materials on surface. Hence the first stage of
excavations at the Wacheqsa sector was an informal purposive sampling as excavation units
were laid out either according to their proximity to Fung’s original excavations, or to their
proximity to archaeological features such as exposed walls or platforms. Concentrations of
archaeological materials were not distinguishable on surface as the surface land was entirely
covered by the 1945 landslide. These excavations represent only 66m² (0.45%) of the total
extent of the site. The main difficulty with this sampling strategy is that it lacked the potential
of generalization for an entire population of 14600 m², however it allowed me to hypothesize
preliminarily the Wacheqsa sector as a multicomponent deep stratified archaeological unit.
64
These units impose a challenge in sampling them as,
“even under the most favorable excavation conditions deeply
buried occupations cannot be as extensively excavated as surface
occupations because of the great expense in opening large
exposures and because of the inordinate future commitment to
pursuing such a vast objective” (Brown 1975:157)
Given the deeply stratified nature of the Wacheqsa sector, it was necessary to
elucidate a sampling program that would allow me to make general inferences about the
activities developed in the Wacheqsa sector taking into account not only the spatial horizontal
variation of the archaeological record but also its vertical characteristics. Additionally, the
elucidation of a sampling strategy at the Wacheqsa sector needed to take into consideration the
site’s conservation, time and finances. Chavín de Huántar is an UNESCO world heritage site
with specific regulations on what can be done and what cannot be done in terms of excavation
and conservation, similarly large area excavations in deeply stratified archaeological units
require not only the investment of time but also financial resources that were not under the
possibilities of the project. As a result I decided to sample systematically the central section of
the Wacheqsa Sector, an area of 3500 m². Systematic sampling “simplify[ies] or speeds up the
process of selection by choosing units at equal intervals throughout the sampling frame with
only the first one being selected at random” (Orton 2000:21). Fieldwork in years 2003 and
especially 2004 hinted at the probability of the existence of stone structures associated with
Janabarriu-like ceramics covered by platforms. With the goals of finding more of these
structures, investigating in detail their associations, and examining the cultural deposits prior
to the construction of these structures, the south central section of the Wacheqsa sector was
targeted as a systematic sample area for excavations.
A sampling area of 400 m² (20 x 20 m) was defined in the south central section, not
only for the purpose of finding potential structures but also in order to cover the south section
of this sector from which I had no information (figure 32). The 20 x 20 m unit (WQ-7) was
divided into 4 segments which had four 2 x 1 m units arranged equidistant from each other.
The units were deemed to be big enough to provide a good likelihood of feature discovery,
and small enough not to be too time-costly to excavate. Once an archaeological context was
identified, the units could be expanded in order to accurately record its associations. Not all
the units were excavated; only eight were investigated and expanded reaching in some cases 6
m². This sampling is purposive and flexible because the criteria for the location of the 20 x 20
65
m is none other than the knowledge gained from past field season. It is also systematic
because there is formal criterion of location of small units within the chosen sampled area and
it is also flexible as I was able to expand excavation units according to the findings identified.
With this sampling strategy I excavated an additional 0.22 % of the entire site (32 m²). In
addition to this systematic sampling program, an informal purposive sampling strategy was
used in order to add more information to the one that was being retrieved in WQ-7. These
units (WQ-8, WQ-9 and WQ-10) added an extra 0.24% (36 m²) to the total area sampled. In
total, from years 2003 to 2005, an area of 0.91% (132 m²) of the entire Wacheqsa sector has
been excavated
5.1 Units Excavated
In this section I present crude stratigraphic descriptions of all layers excavated in the
Wacheqsa sector. At the same time the exact location of the NW corner of every unit is given,
in reference to the site grid; all color codes given are Munsell soil color codes.
5.1.1 WQ-01 (N 849, E 484.5)
This unit had an area of 6m² (4 x 1.5 m) and reached a depth of 1.90 m below surface.
It was composed by two strata deposited on top of a stone platform (figure 33).
Layer Soil Description Color Notes 01 Matrix mixed with small rocks. At a depth of 0.50
m medium size rocks appear in an irregular form. Black (2.5Y, 2.5/1) It was
deposited by
the natural
aluvión slide
of 1945. 02 Matrix mixed with a low density of angular rocks
and cobbles of different sizes. Reddish Brown (5Y 5/3)
Agricultural
land. 03 Stone platform built with angular rocks of medium
and small sizes joined together without any clay. It
covered the entire extension of the initial 4x1 m
trench. The excavation had to stop in this feature in order to preserve the platform.
The
platform
was
designated
as Feature 1
in this unit.
Table 05: Stratigraphic summary of unit WQ-01
5.1.2 WQ-01 (west extension) (N 847, E 483)
This unit had also 4m² (2 x 2 m) and was located adjacent to the west of Unit WQ-01.
The excavations in this unit reached a depth of 3.56 m below surface (figures 34-36)
66
Layer Description Color Notes 01 Matrix mixed with small rocks. At a depth of
0.50 m medium size rocks appear in an irregular
form.
Black (2.5Y, 2.5/1)
Aluvión slide of 1945.
02 Coarse matrix mixed with a low density of angular rocks and cobbles of different sizes.
Reddish Brown (5Y 5/3)
Agricultural land
03 Loose clayish matrix mixed with coarse sand and abundant small and middle sized angular rocks
Very dark brown (10YR 3/6)
Agricultural land
04 Loose coarse sand mixed with dirt and small angular rocks
Brown (10YR 3/6)
Feature 01
Stone wall conformed by three superimposed
rows of large sized cobbles; the upper row
includes a batan (large concave stone used as a
surface to grind grains ) reused as a wall component. The stones are joined with mud.
This feature has a
north-south
orientation and
served as a retaining wall for
the platform
recorded in WQ-1
(Feature 1). 05 Semi compact soil composed by coarse sand
mixed with middle and large sized angular rocks Very dark gray (10YR 3/1)
Date AA75385
was taken from
this level. 06 Ashes, burnt earth and carbon Very dark gray
(10YR 3/1) Located in the south west
section of the
unit. 07 Rocky deposit, composed mainly by small and
medium sized angular rocks, with almost no soil
among the rocks
08 Clayish matrix mixed with small and medium sized angular rocks
Brown (10YR 3/6)
09 Coarse sand mixed with abundant large sized cobbles
Light brownish gray (10YR 6/2)
Sterile soil
Table 06: Stratigraphic summary of WQ1 (west extension)
5.1.3 WQ – 02 (N 852, E 483)
This unit of excavation had an area of 16m² (4 x 4m). It was located on the northern
edge of the Wacheqsa sector, intersecting a drop of 2.10 m that marks the end of the
Wacheqsa northern platform. This drop is delimited on its north side by a zigzag stone wall
that extends 100 m along this northern end; this wall is on the lower end of the drop. The unit
was placed in this area to document the stratigraphy of the northern end of the Wacheqsa
sector and to uncover the archaeological deposits that were associated with the zigzag. Unlike
the rest of the units excavated in the Wacheqsa sector (with the exception of unit WQ-5), it
was not oriented towards the north but towards the NE in order to intersect the drop more
efficiently. This unit was divided in four sections of 4 x 1 m each, numbered from west to east,
in order to have a better control of the drop (figures 37-39).
67
Layer Description Color Notes 01 Matrix mixed with small rocks. At a depth of 0.50 m
medium size rocks appear in an irregular form. Black (2.5Y, 2.5/1)
Aluvión slide of 1945.
02 Matrix mixed with a low density of angular rocks and cobbles of different sizes.
Reddish Brown (5Y 5/3)
03 Medium and large sized cobbles mixed with angular rocks and a clayish soil
Reddish Brown (5Y 5/3)
Similar two layer 2
04 Damaged white clay floor. It was found in a very poor condition most likely due to the fill (Layer 3)
that was sitting upon it
Floor 1
Feature 01 NE-SW stone wall, 0.47 m wide. The rocks that form the wall are quarried rectangular ones that are
joined with clay. This wall sits on the surface of
Layer 4 and an extension of 1.75 m was exposed and
was found in quads 3 and 4.
05 Clay floor found immediately under Layer 4. Found in much better conditions than the previous floor (Floor 1) exposed in the excavation of Layer 4.
Floor 2
06 Clay floor, similar as the preceding one Floor 3. Date AA75386 .
07 Semi-compact soil mixed with a low density deposit
of small and medium sized angular rocks and
cobbles
Pale brown (10YR 6/3)
08 Very compact soil mixed with a low density of small angular rocks
Brown (10YR 3/4)
09 Coarse sand mixed with abundant large sized cobbles
Light brownish gray (10YR 6/2)
Sterile soil
Table 07: Stratigraphic summary of WQ2
5.1.4 WQ-03 (N 830, E 481)
This unit had an area of 16 m² (4 x 4m). The purpose of this unit was to sample the
central north section of the Wacheqsa sector. The excavations in this unit reached a depth of
3.40 m below surface and each square meter was given a number in order to keep track of the
findings and/or contexts (figures 40-43).
Layer Description Color Notes 01 Matrix mixed with small rocks. At a depth of
0.50 m medium size rocks appear in an
irregular form.
Black (2.5Y, 2.5/1)
Aluvión slide of 1945.
02a Platform floor made of medium sized quarried rocks. The disposition of the rocks on the
surface has a formal arrangement
Reddish (2.5 YR 4/3)
02b Very compact clayish soil Dark greenish
gray (10GY 4/1)
Table 08: Stratigraphic summary of WQ3
68
Layer Description Color Notes Feature 1
Stone wall composed by large sized angular rocks. Same wall that appears in WQ-01
(Feature 1) and WQ-2 (Feature 1). It has a
north-south orientation; it defines the eastern
boundary of the stone platform that appears in
the west half of the unit (layer 2a). The wall has
a doorway located at the south end of the wall.
03 Floor associated at Feature 1 04 Semi compact matrix mixed with medium and
small sized angular rocks. Reddish brown (2.5Y 5/4)
Floor 1
05 Clayish matrix mixed with medium sized cobbles and small fragments of slate
Dark brown (7.5 YR 3/3)
06 Coarse sand mixed with abundant large sized cobbles
Light brownish gray (10YR 6/2)
Sterile soil
Table 08 (continuation): Stratigraphic summary of WQ3
5.1.5 WQ-4 (N 821, E 466)
This unit was located at 14 m southwest of WQ-3, and equally had an area of 16 m² (4
x 4 m). Each square meter was given a number in order to keep track of the findings and/or
contexts. This unit reached a total depth of 4.40 m below surface (figures 44-47).
Layer Description Color Notes 01 Matrix mixed with small rocks. At a depth of 0.50
m medium size rocks appear in an irregular form. Black (2.5Y, 2.5/1)
Aluvión slide
of 1945.
Excavated in
the entire
unit. 02 Platform floor made of medium sized quarried
rocks. The disposition of the rocks on the surface
has a formal arrangement
The soil between the
rocks was
reddish (2.5
YR 4/3)
Excavated in the entire
unit.
Covered
Feature 1 03 Very compact clayish soil Dark greenish
gray (10GY
4/1)
Present in quads 13,
14, 15 and
16. Feature 01 Rectangular room partially excavated; only the
southwestern section was found. Its walls were built
with middle sized quarried rectangular stones, the
width of the walls ranged between 0.6 m. and 0.7 m.
The N-S wall had three superimposed row of stones while the W-S had only two, these stones were
joined together with mud. The room was totally
covered by Layer 2.
Located in quads 3, 4, 7, 8 and 11
and 12
Table 09: Stratigraphic summary of WQ4
69
Layer Description Color Notes Feature 02
Stone wall composed by middle sized rectangular quarried stones joined together with mud mortar.
Located on the south
profile of the
unit. 04 Clayish compact floor. Dark reddish
brown (5YR
3/2)
Floor 1. Only excavated in
quads 6, 10,
15 and 16. 05 Semi compact matrix mixed with medium and small
sized angular rocks. Reddish brown (2.5Y 5/4)
Only excavated in
quads 6, 10,
15 and 16 05a Loose clayish soil mixed with a low quantity of
small sized angular rocks Dark brown (10YR 3/3)
Only excavated in
quad 6. 06 Clayish and sandy semi compact soil mixed with
abundant small sized cobbles and angular rocks Reddish brown (2.5 YR 4/3)
Only excavated in
quads 6, 15
and 16. 07 Medium and small sized angular rocks and cobbles
mixed with loose fine grain sandy soil Brown (10YR 3/4)
Only excavated in
quads 15 and
16. 08 Very compact clayish soil mixed with fine sand and
small sized angular rocks Dark brown (10YR 2/2)
Excavated in quads 6, 15
and 16. Feature 03
Large boulder sitting on top of Layer 8 and covered by Layers 4-7
Black (2.5Y, 2.5/1)
This feature
extends to
quads 10, 11,
14 and 15. Feature
04 Burnt soil immediately next to Feature 1 Found in
quad 6. 09 Very compact clayish soil. No rocks of any kind
were found Reddish brown (2.5Y 5/4)
Only excavated in
quads 15 and
16 10 Loose soil mixed with abundant small sized angular
rocks and a small amount of medium sized angular
rocks
Light gray (2.5Y 7/2)
Only
excavated in
quads 15 and 16
11 Semi compact clayish soil mixed with coarse grain sand and abundant small sized angular rocks.
Reddish brown (2.5Y 5/4)
Only
excavated in
quads 15 and
16 12 Loose soil, mixed with fine grain sand and abundant
small sized angular rocks. Dark reddish brown (2.5YR
2.5/3)
Only
excavated in
quads 15 and 16
13 Coarse sand mixed with abundant large sized cobbles
Light brownish gray (10YR 6/2)
Sterile soil
Table 09 (continuation): Stratigraphic summary of WQ4
70
5.1.6 WQ-5 (N746, E432)
This unit had an area of 4 m² (2 x 2) and was located at the bottom of the massive wall
that marks the south border of the Wacheqsa sector. The purpose of this unit was to
understand the nature of the deposits that lie under the wall in order to relate them with the
rest of the stratigraphy recorded in the Wacheqsa sector. The excavation in this unit reached a
depth of 2.27 m below surface (figure 48).
Layer Description Color Notes 01 Matrix mixed with small rocks. At a depth of 0.50 m
medium size rocks appear in an irregular form. Black (2.5Y, 2.5/1)
Aluvión slide of 1945.
02 Loose soil mixed with small angular rocks Reddish brown (2.5
YR 5/4).
03 This layer was located exclusively at the bottom of the southeast section of the massive wall. It had a
semicircular shape of approximately 0.30 m of
diameter. There was the presence of modern
materials (fragments of plastic bags) on the surface
as well as remains of burnt animal bones (deer?).
The surface of the layer has a reddish color as if was
the subject of fire.
The west boundary of this
layer is marked
by Feature 1.
Feature 1
Double face north-south wall of 1.20 m long and
0.40 m wide, located perpendicularly to the massive
east-west wall. This wall separated layers 3a and 3b.
The wall was build using
rectangular and
quadrangular
stones. 03a Very compact clayish soil Dark brown
(7.5 YR 3/2)
04 Loose soil matrix with abundant small sized angular rocks
Yellowish brown (10.5
YR 5/4)
05 Organized massive stone fill. Vertical lines of
medium sized cobbles separated massive fills
composed by small sized cobbles and angular rocks.
The spaces in between the vertical accommodations of cobbles were of 0.25 m in average
The excavation
stopped due to
the instability of
the profiles.
Table 10: Stratigraphic summary of WQ5
5.1.7 WQ-6 (N786, E430)
This unit is located at the west section of the Wacheqsa sector and had an area of 4 m²
(2 x 2 m). The purpose of this unit was to understand the stratigraphic depositions in this
particular area in relation with the other units excavated. The excavation in this unit reached a
depth of 3.40 m below surface (figure 49).
71
Layer Description Color Notes 01 Matrix mixed with small rocks. At a depth of
0.50 m medium size rocks appear in an
irregular form.
Black (2.5Y, 2.5/1) Aluvión slide of 1945.
02 Very compact matrix. Reddish (2.5Y 5/4) 03 Very compact matrix. Greenish (5Y 6/4) 04 Loose matrix Greenish (5Y 6/4) 05 Semi compact matrix mixed with medium
sized angular rocks. Olive brown (2.5Y 4/3)
06 Semi compact matrix mixed with medium quantities of small sized angular rocks.
Olive gray (5Y 4/2)
07 Very compact matrix mixed with a few small angular rocks and cobbles.
Dark reddish brown (2.5YR 5/3)
08 Loose matrix mixed with abundant small sized angular rocks.
Reddish (2.5Y 5/4)
09 Semi compact matrix with a low density of small sized angular rocks.
Reddish (2.5Y 5/4)
10 Very loose soil mixed with abundant small sized angular rocks.
Dark brown (10YR 3/3)
11 Very loose soil mixed with abundant small sized angular rocks.
Grayish brown (10YR 4/2)
12 Abundant medium sized angular rocks mixed
with a loose matrix. Very dark gray (10YR
3/1)
13 Semi compact matrix mixed with abundant small sized angular stones.
Dark reddish brown (2.5YR 5/3)
It surrounded Feature 1.
Feature 3
Elongated fireplace, reddish surface, abundant charcoal.
Date GX-
31647 was
obtained
from this
feature. 14 Coarse sand mixed with abundant large sized
cobbles Light brownish gray (10YR 6/2)
Sterile soil
Table 11 Stratigraphic summary of WQ6
5.1.8 WQ – 07 (N771, E434)
This unit was located at the south central section of the Wacheqsa sector and had an
area of 400 m². The unit was divided and numbered clockwise into four sectors starting from
the northwest one. This unit was systematically sampled, placing 2 x 1 units every five meters
within each sector. When required the units were expanded (figure 50).
5.1.8.1 SECTOR I
5.1.8.1.1 Unit 1 (N777, E435)
This unit was excavated until reaching 5.10 m below surface and had an area of 2 m²
(2 x 1 m) with a west-east orientation (figures 51-53).
72
Layer Description Color Notes 01 Matrix mixed with small rocks. At a depth of 0.50 m
medium size rocks appear in an irregular form Black (2.5Y, 2.5/1)
It was deposited by
the natural
aluvión slide of
1945.
Excavated in
the entire unit. 02 Dry and compact clayish matrix mixed with small size
rocks. It has several micro layers deposited
homogeneously one above each other given the
impression of a layer originated by water
sedimentation, produced either by a low gravity canal
or by water accumulation over time
Very dark gray
(Gley 1 3N) Unevenly
distributed,
located only at
the east and
south sides of
the unit 03 Semi compact coarse sand Yellowish (10YR
5/5) Similar distribution
than previous
layer. 04 Similar than layer 02 05 Similar than layer 03 06 Similar to layers 02 and 04 Very dark gray
(Gley 1 3N)
07 Loose clayish matrix. Gray (10YR 5/1) Its distribution and orientation
is similar to
that of layer
two, three, four,
five and six. 08 Compact matrix mixed with small angular rocks and
cobbles. Dark brownish
(10YR 4/2) Agricultural
land. 09 Semi compact clayish matrix mixed with small sized
angular rocks and cobbles. Dark reddish
brown (5YR 3/3)
10 Stone platform. Surface made of large sized cobbles arranged with their flat faces up. The interior of the
platform was made of a clayish matrix mixed with
small and medium sized angular rocks.
Olive (5Y 5/3)
11 Rocky layer, composed by medium sized angular rocks. In several cases the rocks show signs of
decomposition giving the layer an orange-reddish
coloration. The soil among the rocks was wet coarse
sand
Orange-reddish (10YR 4/3)
A block of crude limestone
was found in
this layer,
weighing 3 kg 12 Semi compact matrix mixed with large sized angular
rocks Dark yellowish
brown (10YR
3/4)
The excavation
had to be
stopped at this
level as the
profiles were
not stable
enough.
Table 12: Stratigraphic summary of WQ7-SI-U1
73
5.1.8.1.2 Unit 4 (N763, E441)
This unit was excavated to 3.50 m below surface and had an area of 2 m² (2 x 1 m)
with a west-east orientation.
Layer Description Color Notes 01 Matrix mixed with small rocks. At a depth of 0.50
m medium size rocks appear in an irregular form Black (2.5Y, 2.5/1) Aluvión slide
of 1945.
Excavated in the entire unit.
02 Semi compact clayish soil mixed with a low density of angular rocks of different sizes
Reddish brown (5Y 5/3)
Agricultural Land
03 Loose matrix mixed with abundant middle sized cobbles and angular rocks
Grayish brown (2.5Y 5/2)
04 Compact matrix mixed with abundant middle sized cobbles and angular rocks.
Grayish brown (2.5Y 5/2)
05 Rocky layer composed by abundant middle sized angular rocks mixed with a compact matrix.
Dark gray (10YR 4/1)
06 Semi compact clayish matrix mixed with abundant small cobbles and angular rocks
Grayish brown (10YR 5/2)
07 Compact matrix mixed with abundant small and middle sized angular rocks
Grayish brown (10YR 4/2)
An 80 kg block of
crude
limestone
was
recovered 08 Very compact clayish matrix mixed with small
sized angular rocks Dark gray (2.5Y 4/1)
09 Similar to layer 6 10 Stone platform. Surface made of large sized cobbles
arranged with their flat faces up. The interior of the
platform was made of a clayish matrix mixed with
small and medium sized angular rocks.
Olive (5Y 5/3) Similar to layer 10 in
unit 1. The
excavation
was stopped
at this layer
due to the
instability of
the profiles.
Table 13: Stratigraphic summary of WQ7-SI-U4
5.1.8.2 SECTOR II
5.1.8.2.1 Unit 1 (N770, E444)
This unit had an area of 6 m² (3 x 2 m) with a west-east orientation. This unit was
excavated to a depth of 3.40 m below surface. In order to have a better control of the
stratigraphy and features to be found in the unit, the unit was subdivided in 6 quads of one
meter each (figures 54 and 55).
74
Layer Description Color Notes 01 Matrix mixed with small rocks. At a depth of 0.50
m medium size rocks appear in an irregular form Black (2.5Y, 2.5/1)
Aluvión slide of 1945.
Excavated in
the entire unit. 02 Semi compact clayish soil mixed with a low
density of angular rocks of different sizes Reddish brown (5Y
5/3)
This layer was
excavated in
all quads. Feature 1 N-S wall, with a flat face toward the west. The wall
was built using medium sized cut stones This Feature
was found in
quads 2, 3 and
5. Feature 2 W-E wall composed by small sized angular rocks
lacking the formality of Feature 1. This wall joins
the southern end of Feature 1 forming a corner
This Feature
was found in
quads 1 and 2 03a Dried semi compact matrix mixed with angular
rocks of different sizes mixed with a small amount
of cobbles.
Dark brownish gray
(2.5Y 4/2)
Located towards the
east of Feature
1 03b Similar to layer 3a but with a major amount of
rocks Dark brownish gray
(2.5Y 4/2)
Located towards the
west of
Feature 1 04 Loose clayish matrix mixed with small sized
angular rocks Very dark grayish brown
(10YR 3/2)
Only excavated in
quads 1 and 2 05 Large amount of small angular rocks, loose
consistency and easy to excavate Olive brown (2.5Y 4/3)
Only
excavated in
quads 1 and 4. 06 Clayish matrix mixed with angular rocks of various
sizes mixed with small sized cobbles. Rocks of
larger size are located on the two upper thirds of
the layer.
Dark grayish brown (2.5Y
3/2)
Only
excavated in
quads 1 and 4.
07 Semi compact matrix mixed with angular rocks of
different sizes. Very dark
grayish brown (2.5Y 3/2)
Only
excavated in
quads 1 and 4. 08 Semi compact matrix mixed with abundant
medium sized angular rocks and a little amount of
small sized cobbles
Dark brown
clayish wet
soil (7.5YR
3/2)
Only
excavated in
quads 1 and 4.
09 Similar to the previous layer but with a major amount of cobbles
Dark grayish brown
The excavation
was stopped at
this layer due
to the
instability of the profiles.
Table 14: Stratigraphic summary of WQ7-SII-U1
75
5.1.8.2.2 Unit 4 (N764, E 451)
This unit was excavated reaching 3.20 m below the surface and had an area of 4 m² (2 x 2 m).
The unit was divided into four quads in order to have a better understanding of the
stratigraphy and features to be recorded.
Layer Description Color Notes 01 Matrix mixed with small rocks. At a depth of 0.50
m medium size rocks appear in an irregular form Black (2.5Y, 2.5/1)
Aluvión slide of 1945.
Excavated in
the entire unit. 02 Semi compact clayish soil mixed with a low density
of angular rocks of different sizes Reddish brown (5Y
5/3)
This layer was
excavated in
all quads. 03 Loose matrix mixed with abundant middle sized
cobbles and angular rocks Grayish brown (2.5Y
5/2)
Only excavated in
quads 1 and 2 04 Compact matrix mixed with abundant middle sized
cobbles and angular rocks. Grayish brown (2.5Y 5/2)
Only
excavated in
quads 1 and 2 05 Very compact clayish matrix. Very dark
gray (5YR
3/1)
Only
excavated in
quads 1 and 2 06 Loose matrix mixed with medium sized angular
rocks Dark brown (7.5 YR 4/2)
Only
excavated in
quads 1 and 2 07 Very compact matrix mixed with angular rocks and
cobbles of different sizes Reddish gray (5 YR 5/2)
Only excavated in
quads 1 and 2 08 Very compact matrix mixed with angular rocks of
different sizes Brown (10YR 4/3)
Only
excavated in
quads 1 and 2.
The
excavation was stopped
due to the
instability of
the profiles.
Table 15: Stratigraphic summary of WQ7-SII-U2
5.1.8.3 SECTOR III
5.1.8.3.1 Unit 1 (N760, E 435)
Unit with an area of 2² m (2 x 1 m), it had a west-east orientation and was excavated until
reaching a depth of 2.7 m below surface.
76
Layer Description Color Notes 01 Matrix mixed with small rocks. At a depth of
0.50 m medium size rocks appear in an irregular
form
Black (2.5Y, 2.5/1) Aluvión slide
of 1945.
Excavated in the entire
unit. 02 Semi compact clayish soil mixed with a low
density of angular rocks of different sizes Reddish brown (5Y 5/3)
03 Loose matrix mixed with abundant middle sized cobbles and angular rocks
Grayish brown (2.5Y 5/2)
04 Compact matrix mixed with abundant middle sized cobbles and angular rocks.
Grayish brown (2.5Y 5/2)
05 Loose matrix mixed with abundant large sized angular rocks
Dark brown (2.5Y 4/2)
06 Similar than previous but with a different coloration
Olive brown (2.5Y 4/3)
07 Large sized angular rocks (4) and large sized
cobbles (3) concentrated in the west section of
the unit mixed with coarse sand.
Grayish dark brown (2.5Y 4/2)
There was not formal
organization
of these
elements. 08 Compact matrix mixed with large sized angular
rocks. Dark grayish brown (2.5Y 3/2)
09 Compact matrix mixed with decomposed cobbles
Brown (10YR 3/4) The excavation
was stopped
at this layer.
Table 16: Stratigraphic summary of WQ7-SIII-U1
5.1.8.3.2 Unit 2 (N758, E441)
This unit had an area of 4 m² (2 x 2 m). The excavation in this unit reached a depth of 1.10 m
below surface (figure 56).
Layer Description Color Notes 01 Matrix mixed with small rocks. At a depth of 0.50
m medium size rocks appear in an irregular form Black (2.5Y, 2.5/1)
Aluvión slide of 1945.
02 Semi compact clayish soil mixed with a low density of angular rocks of different sizes
Reddish brown (5Y 5/3)
03 Loose matrix mixed with abundant middle sized cobbles and angular rocks
Grayish brown (2.5Y 5/2)
This layer was sitting
on top of
Feature 1 Feature 01
Stone platform floor made of middle and large sized cut rocks.
The excavation
was stopped
at this
Feature
Table 17: Stratigraphic summary of WQ7-SIII-U2
77
5.1.8.3.3 Unit 4 (N753, E451)
Unit with an area of 2² m (2 x 1 m), it had a west-east orientation. The excavation in this unit
reached a depth of 3.0 m below surface (figure 57).
Layer Description Color Notes 01 Matrix mixed with small rocks. At a depth of 0.50
m medium size rocks appear in an irregular form Black (2.5Y, 2.5/1) Aluvión
slide of
1945. 02 Semi compact clayish soil mixed with a low
density of angular rocks of different sizes Reddish brown (5Y 5/3)
03 Loose matrix mixed with abundant middle sized
cobbles and angular rocks Grayish brown (2.5Y
5/2)
04 Small angular rocks and cobbles mixed with dry coarse sand
Reddish brown (5YR 4/3)
Located in the central
section of
the unit 05 Loose fine grain sand mixed with small angular
rocks. Dark grayish brown color (10YR 4/2)
06 Very compact matrix mixed with coarse sand and a large amount of medium sized angular rocks.
Gray (7.5YR 5/1) Located towards the
west side of
the unit. 07 Compact clayish matrix mixed with coarse sand
and small and medium sized angular rocks. Dark grayish brown (2.5Y 4/2)
07a Similar than the previous layer but less compact. Dark grayish brown (2.5Y 4/2)
08 Loose matrix mixed with abundant large and small sized angular rocks.
Brown (10YR 4/3) Located towards the
east side of
the unit 09 Clayish compact matrix mixed with a large
amount of small and medium sized angular rocks Reddish brown (7.5 YR 4/4)
10 Semi compact clayish matrix mixed with coarse
sand and a large amount of medium sized angular
rocks.
Reddish brown (2.5 YR 4/3)
11 Semi compact matrix mixed with abundant medium sized angular rocks
Dark grayish brown (10YR 4/2)
12 Compact matrix mixed with coarse sand and abundant large sized angular rocks.
Black (5YR 2.5/1)
13 Semi compact matrix mixed with coarse sand and abundant small sized angular rocks
Reddish brown (5YR 5/4)
14 Loose matrix mixed with fine sand and middle sized angular rocks
Very dark brown (7.5 YR 2.5/2)
15 Similar to layer 12 Black (5YR 2.5/1) 16 Compact matrix mixed with abundant small sized
angular rocks Weak red (10YR 4/3)
17 Compacted fine sand Greenish gray (Gley1 5/10Y)
It was not excavated.
Table 18: Stratigraphic summary of WQ7-SIII-U4
78
5.1.8.3.4 Unit 4A (N756, E 440)
This unit had an area of 6² m (3 x 2 m). The excavation in this unit reached a depth of 4.20 m
below surface. In order to have a better control of the stratigraphy and features to be found in
the unit, the unit was subdivided in 6 quads of one meter each (figures 57-60).
Layer Description Color Notes 01 Matrix mixed with small rocks. At a depth of
0.50 m medium size rocks appear in an
irregular form
Black (2.5Y, 2.5/1) Aluvión slide of 1945..
02 Semi compact clayish soil mixed with a low density of angular rocks of different sizes
Reddish brown (5Y 5/3)
03 Loose matrix mixed with abundant middle sized cobbles and angular rocks
Grayish brown (2.5Y 5/2)
05 Loose fine grain sand mixed with small angular rocks.
Dark grayish brown (10YR 4/2)
06 Very compact matrix mixed with coarse sand and a large amount of medium sized angular
rocks
Gray (7.5YR 5/1) This layer was recorded in all
quads 07 Semi compact clayish matrix mixed with small
cobbles and medium sized angular rocks. Dark grayish soil (2.5Y 4/2)
This layer was
recorded in all
quads 07b Similar to the previous layer but found below
layers 8, 10 and 13. Dark grayish soil (2.5Y 4/2)
Found in quads 1, 2, 3
and 4. 08 Compact coarse sand mixed with abundant
large sized angular rocks. Brown (10YR 4/3) Recorded in
quads 1, 2, 5
and 6. Date AA 75389 was
taken from this
layer. 09 Clayish matrix mixed with abundant medium
sized angular rocks Reddish brown (7.5YR 4/4)
It was recorded in
quads 3, 4, 5
and 6. 10 Semi compact clayish matrix mixed with fine
sand and a few small sized angular rocks Reddish brown (2.5YR 4/3)
It is located in quads 2, 3 and
5. 10a Compact clayish matrix mixed with fine sand
and gravel. Olive brown (2.5YR
4/3) It was located t
in quads 2, 4 and 6.
11 Loose matrix mixed with abundant large and small sized angular rocks.
Brown (10YR 4/3) It was located in quads 2, 4
and 6. 12 Very similar to layer 7 Dark grayish soil
(2.5Y 4/2) It was
recorded in
quads 1, 3 and
5
Table 19: Stratigraphic summary of WQ7-SIII-U4A
79
Layer Description Color Notes 13 Loose fine sand. Gray (5Y 6/1) It was
recorded in
quads 3 and 5 14 Loose coarse sand Dark yellowish brown
(10YR 4/4) Recorded in quad 6
15 Loose matrix mixed with abundant large and small sized angular rocks. Similar to layer 11.
Brown (10YR 4/3) It was recorded in
quads 1, 3, 4, 5
and 6. 17 Compacted fine sand Greenish gray (Gley1
5/10Y) It was
recorded in all
quads 18 Compacted coarse sand Light brownish gray
(2.5Y 6/2) Located in quads 3, 4, 5
and 6 19 Loose matrix mixed with medium sized
angular rocks. It has an irregular circular
shape with an approximate diameter of 1 m
Brown (7.5YR 4/3)
20 Coarse compacted sand Light brownish gray (2.5Y 6/2)
Located in quads 1, 2
21 Compacted sand Greenish gray (Gley1 5/10Y)
Located in quads 3, 4, 5
and 6 22 Medium and large sized angular rocks mixed
with compacted sand. Greenish gray (Gley1 5/10Y)
Located in quads 3, 4, 5 and 6
23 Clayish semi compact matrix mixed with decomposed rocks and cobbles
Yellowish brown (10YR 5/6)
Located in quads 3, 4, 5
and 6 24 Coarse sand mixed with abundant large sized
cobbles Light brownish gray (10YR 6/2)
Sterile soil
Table 19 (continuation): Stratigraphic summary of WQ7-SIII-U4A
5.1.8.4 SECTOR IV
5.1.8.4.1 Unit 3 (N754, E445)
This unit had an area of 4 m² (2 x 2 m). The excavation in this unit reached a depth of 4.03 m
below surface. This unit was subdivided in four quads enumerated from west to east starting in
the NW quad (figure 61).
Layer Description Color Notes 01 Matrix mixed with small rocks. At a depth of
0.50 m medium size rocks appear in an
irregular form
Black (2.5Y, 2.5/1) Aluvión slide of 1945.
02 Semi compact clayish soil mixed with a low density of angular rocks of different sizes
Reddish brown (5Y 5/3)
Table 20: Stratigraphic summary of WQ7-SIV-U3
80
Layer Description Color Notes 03 Loose matrix mixed with abundant middle
sized cobbles and angular rocks Grayish brown (2.5Y 5/2)
04 Very compact soil mixed with small sized cobbles
Grayish brown (2.5Y 5/2)
Recorded in
quads 1 and 3 and
partially units
partially in quads
2 and 4. 05 Loose matrix mixed with large sized
angular rocks. Dark gray (10YR 4/1)
Recorded in units 2 and 4
06 Compact clayish surface. Brown (10YR 4/3) Located partially
in quads 1 and 2.
It was not
excavated. 07 Compact matrix with very small angular
rocks and coarse sand Grayish brown (10YR 5/2)
Located in quad 1
and partially in
quad 3 08 Compact matrix with fine sand, coarse sand
and small sized angular rocks Dark grayish brown (10YR 4/2)
Located in quads
2, 4 and partially
in quad 3. 09 Compact matrix with medium sized angular
rocks. Grayish brown (2.5Y 5/2)
Located in quads 1, 2, 4 and
partially in quad
3. 10 Compact matrix formed by fine sand and
coarse sand. There was a total absence of
rocks in this layer.
Dark gray (10YR 4/1)
Located partially in quad 3.
11 Compact matrix formed by coarse sand
mixed with abundant small sized angular
rocks
Grayish brown (10YR 5/2)
Located in quads
1, 2 and partially
in quads 3 and 4 12 Loose matrix formed by fine sand mixed
with coarse sand. It was similar to layer 8 Dark grayish
brown (10YR 4/2) It was partially
located in quad 4 13 Loose matrix mixed with abundant large
and small sized angular rocks. Brown (10YR 4/3) Located in quads
1, 3 and partially
in quads 2 and 4. 14 Clayish matrix formed by coarse sand
mixed with abundant gravel. Brown (7.5YR 5/2) Partially located
in quads 3 and 4. 15 Loose matrix formed by fine sand and
coarse sand mixed with gravel. In this layer
a line of medium size cobbles was
identified and left as a witness.
Brown (10YR 4/3) Present in all quads
16 Loose matrix composed by fine sand mixed with gravel.
Dark brown (2.5Y 6/2)
Partially present in quads 1, 2, 3
and 4. 17 Clayish matrix formed by fine sand mixed
with small sized angular rocks. Light brownish gray (10YR 6/2)
Partially present in quads 1, 2, 3 and 4.
Table 20 (continuation): Stratigraphic summary of WQ7-SIV-U3
81
Layer Description Color Notes 18 Loose matrix formed by fine sand mixed with
gravel and medium sized angular rocks. Gray (2.5Y 5/1) Partially
present in
quad 3. 19 Compacted fine sand Greenish gray
(Gley1 5/10Y) Partially present in
quads 1, 2, 3
and 4. 20 Compacted fine sand. Gray (Gley1 5/N) Partially
present in
quads 1, 2, 3
and 4. 21 Compacted coarse sand Light brownish
gray (2.5Y 6/2) Partially
present in
quad 3. 22 Coarse and fine sand mixed with small sized
angular rocks. Brown (7.5YR 5/6) Partially
present in
quads 1, 2, 3
and 4. 23 Coarse sand mixed with small sized angular and
rocks and cobbles. Pale olive (5Y 6/2) Partially
present in
quads 1, 2, 3
and 4. 24 Deposit of fine sand mixed with gravel and
small angular rocks. This deposit had a gray coloration (5Y
6/1).
Partially present in
quads 1, 2, 3
and 4. 25 Clayish semi compact matrix mixed with
decomposed rocks and cobbles Yellowish brown (10YR 5/6)
Partially present in
quads 1, 2, 3
and 4. 26 Fine sand mixed with gravel. Pale brown (10YR
6/3) Partially present in
quads 1, 2, 3
and 4. 27 Coarse sand mixed with gravel. light brownish gray
(2.5Y 6/2) Partially present in
quads 1, 2, 3
and 4. 28 Coarse sand mixed with abundant large sized
cobbles Light brownish gray (10YR 6/2)
Sterile soil
Table 20 (continuation): Stratigraphic summary of WQ7-SIV-U3
5.1.8.4.2 Unit 4 (N754, E441)
This unit had an area of 4² m (2 x 2 m). The excavation in this unit reached a depth of 4.20 m
below surface (figure 62).
82
Layer Description Color Notes 01 Matrix mixed with small rocks. At a depth
of 0.50 m medium size rocks appear in an
irregular form
Black (2.5Y, 2.5/1) Aluvión slide of 1945.
02 Semi compact clayish soil mixed with a low density of angular rocks of different
sizes
Reddish brown (5Y 5/3)
Excavated in the entire unit.
03 Loose matrix mixed with abundant middle
sized cobbles and angular rocks Grayish brown (2.5Y
5/2) Excavated in the
entire unit. 04 Clayish matrix mixed with small and
medium sized angular rocks. Dark grayish brown (2.5Y 4/2)
Excavated in the entire unit.
05 Semi compact deposit formed by coarse
sand mixed with gravel and medium sized
angular rocks.
Very dark gray (5YR 3/1)
Located in quad
1, partially
located in quads 2
and 3. 05a Compact clayish matrix without rocks of
any type. Light brownish gray (2.5Y 6/2)
Partially located in quads 1 and 2.
06 Semi compact matrix formed by gravel mixed with fine sand and medium sized
angular rocks.
Brown (7.5YR 4/2) Partially present in quads 2, 3 and
4. 06a Semi compact matrix mixed with medium
sized angular rocks. Light brownish gray (10YR 6/2)
Partially located in quad 2.
07 Loose matrix mixed with abundant presence of angular rocks.
Reddish gray (5YR 5/2) Partially located in quads 2 and 4.
08 Loose sandy matrix, mixed with gravel. Reddish gray (2.5YR 6/1)
Located in quad 1, partially
located in quads 2 and 3.
09 Semi compact matrix mixed with small sized angular rocks
dark gray (5YR 4/1) Located in quads
1, 2 and partially
in quads 3 and 4. 10 Loose matrix mixed with abundant large
and small sized angular rocks. Brown (10YR 4/3) Located in quads
1, 2 and partially
in quads 3 and 4.
Date AA75382
was taken from
this layer 11 Very compact clayish matrix mixed with
fine sand. Gray (7.5YR 5/1) Located in quad 4
and partially in
quad 3. 12 Compact clayish mixed with a small
amount of medium sized angular rocks. Gray (10YR 6/1)
13 Semi compact matrix, mixed with abundant angular rocks of various sizes
mixed with coarse sand
Light brownish gray (10YR 6/2)
14 Loose matrix formed by coarse sand mixed with medium and large sized angular rocks.
Light brownish gray (10YR 6/2)
Date AA75384
was taken from
this layer
Table 21: Stratigraphic summary of WQ7-SIV-U4
83
Layer Description Color Notes 15 Compacted fine sand Gray (Gley 1
5/10Y)
16 Loose matrix mixed with small sized angular rocks. Dark brown (7.5YR 3/3)
17 Fine sand mixed with abundant gravel. Gray (2.5Y 6/1) 18 Similar to layer 15. Greenish gray
(Gley 1 5/10Y)
19 Semi compact clayish matrix located in the central
portion of the unit, below layer 16; it had an
irregular circular shape.
Dark brown (7.5YR 3/3)
20 Fine sand mixed with gravel Grayish green (Gley 1 4/10Y)
Date AA75383
was taken
from this
layer 21 Coarse sand mixed with abundant large sized
cobbles Light brownish gray (10YR 6/2)
Sterile soil
Table 21 (continuation): Stratigraphic summary of WQ7-SIV-U4
5.1.9 WQ-8 (N794, E453)
This unit had 16m² (4 x 4 m) and it was located 19 m north of WQ-7. In order to have
a better control of the stratigraphy and features to be found in the unit, the unit was subdivided
in 16 quads of one meter each (figure 63).
Layer Description Color Notes
01 Matrix mixed with small rocks. At a depth
of 0.50 m medium size rocks appear in an
irregular form
Black (2.5Y,
2.5/1) Aluvión slide of
1945. Excavated in
the entire unit. 02 Platform floor made of medium sized
quarried rocks. The disposition of the rocks
on the surface has a formal arrangement
The soil between
the rocks was
reddish (2.5 YR
4/3)
Present in all quads.
It covered a set of
stone structures.
03 Very compact clayish soil Dark greenish gray (10GY 4/1)
Located in quads 1, 2, 3, 5, 6, 7, 8, 9, 10,
11 and 12 but
excavated only from
quad 5 to quad 12.
In these latter quads a passage associated
to a set of stone
structures was
recorded. Date
AA75390 was taken
from this layer.
Table 22: Stratigraphic summary of WQ8
84
Layer Description Color Notes 04 Clayish compact floor associated to the
architecture exposed in this unit Dark reddish brown (5YR
3/2)
Floor 1. Until this layer, the unit was
excavated in its
entirety. A 1m² cut
was excavated in
quad 9 in order to
explore the deposits
under the floor. Feature 1 W-E wall recorded in quads 1, 2, 3 and 4.
This wall had an average with of 0.55 m
and crossed the entire W-E axis of the unit.
It had a double face and was built using
rectangular cut stones.
Feature 2 W-E wall joined to the south face of Feature 1. This wall was an addition to Feature 1. Feature 2 had an average with of
0.8 m. It was made of similar materials
than Feature 1.
This feature was present in quads 5, 6, 7 and 8.
Feature 3 Platform located in quads 13, 14, 15 and 16. It was delimited in its north edge by a
set of nicely cut rectangular stones.
Towards its west half (quads 13 and 14) the
surface is composed by the stones that
form the platform while on the east half
(quads 15 and 16) there was clay floor.
This platform was not excavated
Feature 4 Rectangular cist (0.52 x 0.6 m) located in quads 11 and 12. This cist was built next to
Feature 3, using the platform face as its
south wall. The cist lacked formality and
was built with reused cut stones similar to
the ones that formed Features 1 and 2.
Feature 5 Small N-S wall located in quad 4 and 6. This wall was 1.27 m long and 0.53 m
(average). It was joined to the north face of
Feature 1 and looked like it was retaining
an unexcavated fill located at its east side
05 Loose clayish deposit mixed with gravel. Dark reddish brown (5YR
3/2)
Floor 1, associated
Only excavated in
quad 9 06 Medium and large sized cobbles and
angular rocks mixed with coarse sand. This
layer had a moist texture.
Dark reddish (2.5YR 3/6)
Only excavated in quad 9
07 Wet clayish matrix mixed with small angular rocks and gravel.
Dark reddish brown (2.5 YR
3/4)
Only excavated in quad 9
08 Clayish compact matrix. Olive (5Y 4/3) Only excavated in quad 9
09 Coarse sand mixed with abundant large sized cobbles
Light brownish gray (10YR
6/2)
Sterile soil
Table 22 (continuation): Stratigraphic summary of WQ8
85
5.1.10 WQ-09 (N799, E509)
This unit had an area of 4m² (2 x 2) and was located at the back of Marino Gonzales´s
house at the eastern side of the Wacheqsa sector. This unit was excavated until reaching 3.04
m below surface
Layer Description Color Notes 01 Matrix mixed with small rocks. At a depth of 0.50
m medium size rocks appear in an irregular form Black (2.5Y, 2.5/1)
It was deposited by
the natural
aluvión slide
of 1945.
Excavated in
the entire unit. 02 Very compact deposit formed by a clayish matrix
mixed with coarse sand and medium sized angular
rocks and cobbles.
Brown (10YR 4/3)
It covered a
partially
destroyed
stone floor. Feature 1 Stone floor located at the south section of the unit.
Composed by medium sized flat angular rocks, that
had an average width of 2 cm
03 Loose coarse sand, small and medium sized cobbles and gravel.
Dark grey (2.5 Y 4/1)
Located towards the
north portion
of the unit 04 Coarse sand mixed with abundant large sized
cobbles Light brownish
gray (10YR 6/2)
Sterile soil
Table 23: Stratigraphic summary of WQ9
5.1.11 WQ-10 (N751, E438)
This unit had an area of 6 m² (6 x 1). It was a trench located immediately south of
WQ-7, perpendicular to the wall that marks the north end of the Wacheqsa sector. The main
purpose of this unit was to establish the stratigraphic relationships between the deposits
uncovered in sectors III and IV of WQ-7. In order to have a better control of the architecture
and deposits to be recorded, the unit was divided in 6 quads of one meter each starting from
the northern quad.
Layer Description Color Notes 01 Matrix mixed with small rocks. Black (2.5Y,
2.5/1) Aluvión slide of 1945.
Feature 1 W-E wall located in quad 2. This wall was built
using middle sized cut stones as well as middle
sized boulders.
Table 24: Stratigraphic summary of WQ10
86
Layer Description Color Notes 02 Semi compact clayish soil mixed with a low
density of angular rocks of different sizes Reddish brown (5Y 5/3)
Located immediately at
the north of
Feature 1 in
quads 1 and 2 03 Loose matrix mixed with abundant middle sized
cobbles and angular rocks Grayish brown (2.5Y 5/2)
04 This is a massive fill formed by medium and large sized angular rocks and cobbles.
Located immediately at
the south of
Feature 1, in
quads 3 and 4. 05 Massive fill composed only by small and
medium sized cobbles. Larger cobles were
placed at the junction with layer 4 separating
both layers.
Located at the south of layer 4 and under
the wall that
delimits the
south edge of
the Wacheqsa
sector in quads
5 and 6 06 Dried compact matrix mixed with coarse sand. It
was located Gray (7.5YR 5/1)
Located at the north of
Feature 1, in
quads 1 and 2
under layer 3. 07 Clayish matrix mixed with medium sized
angular rocks and large sized cobbles. Very dark gray (10YR 3/1)
Located at the
south of
Feature 1,
under layers 4
and 5 in quads 3, 4, 5 and 6
08 Loose matrix mixed with medium sized angular rocks and large sized cobbles
Brown (10 YR 4/3)
09 Coarse sand mixed with large sized cobbles Light brownish gray (10YR 6/2)
The excavation had
to stop due to
the instability
of the profiles.
It could not be
confirmed if it
was the sterile
soil.
Table 24 (continuation): Stratigraphic summary of WQ10
Strata from Units WQ-5, WQ-9 and WQ-10 were not considered when calculating
densities, Boone index or kernel density estimations. The reason for this exclusion lies in their
characteristic. These units were located on architectural features such as the massive wall that
87
serves as a southern frontier of the Wacheqsa sector which had massive architectural sterile
loose fills made of large sized cobbles (WQ5 and WQ10) and on a stone floor constructed on
top of sterile soil (WQ9). These units were more informative regarding construction
techniques than regarding the social nature of the occupation in the Wacheqsa sector.
In total 200 layers and 23 features have been considered for analysis in the present
dissertation, recorded in a total excavated area of 132 m² (0.9% of the entire site).
This high level of stratigraphic complexity posed and interpretative challenge for the
elucidation of the nature of the archaeological record, the way the different components of this
record interact with each other and the nature of the social activities that originated the record.
These topics will be treated in the next chapter.
88
CHAPTER 6
INTRASITE COMPLEXITY
This chapter is divided in three sections. The first section provides an account of the
phases and spatial analytical units identified in the Wacheqsa Sector. The horizontal and
vertical spatial relationships within the analytical units identified were inferred from the
computer-aided design (CAD) stratigraphic model described in Chapter 4 and translated into a
Harris Stratigraphic Matrix. Each analytical unit has been recognized through examination of
the following strata characteristics: nature of the sediment, count density of the total
archaeological materials per m³, and density of each of the classes of archaeological materials
per m³ as stated in chapter 4; on the other hand phases have been assigned in the Wacheqsa
sector based upon the examination of superposition of strata, analytical units and ceramic
features present in each of them.
The second segment covers the results of Boone index calculations as measurements
of heterogeneity for deposits and analytical units as a whole. This measure has the potential to
reinforce the patterns of spatial segregation inferred from the analysis of deposits, volumes
and densities.
The third section of this chapter exclusively deals with the intrasite patterning of
ceramic vessels identified in all prehistoric analytical units, emphasizing size and thickness as
explained in chapter 4. The variability of ceramic form and size is notable between the
different spatial analytical units.
Also as mentioned in Chapter 4, each deposit has been codified with a correlative
number within each analytical unit identified. For example deposits 101, 102 …n belong to
the same analytical and deposits 201, 202 …n belong to another one.
6.1 Spatial Intrasite Complexity
The Wacheqsa Sector has been divided in three phases and eight analytical units,
which are described in detail in the following pages.
Dates Phase Analytical Units Modern era Modern Agricultural land, Modern Canal and Aluvión 800 -500 BC Janabarriu Midden, Stone Rooms, Late Platforms 1200 – 800 BC Urabarriu Early Platforms, Water Flood
Table 25: Chronological chart of the Wacheqsa Sector
89
6.1.1 Prehistoric Occupation
It is composed of two phases, five analytical units, seven features and 151
stratigraphic layers (figure 64). It covers the entire Wacheqsa sector and was formed over the
course of 700 years, from about 1200 to 500 BC (calibrated C14 years).
6.1.1.1 Urabarriu Phase
This phase is composed of two analytical units, containing 57 stratigraphic layers and
four features identified separately across the units excavated. It was identified in both the
north and south sections of the Wacheqsa sector. Both analytical units are spatially
differentiated, found in different excavation units and located on top of sterile soil; they
represent the earliest occupation in the Wacheqsa Sector. This phase spans from 1200-800 BC
(calibrated C14 years)
6.1.1.1.1 Water Flood
This analytical unit is located at the south end of the Wacheqsa Sector and has been
identified in WQ7-SIIIU4A, SIVU3 and SIVU4. It encompasses an estimated area of 48 m²
and an estimated volume of 46 m³. The area estimate has been calculated by adding the area of
units excavated with the area that is located between these units, and the estimated volume has
been calculated by multiplying the estimated area with the average bottom depth of the unit (I
followed the same procedure for further similar calculations). The average depth of this unit –
inferred from the depth of the surface of the upper stratum of each unit excavated- was 3.05 m
below surface. It had 20 strata distributed among the excavation units mentioned above:
Unit Layer Harris Matrix Code
WQ7SIII4A 19 300 WQ7SIV3 15 301 WQ7SIV4 20 302 WQ7SIV3 18 303
WQ7SIII4A 17 304 WQ7SIV4 22 305 WQ7SIV3 21 306
WQ7SIII4A 18 307 WQ7SIV4 23 308 WQ7SIV3 20 309
WQ7SIII4A 22 310 WQ7SIII4A 23 311 WQ7SIII4A 19 312 WQ7SIV3 24 313 WQ7SIV4 20 314
Table 26: Water Flood analytical unit strata
90
Unit Layer Harris Matrix Code WQ7SIV3 25 315 WQ7SIV3 26 316
WQ7SIII4A 24 317 WQ7SIV3 27 318 WQ7SIV4 21 319
Table 26 (continuation): Water Flood analytical unit strata
It was characterized by a succession of compact layers of grey and greenish sand
alternated with fine and coarse gravel. No concavity or structure containing these strata has
been found, hence I hesitate to name this analytical unit “Prehistoric Canal” or “Urabarriu
Canal”, on the other hand the characteristics of the strata recorded indicate the flowing of
water (as suggested by Tello in 1940), thus I have decided to label this as “Water Flood”.
At least three major flooding events or major water currents have been identified in this
analytical unit. Strata 300, 301, 302 are part of the same depositional episode; strata 304, 305,
306, are part of another depositional event, and strata 312, 313 and 314 conform to a different
depositional episode. The first major depositional event is identified as Water Flood 1 and
formed by layers 300, 301 and 302. The second major depositional event has been identified
as Water Flood 2 and the third one as Water Flood 3. The slope of these flooding events was
towards the east (figures 65 and 66). Among these principal depositional episodes there are
eight sub-flood events identified independently in the units excavated that could not have been
correlated as part of a larger event. In total, 5.72 m³ of this analytical unit were excavated,
recovering 330 artifacts with an average density of 57 artifacts per m³.
Material Total Recovered Density Animal bones 7.6 kg 1.33 kg/m³ Decorated ceramic sherds 66 11.54 sherds/m³ Diagnostic ceramic sherds 217 37.9 sherds/m³ Burnt Clay 22 3.85 fragments/m³ Obsidian 20 3.49 fragments/m³ Anthracite 0 0 fragments/m³ Shells 0 0 fragments/m³ Projectile points 1 0.2 fragments/m³ Worked bones 1 0.2 fragments/m³ Chrysocolla 0 0 fragments/m³ Lithics 9 2 fragments/m³
Table 27: Densities of archaeological materials from the Water Flood analytical unit
91
The ceramics identified in this analytical unit resemble the ones defined as Urabarriu
by Richard Burger (Burger 1984; Burger 1998), no Janabarriu related ceramic component was
present in these deposits.
6.1.1.1.2 Early Platforms
Located in the north and central sections of the Wacheqsa Sector and has been
identified in units WQ1, WQ2, WQ3, WQ4, and WQ6. It encompasses an estimated area of
1100 m² and an average depth of this unit was of 2.04 m below surface (figure 67). This
analytical unit has 57 stratigraphic layers and four features.
Unit Layer Harris Matrix Code WQ4 05 500 WQ6 04 501 WQ8 05 502 WQ4 5a 503 WQ6 05 504 WQ3 04 505 WQ4 04 506 WQ6 06 507 WQ8 06 508 WQ4 07 509 WQ3 05 510 WQ6 07 511 WQ4 F3 512 WQ4 F4 513 WQ6 08 514 WQ6 09 515 WQ4 08 516 WQ6 10 517 WQ4 09 518 WQ6 11 519 WQ4 10 520 WQ6 12 521 WQ4 11 522 WQ4 12 523 WQ6 13 524 WQ8 07 525 WQ8 08 526 WQ6 14 527 WQ4 13 528 WQ3 06 529 WQ8 09 539 WQ1 03 604 WQ1 04 605 WQ1 06 606 WQ1 05 607
Table 28: Early Platform analytical unit strata
92
Unit Layer Harris Matrix Code WQ2 03 608 WQ1 F3 609 WQ1 07 610 WQ2 04 611 WQ2 05 612 WQ1 08 615 WQ2 06 613 WQ2 07 614 WQ2 08 616 WQ1 09 617 WQ2 09 618
Table 28 (continuation): Early Platform analytical unit strata
It is characterized by a sequence of deposits of almost flat surfaces with small sized
angular rocks as fills and a low density of archaeological materials. In total 20.3 m³ were
excavated, recovering 978 artifacts (figures 68-72) with an average density of 48 artifacts per
m³.
Material Total Recovered Density Animal bones 6.8 Kg 0.33 kg/m³ Decorated ceramic sherds 266 13.10 sherds/m³ Diagnostic ceramic sherds 591 29.11 sherds/m³ Burnt clay 50 2.42 fragments/ m³ Obsidian 31 2 fragments/m³ Anthracite 1 0.1 fragments/m³ Shells 3 0.2 fragments/m³ Projectile points 0 0 fragments/m³ Worked bones 8 0.4 fragments/m³ Chrysocolla 0 0 fragments/m³ Lithics 45 2 fragments/m³
Table 29: Densities of archaeological materials from Early Platforms analytical unit
6.1.1.2 Janabarriu Phase
This phase is composed of three analytical units, 97 strata and 13 features. It is
distributed over the entire area and is the latest prehistoric occupation in the Wacheqsa Sector.
6.1.1.2.1 Midden
This analytical unit is located at the south end of the Wacheqsa Sector, on top of the
Water Flood analytical unit, and has been identified in units WQ7-SIIIU4, SIIIU4A, SIVU3
and SIVU4. It encompasses an estimated area of 48 m², and an estimated volume of 83 m³,
close to 2 m average thickness. The surface of this analytical unit is located at 1 m below
surface on average; it had 44 strata distributed among all units excavated. It is characterized
93
by a semi-compact to compact matrix mixed with middle and large sized angular stones and
cobbles and a high density of fragmented archaeological materials.
Unit Layer Harris Matrix Code
WQ7SIV4 4 112 WQ7SIII4 6 113 WQ7SIV3 6 114 WQ7SIV4 6 115 WQ7SIV4 6a 116 WQ7SIII4A 7 117 WQ7SIV3 7 118 WQ7SIV4 7 119 WQ7SIII4 4 120 WQ7SIII4 5 121 WQ7SIV3 8 122 WQ7SIV4 5 123 WQ7SIII4A 9 124 WQ7SIV3 10 125 WQ7SIV4 8 126 WQ7SIII4A 10 127 WQ7SIII4A 10a 128 WQ7SIV3 9 129 WQ7SIV4 9 130 WQ7SIV3 12 131 WQ7SIV4 11 132 WQ7SIV3 11 133 WQ7SIII4 7 134 WQ7SIII4A 8 135 WQ7SIV3 14 136 WQ7SIII4 7a 137 WQ7SIII4A 7b 138 WQ7SIII4 9 139 WQ7SIII4 8 140 WQ7SIII4A 11 141 WQ7SIII4A 12 142 WQ7SIV3 13 143 WQ7SIV4 10 144 WQ7SIII4 10 145 WQ7SIV3 15 146 WQ7SIV4 12 147 WQ7SIV4 13 148 WQ7SIV3 16 149 WQ7SIV3 18 150 WQ7SIII4A 14 151 WQ7SIII4A 15 152 WQ7SIV3 17 153 WQ7SIV4 14 154 WQ7SIII4A 19 155
Table 30: Midden analytical unit strata
94
As mentioned in Chapter 3, this Midden was identified by Julio C. Tello when he
excavated the Wacheqsa sector in 1940 (Tello 1960).
Three major depositional episodes have been identified in all units excavated. The
first and oldest depositional event is represented by strata 152, 153 and 154 and has a gradient
towards the east; this depositional event has been recognized as Midden I. A second major
depositional event has been identified as Midden II, also with a gradient towards the east,
represented in strata 140, 141, 142, 143 and 144.
The latest main depositional event is represented by Midden III, identified in strata
112, 113 and 114. As Midden I and II, Midden III has a gradient towards the east. Between
Middens I and II there are seven midden-type depositional events represented by strata 146,
147, 148, 149 and 150. Between Midden II and Midden III there are 25 small midden-type
depositional events represented in strata 115, 116, 117, 118, 119, 120, 121, 122, 123, 124,
125, 126, 127, 128, 129, 130, 131, 132, 133, 134, 135, 136, 137, 138 and 139. Middens I, II
and III have been identified on the grounds of similarities in the nature of the sediments
among the four units excavated. In other words they are the same stratum identified in
different units. Between each one of these large depositional events, small ones have been
located. It is necessary to mention that even though these small midden-type deposits were not
located in all units, there is the possibility that some of them are part of large depositional
events that have not been located in the sampled area (figures 73-74).
In total 22.03 m³ were excavated, recovering 15814 fragments of archaeological
materials, having a general density of 718.18 fragments per m³ (figures 75-81).
Material Total Recovered Density Animal bones 184.2 kg 8.27 kg/m³ Decorated ceramic sherds 4417 200.49 sherds/m³ Diagnostic ceramic sherds 8567 388.87 sherds/m³ Burnt clay 2136 96.96 fragments/ m³ Obsidian 554 25.15 fragments/ m³ Anthracite 138 6.26 fragments/ m³ Shells 6 0.27 fragments/ m³ Projectile points 15 0.68 fragments/ m³ Worked bones 62 2.81 fragments/ m³ Chrysocolla 15 0.68 fragments/ m³ Lithics 235 10.66 fragments/ m³
Table 31: Densities of archaeological materials
A striking characteristic is the high density of burnt clay recovered in this analytical
unit, 79 fragments per m³ which is 11 times more the density of the Stone Rooms analytical
95
unit, the analytical unit with the next highest density of burnt clay fragments. This class of
artifact represents fragments of floors, walls, columns and even a fragment of a molded frieze
(fig 78, 82 and 83). Most of these fragments have portions of vitrified surface, which mean
that they were exposed to high temperatures. In some cases fragments were plastered with
white and red color. The majority of fragments show the imprints of canes in their interiors,
suggesting that the construction technique used in the structures involved clay tempered with
straw and fired to a very hard consistency. The presence of these types of structures should
not be surprising, since in 1944. Bennett reported the existence of a structure made of “burnt
clay plaster walls” (Bennett 1944:77) located on top of Mound D, and Tello reported the
occurrence of these fragments on deposits located in front of the façade of Building A (Tello
1960).
6.1.1.2.2 Late Platforms
This analytical unit is located in the south central section of the Wacheqsa sector, and
has been identified in units WQ7-SIU1, SIU4, SIIU1, SIIU4, and SIIIU1. It encompasses an
estimated area of 223 m² with an estimated volume of 276 m³.
Unit Layer Harris Matrix Code WQ7SII1 3a 416 WQ7SIII1 3 417 WQ7SI4 3 418 WQ7SII4 3 419 WQ7SII1 3b 420 WQ7SIII1 4 421 WQ7SI4 4 422 WQ7SII4 4 423 WQ7SI4 5 424 WQ7SIII1 5 425 WQ7SII1 4 426 WQ7SII4 5 427 WQ7SII4 6 428 WQ7SIII1 6 429 WQ7SII1 5 430 WQ7SI4 6 431 WQ7SII4 7 432 WQ7SIII1 7 433 WQ7SII1 6 434 WQ7SI4 7 435 WQ7SII4 8 436 WQ7SIII1 8 437 WQ7SII1 7 438 WQ7SI4 8 439
Table 32: Late Platforms analytic unit strata
96
Unit Layer Harris Matrix Code WQ7SIII1 9 440 WQ7SII1 8 441 WQ7SI4 9 442 WQ7SI1 9 443 WQ7SI4 10 444 WQ7SII1 9 445 WQ7SIII1 10 446 WQ7SI1 10 447 WQ7SI1 11 448 WQ7SI1 12 449
Table 32 (continuation): Late Platforms analytic unit strata
It is situated between 1.5 m and 2 m below surface; it had had 34 strata distributed
among all the units excavated which are characterized by compact matrices mixed with
abundant middle and large sized angular stones and cobbles and a very low density of
fragmented archaeological materials. In total 5.88 m³ were excavated, recovering 102
fragments of archaeological materials, having a general density of 17.3 fragments per m³. This
analytical unit has the lowest density of archaeological materials.
Material Total Recovered Density Animal bones 3.21 kg 0.55 kg/m³ Decorated ceramic sherds 29 4.93 sherds/m³ Diagnostic ceramic sherds 35 5.95 sherds/m³ Burnt clay 1 0.17 fragments/ m³ Obsidian 3 0.51 fragments/ m³ Anthracite 7 1.19 fragments/ m³ Shells 0 0 fragments/ m³ Projectile points 1 0.17 fragments/ m³ Worked bones 1 0.17 fragments/ m³ Chrysocolla 0 0 fragments/ m³ Lithics 53 9.01 fragments/ m³
Table 33: Densities of archaeological materials
6.1.1.2.3 Stone Rooms
This analytical unit is located in the central section of the Wacheqsa sector, above the
Early Platforms analytical unit, and has been identified in units WQ3, WQ4, WQ6 and WQ8.
It encompasses an estimated area of 1717 m², and an estimated volume of 668.47 m³.
97
Unit Layer Harris Matrix Code WQ3 02a 204 WQ6 02 205 WQ8 02 206 WQ4 02 207 WQ3 02 208 WQ4 03 209 WQ8 03 210 WQ8 F4 211 WQ8 F3 212 WQ8 F2 213 WQ8 F5 214 WQ4 F1 215 WQ4 F2 216 WQ3 F1 217 WQ6 F2 218 WQ6 F1 219 WQ8 F1 220 WQ4 04 221 WQ3 03 222 WQ6 03 223 WQ8 04 224
Table 34: Stone Rooms analytic unit strata
This analytical unit is located between 1.8 m and 2 m below the surface; it had 10
stratigraphic layers and 12 features among all units excavated. Stone rooms and
interconnected alleys are part of the earliest occupation in this analytical unit. Alleys were
covered by fill made of loose matrix mixed with middle sized cobbles and angular stones.
Once alleys were filled, rooms and alley’s fills were covered by a platform almost entirely
made of quadrangular middle sized rocks mixed with scarce loose matrix. Rooms and alleys
were cleansed before being filled, archaeological materials recovered in this analytical unit
come from alleys and platforms fills (figures 84-89). There are two types of fills in this
analytical unit, the fill that covered the alleys, which is composed by a very compact clayish
matrix mixed with small medium sized angular rocks located on top of stone room’s floors;
and the platform fill that covered room and alley deposits, this fill was composed of medium
sized quarried angular stones.
In total 22.3 m³ were excavated, recovering 3425 fragments of archaeological
materials, giving an overall density of 154 fragments per m³
98
Material Total Recovered Density Animal bones 8.22 kg 0.37 kg/m³ Decorated ceramic sherds 1015 46.07 sherds/m³ Diagnostic ceramic sherds 2039 91.43 sherds/m³ Burnt clay 130 5.83 fragments/ m³ Obsidian 123 5.52 fragments/ m³ Anthracite 31 13.90 fragments/ m³ Shells 2 0.9 fragments/ m³ Projectile points 1 0.04 fragments/ m³ Worked bones 18 0.81 fragments/ m³ Chrysocolla 7 0.31 fragments/ m³ Lithics 54 2.42 fragments/ m³
Table 35: Densities of archaeological materials
6.1.2 Modern Phase
The Modern Phase has been extensively identified in all units excavated. It is
composed of three analytical units containing 35 stratigraphic layers.
6.1.2.1 Aluvión
This analytical unit is composed of 15 strata that share the same characteristics, which
is a dark gray matrix mixed with small rocks but with larger rocks at the bottom of each
stratum. This analytical unit was formed on January 17, 1945, and it is the product of a
landslide that covered the entire site of Chavín de Huántar as well as parts of the town. As
mentioned in Chapter 3, this natural disaster transformed the topography of the Wacheqsa
sector. Layers in this analytical unit were not sifted in order to improve excavation efficiency.
Unit Layer Harris Matrix Code WQ7SIII4 01 101 WQ7SIII4A 01 102 WQ7SIV3 01 103 WQ7SIV4 01 104 WQ3 01 200 WQ4 01 201 WQ6 01 202 WQ8 01 203 WQ7SI1 01 400 WQ7SI4 01 401 WQ7SIII1 01 402 WQ7SII4 01 403 WQ7SII1 01 404 WQ1 01 600 WQ2 01 601
Table 36: Aluvión analytical unit strata
99
6.1.2.2 Modern Canal
This analytical unit was only found in unit WQ7, SIU1. It was composed by six
superimposed strata (405, 406, 407, 408, 409 and 410). These layers were deposited in a
concavity that intruded into Layer 427. The concavity had an elongated shape 1.73 m long
oriented southwest/northeast with a gentle slope of five degrees downwards toward the
southwest. The layers deposited in the concavity were an alternation of very dark gray clayish
soil with semi compact yellowish coarse sand similar to the deposits of the Water Flood
analytical unit. The concavity has been identified as an irrigation canal and the layers as
evidence of seasonal episodes of water flow with their associated gravel sedimentation. The
layer intruded into (427) is part of the agricultural land that was subject to irrigation prior to
the landslide of 1945. Only 0.31 m³ of this analytical unit was excavated and 50 fragmented
objects were recovered.
Unit Layer Harris Matrix Code WQ7SI1 2 405 WQ7SI1 3 406 WQ7SI1 4 407 WQ7SI1 5 408 WQ7SI1 6 409 WQ7SI1 7 410
Table 37: Layers that are part of the Modern Canal analytical unit
6.1.2.3 Agricultural Land
This analytical unit is sitting on top of the prehistoric occupation of the Wacheqsa
Sector. It is composed of 11 strata, 105, 106, 107, 108, 411, 412, 413, 414, 415 602 and 603
independently identified in units WQ1, WQ2, WQ4, WQ5, WQ6, WQ7SIU1, WQ7SIU4,
WQ7SIIU1, WQ7SIIU4, WQ7SIIIU1, WQ7SIIIU4, WQ7SIIIU4A, WQ7SIVU3, WQ7SIVU4
and WQ8. The surface of this analytical unit represented the topographic configuration of the
Wacheqsa sector before the 1945 landslide. Excavations have confirmed the hypothesis that
the topography of the Wacheqsa sector was different from present times as seen in figure 90.
The section represented in this figure was constructed from the surface elevations of the
agricultural land strata identified across excavations. This analytical unit’s surface represents
the surface configuration of the Wacheqsa sector prior the 1947 landslide, used as agricultural
land and also as living places in the first half of the 20th
Century,
100
Currently, [November 18, 1940] maize is being harvested in this area.
Its surface shows signs of furrows left by plowing activities and a
multitude of small rocks, roots and maize roots are present. This
agricultural field is filthy and neglected (Tello 1940:27)
The following table shows general densities of archaeological materials per analytical
unit, each number is referred to a measure per cubic meter:
Analytical Unit DE DI O BC A S L PP WB C Midden 200 389 31 79 5 0.22 8 1 2 1 Stone Rooms 46 91 9 6 1 0.1 2 0.04 1 0.3 Late Platforms 5 6 0.3 0 0.7 0 6 0.1 0.1 0 Water Flood 12 38 4 4 0 0 2 0.2 0.2 0 Early Platforms 13 29 2 2 0.1 0.2 2 0 0.4 0
Table 39: Densities of archaeological classes per analytical unit, De, decorated ceramic sherds; Di,
diagnostic ceramic sherds; O, Obsidian fragments; BC, burnt clay fragments; A, anthracite; fragments;
S, shell fragments; L, lithic fragments; PP, projectile points; WB, worked bone fragments; C,
chrysocolla fragments
Finally, figure 91 shows the stratigraphic Harrix Matrix with all strata and features
organized according to the methods described in the previous chapter and figures 92 and 93
summarize the volume excavated per analytical unit as well as the density of materials
retrieved. .
6.2 Boone Index Measurement
As described in Chapter 4, I am suing this measurement to test the null hypothesis that
all spatial analytical units are composed by deposits that have similar Hi values. In order to do
that I have compared the individual provenience units (deposits) of artifacts with the
aggregated distribution of all deposits combined. This comparison allowed the creation of a
measure of diversity of all provenience units in relation to artifact classes. In this sense -- as
noted in the previous section -- ten classes of artifacts have been identified in the Wacheqsa
sector, all of which have been weighed using the Boone index.
This measure of diversity serves the purpose of observing if the deposits of the
analytical units previously identified form clusters along Hi values. Figure 94 shows clusters
of analytical units are formed along Hi values with bounding lines that encompass 50% of
deposits per analytical unit. The analytical unit with the lowest Hi value is the Midden while
the Late Platform value is the one with the highest Hi value. Early Platform and Water Flood
101
analytical units almost entirely overlap and show very similar behaviors with Hi values that
range from 0.2-0.85. The Stone Room analytical unit is tightly clustered along the 0.4 and
0.42 values partially overlapping the Midden analytical unit. At first look it is observed a clear
segregation of Hi values per analytical unit, but how much of this pattern is due to a sample
size bias?
Figure 95 plots the sample size of the five prehistoric analytical units, a careful look at
this plot compared with figure 96 shows that the difference of sample size among each
analytical unit is actually the segregation shown when calculating the Boone index. Analytical
units with small sample sizes have high Hi values while analytical units with high sample
sizes have low Hi values. As mentioned in Chapter 4 in, the first step taken in order to control
for sample size bias was to calculate a 90% confidence interval; strata outside the confidence
interval supposedly are anomalies to the expected pattern of sample size given by the
confidence interval (figure 97). The confidence interval gave me the parameters under which
strata could be categorized as an outlier taken into account sample bias. Still, I needed to make
sure that the clustering of analytical units observed when the Hi values were calculated was
not due to sample bias. For this reason I repeatedly sampled from the observed population
using a Monte Carlo routine in order to determine whether Hi values calculated could
reasonably be due to sample size bias. As explained in Chapter 4, Monte Carlo routines are
particularly useful for testing the significance of a test (in this case the Boone index). Using
this routine I drew 10,000 samples to see what the expectable range of variability would be,
based on the actual sample from which I was working
Figure 98 shows the probability of Hi values [p(Hi)] after the Monte-Carlo routine was
performed, the distribution of Hi are drastically different from those shown in figure 94, Hi
values are not similar to p(Hi) values. This confirms that that original Hi values were
generated by sample size but also indicates that in general terms all prehistoric analytical units
are highly diverse. Figure 99 shows logged distribution of p(Hi) values per unit. The general
mean and median p(Hi) values are 0.225 and 0.1275 respectively.
Two observations stand out. The first one is that the five prehistoric analytical units
have low p(Hi) values, and the second is that the Stone Rooms and Early Platforms have
significantly different p(Hi) values and the rest of analytical unit´s values are between them,
confirmed by a Student’s T test performed on p(Hi) values of each analytical unit as shown in
Tables 40-42.
102
t Alpha 1,98667 0,05
Abs(Dif)-LSD Late Platforms Water Flood Early Platforms Midden Stone Rooms Late Platforms -0,15680 -0,09393 -0,04499 -0,02606 0,03875 Water Flood -0,09393 -0,17689 -0,12982 -0,11148 -0,04396 Early Platforms -0,04499 -0,12982 -0,11734 -0,09700 -0,03820 Midden -0,02606 -0,11148 -0,09700 -0,09917 -0,04337 Stone Rooms 0,03875 -0,04396 -0,03820 -0,04337 -0,18553
Table 40: Student´s t test Least Significant Difference threshold matrix. Positive values show pairs of
means that are significantly different.
Analytical Unit Student’s T Value Mean Late Platforms A 0.28521429 Water Flood AB 0.21200000
Early Platforms AB 0.19172000 Midden AB 0.18008571
Stone Rooms B 0.07470000
Table 41: Student´s T test connecting lines report. Levels not connected by same letter are significantly
different
Level Level Difference Lower CL Upper CL p-Value Late Platforms Stone Rooms 0,2105143 0,038750 0,3822787 0,0168701 Water Flood Stone Rooms 0,1373000 -0,043961 0,3185612 0,1358649 Early Platforms Stone Rooms 0,1170200 -0,038203 0,2722428 0,1377046 Miden Stone Rooms 0,1053857 -0,043367 0,2541381 0,1627272 Late Platforms Midden 0,1051286 -0,026059 0,2363158 0,1148805 Late Platforms Early Platforms 0,0934943 -0,044987 0,2319752 0,1832020 Late Platforms Water Flood 0,0732143 -0,093934 0,2403623 0,3865037 Water Flood Midden 0,0319143 -0,111483 0,1753113 0,6594398 Water Flood Early Platforms 0,0202800 -0,129819 0,1703785 0,7889888 Early Platforms Midden 0,0116343 -0,096999 0,1202676 0,8319903
Table 42: Student´s T test ordered differences report
Stone Rooms’ analytical unit is the most diverse of the five, while the Late Platforms
unit is the least diverse of the five. Water Flood, Early Platforms and the Midden are located
between the ones mentioned before. The null hypothesis (all spatial analytical units are
composed by deposits that have similar Hi values) is therefore rejected. Even though all the
means of the analytical units are below the 0.3 value (and therefore suggesting high diversity
in their contents), within this diversity there are differences among these analytical units,
difference that are not due to sample size bias but rather to behavioral reasons will be
discussed in the next chapter.
103
6.3 Ceramic Intrasite Complexity
A subset of 3020 diagnostic sherds were used to generate univariate and bivariate kernel
density estimates of rim diameter and wall thickness in order to identify patterns of
distribution and variation of ceramic vessels within the spatial analytical units described in
previously. The following table addresses the distribution of the ceramic sample per spatial
analytical unit
Analytical Unit n Early Platform 118 Late Platform 42 Midden 2441 Room 317 Water Flood 102
Table 43: Distribution of sampled sherds
The next Table addresses the variation of ceramic forms per analytical unit
Analytical Unit Bowls OSC Jars Bottles Cups Plates Total Early Platform 36 58 16 5 0 2 118 Late Platform 14 20 6 1 1 0 42 Midden 1263 656 338 101 37 46 2441 Room 143 111 51 4 1 7 317 Water Flood 40 37 14 8 1 2 102 Total 1496 882 425 119 40 57 3020
Table 44: Distribution of ceramic types sampled per analytical unit
The next Table shows the same distributions showed above but with percentages,
Analytical Unit Bowls OSC Jars Bottles Cups Plates Early Platform
Late Platform
Midden
Room
Water Flood
30.51 49.15 13.56
33.33 47.62 14.29
51.74 26.87 11.19
45.11 36.27 16.09
39.22 36.26 13.73
5.08 0.00 1.69
2.27 2.38 0.00
3.34 1.52 1.88
1.26 0.32 2.21
7.84 0.98 1.96
Table 45: Percentages of ceramic types sampled per analytical unit
Interesting enough, early and late platforms are characterized by a high Predominance
of OSC’s while Midden, Stone Room and Water Flood Analytical Units are characterized by
the Predominance of bowls. Jars are the third most prevalent class in all analytical units, the
fourth being bottles. Cups and plates have very low percentages and are almost inexistent in
the sample (figure 100).
104
Having the basic distribution of vessel forms, the next step is to recognize specific
patterns of size/thickness variability in analytical units. In doing that I used modal clusters
extracted from bivariate kernel density estimations. As explained in Chapter 4, I used
univariate and bivariate KDE because I wanted to test if the modalities observed in univariate
KDE were replicated using bivariate KDE via thickness in addition to diameter. I had the
expectation that the modalities reflected by using diameter would resist when adding thickness
as a new dimension. The results indicate that when thickness is added a slightly different set of
modes appear. Thickness acts as a controlling measurement clustering modes according to the
diameter/thickness association.
6.3.1 OSC
Overall sample of OSC (n=751). Table 46 shows the results of KDE for diameter and
thickness measurements. Figures 101 and 102 indicate these modes and the way they overlap.
KDE Mode 1 Mode 2 Mode 3 Diameter 15 cm 25 cm absent Thickness 0.5 1.2 cm absent
Table 46: Results of univariate KDE for diameter and thickness measurements
When these two measurements (diameter and thickness) are plotted in a bivariate
kernel density estimation plot, three mode peaks appear as shown in table 47 and figure 103.
Thickness Diameter Count
0.61074 11.68 60 0.8906 14.56 229
1.17046 26.08 436
Table 47: Modal clustering table of OSC’s
Small Medium Large 12/0.6 16/0.9 28/1.3
Table 48: Measurements of total population of OCS’s types
Types Predominance Small and large Large
Table 49: Types of OCS’s
105
There is a third cluster that is hinted in the plateau between the two peaks in the
bivariate KDE plot. Conservatively there are two major size groups of OSC in the overall
population. But in order to address specific variation in the ranges of diameters and thickness, I
will segregate this population in each analytical unit identified previously.
6.3.1.1 Midden
There are 549 OSC sherds in the Midden unit. The following table illustrates the
results of KDE for diameter and thickness measurements. Figures 104 and 105 point out these
modes and the way they overlap.
KDE Mode 1 Mode 2 Mode 3 Diameter 10 cm 30 cm absent Thickness 0.5 1.3 cm absent
Table 50: Results of univariate KDE for diameter and thickness measurements
When these two measurements (diameter and thickness) are plotted in a bivariate kernel
density estimation plot, three mode peaks appear as show in table 51 and figure 106. The
plateau observed in the bivariate KDE plot of the total population of OSC is presented here as
an independent peak.
Thickness Diameter Count 0.5612 11.68 97 0.8888 15.52 61 1.3256 28 377
Table 51: Modal clustering table of Midden’s OSC
Small Medium Large 12/0.6 16/0.9 28/1.3
Table 52: Measurements of OCS’s types from the Midden Analytical Unit
Types Predominance
Small, medium and large Large
Table 53: Types of OCS’s from the Midden Analytical Unit
Three types of OSC are identified in the Midden analytical unit, small ones, medium
ones and large ones, each of these sizes are associated with a particular value of thickness that
increases as the diameter gets larger.
106
6.3.1.2 Stone Rooms
There are 103 OSC sherds in the Stone Rooms unit. Table 54 shows the results of
KDE for diameter and thickness measurements. Figures 107 and 108 indicate these modes and
the way they overlap.
KDE Mode 1 Mode 2 Mode 3 Diameter 15 cm 25 cm 40 Thickness 0.6 cm 1.0 cm
Table 54: Results of univariate KDE for diameter and thickness measurements
When diameter and thickness are plotted together in a bivariate KDE plot, the results
indicate the presence of three clusters of diameter/thickness relationship as seen in Table 55
and fig 109.
Thickness Diameter Count 0.58348 10.56 17 0.95284 16.32 25 1.13752 25.28 30
Table 55: Modal clustering table of Stone Rooms OSC’s
The same segregation of small, medium and large is repeated in the room analytical
unit. There is a higher presence of large and thicker OSC in this analytical unit, followed
closely by medium size OSC.
Small Medium Large 11/0.6 16/1.0 25/1.1
Table 56: Measurements of OCS’s types from the Stone Rooms Analytical Unit
The third mode observed in the univariate diameter KDE did not resist the bivariate
KDE analysis.
6.3.1.3 Early Platforms
There are 52 OSC sherds in the Early Platforms unit. Table 57 shows the results of
KDE for diameter and thickness measurements. Figures 110 and 111 indicate these modes and
the way they overlap.
KDE Mode 1 Mode 2 Mode 3 Diameter 15 cm 25 cm absent Thickness 0.8 cm 1.2 cm 1.6
Table 57: Results of univariate KDE for diameter and thickness measurements
107
When diameter and thickness are plotted together in a bivariate KDE plot, the
results indicate the presence of two of diameter/thickness relationship as seen in Table
59 and figure 112
Thickness Diameter Count 0.8746 14.6 37 1.0412 24.72 7
Table 58: Modal clustering table of Early Platfroms OSC’s
OSC population in the Early Platform context can be dissected into two types small
and large with a high Predominance of small ones.
Small Large 15/0.9 25/1.0
Table 59: Measurements of OCS’s types from the Early Platforms Analytical Unit
Types Predominance
Small and large Small
Table 60: Types of OCS’s from the Early Platforms Analytical Unit
6.3.1.4 Water Flood
There are 30 OSC sherds in the Water Flood unit. Table 61 shows the results of KDE
for diameter and thickness measurements. Figures 113 and 114 show these modes and the way
they overlap.
KDE Mode 1 Mode 2 Mode 3 Diameter 12 27 absent Thickness 0.8 cm 1.2 cm absent
Table 61: Results of univariate KDE for diameter and thickness measurements
When diameter and thickness are plotted together in a bivariate KDE plot, the results
indicate the presence of two of diameter/thickness relationship as seen in Table 62 and figure
115.
Thickness Diameter Count 0.7998 12.68 15 1.1268 26.72 6
Table 62: Modal clustering table of Water Flood OSC’s
Two types of OSC are quantitative detectable in this analytical unit, large and small
with a Predominance of small ones.
108
Small Large 13/0.8 27/1.1
Table 63: Measurements of OCS’s types from the Water Flood Analytical Unit
Types Predominance
Small and large Small
Table 64: Types of OCS’s from the Water Flood Analytical Unit
6.3.1.5 Late Platforms
This analytical unit by definition is characterized by its low density of archaeological
materials per cubic meter. There are 17 OSC sherds in the Early Platforms unit. Table 65
shows the results of KDE for diameter and thickness measurements. Figures 116 and 117
indicate shows these modes and the way they overlap.
KDE Mode 1 Mode 2 Mode 3 Diameter 13 cm 20 cm absent Thickness 0.3 cm 1.1 cm absent
Table 65: Results of univariate KDE for diameter and thickness measurements
When diameter and thickness are plotted together in a bivariate KDE plot, the results
indicate the presence of three of diameter/thickness relationship as seen in Table 66 and figure
118. It has to be noted that given the small sample from this analytical unit, the clusters
obtained have to be considered as highly preliminary until the sample size is expanded with
further excavations.
Thickness Diameter Count 0.735 14.2 3 0.811 19.8 6 1.02 11.12 5
Table 66: Modal clustering table of Early Platforms OSC’s
There are three types of OSC coming from this context: small, medium and large.
With a high Predominance of medium sized vessels.
Small-Thick Medium Large 11/1.0 20/0.8 25/1.1
Table 67: Measurements of OCS’s types from the Late Platforms Analytical Unit
Types Predominance
Small, medium and large Large
Table 68: Types of OCS’s from the Late Platforms Analytical Unit
109
Summary of OSC per Analytical Unit
Analytical Unit Small Medium Large Midden 12/0.57 16/0.9 28/1.3 Room 11/0.58 16/1.0 26/1.1 Early Platform 14.5/0.87 18/1.5 25/1.0 Water Flood 13/0.80 absent 27/1.1
Late Platform 11/1.0 20/0.8 25/1.1
Table 69: Measurements of different types of ollas sin cuello per analytical unit
Analytical Unit Types Predominance
Midden Large, medium, small Large Stone Rooms Large, medium, small Large, medium
Early Platform Large and small Small Water Flood Large and small Small
Late Platforms Medium and small Small
Table 70: Sizes of ollas sin cuello per analytical unit
6.3.2 Bowls
General population of bowls (n=1334). Table 71 shows the results of KDE for
diameter and thickness measurements. Figures 119 and 120 show these modes and the way
they overlap.
KDE Mode 1 Mode 2 Mode 3 Diameter 17 Absent Absent Thickness 0.5 cm Absent Absent
Table 71: Results of bowls univariate KDE for diameter and thickness measurements
When diameter and thickness are plotted together in a bivariate KDE plot, the results
indicate the presence of one diameter/thickness relationship as seen in Table 72 and figure 121
Thickness Diameter Count 0.4318 15.2 1275
Table 72: Modal clustering of total bowl sample
A medium bowl size predominates in the total bowl sample, characterized by a
diameter of 15.2 cm and a thickness of 0.43 cm.
110
6.3.2.1 Midden
The bowl midden sample is composed of 1114 rim fragments. Table 73 shows the
results of KDE for diameter and thickness measurements. Figures 122 and 123 show these
modes and the way they overlap.
KDE Mode 1 Mode 2 Mode 3 Diameter 17 Absent Absent Thickness 0.4 0.6 Absent
Table 73: Results of univariate KDE for diameter and thickness measurements
When diameter and thickness are plotted together in a bivariate KDE plot, the results
indicate the presence of two of diameter/thickness relationship as seen in Table 74 and figure
124.
Thickness Diameter Count
0.4318 15.2 1063
Table 74: Modal clustering of bowl sample from Midden
This is the same concentration evident in the general bivariate bowl kernel density
estimation. Given that midden bowls make 84.66% of the total bowl sample, this is hardly
surprising. The population of bowls in the Midden Analytical Unit is characterized by a very
predominant presence of medium size vessels of 15.2/0.43 cm.
6.3.2.2 Stone Rooms
The bowl Stone Rooms sample is composed of 143 rim fragments. Table 75 shows the
results of KDE for diameter and thickness measurements. Figures 125 and 126 show these
modes and the way they overlap.
KDE Mode 1 Mode 2 Mode 3 Diameter 15 Absent Absent Thickness 0.5 Absent Absent
Table 75: Results of univariate KDE for diameter and thickness measurements
When diameter and thickness are plotted together in a bivariate KDE plot, the results
indicate the presence of one diameter/thickness relationship as seen in Table 76 and figure 127
111
Thickness Diameter Count 0.4972 14.8 126
Table 76: Modal clustering table of bowl sample from Stone Rooms
The population of bowls in the Stone Rooms analytical unit is characterized by a very
predominant presence of medium size vessels of 15.2/0.50 cm, very similar to the sample from
the Midden analytical unit.
6.3.2.3 Early Platforms
The Early Platform bowl sample is composed of 36 rim fragments. Table 77 shows
the results of KDE for diameter and thickness measurements. Figures 128 and 129 show these
modes and the way they overlap.
KDE Mode 1 Mode 2 Mode 3 Diameter 15 Absent Absent Thickness 0.5 cm 0.8 cm Absent
Table 77: Results of univariate KDE for diameter and thickness measurements
When diameter and thickness are plotted together in a bivariate KDE plot, the results
indicate the presence of two of diameter/thickness relationship as seen in Table 78 and figure
130.
Thickness Diameter Count 0.53 10.8 20 0.57 14.64 9
Table 78: Modal clustering of bowl sample from Early Platforms
Small bowls are the most predominant size vessel in this analytical unit.
Small Medium 11/0.5 15-/0.6
Table 79: Measurements of bowl types from Early Platforms
Types Predominance Small and medium Small
Table 80: Types of bowls from Early Platforms
112
6.3.2.4 Water Flood
The bowl Early Platform population is composed of 40 rim fragments. Table 81
shows the results of KDE for diameter and thickness measurements. Figures 131 and 132
show these modes and the way they overlap.
KDE Mode 1 Mode 2 Mode 3 Diameter 15 25 Absent Thickness 0.5 cm 0.8 cm Absent
Table 81: Results of univariate KDE for diameter and thickness measurements
When diameter and thickness are plotted together in a bivariate KDE plot, the results
indicate the presence of two diameter/thickness relationships as seen in Table 82 and figure
133.
Thickness Diameter Count 0.3898 15.12 10 0.4724 23.28 18
Table 82: Modal clustering table of bowls from Water Flood
These values can be summarized in the following tables:
Types Predominance Medium and large Large
Table 83: Measurements of bowl types from Water Flood
Medium Large 15/0.39 23/0.47
Table 84: Types of bowls from Water Flood
6.3.2.5 Late Platforms
A population of 14 bowl rims was recovered from this analytical unit. Table 85 shows
the results of KDE for diameter and thickness measurements. Figures 134 and 135 show these
modes and the way they overlap.
KDE Mode 1 Mode 2 Mode 3 Diameter 15 25 Absent Thickness 0.5 cm 0.8 cm Absent
Table 85: Results of univariate KDE for diameter and thickness measurements
When diameter and thickness are plotted together in a bivariate KDE plot, results
indicate the presence of two diameter/thickness relationships as seen in Table 86 and figure
136.
113
Thickness Diameter Count 0.4928 17.2 5 0.5792 12.16 4
Table 86: Modal clustering table of bowls from Late Platforms
All univariate and bivariate kernel density estimations from this analytical unit must
be taken with extreme caution given the tiny sample size.
Types Predominance Small, medium medium
Table 87: Measurements of bowls types from Late Platforms
Small Medium 12/0.50 17/0.49
Summary of bowls:
Table 88: Types of bowls from Late Platforms
Analytical Unit Small Medium Medium Large
Midden absent 15/0.43 absent Room absent 15/0.50 absent Early Platform 11/0.52 15/0.60 absent Water Flood absent 15/0.40 23/0.47
Late Platform 12/0.50 17/0.49 absent
Table 89: Measurements of different types of bowls per analytical unit
Analytical Unit Types Predominance Midden Medium Medium Room Medium Medium Early Platform Small, medium Small Water Flood Medium and large Large Late Platforms Small, medium Medium
Table 90: Sizes of bowls per analytical unit
6.3.3 Jars
The total jar sample is composed of 425 rim sherds. Table 91 shows the results of
KDE for diameter and thickness measurements. Figures 137 and 138 show these modes and
the way they overlap.
KDE Mode 1 Mode 2 Mode 3 Diameter 10 Absent Absent Thickness 0.4 cm 0.6 cm 0.9
Table 91: Results of univariate KDE for diameter and thickness measurements
114
When diameter and thickness are plotted together in a bivariate KDE plot, the results
indicate the presence of one diameter/thickness relationship as seen in Table 92 and figure
139.
Thickness Diameter Count 0.5392 9.24 355
Table 92: Modal clustering table of jars
In order to identify the patterning of jar distribution within the prehistoric analytical
units of the Wacheqsa sector, each analytical unit will be independently analyzed as done with
previous ceramic types.
6.3.3.1 Midden
Jar population in this analytical unit is composed of 338 fragments. Table 93 shows
the results of KDE for diameter and thickness measurements. Figures 140 and 141 show these
modes and the way they overlap.
KDE Mode 1 Mode 2 Mode 3 Diameter 10 Absent Absent Thickness 0.4 cm 0.6 cm 0.9
Table 93: Results of univariate KDE for diameter and thickness measurements
When diameter and thickness are plotted together in a bivariate KDE plot, the results
indicate the presence of one diameter/thickness relationship as seen in Table 94 and figure
142.
Thickness Diameter Count 0.5392 8.46 282
Table 94: Modal clustering table of jars from Midden
6.3.3.2 Stone Rooms
The jar sample in this analytical unit is made up of 51 rim sherds. Table 95 shows the
results of KDE for diameter and thickness measurements. Figures 143 and 144 show these
modes and the way they overlap.
KDE Mode 1 Mode 2 Mode 3 Diameter 10 cm 15 cm 23 cm Thickness 0.4 cm 0.6 cm 0.9
Table 95: Results of univariate KDE for diameter and thickness measurements
115
When diameter and thickness are plotted together in a bivariate KDE plot, the results
indicate the presence of one diameter/thickness relationship as seen in Table 96 and figure
145.
Thickness Diameter Count 0.5347 11.65 36
Table 96: Modal clustering table of jars from Stone Rooms
6.3.3.3 Early Platforms
The sample size in this analytical unit is very low, and the results should be
considered as preliminary and be taken with caution as the jar population is composed of only
16 rim sherds. Table 97 shows the results of KDE for diameter and thickness measurements.
Figures 146 and 147 show these modes and the way they overlap.
KDE Mode 1 Mode 2 Mode 3 Diameter 6 cm 15 cm absent Thickness 0.4 cm 0.8 cm absent
Table 97: Results of univariate KDE for diameter and thickness measurements
When diameter and thickness are plotted together in a bivariate KDE plot, the results
indicate the presence of two diameter/thickness relationship as seen in Table 98 and figure
148.
Thickness Diameter Count 0.4112 6 5 0.7136 7.8 4
Table 98: Modal clustering table of jars from Early Platforms
These results can be organized in the following manner,
Very Small Small 6/0.41 8/0.7
Table 99: Measurements of jars types from Early Platforms
Types Predominance
Very small and small Very small
Table 100: Types of jars from Early Platforms
6.3.3.4 Water Flood
The same caution stated for the previous analytical unit has to be repeated for the
Water Flood analytical unit as the population of jars is composed of 14 rim sherds. Table 101
116
shows the results of KDE for diameter and thickness measurements. Figures 149 and 150
show these modes and the way they overlap.
KDE Mode 1 Mode 2 Mode 3 Diameter 6 cm 11 cm absent Thickness 0.2 cm 0.6 cm absent
Table 101: Results of univariate KDE for diameter and thickness measurements
When diameter and thickness are plotted together in a bivariate KDE plot, the results
indicate the presence of three diameter/thickness relationships as seen in Table 102 and figure
151.
Thickness Diameter Count 0.56 9.96 3 0.59 6 2 0.61 8.04 2
Table 102: Modal clustering table of jars from Water Flood
These values can also be organized in the subsequent way,
Very Small Small Medium 6/0.59 8/0.6 10/0.56
Table 103: Measurements of jars types from Water Flood
Types Predominance
Very small, small and medium Medium
Table 104: Measurements of jars types from Water Flood
6.3.3.5 Late Platforms
The jar sample in this analytical unit is composed of only five cases; with such a small
sample there is no need for analysis beyond examining the data table to find the patterning of
jars within this analytical unit. Nevertheless it is necessary to state that these values should be
taken with caution and considered preliminary until the sample is expanded.
Diameter Thickness 8 0.32 18 0.9 8 0.32 6 0.42 6 0.92
Table 105: Measurements of jar sample from Late Platforms
117
These values can be organized in the following way,
Very small Small Large 6/0.6-0.9 8/0.32 18/0.9
Table 106: Measurements of jars types from Late Platforms
Types Predominance
Very small, small and large Medium
Table 107: Measurements of jars types from Late Platforms
As a summary of the intrasite jar distribution in the Wacheqsa sector, the following
tables are offered,
Analytical Unit Very Small Small Medium Large Midden absent 8/0.5 absent absent Stone Rooms absent absent 12/0.5 absent Early Platforms 6/0.4 8/0.7 absent absent Water Flood 6/0.6 8/0.6 10/0.6 absent Late Platforms 6/0.6-0.9 8/0.3 absent 18/0.9
Table 108: Measurements of overall jars types per analytical unit
Analytical Unit Types Predominance
Midden Small Small
Stone Rooms Medium Medium
Early Platforms Very small and small Very small
Water Flood Very small, small and medium Medium
Late Platforms Very small, small and large Medium
Table 109: Jar types per analytical unit
6.3.4 Bottles
The total population of bottles is composed of 119 rim sherds, from which 101
(84.9%) are part of the Midden Analytical Unit. Given the reduced number of the bottle rim
sherds in the other analytical units (Early Platform 5, Late Platform 1, Stone Rooms 4 and
Water Flood 8), only the data set tables will be shown and patterns will be extracted from
them without doing kernel density estimations. It could also be argued that the non-Midden
units do not contain a significant proportion of the bottle population. Figures 152 and 153
show these modes and the way they overlap.
118
KDE Mode 1 Mode 2 Mode 3 Diameter 4 cm 6 cm 8 cm Thickness 0.4 cm 0.6 cm 0.9 cm
Table 110: Results of univariate KDE for diameter and thickness measurements
When diameter and thickness are plotted together in a bivariate KDE plot, the results
indicate the presence of two diameter/thickness relationships as seen in Table 111 and figure
154.
Thickness Diameter Count 0.3718 3.92 82 0.4822 5.84 20
Table 111: Modal clustering table of bottles
In order to identify the patterning of bottle distribution within the prehistoric
analytical units of the Wacheqsa sector, each analytical unit will be independently analyzed as
done with previous ceramic types, but before going further I consider it necessary to indicate
that with bottles I am dealing primarily with spout variation, which is rather different from the
inference of overall size variability implied for the other vessel forms. The results obtained,
rather than being significant in the sense of size bottle variation, are going to be meaningful in
the segregation of spout variation in the bottle sample as wide and narrow spouts can reflect
either vessels of quite similar or varied overall size.
6.3.4.1 Midden
As mentioned before a sample size of 101 rim sherds makes up the bottle population of the
Midden Analytical Unit. Given the predominance of the Midden component in the overall
sample, a similar tendency for bottles is not surprising. Figures 155 and 156 show these modes
and the way they overlap.
KDE Mode 1 Mode 2 Mode 3 Diameter 4 cm 6 cm 8 cm Thickness 0.4 cm 0.6 cm 0.9 cm
Table 112: Results of univariate KDE for diameter and thickness measurements
When diameter and thickness are plotted together in a bivariate KDE plot, the results
indicate the presence of two diameter/thickness relationship as seen in Table 113 and figure
157.
119
Thickness Diameter Count 0.358 3.92 67
0.4822 5.96 17
Table 113: Modal clustering table of bottles from Midden
Two types of spouts have been identified in this analytical unit, with a Predominance
of small-spouted sized bottles:
Small Medium 4/0.4 6/0.5
Table 114: Measurements of bottle’s spouts types from Midden
Types Predominance
Small and medium Small
Table 115: Types of bottle’s spouts types from Midden
6.3.4.2 Stone Rooms
The sample size in this analytical unit is composed of only four rim fragments. Given
the small nature of the sample size, no univariate or bivariate kernel density estimations were
made.
Diameter Thickness 6 0.6 3 0.42 3 0.36 3 0.51
Table 116: Measurements of bottle sample from Stone Rooms
The data from this analytical unit correspond to the small and medium spout types identified
for bottles in the Midden analytical unit. The small sized ones are more numerous if that term
can be applied to such a small sample size.
6.3.4.3 Early Platforms
As in the Stone Rooms Analytical Unit, the sample size does not justify the use of
kernel density estimations for the bottles class in this analytical unit. In this case the total
population of bottle cases is made up of six rims,
120
Diameter Thickness 4 0.49 4 0.65 4 0.57 4 0.33 4 0.25 4 0.4
Table 117: Measurements of bottle sample from Early Platforms
All the cases from this analytical unit can be related with the small spout type
recognized for bottles in the Midden analytical unit.
6.3.4.4 Water Flood
As in the previous analytical units, the sample size of bottle population from this
analytical unit does not warrant the use of kernel density estimations.
Diameter Thickness
4 0.43 4 0.38 4 0.41 3 0.34 6 0.36 6 0.34
Table 118: Measurements of bottle sample from Water Flood
The values from the table above indicate the presence of small and medium-sized
spouts, according to the values established for the Midden analytical unit.
6.3.4.5 Late Platform
The population of rim bottles from this analytical unit is composed by only once case
which measurements are 5/0.24. These numbers can be placed in the medium-sized spout
identified for the Midden analytical unit.
The following summary of bottle sizes per analytical unit is offered. It has to be stated
again that the values coming from Stone Rooms, Early Platforms, Water Flood and Late
Platforms must be taken with caution given their sample sizes.
121
Analytical Unit Small Medium Midden 4/0.4 6/0.5 Room 4/0.4-0.5 6/0.6 Early Platforms 4/0.3-0.7 absent Water Flood 3-4/0.3-0.4 6/0.3-0.4 Late Platforms absent 5/0.2
Table 119: Measurements of bottle spouts per analytical unit
Analytical Unit Types Predominance
Midden Small, medium Small Stone Rooms Small, medium Small
Early Platform Small Small Water Flood Small, medium Small
Late Platforms Medium Medium
Table 120: Bottle spouts types per analytical unit
6.3.5 Cups
The entire population of cups is composed by 40 rim sherds, of which 37 (92.5%)
belong to the Midden Analytical Unit. Given the reduced number of the cup rim sherds in the
other analytical units (Early Platform 0, Late Platform 1, Stone Rooms 1 and Water Flood 1),
only the values from those contexts will given. The extreme low frequency of this vessel type
suggests that their presence was not significant in the Wacheqsa cup population. Table 121
shows the values obtained with univariate kernel density estimations of diameter and thickness
(figures 158 and 159)
KDE Mode 1 Mode 2 Mode 3 Diameter 4 cm 6 cm 8 cm Thickness 0.4 absent absent
Table 121: Results of univariate KDE for diameter and thickness measurements
When diameter and thickness are plotted together in a bivariate KDE plot, the results
indicate the presence of five diameter/thickness relationship as seen in Table 122 and figure
160.
Thickness Diameter Count 0.2996 5.94 5 0.3968 8.04 7 0.44 3.98 10 0.5372 5.94 7 0.6236 8.04 2
Table 122: Modal clustering table of cups
122
It has to be noted that the small mode is incredible small and may represent miniature
vessels. It is expected that these same values will be replicated in the Midden analytical unit
given the preponderance of Midden cups in the entire cup population.
6.3.5.1 Midden
As mentioned before the population of cups is composed of 37 rim sherds. As
expected univariate and bivariate kernel density estimations (figures 161 and 162) replicate
the patterns observed for the entire population of cups and the modal clustering analysis
effectively replicate the values shown before (figure 163):
Thickness Diameter Count 0.2996 5.52 5 0.386 8.04 7
0.4184 3.84 10 0.548 5.94 3
0.6236 8.04 2
Table 123: Modal clustering table of cups from Midden
These values can be better understood looking at the following tables,
Small Medium Large 4/0.4 6/0.3-0.6 8/0.4-0.6
Table 124: Measurements of cup types from Midden
Types Predominance
Small, medium and large Small
Table 125: Types of cups from Midden
6.3.5.2 Stone Rooms
Only one cup rim sherd makes up the sample size of this analytical unit. Its value is
4/0.43 cm. It fits in the category of small cups from the Midden analytical unit.
6.3.5.3 Early Platforms
No cups were observed in this analytical unit.
6.3.5.4 Water Flood
The cup population of this analytical unit is composed of only one rim fragment. Its
value is 4/0.31 and it is located within the range of small cups from the Midden analytical unit.
123
6.3.5.5 Late Platforms
As the Stone Room and Water Flood analytical units, the population of cups is
composed by only one rim. Its value is 6/0.53 which locates it within the medium-thick range
of Midden cups.
With the reservations stated before regarding sample size, the cup population of the Wacheqsa
sector can be organized in the following ways,
Analytical Unit Small Medium Large Midden 4/0.4 6/0.3-0.6 8/0.4-0.6 Room 4/0.4 absent absent Early Platforms absent absent absent Water Flood 4/0.3 absent absent Late Platforms absent 6/0.5 absent
Table 126: Measurements of types of cups per analytical unit
Analytical Unit Types Predominance
Midden Small, medium and large Small Stone Rooms Small Small
Early Platform Water Flood Small Small
Late Platforms Medium Medium
Table 127: Types of cups per analytical unit
6.3.6 Plates
The entire population of plates is composed of 57 rim sherds, of which 46 (80.1%)
belong to the Midden Analytical Unit. Given the reduced number of plate rim sherds in the
other analytical units (Early Platform = 2, Late Platform = 0, Stone Rooms = 7 and Water
Flood = 2), only the values from those contexts will be mentioned. Also, the extremely low
densities suggest that plates do not make up a significant proportion of the overall sample.
Table 128 shows the values obtained with univariate kernel density estimations of diameter
and thickness (figures 164-166)
KDE Mode 1 Mode 2 Mode 3 Diameter 15 cm 25 cm absent Thickness 0.6 cm 0.8 cm absent
Table 128: Results of univariate KDE for diameter and thickness measurements
124
When diameter and thickness are plotted together in a bivariate KDE plot, the results
indicate two diameter/thickness relationship as seen in Table 129 and figure 166
Thickness Diameter Count 0.5724 12.4 6 0.615 22 40
Table 129: Modal clustering table of plates
There are two modes in the plate sample that I expect to be replicated in the Midden
analytical unit given the predominance of Midden plates in the entire plate sample.
6.3.6.1 Midden
As mentioned before the population of Midden plates is composed of 46 rim sherds.
As expected, the univariate and bivariate kernel density estimates fairly well replicate the
patterns observed for the entire sample of plates, and the modal clustering analysis nearly
replicate the values showed before (figures 167-169),
Thickness Diameter Count 0,615 22 38
0,9132 22 3
Table 130: Modal cluster table of plates from Midden
The main difference lies in the absence of the mode represented in the values 12/0.57
which belongs to a different analytical unit. That being said, large plates were the only mode
identified in the sample of Midden plates examined
Small Medium Large absent absent 22/0.6-0.9
Table 131: Measurements of type of plates from Midden
6.3.6.2 Stone Rooms
The population of plates is composed by six rim sherds with the following values,
Diameter Thickness 22 0.82 26 0.63 24 0.82 18 0.63 10 0.4 10 0.27
Table 132: Measurements of plate rims from Stone Rooms
125
These measurements can be organized in the following manner:
Small Medium Large 10/0.4 18/0.6 22-26/0.9
Table 133: Measurements of types of plates from Stone Rooms
Types Predominance
Small, medium and large Large
Table 134: Types of plates from Stone Rooms
This classification has to be taken with extreme caution given the small size of the
population of plates in this analytical unit.
6.3.6.3 Early Platforms
The population of plates in this analytical unit is composed by two rim sherds with the
following measurements,
Diameter Thickness 12 0.56 24 0.69
Table 135: Measurements of plate rims from Early Platforms
These values fall with the range of small and large plates identified for the overall plate
sample.
6.3.6.4 Water Flood
Again, only two cases are present in the Water Flood analytical unit with the
following values,
Diameter Thickness 26 0.51 14 0.92
Table 136: Measurements of plate rims from Water Flood
These measurements fall within the small and large plates identified in the overall
plate sample.
6.3.6.5 Late Platforms
No plates were identified in this analytical unit.
126
The following is a synopsis of plate distribution per unit. It cannot be stated too strongly the
importance of being cautious with the values that come from Stone Rooms, Early Platforms,
Water Flood and Late Platforms units given their sample sizes.
Analytical Unit Small Medium Large Midden absent absent 22/0.9 Room 10/0.4 18/0.6 22-26/0.9 Early Platforms 12/0.6 absent 24/0.7 Water Flood 14/0.9 absent 26/0.51 Late Platforms Absent absent absent
Table 137: Measurements of type of plates per analytical unit
Analytical Unit Types Predominance
Midden Large Large Stone Rooms Small, medium and large Large
Early Platform Small and large Small and large Water Flood Small and large Small and large
Late Platforms absent absent
Table 138: Types of plates per analytical unit
There is a clear variation in the ceramic assemblage analyzed that may reflect different
functions and uses of each of the prehistoric analytical units identified.
Vessel Sizes Predominance OSC Small, medium and large Large Bowl Medium Medium
Jar Small Small Bottle Small and medium Small Cup Small, medium and large Small Plate Large Large
Table 139: Summary of vessels type from Midden
Vessel Sizes Predominance OSC Small, medium and large Large Bowl Medium Medium
Jar Medium Medium Bottle Small, medium Small Cup Small Small Plate Small, medium and large Large
Table 140: Summary of vessels type from Stone Rooms
127
Vessel Sizes Predominance OSC Small and large Small Bowl Small and medium Small
Jar Very small and small Very small and small Bottle Small Small Cup absent Plate Small and large Small and large
Table 141: Summary of vessels type from Early Platforms
Vessel Sizes Predominance OSC Small and large Large Bowl Medium and large Large
Jar Very small, small and medium Medium Bottle Small Medium Cup Small Small Plate Small and large Small and large
Table 142: Summary of vessels type from Water Flood
Vessel Sizes Predominance OSC Small, medium and large Large Bowl Small, medium Medium
Jar Very small, small and large Medium Bottle Medium Medium Cup Medium Medium Plate absent absent
Table 143: Summary of vessels type from Late Platforms
In this chapter I have provided evidence of the intrasite variability that exists in the Wacheqsa
sector that can be summarized as follows:
• Eight spatial analytical units have been identified on the grounds of spatial
distribution of strata, examination of types of soil and density estimates of
archaeological artifacts.
• These analytical units are segregated in three chronological phases.
• Prehistoric analytical units have high values of diversity according to the Boone Index
calculated per each stratum. All analytical units present high diversity of
archaeological materials suggesting that the activities carried on in these spaces
involved a wide array of materials. No single unit is characterized by an exclusive
value of homogeneity.
• Nevertheless, in spite of this shared diversity, prehistoric units show intrasite
variation. This variation can be inferred examining the density values per each class of
archaeological materials but in order to test this intrasite diversity I have used kernel
128
density estimations on a subset of 3020 ceramic sherds trying to identify the patterns
of variation within each analytical unit and among themselves.
One cannot help but wonder what is the relevance of these inferences? What are the
social implications of this data? The answers to these questions are in the next chapter in
which I discuss the implications of this variability as well as the anthropological meaning of
the intrasite variation identified.
129
CHAPTER 7
THE WACHEQSA SECTOR AS A MULTICOMPONENT AREA
This chapter discusses the results obtained in the present research, contextualizing
them into archaeological and anthropological frameworks. First I explain the activities that
have been inferred from the materials analyzed, and then I discuss the implications of these
activities for the understanding of Chavín de Huántar. Next, I lay out the contributions of the
Wacheqsa Sector for the general comprehension of the chronological framework of Chavín de
Huántar and the Middle and Late Formative periods. Lastly, I summarize the results of my
research.
7.1 Inferred Activities
I will identify activities in the Wacheqsa sector primarily based on ceramic bivariate
kernel density estimations, modal clustering, vessel modalities, and frequencies and densities
of different classes of archaeological materials retrieved during the excavation process. Based
on these results I have inferred two different activities from the archaeological record of the
Wacheqsa sector. I argue that each activity is a reflection of a different power strategy
exercised by the authorities of Chavín de Huántar.
7.1.1 Feasting activities
Evidence for feasting activities has been identified in the Midden analytical unit. In
this unit densities of archaeological materials are extremely high when compared with
densities of other analytical units, which immediately separates this unit from the rest. The
ceramic assemblage from the Midden corresponds to the types previously defined by Richard
Burger as Janabarriu (Burger 1984; Burger 1992; Burger 1998; Burger and Matos 2002)
Middens can either be formed by the aggregation of waste from different areas of a
site or by the aggregation of waste coming from one specific area (Boone 1987); they do not
reflect an in situ location of the archaeological materials but rather a secondary one. If a
midden is the aggregation of materials brought from a different place or different places, one
can expect to find materials that are the results of several/multiple different activities. When
the physical characteristics of these deposits are variable, one can argue for different behaviors
producing different records. On the other hand if densities of archaeological materials as well
of their distributions are homogeneous, one can argue for a common behavior or activity as
130
source of the waste present in the deposits. In the case of the Midden analytical unit of the
Wacheqsa sector, deposits reflect a homogenous behavior in terms of soil composition (see
Chapter 5) as well as density of archaeological materials per stratum excavated (see Chapter
6). The physical characteristics of the Midden analytical unit, as well as the consistent density
of archaeological materials in the included strata point towards the identification of the
Midden mostly as a product of a singular activity that was repeated over time. Most of the
strata were compacted with abundant soil matrix between the refuse materials, suggesting that
its deposition was gradual. The question at hand is what kind of activity this was and where
the deposits were coming from. I will argue that deposits from the Midden analytical unit were
formed from waste produced by collective consumption of food and drink, in other words,
suprahousehold feasts.
As Dietler and Hayden indicate, it is important to differentiate communal
consumption from “everyday domestic meals and from the simple exchange of food without
communal consumption” (Dietler and Hayden 2001:3). Large quantity of food items,
unusually large numbers of serving and cooking vessels of large size, exotic items as well as
narcotic paraphernalia together are indicators of feasting activities (Blitz 1993; DeBoer 2001;
Dietler and Herbich 2001; Hayden 2001; Mills 1999; Potter 2000; Rosenswig 2007). I will
examine these indicators in the following pages, based on the data retrieved from the Midden
analytical unit at the Wacheqsa sector
7.1.1.1 Ceramics
Six general types of ceramic vessels have been identified in the five prehistoric
analytical units, and I have demonstrated intrasite variability in their distributions. Among
these types, ollas sin cuello, bowls, and jars are consistently present in all prehistoric
analytical units, while bottles, cups and plates are not significantly present in any but the
Midden analytical unit
The most ubiquitous ceramic vessel present in the Midden is the bowl, with a
unimodal distribution of medium sized bowl that makes up more than 50% of the analyzed
Midden sample. The second ubiquitous type is the olla sin cuello that make up 26% of the
analyzed sample with bimodal distribution but with a high predominance of large sized
vessels. Third is the jar type that makes up 11% of the ceramic sample analyzed with a
131
unimodal distribution of small jars24
. The other types of ceramics are under 5% of the sample
analyzed.
The elevated number of bowls and specially the large quantity of medium-sized bowls
indicate a) the consumption of large amounts of food, and b) the use of a standardized serving
ration. Bowls are serving vessels that can be either used for solid food or liquids (DeBoer
2001; Lumbreras 2005). Ollas sin cuello can be either used as cooking or storage (Blitz 1993;
Lumbreras 2005). Bowls posses unrestricted mouth orifices “unrestricted vessels are an
advantage not only in getting the contents out, but also in putting materials in” (Rice 1987:
241). Additionally unrestricted vessels, such as bowls, show the contents of the vessels which
is important for serving vessels.
The bimodal distribution of osc’s, points towards the cooking/storage of either
different types of food and/or for smaller and larger amounts of people. The size of an olla not
only reflects the amount of food being cooked/stored but also the type of food prepared/stored.
Size in ollas can be a function of amount and type of aliments,
“The relation between use and capacity of a vessel can be conceived in
terms of the kind of materials the vessel contains, the amount, the
length of time it is to be contained, the number of anticipated users of
the material during that time and micro environmental factors such as
availability of water and other necessities” (Rice 1987:225)
As ollas sin cuello, jars were fundamentally storing/serving vessels but primarily for
liquids. Jars are restricted vessels with necks that prevent the contents to be spilled. They are
significant in the ceramic midden assemblage and as seen in the previous chapter the most
typical jar in this analytical unit is the mid-sized one. Interesting enough, bottles only make
3% of the sample analyzed indicating that jars, rather than bottles, were the more popular type
of vessel for storing and/or serving liquids.
Food may have been cooked and/or carried in ollas sin cuellos to the feasting facility
where it would have been distributed in bowls. It is not uncommon to find fragments of ollas
sin cuello in the analyzed sample with evidence of firing on their surfaces. Jars may have been
used for storing beverages consumed during the feasting activity from which liquids could
have been poured into bowls, or more probably cups.25
(An interesting analogy is how
“ceremonial” drinking now occurs in the Andes – liquids are poured from containers into a
24
Regarding jars there is no a direct correlation between the volume of liquids that a jar can contain and
its orifice diameter. A direct correlation has been found between height and volume (Mills 1999) but
regrettably jar sherds analyzed are not big enough to provide information regarding jar heights. 25
Cups only make 1.5% of the sample analyzed.
132
very restricted number of cups/glasses, in sequential or multi-partnered reuses of shared
consuming vessels. The archaeological record of this would yield very few consumer
containers compared with the source bottles). Even though there are very few plates in the
analyzed ceramic subset, it is interesting that the higher mode of plates represented is the large
one. Plates may have been used for serving special types of food or for serving food for
especial attendees.
There are unusually large numbers of cooking and serving vessels in the Midden
analytical unit, midden bowls conform 84% of the entire bowl population and midden osc’s
make up 74% of the osc population.
7.1.1.2 Faunal Remains
The ceramic data is more compelling when cross referenced with the weight of faunal
remains recovered in the Midden which is extremely high when compared with the rest of
prehistoric analytical units, as seen in fig 170 and tables 145-146. Student’s t tests of these
measurements indicate that the means of the deposits excavated are significantly different.
t Alpha 1,98498 0,05
Abs(Dif)-LSD Midden Water Flood Early Platforms Rooms Late Platforms Midden -2670,7 2554,7 5657,6 4536,2 5236,1 Water Flood 2554,7 -4763,8 -1917,2 -2755,8 -2167,4 Early Platforms 5657,6 -1917,2 -2884,6 -3974,7 -3286,3 Rooms 4536,2 -2755,8 -3974,7 -4996,3 -4419,2 Late Platforms 5236,1 -2167,4 -3286,3 -4419,2 -4079,5
Table 144: Student’s t test. Least significant difference threshold matrix. Positive values show pairs of
means that are significantly different.
Level - Level Difference Lower CL Upper CL p-Value Midden Late Platforms 8683,948 5236,14 12131,76 0,0000026 Midden Rooms 8542,161 4536,17 12548,15 0,0000529 Midden Early Platforms 8437,329 5657,62 11217,04 0,0000026 Midden Water Flood 6416,508 2554,74 10278,28 0,0013658 Water Flood Late Platforms 2267,440 -2167,44 6702,33 0,3127186 Water Flood Rooms 2125,653 -2755,82 7007,13 0,3895406 Water Flood Early Platforms 2020,821 -1917,15 5958,79 0,3109436 Early Platforms Late Platforms 246,620 -3286,33 3779,57 0,8900858 Rooms Late Platforms 141,787 -4419,23 4702,81 0,9509248 Early Platforms Rooms 104,832 -3974,67 4184,33 0,9594244
Table 145: Student’s t test. Ordered differences report
133
A measure of density is more reliable than a measure of weight as weight can be
influenced by the excavation volume of each deposit while density measures the amount of
remains per standardized unit, in this case a cubic meter. These results could be considered
evidence of substantial food consumption that align with what Mercer and Hayden have
suggested regarding faunal remains as a signature for banquets as feasting activities. “Feasting
foods as well as actual feasts can often be recognized by copious food leftovers and much
greater wastage than usual” (Mercer 1985:100) and that “feasting refuse tends to occur in
considerable quantities in single deposits” (Hayden 1995:138). The Midden analytical unit is
composed of strata of similar characteristics in which faunal remains tend to occur in
considerable quantities. However, the faunal evidence alone is not conclusive for feasting, but
when cross-referenced with the vessels modalities explained above, the argument is more
robust. Special care was taken during the excavation process in separating possible human
remains, not finding any human remain as part of the Midden bone assemblage. Further
analyses will indicate what kinds of animals were consumed during feasting activities at
Chavín de Huántar, but I believe it is safe to hypothesize that a large proportion are probably
camelids (Burger 1998; Miller and Burger 1995).
7.1.1.3 Narcotic Paraphernalia
Bone artifacts (n=62) were also present in the Midden archaeological assemblage.
Many of which are small fragmented spoons and bone tubes of different sizes (figures 171-
174). The presence of polished bone tubes and small spoons is interesting and may indicate
the presence of drug consumption during feasting activities,
“The equipment for the inhalation of psychoactive powders consists
of a distinct set of implements: a small tray, a snuffing tube, a spoon,
and leather pouches as containers for the powders” (Torres and Repke
2006:11)
Based on ethnographic evidence from South America and the Caribbean, Torres and
Repke point out two general methods for snuffing powders: self-administered and
cooperative. Self administered consumption requires inhalers made of bird bone and wood
which can be single tubes or double tubes, while the collaborative way usually requires long
bird bones preloaded with powder that one individual blows into the nose of other individual
(Torres and Repke 2006). Evidence from the Midden deposit suggests self administered
consumption of psychoactive elements, possibly anadenanthera as suggested by Torres,
Repke and Rick (Rick 2006; Torres and Repke 2006). Anadenanthera beans need to be
134
roasted until desiccated, then the outer case is removed and the inner tissue is ground until a
fine powder is produced, large particles are removed and then the powder is ground again,
finally the substance is ready to be mixed with lime in order to be consumed (Torres and
Repke 2006). Similar equipment has been also found in La Banda and has also been
interpreted as evidence of consumption of psychoactive substances (Rick 2006).
7.1.1.4 Exotic Items
Mollusks are present in the midden deposit as well, although extremely rare. Six
fragments of Mesodesma donaceum were found. Richard Burger pointed out that the presence
of mollusks at Chavín de Huántar was due mostly to ceremonial reasons rather than nutritional
motivations as it is very costly to bring them from the coast and keep them fresh during the
trip (Burger 1998). According to Sheila Pozorski, it is possible to carry fresh seafood from the
coast to Chavín if they are maintained in a saline and moist environment which aggravates the
cost of transport (cited by Burger as personal communication in Burger 1998: 242). Travel
between the coast and Chavín can be done in four to seven days (Burger 1998), which would
allow the shellfish to be fresh on arrival to Chavín if the above stated precautions are taken.
The extreme low density of shells in the Midden can be explained by their fragile nature.
Large amounts of waste were thrown into the midden, which in turn will most likely
disintegrate fragile materials reducing the frequencies of retrievable materials. Another line of
evidence that reinforces the interpretation of shells as ritual or even elite artifacts is that 90%
of the sample in previous excavations of domestic units was located in structures identified as
belonging to elite households (Burger 1984; Burger 1998). Rare items also include slate
projectile points, chrysocolla beads and anthracite mirrors which are ubiquitous only in the
Midden analytical unit. There is also a very small amount of what can be considered bone
pendants, probably worn around the neck.
Examined together, the quantity and quality of the archaeological materials suggest
that the Midden analytical unit resulted from activities that involved suprahousehold
consumption of food and probably use of psychoactive substances
“Feasting activities by their very nature produce copious amounts of
distinctive refuse at the locations where they occur, and feasting
locations are often associated with notable ritual structures” (Dietler
and Hayden 2001:9)
The evidence of large ollas sin cuello, medium sized bowls, abundant jars, unusually
high densities of faunal remains, narcotic paraphernalia and exotic artifacts points towards
135
supra household consumption of food and liquids. Isolated, these remains are not strong
indicators of this activity, but found together in stratigraphically controlled deposits makes the
case compelling for the occurrence of feasting at Chavín de Huántar. Nevertheless, there is
only one set of materials that do not correspond in a clear way to supra household food
consumption activities: large fragments of columns, walls, and floors made of clay that in
some cases show signs of fire. As explained in Chapter 6, these fragments are present in the
Midden analytical unit with a mean density of 96.96 fragments per cubic meter. They came
from larger structures, finely plastered in white and red colors. Tello found similar materials in
front of Building A that most likely were the remains of architecture and plaster that was part
of the façade of Building A although Tello hypothesized that they were product of a big fire at
Chavín de Huántar.
“In front of the staircase [south staircase in Building´s A façade]
abundant pieces of burnt clay appear, several of them heavily burnt
and amorphous, suggesting that the combustion was very high and
does not correspond to preparation techniques of wall plastering
through firing. Most likely, either an intentional or unintentional fire
was responsible of firing evidence in these materials26
”. (Tello Ms
[1940]: 50)
A small fraction of the wall and floor materials found in the Midden analytical unit,
show signs of vitrification on their surfaces, suggesting high temperatures in their
combustion,. How can these elements be related with feasting activities? Would these
fragments be part of the structures were banquets occurred? Was some sort of ritual
destruction involved in banquets? Or simply put, do these materials come from a different set
of contexts? Ceremonial breaking of vessels has been recorded for the Middle Horizon (700 -
1100 AD) site of Conchopata in which ritual vessels were broken and then buried or disposed,
all this in a context of supra household feasting (Cook 2004). In extending this analogy to the
architectural fragments mentioned above, would it be that portions of structures associated
with feasting activities were disposed after the conclusion of the ceremony? These are more
open questions than conclusive statements given the lack of comparative examples.
Nevertheless in spite of the presence of these materials I find that the evidence for
feasting still stands. Ceramic modalities, extremely high densities of faunal remains and
evidence of psychoactive consumption together and not considered independently, point
towards feasting as one reasonable explanation.
26
My translation
136
7.1.2 Domestic Activities
7.1.2.1 Early Platforms
Early platforms contain evidence of domestic activities and may have been used as
domestic areas before the Black and White stage. Decorated ceramics can be related with the
types defined by Richard Burger as Urabarriu (Burger 1984; Burger 1998).
The olla sin cuello is the most prevalent ceramic type of the Early Platforms
consisting of 49% of the Early Platforms ceramic sample. The distribution of this type is
characterized by size bimodality of small and large sized vessels, with a bias towards the
small-sized mode. Bowls make up 30% of the Early Platforms sample size and have a bimodal
distribution represented by small and medium sized vessels with a strong bias towards small
bowls. Jars make up 14% of the Early Platform ceramic sample and have a unimodal
distribution represented by small jars. Bottles and plates together make up less than 7% of the
sample of the Early Platform analytical unit.
Based on the ceramic assemblage analyzed it can be inferred that mainly cooking
activities were carried out on these platforms. The small osc predominant mode points towards
the preparation of food for small numbers of people in comparison to the food prepared for the
feasting parties represented in the Midden analytical unit. This interpretation is reinforced by
the small-sized mode prevalent in the bowl distribution; hence consumption of food was
probably here carried out at the household level. That being said, it is important to point out
the presence of a small but nevertheless consistent large vessel size mode in the ollas sin
cuello form. Equally important is the presence of a second, medium size bowl mode. These
results may indicate the occasional consumption of food at a slightly larger scale. Faunal
remains are also present in this analytical unit but on a smaller scale than in the Midden. Only
32 kg of animal bone were recovered with a mean density of 1.5 kg per m³ excavated in this
analytical unit, contrasting with the Midden unit.
Three fragments of Mesodesma donaceum were recovered. As discussed before
preservation of these remains is affected by their fragile character and the humid conditions of
the highlands. Nevertheless the presence of these objects in this analytical unit is interesting.
As discussed previously shells can be considered as status artifacts in the highlands given the
costly effort in bringing them from the coast and in conserving their freshness during their
transport. But even, if the shells were not consumed and only used for artisan shell usage -
elaboration of shell artifacts for example- shells would still be considered exotic and probably
status artifacts given their costly transportation. Shell artifacts are even more difficult to find
137
in the archaeological record (there were no shell artifacts identified in the Wacheqsa sector)
because the shell is reduced in its form. Small size shell artifacts would have disappeared from
the archaeological record given their small size, fragile structure and humid conditions of
Chavín de Huántar. The waste produced by shell work would similarly have the same low
chances of survival in the archaeological record.
Fragments of burnt clay have been also retrieved in this analytical unit. These
fragments are smaller than the ones retrieved from the Midden analytical unit and do not show
signs of vitrification. As discussed before, fragments recovered in the Midden deposits show
signs of being exposed to very high temperatures that left evidence of vitrification on their
surfaces and were part of non domestic architectural features such as large, formal columns,
plastered walls and floors. The fragments recovered in the Early Platforms differ sharply,
being characterized by their small size but overall narrow thickness (<1.0 cm), suggesting a
distinct architectural origin rather than being the result of site formation processes27
. Some of
them have imprints of canes which indicate that probably were part of structures made of dried
mud with large wooden canes providing the internal structure. These types of structures are
still ubiquitous in the modern town of Chavín de Huántar, where low stone walls serve as
foundations for a wattle-and-daub superstructure. Regrettably, the foundations of the domestic
units have yet to be found. Only a small hearth found in stratum 524 has been recorded. One
of the reasons for the lack of structures in the Early Platforms, could be that they did not have
solid foundations having their bases disappeared or that the sampling program could not find
any structure given the organization of the excavation sampling design.
7.1.2.2 Stone Rooms
Bowls make up almost 50% of the ceramic assemblage analyzed in the Stone Room
analytical unit, having a unimodal distribution representing the medium sized bowls. Bowls
dominate the ceramic assemblage, emphasizing the serving nature of the activities that
originated the deposits. Ollas sin cuello make up 36% of the ceramic assemblage. Having a
multimodal distribution as revealed in the bivariate kernel density plots. The predominant
mode is the one represented by large-sized vessels closely followed by medium-sized ones.
Jars have a unimodal distribution with a mode represented by medium-sized vessels. Bottles,
27
Several fragments of burnt clay retrieved in the Midden analytical unit are of small size but thicker
(>1.0 cm) suggesting that they were part of thicker and more solid walls.
138
cups and plates together make up less than 5% of the ceramic assemblage. Judging from the
ceramic evidence, it can be said that the consumption of food was one of the activities carried
on in these rooms. Medium sized bowls were likely used for this function.
The Early Platform osc ceramic assemblage indicates the consumption of foodstuffs at
a household level while the ceramic assemblage in the Stone Rooms analytical unit may
indicate consumption of food by a larger number of people. Medium sized bowls and large
ollas sin cuello characterize the ceramic assemblage from the Stone Rooms analytical unit
while small ollas sin cuello and small bowls are predominant in the Early Platforms.
Cautiously, it can be suggested that change in ceramic modalities can be interpreted as an
increase in the population living in the Wacheqsa sector or a change of the type of inhabitants
living there. A change from small to large ollas sin cuello and bowls indicates that more food
and liquids were being produced and stored, if more food was being cooked (ollas sin cuello)
and served (bowls) it can be argued that more people were being fed or that a small number of
people was abundantly consuming food. An accurate estimation of mnv (minimum number of
vessels) will shed light towards this complex issue.
Additionally, prestige items such as obsidian, shells and anthracite mirrors are present
in this analytical unit. Anthracite mirrors are almost absent in the Early Platforms but are
present in the Stone Room analytical unit, seven fragments were recovered. Although this is a
large number when compared to the Early Platforms, it is small when compared to the
anthracite Midden assemblage (n=139).
Other special elements in the record of this analytical unit are beads made of
chrysocolla. Chrysocolla is a hydrated copper silicate often used as an ornamental stone. I put
special importance on three pieces of unworked chrysocolla, the only fragments of this type
found in the Wacheqsa area (figure 86). Another special material is a fist-sized fragment of
native copper ore (figure 87). There are three reported cooper ore sources less than a kilometer
from Chavín de Huántar that could have been perfectly exploited in order to get raw materials
for the production of cooper artifacts. These items are important because they give indications
of the occurrence of metallurgy at the Wacheqsa sector during the Janabarriu phase.
In terms of diversity, the Stone Rooms analytical unit also shows the lowest p(Hi)
mean, indicating that their deposits were highly diverse within the already high diversity
stated for all five prehistoric analytical units. This points out that the unique character of
artifact distribution in this analytical unit is best explained in reference to the activities carried
on inside the rooms rather than sample size bias.
139
In summarizing the information presented in the last few pages, I propose that during
the Urabarriu phase a support population lived in the Wacheqsa sector
7.1.3 Water Flood
As mentioned in the previous chapter, this analytical unit is composed of strata
representing water-originated sedimentation. It is, thus, likely that the archaeological materials
present in these strata were transported by water from other areas. Interestingly enough, all
decorated ceramics recovered in this analytical unit are related to ceramic types identified as
Urabarriu (Burger 1984; Burger 1998), a conclusion that is reinforced by the stratigraphic
position of these deposits below the Midden deposits. It is difficult to assess if all the materials
were carried from a single area or from different areas, but there is a clear pattern in the
frequency distribution of ceramic types that may suggest a common origin. This pattern is
represented in the following way. Bowls and ollas sin cuello have almost equal proportions in
the ceramic assemblage analyzed (39% and 36% respectively). According to kernel density
estimations, bowls present a bimodal distribution of medium and large sizes with a bias
towards the large size. Ollas sin cuello show a bimodal distribution too, representing small and
large-sized vessels with a bias towards small-sized ones, large size bowls and small ollas sin
cuello dominate the ceramic assemblage of this analytical unit. Obsidian and fragments of
burnt clay are present in these deposits, the burnt clay fragments are similar to those found in
the Early Platforms rather than in the Midden, suggesting that these fragments did not come
from any sort of large architecture.
7.1.4 Intermediate Area
Archaeological materials are extremely scarce in the Late Platforms analytical unit.
Their density is only 17 artifacts per m³ excavated. The small number of ceramics analyzed
(n=42) point towards an emphasis on cooking activities for a small number of people, given
the bias towards small ollas sin cuello and small bowls, but caution needs to be exercised
given the extreme low frequencies of materials recovered.
Only one fragment of burnt clay was found indicating that virtually no structures were
associated with this analytical unit. It seems that this area was kept clean for the most part,
given the extremely low density of archaeological materials. The Late Platforms analytical
unit is spatially located between the Midden and Stone Rooms analytical units. It may have
served just as an area of intercommunication or separation between the Midden and Stone
140
Room analytical units. Further excavations in this particular unit will shed light towards its
actual function.
7.2 Implications and Relevance
As mentioned at the onset of this chapter, one of my goals is to discuss the
implications of these inferred activities for the understanding of Chavín de Huántar. In that
sense this section brings together the specific data interpretation of the Wacheqsa sector in
relation with the ceremonial center, elaborating an anthropological framework for the
Wacheqsa sector and the ceremonial center combined. After this discussion is concluded I will
then discuss the carbon dates retrieved from the five prehistoric analytical units in order to
assess their distribution in absolute chronological dates.
7.2.1 Feasting and Power
Evidence from the Midden analytical unit suggests that suprahousehold food
consumption occurred at Chavín de Huántar. Feasting seems to be a plausible interpretation of
the archaeological record of this analytical unit, however special care has to be taken in
making a conclusive statement until faunal remains had been analyzed. Nevertheless, the
occurrence of supra household feasting at Chavín de Huántar would not be surprising.
The occurrence of feasting at Chavín the Huántar carries implications for the
interpretation of power strategies and corporate activities sponsored at Chavín during the
Andean Formative. Potential feasting facilities are present at Chavín de Huántar: a) a
quadrangular plaza located in front of Building A and surrounded by Buildings E and F
resembling the form of the U-shaped buildings of the central coast; this plaza could hold up to
1500 people (Burger 1992) and b) a circular plaza located in front of Building B, between
buildings C and A. Following Kembel’s architectural sequence (Kembel 2001), these two
plazas were contemporary, constructed as part of the Black and White phase. As mentioned in
Chapter 1, the Black and White phase was the architectural phase in which the largest
construction effort was made in Chavín de Huántar;. This architectural project involved the
modification of the Mosna River’s course in order to gain space for the construction of the
quadrangular plaza (Kembel 2001; Kembel and Rick 2004; Rick 2005). Additional evidence
of this displacement of the river has been retrieved by Daniel Contreras in his investigations at
Chavín de Huántar (Contreras 2007). Unlike the quadrangular plaza that is in the open, the
Circular Plaza is located in a more intimate environment (an intimacy reinforced by its modest
141
21 m diameter). Its visual isolation is evident in the plaza floor’s location around two meters
below the East Atrium. The plaza is surrounded by engraved stone slabs of anthropomorphic
and zoomorphic images of the Chavín religious world, creating a sacred space where much of
the world is distant; this plaza may have held up to 500 people (Burger 1992) . These plazas
have been interpreted as locations for ceremonies and festivities related with to rituals
performed and staged at the ceremonial center (Burger 1992; Kembel 2001; Kembel and Rick
2004; Lumbreras 1989, 1993; Rick 2005; Rowe 1962a; Tello 1960) and they were reasonable
locations for supra household feasting activities. Additionally it is important to note that
during the Black and White stage there was an architectural emphasis towards the construction
of open spaces:
“Prior to the Black and White stage, the focus on galleries, small
gallery patios outside their entrances, and their restricted access,
suggest they were designed for individuals or small groups to use the
space. The shift to the Black and White stage, with its plazas,
terraces, and flanking mounds with staircases ascending directly from
ground level, likely represents a shift in the emphasis to creating
larger open spaces to accommodate more people” (Kembel 2001:223)
It is worthy of note that the Circular Plaza is flanked by the Ofrendas and Caracolas
Galleries. The first one was the locus of a ritual offering of 681 ceramic vessels accompanied
with food remains and probably liquids as well as shells, stone and wooden artifacts
(Lumbreras 1989, 1993) . The Caracolas Gallery was the repository of 20 Strombus shell
trumpets that showed evidence of heavy, extended use and which most likely were used in the
ceremonies carried on in the ceremonial center (Rick 2005). In the Ofrendas Gallery only two
Janabarriu related ceramic vessels were found while the rest of the ceramics are representative
of ceramic styles from the north coast, northern highlands and central coast. In the Wacheqsa
midden deposits the typical Janabarriu designs of stamped circles, stamped concentric circles,
impressed concentric circles with dots and stamped S’s overwhelmingly dominate the
decorated ceramic assemblage while designs related to the iconography depicted in the
ceremonial center architecture or related to the ceramic styles identified in the Ofrendas
gallery are not ubiquitous. The presence of highly religiously loaded ceramics in the Ofrendas
gallery and the presence of foreign pottery and the near absence of Janabarriu ceramics
suggests the different nature of the Janabarriu ceramics when compared to the ceramics found
in the Ofrendas gallery. The architectural environment in which the Circular Plaza is located
seems to be of a more sacred nature than the quadrangular plaza and the Midden deposits do
not show a significant presence of sacred items (when I use this term sacred I mean similar to
142
the objects identified in the Ofrendas Gallery or even in the Caracolas gallery). Janabarriu
ceramics are not common in religious contexts at Chavín de Huántar. Granted, that the only
evidence of in situ ceramic remains in the architecture comes from the Ofrendas Gallery. The
extreme low frequency of Janabarriu ceramics in the Ofrendas Gallery suggests their different
nature when compared to the Ofrendas or Draganiano styles that are heavily loaded with
zoomorphic and anthropomorphic figures. Even though the Plaza Mayor can be considered a
ceremonial area, the difference between its size and that of the Circular Plaza suggests than
not all the participants of ceremonies in the quadrangular plaza were allowed to enter to the
circular plaza.
It is probable but not conclusive that the midden represents evidence of feasting
activities carried out in a space other than the Circular Plaza, probably the Plaza Mayor.
Nevertheless is necessary to be cautious. Even thought the Plaza Mayor seems to be an
interesting candidate there is no evidence that links the Plaza with the Midden deposits other
than educated assumptions.
But, why would the authorities of Chavín have invested in holding large feasts? In
Chapter 2 I illustrated how the ceremonial centers populating the Andes during the Late
Formative shared a basic iconographic set that has been characterized as chavinoid or
cupisnicoid. Richard Burger has suggested a peer polity interaction among ceremonial centers
during the Early and Middle Formative and John Rick and Silvia Kembel have argued for the
presence of a competitive condition among ceremonial centers during this period (Burger
1988; Kembel and Rick 2004; Rick 2005). Competition is one of the many ways that peer
polity interaction can manifest itself (Renfrew and Cherry 1986). The reason for this
competitive process lies in the need to gain prestige in order for the survival of the belief
system sponsored by authorities. Prestige is connected to reputation; reputation is internalized
by a social group that develops acceptance or rejection towards the religious system that is
materialized in the ceremonial center that possesses prestige, “competition for prestige
consists of rivalry for continual public recognition by supporters […] vying for prestige is the
equivalent of competing for people or their labor, power and support” (Clark and Blake
1996:260). Translating this to the case of a ceremonial center, competition could increase a
center’s prestige in order to lure more people into their religious system. In most cases
newcomers would contribute economically to the center in the form of labor or offerings.
Individuals had to decide which center to support or contribute to. Why were these social
systems inclined to ally with any center? The answer to this question is derived from the
143
definition of the concept of power. Power can be understood as the ability to do or influence
something or anything, or to operate upon a person or thing, it is “the probability that one
actor within the social relationship will be in a position to carry out his own will despite
resistance, regardless of the basis on which this probability rests” (Weber 1978:53). The
capacity to influence different ceremonial centers was also responsible for a center’s prestige.
Ceremonial centers presumably were initiating newcomers into their religious cult
(Kembel and Rick 2004; Rick 2005); this model indicates that elites were arriving at Chavín
de Huántar in order to be initiated in a belief system that would give them legitimate authority
and therefore rightful power. The system was supported on the “strategic manipulation of
traditional concepts” (Rick 2005:3) such as shamanism. Authority emerged in Chavin as a
result of this manipulation that in a sense would have created a collective consciousness
among members of the belief system
“… some aspects of Chavín iconography and perhaps ritual activity
derive from shamanistic origins, but it is doubtful that this monument
and its features can be seen as a result of system-serving activities of
a problem solving group of shamans. I believe that the familiarity of
shamanism, and its pre-existing acknowledgement of human contact
with powerful natural elements is a credible foundation for arguments
that those involved with Chavín practice (priests at the site, inductees
into the cult) could be imbued with nature-derived powers, or perhaps
even descendant from powerful natural ancestors” (Rick 2005:80)
According to Rick (Rick 2005), the use of drugs may have served to present an
alternative reality in which the connections with nature-derived powers and the relationship
with powerful ancestors would appear real and legitimate. In this regard, authority is
legitimized through the manipulation of ancient practices; people from different places would
have traveled to Chavín in order to participate in the system (Rick 2005). A major issue for
this Chavín strategy would have been how to materialize power and prestige in a competitive
environment. DeMarrais et al. (DeMarrais, et al. 1996) have argued that power and authority
can be materialized through ceremonial events, symbolic objects, public monuments and
writing. In the case of Chavín, obviously, writing has to be left out as a materialization of
power that was lacking. Nonetheless, at Chavín de Huántar the impressive architecture can be
considered by itself a statement of authority, and as previously discussed the plazas served as
places where attendees to ceremonies gathered, witnessed and participated in events prepared
by Chavín authorities. In this regard, food events can be considered one of the activities
sponsored by Chavín authorities as a part of their political strategy. Supra-household feasting
144
provides an opportunity for display of success as it implies an effective organization of labor
as well as the presence of a well-supported economic base, “trying to impress attendees means
obtaining and preparing labor intensive foods, drinks, serving vessels, prestige items and ritual
items” (Hayden 2001:30). As expensive as the organization and practice of feasting activities
may have been for the authorities of Chavín, they had an adaptive value; adaptive values are
“referred to behavior that generates some practical benefit for survival, reproduction, health or
standard of living” (Hayden 2001:28). In a context of regional competition, any possible
displays of success are a good way to advertise the benefits of a system of beliefs, especially
when trying to convince others to be part of it.
Numerous aspects of feasting operate as public counting and ordering
devices, which in turn reduce the vagueness of social and political
situation by promoting social comparison. For example, depending
upon the quantity and quality or resources mobilized for communal
feasts and the frequency which they are mobilized, feasting can be a
quantitative measure of the abilities of the host as an efficient, skilful,
vital and generous leader (Potter 2000:472)
Sponsored feasts can serve as environments in which ritual and knowledge are
controlled and also manipulated. Jar frequencies suggest that along with large amounts of
food, large amounts of beverages were consumed, alcoholic beverages being likely as they are
pervasive in world-wide feasting ceremonies (Blitz 1993; DeBoer 2001; Dietler and Hayden
2001; Hayden 1995; Jennings, et al. 2005; Potter 2000). Jars are to be the more appropriate
recipients for liquids as they have necks and restricted mouth orifices which are useful for
keeping the contents and are adaptations for containing liquids (Rice 1987). Bone tubes
provide evidence of the presence of psychoactive drugs involved in feasting activities.
Feasting at Chavín de Huántar was a way to materialize power. It was an avenue for
authorities’ propaganda, a way to control ritual knowledge and entice people into the system,
an opportunity for display of success. It was part of the convincing system created in order to
attract followers and contributions that came with them. Feasts and other give-aways are
settings where rank is made clear by serving order – like present-day pachamancas, the order
in which a person get his/her food and the amount of food given directly reflects rank in most
cases. Feasting was a conscious, important way of creating a competitive advantage that
allowed the authorities of Chavín de Huántar to maintain the prestige of the ceremonial center
and the flow of practitioners and their contributions to the site. I propose that Chavín elites
sponsored feasts to gain prestige and attract followers. Feasting may have occurred in the
rectangular plaza or even in the circular plaza of the ceremonial center as they seem to be the
145
most appropriate places. But others areas still unknown may have been the focus of these
activities. Authorities provided food and beverages to elites that traveled to Chavín to be
inducted into the Chavín religious system, developed by the authorities of the center in order
to self-serve themselves from it.
7.2.2 The Wacheqsa Sector as a domestic area
Before recent research in the Wacheqsa sector, domestic activities were only
suggested but never confirmed. Bennett and Tello (Bennett 1944; Tello 1960) thought that the
deposits they excavated were formed by domestic refuse but as I have shown, they only
excavated the Midden analytical unit. Other documentation of the presence of a domestic area
in the Wacheqsa sector came from Rosa Fung, who stated that the northern edge of the
Wacheqsa sector was a domestic area located on top of a pre-Chavín occupation “During our
last excavations at Chavín de Huantar, in domestic areas, we have found Kotosh-Kotosh
ceramics in deep strata, but it cannot be said that the superimposed Chavín ceramics descend
from them” (Fung 1975:199). It is important to point out that at the time Rosa Fung excavated
at Chavín de Huántar, the Urabarriu phase had not yet been recognized. The ceramic sequence
at Chavín was based on the vessels found in the Ofrendas Gallery and the Rocas Canal, and
with correlations mainly with the Kotosh sequence at Huánuco (Lumbreras and Amat 1965,
Lumbreras 1972, Izumi 1962). Burger´s excavations retrieved imported Kotosh-Kotosh
vessels associated with Urabarriu ceramics, indicating that ceramics identified by Burger as
Urabarriu were probably contemporary with the Kotosh-Kotosh phase at Kotosh. I concur
with Burger when he states that Urabarriu and Kotosh-Kotosh phases were contemporaries.
This in turn would indicate that the domestic units reported by Rosa Fung were actually
Chavín rather than pre Chavín.
In describing the extent of the earliest domestic settlement at Chavín de Huántar,
Richard Burger stated that
“The residential zone nearest the old temple extended to the banks of
the Huacheqsa River. Part of the sector nearest the monumental
architecture was probably occupied by the people responsible for the
religious activities and the construction and maintenance of the
buildings” (Burger 1992: 159)
The evidence recovered in my research indicates that Fung and Burger are correct in
stating the existence of an early domestic area in the northern area of the Wacheqsa sector, a
domestic area that occupied the north half of the Wacheqsa sector as suggested by the
146
extension of the Early Platforms analytical unit. But who were the inhabitants and what was
their relationship with the temple?
As quoted previously, Burger has suggested that people living in this area were either
responsible for the religious activities or responsible for the construction and maintenance of
the temple. This statement has two inferences, on one hand, it can be understood that
authorities lived in this sector, and on the other, that the laborers lived there. If authorities
lived in the Wacheqsa sector during the Urabarriu phase, one can expect the presence of
expensive constructions, high quality ceramics and a fair quantity of prestige items. Ceramics
recovered in the Early Platforms are not sumptuous; the great majority is formed by small
ollas sin cuello. House foundations have not been found but a fair number of wall fragments
made of clay have been retrieved suggesting that house foundations were made of non-
resistant materials. Food consumption was at the household level given the small predominant
mode of osc and bowls existing in the ceramic assemblage, inhabitants had access to special
materials such as shells and obsidian, but there is an absence of chrysocolla and anthracite
artifacts.
During Janabarriu times the situation changed drastically and the Stone Room
analytical unit provides relevant information. Structures are ubiquitous above the early
Urabarriu domestic settlement; they were made of two or three rows of quadrangular medium-
sized rocks. It seems that clay or mud was still used for construction of walls given the
evidence recovered. There is an increase in obsidian artifacts and the additional presence of
chrysocolla and anthracite artifacts previously unseen in the Early Platforms analytical Unit. It
is important to mention that chrysocolla and copper ore were present as raw materials which is
another difference from contrast to the Early Platforms analytical unit. Certainly the raw
material evidence is too limited to conclusively state that artisans were living in these
structures but it is a possibility that needs to be seriously considered. Nevertheless it can be
said that inhabitants of the stone structures had more access to special items such as obsidian
and anthracite mirrors as well as to raw materials like chrysocolla and copper ore. As
previously explained the ceramic assemblage analyzed from this analytical unit indicates that
food was consumed on a larger scale than the food consumed at the Early Platforms. The
inhabitants of the Stone Rooms analytical unit had durable structures and had access to items
such as anthracite, chrysocolla and copper ore that were not present in the Early Platforms
analytical unit. The appearance of previously unknown archaeological materials in the Stone
Rooms analytical unit is contemporary with the Black and White stage. During this stage the
147
ceremonial center peaked in terms of space utilization and monumentality. This is represented
by major transformations imposed on the landscape in order to accommodate new structures,
and is best exemplified in the modification of the Mosna River course and the leveling of the
terrain for the construction of the Plaza Mayor (Kembel and Rick 2004; Rick 2005).
But, why to be located in the Wacheqsa sector? The Wacheqsa sector was part of the
ceremonial center, located in a tinkuy, a geographical point where two rivers meet. Tinkuys
have been interpreted as important landmarks in Andean landscapes (Burger 1992; Lumbreras
1989). The Wacheqsa sector is also located in an area that can easily be monitored from the
top of Buildings C and D. In this regard, Kembel and Rick have argued that constant
construction at the ceremonial center could have been a representation of the Chavín rulers’
prestige and power and that it was part of the well-crafted, convincing system created by the
rulers (Kembel and Rick 2004). Analogously, the continuous display of activities related to the
maintenance of the ceremonial center in the Wacheqsa area could have served the same ends.
Given the growing nature of the ceremonial center, it would not be surprising to find evidence
of a permanent support population living in the Wacheqsa area. Among the various questions
that may be asked regarding the relationship between these inhabitants and the authorities of
the ceremonial center I would like to select two. What were the social statuses of the people
living in this area, and what social mechanisms did the authorities of Chavin utilize in order
to exert control over them?
Residents of the Wacheqsa sector were occupying an area immediately adjacent to the
monumental core, an area enclosed by the Wacheqsa and Mosna Rivers. As previously
mentioned, this fact allowed elites of the ceremonial center close control over the activities
developed in the Wacheqsa area. Residents were probably attached to the ceremonial center,
fulfilling the needs of those who were in charge of it. Inhabitants of the Wacheqsa sector
probably came voluntarily to Chavín after being enticed by a convincing system developed by
elites – a strategy that John Rick calls a devotional system. In this devotional system “the
societies’ individuals were deeply committed to, and willing to invest resources in a temple
center because of their intrinsic adherence to the religious concepts on which the system rests”
(Rick In press). According to John Rick, this system would have served for gaining local
support rather than attracting elites from distant areas.
148
“Ritual at Chavín was not only directed toward the delivery of
information from priests to recipients, as in oracular consultation, but
also in ceremonies meant to increase the devotion of a supporting
populace. In addition, the site served as a convincing system
particularly aimed at a small but key group of initiates to the cult, of
both local and distant origin”. (Rick 2005:108, emphasis added)
Indeed, the devotional system hypothesis is useful at the local level because it
explains how people were mobilized to work for the temple. This convincing system would
have produced different results according to the specific material conditions and motivations
of the people who were exposed to this system and the emergence of a devotional system
should be evident in areas local to the temple. This system could be related to a persuasive
theory of labor control that
Focus[es] on the proactive role of nascent elites who use a variety of
strategies such as the assumption of ideological power, cooption of
separate divine descent, the control and strategic redistribution of
exotic goods, the creation of economies of scale, and so forth.
(Stanish 2003:224)
The Wacheqsa area is enclosed by the junction of the Wacheqsa and Mosna rivers,
which provides a space of physical confinement for the people established there (Burger
1992). It also provides a space where the authorities can observe the inhabitants’ activities,
perhaps with control in mind. Arnold suggested that,
As rising elites begin to accrue power, privilege and status, they draw
increasingly economically dependent sectors of the population into
important production roles of labor intensive group activities. If rising
elites learn to control the information or technology critical to
economic success and thus orchestrate networks of interdependencies
that limit power outside their small circle, then nonelites become
marginalized from positions of substantial political or economic
influence (Arnold 1995: 208)
In Chavín, the religious was a specialized institution. It acted not only in matters of
religion, but also assumed the traditional roles and responsibilities of both a political, as well
as an economic system. The Chavín system was therefore a multiple system as it was
comprised of the functions of segregated systems in one agglomeration, the core of which
was fixed in the religious aspect but that also comprised the other institutions’ functions.
While these institutional activities are present in every social system, the preeminence of
one above the others gives a system its specific identity. Employing the multiple, the power
structure in the Chavín system acted primarily to normalize and validate the interests of
149
authorities. Its tools for enforcement and propagation of the power structure primarily
included deceptive actions, such as the abovementioned devotional system (to attract
workers), the deliberate physical location of laborers in a controllable area, the creation of
an outer world through impressive monumental architecture and the creation of a different
convincing system meant to entice authorities or elites from other areas.
Going a little further in my interpretation, I could hypothesize that under Chavín’s
multiple system, the rulers’ visual control and the strategic location of laborers would both
reinforce the hierarchical relationship between the elites and the ruled and also create a
compelling environment in which laborers would feel bound to increase their productivity.
Stanish noted that “in the absence of pressures or inducements to the contrary, households
substantially underproduce and underconsume relative to their economic capacity” (Stanish
2003: 23).
Elites managed labor, information and transportation. They also chose what to give
and what to place in circulation, therefore subjecting people to their authority. As mentioned
before, there were different facets of the Chavín multiple system, one of those facets being the
economic one. The inclusion of laborers in an internal circulation network of prestige-building
events within the system would produce and reinforce the dependent relationship between
workers and rulers but, in return, would have also provided workers with a different status in
comparison with those who do not participate in this internal system. That would explain the
presence of maritime resources, obsidian, chrysocolla, anthracite and copper ore in the Early
Platforms and Stone Rooms analytical units. During the Janabarriu phase, the inhabitants of
the Wacheqsa sector were part of at a craft specialist establishment located in the northern half
of it.
These theoretically driven reconstructions have the purpose of putting the domestic
settlements of the Wacheqsa sector and its midden in a broader anthropological perspective in
relation with the ceremonial center. The Wacheqsa sector was not isolated from what
happened at the monumental core, rather these two sectors were intimately related as I have
tried to explain in the previous pages. As I have mentioned in this chapter and in the previous
one, the analytical units identified in the Wacheqsa sector not only are spatially but also
chronologically segregated. Each of the analytical units is associated with a specific ceramic
assemblage which has chronological implications. Early Platforms and Water Flood analytical
units are associated with ceramics related to the Urabarriu ceramic phase, while Midden, Late
150
Platforms and Stone Rooms units are associated with Janabarriu ceramic phase. These two
ceramic phases were previously defined by Burger (Burger 1984; Burger 1998). As mentioned
in Chapter 2, Burger placed the Urabarriu phase between 850-460 BC and the Janabarriu
phase between 390-200 BC. He also proposed a third phase located between the Urabarriu and
Janabarriu ones, with a range of 460-390 named Chakinani. I have not found any analytical
unit exclusively associated with ceramics identified with this particular ceramic phase which
leads me to suspect that the Chakinani phase is not a good chronological marker, especially
after examining the carbon dates obtained from the Wacheqsa sector.
In the next section I discuss a set of 10 radiocarbon dates obtained from the Wacheqsa
sector, attempting to shed light on the absolute chronological markers of the occupations
identified in the Wacheqsa sector, as well as their associated materials. I believe these dates
are highly relevant, not only for the relationship between the ceremonial center and the
Wacheqsa sector, but also for the overall chronology of Chavín, the ceramic sequence and the
regional relationship between Chavín de Huántar and other sites from the Andean Formative.
7.2.3 Radiocarbon Dates
Ten carbon samples were submitted for carbon dating with results falling within the
range of 1200 – 800 BC for the Pre Black-and-White/Urabarriu phase and 800-500 BC for the
Black and White/Janabarriu phase. These dates have been calibrated using C14 calibration
software called OxCal V 4.0 (http://c14.arch.ox.ac.uk/embed.php?File=oxcal.html) and the
results are presented using a probability range of 99.7% (three sigmas). The curve used for
calibration was the south hemisphere SHCal 04 recommended by Leon (Leon 2006). In the
next few pages I will revise these dates in relation with each prehistoric analytical unit
identified.
7.2.3.1 Early Platforms
Four dates have been obtained from the Early Platforms analytical unit. The oldest
one (GX-31647) has a range of 1209-969 BC and it was obtained from a hearth located on top
of sterile soil in unit WQ-6. This hearth represents the oldest occupation of the Wacheqsa
sector. The ceramics associated with this hearth can be related to those identified as belonging
to the Urabarriu phase by Burger (1984, 1998), especially a bottle rim that is identical to type
B1 in Burger’s typology (figures 175-177).
151
The second date (AA-75385) obtained from the Early Platforms analytical unit comes
from a layer (layer 5) characterized by a semi-compact coarse sand mixed with medium and
large sized angular rocks located in WQ-1. This layer has been dated between 1000 and 833
BC and the ceramics recovered there are similar to the ones identified by Richard Burger as
belonging to the Urabarriu phase (figure 178) (Burger 1984; Burger 1998)
A third date (AA-75386) was obtained from the top of a clay floor (floor 3) associated
with a zig-zag wall located in the northern edge of the Wacheqsa sector in Unit WQ-2. The
date obtained falls within the range of 895-794 BC. A ceramic sherd that does not resemble
any Janabarriu-related type was found on top of this layer in direct association with the carbon
sample (figure 179).
A fourth date (AA-75387) was obtained from a clayish and sandy semi compact soil
mixed with abundant small sized cobbles and angular rocks, located in WQ-4 (layer 6). The
date obtained falls within the range of 836-538 BC. As the rest of the strata dated from this
analytical unit, there are no Janabarriu-like ceramics present in its ceramic assemblage, only
Urabarriu- like ones (figures 72, 180-181).
As previously stated there are no Janabarriu ceramics present in any of the strata
recorded from this analytical unit. These dates seem fairly consistent with the exception of
date AA-75387 which has too broad a range that overlaps with the Stone Rooms analytical
unit located stratigraphically above the Early Platforms analytical unit. Stratigraphic
relationships position this layer - where date AA-74387 was retrieved- below the Stone
Rooms analytical unit which as I will show has been dated between 794 – 413 BC and has an
overwhelming presence of Janabarriu like ceramics. Layer 6 in WQ4 is located
stratigraphically below Layer 3 in WQ8. Layer 3 in WQ8 is part of the same stone
architectural group present in WQ4 that is also present in WQ3 and WQ-6 which has been
identified as the Stone Rooms analytical unit. There is one layer between the floor associated
with the stone architectural group and layer 6 in WQ4. Two units stratigraphically located one
above the other may overlap in terms of carbon date ranges but cannot overlap in terms of use,
especially when there is another stratum between them. As McBird stated “Radiocarbon dates
from the Early Horizon are particularly hard to interpret due to two important short-term
increases in the C14:C12 ratio peaking at 370 and 730 BC” (Bird 1987:297). These short term
increases in the aforementioned ratio create a horizontal plateau that prevents more precise
dates within this time range (730-370 BC).
152
Consequently given the ranges of the other deposits dated in this analytical unit and
the range of the Stone Rooms analytical unit, a terminal date of 800 BC for the
construction/use of the Early Platforms seems to be appropriate. Dates obtained from the Early
Platform analytical and stratigraphic relationships indicate that it was probably formed
between 1200 and 800 BC. Based on this date, it is logical to state the Early Platform
analytical unit is associated with Urabarriu-related ceramics and was occupied before the
ceremonial center’s Black and White/Janabarriu stage.
7.2.3.2 Water Flood
One date (AA-75383) was obtained from the earliest water flood episode and fell
within the range of 900-798 BC. I would have expected for this deposit to be in the range of
GX-31647 (1209 -969 BC) as it sits on top of sterile soil. The explanation that I can offer is
that the area where the Water Flood analytical unit is located had a previous use and later
transformed into some sort of canal or water repository, probably for water coming from the
Wacheqsa River. The ceramics associated with this deposit resemble the ones identified as
belonging to the Urabarriu phase by Burger (figure 182) (Burger 1984; Burger 1998)
7.2.3.3 Midden
Three dates were obtained from this analytical unit, dating it within the range of 800-
500 BC. Date AA75389 gave a range of 811-551 BC and was retrieved from WQ-7, SIII, U4,
L8. The ceramics associated with this deposit are similar to the ones identified by Burger as
Janabarriu (figure 183) (Burger 1984; Burger 1998). Date AA75384 falls within the range of
791-521 BC. It was retrieved from WQ7, SIV, U4, L14. The ceramics associated to this
deposit are similar to the ones identified by Burger as Janabarriu (figure 184) (Burger 1984;
Burger 1998) and Date AA75382 falls within the range of 770- 486 BC. It was retrieved from
WQ7-SIV, U4, L10. The ceramics associated to this deposit are similar to the ones identified
by Burger as Janabarriu (figure 185) (Burger 1984; Burger 1998)
Dates point towards the formation of the Midden analytical unit between 800 and 500
BC, establishing the feasting activities inferred from this unit as contemporary with the
architectural Black and White stage of the ceremonial center.
153
7.2.3.4 Stone Rooms
One date (AA75390) has been retrieved from this analytical unit. It came from a deposit
located on top of a floor associated to a stone room alley located in unit WQ-8. This deposit
has been dated within the range of 790-510 BC and the ceramics associated are similar to the
ones identified by Burger as Janabarriu (figure 186) (Burger 1984; Burger 1998)
7.2.3.5 Late Platforms
One date (AA 75388) was retrieved from this analytical unit. It came from WQ7,SIII, U1
layer 9 which was a very compact surface whose fill was composed of a cobble fill. This
stratum had a special character as it was the one with major density of decorated ceramic
sherds and therefore gave me indications of its relative chronological position. The ceramics
retrieved from this layer were Janabarriu-related ceramics although there is particular
punctuated sherd that may be Urabarriu related (figure 187). This deposit gave a date of 980 –
772 BC. It is most likely dating the first construction episodes of the Late Platforms analytical
unit given its stratigraphic depth. This early date is quite surprising. I would have expected for
these platforms to be at least contemporary with the early dates of the Midden analytical unit.
These dates may indicate a partial overlapping between the Late Platform analytical unit and
the Early Platforms one. This date does not mean that the Late Platform analytical unit
functioned during the range established by this date as the deposit dated is one of the earliest
ones in stratigraphic terms. Upper strata in this analytical unit are located at the same
stratigraphic level as deposits in the Midden and even Stone Rooms analytical units. What is
certainly remarkable is such an early date for Janabarriu related ceramics that might have
overlapped in time with the last portion of the Urabarriu phase. It would not be surprising to
find late Urabarriu contexts contemporary with early Janabarriu ones as people do not stop
producing types and designs in an abrupt way, transitional times are the ones in which early
and late styles meet and coexist together. However caution needs to be exercised regarding
this date until more contexts in this analytical unit are dated in order to corroborate this early
chronological marker.
The dates from the Wacheqsa sector point towards the segregation of two ceramic
components distributed in five analytical units. Urabarriu ceramics have been found in the
Early Platforms and Water Flood analytical units and these analytical units have a
chronological range of 1200-800 BC. Dates obtained from the Wacheqsa sector put the earlier
of these two components at Chavín de Huántar as early as 1200 BC which confirms Rick and
154
Kembel’s assertion of the antiquity of Chavín de Huántar (Kembel 2001; Kembel and Rick
2004; Rick 2005, 2006). A close examination of the Urabarriu dates published by Burger will
also support this claim. Dates obtained from excavations outside the ceremonial center, ISGS-
493 (1446-802 BC), ISGS-486 (1122-801 BC) and UCR-694 (1133-746 BC), although
problematic because of their broad standard deviation, are quite consistent with the Urabarriu
related dates from the Wacheqsa sector. Date UCR-705 (906-411 BC) is problematic given the
broadness of its standard deviation falling right into what I consider the chronological limits of
the Janabarriu phase. Date ISGS-510 (361-272 BC) falls right in what I consider post-Chavín
times and has been rejected by Burger (Burger 1981). But if we only consider the 1200 BC
limit for Burger’s Urabarriu phase, this is a phase that lasts at least 400 years (figure 188).
There would have to have been an unusual stability within the social processes occurring at
Chavín de Huántar to maintain the ceramic unity of a phase for that long. Lumbreras
suggested that the ceramic styles defined as Urabarriu are not understood in detail and that
there probably is more than one phase in the Urabarriu assemblage excavated by Burger which
would make sense given the notably long 400 year period (or 700 year period if we consider
the outer limits of sample ISGS-493): “it is evident that the Urabarriu phase as presented by
Burger is unsatisfactorily known and certainly contains components of more than one phase
that are susceptible to be analyzed” (Lumbreras 1993:354). Nevertheless, these dates obtained
in the site periphery need to be related to dates obtained inside the monumental core to
establish the absolute chronology of the pre-Black and White architectural stages identified
by Silvia Kembel (Kembel 2001)
Janabarriu-related ceramics appear in the Wacheqsa archaeological record in three
analytical units: Midden, Late Plaforms and Stone Rooms. Unfortunately the Janabarriu dates
fall within the calibration plateau mentioned above making it difficult to precisely determine
the overall date range. Nevertheless the dates and their ranges all come before the 400-200 BC
span established by Richard Burger as the chronological range for the Janabarriu phase at
Chavín de Huántar. In a broader sense, it can be stated that the Midden was formed between
800 and 500 BC while the Stone Room analytical unit probably dates to around 790-510 BC.
The Wacheqsa dates establish Janabarriu-related ceramics as contemporary with the
Black and White Stage as defined by Kembel (Kembel 2001). It also puts the major
topographical transformation event in the Wacheqsa sector contemporary with the Black and
White Stage. Feasting activities identified at the Midden analytical unit and the Janabarriu
domestic settlement identified were also contemporary with the Black and White stage. As
155
suggested in Chapter 2, I believe it is fair to equate the Black and White stage with the
Janabarriu ceramic phase as dates indicate that contexts associated with Janabarriu related
ceramics are contemporary with the maximum construction effort at Chavín de Huántar
represented in the Black and White stage. Also in Chapter 2 I explored the relocation of the
Janabarriu ceramic phase that has been explicitly suggested by Bischoff and Inokuchi (Bischof
1998; Inokuchi 1998) and suggested by a close examination of the dates and ceramic
assemblages recovered in the sites of Kotosh, Garagay, La Pampa, Pacopampa and Kuntur
Wasi (Izumi and Sono 1963; Onuki 1995; Ravines, et al. 1982; Seki, et al. 2006; Terada
1979). This relocation of Janabarriu is not surprising considering the limitations in the original
dating by Burger who stated that “the principal weakness of the hypothetical [Chavín ceramic]
sequence comes from inconsistent readings on the Janabarriu samples” (Burger 1981:595)
7.2.4 Regional chronological implications
Chavín has been understood either as a Mother Culture, as a derivation from Cupisnique or as
synthesis from of north and central coast developments (Burger 1988; Burger 1992; Larco
1945; Tello 1942). Dates obtained from the Wacheqsa sector and even the Urabarriu dates
obtained by Richard Burger put Chavín in a contemporary relationship with most of the
centers of the north and central coast, at variance with what Richard Burger has previously
stated:
“An evaluation of the radiocarbon measurements available suggests that the ceremonial centers of Caballo Muerto, Haldas, and Garagay were prospering on the coast of Peru between 1200 B.C. and 900 B.C. In contrast, the earliest phase of the religious center at Chavin de Huantar is estimated as lasting from 850 B.C. to 460 B.C., on the
basis of two sets of radiocarbon analyses28
” (Burger 1981:596)
Granted the U-shaped building architectural tradition seems to come very early in the
Early Formative according to the dates of La Florida and Mina Perdida (Burger and Gordon
1998; Patterson 1985), but nevertheless, dates from Cardal and Garagay place them roughly at
the same time that the Urabarriu phase at Chavín de Huántar in the Middle Formative.
In the North Coast the situation is no different. As mentioned in Chapter 2, a re-
evaluation of carbon dates made by Bischof points towards the placement of Huaca de Los
28 However, it is important to indicate that Burger used uncalibrated carbon dates in making these
assertions. The use of uncalibrated dates adds a different level of complexity that has to be taken into
consideration as it assumes that the level of atmospheric C14 has always remained constant. The
calibration of radiocarbon dates have a consistent aging effect (Ziólkowski et al 1994).
156
Reyes into the Middle Formative rather than into the Early Formative as postulated by
Pozorski (Bischof 1998; Pozorski 1975); the architectural critique of Conklin (Conklin 1985)
also reinforces Bischof’s argument. Huaca de los Reyes –or to a larger extent the architectural
complex of Caballo Muerto- has been understood as the typical, and possibly the most
important Cupisnique center during the Andean Formative. There are grounds to see Chavín
de Huántar as contemporary with Huaca de los Reyes and that the chavinoid murals and
sunken plaza from Huaca de los Reyes did not precede, but were contemporary with Chavín
de Huántar.
If Chavín was contemporary with Cupisnique, Larco’s hypothesis of Chavín as a
product of Cupisnique diffusion does not stand. Late Archaic ceremonial centers as well as
Early Formative centers from Casma and the central coast contradict Tello’s hypothesis of
Chavín as a Mother Culture. It is pretty clear that Chavín was not the source from where
civilization radiated but is also clear that Chavín was not a derivative of Cupisnique or a
product of north coast and central coast population displacement as suggested by Larco and
Burger respectively (Burger 1981, 1988; Larco 1945). In this regard my findings are in
agreement with Kembel when she states that
“Neither a precursor to the monumental centers of the late Initial
Period [Middle Formative] and the early Early Horizon [Late
Formative], nor the late consequence of their collapse, the
monumental center at Chavín de Huántar appears to have been coeval
to these centers, and part of a network of centers that declined by the
middle of the first millennium B.C.” (Kembel 2001)
The dates from the Wacheqsa sector indicate that Chavín de Huántar emerges at the
onset of the Middle Formative, contemporary with the complex monumental architecture of
the north highlands, north coast and central coast.
In this Chapter I have provided evidence regarding the use of the Wacheqsa sector as
a domestic settlement and waste area for suprahousehold eating and drinking activities during
the Middle and Late Formative periods. I have located these activities not only in spatial but
also in chronological terms which in turn label the Wacheqsa Sector as a multicomponent area
within the ceremonial center of Chavín de Huántar. I have argued that these activities have
implications for the understanding of power relationships in Chavín de Huántar, one aimed
towards local populations and the other towards foreign elites. The dating of these activities is
also tremendously important for the chronological location of Chavín within the Andean
157
Formative and for the understanding of processes of regional interaction during this period. I
expand these conclusions in the next Chapter, hoping that they will instigate further debates
regarding Chavín de Huántar and the Andean Formative in the future.
158
CHAPTER 8
CONCLUSIONS
My research presents a new methodology for the investigation of intrasite space
organization of stratigraphic components recorded through sampling programs. I have
sampled 132 m² (0.9%) of a total population of 1.4 ha, recorded 200 strata, and 23 features.
These elements have been grouped into eight analytical units which encompass three
chronological phases. I have demonstrated that careful sampling programs can be extremely
advantageous in investigating intrasite variation, in particular when all stratigraphic record is
modeled using computer aid design (CAD) which allows the spatial reconstruction of
stratigraphic spatial distribution in a flexible model. I emphasize the word flexible as the
model can be observed from different visual perspectives that provide views that in many
cases are unachievable in the real world. As emphasized in this dissertation, stratigraphic
modeling is particularly useful at the level of intrasite examination of deposits and the
relationships among themselves, in particular when identical strata are identified in different
areas within the site. Sampling programs allow this identification and CAD modeling tools
allow the spatial reconstruction of this relationship which has been recorded in a dispersed
way during the sampling program.
Why is this important? The advantages of this joint methodology are particularly
relevant not only in as a research tool but also as a conservation one. Complete site excavation
is a problematic endeavor, not only in financial aspects but also in conservation ones. In this
regard sampling programs can be particularly efficient in extracting information when area
excavations are not feasible due to financial constraints or conservation issues. It is necessary
to emphasize the archaeologist’s responsibility in protecting archaeological sites and in many
cases this protection implies restrictive excavations. The site of Chavín de Huántar is part of
the UNESCO world heritage list and therefore special care had to be taken in carrying my
research, being this one of the reasons why I decided to sample the sector instead of planning
for large area excavations. Currently a Management Plan is being elaborated for Chavín, plan
that –among many things- will regulate what can be excavated and how it can be excavated.
With the increase number of management plans being developed for different archaeological
sites across the world, more restrictions will be enforced by these plans, and sampling
programs emerge not only as excavation tools but also as conservation one. But with each
sampling program there is a risk. Large sampling programs face the problem of organizing
159
large data sets that in many cases are spatially dispersed. For regional sampling programs,
digital modeling tools such as GIS are most recommended given its potential for working at a
regional level, but at the site level and specially at the subsurface level, CAD modeling is the
most adequate three dimensional modeling tool as allows the modeling of small strata and
large strata with equal level of detail and flexibility and allows the management of large
stratigraphic data sets obtained through sampling than can be better comprehended with the
used of CAD modeling technologies. But most important, the conclusions I present in this
chapter can be tested in the field as 99% of the Wacheqsa sector is untouched and awaiting
further careful research.
With the information extracted through careful purposive and systematic sampling,
together with the three dimensional stratigraphic modeling, I have segregated eight spatial
analytical units, five prehistoric and three modern ones. I have demonstrated that all
prehistoric analytical units (Early Platforms, Water Flood, Midden, Stone Rooms and Late
Platforms) show indexes of high diversity according to the calculated Boone index for each
deposit. But there is variability within this diversity. Deposits in each analytical unit in spite of
showing nearly similar values of diversity reflect variation in their archaeological
assemblages. I used bivariate kernel density estimations in order to investigate ceramic
modalities for each prehistoric analytical unit and comprehend the nature of the activities
developed in each unit cross referencing this line of evidence with the distribution of
archaeological materials in each prehistoric unit.
Using these methods, I have determined that during 1200-800 BC the north section of
the Wacheqsa sector was the focus of a settlement inhabited by people whose ceramic
assemblage can roughly be labeled as Urabarriu. They were cooked in small ollas sin cuello,
ate on small bowls, and had access to exotic resources such as shells, chrysocolla and
obsidian. The hypothesis of a settlement located in this section was previously suggested by
Wendell Bennett, Julio C. Tello, Rosa Fung, Richard Burger and Luis Lumbreras (Bennett
1944; Burger 1984; Burger 1998; Fung 1975, 2006; Lumbreras 1989; Tello 1960) and my
research has demonstrated that there was an occupation of domestic nature judging by the
archaeological materials recovered. Towards the end of the Urabarriu phase around 1000-800
BC, the southern portion of the Wacheqsa sector was used as a water flooding area, probably
canalizing water coming from the Wacheqsa River, as inferred by Tello (Tello 1960). The
ceramics found in this canal can be related to the ceramics defined by Richard Burger as
Urabarriu. These two analytical units encompass the Urabarriu phase at the Wacheqsa sector
160
that spanned from 1200 to 800 BC according to the associated carbon dates. Then,
approximately by 800 BC, there was a reconfiguration of the occupation in the Wacheqsa
sector. Inhabitants cooked their aliments in large ollas, used medium sized bowls as serving
vessels, built their structures using solid foundations, and continued to have access to shells,
chrysocolla, and obsidian, but also had access to copper ore and anthracite mirrors.
Chrysocolla and cooper ore found in these structures are present in the form of raw materials.
It is tempting to hypothesize that people living in the Wacheqsa sector around 800 BC were
craftspeople producing artifacts required by the authorities of Chavín, it would not be
surprising at all given the geographic convenience of the Wacheqsa sector for holding
craftspeople as it is an area that could easily be monitored from the ceremonial center and it is
conveniently close.
I have hypothesized that the inhabitants of the Wacheqsa sector – during Urabarriu
and Janabarriu times – were people affiliated with the ceremonial center, who were
developing specifics activities yet to be determined. I base this assertion on the on the
geographical division imposed by the Mosna and Wacheqsa rivers between areas inside and
outside of the ceremonial center, the visual control that can be exerted from monumental core
Buildings C and D, the evidence of structures and the evidence of food consumption
represented in large cooking and serving vessels as well as the presence of faunal remains in
the archaeological assemblage.
Around the time of the Janabarriu domestic settlement, a large amount of garbage in
the form of a large midden was discarded on top of the Urabarriu period canal. Large
cooking vessels, medium sized serving vessels, a large number of jars plus an unusually high
density of faunal remains suggest that together these materials are the remains of supra
household eating and drinking events. Additionally indirect evidence of psychoactive drug
consumption during these events in the form of small polished bone tubes have been found
associated with the elements mentioned above, probably related with the consumption of
anadenanthera as suggested by other researchers (Rick 2006; Torres and Repke 2006). The
ceramics recovered in the midden are related to the ceramic assemblage defined by Burger as
Janabarriu, locating the Janabarriu phase at around 500-800 BC. There is no further
archaeological occupation in the Wacheqsa sector in other prehistoric periods, although a
handful of Huaraz and Recuay sherds were found in the agricultural land analytical unit.
In terms of the relation of these analytical units and their phases with the architectural
core, the Urabarriu occupation at the Wacheqsa sector cannot be equated yet with a specific
161
architectural stage at the monumental core. However, it can be stated that the Early Platforms
and Water Flood analytical units are Pre-Black and White stage given the carbon dates
available and their pre-Janabarriu ceramic assemblage; further dates from the monumental
core will accurately relate the early architectural stages of the monumental core to the
Urabarriu phase from the Wacheqsa sector. On the other hand, it is possible to correlate the
Midden, Late Platforms and Stone Rooms analytical units to the Black and White stage at the
monumental core. Dates from the Midden and Stone Rooms analytical units are consistent
with the dates coming from Black and White structures at the monumental core. Also the
occurrence of feasting activities during the Black and White stage is consistent with the
change of emphasis in the monumental architecture, shifting the focus from small gallery
patios and volumetric construction to large open areas in which rituals, ceremonies and feasts
might have taken place but caution is needed in linking the feasting remains found in the
midden with any of the plazas from Chavín. Although this hypothesis seems logical, there is
no evidence of such activities found in any of the plazas, probably because of the heavy post-
Chavín occupation of the Circular Plaza and the use of the Plaza Mayor as agricultural land
for several centuries.
The construction of the Late Platforms analytical unit started at the beginning of the
Janabarriu phase based on one associated carbon date and the ceramic assemblage found in
one of its deposits. The extreme low density values of archaeological materials precludes me
from making more inferences other than to state that this analytical unit was kept clean for
almost the whole duration of the Janabarriu phase. This was an intermediate area between the
Stone Rooms and Midden analytical units, probably a small buffer zone between the
residential area and the large midden deposit. The Midden, Stone Rooms and Late Platform
analytical units constitute the Janabarriu phase deposits of the Wacheqsa sector.
The correlation between the largest architectural project and the Janabarriu ceramic
style(s) and its regional distribution serves to partially solve the problem regarding the
disjunction between ceremonial architecture and Janabarriu-related ceramics (Kembel 2001;
Kembel and Rick 2004). The ceramic assemblage that belongs to the Janabarriu phase is for
the most part contemporary with the Black and White phase and represents a time of intense
regional interaction that Burger has labeled as the “Chavín Horizon”. He placed this horizon
chronologically between 400 and 200 BC, which as I showed in this dissertation is a time
range that is definitely post-Chavín. Radiocarbon dates from the Wacheqsa sector indicate that
Chavín de Huántar was contemporary with the Cupisnique settlements from the North Coast
162
and with the U-Shaped buildings from the central coast, not to mention the early architectural
phases of Kuntur Wasi and Pacopampa. This contemporaniety requires a reevaluation of the
concept of Burger’s Chavín Horizon. Chavín de Huántar was neither the origin of social
complexity in the Andes, nor a product of coastal migration nor a synthesis of early social
developments derived from the north and central coast. There is new evidence that suggests a
long settlement history that goes back to the Late Preceramic in La Banda (Wolf 2005)and in
the modern town of Chavín de Huántar (Rick and Mesia 2006) which in turn indicates that
Chavín de Huántar had a support population inhabiting the Conchucos valley before the
Formative, a population that is likely to have been ancestral to those who constructed the
ceremonial center of Chavín de Huántar.
My research has also allowed me to propose that the existing ceramic sequence at
Chavín de Huántar needs to be revised in the light of the new evidence available.
Quantitatively and qualitatively more data has been retrieved since the late 1970’s that needs
to be included in a reconstruction of the existing ceramic sequence. The Urabarriu phase is not
only older than thought by Burger but also broader in terms of duration. In general terms I
have observed two broad ceramic assemblages in the Wacheqsa sector, Urabarriu and
Janabarriu; the phase labeled as Chakinani is not present in the Wacheqsa sector. As
mentioned in the precedent chapter, a refined subdivision of the ceramics belonging to the
Urabarriu ceramic phase is needed in order to accurately assess ceramic changes in a 400 year
period. As stated above, further ates from early architectural stages will help to illuminate the
relationship within the Wacheqsa sector and the monumental core during the Urabarriu phase.
But the Urabarriu ceramic phase is not the only one that needs to be revised. Excavations at
the Circular Plaza (Lumbreras 1989) uncovered a layer formed by the abandonment of the
ceremonial center associated with Janabarriu related ceramics on top of it (Lumbreras
1989).This evidence suggests the existence of a late Janabarriu component that is actually
from after the formal ritual use of the ceremonial center. The segregation of Chavín and any
post-Chavín Janabarriu components is important not only for understanding the ceramic
sequence of Chavín de Huántar but also for the comprehension of the abandonment of Chavín
as a ceremonial center and for the refinement of regional relative chronologies in order to
accurately identify the late Janabarriu ceramic assemblages in other formative sites.
In terms of political strategies, I propose that at the Wacheqsa sector there is evidence
of two convincing strategies developed by the authorities of Chavín de Huántar, one aimed
towards gaining local support and the other aimed towards gaining the support of elite’s from
163
out of Chavín. Large eating and drinking parties were material manifestations of power and
prestige, a symbol of labor organization and access to resources. Chavín was offering wealth29
,
but different types of wealth to different classes of people. As proposed by Rick and Kembel
(Kembel and Rick 2004; Rick 2005) elites from outside the Conchucos area came to Chavín in
order to be initiated in the Chavín religious system and legitimize their authority in their local
areas; they obtained the wealth of authority and authority became a commodity. On the other
hand, inhabitants of the Wacheqsa sector worked for the ceremonial center, having access to
resources most likely provided by the authorities of Chavín in exchange for their labor and in
addition they obtained religious fulfillment by their participation in the maintenance of the
temple. The management of these two strategies by the authorities of Chavín was the key to
their success during the Middle and Late Formative.
The results of the present research lead me to conclude that the Wacheqsa sector was
occupied for 700 years (1200-500 BC) during the Middle and Late Formative periods and its
occupation was contemporary with the ceremonial center of Chavín de Huántar. It ceased to
be occupied and used as a dumping area by no later than 500 BC, which is consistent with
estimates for the collapse of the ceremonial center (Kembel 2001; Kembel and Rick 2004;
Rick 2005) .
Further research on the Wacheqsa sector needs to segregate the different ceramic
styles present in the Urabarriu and Janabarriu pottery assemblages. The vast sample size
excavated from stratigraphically controlled deposits (6272 decorated sherds and 12017
diagnostic ones) provides an excellent opportunity to quantify, segregate and identify stylistic
variation within each analytical unit’s chronological level. Needless to say this analysis needs
to be heavily quantitative in order to arrive at sound qualitative results. The same needs to be
done with the 215 kilos of faunal remains recovered in order to make statistically significant
assertions regarding intrasite consumption patterns in the Wacheqsa sector. These two lines of
research will be crucial in the years to come to add more relevant information to that provided
in this dissertation.
The topics discussed in the present dissertation are important for the understanding of
the social processes that occurred in the Wacheqsa sector and Chavín de Huántar during the
Middle and Late Formative. Equally these topics are important for the general understanding
of the Formative period. Hopefully some of the information presented here will shed light
towards a better comprehension of this fascinating period in Andean prehistory.
29 I am using Stanish’s definition of wealth “what people will work for” (Stanish 2004: 20)
164
APPENDIX A: ILLUSTRATIONS
166
Figure 01: Satellite photograph of Chavín de Huántar showing the different sectors of the site. The modern town is located at the north of the site
165
Figure 02: Map of the monumental core of Chavín de Huántar. Redrawn from Kembel 2001: Figure 1.3.
167
Figure 03: The Wacheqsa sector viewed from the site of Shallapa, west of the Chavín de Huántar.
Figure 04: The Wacheqsa sector viewed from the top of Mound D at Chavín de Huántar.
168
Figure 05: Garagay date associated to Janabarriu-like ceramics
Figure 06: Garagay dates (arrow indicates date associated with Janabarriu-like ceramics)
169
Figure 07: Examples of Janabarriu-like ceramics from the Wacheqsa sector
Figure 08: Chavín dates recovered by Burger. Black circle indicates the only Janabarriu valid date. Red circle indicates the other two Janabarriu dates. Blue circle indicates Chakinani dates. Note how
the Janabarriu date is earlier than the Chakinani dates
170
Figure 09: Huarás date GIF-1079
Figure 10: Dates from the site of Kotosh. Black circle encloses dates from the Kotosh-Chavín
phase. Blue circle indicates dates from the Kotosh-Kotosh phase.
171
Figure 11: Dates from the site of La Pampa. Black circle encloses dates from La Pampa phase. The
rest of the dates are from the Yesopampa phase.
Figure 12: Kunturwasi dates from the Kunturwasi phase.
Figure 13: Huarás dates published by Lau (2002). Note how they occupied time slot between 390-
200 BC.
Figure 14: Map of the Wacheqsa sector
172
Figure 151: The modern town of San Pedro de Chavín, the Wacheqsa River and the northern portion
of the Wacheqsa sector after the 1945 landslide.
Figure 16: Tello’s excavations at the Wacheqsa sector. Note the pirka wall behind the workers as well as the two eucalyptus trees. At the background the quebrada of the Wacheqsa River can be
seen2.
1
Tello Archive, MNAAHP, AT-595-2001.
174
Figure 17:3Terracing of the Wacheqsa sector. Note the cop plantations on its surface.
2 Tello Archive, MNAAHP, AT-594-2001. 3
Tello Archive, MNAAHP, AT-595-2001
175
Figure 18: Agricultural field’s delimited by pirkas4
Figure 19: Agricultural terraces. Notice the old chapel on top of Mound B. 5
4 Tello Archive, MNAAHP, AT-595-2001
176
Figure 20: Retention wall along the Wacheqsa River. Notice the houses on the Wacheqsa sector. 6
Figure 21: Retention wall along the Wacheqsa River 7
5 Tello Archive, MNAAHP, AT-595-2001 6 Tello Archive, MNAAHP, AT-595-2001 7
Tello Archive, MNAAHP, AT-595-2001
Figure 22: Map of Bennett’s, Tello´s and Fung´s inferred excavations. Fung´s units are inferred from personal communication.
177
Figure 23: Tello´s excavation profile and associated ceramics recovered.
Figure 24: Ceramics recovered by Rosa Fung. Unit H2, Layer 5, Level 2
179
Figure 25: Ceramics excavated by Rosa Fung. Test Pit 3, Layer 2, and Level 2
180
Figure 26: Screenshot of the process of modeling stratigraphy from the Wacheqsa sector. Green
surface represents the modern surface.
Figure 27 Texture subsurface strata modeled with Autodesk Land
181
Figure 28: Wireframe model of strata from the Wacheqsa sector
Figure 29: Same strata as above after textures are applied.
182
Figure 30: Excavations at the Wacheqsa Sector
183
Figure 31: Wacheqsa sector before excavations started in 2003 seen from the southwest. On the
foreground to the south (from left to right) Marino Gonzales’ house, Mound D and Mound C.
Figure 32: Systematic sampling strategy used in year 2005.
185
Figure 33: Location of units WQ1 (left) and WQ2 (right) on the north edge of the Wacheqsa sector.
Notice Building B in the background.
Figure 34: Stone platform located on WQ1.
186
Figure 35: Excavation of Feature 1 in WQ1-AW
187
Figure 36: South profile of WQ1 and WQ1-WE
188
Figure 37: Visible walls at the northern edge of the Wacheqsa sector
189
Figure 38: Exposed deposits in WQ2 before excavations started
Figure 39: Feature 02 in WQ2, Layer 03 and exposed section of Floor 2.
190
Figure 40: Excavation of WQ 3 (left), and WQ 4 (right). On the background, Mound B is observed
Figure 41: Excavation of aluvión layer in WQ 3. Notice the Wacheqsa River on the background.
191
Figure 42: WQ 3, stone platform (Layer 2a) associated to wall (Feature 01).
Figure 43: Stone platform, wall (Feature 01) and floor associated
192
Figure 44: Plan of exposed architecture in WQ4
Figure 45: Stone room, associated floor and wall (Feature 2) that delimits an alley
193
Figure 46: Layer 8 and associated Features 03 and 04
194
Figure 47: WQ-4, East Profile
195
Figure 48: Location of WQ 5
Figure 49: Hearth excavated in WQ 6
196
Figure 50: Panoramic view of WQ 7
Figure 51: Profile of WQ7, SIU1
197
Figure 52: Plant drawing of damaged stone platform (Layer 09)
Figure 53: Stratigraphic section of WQ7, SIU1
198
Figure 54: Profile of WQ 7, SIIU1
Figure 55:. Plan of architecture exposed in WQ 7, SIIU1
199
Figure 56: Stone platform in WQ7, SIII, U2
Figure 57: Profile of WQ7, SIII, U4
200
Figure 58: East profile of WQ7, SIII, U4A
Figure 59: Excavation unit WQ7, SIII, U4A
201
Figure 60: Excavation units WQ 7, SIII, 2, 4A and 4 (from left to right)
202
Figure 61: South profile of WQ7-SIV-U3
203
Figure 62: Stratigraphic detail of WQ7-SIV-U4
204
Figure 63: WQ8, plan of architecture exposed
204
Figure 64: Spatial distribution of analytical units in the Wacheqsa Sector
205
Figure 65: Water Flood analytical unit
206
Figure 66: Water Flood analytical unit. Common depositional events in units excavated.
207
Figure 67: Detail of WQ 4. Stone Rooms analytical unit on top Early Platforms
208
Figure 68: Ceramics recovered in unit WQ-1 WE, layer 8
Figure 69: Stone miniature found in WQ-4, layer 6
209
Figure 70: Ceramics recovered in WQ4-layer 7
Figure 71: Ceramics recovered in WQ4-layer 6
210
Figure 72: Idem
211
Figure 73: Stratigraphic relationship Midden analytical and Water Flood analytical units
212
Figure 74: Stratigraphic modelling of Midden Analytical Unit.
213
Figure 75: Slate projectile points recovered in the Midden Analytical Unit
Figure 76: Bone artefacts recovered in the Midden Analytical Unit
214
Figure 77: Slate projectile point recovered in WQ7, SIII, U4A, L10
Figure 78: Molded Frieze recovered in WQ7, SIII, U4A, L16
215
WQ7, SIII, U4, L5
WQ7, SIII, U4A, L7
WQ7, SIII, U4, L7a
WQ7, SIII, U4A, L15
WQ7, SIV, L13
Figure 79: Ceramics found in Midden Analytical unit
216
WQ7, SIII, U4A, L15
WQ7, SIII, U4A, L12
WQ7, SIV, U4, L14
Figure 80: Ceramics found in the Midden Analytical unit
217
Figure 81: Ceramics found in the Midden Analytical unit
Figure 82: Fragments of columns
218
Figure 83: Fragments of burnt architectural features (floors)
Figure 84: Ceramics retrieved from WQ4, L2
219
WQ-4, L 2
Figure 85: Ceramics retrieved from the Stone Rooms analytical unit
220
Figure 86: Fragment of unworked chrysocolla.
Figure 87: Fragment of cooper ore
221
Figure 88: Spatial distribution of Stone Rooms analytical unit in relation with Late Platforms and Midden units
222
Figure 89: WQ8, stratigraphic relationship between Early Platforms and Stone Room analytical units
223
Figure 90: Aluvion and Agricultural Land sections
224
Figure 91: Stratrigraphic Harris Matrix of strata recorded
225
n
Cu
bic
Mete
rs
Volume Excavated
25
20
15
10
5
0
Water Flood Early Platforms Midden Late Platforms Stone Rooms
Prehistoric Analytical Units
Figure 92: Volume excavated per Analytical Unit
Density of Archaeological Materials per m³
800 700 600 500 400 Series1 300 200 100
0 Water Flood Early Midden Late Platforms Stone Rooms Platforms
Prehistoric Analytical Units
Figure 93: Density of archaeological materials per Analytical Unit
226
Hi
1
0.9
0.8
Analytical Unit
Early Platforms
Late Platforms
Midden
Rooms
Water Flood
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
1 2 3
4 5 6 7 8 10
20 30
40 50 60
80 100
200
300 400 500 700 1000
2000 3000 4000
Sample Size
Figure 94: Analytical Units’ Boone Index. Note the spatial clustering of deposits bounded by
colored lines.
227
log
(sa
mp
Siz
e)
Hi
1
5
10
5
0
10
0
50
0
50
00
0
.0
0.2
0
.4
0.6
0
.8
1.0
EP LP MD RM WF
Figure 95: Distribution of analytical units according to sample size (log)
EP LP MD RM WF
Analytical Unit
Figure 96: Distribution of analytical units according to Hi values
228
Hi
0.0
0
.2
0.4
0
.6
0.8
1
.0
p(H
i)
0.0
0
.2
0.4
0
.6
0.8
1
.0
1 5 10 50 100 500 1000 5000
sample size
Figure 97: Confidence interval (90%) of Hi values. Upper section contains 85% of the confidence
interval; lower section contains 5% of it
1 5 10 50 100 500 1000 5000
sampSize
Figure 98: Probabilities of Hi [p(Hi)] values after 10000 repetitions using a Monte-Carlo routine.
229
p(H
i)
0.0
05
0.0
10
0.0
20
0.0
50
0.1
00
0.2
00
0
.50
0
1.0
00
EP LP MD RM WF
Figure 99: Distribution (log) of p(Hi) values per analytical unit
60.00
50.00
40.00
30.00
20.00
10.00
Bowls
OSC
Jars
Bottles
Cups
Plates
0.00
Early Platform Late Platform Midden Room Water Flood
Figure 100: Percentage of ceramic types sampled per analytical unit
230
D e
nsity
De
nsity
0.0
0
.2
0.4
0
.6
0.8
1
.0
0.0
0
0.0
1
0.0
2
0.0
3
Rim Diameters of OSC's
0 10 20 30 40 50 60
N = 749 Bandwidth = 2 Range of Rim Diameters of OSC's
Figure 101: KDE plot of rim diameters of ollas sin cuello
Rim T hicknesses of OSC's
0.0 0.5 1.0 1.5 2.0
N = 727 Bandwidth = 0.1 Range of Rim Thicknesses of OSC 's
Figure 102: KDE plot of rim thicknesses of ollas sin cuello
231
D e
nsity
Dia
mete
r
0.0
0
0.0
1
0.0
2
0.0
3
0.0
4
50
40
30
20
10
0
0 0,2 0,4 0,6 0,8 1 1,2 1,4 1,6 1,8 2 2,2 2,4
Thickness
.1 .2 .3 .4 .5 .6 .7 .8 .9 Quantile Density Contours
Figure 103: Bivariate KDE plot of ollas sin cuello. Arrows point to the modalities identified
Rim Diameters of OSC's from M idden
0 10 20 30 40 50 60
N = 548 Bandwidth= 2 Range of Rim D iameters from Midden
Figure 104: Rim diameters of ollas sin cuello from Midden
232
D e
nsity
Dia
me
ter
0.0
0
.2
0.4
0
.6
0.8
1
.0
Rim T hicknesses of OSC's from M idden
0.0 0.5 1.0 1.5 2.0
N = 539 Bandwidth = 0.1 Range of Rim Thicknesses from Midden
Figure 105: Rim thicknesses of ollas sin cuello from Midden
60
50
40
30
20
10
0
0 .2 .4 .6 .8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
Thickness
.1 .2 .3 .4 .5 .6 .7 .8 .9 Quantile Density Contours
Figure 106: Bivariate KDE of OSC’s from Stone Rooms. Arrows indicate the modes identified
233
D e
nsity
De
nsity
0.0
0
0.0
1
0.0
2
0.0
3
0.0
4
0.0
0
.2
0.4
0
.6
0.8
1
.0
1.2
Rim Diameters of OSC's from Stone Rooms
0 10 20 30 40 50
N = 103 Bandwidth = 3 Range of Rim Diameters from Stone Rooms
Figure 107: Rim diameters of ollas sin cuello from Stone Rooms
Rim T hicknesses of OSC's from Stone Rooms
0.0 0.5 1.0 1.5 2.0
N = 93 Bandwidth = 0.1 Range of Rim Thicknesses from Stone Rooms
Figure 108: Rim thicknesses of ollas sin cuello from Stone Rooms
234
Dia
me
ter
De
nsity
0.0
0
0.0
2
0.0
4
0.0
6
0.0
8
50
45
40
35
30
25
20
15
10
5
0
0 .2 .4 .6 .8 1 1.2 1.4 1.6 1.8 2
Thickness
.1 .2 .3 .4 .5 .6 .7 .8 .9 Quantile Density Contours
Figure 109: Bivariate KDE of OSC’s from Stone Rooms. Arrows indicate the modes identified
Rim Diameters of OSC's from Early Platforms
0 10 20 30
N = 52 Bandwidth = 2.5 Range of Rim Diameters from Early Platforms
Figure 110: Rim diameters of ollas sin cuello from Early Platforms
235
D e
nsity
Dia
me
ter
0.0
0
.2
0.4
0
.6
0.8
1
.0
1.2
1
.4
Rim T hicknesses of OSC's from Early Platforms
0.5 1.0 1.5 2.0
N = 49 Bandwidth = 0.1 Range of Rim Thicknesses from Early Platforms
Figure 111: Rim thicknesses of ollas sin cuello from Early Platforms
30
25
20
15
10
.4 .6 .8 1 1.2 1.4 1.6 1.8
Thickness
.1 .2 .3 .4 .5 .6 .7 .8 .9 Quantile Density Contours
Figure 112: Bivariate KDE of OSC’s from Early Platforms. Arrows indicate the modes identified
236
D e
nsity
De
nsity
0.0
0
0.0
1
0.0
2
0.0
3
0.0
4
0.0
0
.2
0.4
0
.6
0.8
1
.0
Rim Diameters of OSC's from Water Flood
0 10 20 30 40 50
N = 30 Bandwidth = 5
Range of Rim D iameters from Water Flood
Figure 113: Rim diameters of ollas sin cuello from Water Flood
Rim T hicknessesof OSC's from Water Flood
0.5 1.0 1.5 2.0
N = 30 Bandwidth = 0.15 Range of Rim Thicknesses from Water Flood
Figure 114: Rim thicknesses of ollas sin cuello from Water Flood
237
Dia
me
ter
De
nsity
0.0
0
0.0
2
0.0
4
0.0
6
0.0
8
40
35
30
25
20
15
10
5
0
0 .2 .4 .6 .8 1 1.2 1.4 1.6 1.8 2
Thickness
.1 .2 .3 .4 .5 .6 .7 .8 .9 Quantile Density Contours
Figure 115: Bivariate KDE of OSC’s from Water Flood. Arrows indicate the modes identified
Rim Diameters of OSC's from Late Platforms
5 10 15 20 25 30
N = 16 Bandwidth = 2 Range of Rim Diameters from Late Platforms
Figure 116: Rim diameters of ollas sin cuello from Late Platforms
238
De
nsity
Dia
mete
r
0.0
0
.2
0.4
0
.6
0.8
1
.0
1.2
1
.4
Rim T hicknesses of OSC's from Late Platforms
0.0 0.5 1.0 1.5
N = 16 Bandwidth = 0.19 Range of Rim Thicknesses from Late Platforms
Figure 117: Rim thicknesses of ollas sin cuello from Late Platforms
25
20
15
10
.2 .4 .6 .8 1 1.2 1.4
Thickness
.1 .2 .3 .4 .5 .6 .7 .8 .9 Quantile Density Contours
Figure 118: Bivariate KDE of OSC’s from Late Platforms. Arrows indicate the modes identified
239
Density
De
nsity
0.0
0
0.0
1
0.0
2
0.0
3
0.0
4
0.0
5
0.0
6
0.0
0.5
1.0
1.5
2.0
Bowl Rim Diameter KDE
0 10 20 30 40 50 60
N = 1334 Bandwidth = 2 Bowl Diameter Range
Figure 119: Rim diameters of bowls
Bowl Rim T hickness KDE
0.0 0.5 1.0 1.5 2.0
N = 1284 Bandwidth = 0.1 Bowl Thickness Range
Figure 120: Rim thicknesses of bowls
240
Dia
me
ter
Density
0.0
0
0.0
1
0.0
2
0.0
3
0.0
4
0.0
5
0.0
6
0.0
7
60
50
40
30
20
10
0 .2 .4 .6 .8 1 1.2 1.4 1.6 1.8 2
Thickness
.1 .2 .3 .4 .5 .6 .7 .8 .9 Quantile Density Contours
Figure 121: Bivariate KDE of bowls. Arrow indicates the mode identified
Rim Diameters of Bowls from Midden
0 10 20 30 40 50 60
N = 1114 Bandwidth = 2
Range of Rim Diameters of Bowls from Midden
Figure 122: Rim diameters of bowls from Midden
241
Dia
me
ter
De
nsity
0.0
0
.5
1.0
1
.5
2.0
Rim T hicknesses of Bowls from Midden
0.0 0.5 1.0 1.5
N = 1068 Bandwidth = 0.05 Range of Rim Thicknesses of Bowls from Midden
Figure 123: Rim thicknesses of bowls from Midden
60
50
40
30
20
10
0 .2 .4 .6 .8 1 1.2 1.4 1.6 1.8 2
Thickness
.1 .2 .3 .4 .5 .6 .7 .8 .9 Quantile Density Contours
Figure 124: Bivariate KDE of bowls from Midden. Arrow indicates the mode identified
242
Density
De
nsity
0.0
0.5
1.0
1.5
2.0
0.0
0
0.0
1
0.0
2
0.0
3
0.0
4
0.0
5
0.0
6
Rim Diameters of Bowls from Stone Rooms KDE
0 10 20 30 40 50 60
N = 140 Bandwidth = 3 Range of Rim Diameters of Bowls from Stone Rooms
Figure 125: Rim diameters of bowls from Stone Rooms
Rim T hicknesses of Bowls from Stone Rooms KDE
0.0 0.5 1.0 1.5
N = 137 Bandwidth = 0.1
Range of Rim Thicknesses of Bowls from Stone Rooms
Figure 126: Rim thicknesses of bowls from Stone Rooms
243
Dia
me
ter
Den
sity
0.0
0
0.0
1
0.0
2
0.0
3
0.0
4
0.0
5
50
40
30
20
10
0
0 .2 .4 .6 .8 1 1.2 1.4 1.6 1.8 2
Thickness
.1 .2 .3 .4 .5 .6 .7 .8 .9 Quantile Density Contours
Figure 127: Bivariate KDE of bowls from Stone Rooms. Arrow indicates the mode identified
Rim Diameters of Bowls from Early Platforms KDE
0 10 20 30 40 50 60
N = 34 Bandwidth = 3.5 Range of Rim Diameters from Early Platforms
Figure 128: Rim diameters of bowls from Early Platforms
244
Dia
me
ter
De
nsity
0
1
2
3
4
Rim T hicknesses of Bowls from Early Platforms KDE
0.2 0.4 0.6 0.8
N = 34 Bandwidth = 0.05 Range of Rim Thicknesses from Early Platforms
Figure 129: Rim thicknesses of bowls from Early Platforms
50
40
30
20
10
.4 .5 .6 .7 .8 .9
Thickness
.1 .2 .3 .4 .5 .6 .7 .8 .9 Quantile Density Contours
Figure 130: Bivariate KDE of bowls from Early Platforms. Arrows indicate the mode identified
245
Den
sity
De
nsity
0.0
0
0.0
1
0.0
2
0.0
3
0.0
4
0.0
5
0.0
6
0.0
0.5
1
.0
1.5
2.0
2.5
3.0
Rim Diameters of Bowls from Water Flood
0 5 10 15 20 25 30 35
N = 34 Bandwidth = 2
Range of Rim Diameters from Water Flood
Figure 131: Rim diameters of bowls from Water Flood
Rim T hicknesses of Bowls from Water Flood
0.2 0.4 0.6 0.8 1.0
N = 33 Bandwidth = 0.05 Range of Rim Thicknesses from Water Flood
Figure 132: Rim thicknesses of bowls from Water Flood
246
Dia
me
ter
De
nsity
0.0
0
0.0
2
0.0
4
0.0
6
0.0
8
35
30
25
20
15
10
5
0
0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1
Thickness
.1 .2 .3 .4 .5 .6 .7 .8 .9 Quantile Density Contours
Figure 133: Bivariate KDE of bowls from Water Flood. Arrows indicate the modes identified
Rim Diameters of Bowls from Late Platforms
5 10 15 20 25 30 35
N = 12 Bandwidth = 2.5 Range of Rim Diameters from Late Platforms
Figure 134: Rim diameters of bowls from Late Platforms
247
Den
sity
Dia
mete
r
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3
.5
Rim T hicknesses of Bowls from Late Platforms
0.2 0.4 0.6 0.8 1.0
N = 12 Bandwidth = 0.06 Range of Rim Thicknesses from Late Platforms
Figure 135: Rim thicknesses of bowls from Late Platforms
30
25
20
15
10
0 .2 .4 .6 .8 1
Thickness
.1 .2 .3 .4 .5 .6 .7 .8 .9 Quantile Density Contours
Figure 136: Bivariate KDE of bowls from Late Platforms. Arrows indicate the modes identified
248
Den
sity
Density
0.0
0
0.0
2
0.0
4
0.0
6
0.0
8
0.1
0
0.0
0.5
1.0
1.5
2.0
2.5
Rim Diameters of Jars
0 10 20 30 40 50
N = 379 Bandwidth = 2 Range of Rim Diameters of Jars
Figure 137: Rim diameters of jars
Rim T hicknesses of Jars
0.2 0.4 0.6 0.8 1.0 1.2
N = 370 Bandwidth = 0.05
Range of Rim Thicknesses of Jars
Figure 138: Rim thicknesses of jars
249
Dia
me
ter
Den
sity
0.0
0
0.0
2
0.0
4
0.0
6
0.0
8
0.1
0
40
30
20
10
0
0 .2 .4 .6 .8 1 1.2 1.4
Thickness
.1 .2 .3 .4 .5 .6 .7 .8 .9 Quantile Density Contours
Figure 139: Bivariate KDE of jars. Arrow indicates the mode identified
Rim Diameters of Jars from Midden
0 10 20 30 40 50
N = 299 Bandwidth = 2 Range of Rim Diameters of Jars from Midden
Figure 140: Rim diameters of jars from Midden
250
Dia
me
ter
Density
0.0
0.5
1.0
1
.5
2.0
2
.5
Rim T hicknesses of Jars from Midden
0.2 0.4 0.6 0.8 1.0 1.2
N = 294 Bandwidth = 0.05
Range of Rim Thicknesses of Jars from Midden
Figure 141: Rim thicknesses of jars from Midden
40
30
20
10
0
0 .2 .4 .6 .8 1 1.2 1.4
Thickness
.1 .2 .3 .4 .5 .6 .7 .8 .9 Quantile Density Contours
Figure 142: Bivariate KDE of jars. Arrow indicates the mode identified
251
Density
De
nsity
0.0
0
0.0
2
0.0
4
0.0
6
0.0
8
0
1
2
3
Rim Diameters of Jars from Stone Rooms
0 5 10 15 20 25
N = 50 Bandwidth = 1.5
Range of Rim Diameters of Jars from Stone Rooms
Figure 143: Rim diameters of jars from Stone Rooms
Rim T hicknesses of Jars from Stone Rooms
0.2 0.4 0.6 0.8 1.0
N = 47 Bandwidth = 0.05 Range of Rim Thicknesses of Jars from Stone Rooms
Figure 144: Rim thicknesses of jars from Stone Rooms
252
Dia
me
ter
Den
sity
0.0
0
0.0
2
0.0
4
0.0
6
0.0
8
0.1
0
0.1
2
30
25
20
15
10
5
0
0 .2 .4 .6 .8 1 1.2
Thickness
.1 .2 .3 .4 .5 .6 .7 .8 .9 Quantile Density Contours
Figure 145: Bivariate KDE of jars from Stone Room. Arrow indicates the mode identified
Rim Diameters of Jars from Early Platforms
0 5 10 15 20
N = 15 Bandwidth = 2 Range of Rim Diameters of Jars from Early Platforms
Figure 146: Rim diameters of jars from Early Platforms
253
Dia
me
ter
Density
0.0
0.5
1.0
1.5
Rim T hicknesses of Jars from Early Platforms
0.0 0.2 0.4 0.6 0.8 1.0
N = 14 Bandwidth = 0.1
Range of Rim Thicknesses of Jars from Early Platforms
Figure 147: Rim thicknesses of jars from Early Platforms
20
15
10
5
0
0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1
Thickness
.1 .2 .3 .4 .5 .6 .7 .8 .9 Quantile Density Contours
Figure 148: Bivariate KDE of jars from Early Platforms. Arrows indicate the modes identified
254
De
nsity
Density
0.0
0
0.0
2
0.0
4
0.0
6
0.0
8
0.1
0
0.1
2
0.0
0.5
1.0
1.5
2.0
2.5
Rim Diameters of Jars from Water Flood
2 4 6 8 10 12 14 16
N = 10 Bandwidth = 1.3
Range of Rim Diameters of Jars from Water Flood
Figure 149: Rim diameters of jars from Water Flood
Rim T hicknesses of Jars from Water Flood
0.0 0.2 0.4 0.6 0.8 1.0
N = 10 Bandwidth = 0.1 Range of Rim Thicknesses of Jars from Water Flood
Figure 150: Rim thicknesses of jars from Water Flood
255
Density
Dia
me
ter
0.0
0
0.0
5
0.1
0
0.1
5
0.2
0
0.2
5
0.3
0
0.3
5
15 14
13
12
11
10
9
8
7
6
5
4
3
2
1
0 0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1 Thickness
.1 .2 .3 .4 .5 .6 .7 .8 .9 Quantile Density Contours
Figure 151: Bivariate KDE of jars from Water Flood. Arrows indicate the modes identified
Rim Diameters from Bottles
0 2 4 6 8 10
N = 113 Bandwidth = 0.7 Range of Rim Diameters from Bottles
Figure 152: Rim diameters from bottles
256
Density
Dia
mete
r
0.0
0.5
1.0
1.5
2.0
2.5
3.0
Rim T hicknesses from Bottles
0.2 0.4 0.6 0.8 1.0
N = 110 Bandwidth = 0.04 Range of Rim Thicknesses from Bottles
Figure 153: Rim thicknesses from bottles
10
9
8
7
6
5
4
3
2
1
0
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1
Thickness
.1 .2 .3 .4 .5 .6 .7 .8 .9 Quantile Density Contours
Figure 154: Bivariate KDE of bottles. Arrows indicated the modes identified
257
Den
sity
De
nsity
0.0
0
0.0
5
0.1
0
0.1
5
0.2
0
0.2
5
0.3
0
0.3
5
0.0
0.5
1.0
1
.5
2.0
2.5
3.0
Rim Diameters of Bottles from Midden
0 2 4 6 8 10
N = 95 Bandwidth = 0.7 Range of Rim Diameters of Bottles from Midden
Figure 155: Rim diameters of bottles from Midden
Rim T hicknesses of Bottles from Midden
0.2 0.4 0.6 0.8 1.0
N = 93 Bandwidth = 0.04 Range of Rim Thicknesses of Bottles from Midden
Figure 156: Rim thicknesses of bottles from Midden
258
Dia
me
ter
D e
nsity
0.0
0
0.0
5
0.1
0
0.1
5
10
9
8
7
6
5
4
3
2
1
0
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1
Thickness
.1 .2 .3 .4 .5 .6 .7 .8 .9 Quantile Density Contours
Figure 157: Bivariate KDE of bottles from Midden. Arrows indicated the modes identified
Rim Diameters from Cups
0 2 4 6 8 10 12
N = 35 Bandwidth = 0.9 Range of Rim D iameters from C ups
Figure 158: Rim diameters of cups
259
Dia
me
ter
De
nsity
0.0
0
.5
1.0
1
.5
2.0
2
.5
Rim T hicknesses from Cups
0.0 0.2 0.4 0.6 0.8
N = 33 Bandwidth = 0.07 Range of Rim Thicknesses from C ups
Figure 159: Rim thicknesses from cups
10
9
8
7
6
5
4
3
2
0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8
Thickness
.1 .2 .3 .4 .5 .6 .7 .8 .9 Quantile Density Contours
Figure 160: Bivariate KDE of cups. Arrows indicate modes identified
260
Density
Density
0.0
0.5
1
.0
1.5
2
.0
2.5
0.0
0
0.0
5
0.1
0
0.1
5
Rim Diameters of Cups from Midden
0 2 4 6 8 10 12
N = 33 Bandwidth = 0.9 Range of Rim Diameters of Cups from Midden
Figure 161: Rim diameters of cups from Midden
Rim T hicknesses Cups from Midden
0.0 0.2 0.4 0.6 0.8
N = 33 Bandwidth = 0.07
Range of Rim Thicknesses of Cups from Midden
Figure 162: Rim thicknesses of cups from Midden
261
Dia
me
ter
D e
nsity
0.0
0
0.0
1
0.0
2
0.0
3
0.0
4
0.0
5
0.0
6
10
9
8
7
6
5
4
3
0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8
Thickness
.1 .2 .3 .4 .5 .6 .7 .8 .9 Quantile Density Contours
Figure 163: Bivariate KDE of cups from Midden. Arrows indicate modes identified
Rim Diameters from Plates
0 10 20 30 40
N = 54 Bandwidth = 2 Range of Rim Diameters from Plates
Figure 164: Rim diameters of plates
262
Dia
me
ter
D e
nsity
0.0
0
.5
1.0
1
.5
2.0
Rim T hicknesses from Plates
0.2 0.4 0.6 0.8 1.0 1.2
N = 53 Bandwidth = 0.07
Range of Rim Thicknesses from Plates
Figure 165: Rim thicknesses of plates
40
35
30
25
20
15
10
5
0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1
Thickness
.1 .2 .3 .4 .5 .6 .7 .8 .9 Quantile Density Contours
Figure 166: Bivariate KDE of plates. Arrows indicate the modes identified
263
De
nsity
De
nsity
0.0
0
.5
1.0
1
.5
2.0
0
.00
0
.01
0
.02
0
.03
0
.04
0
.05
0
.06
0
.07
Rim Diameters of Plates from M idden
0 10 20 30 40
N = 44 Bandwidth = 2 Range of Rim D iameters of Plates from Midden
Figure 167: Rim Diameters of plates from Midden
Rim T hicknesses of Plates from M idden
0.2 0.4 0.6 0.8 1.0 1.2
N = 43 Bandwidth = 0.07 Range of Rim Thicknesses of Plates from Midden
Figure 168: Rim Thickness of plates from Midden
264
Bo
ne
Weig
ht
Dia
me
ter
Ea
rly P
latf
orm
s
La
te P
latform
s
Mid
den
Room
s
Wate
r F
lood
40
35
30
25
20
15
10
5
0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1
Thickness
Figure 169: Bivariate KDE of plates. Arrow indicate the mode identified
100000
40000
10000
4000
1000
400
100
40
10
4
2
Analytical Unit
Figure 170: Density of faunal remains per analytical unit (log)
265
Figure 171: Fragment of a small spoon retrieved in WQ7,SIV, U3, L11
Figure 171: Fragment of a small spoon retrieved in WQ7,S III, U4A, L16
266
Figure 172: Polished bone tube retrieved in WQ7,SIII, U4A, L19
Figure 173: Polished bone tube retrieved in WQ7, SIII, U4A, L18
267
Figure 174: Polished bone tube retrieved in WQ7, SIII, U4A, L11
Figure 174: Polished bone tube retrieved in WQ7, SIII, U4A, L12
268
Figure 175: Bottle fragment found associated to hearth in WQ-6
Figure 176: Ceramics associated to hearth in WQ-6
269
Figure 177: Ceramic associated to hearth in WQ-6
Figure 178: Ceramics from WQ1, layer 5
270
Figure 179: Ceramic sherd from WQ1, Floor 3.
Figure 180: Ceramic sherds from WQ4, layer 6
271
Figure 181: Ceramic sherds from WQ5, layer 6
Figure 182: Ceramics sherds from WQ7, SIV, U4, layer 20
272
Figure 183: Ceramics retrieved from WQ-7, SIII, U4, layer 8
Figure 184: Ceramics retrieved from WQ-7, SIV, U4, layer 14
273
Figure 185: Ceramics retrieved from WQ-7, SIV, U4, layer 10
Figure 186: Ceramics retrieved from WQ8, layer 3
274
Figure 187: Ceramics retrieved from WQ-7,SIII, U1, layer 9
Figures 188: Urabarriu dates retrieved by Burger
275
Figure 189: Wacheqsa Sector radiocarbon dates
276
BIBLIOGRAPHY
Alva, W.
1988a Excavaciones en el Santuario del Tiempo Formativo Udima-Poro Poro en la
Sierra Norte del Perú. In: Beitrage Zur Allgemeinen Und Vergleichden, edited by
Kava, pp. 301-349. vol. 8. Verlag Philipp Von Zabern-Mainz Am Rhein, Mainz.
1988b Investigaciones en el Complejo Formativo con Arquitectura Monumental de
Purulén, Costa Norte del Perú. In: Beitrage Zur Allgemeinen Und Vergleichden, edited
by Kava, pp. 283-298. vol. 8. Verlag Philipp Von Zabern-Mainz Am Rhein, Mainz.
Alvey, B.
1993 Interpreting Archaeology with Hindsight: The Use of Three Dimensions in
Graphic Recording and Site Analysis. In: Practices of Archaeological Stratigraphy,
edited by E. Harris, M. Brown III and G. Brown, pp. 218-228. Academic Press,
Cambridge.
Amat, H.
2004 Huarás y Recuay en la secuencia cultural del Callejón de Conchucos. In:
Arqueología de la Sierra de Ancash, edited by B. Ibarra, pp. 97-120. Instituto Cultural
Runa, Lima.
Autodesk
2006 Autodesk Land Desktop: Getting Started. Autodesk Inc.
Barreto, D.
1984 Las Investigaciones en el Templete de Limoncarro. In: Beitrage Zur
Allgemeinen Und Vergleichden, edited by V. P. V. Z.-M. A. Rhein, pp. 541-547.
KAVA, Mainz.
Baxter, M.
2003 Statistics in Archaeology. First ed. Arnold Publishers, New York.
277
Baxter, M. J., C. C. Beardah and R. V. S. Wright
1997 Some Archaeological Applications of Kernel Density Estimates. Journal of
Archaeological Science 24(4):347-354.
Beex, W.
1994 From Excavation Drawings to Archaeological Playground: CAD Applications
for Excavations. Paper presented at the Computer Applications and Quantitative
Methods in Archaeology.
Bennett, W.
1943 The Position of Chavin in Andean Sequences. Proceedings of the American
Philosophical Society 86(2) Symposium on Recent Advances in American
Archaeology): 323-327..
1944 The North Highlands of Peru, Excavations in the Callejon de Huaylas and
Chavin de Huantar. Anthropological Papers of the American Museum of Natural
History 39. The American Museum of Natural History, New York.
Binford, L. R.
1964 A Consideration of Archaeological Research Design. American Antiquity
29(4):425-441.
1968 Archaeological Perspectives. In New Perspectives in Archaeology, edited by
S. Binford and L. R. Binford, pp. 5-32. Aldine Publishing, Chicago.
Bird, R. M.
1987 A Postulated Tsunami and Its Effects on Cultural Development in the
Peruvian Early Horizon. American Antiquity 52(2):285-303.
278
Bischof, H.
1998 El Periodo Inicial, el Horizonte Temprano, el Estilo Chavín y la Realidad del
Proceso Formativo en los Andes Centrales. Paper presented at the I Encuentro
Internacional de Peruanistas. Estado de los Estudios Histórico-Sociales sobre el Perú a
fines del siglo XX.
Blitz, J. H.
1993 Big Pots for Big Shots: Feasting and Storage in a Mississippian Community.
American Antiquity 58(1):80.
Bonavia, D.
1991 De los Orígenes al Siglo XV. First ed. Perú, Hombre e Historia I. Edubanco,
Lima.
Bonnier, E.
1983 Piruru: Nuevas Evidencias de la Ocupación Temprana en Tantamayo, Perú.
Gaceta Arqueológica Andina 8:8-10.
1997 Preceramic Architecture in the Andes: The Mito Tradition. In Archaeologica
Peruana 2, edited by E. Bonnier and H. Bischof, pp. 120-144, Mannheim.
Boone, J. L., III
1987 Defining and Measuring Midden Catchment. American Antiquity 52(2):336.
Brown, J.
1975 Deep-site Excavations Strategy as a Sampling Problem. In: Sampling in
Archaeology, edited by J. Mueller, pp. 155-169. The University of Arizona Press,
Tucson.
Burger, R.
1981 The Radiocarbon Evidence for the Temporal Priority of Chavin de Huantar.
American Antiquity 46(3):592.
279
1984 The Prehistoric Occupation at Chavín de Huántar, Peru. University of
California Publications, Anthropology 14. University of Berkeley, Berkeley.
1985 Concluding Remarks: Early Peruvian Civilization and Its Relation to the
Chavín Horizon. In: Early Ceremonial Architecture in the Andes, edited by C.
Donnan, pp. 269-289. Dumbarton Oaks, Washington.
1988 Unity and Heterogeneity within the Chavin Horizon. In: Peruvian Prehistory:
An Overview of Pre-Inca and Inca Society, edited by R. Keatinge, pp. 99-144.
Cambridge University Press, Cambridge.
1992 Chavin and the origins of Andean civilization. Thames and Hudson, London.
1993 The Chavin Horizon: Stylistic Chimera or Socioeconomic Metamorphosis?
In: Latin American Horizons, edited by D. S. Rice, pp. 41:82. Dumbarton Oaks,
Washington.
1998 Excavaciones en Chavín de Huántar. Pontificia Universidad Católica del
Perú, Lima.
Burger, R. and R. B. Gordon
1998 Early Central Andean Metalworking from Mina Perdida, Peru. Science
282(5391):1108.
Burger, R. and R. Matos
2002 Atalla: A Center on the Periphery of the Chavin Horizon. Latin American
Antiquity 13(2):153.
Burger, R. and L. Salazar-Burger
1991 The Second Season of Investigations at the Initial Period Center of Cardal,
Peru. Journal of Field Archaeology 18(3):275.
280
Burger, R. and L. Salazar
1980 Ritual and Religion at Huaricoto. Archaeology 33(6):26-32.
1985 The Early Ceremonial Center of Huaricoto. In: Early Ceremonial
Architecture at the Andes, edited by C. Donnan, pp. 111-138. Dumbarton Oaks,
Washington.
Campo, P.
2004 Computer... What For? Virtuality vs. Reality in Archaeology. In [Enter the
past]: The e-way into the Four Dimensions of Cultural Heritage: CAA 03 – Computer
Applications and Quantitative Methods in Archaeology., edited by W.
Stadtarchäologie, pp. 192-196. Archaeopress, Vienna.
Cattani, M., A. Fiorini and B. Rondelli
2004 Computer Applications for a Reconstruction of Archaeological Stratigraphy
as a Predictive Model in Urban and Territorial Contexts. Paper presented at the [Enter
the past]: The e-way into the Four Dimensions of Cultural Heritage: CAA 03 –
Computer Applications and Quantitative Methods in Archaeology.
Cieza de León, P.
1553 [1984] Crónica del Perú. Primera Parte. Pontificia Universidad Católica del
Perú, Academia Nacional de Historia, Lima.
Clark, J. and M. Blake
1996 The Power of Prestige: Competitive Generosity and the Emergence of Rank
Societies in Mesoamerica. In: Contemporary Theory in Archaeology: A Reader,
edited by R. Preucel and I. Hodder, pp. 258-281. Blackwell Publishing.
Clark, P.
2000 Post-Excavation Analysis: Moving from the context to the Phase. In:
Interpreting Stratigraphy, Site Evaluation, Recording Procedures and Stratigraphic
Analysis, edited by S. Roskams, pp. 157-160, Oxford.
281
Colley, S. M.
1988 Three-Dimensional Computer Graphics for Archaeological Data Exploration:
An Example from Saxon Southampton. Journal of Archaeological Science 15(1):99.
Conklin, W.
1985 The Architecture of Huaca de los Reyes. In: Early Ceremonial Architecture in
the Andes, edited by C. Donnan, pp. 139-164. Dumbarton Oaks, Washington.
Contreras, D.
2007 Geomorphologic and Sociopolitical Change at Chavín de Huantar. Ph.D
Dissertation, Stanford University.
Cook, A.
2004 Wari Art and Society. In: Andean Archaeology, edited by H. Silverman, pp.
146-166. Blackwell Publishing, Oxford.
Cowgill, G.
1970 Some sampling and reliability problems in archaeology. In Archeologie et
calculateurs: problemes semiologiques et mathematiques., pp. 161-175. Colloques
Internationaux du Centre National de la Recherche Scientifique. Centre National de la
Recherche Scientifique, Paris.
Cruz-Uribe, K.
1988 The use and meaning of species diversity and richness in archaeological
faunas. Journal of Archaeological Science 15(2):179-96.
DeBoer, W.
2001 The Big Drink: Feast and Forum in the Upper Amazon. In: Feasts,
Archaeological and Ethnographic Perspectives on Food, Politics and Power, edited
by M. Dietler and B. Hayden, pp. 215-239. Smithsonian Institution Press, Washington
DC.
282
DeMarrais, E., L. J. Castillo and T. Earle
1996 Ideology, Materialization, and Power Strategies. Current Anthropology
37(1):15-31.
Dietler, M. and B. Hayden
2001 Digesting the Feast-Good to Eat, Good to Drink, Good to Think: An
Introduction. In: Feasts, Archaeological and Ethnographic Perspectives on Food,
Politics and Power, edited by M. Dietler and B. Hayden, pp. 1-22. Smithsonian
Institution Press, Washington.
Dietler, M. and I. Herbich
2001 Feasts and Labor Mobilization: Dissecting a Fundamental Economic Practice.
In: Feasts, Archaeological and Ethnographic Perspectives on Food, Politics and
Power, edited by M. Dietler and B. Hayden, pp. 240-266. Smithsonian Institution
Press, Washington DC.
Doneus, M. and W. Neubauer
2004 Digital Recording of Stratigraphic Excavations. In: [Enter the past]: The e-
way into the Four Dimensions of Cultural Heritage: CAA 03 – Computer Applications
and Quantitative Methods in Archaeology, edited by W. Stadtarchäologie, pp. 113-
116. Archaeopress, Viena.
Drennan, R.
1996 Statistics for Archaeologists, A Common Sense Approach. First ed. Plenum
Press, New York.
Durkheim, É.
1947 The Elementary Forms of the Religious Life: A Study in Religious Sociology.
Free Press, New York.
Eiteljorg II, H.
2007 Archaeological Computing. Center for the Study of Architecture.
283
Elera, C.
1998 The Poemape Site and the Cupisnique Culture: A Case Study on the Origin
and Development of Complex Society in the Central Andes, Peru. Ph D, University of
Calgary.
Espejo, J.
1941 Exploraciones Arqueológicas en las cabeceras del Pukcha. Bachelor in
Humanities, Universidad Nacional Mayor de San Marcos.
Fung, R.
1975 Excavaciones en Pacopampa, Cajamarca. Revista del Museo Nacional
41:129-210.
2006 Personal Communication, Lima.
Guipert, G., M. A. d. Lumley, H. d. Lumley and B. Mafart
2004 Three-Dimensional Imagery: A New Look of the Tauntavel Man. In: [Enter
the past]: The e-way into the Four Dimensions of Cultural Heritage: CAA 03 –
Computer Applications and Quantitative Methods in Archaeology., edited by W.
Stadtarchäologie, pp. 100-102. Archaeopress, Vienna.
Haas, J., W. Creamer and A. Ruiz
2004 Dating the Late Archaic occupation of the Norte Chico region in Peru. Nature
432(7020):1020-1023.
Harris, E.
1992 Principles of Archaeological Stratigraphy. Second ed. Academic Press,
London.
Hass, J. and W. Creamer
2004 Cultural transformations in the Central Andean Late Archaic. In: Andean
Archaeology, edited by H. Silverman, pp. 35-50. Blackwell Malden.
284
2006 Crucible of Andean Civilization: The Peruvian Coast from 3000 to 1800 BC.
Current Anthropology 47(5):745-776.
Hayden, B.
1995 Feasting in Prehistoric and Traditional Societies. In: Food and the Status
Quest, edited by P. Wiessner and W. Schiefenhovel, pp. 127-148. First ed. Berghahn
Books, Oxford.
2001 Fabulous Feasts: A Prolegomenon to the Importance of Feasting. In: Feasts,
Archaeological and Ethnographic Perspectives on Food, Politics and Power, edited
by M. Dietler and B. Hayden, pp. 23. Smithsonian Institution Press, Washington DC.
Herzog, I.
2004 Group and Conquer – A Method for Displaying Large Stratigraphic Data Sets.
In: Enter the Past. The E-way into the Four Dimensions of Cultural Heritage, edited
by W. Stadtarchäologie, pp. 423-426. Archaeopress, Vienna.
Inokuchi, K.
1998 La Cerámica de Kuntur Wasi y el problema Chavín. Boletín de Arqueología
PUCP 2:161-180.
Izumi, S. and T. Sono
1963 Andes 2, Excavations at Kotosh Peru 1960. University of Tokyo, Tokyo.
Jennings, J., K. L. Antrobus, S. J. Atencio, E. Glavich, R. Johnson, G. Loffler and C. Luu
2005 "Drinking Beer in a Blissful Mood": Alcohol Production, Operational Chains,
and Feasting in the Ancient World. Current Anthropology 46(2):275-303.
Kaulicke, P.
1975 Pandanche. Un Caso del Formativo de los Andes de Cajamarca. Seminario
de Historia Rural Andina. Universidad nacional Mayor de San Marcos, Lima.
1994 Los Orígenes de la civilización andina I. Editorial Brasa, Lima.
285
Kembel, S.
2001 Architectural Sequence and Chronology at Chavin de Huantar, Peru. Ph D,
Stanford University.
Kembel, S. and J. Rick
2004 Building Authority at Chavín de Huántar: Models of Social Development in
the Initial Period and Early Horizon. In: Andean Archaeology, edited by H. Silverman,
pp. 51-76. Blackwell Publishing, Oxford.
Kintigh, K.
1989 Sample Size, Significance, and Measures of Diversity. In: Quantifying
Diversity in Archaeology, edited by R. Leonard and G. Jones, pp. 25-36. Cambridge
University Press, Cambridge.
Lanning, E.
1953 Chronological and Cultural Relationships of Early Pottery Styles in Ancient
Peru. Ph D, Dissertation University of California Los Angeles.
Larco, R.
1945 Los Cupisniques. Sociedad Geográfica Americana, Buenos Aires.
1948 Cronología Arqueológica del Norte del Perú. Sociedad Geográfica
Americana, Buenos Aires.
Lathrap, D.
1960a Alternative Models of Populations Movements in the Tropical Lowlands of
South America. Paper presented at the XXXIX Congreso Internacional de
Americanistas, Lima.
1960b The Upper Amazon. Thames and Hudson, London.
286
Lau, G. F.
2002 The Recuay Culture of Peru's North-Central Highlands: A Reappraisal of
Chronology and Its Implications. Journal of Field Archaeology 29(1/2):177-202.
Leon, E.
2006 Radiocarbono y calibración: Potencialidades para la Arqueología.
Arqueología y Sociedad (17):67-89.
Lock, G.
2003 Using Computers in Archaeology, Towards Virtual Pasts. First ed. Routledge,
London.
Longacre, W.
1999 Standardization and Specialization: What's the Link? In: Pottery and People,
A Dynamic Interaction, edited by J. Skibo and G. Feinman, pp. 44-58. Foundations of
Archaeological Inquiry. University of Utah Press, Utah.
Losier, L. M., J. Pouliot and M. Fortin
2007 3D geometrical modeling of excavation units at the archaeological site of Tell
‘Acharneh (Syria). Journal of Archaeological Science 34(2):272.
Lumbreras, L.
1989 Chavín de Huántar en el Nacimiento de la Civilización Andina. Instituto de
Estudios Andinos, Lima.
1993 Chavín de Huántar, Excavaciones en la Galería de las Ofrendas. KAVA,
Mainz.
2005 Arqueología y Sociedad. Instituto de Estudios Peruanos, Lima.
Lumbreras, L. and H. Amat
1965 Informe Preliminar sobre las Galerías interiores de Chavín. Revista del Museo
Nacional 34:143-197.
287
Malinverni, E., G. Gagliardini and G. Fangi
2002 Virtualization of an Archaeological Site. Paper presented at the Close-range
Imaging, Long Range Vision.
Manly, B.
1991 Randomization and Monte Carlo Methods in Biology. Chapman and Hall,
London.
Marx, K.
1973 On Society and Social Change. The Heritage of Sociology. The University of
Chicago Press, Chicago.
Meister, C. and A. Bastian
2004 Quicktime Virtual Reality (QTVR) and the Documentation of Rock Art
Localities In: [Enter the past]: The e-way into the Four Dimensions of Cultural
Heritage: CAA 03 – Computer Applications and Quantitative Methods in
Archaeology., edited by W. Stadtarchäologie, pp. 548-551. Archaeopress, Vienna.
Meister, M.
2004 On Using State of the Art Computer Game Engines to Visualize
Archaeological Structures in Interactive Teaching and Research. In: [Enter the past]:
The e-way into the Four Dimensions of Cultural Heritage: CAA 03 – Computer
Applications and Quantitative Methods in Archaeology., edited by W.
Stadtarchäologie, pp. 505-509. Archaeopress, Vienna.
Mejía, T.
1945 El Estado Actual del Conocimiento sobre las Ruinas de Chavín. Apuntes
hechos por M.T Mejía Xesspe, por indicación del Dr. Julio C. Tello, con motivo de la
catástrofe ocurrida en 1945, pp. 12. vol. Bulto 53, Cuadernillo 11. Archivo Tello,
Museo de Arqueología y Antropología de la Universidad Nacional Mayor de San
Marcos, Lima.
288
Mercer, R.
1985 A Neolithic fortress and funeral center. Scientific American 252(3):94-101.
Mesia, C.
2000 Anchucaya: Una Aproximación Teórica a un Complejo con Planta en U.
Arqueológicas 24:45-52.
2006 Julio C. Tello: Teoría y Práctica en la Arqueología Andina. Arqueología y
Sociedad 17:141-158.
Middendorf, E.
1893 [1974] Perú. Observaciones y Estudios del País y sus habitantes durante una
permanencia de 25 años. 3 vols. Universidad Nacional Mayor de San Marcos, Lima.
Miller, G. R. and R. L. Burger
1995 Our Father the Cayman, Our Dinner the Llama: Animal Utilization at Chavin
de Huantar, Peru. American Antiquity 60(3):421.
Mills, B.
1999 Ceramics and Social Contexts of Food Consumption. In Pottery and People,
A Dynamic Interaction, edited by J. Skibo and G. Feinman, pp. 99-114. Foundations
of Archaeological Inquiry. University of Utah Press, Utah.
Mogrovejo, T. A. d.
1593 [1920] Diario de la Segunda Visita que hizo de su Arquidiócesis el Ilst. señor
don Toribio Alfonso de Mogrovejo, Arzobispo de Los Reyes. Libro de Visitas, 1593.
Revista del Archivo Nacional del Perú 1:401-419.
Morales, D.
1993 Historia Arqueológica del Perú. First ed. Compendio Histórico del Perú I.
Editorial Milla Batres, Lima.
289
Neils, F., E. Dees, M. Mosely, S. Pozorski, T. Pozorski and R. Feldman
1979a El Niño: The Catastrophic Flooding of Coastal Peru. Bulletin of the Field
Museum of Natural History 50(7):4-14.
1979b El Niño: The Catastrophic Flooding of Coastal Peru. Bulletin of the Field
Museum of Natural History 50(8):4-10.
Onuki, Y.
1995 Kuntur Wasi y Cerro Blanco, Dos sitios del Formativo en el Norte del Perú.
Hokusen-Sha, Tokyo.
2001 Una Perspectiva del Período Formativo de la Sierra Norte del Perú. In:
Historia de la Cultura Peruana, pp. 103-126. vol. I. Congreso de la República del
Perú, Lima.
Orton, C.
2000 Sampling in Archaeology. First ed. Cambridge University Press, Cambridge.
Ottaway, S., L. Sawyer and A. Miller
1986 Archaeology Gets Graphic. Nature 323(16):651-652.
Patterson, T.
1985 The Huaca La Florida, Rimac Valley, Perú. In: Early Ceremonial
Architecture in the Andes, edited by C. Donnan, pp. 59-69. Dumbarton Oaks,
Washington.
Pearson, N. and T. Williams
1993 Single-Context Planning: Its Role in On Site Recording Procedures and in
Post-excavation analysis at York. In: Practices of Archaeological Stratigraphy, edited
by E. Harris, M. Brown III and G. Brown, pp. 89-103. Academic Press, London.
Polo, J. T.
1900 La Piedra de Chavín. Boletín de la Sociedad Geográfica de Lima 9:199-232.
290
Poma de Ayala, G.
1992 [1613] El Primer Nueva Corónica y Buen Gobierno. Siglo Veintiuno,
Mexico DF.
Potter, J. M.
2000 Pots, Parties, and Politics: Communal Feasting in the American Southwest.
American Antiquity 65(3):471-492.
Potzolu, C., A. Mc Carthy and L. Ristvet
2004 Volumes of History: Volume Calculations from 3D Sections at the Tell Leilan
City Gates Excavation. In: [Enter the past]: The e-way into the Four Dimensions of
Cultural Heritage: CAA 03 – Computer Applications and Quantitative Methods in
Archaeology, edited by W. Stadtarchäologie, pp. 434-438. vol. 1227. Archaeopress,
Viena.
Pozorski, S. and T. Pozorski
1987 Early Settlement and Subsistence in the Casma Valley, Peru. University of
Iowa Press, Iowa.
1998 La Dinámica del Valle de Casma durante el Período Inicial. Boletín de
Arqueología PUCP 2:83-100.
Pozorski, T.
1975 El Complejo Caballo Muerto y los Frisos de Barro de la Huaca de los Reyes.
Revista del Museo Nacional 41:211-252.
Ravines, R., H. Engelstad, V. Palomino and D. Sandweiss
1982 Materiales Arqueológicos de Garagay. Revista del Museo Nacional 46:135-
233.
Ravines, R. and W. Isbell
1975 Garagay: Sitio Temprano en el Valle de Lima. Revista del Museo Nacional
41:253-272.
291
Reid, J. J., M. B. Schiffer and W. L. Rathje
1975 Behavioral Archaeology: Four Strategies. American Anthropologist
77(4):864-869.
Reilly, P.
1989 Data Visualization in Archaeology. IBM Systems Journal 28(4):559-579.
Rice, P.
1987 Pottery Anaysis, a socurcebook. The University of Chicago Press, Chicago.
Rick, J.
2005 The Evolution of Authority and Power at Chavin de Huantar, Peru. In:
Foundations of Power in the Prehispanic Andes, edited by Vaughn K., D. Ogburn and
C. Conlee. pp. 71-89. vol. 14. American Anthropological Association, Los Angeles.
2006 Chavín de Huántar: Evidence for an Evolved Shamanism In: Mesas and
Cosmologies in the Central Andes, edited by D. Sharon. pp. 101-112. vol. 44. San
Diego Museum, San Diego.
In press. Context, Construction and Ritual in the Development of Authority at Chavín
de Huántar. In: Transformation in Chavín: Art and Culture, edited by W. Conklin and
J. Quilter. University of Florida Press, Gainesville.
Rick, J., S. Kembel, R. Mendoza and J. Kembel
1998 La Arquitectura del Complejo Ceremonial de Chavín de Huántar:
Documentación Tridimensional y sus Implicancias. Boletín de Arqueología PUCP
2:181-214.
Rick, J. and C. Mesia
2006 Informe Preliminar de las Excavaciones realizadas en la Plaza de Armas de
Chavín de Huántar. Informe presentado al Instituto Nacional de Cultura, Lima.
292
Rivero de Ustariz, M.
1851 Antigüedades Peruanas, Lima.
Robertson, I.
2007 Personal communication.
Rosas, H.
1970 La Secuencia Cultural del Periodo Formativo en Ancón. Bachelor Thesis in
Archaeology, Universidad Nacional Mayor de San Marcos.
Rosenswig, R. M.
2007 Beyond identifying elites: Feasting as a means to understand early Middle
Formative society on the Pacific Coast of Mexico. Journal of Anthropological
Archaeology 26(1):1.
Rowe, J.
1962a Chavin Art, an Inquiry into its Form and Meaning. The Museum of Primitive
Art, New York.
1962b Stages and Periods in Archaeological Interpretation. Southwestern Journal of
Anthropology 18(1):40-54.
Salinas y Córdoba, F. B. d.
1957 [1630] Memorial de las Historias del Nuevo Mundo, Pirú. Universidad
Nacional Mayor de San Marcos, Lima.
Sall, J., L. Creighton and A. Lehman
2005 JMP Start Statistics, A Guide to Statistics and Data Analysis Using JMP and
JMP IN Software. Third ed. Tompson Brooks Cole, Belmont, CA.
Scheele, H.
1970 The Chavin Occupation of the Central Coast Peru. Ph.D. Dissertation in
Anthropology, Harvard University.
293
Schiffer, M. B.
1972 Archaeological Context and Systemic Context. American Antiquity 37(2):156-
165.
1983 Toward the Identification of Formation Processes. American Antiquity
48(4):675-706.
1988 The Structure of Archaeological Theory. American Antiquity 53(3):461-485.
Seki, Y.
1998 El Periodo Formativo en el Valle de Cajamarca. Boletín de Arqueología
PUCP 2:147-160.
Seki, Y., W. Tosso, J. Villanueva and K. Inokuchi
2006 Proyecto Arqueológico Pacopampa'05: Avances y Correlaciones Regionales.
Arqueología y Sociedad 17:149-177.
Shady, R.
1997 La Ciudad Sagrada de Caral-Supe en los Albores de la Civilización en el
Perú. Universidad Nacional Mayor de San Marcos, Lima.
2004 Caral: La Ciudad del Fuego Sagrado. Centura, Lima.
Shady, R., J. Haas and W. Creamer
2001 Dating Caral, a Preceramic Site in the Supe Valley on the Central Coast of
Peru. Science 292(5517):723-726.
Shady, R. and C. Leyva (editors)
2003 La Ciudad Sagrada de Caral-Supe: Los Origines de la Civilización Andina y
la Formación del Estado Prístino en el Antiguo Perú. First ed. Instituto Nacional de
Cultura, Lima.
294
Shanks, M. and C. Tilley
1992 Re-Constructing Archaeology, Theory and Practice. Second ed. Routledge,
London.
Shennan, S.
2006 Quantifying Archaeology. Second ed. University of Iowa Press, Iowa.
Shimada, I., C. Elera and M. Shimada
1983 Excavaciones efectuadas en el Centro de Huaca Lucía-Cholope del Horizonte
Temprano, Batan Grande, Costa Norte del Perú. Arqueológicas 19:109-208.
Silva, J. and R. García
1997 Huachipa-Jicamarca: Cronología y Desarrollo Sociopolítico en el Rímac.
Instituto Francés de Estudios Andinos 26(2):195-228.
Stanish, C.
2003 Ancient Titicaca. The Evolution of Complex Society in Sourthen Peru and
Northern Bolivia. University of California Press, Los Angeles.
Steward, J.
1948 A Functional-Developmental Classification of American High Cultures. In: A
Reappraisal of Peruvian Archaeology: Memoirs of the Society for American
Archaeology, edited by W. Bennett, pp. 103-105. Society for American Archaeology
and the Institute of Andean Research, Wisconsin.
Strong, D.
1948 Cultural Epochs and Refuse Stratigraphy in Peruvian Archaeology. In: A
Reappraisal of Peruvian Archaeology: Memoirs of the Society for American
Archaeology, edited by W. Bennett, pp. 93-103. Society for American Archaeology
and the Institute of Andean Research, Wisconsin.
295
Strong, D. and C. Evans
1952 Cultural Stratigraphy in the Virú Valley Northen Peru: The Formative and
Florescent Epochs. Columbia University Press, New York.
Tellenbach, M.
1986 Las Excavaciones En El Asentamiento Formativo De Montegrande, Valle De
Jequetepeque En El Norte Del Peru. Materialien Zur Allgemeinen Und
Vergleichenden Archaologie 39. Verlag Philipp von Zabern GmbH, Mainz am Rhein.
Tello, J. C.
1923 Wiracocha. Reimpreso de la Revista Inca ed. Universidad Nacional Mayor de
San Marcos, Lima.
1929 Antiguo Perú, Primera Época. 1 ed. Comision Organizadora del Segundo,
Lima.
1940 Expedición a Chavín (del 7 de Noviembre al 14 de Diciembre de 1940), pp.
101. vol. Bulto 53, Cuadernillo 5. Universidad Nacional Mayor de San Marcos.
Archivo Tello, Lima.
1942 Origen y Desarrollo de las Civilizaciones Prehistóricas Andinas. Liberia e
Imprenta Gil, Lima.
1943 Discovery of the Chavin Culture in Peru. American Antiquity 9(1, Countries
South of the Rio Grande):135.
1956 Arqueología del Valle de Casma, Culturas: Chavín, Santa o Huaylas Yunga y
Sub Chimú. Universidad Nacional Mayor de San Marcos, Lima.
1960 Chavín Cultura Matriz de la Civilización Andina. Universidad Nacional
Mayor de San Marcos, Lima.
296
Terada, K.
1979 Excavations at La Pampa in the north highlands of Peru, 1975. Univesity the
Tokyo, Tokyo.
Terada, K. and Y. Onuki
1982 Excavations at Huacaloma in the Cajamarca Valley, Peru 1979. First ed.
University of Tokyo Press, Tokyo.
Torres, C. and D. Repke
2006 Anadenanthera : Visionary Plant of Ancient South America. Haworth Herbal
Press, New York.
Uhle, M.
1902 Types of Culture in Peru. American Anthropologist 4(4):753.
Vázquez de Espinoza, F. A.
1616 [1948] Compendio y Descripción de las Indias Orientales, Washington.
Vega-Centeno, R.
2007 Construction, labor organization, and feasting during the Late Archaic Period
in the Central Andes. Journal of Anthropological Archaeology 26(2):150-171.
Wand, M. and C. Jones
1995 Kernel Smoothing. First ed. Chapman & Hall, London.
Warburton, D.
2003 Archaeological Stratigraphy, A Near Eastern Approach. First ed. Recherches
Et Publications Neuchatel, Bern.
Weber, M.
1978 Economy and Society I. University of California Press, Los Angeles.
297
Wiener, C.
1880 [1993] Perú y Bolivia. Instituto Francés de Estudios Andinos, Lima.
Wiley, G.
1948 A Functional Analysis of Horizon Styles in Peruvian Archaeology. In: A
Reappraisal of Peruvian Archaeology, edited by W. Bennett, pp. 8-15. Memoirs of the
Society of American Archaeology. vol. 4. Society of American Archaeology,
Menasha.
Williams, C.
1980 Arquitectura y Urbanismo en el Antiguo Perú. In Historia del Perú, edited by
J. Mejía Baca, pp. 369-385. vol. 8. Editorial Mejía Baca, Lima.
Wolf, J.
2005 Redefining Chavín's Ceramic Sequence: Insights from a Formative Urban
Settlement at Chavín de Huántar. Paper presented at the 70th Annual Meeting of the
Society of American Archaeology, Salt Lake City, Utah.
Zhukovsky, M.
2001 Handling Digital 3-D Record of Archaeological Excavation Data. Paper
presented at the Archaeological Informatics: Pushing the Envelope CAA 2001.
Zilkowski, M., M. Pazdur, A. Krzanowski, and A. Michczyfiski
1994 Andes. Radiocarbon Database for Bolivia, Ecuador and Peru. Andean
Archaeological Mission of the Institute of Archaeology, Warsaw University &
Gliwice Radiocarbon Laboratory of the Institute of Physics, Silesian Technical
University, Warsaw-Gliwice.
Zoubek, T. A.
1997 The Initial Period Occupation of Huaca El Gallo/Huaca La Gallina, Virú
Valley, Peru and its Implications for the Guañape Phase Social Complexity Ph.D
Dissertation, Yale University.
298
Zubrow, E.
2006 Digital Archaeology: A Historical Context. In: Digital Archaeology, Bringing
Method and Theory, edited by T. Evans and P. Daly, pp. 10-32. Routledge, London.