The Butana Group Ceramics and their Place in the Neolithic and Post-Neolithic of Northeast Africa 9781407310671, 9781407340395

Citation preview

BAR S2459 2013 WINCHELL THE BUTANA GROUP CERAMICS

B A R

Cambridge Monographs in African Archaeology 83

Series Editors: Laurence Smith, Brian Stewart and Stephanie Wynne-Jone

The Butana Group Ceramics and their Place in the Neolithic and Post-Neolithic of Northeast Africa Frank Winchell

BAR International Series 2459 2013

Cambridge Monographs in African Archaeology 83

Series Editors: Laurence Smith, Brian Stewart and Stephanie Wynne-Jone

The Butana Group Ceramics and their Place in the Neolithic and Post-Neolithic of Northeast Africa Frank Winchell

BAR International Series 2459 2013

ISBN 9781407310671 paperback ISBN 9781407340395 e-format DOI https://doi.org/10.30861/9781407310671 A catalogue record for this book is available from the British Library

BAR

PUBLISHING

Table of Contents List of Figures iv List of Tables vii Acknowledgements ix Preface A Brief Review of Literature Since the Publication of this Work x Chapter 1 Introduction 1 Chapter 2 Background 8 Introduction 8 Geography of Southern Atbai 8 The Land of the Southern Atbai 8 People and Towns of the Southern Atbai 9 Location of the Malawiya/Butana Transition and Butana Group Sites 9 Archaeological Work Carried Out in the Southern Atbai 9 The Prehistory of the Southern Atbai during the Holocene 10 Pre-Saroba 10 The Atbai Ceramic Tradition 12 The Saroba Phase 13 The Kassala Phase 13 The Butana Group of the Early Kassala Phase. 13 The Gash Group of the Middle Kassala Phase. 15 The Mokram Group of the Late Kassala Phase. 15 The Taka Phase 16 Post Atbai Ceramic Tradition Sites 17 Description of the Sites Used in the Ceramic Analysis 17 KG28B 17 KG23 18 KG7 18 KG29S and KG29N 18 KG96 20 KG5 21 KG1 21 Chapter 3 Methodology 22 Introduction 22 Theoretical Considerations Used in Creating the Ceramic Typology: The Type-Variety Method 22 Methodological Procedures in Establishing a Typology for the Butana Group Ceramics 24 Explanation of Various Components Within the Type Descriptions 30 Chapter 4 Definition of the Malawiya/Butana Transition and Butana Group Based on the Ceramics Types and Their Temporal Significance 32 Introduction 32 The Use of Seriation in the Establishment of Ceramic Types Over Time and its Use in Sequencing Levels or Sites Over Time 34 Initial Problems in Setting Up A Seriation for the Butana Group Ceramics and Sites 34 Setting Up A Seriation for the Butana Group Ceramics and Site Deposits 35 The Sites 36 KG23c 36 i

KG7a 41 KG29S 45 KG96a 49 KG5a 53 KGla 56 KG29N 59 KG28B 64 Definition of the Malawiya/Butana Transition and Butana Group Based on the Ceramic Types and Their Temporal Significance 66 The Malawiya/Butana Transition Ceramics 66 The Butana Group 66 Partitioning the Butana Group into an Early and Late Interval 70 Chapter 5 Comparisons of Culture Groups of Northeast Africa During the Neolithic and Post-Neolithic Era 71 Introduction 71 The Geographic Setting of Northeast Africa 71 The Nile and Its Journey Through the Sahel and the Sahara 73 Climatic Changes and Past Environments of Northeast Africa since the Holocene 74 An Overview of the Neolithic and post-Neolithic Cultures of Northeast Africa 76 The Khartoum Mesolithic (Early Khartoum) 77 The Neolithic and post-Neolithic Settlements of Lower Egypt 81 The Fayum A Sites. 82 Merimda Beni Salama. 83 El Omari of the Lower Nile. 84 The Lower Egyptian Post-Neolithic at Maadi. 84 The Egyptian Predynastic of Upper Egypt 85 The Badarian of the Early Predynastic. 86 The Amratian of the Middle Predynastic (Naqada I). 87 The Gerzean of the Late to Terminal Predynastic (Naqada II & III). 88 The A-Group of Lower Nubia 89 The Abkan and Karat Groups of the 2nd Cataract and the Dongola Reach 92 The Abkan Group of Batn el Hajar. 92 The Karat Group of Dongola Reach. 94 The Khartoum Neolithic of the Central Nile Valley 95 The Late Neolithic Development in the Central Nile Valley 98 The Late Neolithic of the Khartoum Province. 99 Kadada of the Shendi Reach. 100 Jebel Moya, Jebel Tomat, and Rabak of the South Gezira Plain. 100 Jebel Moya. 100 Jebel Tomat. 101 Rabak. 101 The late Neolithic of Shaqadud Cave 102 Chapter 6 Comparisons of the Ceramics Associated with the Culture Group of Northeast Africa During the Neolithic and Post-Neolithic Era 103 The Ceramics of the Badarian Culture. Khartoum Neolithic. Malawiya Group. and Malawiya/Butana Transition 103 The Ceramics of the Badarian Culture 103 The Ceramics of the Khartoum Neolithic 106 The Ceramics of the Malawiya Group and the Malawiya/Butana Transition 110 The Ceramics of the Malawiya Group. 110 The Ceramics of the Malawiya/Butana Transition. 110 Comparisons and General Comments Between the Badarian, Khartoum Neolithic, Malawiya, and Malawiya/Butana Transition Ceramics 111 The Ceramics of the A-Group. Abkan Group. Karat Group. and Late Neolithic of the Central Nile Valley and the Butana Group 115 The Ceramics of the A-Group 115 The Early A-Group Ceramics in the 1st Cataract Area. 115 ii

The Classic and Terminal A-Group Ceramics in the 2nd Cataract Area. 117 The Ceramics of the Abkan Group 121 The Ceramics of the Karat Group 123 The Late Neolithic Ceramics of Kadada, el Geili, and Esh Shaheinab in the Central Nile Valley 123 The Late Neolithic Ceramics of Kadada. 123 The Late Neolithic Ceramics at el Geili. 124 The Late Neolithic Ceramics at Esh Shaheinab. 124 The Ceramics of the Early Butana Group 126 Comparisons and General Comments Between the Post and Late Neolithic Ceramics of the Nile Valley and the Early Butana Group Ceramics 127 The Ceramics of the Jebel Moya Complex, Shaqadud Cave, the Gash Group, Agordat, Erkowit, and the Late Butana Group 130 The Ceramics of the Jebel Moya Complex 131 The Ceramics of Jebel Moya. 131 The Ceramics of Rabak. 133 The Ceramics of Shaqadud Cave 135 The Ceramics of the Gash Group 137 The Ceramics of the Late Butana Group 138 Comparisons and General Comments Between the Jebel Moya Complex, Shaqadud Cave, the Gash Group and the late Butana Group Ceramics 139 The Ceramics of Agordat and Erkowit 143 The Ceramics of Agordat. 143 The Ceramics of Erkowit. 144 Chapter 7 Conclusion on the Butana Group Ceramics in Respect to the Nile Valley and its Place in the Prehistory of the Eastern Sudan and Northeast Africa 145 Appendix A Field Manual for Classifying Sherds Using the Type-Variety Method 148 Introduction 148 Types and Varieties 148 Budahah Smoothed Type - variety: Budahah 148 Saroba Combed Impressed Type - variety: Tridget 150 Nulmswi Burnished Undecorated Type - variety: Nulmswi 152 Sarb’ut Ripple Marked Type - variety: Sarb’ut 154 Butana Combed Type - variety: Butana 156 Butana Combed Type - variety: Urdumallah 159 Brindetta Red-topped Type - variety: Brindetta 163 Brindetta Red-topped Type - variety: Bosufru 166 Brindetta Red-topped Type - variety: Unspecified 167 Alabaska Punctated Type - variety: Alabaska 168 Socodad Wiped Type - variety: Socodad 169 Bahri Smoothed Type - variety: Bahri 173 Dratmuir Burnished-Undecorated Type - variety: Dratmuir 177 Dratmuira Chatter Marked Type - variety: Dratmuira 182 Zirhalla Pattern Burnished Type - variety: Zirhalla 185 Sirjino Simple Dentated Type - variety: Sirjino 187 Khordhag Plain Type - variety: Khordhag 189 Varia 192 The Ceramic Groups Based on the Types and Varieties 197 Remarks on the Ceramics Groups, Types, and Varieties Associated with the Malawiya/ Butana Transition and Butana Group. 202 Some Closing Remarks on the Formation of Ceramic Complexes within the Malawiya/ Butana transition and the Butana Group and Their Relevance in Determining the Temporal Order of the Site Occupations Used in This Study. 208 Bibliography 211

iii

List of Figures Figure 1.1. The study area of Northeast Africa. Figure 1.2. The Southern Atbai and surrounding area. Figure 1.3. The culture chronologies of northeast Africa. Figure 1.4. The Malawiya/Butana transition and Butana Group sites of the Southern Atbai. Figure 2.1. The Sudan and surrounding area. Figure 2.2. The Archaeological sequence in the Southern Atbai. Figure 2.3. The site of KG23. Figure 2.4. The site of KG7. Figure 3.1. Lip decoration. Figure 3.2. Simple contour vessel shapes. Figure 3.3. Inflected contour vessel shapes. Figure 3.4. Rim profiles. Figure 3.5. Lip profiles. Figure 4.1. Seriation of the 14 levels at KG23c based on the relative frequencies of the five major ceramic types. Figure 4.2. Seriation of the 8 levels at KG7a based on the relative frequencies of the five major ceramics types. Figure 4.3. Seriation of the 4 level groups at KG23c and the 3 level groups at KG7a based on the relative frequencies of the five major ceramics types. Figure 4.4. Seriation of the 4 level groups at KG23c and the collapsed deposit at KG7a based on the relative frequencies of the five major ceramic types. Figure 4.5. Seriation of the 6 levels at KG29S based on the relative frequencies of the four major ceramic types. Figure 4.6. Seriation of the 4 level groups at KG23c and the collapsed deposit at KG29S based on the relative frequencies of the four major ceramics types. Figure 4.7. Seriation of the 6 levels at KG96a based on the relative frequencies of the five major ceramic types. Figure 4.8. Seriation of the 4 level groups at KG23c and the collapsed deposit at KG96a based on the relative frequencies of the five major ceramic types. Figure 4.9. Seriation of the five levels at KG5a based on the relative frequencies of the five major ceramic types. Figure 4.10. Seriation of the five levels at KG5a based on the relative frequencies of the five major ceramic types. Figure 4.11. Seriation of the 4 levels at KG1a based on the relative frequencies of the five major ceramic types. Figure 4.12. Seriation of the 4 level groups at KG23c and the collapsed deposit at KG1a based on the relative frequencies of the five major ceramic types. Figure 4.13. Seriation of the 5 levels at KG29N based on the relative frequencies of the five major ceramic types. Figure 4.14. Seriation of the 4 level groups at KG23c and the collapsed deposit at KG29N based on the relative frequencies of the five major ceramic types. Figure 4.15. Seriation of the 3 levels at KG28B based on the relative frequencies of the three major ceramic types.

iv

2 3 4 5 10 11 19 20 28 28 29 29 29 39 43 44 45 48 49 52 54 56 57 59 60 62 63 65

Figure 4.16. Seriation of the 4 level groups at KG23c and the remaining six Butana Group sites based on the relative frequencies of the five principal Butana Group ceramic types. 68 Figure 5.1. Geographic areas of northeast Africa. 72 Figure 5.2. General location of important Neolithic and post-Neolithic archeological groups and cultures in northeast Africa. 78 Figure 6.1. Principal component distribution of the late Neolithic, Abkan, A-Group, and early Butana Group. 131 Figure 6.2. Principal component distribution of the Shaqadud Cave complex, Jebel Moya complex, early Gash Group, and Late Butana Group ceramics. 143 Figure A.1. Rim sherds of the Budahah Smoothed type. 149 Figure A.2. Distribution of general vessel shapes affiliated with the Budahah Smoothed type. 150 Figure A.3. Body sherds of the Saroba Combed Impressed type. 151 Figure A.4. Rim sherds of the Nulmswi Burnished (top) and Sarb’ut Ripple Marked types. 152 Figure A.5. Distribution of general vessel shapes affiliated with Nulmswi Burnished type. 153 Figure A.6. Distribution of general vessel shapes affiliated with the Sarb’ut Ripple Marked type. 155 Figure A.7. Rim sherds of the Butana Combed type: Variety Butana. 157 Figure A.8. Rim sherds of the Butana Combed type: variety Butana. 157 Figure A.9. Distribution of general vessel shapes affiliated with the Butana variety of the Butana Combed type. 158 Figure A.10. Rim sherds of the Butana Combed type: variety Urdumallah. 160 Figure A.11. Rim sherds of the Butana Combed type: variety Urdumallah. 160 Figure A.12. Rim sherds of the Butana Combed type: variety Urdumallah. 160 Figure A.13. Distribution of general vessel shapes affiliated with the Urdumallah variety of the Butana Combed type. 162 Figure A.14. Rim sherds of the Brindetta Red-topped type: varieties Brindetta and Bosufru. 164 Figure A.15. Rim sherds of the Brindetta Red-topped type: variety unspecified. 164 Figure A.16. Distribution of general vessel shapes affiliated with the Brindetta Red-topped type. 165 Figure A.17. Body sherds of the Alabaska Punctated type. 168 Figure A.18. Rim sherds of the Socodad Wiped type. 169 Figure A.19. Rim sherds of the Socodad Wiped type. 170 Figure A.20. Rim sherds of the Socodad Wiped type with the Atbara Mode Rim Band. 170 Figure A.21. Distribution of general vessel shapes affiliated with the Socodad Wiped type. 171 Figure A.22. Rim sherds of the Bahri Smoothed type. 173 Figure A.23. Rim sherds of the Bahri Smoothed type. 173 Figure A.24. Rim sherds of the Bahri Smoothed type. 174 Figure A.25. Distribution of general vessel shapes affiliated with the Bahri Smoothed type. 175 Figure A.26. Rim sherds of the Dratmuir Burnished type. 178 Figure A.27. Rim sherds of the Dratmuir Burnished type. 178 Figure A.28. Rim sherds of the Dratmuir Burnished type. 178 Figure A.29. Rim sherds of the Dratmuir Burnished type. 178 Figure A.30. Distribution of general vessel shapes affiliated with the Dratmuir Burnished type. 180 Figure A.31. Rim sherds of the Dratmuira Chatter Marked type. 182 Figure A.32. Distribution of general vessel shapes affiliated with the Dratmuira Chatter Marked type. 183 Figure A.33. Rim sherds of the Zirhalla Pattern Burnished type. 185 Figure A.34. Distribution of general vessel shapes affiliated with the Zirhalla Patterned Burnished type. 186 Figure A.35. Rim sherds of the Sirjino Simple Dentated type. 187 Figure A.36. Distribution of general vessel shape shapes affiliated with the Sirjino Simple Dentated type. 189 Figure A.37. Rim sherds of the Khordhag Plain type. 190 v

Figure A.38. Distribution of general vessel shapes affiliated with the Khordhag Plain type. Figure A.39. Distribution of general vessel shapes affiliated with the Nulmswi ceramic group. Figure A.40. Distribution of general vessel shapes affiliated with the Butana ceramic group. Figure A.41. Distribution of general vessel shapes affiliated with the Dratmuir ceramic group. Figure A.42. Distribution of general vessel shapes affiliated with the Malawiya/Butana transition at KG28B. Figure A.43. Distribution of general vessel shapes affiliated with the early Butana Group interval at KG23c. Figure A.44. Distribution of general vessel shapes affiliated with the Butana and Brindetta ceramic groups in the early Butana interval at KG23c. Figure A.45. Distribution of general vessel shapes affiliated with the Socodad ceramic group in the early Butana Group interval at KG23c. Figure A.46. Distribution of general vessel shapes affiliated with the Dratmuir and Bahri ceramic groups in the early Butana Group interval at KG23. Figure A.47. Distribution of general vessel shapes affiliated with the late Butana Group interval at KG23c. Figure A.48. Distribution of general vessel shape affiliated with the Butana and Brindetta ceramic groups in the late Butana Group interval at KG23c. Figure A.49. Distribution of general vessel shapes affiliated with the Socodad ceramic in the late Butana Group interval at KG23c. Figure A.50. Distribution of general vessel shapes affiliated with the Dratmuir and Bahri ceramic groups in the late Butana Group interval at KG23c.

vi

191 198 199 200 201 202 203 204 205 206 206 207 207

List of Tables Table 1.1. Butana Group Radiocarbon Dates Table 3.1. Hierarchical Relationship Between Exterior Surface Treatment and Decoration Table 3.2. Breakdown of Descriptive Categories By Analytical Components Table 3.3. Variable Number Designations and Corresponding Descriptions Table 4.1. Frequencies of the Ceramic Types Through the Levels at KG23c Table 4.2. Relative Frequencies and 2 Standard Error Margins of the Ceramic Types Through the Levels at KG23c Table 4.3. Relative Frequencies of the Five Major Butana Group Types By Level at KG23c Table 4.4. Relative Frequencies and Corresponding 2 Standard Errors of the Ceramic Types Through the Four Level Groups at KG23c Table 4.5. Frequencies of the Ceramic Types Through the Levels at KG7a Table 4.6. Relative Frequencies and 2 Standard Error Margins of the Ceramic Types Through the Levels at KG7a Table 4.7. Relative Frequencies and 2 Standard Error Margins of the Five Major Butana Group Types Among the 4 Level Group at KG23c and the Collapsed Deposit at KG7a Table 4.8. Frequencies of the Ceramic Types Through the Levels at KG29S Table 4.9. Relative Frequencies and 2 Standard Error Margins of the Ceramic Types Through the Levels at KG29S Table 4.10. Relative Frequencies and 2 Standard Error Margins of the Four Major Butana Group Ceramic Types at KG29S in Relation to the 4 Level Groups at KG23c Table 4.11. Frequencies of the Ceramic Types Through the Levels at KG96a Table 4.12. Relative Frequencies and 2 Standard Error Margins of the Ceramic Types Through the Levels at KG96a Table 4.13. Relative Frequencies and 2 Standard Error Margins of the Five Major Butana Group Type by the Two Level Groups at KG96a Table 4.14. Relative Frequencies and 2 Standard Error Margins of the Five Major Butana Group Types Among the 4 Level Groups at KG23c and the Collapsed Deposit at KG96a Table 4.15. Frequencies of the Ceramic Types Through the Levels at KG5a Table 4.16. Relative Frequencies and 2 Standard Error Margins of the Ceramic Types Through the Levels at KG5a Table 4.17. Relative Frequencies and 2 Standard Error Margins of the Five Major Butana Group Types Among the 4 Level Groups at KG23c and the Collapsed Deposit at KG5a Table 4.18. Frequencies of the Ceramic Types Through the Levels at KG1a Table 4.19. Relative Frequencies and 2 Standard Error Margins of the Ceramic Types Through the Levels at KG1a Table 4.20. Relative Frequencies and 2 Standard Error Margins of the Five Major Butana Group Types Among the 4 Level Groups at KG23c and the Collapsed Deposit at KG1a Table 4.21. Frequencies of the Ceramic Types Through the Levels at KG29N Table 4.22. Relative Frequencies and 2 Standard Error Margins of the Ceramic Types Through the Levels at KG29N Table 4.23. Relative Frequencies and 2 Standard Error Margins of the Five Major Butana Group Types Among the 4 Level Groups at KG23c and the Collapsed Deposit at KG29N Table 4.24. Frequencies of the Ceramic Types Through the Levels at KG28B

vii

5 24 25 27 37 37 40 40 46 46 46 47 47 50 50 50 53 54 55 55 57 58 61 61 61 61 63 64

Table 4.25. Relative Frequencies and 2 Standard Error Margins of the Ceramic Types Through the Levels at KG28B Table 4.26. Relative Frequencies of the Butana Combed and Sarb’ut Ripple Marked Types in Relation to the Remaining Types at KG28B and the 4 Level Groups at KG23c Table 4.27. Percentage Breakdown of the 10 Primary Butana Group Types Among the Seven Butana Group Sites Table 4.28. Relative Frequencies and 2 Standard Error Margins of the Ten Primary Butana Group Types Among the 4 Level Groups at KG23c and the Other Six Butana Group Sites Table 6.1. Comparative Table on Decorated Ceramics at a Midden Deposit at Hemamieh Table 6.2. Hierarchical Structure of Prehistoric Impressed Ceramics of the Central Nile Valley Table 6.3.Frequencies of Khartoum Neolithic Sherds Based on Exterior Surface Treatment Table 6.4. Breakdown of the Malawiya/Butana Transition Ceramics by Types, Exterior Surface Treatment, and Decoration Table 6.5. General Ceramic Characteristics Among the Badarian, Khartoum Neolithic Malawiya, and Malawiya/Butana Transition Table 6.6. Comparison Between Decorated and Undecorated Ceramics Among the Badarian, Khartoum Neolithic, Malawiya, and Malawiya/Butana Transition Table 6.7. A Compilation of Reisner’s Classification of Early A-Group Phase Ceramics by Ware Group, Ceramic Class, and Vessel Shape Table 6.8. Classic and Terminal A-Group Type Groups and Corresponding Decorations Within Each Type Group Based on Nordstrom’s Ceramic Typology Table 6.9. Breakdown of the Abkan Group Ceramics by Ware Group, Ware and Corresponding Decorations Within Each Ware Based on Nordstrom’s Typology Table 6.10. Tally of Late Neolithic Ceramics at Esh Shaheinab By Category, Decorative Technique, and Vessel Shape Table 6.11. Early Butana Group Ceramics by Type, Exterior Surface, Decoration Table 6.12. General Ceramic Characteristics of the Ceramic-Bearing Groups in the Nile Valley and the Butana Group of the Southern Atbai Table 6.13. Comparison Between Contemporary Ceramic-Bearing Groups in the Nile Valley and the Butana Group in the Southern Atbai Using Basic Ceramic Decorations Table 6.14. Khartoum Neolithic and Late Neolithic Ceramics by Decoration and Stratigraphic Positioning at Rabak Table 6.15. Typological Breakdown of Ceramics From Shaqadud Cave Table 6.16. Ceramic Categories of the Early Gash or Proto-Gash Group From Trench II at Mahal Teglinos Table 6.17. Breakdown of Early and Late Butana Group Ceramics by Type, Exterior Surface Treatment, and Decoration Table 6.18. General Ceramic Characteristics of the Jebel Moya Complex, Early Shaqadud Cave, Early Gash, and Late Butana Group Table 6.19. Decorative Techniques Among the Ceramic Associated with the Jebel Moya Complex, Shaqadud Cave, Early Gash Group, and the Late Butana Group Table A.1. Listing of the Eight Ceramic Groups and Affiliated Types and Varieties Table A.2. Breakdown of General Characteristics Associated with the Ceramic Groups of the Malawiya/Butana Transition and Butana Group Table A.3. Frequencies and Percentages (including 2 standard errors) of the Early, Late and Transitional Butana Group Ceramic Complexes in the Early and Late Butana Group Intervals Based on the Number of Accountable Vessels at KG23c Table A.4. Frequencies and Percentages (including 2 standard errors) of the Early, Late and Transitional Butana Group Ceramic Complexes in the Early and Late Butana Group Intervals Based on the Number of Accountable Vessels at KG29S, KG7a, KG1a, KG96s, KG5a, and KG29N

viii

64 66 67 69 104 107 108 110 112 112 116 119 122 125 126 128 128 133 134 137 138 139 139 197 201

208

209

Acknowledgements I would like to start off and thank everyone who was on my dissertation committee who helped me to complete this work. First and foremost, I would like to thank Anthony Marks who supported me on my ceramic analysis and oversaw the writing of this dissertation. I would also like to thank Robin Robertson who guided me through the ceramic analysis and commented on the type descriptions, and Harold Hietala who helped me on the statistics. Fred Wendorf also was very helpful on his input on the Neolithic of the Sahara and was my interim chairperson for the pre-oral exam proposal write up. Abbas Mohammed-Ali, Rodolfo Fattovich, and Isabella Caneva deserve special thanks for their involvement with my ceramic analysis as well. I would also like to thank Edward Fry for editing the many editions which I gave him over the years. I would also like to thank the various faculty members at SMU anthropology for their insight and support. This includes Thomas Johnson, David Freidel, Carolyn Sargent, Ronald Wetherington, David Meltzer, David Wilson, Robert Kemper, Angela Close, Ladislov Novak, Garth Sampson, Ann Woosely, Patricia Crown, and Benjamin Wallace. I would also like to thank Mr. Weber for his generous Weber Grant and my colleagues Karim Sadr and Steven M’Butu for their valuable input. I would also like to thank my mother, Mary Ruth Winchell; and especially my first wife, Cristi Winchell, for the years of patience she gave to see the process and writing of the dissertation come to final fruition. For the recent publication of this manuscript, Laurence Smith, and his colleague David Davison, at Cambridge Monographs in African Archaeology are deeply thanked for accepting this work for publication, and the helpful and constructive editorial comments they provided. Danko Josić, who prepared the manuscript for final publication, also is appreciated. My second wife, Inna Tulchinskaya-Winchell, also deserves credit for giving me the last inspirational push for finalizing the manuscript for publication. Michael Brass at the Institute for Archaeology, University College London deserves final credit for resurrecting this work from an otherwise unpublished dissertation. It was Michael’s idea that I should get it published. Michael did much to make the initial contacts with the publishers at Cambridge Monographs in African Archaeology and in digitizing the original text. Michael also furnished me with many of the recent publications that I used in the Preface section of this work. May Michael and his young colleagues continue to endeavor in the research, analysis, and hopefully, future excavations in the Sudan, which this work may help to illuminate.

ix

Preface A Brief Review of Literature Since the Publication of this Work This study was completed as a dissertation for a doctorial degree in Anthropology in 1992. Since then, two decades of archeological investigations have been conducted in the Nile Valley, and adjacent areas, broadening the body of research and adding a number of new publications that were not covered in this study. In the Sudan, new sites and associated excavations (mostly in the central Nile Valley in and around the Khartoum area, and in the Dongola Reach area) have contributed significantly to the archaeological database since the early 1990s (See Edwards 2004 and Sadig 2010 for examples on overviews). Nevertheless, the basic themes and observations made in this study, involving the archaeological culture groups, associated paleoenvironments, material culture, settlement patterns, and overall culture history and chronology of Lower and Upper Egypt, Lower and Upper Nubia and the central Nile Valley and adjacent areas of eastern Sudan, remain relatively unchanged. In dealing with the development associated with the first ceramic cultures of northern Africa, one basic pattern not mentioned and improved upon since this study was completed, is that Dotted Wavy Line ceramics were dominant in the Sahara and Sahel west of the Nile and were contemporary with the earliest manifestations of Wavy Line ceramics in the central Nile Valley (See Caneva 1991: 267, Mohammed-Ali and Khabir 2003, Garcea and Hildebrand 2009: 314-315, Jesse 2010). There, it appears that Dotted Wavy Line ceramics come in later, perhaps indicating that populations, or ideas, from the Sahara and Sahel moved west into the central Nile Valley (Caneva 1991: 265-267). Thus, it appears that at least two separate ceramic traditions apparently evolved independently out of northern Africa, one in the central Nile Valley, and the other(s), across the Sahara and Sahel west of the Nile Valley (Caneva 1991), and perhaps as far west in Mali during the 10th millennium BC (see Hysecom, et al. 2009). Following the Early Khartoum in the central Nile Valley, literature since the early 1990s (however, see Hassan 1986) tends to divide the Khartoum Neolithic in an early and late period, and where researchers have drop the “Khartoum”; thus, either “Early Neolithic” or “Late Neolithic (see Sadig 2010 for example).1 As it will be presented in greater detail in this study, the Butana Group of the eastern Sudan belongs chronologically to the late Neolithic period, but is culturally separate from the developments in the central Nile Valley. As will be discussed in more detail in this study, the term post-Neolithic (which does not appear in the recent literature) is also used to demarcate those cultures, in particular the Predynastic groups in Lower and Upper Egypt where metallurgy began, but which were, nonetheless, contemporary to the late Neolithic developments further up the Nile in central and eastern Sudan. Perhaps some readers may object to the broad-brush use of early Neolithic and post-Neolithic chronological demarcations, and to which everyone familiar with these terms knows only too well that some cultures were, or were not truly “neolithic” in a technological, foodproducing sense. Furthermore, some archaeological groups, such as the Abkan culture actually transcended both the early and late Neolithic periods. Nevertheless, and as it will be further explained, this study will illuminate more on how these terms were originally defined and how they still can be effectively used for coherent temporal and spatial constructs that further aid in integrating the various ceramic-bearing cultures across northeast Africa, even in light of the more recent literature. The more recent publications have also provided a revised chronology for the early Neolithic (4900-3800 BC)2 and late Neolithic (3800-1250 BC) periods of the central Nile Valley (see Krzyzaniak 1992:243, Sadig 2008:45, and Garcea and Hildebrand 2009 3, Fernández, et al. 2003). In this study, I placed the beginning of the Khartoum Neolithic (i.e. early Neolithic) at 4900 BC, but had the late Neolithic begin slightly earlier at 1 Note that I retain the lower-case terms “early Neolithic” and “late Neolithic” to be more consistent with the use of such terms in the body of the original study. 2 Garcea and Hildebrand place the beginning of the early Neolithic at 5000 BC (see Garcea and Hildebrand 2009:306). 3 However, Garcea does not like using the term late Neolithic at Sai Island (personal communication, 2011).

x

4000 BC and ending sometime around 2500 BC. Through his excavations at es-Sour, Azhari Mustaf Sadig wisely uses the presence of ripple pottery as a general indicator of the late Neolithic in the central Sudan to which he ascribes a date of about 3800 to 3000 BC, and where he notes the Bardarian culture in Upper Egypt has ripple ceramics a few hundred years earlier, around 4000 BC, and where the Terminal Abkan has similar ripple ceramics around 3200 in Nubia (Sadig 2008:45, 2010:170-171, Bietak 1986).4 South of the 2nd Cataract, I place the beginning of the Abkan culture just prior to 4000 BC and it ending a little prior to 3500 BC. In Garcea and Hildebrand (2009), they place the chronological order of the Abkan culture between 5000 and 4000 BC, and cite Nordstrom 1972 and Lange and Nordstrom 2006 (and Gatto 2002, 2006a, b, and c) for placing the Abkan at this time (Garcea and Hildebrand 2009: 307). I also cite Nordstrom 1972 in this study suggesting that the Abkan can be dated as early as 3800 BC, or even earlier. Garcea and Hildebrand also state that there are Abkan-related sites in Nabta Playa that date to 4900 BC, and at Kadruka at 4800 BC (Garcea and Hildebrand 2009, Gatto 2006a). What probably has changed most since the publication of this study in 1992, is the addition of new archaeological investigations which have filled-in some of the gaps along the Sudanese Nile Valley where no research had been conducted prior to publishing the study (See Sadig 2010). In examining Figure 5.2 in this study, the reader will note that there is a considerable gap involving Neolithic and post-Neolithic archaeological data from Melik-en-Nasir where the Abkan Group ends south through the 3rd Cataract and Kerma, up to the Karat Group at the confluence of Wadi El Milk.5 Likewise, from the Karat Group upriver along the Nile past the 4th and 5th Cataract to Kadada, and further up the Atbara River to the Butana Group there was another gap where no archaeological sites were recorded at the time. What comes to light since the publication of this study, and supported by the more recent literature, is that a key element for some of these gaps along the Nile River is that less hospitable environments for human habitation were in existence since the early Neolithic period, and later on up through the late Neolithic, post-Neolithic, and later historic periods. Such areas would include Abu Hamed Reach between the 4th and 5th Cataract, the Abri-Delgo Reach between the 3rd Cataract and Dal Cataract, and Batn el Hajar, between Dal Cataract and 2nd Cataract. Nevertheless, recent investigations between the 2nd and 3rd Cataract, including the archaeological investigations at Sai Island, have found that Abkan Group ceramics (dating from 5000 to 4000 BC) are distributed farther up the Nile from the site of Abka, to Sai Island, Laqiya, Kadruda, and Dongola Reach (see Garcea and Hildebrand 2009:307). A Pre-Kerma archeological component (dating from 3300 to 2600 BC) has a similar distribution area along the Nile from Dongola to the south, northward to Kadruka, Kerma, Arduan, Laqiya, Soleb, Sai, Saras, Buhen and Faras (Garcea and Hildebrand 2009:307&310, Honegger 2004a&b, and 2006). Garcea and Hildebrand (2009) also see some overlap or affinity between the ceramics associated with the Abkan, Pre-Kerma cultures, and the A-Group, of which the latter has a more northern distribution along its traditional territory between the 1st and 2nd Cataracts. The Sai Island sites during the late Neolithic period probably represents a Pre-Kerma type of affiliation but provides ceramic and ethnobotanical data (presence of wheat and barley) that clearly shows cultural interaction with cultures associated with the A-Group and perhaps with Early Dynastic Egypt (Garea and Hildebrand 2009:319). In this study, the Pre-Kerma component or culture of the 3rd Cataract area was not defined nor discussed, since very little was know about this manifestation at that time. Since then, Matthieu Honegger has published a number of articles based on his excavations at Kerma and el-Barga defining what he calls the Pre-Kerma culture that was originally discovered by Bonnet in his excavations at Kerma in the late 1980s (Honegger 2003, 2004a, 2004b, 2004c, 2004d, and 2006). The Pre-Kerma culture has a similar chronological placement (existing between 3300 and 2600 BC) to the Butana Group (existing between 3750 to 2500 BC). The socioeconomic makeup of peoples associated with the Pre-Kerma culture are described as proto-urban, with a village site at Kerma which was 1.5 hectors in size (Garcea and Hildebrand 2009:310). They were cattle herders who may have practiced agriculture, as well, although there is no direct evidence for cultigens at this time (See Garcea and Hildebrand 2009, and Hildebrand 2006-2007). Ceramics associated with the Pre-Kerma culture include both mineral and organic tempered pastes (organic temper was dung), where pots were burnished or polished, sometimes with a red slip and often black-topped. Ripple wares appear to be It is also worth noting that the more recent publications uses the term “Early Neolithic” in place of the “Khartoum Neolithic” for the central Nile region (referred also as the central Nile Valley in this study). 5 Sadig has illustrated that along the Nile from the Delta down through the Khartoum area from 5300 to 3500 BC, there were concentrations of populations centering around the Egyptian Delta area, then around the Great Bend and south through Upper Nubia down to about Melik en Nasir, then picking up smaller pockets of populations around Kerma and the Karat Group at Wadi El Milk, and finally along the Khartoum area (See Map 1.7 on page 22 in Sadig 2010). Gaps where there were fewer to no settlement generally correspond with the gaps depicted in Figure 6.2 of this study. 4

xi

common, and other ceramic decorations include incising, impressing, rocker-stamping, horizontal and vertical combing. The distribution of Pre-Kerma ceramics appears to go as far north as the 2nd Cataract, overlapping with A-Group cultures, and extends as far south as the 4th Cataract. As discussed above, the archeological sites associated with Sai Island may have been a northernmost facies of the Pre-Kerma culture and where the presence of wheat and barley appear there at around 2700 BC, indicating some contact with A-Group cultures or Egyptian groups farther to the north (Honegger 2004a, Garcea and Hildebrand 2009:311). Overall, it would be interesting to make further comparisons between the Butana Group and the Pre-Kerma culture as it appears that both groups were contemporary, notwithstanding that the Butana Group was more than 800 kilometers to the south and east of Kerma. In fact, some preliminary comparisons between ceramics of the Butana Group and Pre-Kerma culture have been made (see Manzo 2010 below). The use of horizontal combing as an exterior surface treatment, and as noted with some of the Pre-Kerma ceramics, is a hallmark of the Butana Group and overall principal trait associated with the Atbai Ceramic Tradition of the eastern Sudan. One important distinction to be made, however, is that ripple wares quickly disappear in the southern Atbai from the Malawiya/Butana transition where they are virtually absent in the Butana Group by 3750 BC. As discussed further in this study, the disappearance of ripple wares in the Butana Group may signal a cultural separation of the Butana Group from other developments along the central Nile Valley at this time. It is fascinating to speculate that perhaps one of the reasons for the Butana Group separating culturally from the Nile could have been due to emerging Kerma culture in the Dongola Reach area, of which peoples associated with the Butana Group retreated from. However, Manzo (2010) contradicts this, and postulates that there is new archeological data, supported with more current linguistic evidence, that shows the Pre-Kerma culture and Butana Group were related. In any case, the Butana Group started out as a larger cultural manifestation than the Pre-Kerma culture (the largest Butana Group sites were 8 to 12 hectors in size in comparison with the Pre-Kerma site at Kerma which was 1.5 hectors), and where the former predated the latter culture by at least several centuries. Nevertheless, both groups appear to have practiced a similar subsistence pattern (however, domesticated animals appear later in the Butana Group) and were probably organized along a chiefdom-like social structure (Fattovich 2010). Both groups may have also been in the beginning stages of domesticating plant resources indigenous to Africa, especially sorghum and millet. Through his synthesis of recent archeological investigations in the Sudan, David Edwards also gives a wellreasoned discussion about an overall reassessment of the Egyptian-centric A-Group phenomenon in Lower Nubia where recent archaeological investigations further south in the Sudan demonstrates that the A-Group may represent a more northern-most manifestation of a Sudanic-centric culture originating in the heavily populated areas of Dongola Reach and further south into the Khartoum region (Edwards 2004: 68-74). The early A-Group cultures near the 1st Cataract may indeed indicate a movement of Naqada II cultures extending south from Upper Egypt, but where the southerly Classic/Middle A-Group cultures were more Sudanese in comparison. Nevertheless, additional work needs to be done in discerning these kinds of cultural differences (Edwards 2004: 70). By the mid third millennium BC, Edwards contends that the concentrated settlements of Dongola Reach associated with the Kerma Culture from Mirgisse to the north, and as far south as Barkal near the 4th Cataract, is the dominant culture south of the Egyptian state, and that the contemporary C-Group that re-inhabits Lower Nubia during this time is actually a northward extension of the early Kerma culture (Edwards 2004: 75-78, 110-111). Thus, Edwards asserts that the origins of C-Group is probably based in the more heavily populated area of Dongola Reach as opposed to Lower Nubia (Edwards 2004; 77, Gratien 1978, 1995, Bietak 1979). In her comprehensive analysis of the A-Group and other Nubian/Sudanese groups in the Sudan and Egypt, Maria Gatto includes site KG28 (associated with the Malaywiya/Butana transition) of the Butana-Gash area as part of a greater Nubian phenomena which she concludes is culturally different from the predynastic developments in central and southern Egypt, but separate from the central Nile Valley (Gatto 2002: 11). She feels that KG28 is more akin to a lower Nubian facies (i.e. Nubian Group), somewhat related to the A-Group, but different from developments in the Khartoum region (i.e. Sudanese Group). Overall, Gatto sees the A-Group as a significant indigenous development in northern Sudan which was a much more widespread Nubian manifestation than being merely restricted along the Nile Valley between the 1st and 2nd Cataracts. She argues that A-Group peoples were also traders who ventured into both the Western and Eastern Deserts of Egypt and the Sudan, including the Atbai region. Gatto also postulates that the A-Group likely developed out of the early Abkan Group, and where the Abkan may also have been the progenitor of the Badarian (or Tasian) culture further north in Egypt (Gatto 2006b). It is interesting to note that in her dissertation (2001), Gatto did a similar synthesis of ceramic-bearing cultures in Northeast Africa as discussed in this work. Neither author appeared to have been aware of each other’s research at the time, however. Nevertheless,

xii

Gatto, as a talented expert in ceramic analysis, has fortified her basic research with a number of recent field investigations in the Sudan and Egypt. Farther to the south in the Meroe region, and just north of the site of Meroe, Azhari Mustafa Sadig has conducted excavations at the late Neolithic site of es-Sour (Sadig 2005, 2008, 2010). The site of es-Sour consist of two low mounds together with an area of 176 by 90 meters and dates sometime between 4300 and 4000 BC (Sadig 2008: 39-41, and Sadig 2010: 34-35).6 A number of pot burials have been recovered there, some buried in ripple-marked vessels. Basically, the ceramics and material culture of es-Sour is comparable to other late Neolithic manifestations in the central Nile area, such as Kadada to the south, and other nearby sites of Al Ghaba and Ushara. North of es-Sour, Sadig recorded another 15 archeological sites along the Atbara-Shendi Road on the east side of the Nile to the confluence of the Atbara River, where 8 of the 15 sites dated the 3rd millennium BC and where the components at these sites (dating before 2500 BC) are comparable to materials associated with the Pre-Kerma culture (Sadig 2010: 95). These particular sites are also not too far down river from the Butana Group sites in the southern Atbai region and could perhaps share some similarities. Archeological investigations along the Atbara near the confluence of the Nile has also been done by Randy Haaland and Sadig where they located three sites that dated to the Early Khartoum period (Early Holocene) (Haaland 1995, 1997, 2005, Haaland and Magid 1995, Edwards 2004: 25-26). Haaland notes that there were thin cultural deposits dating to the Neolithic period (Mid-Holocene) above the more substantive Early Khartoum occupations at the three archaeological sites. Although Haaland describes the ceramics associated with the Early Khartoum period at these three sites, there is little about the ceramics dating to the Neolithic period. Nevertheless, Haaland states that a decrease in rainfall from the Early to Mid-Holocene times resulted in fewer people inhabiting this region near the confluence of these two rivers after the Early Khartoum period (Haaland 2005). In the Gezira plains between the White and Blue Nile, the more recent investigations by Rudolf Gerharz have significantly modified the ceramic sequence at Jebel Moya (Gerharz 1994). Based on Gerharz’s reanalysis, it appears that the earliest occupation at Jebel Moya begins with Phase I (beginning at ca. 5000 BC) which is followed by a nonconforming Phase II (beginning at ca. 3000 BC) and which Phase II develops into a Phase III occupation (ca. 700 BC) and lasts on up to around 100 BC when the site of Jebel Moya was abandoned. The Phase I occupation appears to be a later manifestation of a late Mesolithic (Early Khartoum) occupation characterized by unburnished Dotted Wavy Line and impressed wares. The Phase II occupation appears not to have evolved out the earlier Phase I occupation as there is a hiatus between the two phases at Jebel Moya. The Phase II occupation is more substantial and is characterized by ceramics with everted rims (called “Rabak ware”) similar to the contemporary occupation at Rabak, some 70 kilometers to the west, and a more predominant group of ceramics characterized by rocker-stamped and incised decorations over a burnished exterior. There appears to be some similarity of the Phase II occupation at Jebel Moya with the Pre-Kerma and C-Group cultures further to the north. The phase III occupation at Jebel Moya continues with the same Phase II ceramic tradition with the addition of imported trade items associated with the Napatan and Meroitic cultures, including jewelry of semi-precious stones, ivory, gold, and silver, along with smaller tools and weapons made of iron. Michael Brass who is in the process of reanalyzing the archaeological collections at Jebel Moya, basically agrees with Gerharz’s chronological placements; however, Brass suggests that the early Phase I component could be also be placed perhaps within a Early Neolithic context, and he expects that additional new insights will be obtained when he reanalyzes the ceramics (Brass 2011, and personal communication, 2011). It is also quite conceivable that Brass’ new analysis on the ceramics at Jebel Moya could shed additional light on whether there are any connections with the late Neolithic developments in the southern Atbai. In the southern Atbai region where the focus of this study concentrates on the development of the Butana Group, no new intensive investigations have taken place there since the late 1980s. However, through a reconnaissance survey conducted in November 2010 by the University of Naples (as an outgrowth of Rodolfo Fattovich’s work and direction through the Italian Archaeological Expedition to the Sudan, Kassala) Andrea Manzo and his colleagues did reinvestigate the southern Atbai region in and around Kassala (investigations at Mahal Teglinos stopped in 1995) and further to the west where sites associated with the Butana Group had

The 4300 to 4000 BC occupation at es-Sour would place it at the end of the early Neolithic period, however, the presence of ripple-marked ceramics would characterized the occupation as late Neolithic. As a result, a 4000 BC date for the beginning of the late Neolithic period as argued in this study might be more appropriate. 6

xiii

been previously investigated in the 1980s (Manzo 2010).7 New areas north of the 1980s investigations were also surveyed. In total, Manzo and his colleagues recorded 135 sites, representing the full range occupations associated with the Atbai Ceramic Tradition. Basically, Manzo and his colleagues reconfirmed Fattovich’s, Marks, Mohammed-Ali, and Sadr’s earlier work about the development of the Atbai Ceramic Tradition and the related movements and social evolution of its people. Manzo also noted that an earlier Butana Group component was documented at Mahal Teglinos which underlay the main Gash Group component there (Fattovich, Manzo, and Usai 1994: 15). The earlier Butana Group component at Mahal Teglinos contained ripple-decorated ceramics (Manzo 2010:5). Manzo and his colleagues believe that the Butana Group ceramics at Mahal Teglinos, and the Butana Group occupations further west in the southern Atbai also attest to strong cross-cultural relationship with other groups in Upper Nubia, namely the Pre-Kerma culture of the Dongola Reach area (Manzo 2010: 5 and 31). Manzo and his colleagues further state that the similarities between the Butana Group and pre-Kerma culture ceramics are “remarkable” and that there may have been an overland route from the confluence of the Atbara and Nile rivers to the 4th Cataract area (Manzo 2010: 5 and 31, Honegger 2004d, 85, Figure 3-4, 2004a 39-40). Manzo uses the linguistic correlations of Claude Rilly who argues for a broader Northern East Sudanic language group (of Nilo-Saharan stock--see Rilly 2009) to support similarities observed in the ceramics between the Butana Group and Pre-Kerma culture (Manzo 2010: 31, Rilly 2008: 6-10). Finally, the more recent literature produced since the publication of this study include a number of brief references about the southern Atbai and Atbai Ceramic Tradition, including the Gash, Butana, and Mokram Groups of the Kassala Phase (some of the more recent literature refers to the Mokram Group as the “Jebel Mokram Group”). It is important to note that the recent references about the southern Atbai and related archeological cultures of the Atbai Ceramic Tradition is based on the earlier writings of Mohammed-Ali, Fattovich, Marks, and Sadr of the 1980s and early 1990s. Beginning with Sadig’s 2010 publication, Sadig make the observation that by the 5th millennium (before present), the “Atbai plains east of the Nile in Eastern Sudan” increasing appear to be culturally distinct from the riverine areas (Sadig 2009: 258; Sadig 2010: 50, Mohammed-Ali 1985, Fattovich, Marks, and Mohammed-Ali 1984: 182). Later, Sadig states that the Butana and the Atbai plains east of the Nile Valley provides indirect evidence of domesticated plants associated with the later Neolithic cultures of the Kassala phase, namely the Mokram Group, where seed inclusions in potsherds have been identified as domestic sorghum (Fattovich, Marks, and Mohammed-Ali. 1984; 182, Sadig 2010: 75). It is interesting to note, that in this study there are sherds associated with the earlier Butana Group of the Kassala phase where such seed inclusions could be domesticated varieties of sorghum or millet (see the Khordhag Plain type in Appendix A). Sadig also makes observations that similar-looking finely ground stone mace-heads recovered at es-Sour resemble similar examples found in the eastern Butana and near Kassala (Marks et al. 1985: 47, Sadig 2008: 39; 2010: 144). Such similar looking mace-heads made from porphyry are located on Butana Group sites dating to the 4th millennium B.C. (See Mbutu 1991: 364, Plate 4). In 2003, Abbas Mohammad-Ali and Abdel-Rahim Khabir note in their physicochemical analyses, that a principal tempering material used in ceramics of the Khasm el Girba area was derived from Abyssinian basalts that were transported by the Atbara River (Mohammad-Ali and Khabir 2003: 32). Although this study did not attempt to source the mineral tempering materials used in the making of Butana Group ceramics, it seems perfectly reasonable that the non-plastic materials were derived from igneous rocks that were recovered or found near the Atbara River in the southern Atbai region. In a reanalysis of the ceramics recovered by Arkell in 1954 at Agordat, Alemseged Beldados notes that combed ceramics (combed on both the interior and exterior surfaces) are present at this site, along with a characteristic appliqué technique on the sites of pots that he notes are similar to such motifs found at Gash Group sites dating between 3500 to 3820 B.P. (Beldados 2010: 97, Phillipson 1977: 60). The combed ceramics noted at Agordat are also comparable to the Butana Combed type ceramics described and illustrated in this study, and the appliqué motif can also be similar to the motif (called the Atbara Rim Band Mode) noted on the exterior of wiped ceramics associated with the Socodad Wiped ceramics of the Butana Group (See Appendix A). Beldados also notes that the prevalent cross-hatch decorative pattern present in Agordat ceramics is associated with other ceramics of Atbai Ceramic Tradition (Beldados 2010: 99, Fattovich, Marks, and Mohammed-Ali 1984). Such cross-hatch decorative patterns can perhaps be seen on Butana Group ceramics associated with the type Zirhalla Pattern Burnished described in this study (See Appendix A in Manzo and his colleagues plan to return to the southern Atbai in 2011 and will no double elaborate more about the Butana Group and Atbai Ceramic Tradition. Publications on their work are expected to come out later in 2012.

7

xiv

this study). Rodolfo Fattovich (2002), believes that Agordat is actually a northern outlier, or is within the sphere of influence of Mahal Teglinos and the Gash Group. Brass makes the observation that the presence of new ceramic styles seen across the Sudan and which are present Napatan and Meroitic sites may have some connections with the Butana Group ceramics in general (Brass 2009: 123, Mohammed-Ali and Khabir 2003, Winchell 1992). This would be especially true with the decorated ceramics associated with the late Butana Group, as well as many ceramics types associated with the Mokram Group that post-dates the Butana and Gash Group after 1500 BC. Victor Fernández, Alfredo Jimeno, Mario Menéndez, and Javier Lario note that ceramics decorated with simple zoned and impressed decorations associated with Late Neolithic assemblages from the Blue Nile area of the central Sudan are comparable to the Gash Group of the southern Atbai area of eastern Butana (Fernández et al. 2003: 255, Sadr 1991). Comparable Butana Group ceramics in this study would be associated with the Sirjino Simple Dentated type (see Appendix A). These authors further note that another comparable feature between the Late Neolithic Blue Niles ceramics and the southern Atbai ceramics of the same period is the prevalence of rough wide scraping or combing of the surfaces of pots (Fernández et al. 2003: 260, Fattovich, Marks, Mohammed-Ali 1984). As with similar combed ceramics noted at Agordat, such combed ceramics associated with the Butana Group would be the Butana Combed type noted above. In surmising the basic economic subsistence pattern of late Neolithic cultures in the Blue Nile and Western Butana, Fernández, Jimeno, Menéndez, and Lario, observe similar patterns in the southern Atbai, noting that all of these groups exploit significantly larger proportions of wild savannah species over domesticated ones (Fernández et al. 2003: 263, Fattovich 1991b, Sadr 1991). They also note that evidence for food production does not appear until the later half of the Butana Group (ca. 6000-45000 bp), but being fully established in the Gash Group (ca 4500-3500 bp) (ibid.). These observations made by Fernández, Jimeno, Menéndez, and Lario appear to be consistent with data presented in this study on the subsistence pattern of the Butana Group population; however, there remains the possibility of Butana Group peoples domesticating sorghum and millet from the early Butana Group period from as early as 3750 BC on through the late Butana Group period at about 2500 B.C. The clearest indicator of a continual use of possible domesticated plants in the Butana Group is seen through the seed-tempered sherds associated with the Khordhag Plain type (see Appendix A in this study). In terms of macro settlement patterns and the movements of peoples of the central Sudan at the close of the Late Neolithic period (ca. 2500 BC) and beginning the of Napatan-Meroitic periods (ca 700 BC), Fernández seems to agree with Haaland’s earlier observations that Cushitic-speaking populations were moving out the Eastern Butana/Khashm el Girba area into the central Nile and constructing tumuli in the Khartoum region (Fernández 2003: 417-418, Haaland 1987: 224-231). Again, Fernández sees similar ceramic decorative traits such as exterior cross-hatched pottery between the Nubian Pan Grave culture of Northern Sudan with those associated with the Mokram Group of the southern Atbai in the eastern Butana, who, in turn, may be have been ancestral to the present-day Cushitic-speaking Beja (Fernández 2003: 418, Caneva 2002, Sadr 1990). M.J. Grove (2007) uses Karim Sadr’s 1991 synthesis on nomadism in the southern Atbai and uses a complex series of statistics and computer simulations to demonstrate that the spatial patterning of sites shows an increase in mobility through time between cultural groups associated with the Butana, Gash, Mokram, and Hagiz groups of the Kassala and Taka phases of the Atbai and post-Atbai Ceramic Tradition. Grove shows that among the four groups he analyzed, the Butana Group demonstrated the lease mobility, followed by slightly more mobility in the Gash group. Thus, Grove’s least mobility model for the Butana Group would give credence that this group represented, for the most part, a sedentary society. Finally, Rodolfo Fattovich has recently synthesized the entire Atbai Ceramic Tradition that he, Anthony Marks, and Abbas Mohammed-Ali originally defined in 1984 into the developments of complex societies and states in the northern Horn of Africa (Fattovich 2010). He basically demarcates the northern Horn of Africa cultural developments into at least two distinct trajectories that are indirectly related; one occurring in the Eritrean-Sudanese lowlands, and the other in the Eritrean-Ethiopian highlands, culminating into the Kingdom of Aksum. Fattovich puts forth that the Atbai Ceramic Tradition represents the material manifestations of the Eritrean-Sudanese lowlands which develops in lowland plains of eastern Sudan between Kassala and Khashm el Girba (Fattovich 2010:150). As it will be explained in greater detail in this study, Fattovich surmises that the Atbai Ceramic Tradition is based in the earlier Early Khartoum ceramic tradition and divides the Atbai Ceramic Tradition into three phases and associated groups: (1) Saroba Phase (ca. 5000-3800 BC), including the Malawiya Group; (2) Kassala Phase (ca. 3800-800/700 BC), including the Butana Group (3800-3000 BC) and Gash Group (ca. 3000-1500/1400 BC), and Jebel Mokram Group (termed in this study as the Mokram

xv

Group as it was originally defined) (ca. 1500/1400-800/700 BC); and (3) Taka Phase (ca. 800/700 BC to AD 300/400) including the Hagiz Group (Fattovich, Marks, and Mohammed-Ali 1984, Fattovich 2010: 150). Although Fattovich believes that the origins of the Atbai Ceramic Tradition may be located at the Amm Adam 1 site and related sites in the northern Gash Delta sometime in the 6th millennium BC, he notes the primary early development of the Atbai Ceramic takes place within the Khashm el Girba region associated with the transitional Saroba/Kassala Phase archaeological site of KG28 (ca. 4200-3800 BC), and in particular with the subsequent rise of the very large archaeological sites associated with the Butana Group of the Kassala Phase, such as KG23, KG7, and KG96. These archaeological sites and the sudden dramatic rise of the Butana Group is discussed in greater detail in this study. Fattovich also notes that porphyry mace-heads found on Butana Group sites may indicate some class differentiation, since porphyry had to have been transported about 300 kilometers from their original sources located in the Red Sea Hills (Marks, Mohammed-Ali, and Fattovich 1986, Sadr 1991, Fattovich 2010: 154). Fattovich goes on to say that such porphyry mace-heads have been found in other contemporary cultures in Egypt and Nubia and are associated with the high elite burials. Fattovich postulates that the Butana Group developed into a chiefdom society associated with the later Gash Group of the Kassala region where the latter group was centered at Mahal Teglinos and Shurab el Gash (Fattovich 2010: 154 and 160). Funerary stelae erected at Mahal Teglinos during middle and late times (ca. 2300-1900 BC and 1700-1500/1400 BC) of the Gash Group resemble similar-looking stelae found at early proto-Aksumite tombs (ca. 400-50 BC) of the Eritrean/Ethiopian highlands, to which Fattovich believes could show some cultural connection between the Gash Group and formative development at Aksum (Fattovich 2010: 158, 161). In any case, the archaeological record in Fattovich’s perspective demonstrates that the mixed foraging/farming/herding economy associated with the Butana Group develops into a mixed foraging/herding economy associated with the Gash Group (Fattovich 2010: 161). As discussed in more detail in this study, and as Fattovich (2010) points out, the transition between the Butana and Gash Groups is rooted in a physiographic shift of the Gash River migrating away from its confluence to the Atbara River to the present-day Gash Delta. Thus, by the middle of the 3rd millennium BC, the Khashm el Girba homeland of the Butana Group was essentially abandoned, where peoples slowly migrated with the Gash River into the Kassala region and establishing themselves around the centers of Mahal Teglinos and Shurab el Gash. Due to a continued drying trend in the Eritrean-Sudanese lowlands, the culture of the Gash Group in Kassala diminishes significantly and people disperse into smaller but more widespread agro-pastoral communities associated with the Jebel Mokram Group which basically divided into two cultural economic divisions; one being associated mainly with cultivation and the other with herding (Marks and Sadr 1988, Sadr 1991: 52-56, Fattovich 2010: 161). The Jebel Mokram Group ceramics consist of two principal wares; one associated with the Gash Group Ware, and the other with the Middle Nubia Ware, to where the Middle Nubia Ware indicates an intrusion of people coming in from the Eastern Desert into the central Sudan into Eritrean-Sudanese lowlands of the Khashm el Girba and Kassala regions (Fattovich, Sadr, Vitagliano 1988-1989, Fattovich 1991b, Sadr 1991: 45-48, Fattovich 2010: 150-151). Fattovich notes that there are archaeological sites near Agordat in the middle Barka valley of Eritrea that contain ceramics similar to those found in late Gash Group and Jebel Mokram components. This would suggest that the widespread phenomenon of the Jebel Mokram Group of the Atbai Ceramic Tradition reached into, or influenced other cultures in the central Sudan/Nile area across into the Eritrean/Ethiopian Highlands, and maybe farther north into historic Nubia. Again, this study goes into more detail about the transition from the Butana to Gash to Jebel Mokram Group and which parallels close in what Fattovich expressed in 2010 (See chapter 2 in this study). To close, in light of the more recent literature, the place the Butana Group ceramics have in the Neolithic and Post-Neolithic eras of northeast Africa, and in relation to the other ceramic-bearing cultures in the region, appears to correlate well with the observations and conclusions originally made in this study. One observation not made in this study, and probably warrants additional research is the relationship between the Butana Group and Pre-Kerma culture. Nevertheless, the Butana Group as an archaeological phenomenon, represented a major development in the eastern Sudan that appeared as an autonomous and indigenous manifestation in the first quarter of the fourth millennium BC and lasting into the first half of third millennium BC, totaling an occupation spanning more than a thousand years. There are no abrupt changes within the Butana Group suggesting that it was relatively stable over this period of time. As Fattovich (2010), Sadr (1991), and others hypothesize, the disappearance of the Butana Group in the southern Atbai can be attributed to the northeastern migration of the Gash River, which had originally drained into the Atbara, and which eventually shifted course and drained into the inland Gash Delta by the third millennium BC. As Fattovich points out, and to which this study make an unequivocal conclusion, the ceramics of the Butana Group are directly ancestral to the ceramics associated with the Gash Group at Mahal Teglinos and Shurab el Gash. Why the Butana Group did not spread further downriver along the Atbara to the Nile, may have simply

xvi

been an unintended byproduct of the Gash River shifting its course inland and away from the Atbara River.8 Unlike the Pre-Kerma culture, there also appears to be no outward push of the Butana Group into other areas in the central Sudan, either across land or along the Nile. However, this hypothesis may be modified (or even changed) in the near future, perhaps suggesting that at least during the development of the early phase of the Butana Group, that peoples in the southern Atbai were indeed exporting ideas (or even people) down the Atbara River, and beyond. Nevertheless, the Butana Group ceramics do appear to represent a clear cultural separation from the Central Nile developments, where such decorative traits as rippling the surfaces of ceramics, which manifested in greater numbers along the Nile from Lower Nubia into the Khartoum region, disappeared entirely from the Butana Group in the southern Atbai. Finally, and as this study attests, the Butana Group remains as a significant cultural development, perhaps exporting some of its innovations outside the southern Atbai, and perhaps fostering the first domesticated varieties of sorghum and millets in this part of Africa that were later adopted by the Gash Group at Mahal Teglinos, and perhaps elsewhere. Such innovations would have allowed peoples of the Butana Group to congregate and settle into some of the largest village sites known in northeast Africa at the time. Frank Winchell Washington, D.C.

8 Nevertheless, additional research downriver along the Atbara may prove that some aspects of the Butana Group did indeed move in this direction.

xvii

Chapter 1

Introduction

Our understanding of the Neolithic and post-Neolithic eras in northeast Africa (from ca. 5000 to 2000 BC) has been based primarily on the well known ceramic-bearing groups of the Egyptian Neolithic/Predynastic, the Nubian A-Group, and the Khartoum Neolithic (see Figure 1.1) which have all been found in the Nile Valley (Petrie 1901; Reisner 1910; Brunton and Caton-Thompson 1928; Caton-Thompson and Gardner 1934; Adams 1977; Hoffman 1979; Haaland 1987: Caneva 1988; Hassan 1988). Almost nothing is known about contemporary ceramic-bearing groups situated well to the east of the Nile Valley in the semi-arid region of the eastern Sahel where it has been hypothesized that some cultures after 4000 BC were in the process of developing methods of cultivation utilizing native African plants (Harlan et al. 1976; Phillipson 1985: 115; Clark 1984: 113-126; Stemler 1984: 127-131; Fattovich, Marks, and Mohammed-Ali 1984; Marks and Sadr 1988: 69-90).

A-Group (Wendorf 1968: 1053; Adams 1977: 122; Nordstrom 1972: 12-17, 28; Hassan 1988: 139). On the other hand, the Karat Group appears to have some cultural connection to, or contact with, both the Abkan and the Khartoum Neolithic (Marks and Ferring 1971: 189-190; Nordstrom 1972: 16). In addition to the Abkan and Karat groups are a series of less well-known ceramic-bearing cultures which replaced the Khartoum Neolithic in the central Nile Valley. This particular cultural manifestation has been recently termed the late Neolithic (Caneva 1988). Despite significant cultural and economic differences (cf. Marks and Ferring 1971; Nordstrom 1972; Hoffman 1979; Caneva 1988; Hassan 1988), the common denominator among all of these Neolithic and post-Neolithic ceramic-bearing groups was their close proximity and affiliation with the Nile Valley. Outside the Nile Valley, the Neolithic in northeast Africa east of the Sahara has been viewed as a peripheral phenomenon (Shiner et al. 1971: 3-6; Marks et al. 1985: 263) where ceramic-bearing cultures of the arid Sahelean steppes later acquired the trappings of agriculture and pastoralism (Haaland 1987: 23-24, 223-232).

The Egyptian Neolithic/Predynastic, lasting from about 5000 to 3000 BC, encompassed a series of Neolithic-Chalcolithic cultures which existed along the Egyptian Nile from the Delta to the 1st Cataract (Hays 1984: 66; Hassan 1988: 135). South of the 1st Cataract in lower Nubia, there is the Nubian A-Group (Reisner 1910), or A Horizon (Adams 1977: 118), which involves a series of similar Neolithic cultures which began sometime around 3500 BC and lasted through 3000 BC (Nordstrom 1972: 30; Adams 1977: 7). In the region of the central Nile, the Khartoum Neolithic encompasses a group of ceramic-bearing settlements situated north and south of the modern Sudanese capital of Khartoum Neolithic (Arkell 1953). This particular cultural manifestation was short lived like the A-Group, lasting only 800 years, from approximately 5000 to 4200 BC (Marks et al. 1985: 262). Of these three major cultural developments, only the Khartoum Neolithic appears to have developed in place from the former Khartoum Mesolithic which was the first ceramic-bearing culture in northeast Africa (Arkell 1949; Arkell and Ucko 1965; Hays 1971; Adamson, Clark, and Williams 1974; Marks et al. 1985; Haaland 1987; Caneva 1988).

The site of Shaqadud, (see Figures 1.1 & 1.2), located approximately 50 kilometers southeast of the central Nile Valley and 115 kilometers northeast of Khartoum, has revealed a stratified sequence from the Khartoum Mesolithic (Early Khartoum) through the Khartoum Neolithic (Marks, et al 1985). Besides local variations, the stylistic qualities of the Shaqadud Khartoum ceramics reflect the same progression of stylistic change seen between ceramics of the Early Khartoum and Khartoum Neolithic of the central Nile Valley (ibid. 1985: 265). In the very final part of the Shaqadud sequence, however, a different set of ceramics appears which does not share stylistic affinities with any known Nilotic ceramic bearing culture (ibid. 1985: 277). These particular ceramics date after 2200 BC and represent the first re-occupation of the Shaqadud site after 3000 BC. Based on preliminary results (ibid. 1985), these postKhartoum Neolithic ceramics are stylistically related to yet another series of related ceramic-bearing groups situated approximately 270 kilometers southeast of Shaqadud in the eastern Sahel, in an area known as the southern Atbai (see Figure 1.2). These particular ceramic-bearing groups have been recently defined under a single tradition known as the Atbai Ceramic Tradition (Fattovich, Marks, and Mohammed‑Ali

Other Neolithic era ceramic-bearing groups of northeast Africa, such as the Abkan of the 2nd Cataract district, and the Karat Group, of the Dongola Reach, are not as well known as are the three main Nilotic-Chalcolithic developments (Hassan 1984: 62-63; Shiner 1968: 629,789, Marks and Ferring 1971: 187-188, Nordstrom 1972: 8-17). Nevertheless, it is probable that the Abkan is somewhat related (perhaps ancestral) to the

1

The Butana Group Ceramics

Figure 1.1. The study area of Northeast Africa.

Mohammed-Ali and Jaeger 1989, Fattovich 1989; Sadr 1991) very little has been published about the ceramics of this east Sahelean ceramic-bearing tradition.

1984). As it will be discussed in greater detail below, many of the ceramic-bearing groups associated with this particular ceramic tradition spanned the critical 3000 year period which involved the development of the Neolithic and postNeolithic eras in northeast Africa (see Figure 1.3).

One of the ceramic groups placed into the Atbai Ceramic Tradition (Fattovich, Marks, and Mohammed-Art 1984) was labeled by Shiner as the Butana Industry and was defined originally by a group of similar lithic assemblages occurring on four of the five Butana sites initially found there (Shiner 1971: 335). From the start, however, it was apparent that the ceramics of the Butana Industry were quite distinctive (ibid.: 346-347) and since then, the Butana Industry has been renamed as the Butana Group, which is primarily defined by a number

The initial discovery of many of these particular ceramicbearing groups of the Atbai Ceramic Tradition goes back for some time, since Joel Shiner of the Northern Sudan Project first noted them in 1968 (Shiner 1971). Since the original descriptions (Phillipson 1977; Fattovich, Marks, and Mohammed-Ali 1984; Fattovich and Vitagliano 1987, 1989; Clark 1984; Marks and Sadr 1988; Marks and Fattovich 1989;

2

Introduction

Figure 1.2. The Southern Atbai and surrounding area.

of characteristic ceramics, the most notable one being that of a thick combed ware (Fattovich, Marks, and Mohammed-Ali 1984: 8).

Group to sometime after 3700 BC and lasting perhaps as late as 2500 BC (Marks and Sadr 1988: 88, Sadr 1988: 95). A mean early date for the Butana Group occupation would be approximately 3500 BC.

The ceramics of the Butana Group have yet to be systematically reported, but based upon tentative observations, they appear to represent a non-Nilotic ceramic manifestation (Shiner 1971: 347) which reflected a high point within the Atbai Ceramic Tradition (Fattovich, Marks, and Mohammed-Ali 1984). Overall, the ceramics of the Butana Group contain a rich and varied assortment of pottery types ranging from mundane utilitarian pots to highly decorated vessels. Calibrated radiocarbon dates (see Table 1.1) have bracketed the Butana

The Butana Group, as an archaeological entity, is basically represented by a limited number of large village sites, located on, or just east of, the Atbara River, straddling the extreme eastern Butana region and the western part of the southern Atbai (see Figures 1.2 & 1.4). Indirect evidence for an agricultural based economy is compelling, due to the massive size, depth, and density of midden deposits, including the fact that the location of Butana Group sites are situated in the

3

The Butana Group Ceramics

Figure 1.3. The culture chronologies of northeast Africa.

4

Introduction

Figure 1.4. The Malawiya/Butana transition and Butana Group sites of the Southern Atbai.

Table 1.1. Butana Group Radiocarbon Dates

1

Butana Group Site

Uncalibrated Date1

MASCA Date

Lab Number

Site KG28B Site KG23 Site KG23 Site KG23 Site KG7 Site KG7 Site N125

5168±67 bp 4727±154 bp 4527±253 bp 4519±67 bp 4569±68 bp 4421±93 bp 4410±90 bp

4018 BC 3544 BC 3319 BC 3283 BC 3551 BC 3163 BC 3152 BC

SMU 1193 SMU 1201 SMU 1155 SMU 1188 SMU 115 SMU 1156 TX 445

Mean date falls within 1 standard deviation.

5

The Butana Group Ceramics

most favorable areas for cultivation (Fattovich, Marks, and Mohammed-Ali 1984; Marks and Sadr 1988; Sadr 1991). As of yet, no evidence for cultigens has been found; nevertheless, domestic cattle and ovicaprid bones have been recovered in the upper levels of some Butana Group occupations (Peters 1986).

along the central Nile seem to have disappeared and the area was virtually abandoned after 5000 BP (Mohammed-Ali 1984: 117-118; Marks, et al. 1985: 262-263; Haaland 1987: 220). Conventional wisdom might suggest that some of the central Nilotic groups spread eastward into eastern Sahel, thus sparking a new cultural manifestation in the southern Atbai (cf. Haaland 1987). However, as stated earlier, the Butana Group ceramics appear to have very few stylistic similarities to those of the Neolithic ceramic-bearing groups of the Nile Valley. Of the three major Nilotic Neolithic ceramic-bearing cultures, the Khartoum Neolithic of the central Nile Valley is geographically the closest ceramic group to the Butana Group. This group is also associated with the abandonment of the central Nile after 5000 BP (Mohammed-Ali 1984: 117-118; Marks, et al. 1985: 262­263; Haaland 1987: 220).

When the Butana Group sites were first reported by Shiner (1971: 350), they appeared to represent a scatter of lithics and ceramic debris on “natural hill[s]”. However, upon close inspection, (ibid.) these “natural hill[s]” turned out to be huge midden deposits (laden with pottery and lithic artifacts) some of which cover an area greater than 100,000 square meters (Fattovich, Marks, and Mohammed-Ali 1984: 2). In fact, Shiner believed that one of the Butana Group sites “may very well be the largest prehistoric settlement in Northeast Africa” (Shiner 1971: 359). This claim may or may not be true, but there can be no doubt that some of the large Butana Group settlements were as big or larger than Predynastic villages in Egypt (cf. Hassan 1988: 152-154), three times the size of the largest Khartoum Neolithic occupations (cf. Haaland 1987: 42), and significantly larger and more intensively occupied than any Neolithic A Group site (cf. Nordstrom 1972: 23, 134-249).

Approximately 20,000 ceramic sherds (which will form the data base for this study) from eight sites recovered from the Khashm el Girba area will be examined in detail (see Figure 1.4). One site (KG28B) is considered to be the only known transitional Malawiya-Butana Group site in the southern Atbai. The remaining seven sites (KG23, KG7, KG1, KG29N, KG29S, KG96, and KG5) represent Butana Group occupations and are considered to be the largest and most representative sites in the study area. Overall, a single, inter-site typology will be created from these ceramics. This typology will be made up of ceramic types, of which many will prove to have temporal significance within and between the eight sites used in this study.

Archaeological investigations conducted by a joint effort with the University of Khartoum, Southern Methodist University, and the Oriental Institute of Naples have confirmed Shiner’s original assessment associated with the Butana Group were engaged in some form of craft specialization and long distant trade (Shiner 1971; Fattovich, Marks, and Mohammed-Ali 1984; Fattovich of the Butana Group (Fattovich, Marks, and Mohammed-Ali 1984). Beyond the sheer size of the Butana Group settlements is the undisputed radiocarbon evidence that they were contemporary, in part, with some of the Neolithic and post-Neolithic developments on the Nile (associated with Predynastic Egypt, Nubia, and the central Nile) and were to some degree separate from them. Butana Group assemblages also reveal signs of a complex social organization from its earliest beginnings. Possible ceramic workshops and finely made polished stone maceheads and armlets (originating from porphyry sources located outside the southern Atbai) suggest that people associated with the Butana Group were engaged in some form of craft specialization and long distant trade (Shiner 1971; Fattovich, Marks, and Mohammed-Ali 1984; Fattovich 1985; Marks and Sadr 1988; Sadr 1991). Curiously, there are virtually no material goods from either Nubia or Egypt at any of the known Butana Group sites.

Based on radiocarbon dates, and a tentative evaluation of the ceramic data, it is believed that the above eight sites can be arranged sequentially from about 4000 BC to 2500 BC (Fattovich, Marks, and Mohammed-Ali 1984; Marks and Sadr 1988; Sadr 1991). This can be done by a frequency seriation of the ceramic types (based on stylistic differences) from one site to the next. Based on this ceramic seriation, it will be possible to prove or disprove the Malawiya-Butana transition from the site of KG28B to the earlier Butana Group sites. In addition, if a ceramic seriation can be established from roughly 4000 to 2500 BC, it is likely that there will be stylistic change from early to late Butana Group ceramics. This will be of critical importance in recognizing the so-called late Butana ceramic traits which have been noted in the beginning of the Gash Group east the Khashm el Girba area in the vicinity of Kassala (Fattovich 1989). There is also the possibility that some late Butana elements may have spread as far west as Shaqadud Cave.

Overall, the cultural manifestation of Butana Group implies that the Neolithic and post-Neolithic era of northeast Africa can no longer be examined as an exclusive Nilotic phenomenon. At present, it appears that the Butana Group represents yet another significant and advanced Neolithic manifestation which took place in the far reaches of the eastern Sahel.

In summary, this study will first establish a general typology of the Butana Group ceramics. This typology will then be examined among the eight sites used in this study to determine how the Butana Group ceramics changed overall within an approximate 2000 year time period. The temporal-typological characteristics of the Butana Group ceramics will then be compared and contrasted with the various Neolithic and post-Neolithic ceramic groups of the Nile Valley, in particular those ceramic groups located within the central Nile area of the Sudan. Other comparisons will be made with the Butana Group ceramics concerning the initial development of the Gash Group

As it will be shown later, the Butana Group in many ways developed rather suddenly and dramatically in the southern Atbai which has lead some researchers to believe that this group may have been a result of different populations moving into this area from somewhere outside (Haaland 1987: 230). Indeed, during the development of the Butana Group, cultures

6

Introduction

in the Kassala area and the late occupation at Shaqadud Cave. In the end, this study will make a definitive statement on what the Butana Group ceramics are as an archaeological entity and

how they fit within the overall context of the Neolithic and post Neolithic eras of northeast Africa east of the Sahara, given our present available data.

7

Chapter 2 Background

Introduction

Three rivers, the Setit, Gash, and Baraka, cross the southern Atbai. Their headwaters are located in the Eritrean highlands (where the annual rainfall is significantly higher), and during the summer months, these rivers are credited with discharging massive amounts of silt laden water across and throughout the southern Atbai (ibid.). The Atbara River flanks the western edge of the southern Atbai and flows into the Nile River approximately 300 kilometers north of Khartoum at the town of Atbara.

This chapter will integrate the Malawiya/Butana Transition and Butana Groups within the overall geography, environment, and archaeology of the southern Atbai. Both the Malawiya/Butana Transition and Butana Groups have been defined provisionally as groups of sites each of which shares a number of stylistically similar ceramics (Fattovich, Marks, and Mohammed-Ali 1984). These sites are located in the western sector of the southern Atbai and date between 4000 BC and 2500 BC. The Malawiya/Butana Transition and Butana Groups account for only a small fraction of the prehistory of the southern Atbai and one quickly appreciates the immensity of this region (in time and space) for archaeological study.

The most significant geographic feature in the southern Atbai is the Gash Delta. It is an inland delta, formed by tons of annual sediments discharged by the Gash River. The Gash Delta is the most fertile land in the southern Atbai and is optimal for agriculture, where today cotton, sorghum, millet, oranges, melons, and other assorted vegetables and fruits are grown in abundance. The Gash Delta is approximately 20 kilometers wide and spreads for more than a hundred kilometers north of the town of Kassala. The Gash floods every year in the summer and early fall, from late June to mid September, and inundates the delta area with fresh silts. These silts will stay saturated with water up to six months after the floods (Barbour 1964). The delta’s ability to retain water for long periods after the flood is due to the underlying clays which stop the water from percolating through (ibid.). The high water retentiveness of the Gash sediments causes the soils to stay moist for a large portion of the year, facilitating the growth of plants without extensive use of irrigation. As will be pointed out later, these water saturated silts must have been an attractive feature for prehistoric groups who could have cultivated plaints without relying on irrigation.

Geography of Southern Atbai The Land of the Southern Atbai The Malawiya/Butana Transition and Butana Group ceramics investigated in this study come from eight sites located in the area known as the southern Atbai (see Figures 1.3, 1.4, & 2.1). The southern Atbai is actually the southern tip of the Atbai region which encompasses a large stretch of semi-arid steppe in east central Sudan, in the far eastern Sahel. Its western border is defined by the Atbara River and partly by the Nile River north of the town of Atbara. Its eastern border is defined by the Red Sea Hills in eastern Sudan and by the western edge of the Eritrean Highlands. Traditionally, the Atbai region has been defined as the territory of the Beja, a composite culture group composed of a half dozen or so Arabic speaking tribes of nomads and agro-pastoralists (Barbour 1964). Annual rainfall in the Atbai region varies from as little as 25 millimeters in the north to as much as 800 millimeters in the extreme south.

Apparently, some time before the end of the Middle Holocene the Gash once emptied into the Atbara River around the Khashm el Girba area (0.S.G. Crawford 1951, Barbour 1964, Coltorti et al. 1984). Since then, it has moved north-eastward, until it reached its present position sometime in the early Second millennium BC. As a result, a wide arc of alluvial sediments, interlaced with old paleochannels, spans the southern Atbai from the east bank of the Atbara River across from Khashm el Girba to 100 kilometers north of Kassala. This particular area is filled with many large prehistoric settlements which were situated on lands well suited for rain based agriculture (Marks and Sadr 1988).

The southern Atbai is within a summer rainfall monsoonal zone and receives between 200 to 800 millimeters of rainfall per year. The environment of the southern Atbai is a direct result of the mean annual rainfall and is considerably more hospitable to plant and animal life than is the Northern Atbai, which is more arid. The southern Atbai is located in a low rainfall woodland savanna area (situated on a clay base) which contains several species of Acacia trees and three major grass genera, including Cymbopogon, Sorghum, and Hyparrhenia (Barbour 1964).

8

Background

Unlike the Gash, the Atbara River (which flanks the western border of the southern Atbai) has been downcutting and eroding the landscape. Since recent times, the Atbara has incised itself deeply into the land and near the town of Khashm el Girba there is a scoured strip of badlands, known as the karab, which flanks the river.

cultural boundary for quite some time between the Beja of the Atbai and other Sudanese of the Butana (ibid.). Location of the Malawiya/Butana Transition and Butana Group Sites The single Malawiya/Butana transition site and the seven Butana Group sites used in this study are located in the western sector of southern Atbai (see Figure 1.4). These sites (except for KG29S and KG29N which are located on the west bank of the Atbara River) are located in a 30 by 30 kilometer area beginning just east of the Atbara River across from Khashm el Girba. All of the sites are situated in close proximity to early Gash River paleo-channels or to the present channel of the Atbara River. When these sites were occupied, the Gash River was depositing considerable amounts of silt in the western sector of the southern Atbai.

People and Towns of the Southern Atbai Today, the southern Atbai is mainly inhabited by groups of nomads, agro­pastoralists, and farmers. The nomads consist of the Beja who are divided into a number of tribes and have inhabited much of the steppe and semi-desert areas of the greater Atbai region since Middle Kingdom times (Arkell 1973; Sadr 1991). The two main Beja tribes in the southern Atbai are the Hadendowa and the Beni Amer (Barbour 1964). The Hadendowa migrate in a east-west pattern between the Gash and Atbara, while the Beth Amer travel north-south and cross into Ethiopia during the winter months (ibid.). Some of the Beni Amer are more sedentary and practice a mixed economy where they herd livestock and keep small rain-fed plots around a water-saturated stretch of land called Sarab el Gash which lies about 35 kilometers south of Kassala (Sadr 1991).

Archaeological Work Carried Out in the Southern Atbai Archaeological work in the southern Atbai has been going on for a little over 70 years (Fattovich, Marks, and Mohammed-Ali 1984). The first archaeological investigations were carried out in the Kassala area by Crowfoot in 1917 when he discovered the significant early 3rd millennium BC site of Mahal Teglinos (Crowfoot 1928). From the 1930s to the 1950s (Kirwan 1930s, Sandison 1940s, and Delany 1950s), brief unpublished excavations and surveys were done by the Sudan Antiquities Service in Khartoum which touched upon some of the archaeology in the eastern Butana and southern Atbai (Fattovich, Marks, and Mohammed-Ali 1984). Reconnaissance work, stemming from the Nubian Aswan High Dam Project, was carried out briefly in the 1960s around the Khashm el Girba locality along the Atbara River by Shiner (1971). Shiner discovered several large mounded sites east of Khashm el Girba which he designated as “Butana” and was the first to report on the cultural material of the Butana Group. The tremendous size of the Butana sites, and the richness of the cultural material associated with them, was soon recognized as representing a potentially important formative prehistoric period which had never been reported before (Spratling 1970; Clark 1972; Phillipson 1977). Unfortunately, Shiner’s work was prematurely halted by political events in the Near East and the Sudan. However, his initial work around Khashm el Girba was responsible for stimulating more systematic archaeological investigations in the southern Atbai (Fattovich, Marks, and Mohammed-Alit 1984).

Throughout the Gash Delta and portions of the Sarab el Gash, there are full-time agriculturalists who plant mostly cotton and some millet. The Hadendowa descend upon the Gash Delta to graze their livestock and plant millet during the dry season (Barbour 1964). However, in recent years, more and more of the delta is sewn with cotton which has steadily diminished the traditional winter pasture lands of the Hadendowa (ibid.). The largest town in the southern Atbai is Kassala which lies near the Sudan/Ethiopian border on the east bank of the Gash, just south of the delta. The landscape immediately east of Kassala consists of large granitic jebels or mountains which rise abruptly out of the Atbai steppe. These jebels represent the beginning of the Ethiopian Highlands. Kassala was, and still is, a garrison town placed by the Sudanese to control the Eritrean frontier. Kassala is heavily mixed with Northern Sudanese, sedentary Beja (the Helenga especially), Arabians, Ethiopians, and some Europeans mostly Italians - from Ethiopia (ibid.). The second largest town in the project area is Khashm el Girba which is situated on the west bank of the Atbara River. Technically, Khashm el Girba is not in the southern Atbai but is located in the Butana region, which is a large grassland plain stretching from the west bank of the Atbara River to the Nile Valley. Khashm el Girba is populated by Shukriya and Lawawin Arabs and Western Sudanese (Sadr 1991). There are many Nubians there, as well, who were relocated from Wadi Haifa after the construction of the Aswan High Dam to New Haifa located about 40 kilometers north of Khashm el Girba along the Atbara River. There has never been very many Beja in Khashm el Girba or in the Butana. It is probable that the Atbara River acted as a

In 1981 and 1982, A. Marks, Y. Elamin, and A. MohammedAli (working on a joint project between Southern Methodist University and University of Khartoum) were able to expand upon Shiner’s preliminary findings and successfully completed two seasons of systematic survey and testing in the southern Atbai. During the same time, but on a separate project, R. Fattovich (working from the Istituto Univerisitario Orientale in Naples) began working at

9

The Butana Group Ceramics

Figure 2.1. The Sudan and surrounding area.

Mahal Teglinos and around the Kassala area. In 1983, the Butana Archaeological Project (Marks and MohammedAli) joined with the Italian Archaeological Mission in the Sudan (Fattovich) to produce a preliminary synthesis of the archaeology of the southern Atbai (Fattovich, Marks, and Mohammed-Ali 1984). So far, the Butana Archaeological Project and Italian Archaeological Mission have surveyed over 600 square kilometers and recorded 223 archaeological sites in the southern Atbai (Sadr 1991).

Atbai Ceramic Tradition), and the Atbai Ceramic Tradition (Fattovich, Marks, and Mohammed-Ali 1984: 177). Well within the 1st millennium AD, other ceramic bearing cultures, not related to the Atbai Ceramic Tradition, appear in the southern Atbai (Fattovich, Marks, Mohammed-Ali 1984: 182). Pre-Saroba The pre-Saroba represents two groups within the southern Atbai; one being located in the Khashm el Girba area, while the other is located in the Gash Delta (Fattovich, Marks, and Mohammed-Ali 1984: 178). The Khashm el Girba aspect of the pre-Saroba is represented by only one site (KG 14) and is situated near the Atbara River on the west bank (Marks 1987: 80). The material culture (chipped and ground stone,

The Prehistory of the Southern Atbai during the Holocene The late prehistory (see Figure 2.2) of the southern Atbai (beginning with cultures possessing ceramics) can be divided in two general periods, the pre-Saroba (i.e., pre-

10

Background

Figure 2.2. The Archaeological sequence in the Southern Atbai.

and ceramics) and the dating of this site (6215±75 bp) at first glance suggest that the occupation at KG 14 is related to the general Khartoum Mesolithic (ibid.: 82, 90), a 7th to 6th millennium BC horizon style which extended across a wide area of northern Africa from the Sahara to the central Nile and parts further south and east (Hays 1971, 1976).

However, KG14 does not contain the classic Wavy Line and Dotted-Wavy Line type of ceramics which are common with other Khartoum Mesolithic-like sites. Furthermore, this site contains ceramics “decorated with multiple bands of knobs, formed by pushing a stylus almost through the vessel wall from the interior surface” which are not found

11

The Butana Group Ceramics

on any Khartoum Mesolithic sites (Marks 1987: 90). Unlike Khartoum Mesolithic sites where quartz cobbles were used in the majority of cases for the manufacture of lithic tools, the vast majority chipped stone tools from KG14 were manufactured from chert and agate cobbles, although quartz cobbles are common in the Khashm el Girba area (Marks 1987: 90). Furthermore, there are a high number of blade tools within the lithic assemblage at KG14 which are uncommon at Khartoum Mesolithic sites (ibid.).

Mesolithic (Fattovich, Marks, and Mohammed-Ali 1984: 185). However, as it will be pointed out below, there are many aspects of this tradition which demonstrates that it developed as a separate phenomenon within the southern Atbai. At this time, it is not clear whether the Atbai Ceramic Tradition developed out of the pre-Saroba or not (Fattovich, Marks, and Mohammed-Ali 1984). The Atbai Ceramic Tradition can be defined by a series of three archaeological phases, containing five defined archaeological groups which existed from about the middle of the 5th millennium BC to the later half of the 1st millennium AD (see Figure 2.2). Essentially, the three archaeological phases (Saroba, Kassala, and Jebel Taka) divide the Atbai Ceramic Tradition into temporal parts which represent developmental stages based on different modes of adaptation (Fattovich, Marks, and Mohammed-Ali 1984; Sadr 1991). For example, during the Saroba Phase, cultures within the southern Atbai were based on a hunting and gathering strategy. During the Kassala Phase, this changes and cultures within the area begin to develop new adaptive strategies based more on the cultivation of plants and animal husbandry. In the Jebel Taka Phase, much of the previous practices in cultivation and animal husbandry is sustained, but some groups begin to rely more on a nomadic way of life, depending more upon a reciprocal relationship with other cultures within and outside the southern Atbai.

The faunal remains from KG14 indicate that both large mammals (such as waterbuck, elephant, African buffalo, etc.), small mammals (warthog, bushbuck, jackal, porcupine, etc.) and riverine animals (Nile perch, catfish, mollusk, and tortoise) were being exploited for food, thus indicating that the occupants of this site were extensively utilizing both the savanna and the riverine environment at this time (Marks 1987: 89). Overall, it appears that the climate was moist and warm during the occupation of KG14 (Wickens 1982; Marks 1987). The Khashm el Girba area would have consisted of a grass covered savanna interspersed with trees in steppe areas away from the river, while the river margins would have been covered with a gallery forest (Marks 1987: 89). During this same time, the area north of Kassala was represented by another series of sites which have been termed the Amm Adam Group (Fattovich and Piperno 1982). Unlike KG14, there are no grinding stones associated with any of these sites (Fattovich, Marks, and Mohammed-Ali 1984: 178). Much of ceramics associated with this group, however, also consist of the knobbed ware found at KG14 (Fattovich 1989: 481-485). Some Wavy Line ceramics and a bone harpoon have been recovered from some of these sites, as well (Fattovich 1989: 484). These particular artifacts are common among Khartoum Mesolithic sites along the central Nile and elsewhere, and their presence may prove that there was some connection between the two areas (ibid.).

Within each phase there are one or more archaeological groups which are represented by sites or components which share a particular set of ceramics. Theoretically, it is believed that each group more or less represents a particular culture or social unit of people who were integrated as a population or tribe (Sadr 1991). Each group is believed to have occupied a particular area of the southern Atbai for a certain period of time. For example, some groups were restricted to either the western or eastern part of the southern Atbai. Other groups were located within the interior areas of the southern Atbai between the Atbara and Gash Rivers. Throughout the three phases of the Atbai Ceramic Tradition, it is evident that some of the groups reoccupied territory which had been occupied by former groups. Thus, in the archaeological record, some sites contain several components represented by different groups.

Like KG14, the faunal assemblage associated with the Amm Adam Group suggests that the inhabitants were exploiting riverine resources, as well as hunting out on the steppe (Geraads 1983; Fattovich 1989: 484) Amm Adam Group sites are situated on an open alluvial plain not far from an old paleo-channel of the Gash River (Fattovich 1989: 484). The regional climate and environmental conditions seem to have been the same in this area as they were around Khashm el Girba (Fattovich, Marks, and Mohammed-Ali 1984: 178; Fattovich 1989: 486).

The Saroba Phase contains only the Malawiya Group. The Kassala Phase contains three groups; the Butana, Gash, and Mokram. The Jebel Taka Phase also contains a later aspect of the Mokram and another group called the Hagiz. Basically, all of the archaeological groups within the Atbai Ceramic Tradition can be linked to one another by a particular ceramic trait known as combing (Fattovich, Marks, and Mohammed-Ali 1984: 178). This particular combing trait was used both as secondary forming technique and as a decorative technique. Combing, used as a secondary forming technique, can be seen on many vessel interiors where a combed-like implement was applied on the vessel walls in wide sweeping horizontal bands to obliterate pottery coils and to thin vessel walls. Within the Saroba

The Atbai Ceramic Tradition Overall, the Atbai Ceramic Tradition can be defined “by a time transgressive complex of technological and stylistic components used in the production of ceramics” (Fattovich, Marks, and Mohammed-Ali 1984: 178). In some respects the origins of this tradition (at least through the ceramics), may be related to the much more widespread Khartoum

12

Background

Phase, combing was only used as a secondary forming technique. As a decorative technique, combing was present in both the Kassala and Jebel Taka Phases and consisted of straight, horizontal, multiple strokes done continuously across the entire exterior surface of the pot. In some instances this latter kind of combing was patterned; that is, the combing strokes were positioned in a way that created simple geometric patterns across the pot. Basically, the combing pattern which is ubiquitous on all of the ceramics within all of the groups above gives the Atbai Ceramic Tradition its cohesiveness as a continuous archaeological and cultural phenomenon within the southern Atbai.

Malawiya Group sites (Fattovich 1989: 484). Overall, Malawiya Group ceramics consist of buff colored, sandy paste vessels which were left unburnished and have a series of similarly impressed decorations (Fattovich, Marks, and Mohammed-Ali 1984; Mohammed-Ali and Jaeger 1989). The most common decoration used was a rocker stamped design, consisting of either dotted straight lines or dotted zig-zags (Fattovich, Marks, and Mohammed-Ali 1984: 178179; Mohammed-Ali and Jaeger 1989: 476). In addition to the ceramics, grinding stones seem to have been very common (Marks and Sadr 1988: 72). As at KG14, many of the lithic artifacts were made on blades (Shiner 1971: 317-332). The majority of chipped stone tools were manufactured from chert cobbles, but some agate and quartz cobbles were used, as well (ibid.).

The Saroba Phase The first phase of the Atbai Ceramic Tradition is known as the Saroba Phase and has been dated to about the mid part of the 5th millennium BC (Marks and Sadr 1988: 73). Only one archaeological group (the Malawiya) was affiliated with this phase. The Malawiya Group is represented by a number of sites located in the steppe area approximately midway between the present day Gash and Atbara Rivers (Fattovich, Marks and Mohammed-Ali 1984: 178). Sites affiliated with this group are usually small in size, averaging 5000 m2, and have shallow deposits (Fattovich, Marks, and MohammedAli 1984: 178). Given their location, it appears that peoples of the Malawiya Group of the Saroba Phase were more oriented in their subsistence activities towards the interior steppe of the southern Atbai than to the river margins where no sites of this kind have been found (Fattovich, Marks, and Mohammed-Ali 1984: 178; Marks and Sadr 1988: 74). Indeed, no remains of fish or other riparian fauna has been recovered at any of the Malawiya sites (Marks and Sadr 1988: 74). On the other hand, small to medium sized mammals (such as antelope, oribi, duiker, and dik-dik) adapted to a wooded savanna environment were found in significant numbers (ibid.).

The Kassala Phase The Kassala Phase appears to be an outgrowth of the Saroba Phase and starts at the beginning of the 5th millennium BC and ends sometimes during the 2nd millennium BC (Fattovich, Marks, and Mohammed-Ali 1984; Marks and Sadr 1988). This phase represents the most developed and culturally complex period within the Atbai Ceramic Tradition where a number of different adaptive strategies were carried out over a time span of some three thousand years (Fattovich, Marks, and Mohammed-Ali 1984: 179). Basically, three archaeological groups can be found within the Kassala Phase. The Butana Group seems to be the earliest of the three and sites are primarily restricted in the western half of the southern Atbai. The Gash Group appears to have developed out of the Butana Group and sites of this later group are principally located in the eastern half of the southern Atbai (Fattovich, Marks, and MohammedAli 1984; Fattovich 1989). The Mokram Group appears to be the latest within the Kassala Phase and its sites tend to be located more in the inland areas (in both the western and eastern halves), between the Atbara and Gash Rivers (Sadr 1991). The Mokram Group seems to have some ties with both the later aspects of the Butana and Gash Groups; however, it also contains a significant number cultural affinities to other groups outside the southern Atbai, namely the Pan Grave culture group of the Eastern Desert (ibid.).

Overall, the people of the Malawiya Group appear to have been mobile hunter-gatherers who more or less roamed the savanna areas of the southern Atbai. The distribution of Malawiya sites of the Saroba Phase suggests that there was a shift from more of a riverine adaptation during preSaroba times to an adaptation which was more oriented to the steppe (Fattovich, Marks, and Mohammed-Ali 1984: 178). The presence of Pila wereni (land snails) found in large quantities within the Malawiya Group sites suggest that the environment within the southern Atbai was still quite wet, and that for at least part of the year, these sites may have been situated in swampy areas (Fattovich, Marks, and Mohammed-Ali 1984: 178; Marks and Sadr 1988: 75).

The Butana Group of the Early Kassala Phase. The early Kassala Phase is demarcated by the Butana Group which is represented by ten archaeological sites located in the western half of the southern Atbai. The Butana Group is essentially characterized by large midden-mound sites, averaging 8 hectors in size (Fattovich, Marks, and Mohammed-Ali 1984: 180). Of the ten sites, five seem to have been major villages (ranging from 6 to 12 hectors), while the remaining five (except for one transitional site) seem to have been smaller sites with shallower deposits, some perhaps used as special activity areas (Shiner 1971; Marks and Sadr 1988; Sadr 1991). All of the village sites

The Malawiya Group may possibly be an outgrowth of the pre-Saroba (however, as stated above, their settlement patterns are very different), as the two share ceramic elements which seem to be similar to the Khartoum Mesolithic (Marks and Sadr 1988: 72). Indeed, knobbed decorated type of ceramics which are common in the Amm Adam Group, are found in small quantities on some

13

The Butana Group Ceramics

appear to have been located near one of the many Gash River paleo-channels which criss-crossed this particular portion of the southern Atbai (Sadr 1988: 96). However, the lack of naturally occurring Pila within the Butana Group sites suggests that none of these sites was seasonally inundated (Marks and Sadr 1988: 77-78). Indeed, the location of many of these large village sites along the old Gash River paleo-channels may have been due more to take advantage of the silt deposits (produced by the river) as opposed to being situated next to an active river channel (ibid.).

the earliest data for the domestication of sorghum and millet which has been securely dated to roughly 1500 BC in Mauritania (Munson 1976; Stemler 1984), and perhaps in Mali by 2500 BC (Manning et al. 2010). Other Butana Group artifacts consist mostly of lithic materials which are diverse. Overall, the lithic industry of the Butana Group is made up of small retouched and unretouched flakes produced from small pebble cores (M’Butu, personal communication). A major technique of core reduction was bi-polar, where cores are placed on an anvil and shattered with another stone (ibid.). Flake tools tend to be somewhat amorphous, although some well made crescents appear. Chipped stone picks are a common artifact on Butana Group sites, and they may have been used for breaking up soil for cultivation (Shiner 1971: 341; M’Butu, personal communication). In fact, many of these picks have their tips broken. Other Butana Group stone artifacts include an array of lip plugs and ear spools made of exotic stones not found in the southern Atbai. There are also a series of well made, ground mace heads and large amulet-like rings made from granite and porphyry which were probably quarried from outcrops located in the Red Sea Hills (M’Butu, personal communication). Ground stone probably used as milling tools for the grinding of food grains is also a common artifact found on Butana Group sites.

It is believed that the Butana Group evolved directly out of the Malawiya Group and that the very earliest Butana Group site technically can be considered as a transitional Malawiya/Butana (Saroba/Kassala Transition) manifestation (Fattovich, Marks, and Mohammed-Ali 1984: 179; Marks and Sadr 1988:77). This particular site (KG28B) contains both rocker-stamped decorated sherds (typical of the Malawiya Group), and sherds which were decorated with horizontal combed exterior markings. This horizontal exterior combed pattern is one of the stylistic hallmarks of the Butana Group ceramics. Sherds with exterior ripple decorations also are present. From the relatively small Malawiya/Butana transition site (however, KG28B is about twice the size of a Malawiya Group site) dated to approximately 4000 BC; one or more massive Butana Group settlements seem to have occurred around the mid part of the 4th millennium BC. These Butana Group sites probably represent the first sedentary village communities of the Kassala Period and it is presumed that the inhabitants of these villages were engaged in some form of plant cultivation (Marks and Sadr 1988: 77). Midden deposits from these Butana Group sites tend to be from 1 to 2 meters in depth and are laden with high volumes of ceramics and lithic debris (Fattovich, Marks, and Mohammed-Ali 1984: 180-181).

Faunal remains at Butana sites include wild species such as bovid, hippopotamus, and warthog in the lower levels of the largest sites, while domesticated species such as cattle, goat, and sheep have been recovered in small numbers from the upper levels (Peters 1986). The faunal data suggest that the Butana Group people were initially hunting wild species (primarily on the steppe), while later they continued hunting but acquired some domesticated animals, probably from either the Sahara or the Nile Valley. Cattle, goat, and sheep have been well documented at Esh Shaheinab and Kadero, both of which are located on the central Nile (associated with the Khartoum Neolithic), and it is possible that the domestic mammals found at the Butana sites may have come from this area (Haaland 1981,1986, 1987).

Many new ceramic types (which are not known from the Saroba Phase or earlier pre-Saroba occupations) are associated with these large Butana Group village sites. The two most distinctive types are represented by thin walled, red burnished topped jars with black to buff, unburnished chevron incised bodies and thick walled bowls with horizontally combed exteriors. Other typical Butana Group ceramics types appear, as well, including for the first time in all of the southern Atbai, vessels made with a paste which was tempered with seeds. These seeds may appear to be the first tangible signs of intensive utilization of plant resources native to the southern Atbai. At present, these seeds can be identified as varieties of both sorghum (Sorghum sp.) and millet (Echinochla sp., and Setaria sp.) (D’Andrea 1990). If at some later date these particular seedtempered ceramics prove to contain domesticated varieties of sorghum and/or millet, this would represent the earliest archaeological evidence for the domestication of native African plants anywhere on the continent. This would be significant in light of the current information concerning

Overall, the climate during the early part of the Butana Group occupation was still moist; however, it appears that Butana Group sites were located in the dryer areas of the southern Atbai (Marks and Sadr 1988: 77-78). From the beginning of the 3rd millennium BC on, it appears that the environment overall became increasingly more arid, while at the same time the Gash River was beginning to move out of the western sector of the southern Atbai towards its present position near Kassala (Fattovich, Marks, and Mohammed-Ali 1984: 184; Sadr 1988: 79-82; Marks and Sadr 1988: 76). Also, domesticated animals seem to have made their first appearance in the southern Atbai during this time (Marks and Sadr 1988: 78). The presence of domestic animals throughout the levels (from top to bottom) of some of the Butana Group sites

14

Background

suggests that these particular sites are relatively late (Sadr 1988: 95). Archaeologically, the later Butana Group sites are smaller and have shallower deposits (Marks and Sadr 1988; Sadr 1991). In addition, artifacts, especially the ceramics, associated with these later occupations become less diverse and are not as ornate (Fattovich, Marks, and MohammedAli 1984). Sometime during the 2nd millennium BC (or perhaps earlier) the Butana Group as a cultural phenomenon disappears in the western sector of the southern Atbai.

horizontally combed ceramics (many of which have finger impressed [“pinched”] rims) of the Gash Group seem to be later renditions of Butana combed ceramics. None of the characteristic red-topped ceramics of the Butana Group, however, seem to have survived into the Gash Group. Nonetheless, seed-tempered ceramics are found on Gash Group sites. There are also some outside stylistic elements which may link the some of the Gash Group ceramics with classic Kerma (Fattovich, Marks, and Mohammed-Ali 1984 185; Fattovich 1989: 493).

The Gash Group of the Middle Kassala Phase.

Lithic tools of the Gash Group seem to be technically different from the Butana (Marks 1988, personal communication). Also mace heads and stone picks are absent (Sadr 1988: 102). In addition, there are stone stelae and tombs at Mahal Teglinos which have never been found at any Butana sites (Fattovich 1986).

The next archaeological group in the Kassala Phase is the Gash Group. In all likelihood it appears that the Gash Group originally branched out of a later aspect of the Butana Group; however, no transitional sites have been located (Fattovich, Sadr, and Vitagliano 1988; Sadr 1991). Apparently, no Gash Group sites co-existed with any of the Butana Group sites (Fattovich 1988, personal communication). Gash Group sites are all located well to the east of Khashm el Girba and are situated near to Kassala and the Gash Delta. The Gash Group occurs during the middle part of the Kassala Phase and dates from roughly the mid part of the 3rd millennium BC to the first part of the 2nd millennium BC (Fattovich 1985, Fattovich and Vitagliano 1987; Fattovich, Sadr, Vitagliano 1988; Sadr 1988: 107). The Gash Group consists of large village sites (which were probably occupied year round), consisting of thick midden deposits (Fattovich 1985; Sadr 1988: 111). One site, Mahal Teglinos, may have been a center where high status individuals presided over a village community (Fattovich 1985; Fattovich, Sadr, and Vitagliano 1988). There are many smaller sites within the Gash Group, however, with little or no accumulation of midden deposits, suggesting that these sites were inhabited for only short periods, probably on a seasonal basis (Sadr 1988: 104, 111). Overall, the Gash Group may represent a series of loosely affiliated settlements of people organized into a stratified society (Fattovich 1985; Sadr 1988: 108). Economic divisions from village headman with material wealth (cattle, items of trade, etc.) to trader-middlemen and village agro‑pastoralists may account for several social groups which co-existed in the eastern portion of the southern Atbai. Trade and political interaction with Middle and Late Kingdom societies of Egypt also seem to have affected the local demography in the southern Atbai at this time (Fattovich and Vitagliano 1987; Sadr 1991). Indeed, it appears that some of the population associated with Mahal Teglinos were involved with trade between Egypt, Kerma, and areas to the south, exchanging such goods as electrum, ivory, ebony, and animal skins (Fattovich 1985). It has been proposed that the Kassala area of the southern Atbai was part of the land the Egyptians called Punt (Ibid.).

Overall, people living at some of the larger sites of the Gash Group were probably practicing a mixed economy much like the Butana Group before them and clear evidence of domestic plant cultivation (both Hordeum sp. and Ziziphus sp.) has been documented at Mahal Teglinos (Costantini et al. 1983). Based on faunal remains recovered from Gash Group sites (which consist mostly of domestic cattle, goat, and sheep, as well as wild bovid, gazelle, and dik-dik) the climate, as a whole, seems to have become more arid due to the fact that these animals were probably adapted to a drier grassland type of environment (Geraads 1983; Fattovich, Marks, and Mohammed-Ali 1984: 184). Precipitation levels in the southern Atbai and other parts of the Sudan almost would have reached the present day rainfall levels, suggesting that the environment was becoming very much like it is today (Wickens 1982). Also, by the end of the 3rd millennium BC, the Gash River was probably flowing in its present position (Sadr 1988: 82). As mentioned before, domestic cattle, goats, and sheep were well in place in the southern Atbai by the middle of the 3rd millennium BC, although not abundant, and this coupled with the drying trend may have facilitated the beginning of an agro-pastoral way of life for many people affiliated with the Gash Group (ibid.: 146). It is possible that many of the earlier large Gash villages may have been eastern extensions of later Butana villages which continued to follow the northeast drift of the Gash River (ibid.: 104). However, the ceramic data are not well known from these sites (now situated in central steppe area) and more work needs to be done before the gap between the end of the Butana and beginning of the Gash can be closed (ibid.). The Mokram Group of the Late Kassala Phase.

Gash Group ceramics are characterized by a preponderance of horizontally combed and brushed (wiped) ceramics of which the latter are decorated with a series of banded roulette, incised and/or punctated designs near the top portion of the vessels (Fattovich 1989: 490). The

The Mokram Group represents the third archaeological group in the Kassala Period. Basically, the Mokram Group existed during the Late Kassala Phase roughly from the mid 2nd millennium BC to the later part of the 1st millennium BC (Sadr 1988; 181). This group represents a rather

15

The Butana Group Ceramics

complex amalgam of local Kassala Phase cultural elements, and external characteristics from outside the southern Atbai (ibid.: 120-146).

is little archaeological data to support other alternative adaptive strategies such as hunting-gathering, indicating that these people were primarily engaged in the tending of livestock, mainly cattle, and who grew food crops as a dietary supplement. Sadr (1988: 120-146) has emphasized this particular mode of subsistence (agropastoralism, as opposed to a mixed economy) for the Mokram Group. Essentially, Sadr (1991) believes that a mixed economy is composed of divisions of labor (tending cattle and cultivation) which is divided among individual consanguineous units. On the other hand, an agro-pastoral economy is composed of divisions of labor that are divided among particular social units (Sadr 1991). Sadr’s distinction between mixed and agro-pastoral economies is based on an evolutionary model and he believes that agro-pastoralists evolved out of mixed economy societies (ibid.). Based on this scenario, Sadr believes that people of the Mokram Group were organized along the level of a chiefdom (Sadr 1988: 146).

The Mokram Group generally occurs in the interior steppe region of the southern Atbai, between the Atbara and Gash Rivers (Sadr 1988: 123; Marks and Sadr 1988: 80). Sites are numerous and usually characterized by thin deposits, suggesting probable seasonal occupations (Sadr 1988; 126-130). It appears, however, that some of these Mokram sites were more heavily occupied (with substantial mounds), some of which may have been permanent villages (ibid.: 128, 140). Some Mokram sites also tend to be part of a multi-component occupation overlying earlier Gash components (ibid.: 122). Among the lithic artifacts there are polished axes and stone bracelets made from fine porphyry (ibid.: 120). However, chipped and ground stone artifacts seem to be rather mundane. Ground stone appears on some of the Mokram sites, but is absent from others (ibid.: 132). The presence and absence of ground stone may indicate that peoples of the Mokram Group were engaged in different economic aspects, such as the tending of livestock without growing crops or receiving agricultural products from outside, at different times (ibid.).

The intrusive element of the Pan-Grave cultures within the Mokram Group may also indicate for the first time, since the pre-Saroba Phase, a major foreign cultural influence in the southern Atbai. The infusion of the Pan Grave culture was probably due to the political influences of the New Kingdom and Late Dynasties of Egypt (Bietak 1966; Trigger 1976; Adams 1977; Sadr 1991). Sadr (1991) has proposed that the Pan Grave elements of the Mokram Group were actually composed of mercenaries such as the Medjay (cf. Save-Soderbergh 1941; Bietak 1966; Trigger 1976; Adams 1977) who were allied to Egypt at various times, and who began moving south into areas such as the southern Atbai during the end of the 2nd millennium BC.

Like other Atbai Ceramic Tradition groups, the ceramics of the Mokram Group are the most definitive artifacts. Apparently, many of the Mokram Group ceramics have close stylistic affinities to the Pan-Grave culture of Nubia, indicating that outside influences, or actual people, were entering the southern Atbai from areas such as the central Nile or Eastern Desert during the latter part of the Kassala Phase (Sadr 1988: 120). Some of the common Mokram decorative ceramic elements include a net patterned crossincised configuration and a deep vertical grooved-incised pattern, as well as black-mouthed rim decorations on a red-slipped surface (ibid.). Nonetheless, there is still a significant proportion of horizontally combed ceramics in the Mokram Group, indicating that this particular group had some original ties to either the Butana or Gash Groups.

The Taka Phase The Taka Phase represents the last archaeological period of the Atbai Ceramic Tradition. The Taka Phase lasts from about the middle of the 1st millennium BC to the first half of the 1st millennium AD and encompasses both the Mokram and Hagiz archaeological Groups (Sadr 1991). The first part of the Taka Phase represents a continuation of the Mokram Group and is represented by the Late Mokram Group. The ceramics of the Late Mokram Group still include horizontally combed ceramics with a few new additions, such as fiber tempered pastes.

Faunal remains are rare on Mokram sites because of the thin cultural deposits (Marks and Sadr 1988: 80). When they do appear, however, they consist mostly of domesticated cattle, sheep, and goat (Peters 1986). Fish remains are rare (ibid.). Macrobotanical remains found on some of the Mokram Group sites include varieties of domesticated sorghum and millet (Costantini et al. 1983). Environmentally, the Mokram Group occupations appear to correlate with a drying trend within the southern Atbai and other parts of the Sudan (Wickens 1982; Fattovich, Marks, and MohammedAli 1984: 184). In fact, many of the Mokram sites are associated with artificial wells (Sadr 1983).

The later part of the Taka Phase is primarily dominated by the Hagiz Group. Hagiz Group ceramics are very different from Mokram ceramics (Sadr 1988: 147). Hagiz ceramics tend to be thick and are always heavily fiber tempered. For the most part, they are poorly decorated, usually being haphazardly combed on the interior and the exteriors (ibid.: 147). In addition, a few pre-Axumite sherds have been found on Hagiz Group sites (Fattovich, Marks, and Mohammed-Ali 1984). On the other hand, no pre-Axumite sherds have been found on any Late Mokram sites and this may suggest that the Hagiz Group might have occupied the southern Atbai later than the Mokram Group (Sadr 1988:

Based on the faunal and floral data, and the overall settlement pattern, the Mokram Group people appear to have been agro-pastoralists (Sadr 1988: 146). There

16

Background

Description of the Sites Used in the Ceramic Analysis

149). Nevertheless, no radiocarbon dates are available for either group, thus making their temporal relationship uncertain (ibid.: 149).

The ceramics used in this study were collected from one Malawiya/Butana Transition site; KG28B, and seven Butana Group sites; KG23, KG7, KG29S, KG1, KG96, KG5, and KG29N. These seven Butana Group sites are believed to represent village occupations and are known to have a representative assemblage of ceramics which date almost exclusively to the Butana Group time period.

By the time of the Hagiz Group, there seems to have been another economic change from agro-pastoralism to true nomadism (Sadr 1991). Sadr (1991) believes that the nomads of the Hagiz Group were cattle herders who traded with other groups, possibly people of the Late Mokram Group, to obtain grain products. Pastoralists who depended on agricultural products by trade or exchange is the economic distinction made by Sadr between nomads and agro-pastoralists (ibid.). Archaeologically, the rare occurrence of grinding stones on Hagiz sites seems to support this (ibid.). The only faunal remains recovered from Hagiz sites are the remains of domestic bovids (Fattovich, Marks, and Mohammed-Ali 1984). In contradistinction to the Hagiz sites, there are numbers of ground stones on Late Mokram sites, but no faunal remains (ibid.: 154). Thus, it is possible that people associated with the Late Mokram Group were more intensively engaged with the cultivation of plants, as opposed to the herding of cattle. As a result, it is possible that there may have been a symbiotic relationship between the Hagiz and Late Mokram Groups, if the two groups were contemporary (which may not be the case, ibid.: 168), where the former were nomads and who traded with the later group for agricultural products (ibid.: 147-169).

Other known Butana Group sites in the southern Atbai such as KG50, KG56, KG8, and N 125 were not included in this ceramic study. The site of KG50 did not contain any ceramics and is believed to have been some kind of special purpose site (Sadr 1988: 100). KG56 did contain a small number of ceramics (ca. 250 sherds); however, the Butana Group component at this site appeared to have been heavily mixed with several later cultural manifestations (ibid.). KG8 consisted of a very light scatter of Butana material, including ceramics (less than 100 sherds), and there was no depth to this particular site. Site N125 appeared to have been almost exclusively covered with ceramics; however, this site appeared to have been another special purpose site (possibly a specialized ceramic production center, cf. Shiner 1971: 381) and a representative sample of Butana Group ceramics was not expected to be found there. Nevertheless, this particular site may have great research potential, especially after the Butana Group ceramics, as a whole, have been described and typed.

During the Taka Phase, it appears that the climate in the southern Atbai was more arid than it was in the preceding Kassala Phase and rainfall patterns within this area and with other parts to the Sudan more or less reached their present levels (Wickens 1982; Fattovich, Marks, and MohammedAli 1984: 184; Sadr 1988: 153).

Sites KG28B, KG23, KG7, KG1, KG96, and KG5 are located in the western sector of the southern Atbai, while sites KG29S and KG29N are located just across the river on the west bank of Atbara in the Butana region. All of these sites fall within a 25 km radius of one another (see Figure 1.4). Overall, these sites tend to be mounded (with the exception of KG28B) and are characterized by a rich and diverse assemblage of lithics (both chipped, ground, and polished) ceramics, and faunal remains.

Post Atbai Ceramic Tradition Sites Within the southern Atbai, there are other archaeological sites (primarily consisting of lightly occupied camp sites) which appear not to share any of the typical ceramic characteristics attributed to the Atbai Ceramic Tradition (Fattovich, Marks, and Mohammed-Ali 1984: 182). Most of these sites can be divided into two archaeological groups; the Korak and Gergaf Groups (ibid.). Both groups appear to post-date Late Mokram and Hagiz Groups of the Taka Phase. The Korak Group is characterized by chaff tempered pottery which has been covered with a red slip. Gergaf ceramics are similar to the Korak Group but have cross-hatched and incised decorations just below their rims (ibid.). It is possible that the Gergaf sites date sometime prior to the 15 Century AD, however, a few pieces of glass on some of these sites suggest that they may be more recent (ibid.). It has also been pointed out that the spatial pattern of Gergaf sites matches closely with the distribution of the Hadendowa Beja who presently occupy the southern Atbai (ibid.).

KG28B KG28B is the only known Malawiya/Butana Transition occupation and a single radiocarbon sample from this site dates to 4018 BC±67 years (SMU 1193). It is located near an old paleo-channel of the Gash River in the interior steppe area of the southern Atbai, approximately 22 km east of the Atbara River. The site is represented by a dense scatter of lithics, ceramics, and faunal remains which covers an area approximately 15,000 m2. Fauna are varied and include a number of mammals, birds, reptile and fish (Peters 1986). The majority mammals on KG28B were antelopes, however, suids, and elephants were also identified (ibid.). No domestic mammals were recovered. Catfish, ostrich, guinea fowl, land snails, and snakes were also found. In addition, a considerable number of Pila wareni were also recovered.

17

The Butana Group Ceramics

KG7

Upon close examination, the configuration of the cultural deposit at KG28B is more or less ringed (about 5 meters wide) with a central area devoid of artifacts. Overall, KG28B is believed to be heavily deflated and the deposit appears to be homogeneous with no visible strata or lensing. Over one part of the site, an area of 46 m2 was tested which yielded artifacts down to a depth of 30 cm. Units were divided into various 1x1 m squares and excavated in arbitrary levels.

KG7 is the second largest Butana Group site and consists of another large undulating mound (parts of which may have been deflated) approximately 88,000 m 2 (see Figure 2.4). This site is located on the east bank of the Atbara River and is about 20 km southwest of KG23. Like KG23, KG7 is situated over a large expanse of alluvial sediments, along a particularly wide stretch of the Atbara floodplain, which would have been ideal for cultivation (Sadr 1988: 99). The midden deposit at KG7 is also similar to that of KG23 where the matrix tends to be ashy (M’Butu 1991). Unlike KG23, however, the midden at KG7 is interlaced with gravel deposits which appear to have originated from the Atbara River, during times of floods (ibid.). The midden deposits range in thickness from 80 to 150 cm.

The sherds on this site, on average, are larger than sherds from the other seven Butana Group village sites, and appear to be more uniform in size and style. Ground stone is present, but does not appear in any great quantity. Overall, the collective cultural assemblage seems to be simpler and less diversified than assemblages found on the seven Butana Group sites.

Two areas, totaling 5 m2 of excavations (area a consisting of 4 m2, & area b consisting of 1 m2) were tested at KG7. The cultural remains recovered from these excavations demonstrated that KG7 was similar to KG23 and contained an equally diverse amount of ceramics, lithics, and faunal material (including domestic mammals). Two radiocarbon samples were recovered from area a, both at 80 cm below the surface and produced dates of 3351 BC±68 (SMU 1151) and 3163 BC±93 (SMU 1151).

KG23 KG23 is situated on an old Gash River paleo-channel approximately 19 km east of the Atbara River and just about 3 km southwest of KG28B. This site overlooks a wide alluvial plain of rich, loose sediments which would have been well suited for the cultivation of plants (Sadr 1988: 88, 99). KG23 is the largest known Butana Group site, and is represented by a huge undulating mound approximately 120, 000 m 2 in size (see Figure 2.3). The mound consists of an “ashy powdery” midden deposit (some of which is deflated), which ranges between 150 and 200 cm in depth (M’Butu 1991). Two areas (area a & c) were tested, totaling 8 m2. The excavations encountered no features or house remains; however, a significant amount of daub was recovered (ibid.). In all, the midden deposits which were tested at KG23 yielded a spectacular amount of artifacts, including an array of ground stone mace heads, numerous stone picks, small crescents, other specialized lithic tools, and a wide spectrum of ceramics, ranging from plain to highly decorated wares. The fauna recovered include a number of wild and domestic animals (Peters 1986). Wild mammals include gazelle, small to large antelope, small rodents, medium viverrids, aardvark, gerbil, honey badger, and some carnivores such as leopard and possibly jackal (ibid.). Domestic mammals found on KG23 include ovicaprids and cattle. Catfish, African lungfish, land snails, toads, frogs, birds, and Pila wereni were also found.

KG29S and KG29N KG29S and KG29N are located at the northern edge of the town of Khashm el Girba, approximately 500 m west of the Atbara River in the Karab. Technically, these sites are in the Butana region and are the only known Butana Group sites to be situated on the west side of the Atbara River. KG29S and KG29N are down stream and about 10 km to the west of KG7. At first glance, these two sites (which are separate but adjacent to one another) appear to be the northern and southern part of a larger site which is represented by two separate mounds covering an area approximately 25,000 m2. Nevertheless, there is enough separation between the mounds of KG29S and KG29N (which are now separated by a road) to suggest that they were not part of a single site occupation. The ceramic data from these two sites also suggest that the they were not occupied at the same time. Both sites are located in a strategic position which overlooks the Atbara Karab, and during their respective occupations, they were positioned at one of the few crossing points along the Atbara River where traffic could come and go between the Butana and the southern Atbai (Sadr 1988: 100-101).

There are several reliable radiocarbon dates from KG23. The earliest date (from area c) was 3544 BC±154 years (SMU 1201) and was collected 135 cm below the surface. The second earliest radiocarbon date, (also from area c) was 3283 BC±67 years (SMU 1188) and was from 115 cm below the surface. Another radiocarbon date (from area A) was recorded at 90 cm below the surface at 3319 BC±253 years (SMU 1155).

Only a section of KG29S’ mound is still intact (most was destroyed by recent building activities) and it consists of midden deposits (some of which has been deflated) with an average depth of 80 cm. Much of the extant mound at KG29S has been severely disturbed by burials dug during historic times. The site was tested with eight contiguous 1 m2 units which were taken down in 5 cm arbitrary levels. Like KG23 and KG7, the midden deposits at KG29S

18

Background

Figure 2.3. The site of KG23.

revealed a rich and diverse array of chipped stone and ceramics, but did not yield may faunal remains. It is also important to note that very few grinding stones (common at the other Butana Group sites) were recovered at this site. No radiocarbon samples were recovered.

(however these could be natural alluvial sediments similar to those encountered at KG7a) where the deposits appeared to represent a single cultural episode (M’ Butu 1991). The depth of this deposit varied, but on average it was 60 cm deep. The deposits at KG29N also contained an impressive amount of lithics, including numerous ground stones and ceramics, as well as a large collection of faunal material, in contrast to KG29S. Many of the faunal remains (Peters 1986) consist of domesticated species, such as goat (Capra

The mound at KG29N has also been badly disturbed by recent burial activities and deflation. Six square meters were excavated using 5 cm levels over a possible gravel floor

19

The Butana Group Ceramics

Figure 2.4. The site of KG7.

aegagrus f. hircus) and cattle (Bos primigenius f. taurus). No radiocarbon specimens were recovered.

area (area a) and the cultural deposits averaged 60 cm in depth. Overall, the yield of artifactual material was typical of the Butana Group village sites and a large number of lithics, ceramics, and faunal remains were recovered. Like KG29N, there are many remains of domesticated animals (Sadr 1988: 100). On the other hand, the amount of fine decorated wares, which are plentiful at sites such as KG23, KG7, and KG29S (but not plentiful at KG29N) were found in smaller quantities at KG96.

KG96 KG96 is another large mounded Butana Group site of approximately 70,000 m 2. KG96 is located on the steppe a little more than 5 km southeast of KG23 and is 16 km east of the Atbara River. During the time this site was occupied, it is probable that it was situated along an old paleo-channel of the Gash River.

One radiocarbon sample was recovered at this site at the bottom of the midden and was dated to 984 BC±107 years (SMU 1187). However, this particular sample appears to

A total of 4 m2, dug in 5 cm levels, were excavated in one

20

Background

KG1

be too recent and may actually be associated with a later Mokram Group component which is mainly situated on an adjacent part of the site.

KG1 is the smallest of the seven Butana Group village sites and is located on the east bank of the Atbara River in the floodplain about six kilometers north of KG7. KG1 consists of a cluster of four low mounded areas encompassing a zone approximately 12,000 m2. The mound areas appear to represent midden deposits, some of which have been deflated and disturbed by recent human and animal activity (M’Butu 1991).

KG5 KG5 is the northern most Butana Group site and is located near another early Gash River paleo-channel about 25 km north-northeast of KG23. Like KG96, KG5 is presently situated in the steppe. KG5 consists of a midden mound deposit approximately 45,000 m 2 and is 50 cm deep. Like KG96 and KG29N, the midden at KG5 tends to be shallow relative to the midden deposits at KG23, KG7, and KG29S.

Two square meters were excavated (area a) in 5 cm levels at this site down to a depth of 40 cm. Like the other Butana Group sites, KG1 yielded a typical array of ceramics, Ethics, and faunal remains. No radiocarbon samples were recovered.

A total of 8 m2 were excavated (area a) in 5 cm levels. Like the other Butana Group sites the amount of cultural material recovered at KG5 was impressive; however, like KG96, the ceramics appeared to be less decorative, overall. No radiocarbon samples were recovered.

21

Chapter 3 Methodology

Introduction

method is conceptualized as representing whole vessels. For example, sherds which are different from one another (such as having dissimilar decorations) but are associated with the same vessel are lumped together under a single ceramic type or variety.

The essential purpose of this study was to develop a preliminary ceramic typology derived from eight sites which are believed to be culturally affiliated. The goal behind the formation of this typology is to demonstrate that the ceramics recovered from these eight sites more or less represent a single cultural entity known as the Butana Group. This chapter will disclose the methodological theories and procedures used in typing the ceramics used in this study.

The ultimate goal in using the Type-Variety method is to establish types or varieties which demonstrate some kind of temporal or spatial significance relative to other types or varieties. Types and varieties which are sensitive to either time or space in most cases will be defined by attributes which are basically stylistic and not functional in nature. Stylistic attributes tend to be related to the manner of how a vessel was finished and ultimately decorated as opposed to how it was formed or for what use it was intended.

It must be pointed out that the sherds used in this typology represent only a small fraction of the total number of sherds which are known to exist at each of the eight sites used. Nevertheless, it is felt that for the purpose of this study, the overall sample size (greater than 20,000 sherds) and context of the sherds is representative of the total sherd population from each site. This assumption is based first on the fact that surface collections (not used for the analysis) from all of the sites used in this study were examined and cross­checked for additional variability which may have been overlooked in the sample of sherds used in the typology. No additional variability was noted in the surface collections. Second, all of the sherds used in the typology were recovered from secondary deposits (midden accumulations), generally devoid of features (with the possible exception of KG29N, which had some evidence for a floor). In general, secondary deposits such as middens will contain representative assemblages of artifacts such as ceramics and lithics. To support the contention that the deposits were basically secondary among the eight sites used in this study, is the fact that less than 1 percent of the sherds could be conjoined.

The Type-Variety method is a product of ceramic research in North and Meso-America and its conception and development essentially can be traced through three landmark articles published in American Antiquity between the years 1958 and 1960 (Wheat, Gifford, and Wasley 1958; Phillips 1958; Smith, Willey, and Gifford 1960). Initially, Wheat and others devised a flexible typological scheme where ceramic types could be lumped or split apart into a series of types and varieties. Prior to the inception of the Type-Variety method, ceramic typologies consisted of collections of immutable types which were restricted to a specific geographic region at a particular time. These early ceramic typologies were originally devised in the Southwestern United States (Kidder 1927; Colton and Hargrave 1937, Colton 1953) and were based on a binomial classificatory system (via Linnaeus 1759, cf. Colton 1953: 52), established by Gladwin and Gladwin in 1930. For example, a Pueblo III type of black on white pottery found at Mesa Verde was classified as Mesa Verde Black on White, or a Basket Maker III black on gray pottery found in the Anasazi area of Lino was classified as Lino Black on Gray. This binomial ceramic system worked well in the southwest and archaeologists built good local ceramic sequences based on a number of well defined ceramic types. The number of ceramic types being established in the Southwest, however, escalated at an alarming rate, and it was realized that a new classificatory system was needed to organize the many thousands of types into a more comprehensible and flexible scheme.

Theoretical Considerations Used in Creating the Ceramic Typology: The Type-Variety Method The basic system used in creating the ceramic typology in this study is modeled on the Type-Variety method (Wheat, Gifford, Wasley 1958; Phillips 1958, 1970; Smith, Willey, and Gifford 1960; Sabloff and Smith 1969; 1970; Sabloff 1975; Robertson 1980). The Type-Variety method is essentially based on a holistic approach where the analyst attempts to sort all of the sherds from a particular assemblage into various types and/or varieties. Each ceramic type or variety defined within the Type-Variety

22

Methodology

Wheat, Gifford, and Wasley came up with a solution which organized related types in what they called Type Clusters (Wheat, Gifford, and Wasley 1958). The concept of the Type-Cluster was based on organizing a number of types into one group (Wheat, Gifford, and Wasley 1958: 38-39). Essentially, an unspecified number of similar types (many of them being identical, but defined and established by different archaeologists working in adjacent areas) could be lumped together under the rubric of the Type-Cluster. Within the Type-Cluster, the type with the greatest seniority would be designated as the “established” type, while the remaining types would become varieties (Wheat, Gifford, and Wasley 1958: 38). This simple but fundamental scheme offered archaeologists a way to organize the thousands of known types into a workable typology of perhaps several hundred types with their varietal subsidiaries. Furthermore, ceramic typologists had the option of either spitting or lumping ceramic varieties into any combination of TypeClusters. For example, under the Type-Cluster system if a variety had been designated into one Type-Cluster based upon a particular temporal association, and new archaeological data confirmed that this particular variety was actually occurring at a later time outside the normal temporal range of the type-cluster, then it could be easily designated as a new “established” type for a new TypeCluster. A classic example of this particular problem solved under the Type-Cluster method concerned a newly discovered ceramic variety closely affiliated with a another established type called Fourmile Polychrome. Under the Type-Cluster concept, this new variety was named Fourmile Polychrome: Point of Pines Variety (Wendorf 1950: 43-59). After a few years it was discovered that this particular variety was indeed associated exclusively with the Point of Pines phase which was temporally later than the archaeological phase which contained Fourmile Polychrome ceramics. As a result, the Point of Pines Variety was elevated to a type status called Point of Pines (Wasley 1952; Morris 1957; Wheat, Gifford, and Wasley 1958: 38). Under the Type-Cluster concept, this particular type was labeled Point of Pines Polychrome: Point of Pines Variety.

method. An example of the Type-Variety method devised under Phillips would be the Coles Creek established variety Coles Creek Incised, var. Coles Creek and its affiliated varieties Coles Creek Incised, var. Blakely ; Coles Creek Incised, var. Campbellsville ; Coles Creek Incised, var. Chase ; Coles Creek Incised var. Ely ; Coles Creek Incised, var. Greenhouse (formerly described as a separate type by Ford – who never did accept the TypeVariety method (see Phillips 1970: 23-36); Coles Creek Incised var. Hardy ; Coles Creek Incised, var. Hunt ; Coles Creek Incised, var. Macedonia ; Coles Creek Incised, var. Mott ; Coles Creek Incised, var. Stoner ; and Coles Creek Incised, var. Wade (Phillips 1970: 7076). Under Phillips’ Type-Variety concept, it is important to note that all of these particular varieties of the Coles Creek type were mutually exclusive from one another based on one or more decorative attributes. In 1960, Smith, Willey, and Gifford (1960) applied the Type-Variety method to ceramics in Meso-America. In doing so, they further defined the concept of the ceramic type (essentially reiterating Colton 1953) and the TypeVariety method (Wheat, Gifford, and Wasley 1958; Phillips 1958). They also elaborated on how one establishes types and varieties in an area which had never been studied before. Smith, Willey, and Gifford postulated that within the Type-Variety method, ceramic “units” must be first established (Smith, Willey, and Gifford: 1960: 330). These ceramic units consist of ceramics which share one or more common attributes. “Attributes” according to Smith, Willey, and Gifford were definable characteristics which could be identified on each pot sherd within each ceramic unit (Smith, Willey, and Gifford 1960). In a similar vein, Phillips had earlier invoked his “rule of sortability” which postulated that “if sherds cannot be differentiated by the ordinary methods of hand sorting, they must be assigned to the same type” (Phillips 1958: 119). Thus, the ceramic unit, as defined by Smith, Willey, and Gifford (1960), must be “sortable” by one or more attributes. An example of what an attribute could be would be a particular exterior surface treatment, shape of pot, paste characteristic, etc. In most cases, the exterior surface became the primary locus where the principal attributes were found using the Type-Variety method. This is especially useful when examining and sorting large numbers of pot sherds, since the exterior (and sometimes the interior) surface tends to be the most applicable area on distinguishing typological characteristics, that are not dependent on the overall size, shape, or position of the sherd.

Shortly afterwards, Phillips modified the Type-Cluster concept by redefining it into what he called the TypeVariety System (Phillips 1958). Essentially, Phillips wanted to incorporate the Type-Cluster method to ceramic data recovered in the Southeastern United States. Like the Southwestern United States, a number of ceramic types had been established there (Ford 1935a; Ford and Griffin 1937; Ford and Willey 1939a &b). In order to incorporate the Type-Cluster into the Southeastern ceramic taxonomies, however, Phillips felt that it would be better to replace the “established type” with the term “established variety” (Phillips 1958: 118). As a consequence of this change, the Type-Cluster became what Phillips termed as the “type” (ibid.). Therefore, a number of ceramic varieties would be subsumed under one type name. Like the Type-Cluster method, varieties could be easily lumped or split among an unlimited number of types under the Type-Variety

Once the ceramic units had been established, provisional variety names could be designated for each unit. Afterward, each provisional variety was to be tested in a chronological context, (i.e., stratigraphic situation, or on a separate site basis) to see whether it had temporal significance. If any of the varieties were demonstrated to have chronological significance, then a variety name was assigned to those of the temporally significant ceramic units. Smith, Willey, and Gifford (1960) were careful to point out that all newly

23

The Butana Group Ceramics

Table 3.1. Hierarchical Relationship Between Exterior Surface Treatment and Decoration exterior surface treatment

decoration (secondary)

combed wiped smoothed burnished burnished

combed plain or incised (common) plain or rocker stamped (evenly serrated & plain edge) pattern burnished or incised (rare) plain, chatter marked, simple impressions, or rippled

established varieties were assigned arbitrary place names, followed by a descriptive adjective. This particular naming style was essentially taken after the binomial nomenclature system which had been commonly used for decades in both the southwestern and eastern United States (Kidder 1927, Colton and Hargrave 1937). If a number of varieties shared one or more attributes and were associated in the same temporal context, these varieties would be grouped under a particular type. As Phillips had previously pointed out, the “type” (synonymous with the type-cluster) was an arbitrary unit which contained one or more varieties. The type would be named after one of the varieties, notably after the “established” variety. In Meso-America, primarily at Uaxactun and Barton Ramie, an example of type-varieties named would be Aguacate Orange: Aguacate Variety and Aguacate Orange: Holha Variety (Smith, Willey, and Gifford 1960: 337). For application in the Sudan, the Type-Variety method was used on a group of Late Neolithic ceramics recovered from Shaqadud Cave (see Robertson 1991).

(once labeled by site number and provenience) were piled into one large mass on a table and then sorted. This gross table top sorting procedure was essential and allowed the analyst to see all possible variability within the ceramic assemblage from the start. In essence, grouping sherds into various descriptive categories on a large table top allows the analyst to make fluid, quick, impressionistic decisions based on visual and tactile criteria which facilitates the sorting process. Among the Malawiya/Butana transition and Butana Group ceramics 14 descriptive categories were recognized. Other than one descriptive category, which demarcated sherds having seed tempered pastes, the remaining 13 categories were differentiated by general exterior surface treatment and/or decoration. It is important to note that the descriptive categories were not sorted out along a strict hierarchical ordering; starting first with a particular exterior surface treatment and then subdividing these groups into decorated and undecorated categories. From the start, it was felt that creating descriptive categories along these lines would result in an unwieldy number of sherd groups which would not reflect meaningful typological distinctions. For example, all incised sherds were lumped under one descriptive category irrespective of whether the underlying surface was wiped or smoothed. This decision was based on the fact that there were very few incised sherds with smoothed surfaces and that all of the incised sherds (other than varia) represented a single decorative style. Other decorated sherds also corresponded very closely with a particular exterior surface treatment with few exceptions. These particular exceptions also represented a very small number of sherds which did not warrant further subdivisions. Nevertheless, Table 3.1 provides a basic hierarchical breakdown of the Malawiya/Butana transition and Butana Group ceramics based on primary exterior surface treatment and secondary decorations.

Methodological Procedures in Establishing a Typology for the Butana Group Ceramics A total of 24,501 ceramic sherds were used. All of these were recovered from one particular area within each site. There were 7622 sherds recovered from area c of KG23, 4804 from KG28B, 4025 from area a of KG7, 2861 from KG29S, 2288 from area a of KG96, 1298 from area a of KG5, 1024 from area a of KG1, and 579 from KG29N. The first step in establishing a ceramic typology for the Butana Group ceramics was to sort the sherds into various groups (or descriptive categories) based on exterior surface treatment. As stated by Smith, Willey, and Gifford (1960), descriptive classes of sherds are first differentiated within a particular assemblage. Once the descriptive classes (categories) have been sorted out, they are then tested for typological significance.

Table 3.2 gives a general breakdown of the 13 descriptive categories based on a particular exterior surface treatment and/or decoration. One should note that each descriptive category in table 3.2 is further broken down into 5 analytical components (after Caneva 1988: 80-83). The five analytical components are; 1) the technique used to finish or decorate a vessel, 2) the kind of implement used in finishing or decorating, 3) the particular finishing effect or decorative element, 4) the decorative motif (when applicable), and 5) the overall structure of the decoration (when applicable). Table 3.2 can be used in conjunction

Irrespective of sherd type (such as rims, bases, upper and lower body sherds) all sherds used in this analysis were sorted into descriptive categories. All descriptive categories were considered to be mutually exclusive in respects to one another, thus, a single sherd could belong to only one descriptive category. Each descriptive category was defined by a set of specific attributes which, in turn, defined a particular exterior surface treatment or decoration. It is important to note that all sherds from a particular site

24

Methodology

Table 3.2. Breakdown of Descriptive Categories By Analytical Components (adopted after Caneva 1988:83) no.

technique

implement

elements

motifs

structure

1. 2.

horizontal combed smoothed yielding objects

parallel channels undecorated

horizontal lines plain

continuous plain

3.

burnished

undecorated

plain

plain

4.

wiped

evenly serrated edge smoothed hard, smoothed nonyielding object yielding, bunched object

undecorated

plain

plain

5.

incised on wiped or occasionally smoothed surface

stylus

lines

6.

chatter marked or burnished surface

stylus

short lines

7.

red-slipped and burnished

hard, smoothed nonyielding object

red slip

8.

simple impressions

evenly serrated edge

short dotted lines

9. 10.

simple impressions rocker stamped

stylus evenly serrated edge

dots dots

series of parallel lines punctate packed zig-zag

11.

rocker stamped

plane edge

lines

curved zig-zag

12.

pattern burnished

stylus

lines

13.

rippled

evenly serrated edge

short lines

geometric registers of parallel vertical lines which have been burnished over

with both the explanations of the descriptive categories below and the more detailed type-variety descriptions in Appendix A.

registers of continuous herringbone patterns series of horizontal or vertical track marks red rim band

continuous or banded continuous or banded rim band rim band continuous banded banded or continuous rim band continuous or banded

problem in distinguishing these latter sherds did occur from time to time. In the vast majority of cases, however, burnished sherds could be easily sorted from smoothed sherds. It should be noted that two subdivisions were made within this particular descriptive category based on paste differences, of which all of the plain burnished sherds from KG28B were sorted into another group based on the coarseness of the paste.

The combed category (no. 1) consisted of all sherds which had combed exterior surfaces. Sherds belonging to this category were easily sorted out because of the distinctive combed pattern (usually horizontal and made by broad sweeping strokes across the exterior of the vessel) found on all of the exteriors. These sherds were also distinctive in that they were usually very thick and gritty to the feel.

The wiped category (no. 4) was the fourth and consisted of sherds which had wiped exterior surfaces. Sherds with wiped surfaces (signified by thin, shallow striations) were easy to distinguish and were never mistaken for sherds with combed surfaces (signified by broad, relatively deep striations).

A second descriptive category (no.2) consisted of sherds which were smoothed and left plain. Sherds belonging to this particular category, for the most part, were easily distinguishable. In all cases, sherds in this category had exterior surfaces which were evenly finished, non-glossy, and devoid of any significant striations or wipe marks. Two divisions were made within the smoothed category based on paste differences. Smoothed undecorated sherds from the site of KG28B had coarser pastes than other sherds from the remaining Butana Group sites.

The incised category (no. 5) consisted of sherds with patterned incised lines on their exterior surfaces. It is important to note that the incised category consisted of all sherds with incisions, irrespective of whether they were smoothed or wiped. As it will be discussed in greater detail in the type-variety descriptions below, 99 percent of the incised sherds (of which the majority had been wiped previous to incision) could be confidently group under a single type. The incised patterns consisted of either parallel lines, cross-lines, or zig-zag, chevron (herringbone-like) lines. During the initial sorting of these particular sherds it was discovered that all three patterns listed above were part of a single incised pattern consisting of chevron (herringbone) designs. These particular incised sherds could be easily distinguished as a group and were never

A third descriptive category (no. 3) consisted of sherds which had plain burnished exteriors. Like the smoothed category, these sherds were, for the most part, easy to distinguish. All sherds in this category had exterior surfaces which showed some degree of luster (from a dull sheen to high polish) or, in some cases, had a silky smooth finish. Some of these later sherds had exterior surfaces which were transitional between being burnished or smoothed. Some

25

The Butana Group Ceramics

mistaken for combed sherds. The latter group of sherds always displayed multiple groves which were produced by a multi-toothed comb, while the incised sherds displayed singular incised lines (repeated on the sides of the vessel) which were produced by a pointed, stylus-like implement.

either fell into a miscellaneous or eroded category. The miscellaneous category (labeled varia) consists of odd sherds which did not fit into a particular category described above. The eroded category includes those sherds which have exteriors too weathered to tell what the exterior surface treatment or decoration was originally.

Overall, the combed, smoothed plain, burnished plain, wiped, and incised descriptive categories made up approximately 90 percent of the total number of sherds used in the analysis. Following these main categories there were eight smaller descriptive categories, consisting of sherds with particular decorations on their exterior surfaces.

Once all of the sherds from a particular site were sorted into the various descriptive categories, all sherds which could be conjoined were glued together. The process of fitting sherds together was an important step in the preliminary stages of the ceramic analysis. Basically, if sections of vessels can be constructed by piecing together a number of sherds then this usually signifies that many of the sherds (especially those which could be fitted together) more than likely came from primary contexts. In addition, if there are a number of sherds which can be conjoined, and these fitting sherds come from both the upper and lower levels, then this will, in most cases, signify that the overall deposit in which the sherds came from is somewhat disturbed. Within this study, a very small number of sherds within each of the eight sites could be conjoined. The lack of fitting sherds within each of the eight site samples used signifies that the sherd bearing deposits were more or less secondary in nature. Furthermore, almost all of the sherds recovered from these sites were small, usually not exceeding more than 4 cm in diameter. This, of coarse, should be of no surprise given the fact that it was assumed from the start that the eight site deposits overall were the result of midden or refuse accumulations. Deposits of this nature would occur when refuse was collected from various localities (such as household units) and carried or swept into more common areas such as a midden or dumping area. Unlike whole or partial pots, small sherds of this sort, which cannot be conjoined, probably resulted from being left for some time around or within a household unit where they were broken down by trampling, etc. Then, after a period of time these particular sherds were collected along with other kinds of debris and transported to a more centralized dumping ground.

The first of these (no. 6) consists of sherds with distinctive shallow chatter-like markings on a burnished exterior surface. These sherds are grouped within the chattermarked descriptive category. Another descriptive category (the red-slipped category; no. 7) consists of sherds with red-slipped exteriors which are usually burnished. Another category (simple impressed; no. 8) consists of sherds which have small dentation marks (evenly serrated edge) on their exteriors. These small dentation marks appear to have been made with a fine toothed comb (smaller than the combs used for making the large striations on the combed exterior category). Another category (simple impressed-punctation; no. 9) consists of sherds which have punctation marks on their exteriors which are spaced at close intervals and are arranged in no apparent pattern. The punctation marks appear to have been made with a pointed stylus, but are never made with a finger nail. Another category (rocker stamped; no. 10) consists of sherds which have dentated (evenly serrated edge) rocker stamped impressions on their exteriors. These dentated rocker stamped impressions are several sizes larger than the dentated impressions (simple impressed) found on the other dentated sherd category discussed above. Another category (rocker stamped; no. 11) consists of sherds which have curved (plain edge), rocker stamped exteriors commonly referred to as wolf tooth. Another category (pattern-burnished; no.12) consists of sherds which had burnished lines (laid out in various linear and geometric patterns) on their exteriors. These particular pattern-burnished sherds are quite distinctive and can be easily sorted from plain burnished sherds. The last of the decorated category (ripple; no. 13) consists of sherds which have shallow ripple-like marks on their exteriors. These sherds also have burnished exteriors which more than likely helped produce the shallow ripple marks. Overall, these sherds are very distinctive and one can sort them from the rest of the sherds with little difficulty.

Another critical factor in the process of fitting sherds together is the possibility in joining together seemingly different groups of sherds which in reality belong to the same vessel. As stressed above, ceramic types and varieties defined within the Type-Variety method are meant to represent whole vessels and not groups of pot sherds. This latter fact can be appreciated by an example within this study where two very different sherd descriptive categories were merged into a larger descriptive category once it was discovered that several sherds from each of the two supposed different categories could be fitted together. As mentioned above, there was one descriptive category which consisted of incised (arranged in a chevron pattern) sherds, and another category which consisted of sherds with red slip. Within the incised sherd category, it was noticed that there were no rim sherds within this group. The lack of rims within the incised category signified that these sherds, in particular, were from the body portion of a particular kind

In addition to the decorated categories, there is another descriptive category which consists of sherds which have a very characteristic paste which is filled with charred, round seeds (approximately 2 mm in diameter). As mentioned above, this particular category is the only sherd category in this assemblage which is defined by attributes not specifically tied to a particular exterior surface treatment or decoration. The remaining sherds which were not sorted into the various descriptive categories listed above

26

Methodology

of vessel. After several false starts, it was discovered that a few of the red-slipped rims could be fitted with some of the incised body sherds. Once these particular sherds were fitted together, it was obvious that they were indeed part of the same vessel. Within other sites used in this study, there were still other sherds which had both red-slipped (upper parts) and incised (lower part) portions.

Table 3.3. Variable Number Designations and Corresponding Descriptions variable 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Once all of the sherds were joined which could be, all of the descriptive categories in a site sample were re-examined and checked to see if any other combination of sherds could be associated with a particular vessel or not. All of the sherds within each of the descriptive categories were examined overall to see if they, as a group, were internally consistent. Final resortings of sherds were made at this stage, and when it was felt that the descriptive sherd categories, in general, were more or less correct, sherds were tabulated by number within each descriptive category according to provenience. In this study, sherds were basically counted and recorded by level within each descriptive category per site. Square designates were not part of the provenience information because all squares within a given site were contiguous and the deposits were horizontal and homogeneous in nature. The process of sorting, fitting, and tallying of sherds was repeated for all of the eight sites used in this study. There was a data file on each site which contained rows designating the levels and columns designating the various descriptive categories. Using this kind of format, computer files were generated which recorded the raw counts of sherds by level and descriptive category.

description descriptive category level unit (1x1 m2) interior surface treatment/decoration rim profile rim modification, etc. lip profile lip decoration rim diameter rim orientation lip thickness mid point thickness bottom thickness length rim contour vessel form exterior surface color interior surface color paste color inclusions

any decoration found on the lip; and the ninth listed the diameter of the rim. Rim diameters were taken by matching the plain of intersection along the outer edge of the rim within a series of concentric diameters drawn to the nearest centimeter and were used to define orifice diameter. A tenth variable noted the angle of the rim sherd in relation to the plain of intersection along the outer edge of the rim to the nearest five degrees. Rim angles were used to define the general shape of vessels. For example, bowls with flaring walls were defined by rims which had rim angles greater than 100 degrees. Bowls with vertical walls were defined by rims which had rim angles between 100 and 80 degrees. Jars were defined by rims which had rim angles less than 80 degrees. The eleventh variable gave the measurement (thickness) of the rim across the crown (the juncture between lip and rim portion), while twelfth gave the measurement (thickness) of the mid portion of the rim (at a mid point between the top of the lip and distal edge of the rim sherd) and the thirteenth gave the measurement (thickness) of the bottom of the rim at the distal end of the rim sherd. The fourteenth variable gave the length of the rim sherd (from the tip of the lip to the distal edge of the rim sherd). The fifteenth variable described the overall curvature-shape of the rim in a side or profile view. For example, a curvature shape of the rim could be either simple (hemispherical) or inflectional (S-shaped). It is important to note that all of the rims with inflectional curves were very slight. The sixteenth variable described the extrapolated vessel form (contour) based on the curvature shape of the rim in profile. For example, most of the rims examined reflected simple contour vessels (based on the simple

After all of the sherds from each of the eight sites were recorded into the computer by level and descriptive category, all rims and bases were pulled for further examination. Rims and bases were examined within each descriptive category, in order to see what kinds of vessel shapes existed within each category. It was apparent from the start that there were only two kinds of bases; conical and round. Therefore, the number of both conical and round bases were noted for each of the descriptive categories. In contrast to the bases, the rim sherds were highly variable in many aspects and gave additional information on vessel shape and other pertinent characteristics. As a result, each rim sherd was individually recorded as a case (rowwise) on a separate computer sheet, consisting of twenty variables per case (see Table 3.3). In all, 1770 rims were individually recorded and logged (a copy of the raw data files and code book for the variables will be on hand at SMU Anthropology). The first three variables defined the descriptive category of the rim sherd, level where the rim sherd came, and unit (1x1 meter square) where the rim was found. Other variables were as follows: interior surface treatment, shape-profile of the rim (i.e. direct, thinned, expanding, etc.); the sixth listed any additional modification of the rim sherd such as mending holes, grinding of the lip, etc; the seventh, the shape-profile of the rim lip, the eighth,

27

The Butana Group Ceramics

Figure 3.1. Lip decoration.

Figure 3.2. Simple contour vessel shapes.

28

Methodology

Figure 3.3. Inflected contour vessel shapes.

Figure 3.4. Rim profiles.

Figure 3.5. Lip profiles.

29

The Butana Group Ceramics

curvature shapes of many of the rims examined). However, some rims represented inflected contour vessels (based on the inflected curvature shapes of the some of the rims examined). A smaller number of sherds reflected carinated vessels (based on abrupt angular-shapes of a few of the rims examined). Overall, variables five, seven, nine, ten, eleven, thirteen, fifteen, and sixteen were used to describe the various vessel shapes found within a given descriptive category. The seventeenth and eighteenth variable described the particular exterior and interior color of the rim sherd respectively. The nineteenth variable described the paste color and the twentieth variable described the nature of the nonplastic inclusions in the paste.

invest with the investigating archaeologist the ability to discover or confirm archaeological entities which have meaningful consequences in space and time. In our case, it is the confirmation of both the Malawiya/Butana transition and Butana Group (cf. Shiner 1971; Fattovich, Marks, and Mohammed-Ali 1984) as viable archaeological manifestations in the southern Atbai. The medium for this confirmation are the ceramics, and in this study, they have been defined through the Type-Variety method. In choosing definitions and topical explanations within the basis ceramic type-variety descriptions, the author of this study has used the basic descriptive type-variety format originally prescribed by Phillips (1970) and latter modified by Robertson (1980; 1991) and Winchell and Wootan-Ellis (1991). The various definitive and explanatory components in each type-variety description are used to define as clearly as possible those qualities deemed most characteristic of a particular type-variety using a minimum of technical aids. Suffice it to say, the ceramic descriptions given below are tailored to the non-aided faculties (i.e., sight, feel, etc.) of the analyst so he or she can use the typology in a field setting (cf. Robertson 1991: 124). Of course, in this study such technical aids as calipers, protractors, Wentworth scales, hand lenses, and binocular microscopes were used from time to time. Nevertheless, analysts using the ceramic descriptions provided below are not required to need such technical aids in making cross-cultural comparisons with ceramics provided in this typology.

Explanation of Various Components Within the Type Descriptions Once all of the above information was assembled for all of the 24, 501 sherds used, type descriptions were then compiled. Each type description describes the characteristics and qualities of each of the descriptive categories defined above. It should be noted that, for practical reasons, descriptive categories were converted into provisional type-varieties once they were entered officially into the type descriptions. This may appear at first to be premature, however, the decision to give descriptive categories typevariety designates was based on the realization that most of type-varieties discussed below have never been described before, nor are they noted as other types in any of the assemblages found in other parts of North Africa.

Before presenting the basic descriptions of the typevarieties, definitions and explanations of the various components, headings, and technical words used within the descriptions will be given below.

For all intents and purposes, the ceramic type-varieties (formerly descriptive categories) below can be formulated into what is believed to be several ceramic complexes (this will be explained in greater detail in Appendix A). A ceramic complex is defined as a discrete culture unit composed of like ceramics recovered from two or more components which are located within a particular locality within a given period of time. A ceramic complex can be describe and used in the context of an archaeological phase or sub-phase as defined by Willey and Phillips (1958: 2224). Basically the concept of the phase can be described as:

Each type description begins with the type designation (TYPE) which names the particular ceramic type-variety. The type designation is based on a binomial system. The first word is an arbitrary name and is followed by a word or series of words which describes the particular type. In this study, all type designations are italicized. For example, there is the type Butana Combed, or the type Dratmuira Chatter Marked.

a culture complex possessing traits sufficiently characteristic to distinguish it for purposes of preliminary archaeological classification, from earlier and later manifestations of the cultural development of which it formed a part, and from other contemporaneous complexes (Kidder, Jennings, and Shook 1946: 6 via Willey and Phillips 1958: 22).

After the type name is the variety designation (VARIETY) which signifies the specific variety to which a particular group of sherds (or pots) belongs. All but a few variety designations are the same as the type name in this study. This because all but one of the varieties are newly defined and are deemed as principal (established) type-varieties of the varieties are newly defined and are deemed as principal (established) type-varieties.

As it will be demonstrated later, many of the type-varieties discussed below demonstrate clear temporal-stratigraphic significance within the site deposits as well as between sites.

Variety examples are Butana Combed: var. Butana, Dratmuira Chatter Marked: var. Dratmuira, Brindetta Red-topped Incised: var. Brindetta. At this point, it would be redundant to refer to the ceramics as varieties (even though in theory they are actually varieties as opposed to types) because all but two of the ceramic varieties are principal varieties,

The rationale behind the establishment of ceramic types and varieties; and ultimately ceramic complexes, is to

30

Methodology

where their names are the same as the type names As a result, the ceramics are always referred by their type names and not by their variety names unless specified otherwise. In the future, if other similar type-varieties are found within the Khashm el Girba area or outside, investigators will have the option to give new variety name designations to new ceramic type-varieties which correspond to any or all of the following type-varieties described below. Once other varieties have been established for some or all of the types described below, it would become necessary to refer to ceramics by their varietal names, if investigators want to point out specific kinds of differences on the varietal level. On the other hand, one could refer to the type-varieties as types if one wanted to discuss the ceramics on the type level. In either case, the specific referral to the ceramics as being either types or varieties is based on what level of detail the investigator wishes to discuss.

following the first principal identifying attribute, however, must be present on at least the majority of sherds of a typevariety. In all cases (except for one), the rust principal identifying attribute denotes the particular exterior surface or decoration which characterizes a type-variety. Pastes and Inclusions (PASTES AND INCLUSIONS) define and describe the textural qualities of the ceramic fabric of the sherds affiliated with a type-variety. The definition and description of the pastes and inclusions includes the clays and nonplastic material which was used to produce the ceramics represented by the sherds of a typevariety. Paste is defined as both the clays and nonplastics used to make a workable plastic substance which can be formed into ceramic vessels. In the paste and inclusion descriptions, the paste is generally defined by its texture and tactile qualities. The general colors of the paste is also given.

The word “established” (ESTABLISHED) is more or less self evident and refers to when and where the typevarieties were initially defined. In this study, all ceramic type-varieties (except for the Saroba type) were defined by the author.

Inclusions refer specifically to the nonplastic materials either present in the paste or as added tempering agents. In almost all cases, the inclusions in the Butana Group ceramics are quartz sands (in relatively high quantities) which appear to have been natural sediments in the clays. The presence of natural silt and sand sized inclusions in the pastes suggests that the Butana Group potters probably extracted their clays from sources which had large amounts of naturally occurring silt to sand sized sediments. It is probable that the clays used in the production of Butana Group ceramics were from riverine or alluvial deposits associated with either the Atbara and Gash River systems.

Sample size (SAMPLE SIZE) refers to the total number of sherds used in the description. This number represents all of the sherds which were sorted into a particular typevariety. The number of rims associated with a particular type-variety is also given. In addition, the number of all sherds of a particular type-variety found at a particular site is listed, as well. The principal identifying attributes (PRINCIPAL IDENTIFYING ATTRIBUTES) are the key attributes which define a ceramic type-variety. In all cases, the first principal identifying attribute must occur on all sherds within the type-variety. In some cases the second or third principal identifying attribute will also occur on all sherds, but this is optional. All principal identifying attributes

Surface treatment and decoration (SURFACE TREATMENT AND DECORATION) refers to the manner in which vessels were finished. Each type-variety surface treatment and decoration description defines the exterior and interior surface treatment, as well as any additional decorations, of a type-variety.

31

Chapter 4 Definition of the Malawiya/Butana Transition and Butana Group Based on the Ceramics Types and Their Temporal Significance

Introduction

between these two archaeological manifestations in the southern Atbai.

In this chapter the Malawiya/Butana transition and the Butana Group will be defined by the specific ceramic types they contain. In addition, the temporal significance of each type in terms of both intra-site stratigraphy and across the eight sites used in this study will be presented. In the ceramic type descriptions (Appendix A), each will be individually discussed in greater detail.

Given that approximately 1000 years elapsed from the beginning of the Butana Group (ca. 3500 BC) to its end in the southern Atbai (ca. 2500), it is expected that many ceramic types associated with the Butana Group will show some temporal patterning. Indeed, it has already been stated that toward the later part of the Butana Group (after the introduction of domesticated animals), the frequency of decorated ceramics diminished overall (Fattovich, Marks, and Mohammed-Ali 1984: 182-183). As of yet, it is not clear what particular ceramic types may have been superseded by others within the Butana Group; however, it is suspected that the Butana Group can be essentially partitioned into an early and late period (i.e. complexes) based on the proportion, or presence-absence of certain ceramic types. On a regional level, partitioning the Butana Group ceramics in this manner can greatly enhance the accuracy of determining the relationship of this group with other ceramic-bearing cultures present in northeast Africa during the 4th and 3rd millennia BC. Within the southern Atbai, using the percentage or presence/absence of certain ceramic types can also aid in making finer temporal (as well as cultural or functional) distinctions among sites and/or occupations associated with the Butana Group.

As discussed in Chapter 3, one of the basic tenets in defining a ceramic type within the Type-Variety method is to demonstrate its unique position in either time or space. As it will be discussed in greater detail in Chapter 5, the ceramics associated with the Malawiya/Butana transition and Butana Group are regionally distinct within northeast Africa. Thus, the spatial significance of the ceramic types defined in this study can be established already and it is only a matter of determining which are associated with either the Malawiya/Butana transition or the Butana Group and demonstrating their regular occurrence. It is also important to see if the ceramic types display any kind of temporal patterning within either the Malawiya/ Butana transition or the Butana Group. Determining whether particular ceramic types are present or popular within the Malawiya/Butana transition or within a particular part of the Butana Group can elucidate the internal temporal dynamics of the early part of the Kassala Phase of the Atbai Tradition overall.

In order to understand the internal sequence of the Malawiya/Butana transition and Butana Group ceramics, the types need to be organized along a temporal continuum. This can be done by comparing the various types through the excavated deposits within each site, across from one site to the next, or using a combination of both stratigraphic and cross-site comparisons. In the best of all possible worlds, it is logical to follow that a reliable temporal arrangement of the ceramic types can be provided for the Malawiya/ Butana transition and the Butana Group since there are approximately 20,000 sherds recovered from eight sites from which radiocarbon dates span a period of roughly 1500 years. Conceivably, the sherd counts of each type from each of the eight site deposits can be arranged much like a series of trench profiles to show a sequence of strata, or in our

Indeed, one of the basic questions to be addressed is whether the Butana Group developed out of the earlier Malawiya Group. Essentially, the ceramic types associated with the site of KG28B are considered to represent the Malawiya/Butana transition (cf. Fattovich, Marks, and Mohammed-Ali 1984) and one can determine if any of these particular types survived into the later Butana Group ceramics associated with the other seven sites used in this study. Thus, demonstrating the continuity of one or more major ceramic types from the Malawiya/Butana transition through the Butana Group will prove the cultural connection

32

Definition of the Malawiya/Butana Transition and Butana Group

case, ceramic types from the earliest to latest time period. For example, the ceramic types present or most frequent in the lower most deposits associated with the earliest sites can be contrasted with those which occur in the upper most deposits at the same sites which, in turn, can be contrasted with the lower most deposits from latter sites, and so on. Of course, all of this is contingent upon the assumption that the frequency of sherds per type are truly representative of the temporal positioning of the types within the deposits and that the deposits themselves are undisturbed and represent a long enough duration to allow a frequency change from one type to the next.

Due to the degree of difficulty in juggling both intra-site and inter-site variability, most ceramics types in a ceramic analysis are either defined temporally in the context of a single site deposit or across a series of sites of different ages. Using either intra or inter-site data is analytically expedient, and, in most cases, is accurate and adequate enough to determine the temporal variability of ceramic types. For example, if a there is a deep, stratified deposit at one particular site, a good temporal sequence among a series of types can be established (cf. Marks et al. 1985; Robertson 1991). On the other hand, if the deposits are thin or disturbed among a number of sites which are in turn associated with a particular ceramic group or culture, one can arrange these particular sites in sequential order and determine the temporal nature of the ceramics types in that way (cf. Hassan and Matson 1989).

In reality, there are some serious caveats which, at best, limit the effectiveness of arranging a series of ceramic types along a temporal continuum involving both individual site deposits and across sites over time. The first caveat is the uncertainty in the process of larger portions of vessels being deposited alongside smaller sherds within a deposit at a particular locus (i.e. within an excavated area of several m 2 or so), even when it is a secondary deposit (such as a midden) where there is a greater chance of equal representation of ceramic types at a particular site over time. For example, it is probable that a whole or partially whole vessel of one type will at one point be deposited in a midden among other ceramics types which have been deposited as smaller sherds. In many cases, if the whole or partially whole vessel breaks down into many sherds, these particular sherds can be fitted together which, in turn, notifies the ceramicist that there is an over representation of a particular type within a deposit. As is often the case, however, many of the sherds weather to a point were they can no longer be fitted together and there is no way to differentiate these from the rest of the assemblage and, as a result, an over representation of a particular type goes undetected. In the real world, these kinds of problems can occur often, and can indeed askew the typological patterns of types through a deposit based on sherd frequencies. The typologist can only hope that the deposit is fairly free of differential vessel portion deposition (i.e. larger sections of pots vs. smaller sherds) but as it will be pointed out below, this factor must be taken into consideration.

Part of the difficulty as well as the challenge concerning the Malawiya/Butana transition and Butana Group ceramics, is that there are available data from a number of sites some of which have deep deposits presumed to be stratified. Since the ceramics have never been studied before in any detail, and that there is an opportunity to examine the overall occupation sequence of eight early Kassala Phase sites in the context of the newly defined ceramic types (which presumably would display some temporal patterning), it was determined that both the infra-site and inter-site variability of all of the ceramic types will be examined together. Thus, an attempt will be made to interdigitate the eight site deposits from early to late based on the temporal variability of the 14 defined ceramic types associated with the Malawiya/Butana transition and the Butana Group. As it will be pointed out in greater detail below, conventional graphic displays of the various Malawiya/Butana transition and Butana Group ceramic type frequencies (like those produced by Ford in the Lower Mississippi Valley, cf. Ford 1935a &b; 1938; 1949; 1952) from one site deposit to the next could not be successfully done due to a bewildering amount of circumstances which rendered such an arrangement meaningless. These particular circumstances hinge upon the two caveats of consistent and accurate typological representation by pot sherds, and overall site deposition integrity.

The second caveat is whether the deposit from any site is indeed stratified, allowing the process of superposition to take effect which in turn demarcates particular ceramic types in a temporal context. Again, a wary excavator should in most cases be able to detect a disturbed deposit, however, in the case of soft midden deposits, detecting disturbances caused by wind and/or water erosion or other agencies, especially hundreds of years prior to excavation, can go undetected even by the keenest of archaeologists. In terms of the deposits associated with the eight sites used in this ceramic study, they are all soft, homogeneous midden accumulations and it is probable that some are disturbed. As it will be pointed out below, the ceramic type relative frequencies can help in determining which deposits among the eight sites appear to be disturbed.

Nevertheless, from the analysis of one site deposit (KG23c), a basic temporal arrangement of many of the defined Butana Group ceramic types could be worked out. However, when the basic temporal arrangement of types from the one site was compared with other sites, serious contradictions and irregularities arose. These suggested three things; 1) that the potsherd frequencies of the types, per level, per site were not proportionately uniform (due to differential vessel portion deposition) despite the fact that as discrete categories the ceramic types were sound, 2) that many of the deposits were either of such a short duration that the temporal variability noted among the various types was artificial or that the deposits themselves were badly disturbed, and 3) that both differential sherd representation

33

The Butana Group Ceramics

and lack of good site deposits were affecting the overall arrangement of the ceramic types within and across the eight sites of this study. Finally, beyond these three factors, differences in site function and/or social behavior could also affect the overall outcome of ceramic type occurrences among the eight sites used in this study.

models are based on the premise that all or some of the artifacts used in the seriation will be time sensitive. Time sensitive artifacts tend to be those artifacts which are stylistic or stochastic in nature and in turn will demonstrate a relatively rapid incipience, florescence, and decline through time (Dunnell 1978). It is hypothesized, that many of the Malawiya/Butana transition and Butana Group ceramic types established in this study are time-sensitive since they are essentially defined by stylistic attributes.

Teasing out some kind of coherent, graphic sequence of the types through time in the context of the eight sites used in this study was extraordinarily difficult. Nevertheless, through the use of two essential statistical methods, a palatable solution was finally found. This was based on; 1) the use of confidence intervals (s.e. = square root of the sum total of the proportion times one minus the proportion divided by the total sample size) to place bounds about the relative frequencies of sherd types, which gave reliably correct results for all sample sizes, and 2) the application of a multivariate technique, called principal component analysis, which mathematically arranged the sites and the levels involved into the best possible sequential (i.e. seriation) arrangement based on the ceramic type relative frequencies. Essentially, a seriation is multidimensional scaling in one dimension and that multidimensional scaling is essentially Q-mode principal component analysis (Hietala, personal communication 1992).

Initial Problems in Setting Up A Seriation for the Butana Group Ceramics and Sites Before setting up a seriation for the Malawiya/Butana transition and Butana Group ceramics, a few initial observations must be made about the eight sites used in this study. First, only a fraction of the eight sites were excavated and the ceramic samples selected for the typology may not represent a total sample of all possible ceramics from a particular site. However, it is believed (based on an examination of surface collections and other excavated samples not used in this study) that the types, varieties, and varia defined in this typology do represent nearly all of the known variability of the Butana ceramics observed in the field and laboratory. In the prospect that some rare kinds of ceramics were not detected (most likely noted as varia) in this study, it should still be of no consequence in trying to senate the present ceramics types in the context of the eight sites used.

The Use of Seriation in the Establishment of Ceramic Types Over Time and its Use in Sequencing Levels or Sites Over Time It has been well demonstrated that seriation techniques in archaeology have a dual effect by establishing a reasonable chronological sequence among the units (sites, levels, etc.) used for a desired order, while at the same time, establishing the typological/temporal validity of the artifacts (ceramics, lithic tools, etc.) used in the seriation (Petrie 1901; 1920; 1921; Brainerd 1951; Ascher 1959; Meighen 1959; Dunnell 1970; Ford 1972; Johnson 1972; Marquardt 1978; Hassan and Matson 1989). In many cases, the chronological order of the units is known, while the temporal/typological ordering of the artifacts is unknown, or conversely, the typological pattern of the artifacts is known, while the chronological order of the units is unknown. As in this study, however, there are often times when neither the chronological order of the sites nor the temporal pattern of the types used in the seriation are known. Nevertheless, seriation studies of these sorts can also be done successfully (cf. Ford 1935b; 1938; 1949; 1952; 1969: 1972).

Second, it is evident that the occupations at the eight sites used in this study were not of the same duration (some being very short, while others were very long), nor did they follow one another in a neat, even, sequential order. For example, a site such as KG23c, which has a deep deposit, may have been continuously occupied for over 500 years, while a site with a much shallower deposit, such as KG1a may have been only occupied for a generation or so. Furthermore, it is probable that sites of a short duration such as KG1a, KG96a, and KG5a were in part, or were totally contemporary with a portion of a much longer occupation represented at such sites as KG23c. As a result, a direct sequential order of the Butana Group sites cannot be made by simply arranging sites from earliest to latest based on a seriation of ceramic types. This is further compounded by the prospect that many of the deposits were disturbed. To confront the problem of arranging the sites in a temporal sequence, a comparison will be made with the relative frequencies of each type within and across each site on an individual basis. Using a seriation model, the relative frequencies of a particular type at a particular site should fit within a general monotonic pattern when the relative frequencies of that same type are compared with two or more sites. For example, if the relative frequencies of a given type at a particular site with 4 levels has a pattern of 10, 20, 30, and 40 percent through the levels (from earliest to latest), and another site with two levels has a pattern of 20 and 30 percent (from earliest to latest) associated with

Essentially, all seriations are based on a monotonic or “battleship-shaped” curved pattern derived from the relative frequencies (proportions) of two or more artifacts used among the units desired for sequencing. A seriation model implies that a particular artifact type (if it is timesensitive) will have a monotonic pattern when its relative frequencies (i.e. proportions) are traced through time. For example, the relative frequencies of an artifact type which reveals its whole temporal history will have an incipience, a florescence, and a decline (Marquardt 1978). All seriation

34

Definition of the Malawiya/Butana Transition and Butana Group

the same type through its levels, then a relative temporal positioning of the two sites can be made based on the relative frequencies of that particular type. In using the relative frequencies given in the example, the site with two levels would fit within the middle levels of the site with four levels. This kind of matching, using like types among the levels at different sites, can be repeated until all of the relative frequencies of the types are arranged to their correct monotonic position across all of the sites. Theoretically, a correct monotonic positioning of all of the types within a site should occur in unison so as to position the site (through the levels) into the right chronological context relative to the other sites and so on. As an end result, the monotonic patterns of all of the types (as reflected through their relative frequencies) should pass through all of the eight sites, which are in turn arranged by the monotonic patterns of the types through time.

relative frequencies through the levels are irregular, erratic, or vague, and the application of a standard error statistic renders a truer, accurate assessment of a particular type’s monotonic behavior through the deposit of a particular site. It is important to point out that if a type does display a good monotonic pattern through the deposit of a particular site, this does not necessarily mean that this type is temporally sensitive, for other factors must be considered, as well. Conversely, a type which has been deemed as temporally significant must display a convincing monotonic pattern. In this study, the relative frequencies of each type are bracketed within a 2 standard error range within each of the site deposits. A relative frequency within a 2 standard error range means that there is a 95 percent confidence interval to the relative frequency. A single standard error would produce a 68 % confidence interval. For example, if there is a type which has a relative frequency numeric value of 50 percent and a 2 standard error range between 30 and 70 percent, this would mean that we feel 95 percent confident that the true numeric value of that particular relative frequency at that particular level (at a given site) would be somewhere between 30 and 70 percent. If a series of relative frequencies from a particular type has each a 20 percent standard error range (± 10% from the mean) and the relative frequencies themselves range from 30 to 70 percent through the levels at a particular site, then the pattern produced by the actual relative frequencies will be unreliable, based on the overall series of standard errors, even if the relative frequencies produce a monotonic pattern through the levels or not. Put another way, if the 2 standard error margin is wide enough to allow a vertical line to be placed from the lowest to uppermost levels without going beyond the standard error margin, the monotonic pattern produced by the relative frequencies of a particular types is not sound. On the other hand, if a 2 standard error margin (associated with a series of relative frequencies) is narrow enough not to allow a vertical line to pass through it ( from the lowest to uppermost level) and the overall series of relative frequencies produces a monotonic pattern, then the combination of relative frequencies and the 2 standard errors indicate that the given pattern produced through the levels is temporally significant - given that the site deposit is stratified.

By using this scenario, a site such as KG23 which is believed to have a long occupation and has a number of levels, can be used as a baseline site where the relative frequencies of its types can be compared with the relative frequencies of the same type from the other seven sites. In the final outcome, all seven sites can be arranged temporally in respect to the deposit at KG23c based on seriating the relative frequencies of all 14 types defined in this study. As previously pointed out, all of this of course is based on the assumption that; 1) one or more of the 14 ceramics types defined are temporally sensitive, 2) all eight sites have stratified deposits, and 3) all of the sherds from all of the types broke into proportionally representative pieces based on their popularity through time. Setting Up A Seriation for the Butana Group Ceramics and Site Deposits The first part of the seriation begins with comparing the relative frequencies of all of the types and varieties within the levels of each site deposit. Relative frequencies have been computed by dividing the sherd count of a particular type within each level by the total number of typed sherds within the same level. Varia and eroded sherds (which make up no more than 1 percent of the total sherd population) were not included in the sample used to generate relative frequencies.

As it will be demonstrated later, many of the contradicting or spurious relative frequencies associated with some of the sites can be statistically discounted as false or not significant due to the 2 standard error margin. In the long run, it will be demonstrated that many of the sites used in this study contain no reliable type relative frequencies through their deposits.

Based on a seriation model, it is expected that all or some of the ceramic type relative frequencies will display a whole or partial monotonic pattern through the levels of each site. In order to detect significant monotonic patterns (associated with a particular type’s relative frequencies through the levels at a particular site) from those which are random, the use of a standard error (s.e.) statistic has been implemented with each relative frequency value of each type within all sites. The standard error statistic is used for confidence interval interpretations from which a relative frequency value from a particular type can be evaluated based on the sherd sample size (Hietala, personal communication 1990). As is often the case, many of the

In order to establish a comparative unit of analysis from which the relative frequencies of the types can be derived, all site levels have been condensed into 10 cm increments (i.e. levels). Producing 10 cm levels also provides a larger sample size of ceramic types per level (due to the fact that all of the sites were dug in 5 cm levels) which

35

The Butana Group Ceramics

in turn reduces the 2 standard error value of each relative frequency. In addition, all varieties established in this study have been condensed back into their respective types due the very small sample size associated with the ceramic varieties. Once all of the relative frequencies for each type have been computed for each 10 cm level within each site, the types are then evaluated for probable monotonic patterns through the levels.

percent of sherd sample. With the exceptions of levels thirteen and fourteen, the relative frequencies of this type increase from 13 to 22 percent beginning at the lower levels up to level six. From levels six to one, the pattern is more erratic, however, the relative frequencies of this type generally decline from 25 to 15 percent. When the relative frequencies and their respective 2 standard errors are taken into account, the Socodad Wiped type does not display a convincing monotonic pattern. Nevertheless, if several levels are discounted (namely levels 13, 5, and 4), a pattern can be depicted for this type, suggesting that it had a florescence in the middle - upper levels (levels 4 through 7) at KG23c.

The Sites KG23c KG23c contains the largest collection of typed sherds (n = 7473) of all sites used in the study and has the deepest deposit (depth = 1.4 m) which was condensed into fourteen, 10 cm levels. As previously mentioned, the deposit at KG23c has the greatest potential for containing a long temporal span of Butana Group ceramics, and is the best candidate for providing a baseline series of type relative frequencies from which identical types to which the remaining seven sites can be compared. It is believed that the length of occupation at KG23c was considerable and that the other six Butana Group sites were more or less contemporary with some part of its occupation. Hence, once the ceramic type relative frequencies have been graphed at KG23c, other ceramic type relative frequencies at the other sites can be used to calibrate their relative position in relation to KG23c. The following frequencies and relative frequencies (within a 2 standard error margin) of the types through the 14 levels at KG23c is provided in Tables 4.1 and 4.2. Of the 11 types represented at this site, 5 major types constitute 92 percent of the typed sherd population.

After, the Socodad type, the most frequent type is Dratmuir Burnished (n = 1126) which constitutes 15 percent of the typed sherd sample at KG23c. The relative frequencies of this type stay more or less the same from level fourteen on up to level six, oscillating between 7 and 10 percent. From levels six to two, the relative frequencies increase to 25 percent and then drop off to 20 percent at level one. Overall, the Dratmuir Burnished type with its 2 standard errors, shows a good monotonic pattern through the levels at KG23c. This type, like the Bahri Smoothed type, tends to be most frequent in the upper levels. The last of the 5 major types at KG23c is Butana Combed (n = 1010), which constitutes 13 percent of the sherd sample. The relative frequencies of this type increase from 25 to 34 percent in levels fourteen to eleven, and then steadily decline from level eleven to 7 percent in level one. When the 2 standard errors are taken into consideration, the relative frequencies of this type may actually decrease from the earliest to the latest levels at KG23c. Nevertheless, this type displays a good monotonic pattern through the levels at KG23c.

The largest number of sherds is represented by the Brindetta Red-topped type (n =1750) which constitutes 23 percent of the sherd population at KG23c. On observing the relative frequencies of the Brindetta type through the levels, this type oscillates between 24 and 37 percent from levels fourteen through seven. From levels seven to one, the relative frequencies declines steadily to 12 percent. When the relative frequencies are bracketed within 2 standard errors, a reliable monotonic pattern is maintained for this type through the levels at KG23c.

The remaining six types (Alabaska Punctated, Khordhag Plain, Dratmuira Chatter Marked, Sirjino Simple Dentated, Zirhalla Pattern Burnished, and Sarb’ut Ripple Marked) represent a combined total of 8 percent of the sherd sample at KG23c.

The next most frequent type at KG23c is Bahri Smoothed (n = 1581). This particular type constitutes 21 percent of the sherd sample at KG23c. From levels fourteen to eight (with the exception of level 7), the relative frequencies of this type stay the same, fluctuating between 9 and 12 percent. However, from levels seven to one, there is a steady increase of this type’s relative frequencies from 13 to 35 percent. When the relative frequencies are bracketed between their respective 2 standard errors, a good monotonic pattern for the Bahri Smoothed type can be established through the levels at KG23c.

Of these types, Alabaska Punctated represents the largest sherd count ( n = 267) which constitutes 4 percent of the typed sample at KG23c. The relative frequencies of this type virtually remain at zero until level nine where they first appear at 1 percent and increase up 7 percent through level eight. At level seven, the relative proportions drop off slightly to 4 percent and remain the same through level one. When the 2 standard errors are taken into account, the Alabaska type displays a slight monotonic pattern regardless of the small number of sherds represented. In general, the monotonic pattern of the Alabaska Punctated type is similar (albeit, in much smaller proportions) to the pattern represented by the Brindetta Red-topped type.

The third most frequent type at KG23c is represented by the Socodad Wiped type (n = 1446) which constitutes 19

The Khordhag Plain type represents 2 percent of the sherd sample (n = 147) at KG23c. When the relative

36

Definition of the Malawiya/Butana Transition and Butana Group

Table 4.1. Frequencies of the Ceramic Types Through the Levels at KG23c level 1 2 3 4 5 6 7 8 9 10 11 12 13 14 total

BRtpped. freq. 154 125 64 86 152 184 209 164 222 162 43 68 37 80 1750

B. Smthd. freq. 465 242 151 103 187 117 76 45 56 57 15 32 12 23 1581

S. D. Wpd. Brnshd. freq. freq. 239 269 117 198 141 115 81 74 146 125 183 75 132 55 91 52 110 53 80 47 23 10 28 19 36 13 39 21 1446 1126

B. Cmbd. freq. 88 48 42 44 79 100 77 101 111 115 50 50 46 59 1010

A. Pncttd. freq. 43 29 20 21 29 34 23 39 19 6 1 3 0 0 267

K. Pln. freq. 18 13 12 8 13 9 7 2 17 12 3 10 10 13 147

D.C. Z. Mrkd. Brnshd. freq. freq. 17 3 18 3 7 2 4 4 10 5 8 3 7 5 5 5 4 3 5 2 0 0 0 0 0 0 0 0 85 35

S. Dnttd. freq. 5 8 4 0 2 3 1 0 1 0 0 0 0 0 24

S. Rp. Mrkd. freq. 0 0 0 0 0 0 0 0 0 0 0 1 0 1 2

total freq. 1301 801 558 425 748 716 592 504 596 486 145 211 154 236 7473

Table 4.2. Relative Frequencies and 2 Standard Error Margins of the Ceramic Types Through the Levels at KG23c B. R-tpped. level % (±2s.e.) 1 2 3 4 5 6 7 8 9 10 11 12 13 14

11.8 (±1.8) 15.6 (±2.6) 11.5 (±2.7) 20.2 (±3.9) 20.3 (±2.9) 25.7 (±3.3) 35.5 (±3.9) 32.5 (±4.2) 37.2 (±4.0) 33.3 (±4.3) 29.7 (±7.6) 32.2 (±6.4) 24.0 (±6.9) 33.9 (±6.2)

B. Smthd. % (±2s.e.)

S. Wpd. % (±2s.e.)

D. Brnshd. % (±2s.e.)

B. Cmbd. % (±2s.e.)

A. Pncttd. % (±2s.e.)

K. Pln. % (±2s.e.)

D.C. Mrkd. % (±2s.e.)

Z. Brnshd. % (±2s.e.)

S. Dnttd. % (±2s.e.)

35.7 (±2.7) 30.2 (±3.2) 27.1 (±3.8) 24.2 (±4.2) 25.0 (±3.2) 16.3 (±) 12.8 (±2.7) 8.9 (±2.5) 9.4 (±2.4) 11.7 (±2.9) 10.3 (±5.1) 15.2 (±4.9) 7.8 (±4.3) 9.7 (±3.9)

18.4 (±2.1) 14.6 (±2.5) 25.3 (±3.7) 19.1 (±3.8) 19.5 (±2.9) 25.6 (±3.3) 22.3 (±3.4) 18.1 (±3.4) 18.5 (±3.2) 16.5 (±3.4) 15.9 (±6.1) 13.3 (±4.7) 23.4 (±6.8) 16.5 (±4.8)

20.7 (±2.2) 24.7 (±3.0) 20.6 (±3.4) 17.4 (±3.7) 16.7 (±2.7) 10.5 (±2.3) 9.3 (±2.4) 10.3 (±2.7) 8.9 (±2.3) 9.7 (±2.7) 6.9 (±4.2) 9.0 (±3.9) 8.4 (±4.5) 8.9 (±3.7)

6.8 (±1.4) 6.0 (±1.7) 7.5 (±2.2) 10.4 (±3.0) 10.6 (±2.2) 14.0 (±2.6) 13.0 (±2.8) 20.0 (±3.6) 18.6 (±3.2) 23.7 (±3.9) 34.5 (±7.9) 23.7 (±5.9) 29.9 (±7.4) 25.0 (±5.6)

3.3 (±1.8) 3.6 (±2.6) 3.6 (±2.7) 4.9 (±3.9) 3.9 (±2.9) 4.7 (±3.3) 3.9 (±3.9) 7.7 (±4.2) 3.2 (±4.0) 1.2 (±4.3) 0.7 (±7.6) 1.4 (±6.4) 0.0

1.4 (±2.7) 1.6 (±3.2) 2.2 (±3.8) 1.9 (±4.2) 1.7 (±3.2) 1.3 (±2.8) 1.2 (±3.4) 0.4 (±2.5) 2.9 (±2.4) 2.5 (±2.9) 2.1 (±5.1) 4.7 (±4.9) 6.5 (±4.3) 5.5 (±3.9)

1.3 (±2.1) 2.2 (±2.5) 1.3 (±3.7) 0.9 (±3.8) 1.3 (±2.9) 1.1 (±3.3) 1.2 (±3.4) 1.0 (±3.4) 0.7 (±3.2) 1.0 (±3.4) 0.0

0.2 (±2.2) 0.4 (±3.0) 0.4 (±3.4) 0.9 (±3.7) 0.7 (±2.7) 0.4 (±2.3) 0.8 (±2.4) 1.0 (±2.7) 0.5 (±2.3) 0.4 (±2.7) 0.0

0.4 (±1.4) 1.0 (±1.7) 0.7 (±2.2) 0.0 (±) 0.3 (±2.2) 0.4 (±2.6) 0.2 (±2.8) 0.0

0.0

0.0

37

S. Rp. Mrkd. % (±2s.e.) 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

0.2 (±3.2) 0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.5 (±0.9) 0.0

0.0

0.0

0.0

0.0

0.4 (±0.8)

total % 100 100 100 100 100 100 100 100 100 100 100 100 100 100

The Butana Group Ceramics

frequencies are observed, they register between 5 and 6 percent in levels fourteen through twelve. From level one, the relative frequencies oscillate between 3 and one half of 1 percent. When the 2 standard errors are taken into account, the Khordhag Plain type does not display a convincing monotonic pattern through the levels at KG23c. Nevertheless, this type does appear through the entire deposit at KG23c.

A computer generated principal component statistical application (Statview Factor Analysis - Principal Component Method, cf. Schiffman, Reynolds, and Young 1981) can be used to senate the 14 levels at KG23c based on the monotonic patterning of the 5 major types discussed above. Essentially, the principle component application works by first generating a correlation co-efficient matrix table. Each co-efficient on the matrix table in this case represents the degree of closeness (correlation) of one level in respect to another based on the relative frequencies of the five major types across each level. For example, if two levels are close to one another (i.e. the corresponding relative frequencies of each of the five types across each level are similar) they will have a high co-efficient. Likewise, if two levels are distant from one another (i.e. the corresponding relative frequencies of each type across each level are different) they will have a low co-efficient. In the case of the 14 levels used from KG23c, 98 correlation co-efficients were calculated (14x14=196 and where 196÷2=98). The 98 correlation coefficients from the matrix table were then factored by a principal component method which in turn ranks the 14 levels from KG23c in a seriation (sequential) arrangement.

The remaining 4 types (Dratmuira Chatter Marked, Sirjino Simple Dentated, Zirhalla Pattern Burnished, and Sarb’ut Ripple) account for the last 2 percent of the sherds in the KG23c sample. All have relative frequencies too small to show any monotonic pattern. However, on a presenceabsence basis, some patterns do appear. Both the Dratmuira Chatter Marked (n = 85) and Zirhalla Pattern Burnished (n = 35) types do not appear until level ten at KG23c. Likewise, Sirjino Dentate (n = 24) does not appear until level nine. However, Sarb’ut Ripple Marked (n = 2) only appears in levels 12 and 14. When the relative frequencies of the eleven types are assessed through the levels at KG23c, it appears that many of them have temporal significance based on their monotonic patterns. At this point, it is assumed that the fourteen levels at KG23c are more or less stratified from earliest to latest. On a visual basis, the graphic patterning of many of the type relative frequencies at KG23c reinforces this basic assumption. The most obvious types which demonstrate probable temporal significance are the five major ones; Brindetta Red-topped, Bahri Smoothed, Socodad Wiped, Dratmuir Burnished, and Butana Combed. For example, Butana Combed and Brindetta Red-topped generally decrease in proportion through the levels, while Bahri Smoothed and Dratmuir Burnished increase. The pattern with Socodad Wiped is less certain, however, it is possible that this type increases from the lower levels and then decreases towards the upper levels.

Graphically, when levels (or any other kind of unit) are arranged in a seriation order based on a principal component application, they are represented by two factors on a 2-dimensional scattergram. The horizontal axis is represented by Factor 1, while the vertical axis is represented by factor 2. In the examples below, the relative position of each level on the scattergram are given by two factor co-ordinates which fall between the values of -1 and 1. In all of the principal component scattergrams used in this chapter, only the values pertaining to the first factor (Factor 1) are considered significant in determining if the levels used from any site deposit reflect a good serration order. This is based on the fact that the first factor always accounts for the majority of variability detected between the relative frequencies. Indeed, several independent runs of the same data on a series of other statistical applications including multidimensional scaling, as well as factor and cluster analysis (cf. SPSSx), demonstrated that practically all significant variability can be reduced to a single linear dimension (i.e. along the horizontal axis of the first factor). Nevertheless, it is important to note that a curve along the horizontal axis is acceptable as long as an overall horizontal order of the levels is maintained.

To see whether the monotonic patterns of the various types above are temporally sensitive, the 14 levels at KG23c need to be more or less stratified from early to late. To test whether the 14 levels are stratified, a seriation of the levels can be done by using the relative frequencies of the five major ceramic types (which account for 92 percent of the sherd sample). Essentially a seriation can be accomplished by re-arranging all of the 14 levels in the best possible order based on the best possible monotonic patterns for the five major types. One should note that due to the inherent instability of using sherd counts as a precise means for establishing best possible monotonic patterns for the ceramic types, it is very probable that a perfect sequential ordering of all 14 levels at KG23c will not be produced by a seriation. Nevertheless, if the types used in the seriation reflect basic temporal trends, then the 14 levels at KG23c should senate in roughly the same order that they occur stratigraphically.

Therefore, if the 14 levels from KG23c are stratified (i.e. superposed from early to late), they will be organized in sequential order in either a linear or curved configuration along the horizontal axis of Factor 1. This kind of arrangement among the levels will in turn support the integrity of the temporal patterning observed among the five major ceramic types. Theoretically, the greater horizontal distance between each level in this kind of arrangement indicates a greater temporal distance, which in turn signifies a longer depositional history represented by the levels at a particular site. Thus, if all of the levels are spaced close to one another, this will depict a shorter depositional sequence

38

Definition of the Malawiya/Butana Transition and Butana Group

Figure 4.1. Seriation of the 14 levels at KG23c based on the relative frequencies of the five major ceramic types.

than if they are spaced widely apart along the horizontal axis. Irrespective of the linear arrangement or closeness between levels, if a reasonable sequential order cannot be produced by the principal component application, then the validity of the temporal significance of the ceramic types within a particular deposit must be rejected although good monotonic patterns occur.

On the horizontal scale of Factor 1, levels one, two, and three are tightly clustered in the negative region (from -1 to 0) and are followed by levels four and five (also spaced together closely). In the positive region ( from 0 to 1) levels six and seven occur first and are followed by levels eight through fourteen which are grouped together in another set. Levels one, two, and three all appear to be closely related representing one group (level group 1), as are four and five, representing a second group (level group 2). Levels six and seven are spaced farther apart, but in relation to the rest of the levels, seem to fall into a third group (level group 3).

Figure 4.1, shows a good seriation arrangement of the 14 levels at KG23c (which gives a basic rendition of the superimposed order of the levels), indicating that this particular deposit is stratified.

39

The Butana Group Ceramics

Table 4.3. Relative Frequencies of the Five Major Butana Group Types By Level at KG23c level

B. R-tppd %

B. Smthd. %

S. Wpd. %

D. Brnshd. %.

B. Cmbd. %

total %

Level Group 1

1 2 3

12.7 17.1 12.5

38.3 33.2 29.4

19.7 16.0 27.5

22.1 22.7 22.4

7.2 6.6 8.2

100 100 100

Level Group 2

4 5

22.2 22.1

26.5 27.1

20.9 21.2

19.1 18.1

11.3 11.5

100 100

Level Group 3

6 7

27.9 38.1

17.8 13.8

27.8 24.0

11.4 10.0

15.2 14.0

100 100

Level Group 4

8 9 10 11 12 13 14

36.2 40.2 35.1 30.5 34.5 25.7 36.0

9.9 10.1 12.4 10.6 16.2 8.3 10.4

20.1 19.9 17.4 16.3 14.2 25.0 17.6

11.5 9.6 10.2 7.1 9.6 9.0 9.5

22.3 20.1 24.9 35.5 25.4 31.9 26.6

100 100 100 100 100 100 100

Table 4.4. Relative Frequencies and Corresponding 2 Standard Errors of the Ceramic Types Through the Four Level Groups at KG23c

Level Group 1 Level Group 2 Level Group 3 Level Group 4

B. Cmbd. % (±2s.e.)

B. R-tpp. % (±2s.e.)

B. Smthd. % (±2s.e.)

S. Wpd. % (±2s.e.)

D. Brnsh. % (±2s.e.)

D. C. Mrkd. % (±2s.e.)

K. Pln. % (±2s.e.)

Z. P. Brnsh. % (±2s.e.)

S. Dnttd. % (±2s.e.)

A. Pnctd. % (±2s.e.)

6.7 (±1.0) 10.5 (±1.8) 13.5 (±1.9) 22.5 (±1.7)

12.9 (±1.3) 20.3 (±2.3) 30.0 (±2.5) 33.3 (±2.0)

32.3 (±1.8) 24.7 (±2.5) 14.8 (±2.0) 10.3 (±1.3)

18.7 (±1.5) 19.4 (±2.3) 24.1 (±2.4) 17.5 (±1.6)

21.9 (±1.6) 17.0 (±2.2) 10.0 (±1.7) 0.9 (±1.2)

1.6 (±0.5) 1.2 (±0.6) 1.1 (±0.6) 0.6 (±0.3)

1.6 (±0.5) 1.8 (±0.8) 1.2 (±0.7) 2.9 (±0.7)

0.3 (±0.2) 0.8 (±0.5) 0.6 (±0.4) 0.4 (±0.3)

0.6 (±0.3) 0.2 (±0.2) 0.3 (±0.3) 0.1 (±0.1)

3.5 (±0.7) 4.3 (±1.2) 4.4 (±1.1) 2.9 (±0.7)

S. Rp. Mrkd. % (±2s.e.) 0.0 0.0 0.0 0.1 (±0.1)

total % 100 100 100 100

Of the five major types, Brindetta Red-topped and Butana Combed decrease through time, while Bahri Smoothed and Dratmuir Burnished increase through time within the four level groups. In the same light, Socodad Wiped appears to increase in proportion from level group four to three and then gradually decreases from then on.

Likewise, levels eight through fourteen seem to represent a fourth group (level group 4). Based on the 5 major ceramic type proportions, levels eight through fourteen are less stratified than levels one through seven which show a longer temporal spread. The condensing of the fourteen original levels into four level groups based on the principal component application is reinforced by observing the relative frequencies of the five major Butana Group types by individual levels in Table 4.3. For example, the relative frequencies among levels one through three are very similar to one another, as they are between levels four and five, six and seven, and eight through fourteen. In using the four level groups, as opposed to the 14 individual levels, clearer temporal patterns can be seen among all of the types at KG23c as shown in Table 4.4.

When the six minor types are reviewed in relation to the four level groups, their temporal patterning is less conclusive due to their smaller sample sizes which can be noted by the 2 standard error margins. Nevertheless, some possible patterning can be observed. Alabaska Punctated occurs in all of the level groups, and appears to stay constant (given the 2 standard error margin), however, it is possible that it may peak in both level groups 2 and 3. Khordhag Plain is present through all four level groups, and does not appear to show any kind of temporal patterning. Dratmuira Chatter Marked appears to increase gradually from level group 4 to

40

Definition of the Malawiya/Butana Transition and Butana Group

level group 1. Zirhalla Patterned Burnished also appears to increase slightly from level group 4 to 2 and then decreases from level group 2 to 1. Sirjino Simple Dentated is virtually absent from level groups 4, but is present through the rest of the level groups. Sarb’ut Ripple Marked is only present in level group 4.

level four, the relative frequencies decrease from 26 to 21 percent in level one. When the 2 standard errors are taken into consideration for this type, a vertical line can be placed within the total variance, suggesting that the overall decreasing trend of the relative frequencies through the levels is artificial. When compared to the Socodad Wiped relative frequencies at KG23c, the relative frequencies of the same type at KG7a would correspond roughly with level groups 1 through 3 at KG23c where the Socodad Wiped relative frequencies also appear to diminish as well from 24 to 21 percent.

Based on the findings at KG23c it can be demonstrated that the deposit there is stratified. Within a 2 standard error margin, all of the five major Butana Group types (Brindetta Red-topped, Bahri Smoothed, Dratmuir Burnished, Socodad Wiped, and Butana Combed ) show good temporal patterning through the deposit especially when the 14 levels are condensed into 4 primary level groups. The temporal patterning of the remaining 6 minor types (Alabaska Punctated, Dratmuira Chatter Marked, Khordhag Plain, Sirjino Simple Dentated, and Zirhalla Dentated) is less conclusive, due to the smaller sample sizes of these types. However, it is possible that there may be some temporal patterning associated with Alabaska Punctated, Dratmuira Chatter Marked, and Zirhalla Pattern Burnished. Only 2 sherds of the Sarb’ut Ripple Mark type appears at KG23c, of which both were found in the very lowermost levels of the deposit.

The third largest type is Dratmuir Burnished (n = 645) which constitutes 16 percent of the sherd sample. From levels eight to three, the relative frequencies decrease strongly from 33 to 9 percent. From level three, the relative frequencies increase to 13 percent in level one. Overall, this type at KG7a shows a fairly good monotonic pattern when the 2 standard errors are taken into account. When the Dratmuir relative frequencies at KG7a are compared with KG23c’s, they would fit within the total sequence (level groups 1 - 4), however, in reverse fashion (i.e. level one of KG7a would correspond with level group 4, while level eight of KG7a would correspond with level group 1).

KG7a

After Dratmuir Burnished, the next largest type at KG7a is Brindetta Red-topped (n = 586) which consists of 15 percent of the sherd sample. Like the Dratmuir Burnished type, this particular type displays a strong monotonic pattern when the 2 standard errors are taken into account. From levels eight to two, the type increases from 3 to 26 percent and then decreases back down to 14 percent in level one. When this pattern is compared with the Brindetta Red-topped relative frequencies at KG23c, the two patterns are in reverse with one another. However, at KG23c, the relative frequencies range from 11 to 40 percent, while at KG7a, they range between 3 and 26 percent. As a whole, when the relative frequencies of the Brindetta type at KG7a are matched overall with KG23c, they would fit with level group 1 and 2 of the latter site where the relative frequencies of the same type range from 14 to 22 percent.

Of the seven Butana Group sites, KG7a has the next largest sample of sherds (n = 3914) and contains all ten Butana Group types found at KG23c, including a few sherds of the Sarb’ut Ripple Marked type. KG7a (which has a deposit of 80 cm) was condensed into eight, 10 cm levels, and the frequencies and relative frequencies (with their respective standard errors) for each type can be seen in tables 4.5 and 4.6. Like KG23c, the dominant five types at KG7a are Bahri Smoothed, Socodad Wiped, Dratmuir Burnished, Brindetta Red-topped, and Butana Combed which account for 93 percent of the sherd sample. The Bahri Smoothed type (n = 1092) consists of 28 percent of the sherd sample at KG7a. Through the eight levels at this site, the Bahri Smoothed relative frequencies displays an erratic, non monotonic pattern. In level eight they start at 19 percent, rise to 32 percent in level seven, but decrease to 18 percent in level four. In level five, the relative frequencies rise back to 20 percent and continue to increase up to 37 percent in level one. When compared to the Bahri Smoothed relative frequencies at KG23c, the Bahri Smoothed relative frequencies at KG7a would correspond with level groups 1 and 2 at the former site where the relative frequencies of the same type average between 27 and 32 percent.

The Butana Combed type (n = 385) is the last of the five major types represented at KG7a, and accounts for 10 percent of the sherd sample. The relative frequencies of the Butana Combed type at this site also reflect a convincing monotonic pattern within 2 standard errors. From levels eight to six, the relative frequencies stay roughly the same (fluctuating between 1 and 4 percent) and then increase from levels five through three where the relative frequencies reach 14 percent. From level three, the relative frequencies decrease slightly down to 11 percent in level one. Overall, the relative frequencies of this type at KG7a are monotonic, but like the Brindetta type, do not match well when compared to the same type at KG23c. The Butana Combed relative frequencies at KG7a compare with level groups 1 through 3 at KG23c where the relative frequencies of the same type average between 7 and 15 percent.

The next largest type at KG7a is Socodad Wiped (n = 945) which constitutes 24 percent of the sherd sample. The relative frequencies of this type diminish (with the exception of 2 reversals) through the levels at this site. In level eight, the relative frequencies begin at 32 percent and decrease down to 27 percent in level seven. By level six, the relative frequencies increase back up to 30 percent and decrease as far down as 22 percent in level five. From

41

The Butana Group Ceramics

The remaining six Butana Group types (Dratmuira Chatter Mark, Alabaska Punctated, Sirjino Simple Dentated, Khordhag Organic Plain, Zirhalla Pattern Burnished, and Sarb’ut Ripple Marked) account for only 7 percent of the sherd sample at KG7a. Other than Khordhag Plain, the standard error margins are too wide among these particular types to show any temporal significance.

When the 8 levels at KG7a are seriated (using the five major Butana Group types only) through the principal component method, they organize into 3 level groups all of which are located in the positive sector of Factor 1 between the .2 and 1 interval (see Figure 4.2). Level group 1 consists of levels two and three, which is situated near the .2 interval on the horizontal axis of Factor 1. The second level group consists solely of level one which is located at about .65 on the horizontal axis of Factor 1. The third level group is composed of the remaining levels (4-8) which are situated between .8 and 1 on the horizontal axis of Factor 1. On the horizontal axis, the position of level one in relation to levels two and three seem to be reversed, while levels four through eight do not show any stratigraphic separation.

Of this group Khordhag Organic Plain is the largest type (n = 79) which constitutes a little over 2 percent of the total sherd sample. This particular type’s relative frequencies increase from 2 percent at level eight to 7 percent at level 5. From level five the relative frequencies decrease to 4 percent in level four, and then back up to 11 percent at level three. From level two to one the relative frequencies virtually disappear. When the relative frequencies of this type is compared between KG7a and KG23c, they appear to be roughly the same overall.

Overall, the stratigraphic order of the 8 levels at KG7a is less than satisfactory in providing reliable temporal patterns for the types given the fact that there is some statistical inversion between the levels. Nevertheless, some of the types at KG7a do have convincing monotonic patterns through the eight levels. Unfortunately, the temporal patterning of some of the types at KG7a such as Dratmuir Burnished and Brindetta Red-topped, seem to go in the opposite direction in relation to KG23c, which in turn seems to be contradictory. Nevertheless, when the three level groups of KG7a are factored (using the principal component application) with the four level groups at KG23c, several important aspects can be observed (see Figure 4.3).

Dratmuira Chatter Marked is the second largest type (n = 65) of this group which constitutes almost 2 percent of the total sherd sample. This type at KG7a occurs in the lower levels up until level five, and then virtually drops off from level four on. When the relative frequencies of this type at KG7a is compared with that of KG23c, it appears that the former relative frequencies would fit into level groups 1 to 3 of the former site where the relative frequencies average between 1 and 2 percent.

First, is the apparent reversed order of two of the three level groups associated with KG7a with respect to the four level groups associated with KG23c. This reversed patterning is a result of the reversed monotonic patterns noted among some of the 5 major types at KG7a. Off hand, one could concede that identical ceramic types from both KG23c and KG7a may have behaved differently through time, although this seems highly unlikely. On the other hand, if the deposits from both sites were not of the same age, then it may be that some of the monotonic patterns associated with the types, which at first seem to go in different directions, in fact represent the lower (incipience) and upper halves (decline) of a continuous monotone. For example, a full monotonic pattern (incipience, florescence, and decline) can be produced by placing the relative frequencies of the Brindetta Red-topped type associated with the four level groups at KG23c above the Brindetta Red-topped relative frequencies associated with the 3 level groups at KG7a. This suggests that the deposit at KG7a entirely predates the deposit at KG23c. However, when all of the 5 major types relative frequencies are arranged in this particular manner, many do not form similar composite monotonic patterns. Furthermore, the radiocarbon dates from both deposits indicate that they date to the same age, with KG7a being slightly younger. For example, there are two radiocarbon dates from the lower deposit (within level group 3) of KG7a with mean dates of 3163±93 and 3351±68 BC, and two other dates from the lower deposit of KG23c (within level group 4) with mean dates of 3283±67 and 3544±154 BC (cf. Marks and Sadr 1988: 73). These particular radiocarbon

The Alabaska Punctated type (n = 54) accounts for a little over 1 percent of the total sherds sample from KG7a. Overall, the relative frequencies of this type are too small for any kind of pattern at this site. When compared with KG23c, this particular type is less prevalent at KG7a and may indicate that the Alabaska type was less popular at KG7a. This would suggest that the relative frequencies of this type at KG7a would match the post peak interval of the same type which would have occurred within level group 1 at KG23c. Zirhalla Pattern Burnished (n = 42) accounts for approximately 1 percent of the total sherd sample at KG7a. Like the Alabaska type, the relative frequencies of this type are also too small to be monitored through the levels at KG7a. This particular type is found in the lowest level at KG7a, while it is missing from the 4 lowest levels at KG23c and suggests a similar pattern noted with the Dratmuira type between these two sites. Overall, the pattern noted with the Dratmuira and Zirhalla types would indicate that the lowest level at KG7a would probably have occurred after level group 4 at KG23c. The remaining two types, Sirjino Dentate (n = 17) and Sarb’ut Ripple Marked (n = 4) account for the remaining 1/2 of 1 percent of the typed sherd sample at KG7a. These types are very rare at KG7a, as they are at KG23c. It is important to note that the sherds associated with the Sarb’ut Ripple Marked type occur in the very top level at KG7a, as well as in the middle and lower levels as well.

42

Definition of the Malawiya/Butana Transition and Butana Group

Figure 4.2. Seriation of the 8 levels at KG7a based on the relative frequencies of the five major ceramics types.

dates also provide some additional information as far as stratification of the deposits as well. For example, the two samples recovered from KG7a have a 188 mean year difference between them, even though they were recovered from the same 5 cm level (i.e. level 16, which is 80 cm below surface). On the other hand, the earlier radiocarbon sample at KG23c was stratigraphically lower (recovered at 135 cm below surface) than the younger sample (recovered at 115 cm below surface). In assessing these particular dates, one would have to conclude that the two dates associated with KG23c are more reflective of a stratified deposit than the two dates from KG7a. Furthermore, the fact that two of the four Sarb’ut Ripple Marked sherds from KG7a were recovered from the mid and upper most levels, would also suggest that the deposit at KG7a may be disturbed. As it will be discussed in more detail latter on, these particular ripple sherds are more closely affiliated with the Malawiya/ Butana transition and would only have occurred in the very earliest Butana Group deposits as evidenced by their stratigraphic positioning at KG23c.

A second aspect of the overall arrangement of the level groups from KG7a in relation to those associated with KG23c (see Figure 4.3), is that all of the level groups associated with KG7a occur after level groups four and three of KG23c. This suggests that the deposit associated with KG7a began after 70 cm of accumulation (from levels 14 to 7) had occurred at KG23c. Indeed, the types Dratmuira Chatter Marked, Zirhalla Pattern Burnished, and Sirjino Simple Dentated all appear in the lowermost levels at KG7a, while at KG23c, these types do not occur until level ten and nine. If one were to accept the dates of the two radiocarbon samples from KG23c (which were located in levels 14 and 12) and that this deposit is essentially stratified, then based on the seriation arrangement of the level groups overall, the deposit of KG7a would have to occur after the youngest radiocarbon sample associated with KG23c which has a mean date of 3283 BC. When the two radiocarbon dates at KG7a are averaged, they yield a mean date of 3257 BC and are in agreement with the overall seriation of the level groups from both sites. Indeed, one might be compelled to

43

The Butana Group Ceramics

Figure 4.3. Seriation of the 4 level groups at KG23c and the 3 level groups at KG7a based on the relative frequencies of the five major ceramics types.

sedation of KG7a in relation to the four level groups at KG23c. As it can be seen, the temporal positioning of KG7a fits best between level groups 1 and 2 of KG23c. This is consistent with the individual matching of each of the type relative frequencies through the levels at KG7a with respect to the correspond level group(s) at KG23c. Of course, this is in agreement with the radiocarbon dates provided from both sites. One should note, however, that the temporal patterning of the five type proportions is not perfect. For example, the relative frequencies of the Socodad Wiped type at KG7a appear to be too high. Nevertheless, when the relative frequencies are considered within the 2 standard error range, the Dratmuir Burnished value at KG7a does not appear to be too far off. On the other hand, the Socodad Wiped value is still high irrespective of the standard error range. At this time, there is no explanation for the high proportion of this type at KG7a. It is likely that there were simply more sherd fragments of this particular type at KG7a, perhaps due to differential vessel portion deposition.

accept the more recent date of 3152 BC at KG7a based on this information. Both radiocarbon dates from KG7a were recovered from the very base of the deposit, suggesting that the occupation began sometime around the mean date of 3257 BC. Thus, in relation to the earliest mean radiocarbon date associated with KG23c (3544 BC), it is very conceivable that the bulk of the occupation associated with KG7a did occur several hundred years after the beginning of KG23c. When the levels of KG7a are collapsed into one single unit and another principal component sedation is performed using the proportions of the 5 major Butana Group types among KG7a and the four level groups associated with KG23c, a clearer temporal positioning of KG7a can be observed in Figure 4.4. Table 4.7 gives the temporal patterns of the 5 major type relative frequencies based on the principal component

44

Definition of the Malawiya/Butana Transition and Butana Group

Figure 4.4. Seriation of the 4 level groups at KG23c and the collapsed deposit at KG7a based on the relative frequencies of the five major ceramic types.

sherd sample. In level six, the relative frequency of this type is 52 percent. From level five to one, the relative frequencies fluctuate from 30 to 36 percent. When the 2 standard error margins are taken into consideration the monotonic patterning of this type is not significant. Relative to KG23c, the Dratmuir Burnished type relative frequencies are significantly higher at KG29S. At KG23c, the relative frequencies of Dratmuir Burnished are at their maximum at 24 percent in level group 1, while at KG29S the relative frequencies of this same type never fall below 29 percent.

In any case, there is no reason to suspect that the higher proportion of the Socodad Wiped type at KG7a is due to some cultural or functional difference. KG29S The next largest sample of Butana Group sherds comes from KG29S which contains a total of 2755 sherds. KG29S possesses nine of the ten Butana Group types represented at KG23c and the Sarb’ut Ripple Marked type. The one Butana Group type which is missing from KG29S is Khordhag Plain. The cultural deposit at KG29S was 60 cm deep and was condensed into six, 10 cm levels. Tables 4.8 and 4.9 give the frequencies and relative frequencies (with their respective 2 standard error margins) for all ten types through the six levels at KG29S. Of the 2755 typed sherds at KG29S, 90 percent belong to four of the five major Butana Group types noted at KG23c and KG7a. Unlike KG23c and KG7a, the Brindetta Red-topped type is very rare at KG29S.

The second most numerous type is Socodad Wiped (n = 772) which constitutes 28 percent of the sherd sample. In level six, the relative frequencies begin at 15 percent and increase continuously to 33 percent in level one. When the relative frequencies of this type are bracketed within the 2 standard error margins, there appears to be a good monotonic pattern from level four to level one. In comparison with KG23c, these particular relative frequencies at KG29S could conceivably fit within level groups 4 to 3 at KG23c where the relative frequencies of the same type also increase from 19 to 24 percent.

Dratmuir Burnished represents the most numerous type at KG29S (n = 898) and accounts for 36 percent of the

45

The Butana Group Ceramics

Table 4.5. Frequencies of the Ceramic Types Through the Levels at KG7a level 1 2 3 4 5 6 7 8 total

B. R-tpped. freq. 188 137 100 72 35 38 11 5 586

B. Smthd. freq. 504 123 93 85 62 104 93 28 1092

S. D. Wpd. Brnshd. freq. freq. 291 112 113 119 62 125 77 46 945

B. Cmbd. freq.

172 61 41 88 58 98 80 47 645

146 70 73 57 25 4 4 6 385

A. Pncttd. freq.

K. Pln. freq.

24 12 9 4 1 2 2 0 54

5 1 5 17 21 16 11 3 79

D.C. Z. Mrkd. Brnshd. freq. freq. 4 3 10 7 14 17 6 4 65

9 7 6 5 5 7 1 2 42

S. Dnttd. freq. 7 0 2 2 1 3 0 2 17

S. Rp. Mrkd. freq. 1 0 0 1 0 0 1 1 4

total freq. 1351 526 452 457 284 414 286 144 3914

Table 4.6. Relative Frequencies and 2 Standard Error Margins of the Ceramic Types Through the Levels at KG7a

level

1 2 3 4 5 6 7 8

B. R-tpped. % (±2s.e.)

B. Smthd. % (±2s.e.).

S. Wpd. % (±2s.e.)

D. Brnshd. % (±2s.e.)

B. Cmbd. % (±2s.e.)

A. Pncttd. % (±2s.e.)

K. Pln. % (±2s.e.)

D.C. Mrkd. % (±2s.e.)

Z. Brnshd. % (±2s.e.)

S. Dnttd. % (±2s.e.)

S. Rp. Mrkd. % (±2s.e.)

13.9 (±1.9) 26.0 (±3.8) 22.1 (±3.9) 15.8 (±3.4) 12.3 (±3.9) 9.2 (±2.8) 3.8 (±2.3) 3.5 (±3.1)

37.3 (±2.6) 23.4 (±3.7) 20.6 (±3.8) 18.6 (±3.6) 21.8 (±4.9) 25.1 (±4.3) 32.5 (±5.5) 19.4 (±6.6)

21.5 (±2.2) 21.3 (±3.6) 25.0 (±4.1) 26.0 (±4.1) 21.8 (±4.9) 30.2 (±4.5) 26.9 (±5.2) 31.9 (±7.8)

12.7 (±2.2) 11.6 (±3.0) 9.1 (±3.4) 19.3 (±3.7) 20.4 (±2.7) 23.7 (±2.3) 28.0 (±2.4) 32.6 (±32.6)

10.8 (±1.8) 13.3 (±2.8) 16.2 (±2.7) 12.5 (±3.7) 8.8 (±4.8) 1.0 (±4.2) 1.4 (±5.3) 4.2 (±7.8)

1.8 (±0.7) 2.3 (±1.3) 2.0 (±1.3) 0.9 (±0.9) 0.4 (±0.7) 0.5 (±0.7) 0.7 (±1.0) 0.0

0.4 (±0.3) 0.2 (±0.4) 1.1 (±1.0) 3.7 (±1.8) 7.4 (±3.1) 3.9 (±1.9) 3.8 (±2.3) 2.1 (±2.4)

0.3 (±0.3) 0.6 (±0.7) 2.2 (±1.4) 1.5 (±1.1) 4.9 (±2.6) 4.1 (±2.0) 2.1 (±1.7) 2.8 (±2.7)

0.7 (±0.4) 1.3 (±1.0) 1.3 (±1.1) 1.1 (±1.0) 1.8 (±1.6) 1.7 (±1.3) 0.3 (±0.7) 1.4 (±2.0)

0.5 (±0.4) 0.0

0.1 (±0.1) 0.0

0.4 (±0.6) 0.4 (±0.6) 0.4 (±0.7) 0.7 (±0.8) 0.0

0.0

1.4 (±2.0)

0.2 (±0.4 0.0 0.0 0.3 (±0.7) 0.7 (±1.4)

total % 100 100 100 100 100 100 100 100

Table 4.7. Relative Frequencies and 2 Standard Error Margins of the Five Major Butana Group Types Among the 4 Level Group at KG23c and the Collapsed Deposit at KG7a

Level Group 1 KG7a Level Group 2 Level Group 3 Level Group 4

B. R-tppd. % (± 2 s.e.)

B. Smthd. % (± 2 s.e.)

S. Wpd. % (± 2 s.e.)

D. Brnshd. % (± 2 s.e.)

B. Cmbd. % (± 2 s.e.)

14.0 (±1.6) 16.0 (±1.2) 22.1 (±2.5) 32.5 (±2.7) 35.8 (±2.1)

34.9 (±1.9) 29.9 (±1.5) 26.9 (±2.7) 16.0 (±2.1) 11.1 (±1.3)

20.2 (±1.6) 25.9 (±1.4) 21.1 (±2.5) 26.1 (±2.5) 18.8 (±1.7)

23.7 (±1.7) 17.7 (±1.3) 18.5 (±2.4) 10.8 (±1.8) 9.9 (±1.3)

7.2 (±1.0) 10.5 (±1.0) 11.4 (±1.9) 14.7 (±2.0) 24.5 (±1.8)

46

total % 100 100 100 100 100

Definition of the Malawiya/Butana Transition and Butana Group

Table 4.8. Frequencies of the Ceramic Types Through the Levels at KG29S level 1 2 3 4 5 6 total

B. R-tpped. freq. 15 14 9 3 1 1 43

B. Smthd. freq. 199 102 40 64 31 1 437

S. D. Wpd. Brnshd. freq. freq. 322 216 98 101 31 4 772

289 249 136 171 49 4 898

B. Cmbd. freq.

A. Pncttd. freq.

82 72 68 109 22 4 357

1 0 0 0 0 0 1

K. Pln. freq. 0 0 0 0 0 0 0

D.C. Z. Mrkd. Brnshd. freq. freq. 37 54 12 27 7 1 138

7 9 8 22 5 1 52

S. Dnttd. freq. 24 17 3 7 3 1 55

S. Rp. Mrkd. freq. 1 0 0 1 0 0 2

total freq. 977 733 374 505 149 17 2755

Table 4.9. Relative Frequencies and 2 Standard Error Margins of the Ceramic Types Through the Levels at KG29S

level

1 2 3 4 5 6

B. R-tpped. % (±2s.e.)

B. Smthd. % (±2s.e.).

S. Wpd. % (±2s.e.)

D. Brnshd. % (±2s.e.)

B. Cmbd. % (±2s.e.)

A. Pncttd. % (±2s.e.)

1.5 (±0.8) 1.9 (±1.0) 2.4 (±1.6) 0.6 (±0.7) 0.7 (±1.3) 3.7 (±7.3)

20.4 (±2.6) 13.9 (±2.6) 10.7 (±3.2) 12.7 (±3.0) 20.8 (±6.7) 3.7 (±7.3)

33.0 (±3.0) 29.5 (±3.4) 26.2 (±4.5) 20.0 (±3.6) 20.8 (±6.7) 14.8 (±13.7)

29.6 (±2.9) 34.0 (±3.5) 36.4 (±5.0) 33.9 (±4.2) 32.9 (±7.7) 51.8 (±19.3)

8.4 (±1.8) 9.8 (±2.2) 18.2 (±4.0) 21.6 (±3.7) 14.8 (±5.8) 14.8 (±13.7)

0.1 (±0.2) 0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

The third most frequent type at KG29S is Bahri Smoothed (n=437) which accounts for 16 percent of the sherd sample. From level six to five, the relative frequencies increase from 4 to 20 percent. From level five to three, the relative frequencies decrease to 10 percent, then increase back up to 20 percent at level one. Overall, the series of relative frequencies produced by this type through the levels do not present a convincing monotonic pattern when the 2 standard error margins are taken into consideration. Nevertheless, when the relative frequencies of this type are compared to the relative frequencies of the same type at KG23c, the former relative frequencies tend to fit within level group 4 through 2 at KG23c where the relative frequencies range from 10 to 25 percent.

K. Pln. % (±2s.e.) 0.0

D.C. Mrkd. % (±2s.e.)

Z. Brnshd. % (±2s.e.)

S. Dnttd. % (±2s.e.)

S. Rp. Mrkd. % (±2s.e.)

3.8 (±1.2) 7.4 (±1.9) 3.2 (±1.8) 5.3 (±2.0) 4.7 (±3.5) 3.7 (±7.3)

0.7 (±0.5) 1.2 (±0.8) 2.1 (±1.5) 4.4 (±1.8) 3.4 (±3.0) 3.7 (±7.3)

2.5 (±1.0) 2.3 (±1.1) 0.8 (±0.9) 1.4 (±1.0) 2.0 (±2.3) 3.7 (±7.3)

0.1 (±0.2) 0.0 0.0 0.2 (±0.4) 0.0 0.0

total % 100 100 100 100 100 100

type does appear to steadily decrease. In comparison with KG23c, the Butana Combed relative frequencies at KG29S fit well within level groups 4 through 2 at the former site where the relative frequencies of the same type decrease from 25 to 10 percent. The remaining 6 types at KG29S (Dratmuira Chatter Marked, Sirjino Simple Dentated, Zirhalla Pattern Burnished, Brindetta Red-topped, Alabaska Punctated, and Sarb’ut Ripple Marked) constitute the remaining 10 percent of the sherd sample at KG29S. The largest type within these six is Dratmuira Chatter Marked (n = 138) which accounts for 5 percent of the sherd sample. The relative frequencies of this type (within the 2 standard error margin) are constant through the levels ranging between 4 and 6 percent. In relation to KG23c (where Dratmuira relative frequencies range from 0 to 2 %) the relative frequencies of this type are consistently higher at KG29S.

The Butana Combed type is the fourth most frequent type (n = 357) and represents 13 percent of the total sherd sample. In level six, the relative frequencies begin at 15 percent and increase to 22 percent at level four. From level four, the relative frequencies decrease steadily back down to 8 percent. When the relative frequencies of this type are examined with their 2 standard error margins, the apparent increase of this type from levels six to four does not seem to be reliable. However, from level four to level one, this

Sirjino Simple Dentated (n = 55) accounts for 2 percent of the sherd sample. The relative frequencies of this type are somewhat erratic (ranging between 1 and 6 %) and are present throughout all six levels. In contrast, to KG23c, KG29S contains more than twice the number of sherds belonging to this type.

47

The Butana Group Ceramics

Figure 4.5. Seriation of the 6 levels at KG29S based on the relative frequencies of the four major ceramic types.

The last type at this site is represented by the type Sarb’ut Ripple Marked. Only two sherds of this type were found, coming from the upper levels (levels 1 & 2).

Zirhalla Pattern Burnished (n = 52) accounts for a little less than 2 percent of the sherd sample. When the relative frequencies of this type are observed through the levels at KG29S, they appear to decrease even within their 2 standard error margin. Like Dratmuira Chatter Marked and Sirjino Dentate, KG29S contains a larger number of sherds belonging to the Zirhalla Pattern Burnished type than KG23c.

When the six levels at KG29S are seriated by the principal component analysis (using the four major Butana Group types) there does not appear to be any clear ordering of the levels along the horizontal axis as can be seen in Figure 4.5.

The Brindetta Red-topped type (n = 43) only accounts for a little more than 1 percent of the sherd sample. The low number of sherds of this type at KG29S is in marked contrast to KG23c as well as KG7a where it is present in much greater numbers, and is considered to be a major type at these particular sites. The very small showing of the Brindetta Red-topped type at KG29S may indicate that there is some cultural difference between this site and the sites of KG23c and KG7a.

Indeed, all six levels at KG29S are located between the .7 and 1 interval mark on the horizontal scale of Factor 1. Overall, the lack of any horizontal distribution of the levels at KG29S may indicate that the deposit is either heavily disturbed or does not represent a very long interval of time. However, like KG7a, the graphic display of some of the type relative frequencies (such as Socodad Wiped and Butana Combed) through the six levels at KG29S appear to show some patterning. Unfortunately, no radiocarbon dates were recovered from KG29S. When the levels of KG29S are collapsed into a single unit and a principal component seriation is performed using the proportions of four major Butana Group types at KG29S (Brindetta Red-topped was not included due to the small sample size) in relation to

The Alabaska Punctated type is also under represented at KG29S and only one sherd was recovered. The idea, that the Alabaska Punctated type is somehow related to the Brindetta Red-topped type (the former type probably had red-topped rims as well) may be a factor why there are hardly any sherds of this particular type at KG29S.

48

Definition of the Malawiya/Butana Transition and Butana Group

Figure 4.6. Seriation of the 4 level groups at KG23c and the collapsed deposit at KG29S based on the relative frequencies of the four major ceramics types.

place in relation to the four level groups at KG23c. The deviant position of KG29S in relation to the level groups at KG23c, plus the anomalously high and low proportions of the Dratmuir Burnished and Bahri Smoothed types respectively, as well as the rarity of the Brindetta Red-topped type, may reflect some kinds of cultural differences between KG29S and KG23c. It is doubtful that the phenomenon of differential vessel portion deposition among various types between the sites (as in the case between KG7a and KG23c) can be used as an explanation for these apparent differences.

same types among the four level groups at KG23c, a more clear temporal positioning of KG29S can be observed in Figure 4.6 below. Table 4.10 below gives the temporal pattern of the 4 major ceramic type relative frequencies which correspond to the principal component ordering of KG29S in relation to the four level groups at KG23c. As it can be observed from both Figure 4.6 and Table 4.10, the temporal placement of KG29S fits best between level group 3 and 4 of KG23c. This is consistent with the individual matching of each of the type relative frequencies through the levels at KG29s in respect to the corresponding level group(s) at KG23c. The relative frequencies of the Socodad Wiped and Butana Combed ceramic types at KG29S conform well with the monotonic patterns displayed by the four level groups at KG23c. Nevertheless, the relative frequencies of the Bahri Smoothed and Dratmuir Burnished types do not conform. In fact, both the proportions of these latter types at KG29S are quite anomalous irrespective of the 2 standard error range. In reexamining the scattergram in Figure 4.6, the vertical positioning of KG29S also seems to be out of

KG96a KG96a has the next largest sherd sample (n = 2080) of the Butana Group sites and contains all ten Butana Group types found at KG23c including the Sarb’ut Ripple Marked type. This site has a 60 cm cultural deposit and was condensed into six, 10 centimeter levels. The frequencies and relative frequencies (including the corresponding 2 standard errors) can be observed for all of the types through the six levels at KG96a in Tables 4.11 and 4.12.

49

The Butana Group Ceramics

Table 4.10. Relative Frequencies and 2 Standard Error Margins of the Four Major Butana Group Ceramic Types at KG29S in Relation to the 4 Level Groups at KG23c

Level Group 1 Level Group 2 KG29S Level Group 3 Level Group 4

B. Smthd. % (± 2 s.e.)

S. Wpd. % (± 2 s.e.)

D. Brnshd. % (± 2 s.e.)

B. Cmbd. % (± 2 s.e.)

40.6 (±2.1) 34.6 (±3.3) 17.7 (±1.5) 23.7 (±0.3) 17.2 (±0.2)

23.5 (±1.8) 27.1 (±3.1) 31.2 (±1.9) 38.7 (±3.4) 29.2 (±2.4)

27.5 (±1.9) 23.7 (±2.9) 36.7 (±1.9) 16.0 (±2.6) 15.4 (±1.9)

8.4 (±1.2) 14.7 (±2.4) 14.4 (±1.4) 21.7 (±2.9) 38.2 (±2.6)

total % 100 100 100 100 100

Table 4.11. Frequencies of the Ceramic Types Through the Levels at KG96a level

B. R-tpped. freq.

B. Smthd. freq.

1 2 3 4 5 6 total

66 105 48 16 5 4 244

49 201 76 79 65 64 564

S. D. Wpd. Brnshd. freq. freq. 39 125 72 82 24 49 391

25 56 50 72 41 33 277

B. Cmbd. freq.

A. Pncttd. freq.

K. Pln. freq.

14 62 22 11 4 6 119

1 15 4 2 0 0 22

44 26 58 183 45 47 403

D.C. Z. Mrkd. Brnshd. freq. freq. 3 3 9 10 4 9 38

0 0 2 6 0 3 11

S. Dnttd. freq.

S. Rp. Mrkd. freq.

total freq.

0 1 2 0 1 0 4

0 0 1 3 0 3 7

241 594 344 464 189 248 2080

Table 4.12. Relative Frequencies and 2 Standard Error Margins of the Ceramic Types Through the Levels at KG96a

level

1 2 3 4 5 6

B. R-tpped. % (±2s.e.)

B. Smthd. % (±2s.e.).

S. Wpd. % (±2s.e.)

D. Brnshd. % (±2s.e.)

B. Cmbd. % (±2s.e.)

A. Pncttd. % (±2s.e.)

K. Pln. % (±2s.e.)

D.C. Mrkd. % (±2s.e.)

27.4 (±5.7) 17.7 (±3.1) 14.0 (±3.7) 3.4 (±1.7) 2.6 (±2.3) 1.6 (±1.6)

20.3 (±5.2) 33.8 (±3.9) 22.1 (±4.5) 17.0 (±3.5) 34.4 (±6.9) 37.9 (±6.2)

16.2 (±4.7) 21.0 (±3.3) 20.9 (±4.4) 17.7 (±3.5) 12.7 (±4.8) 19.8 (±5.1)

29.6 (±3.6) 34.0 (±2.4) 36.4 (±3.8) 33.9 (±3.4) 32.9 (±6.0) 51.8 (±4.3)

10.4 (±3.0) 9.4 (±2.5) 14.5 (±2.6) 15.5 (±1.4) 21.7 (±2.1) 13.3 (±2.0)

10.4 (±0.8) 2.5 (±1.3) 1.2 (±1.2) 0.4 (±0.6) 0.0

18.3 (±5.0) 4.4 (±1.7) 16.9 (±4.0) 39.4 (±4.5) 23.8 (±6.2) 19.0 (±5.0)

1.2 (±1.4) 0.5 (±0.6) 2.6 (±1.7) 2.2 (±1.3) 2.1 (±2.1) 3.6 (±2.4)

0.0

Six types (Bahri Smoothed, Khordhag Plain, Socodad Wiped, Dratmuir Burnished, Brindetta Red-topped, and Butana Combed) consist of 96 percent of the sherd sample. Bahri Smoothed is the largest type (n = 564) accounting for 27 percent of the sherd sample. From levels six to four, the relative frequencies of this type decrease from 38 to 17 percent. From level four to two, the relative frequencies increase back up to 34 percent and decrease back down to

Z. Brnshd. % (±2s.e.)

S. Dnttd. % (±2s.e.)

S. Rp. Mrkd. % (±2s.e.)

0.0

0.0

0.0

0.0

0.2 (±0.3) 0.6 (±0.8) 0.0

0.0

0.6 (±0.8) 1.3 (±1.0) 0.0 (±) 1.2 (±1.4)

0.5 (±1.1) 0.0

0.3 (±0.6) 0.6 (±0.7) 0.0 1.2 (±1.4)

total % 100 100 100 100 100 100

20 percent at level one. Overall, the relative frequencies of this type appear to be erratic through the levels at KG96a (even when the 2 standard errors are taken into account), and no monotonic pattern can be discerned. When the relative frequencies of this type are compared with the Bahri Smoothed relative frequencies from KG23c, it would appear that they would fit somewhere between level group 1 and 2 of the latter site.

50

Definition of the Malawiya/Butana Transition and Butana Group

The next largest type at KG96a is Khordhag Plain (n = 403) which accounts for 19 percent of the sherd sample. From level six to four, the relative frequencies of this type increase from 19 to 39 percent. From level four to two, they decrease to 4 percent, and then increase back up to 18 percent at level one. With the exception of level one, the relative frequencies of the Khordhag Plain type at KG96a display a monotonic pattern (even within a 2 standard error margin). However, the relative frequencies of this type are very high at this site in relation to the other Butana Group sites. At KG23c, the Khordhag Plain type relative frequencies never exceed past 10 percent, while at KG96a they reach almost 40 percent. It is conceivable that the very high proportions of Khordhag Plain relative to the rest of the types at KG96a is due to differential vessel portion deposition. On the other hand, the disproportionate number of Khordhag Plain sherds may indicate that at least part of the occupation at KG96a was somewhat different (functionally or culturally) than the other Butana Group occupations associated with the other sites.

sherd sample at KG96a. The relative frequencies of this type increase steadily from 2 percent to 27 percent from level six to one. Overall, the Brindetta type at KG96a appears to have a good monotonic pattern, within a 2 standard error margin. In comparison with KG23c, the relative frequencies of this type at KG96a appear to correspond with level groups 1 and 2 where the relative frequencies of the same type range between 13 and 20 percent. However, the monotonic patterns displayed between the Brindetta types at KG96a and KG23c are opposed to one another like the patterns noted with the Dratmuir Burnished type. The sixth most numerous type at KG96a consists of the Butana Combed type (n=121) which accounts for 6 percent of the sherd sample. The relative frequencies of this type from levels six to four stay constant at 2 percent, and then increase to 10 percent from levels four to two. From level two, the relative frequencies of this type decrease back to 6 percent at level one. Overall, there is a monotonic pattern of the Butana Combed type at KG96a. When the Butana Combed relative frequencies are compared between KG96a and KG23c, KG96a corresponds with level groups 1 and 2 at KG23c, where the Butana Combed relative frequencies range between 11 and 6 percent. However, like the Brindetta and Dratmuir types, the monotonic patterns displayed between the Butana type at both sites are in opposition.

After Khordhag Plain, the next most frequent type is Socodad Wiped (n=392) which constitutes 19 percent of the sherd sample at KG96a. From level six to five, the relative frequencies of this type decline from 20 to 13 percent. From level five, the relative frequencies increase steadily up to level two where they reach 21 percent. From level two to level one, the relative frequencies decrease back down to 16 percent. When the relative frequencies of this type are bracketed within a 2 standard error margin, it appears that they stay more or less constant (not deviating significantly in any direction) and as a result, no monotonic pattern can be discerned for the Socodad Wiped type at KG96a. In comparison to KG23c, the relative frequencies of this type at KG96a could fit anywhere among the four level groups at the former site where the relative frequencies of the same type range from 17 to 24 percent. However, given the fact that the Bahri Smoothed relative frequencies are high at KG96a and correspond with the upper level Bahri Smoothed relative frequencies at KG23c, it is conceivable that Socodad Wiped relative frequencies at this site would also relate to the upper levels (level groups 1 and 2) at KG23c.

The remaining 4 minor Butana Group types at KG96a consists of Dratmuira Chatter Marked, Alabaska Punctated, Zirhalla Pattern Burnished, Alabaska Punctated, and Sirjino Simple Dentated. Sarb’ut Ripple Marked is also represented at KG96a. Altogether, these 5 types constitute 4 percent of the sherd sample at KG96a. Dratmuira Chatter Marked is the most frequent among these 5 types (n = 38) and consists of 2 percent of the sherd sample. This type appears in all the levels at KG96a, and its relative frequencies decrease slightly from 4 to 1 / 2 of 1 percent. When the 2 standard errors are taken into consideration, however, it is apparent that the sample size of this particular type is too small to reflect any meaningful pattern through the levels at KG96a. In comparison to KG23c, the relative frequencies of this type at KG96a correlate best with level groups 1 and 2 of the latter site where the relative frequencies of the same type range between 1 and 2 percent.

Dratmuir Burnished is the fourth largest type (n = 277) at KG96a accounting for 13 percent of the sherd sample. From level six to five, the relative frequencies of this type increase from 13 to 22 percent. From level five to two, the relative frequencies decrease to 9 percent and then increase slightly to 10 percent at level one. When the relative frequencies of this type are bracketed within a two standard error margin, a monotonic pattern does appear. However, in comparison to KG23c, the patterning of the Dratmuir Burnished relative frequencies at KG96a go in the opposite direction. Overall, the relative frequencies of this type at KG96a could fit anywhere within the four level groups at KG23c.

After Dratmuira Chatter Marked, is the Alabaska Punctated type (n = 22) which accounts for 1 percent of the sherd sample. This type is not present in either levels six or five at KG96a, however, the sample size is very small. Overall, the relative frequencies of this type increase from level five to two and then decrease slightly from level two to one. Nevertheless, when the relative frequencies are bracketed within their 2 standard error margin it is apparent that the sample size is too small to reflect a true monotonic pattern. In comparison with KG23c (where the Alabaska Punctated relative frequencies range between 2 and 4 %), the relative frequencies of Alabaska Punctated at KG96a are smaller.

The fifth largest type is represented by the Brindetta Redtopped type (n=244) which constitutes 12 percent of the

51

The Butana Group Ceramics

Figure 4.7. Seriation of the 6 levels at KG96a based on the relative frequencies of the five major ceramic types.

After Alabaska Punctated, the next largest type is Zirhalla Pattern Burnished (n=11) which constitutes 1 / 2 of 1 percent of the sherd sample. This type is present in the lower as well at the upper levels at KG96a. Like Dratmuira Chatter Marked and Alabaska Punctated, the relative frequencies of this type are too small to assess any kind of patterning when their respective 2 standard error margins are taken into account. Like KG23c, this type is also very rare at KG96a.

When the 6 levels at KG96a are seriated by the principal component analysis (using the five major Butana Group types and excluding the Khordhag Plain type which is probably over represented at KG96a), two level groups are apparent in Figure 4.7. The first level group (level group 1) consists of level one which is situated near the .4 mark on the horizontal scale of Factor 1. The second level group consists of levels two through eight which are located between .8 and 1 on the horizontal scale of Factor 1. When the relative frequencies of the five major Butana Group types at KG96a are ordered by level group (level group 1 is assumed to be stratigraphically the youngest) the following observations can be made in Table 4.13.

There are 7 Sarb’ut Ripple Marked sherds at KG96a which account for a little less than 1/2 of 1 percent of the sherd sample. In relation to the other Butana Group sites, KG96a contains the highest number of Sarb’ut Ripple Marked sherds. However, the sample size of this type is very small at any of the Butana Group sites, and never exceeds 1/2 of 1 percent of the sherd sample. The Sarb’ut Ripple Marked sherds at KG96a are situated in levels six, four, and two.

Based on these figures, all of the relative frequencies except for those associated with the Brindetta Red-topped type stay constant from level group 1 and 2, especially when the 2 standard error margins are considered. Thus, the monotonic patterns noted above for these particular types are artificial. Nevertheless, the relative frequencies of Bahri Smoothed still decrease slightly from level group 2 to 1 while the relative frequencies of Brindetta Red-topped increase significantly,

There are only 4 Sirjino Simple Dentated sherds at KG96a. The relative frequencies of this particular type seem to be very similar between KG96a and KG23c where it is also very rare.

52

Definition of the Malawiya/Butana Transition and Butana Group

Table 4.13. Relative Frequencies and 2 Standard Error Margins of the Five Major Butana Group Type by the Two Level Groups at KG96a

Level Group 1 Level Group 2

B. R-tppd % (± 2 s.e.)

B. Smthd. % (± 2 s.e.)

S. Wpd. % (± 2 s.e.)

D. Brnshd. % (± 2 s.e.)

B. Cmbd. % (± 2 s.e.)

34.2 (±6.8) 12.7 (±1.8)

25.4 (±6.3) 36.7 (±2.6)

20.2 (±5.8) 25.1 (±2.3)

13.0 (±4.8) 18.0 (±2.1)

7.3 (±3.7) 7.5 (±1.4)

total % 100 100

Butana Group types observed at KG23c, but does not yield any Sarb’ut Ripple Marked sherds. The deposit at KG5a was 50 cm deep and was condensed into five 10 cm levels. The frequencies, relative frequencies and the corresponding 2 standard errors can be observed for all of the types through the five levels at KG5a in Tables 4.15 and 4.16.

even when the 2 standard errors are taken into account. As mentioned above, the monotonic patterns of these particular types at KG96a seem to be in reverse order relative to KG23c. As with KG7a and KG29S, the infra-site temporal patterning of the Butana Group ceramic types are not reliable at KG96a, thus indicating that the deposit at this latter site might be disturbed. If the deposit at KG96a was stratified, one would expect to find all or most of the Sarb’ut Ripple Marked sherds in the lowermost levels at this site which they are not. Furthermore, a radiocarbon sample recovered within the deposit at KG96a yields a very late date of 984 ± 107 BC (cf. Marks and Sadr 1988: 73) which is more affiliated with a latter Jebel Mokram occupation situated on the same site a little ways from the deposit (Sadr 1988: 93).

Five Butana Group types (Bahri Smoothed, Dratmuir Burnished, Socodad Wiped, Brindetta Red-topped and Butana Combed) comprise 95 percent of the sherd sample at KG5a. The largest type at this site is Bahri Smoothed (n=667) which constitutes 53 percent of the sherd sample. The relative frequencies of this type begin at 44 percent in level five and increase to 63 percent in level one. There is a monotonic pattern (within the 2 standard error margin) for the Bahri Smoothed type at KG5a. When the relative frequencies of this type are compared with those of KG23c, the very upper levels of KG23c (within level group 1) fall slightly short of the higher relative frequencies at KG5a. However, within the 2 standard error margins at both sites, levels five to three at KG5a correspond to level group one at KG23c.

Thus, in the long run it is better to combine all levels of KG96a into one unit and seriate it against the four level groups at KG23c using the proportions of the 5 principal ceramic types. Figure 4.8 gives the relative position of KG96a in respect to the four level groups at KG23c using the principal component application. In observing the principal component seriation of KG96a, the relative age of the deposit there would fit best between level groups 1 and 2 of KG23c, thus indicating that the occupation at KG96a was relatively late. This is reinforced by observing the 5 principal ceramic type relative frequencies among KG96a and the four level groups at KG23c in Table 4.14.

The second most frequent type at KG5a is Dratmuir Burnished (n=263) which represents 21 percent of the sherd sample. The relative frequencies of this type appear to increase slightly from 25 to 29 percent from levels five to four then decrease steadily to 11 percent from levels four to one. When the 2 standard error margins are taken into account, a good monotonic pattern for this type emerges. When the relative frequencies of this type are compared with those of KG23c, all levels at KG5a correspond best with level group 1 at KG23c where the relative frequencies range from 19 to 24 percent.

As it can be seen, the relative frequency values of KG96a fit with the overall monotonic patterns of the five types in relation to the four level groups at KG23c. This is consistent with the individual matching of each of the type relative frequencies through the levels at KG96a with respect to the correspond level group(s) at KG23c. Both the Socodad Wiped and Dratmuir Burnished relative frequencies are a little too high and low respectively. Nevertheless, when the two standard error margins are taken into consideration, these particular values are not too far out of line.

The third largest type at KG5a is Socodad Wiped (n=166) which represents 13 percent of the sherd sample. The relative frequencies of this type fluctuate around 13 percent, and when the 2 standard error margins are taken into account, appears to stay constant. When these relative frequencies are compared to the same type at KG23c, they could conceivably fit within a number of levels at the latter site. However, since the first two largest types at KG5a correspond well with level group 1 at KG23c, it is likely

KG5a After KG96a, KG5a contains the next largest sample (n=1256) of Butana Group sherds. KG5a contains all 10

53

The Butana Group Ceramics

Figure 4.8. Seriation of the 4 level groups at KG23c and the collapsed deposit at KG96a based on the relative frequencies of the five major ceramic types.

Table 4.14. Relative Frequencies and 2 Standard Error Margins of the Five Major Butana Group Types Among the 4 Level Groups at KG23c and the Collapsed Deposit at KG96a

Level Group 1 KG96a Level Group 2 Level Group 3 Level Group 4

B. R-tppd % (± 2 s.e.)

B. Smthd. % (± 2 s.e.)

S. Wpd. % (± 2 s.e.)

D. Brnshd. % (± 2 s.e.)

B. Cmbd. % (± 2 s.e.)

14.0 (±1.6) 15.3 (±1.8) 22.1 (±2.5) 32.5 (±2.7) 35.8 (±2.1)

34.9 (±1.9) 35.4 (±2.4) 26.9 (±2.7) 16.0 (±2.1) 11.1 (±1.3)

20.2 (±1.6) 24.5 (±2.2) 21.1 (±2.5) 26.1 (±2.5) 18.8 (±1.7)

23.7 (±1.7) 17.4 (±1.9) 18.5 (±2.4) 10.8 (±1.8) 9.9 (±1.3)

7.2 (±1.0) 7.5 (±1.3) 11.4 (±1.9) 14.7 (±2.0) 24.5 (±1.8)

54

total % 100 100 100 100 100

Definition of the Malawiya/Butana Transition and Butana Group

Table 4.15. Frequencies of the Ceramic Types Through the Levels at KG5a level 1 2 3 4 5 total

B. R-tpped. freq. 28 9 16 13 12 78

B. Smthd. freq. 202 152 122 101 90 667

S. D. Wpd. Brnshd. freq. freq. 41 29 45 25 26 166

35 40 74 62 52 263

B. Cmbd. freq.

A. Pncttd. freq.

6 7 10 6 19 48

1 3 1 0 0 5

K. Pln. freq. 2 2 1 2 1 8

D.C. Z. Mrkd. Brnshd. freq. freq. 4 3 3 1 3 14

0 0 0 0 1 1

S. Dnttd. freq. 0 1 3 2 0 6

S. Rp. Mrkd. freq. 0 0 0 0 0 0

total freq. 319 246 275 212 204 1256

Table 4.16. Relative Frequencies and 2 Standard Error Margins of the Ceramic Types Through the Levels at KG5a

level

1 2 3 4 5

B. R-tpped. % (±2s.e.)

B. Smthd. % (±2s.e.).

S. Wpd. % (±2s.e.)

D. Brnshd. % (±2s.e.)

B. Cmbd. % (±2s.e.)

A. Pncttd. % (±2s.e.)

K. Pln. % (±2s.e.)

D.C. Mrkd. % (±2s.e.)

8.8 (±3.2) 3.7 (±2.4) 5.8 (±2.8) 6.1 (±3.3) 5.9 (±3.3)

63.3 (±5.4) 61.8 (±6.2) 44.4 (±6.0) 47.6 (±6.9) 44.1 (±7.0)

12.9 (±3.7) 11.8 (±4.1) 16.4 (±4.5) 11.8 (±4.4) 12.7 (±4.7)

11.0 (±3.5) 16.3 (±4.7) 26.9 (±5.3) 29.2 (±6.2) 25.5 (±6.1)

1.9 (±1.5) 2.8 (±2.1) 3.6 (±2.3) 2.8 (±2.3) 9.3 (±4.1)

10.3 (±0.6) 1.2 (±1.4) 0.4 (±0.7) 0.0

0.6 (±0.9) 0.8 (±1.1) 0.4 (±0.7) 0.9 (±1.3) 0.5 (±1.0)

1.3 (±1.2) 1.2 (±1.4) 1.1 (±1.3) 0.5 (±0.9) 1.5 (±1.7)

0.0

Z. Brnshd. % (±2s.e.)

S. Dnttd. % (±2s.e.)

S. Rp. Mrkd. % (±2s.e.)

0.0

0.0

0.0

0.0

0.4 (±0.8) 1.1 (±1.3) 0.9 (±1.3) 0.0

0.0

0.0 0.0 0.5 (±1.0)

0.0 0.0 0.0

total % 100 100 100 100 100

The most numerous of the minor types is Dratmuira Chatter Marked (n=14) which accounts for 1 percent of the total sherd sample. The sample size is too small to indicate any pattern with the relative frequencies through the levels. This type occurs in all 5 levels at KG5a and the relative frequencies correspond to those of the same type at KG23c.

that the Socodad relative frequencies at KG5a also fit best with this level group as well. After, Socodad Wiped, the next largest type is Brindetta Red-topped (n=78), which accounts for 6 percent of the sherd sample at KG5a. Like the Socodad Wiped type, the relative frequencies of the Brindetta type stay pretty much at a constant (within 2 standard errors) fluctuating around 6 percent. The relative frequencies of this type in comparison to KG23c, would correspond best with level group 1 at the later site.

The next most numerous minor type is Khordhag Plain (n=8), which constitutes a little more than 1/2 of 1 percent of the typed sherd sample. The sample size of this type is also too small for any kind of pattern recognition with its relative frequencies. Nevertheless, this type occurs in all five levels at KG5a. There appears to be slightly less of this type at KG5a when compared with KG23c where the relative frequencies of the same type averages 2 percent.

The Butana Combed type (n=48) represents the smallest of the 5 main Butana Group types at KG5a and constitutes 4 percent of the sherd sample. From level five to four, the relative frequencies of Butana Combed decrease from 9 to 3 percent and stay at a constant (with 2 standard errors) through the remaining levels. When the Butana Combed relative frequencies are compared between KG5a and KG23c, they correspond best within level group 1 at KG23c were the relative frequency of the same type is at 7 percent.

After Khordhag Plain, there is the Sirjino Simple Dentated type (n=6) which accounts for less than 1 percent of the total sherd sample. The sample size of this type is small and as a result does not appear in either level one or five. Like KG23c, this type is rare at KG5a. After Sirjino Simple Dentated, there is the Alabaska Punctated type (n=5) which also accounts for less than 1 percent of the total sherd sample. The relative frequencies of this type are much smaller at KG5a when compared to KG23c where the relative frequencies of the same type average 3 percent.

The remaining 5 minor types at KG5a consist of the Dratmuira Chatter Marked, Khordhag Plain, Sirjino Simple Dentated, Alabaska Punctated, and Zirhalla Pattern Burnished. These 5 minor types constitute the remaining 5 percent of the sherd sample at KG5a.

55

The Butana Group Ceramics

Figure 4.9. Seriation of the five levels at KG5a based on the relative frequencies of the five major ceramic types.

Only one sherd belonging to the Zirhalla Patterned Burnished type is from KG5a and this sherd was found in level five. Like KG23c, this type is also rare at KG5a.

Based on the overall monotonic patterns of the 5 major types, KG5a fits best at the top, indicating that the occupation at this site was later than the occupation represented by the level group 1 at KG23c. This is consistent with the individual matching of each of the type relative frequencies through the levels at KG5a in respect to level group 1 at KG23c.

When the five levels at KG5a are seriated by the principal component application using the five principal Butana Group types, there is no significant arrangement of the levels along the horizontal scale as shown in Figure 4.9. Given that the deposit at KG5a is shallow, it is likely that the five levels at KG5a represent a short occupation, and as a result, one would not expect to see any significant temporal change among the Butana Group types overall.

KGla KGla has the next largest sample of Butana Group sherds (n=1019) and contains 7 of the 10 Butana Group types found at KG23c, as well the Sarb’ut Ripple Marked type. The three missing Butana Group types are Alabaska Punctated, Khordhag Plain, and Zirhalla Pattern Burnished. KG1 a has a 40 cm deposit and was condensed into four, 10 cm levels. The frequencies, relative frequencies and the corresponding 2 standard errors for the 8 types at KG5a can be observed through the four levels in Tables 4.18 and Figure 4.19.

When the relative frequencies of the five major types at KG5a are collapsed into a single unit and seriated against to the four level groups at KG23c, the relative positioning of this site can be observed in Figure 4.10. As it can be seen, the relative position of KG5a suggest that it corresponds with the latest part of the deposit (associated with level group 1) at KG23c. Table 4.17 shows the proportions of the five principal Butana Group types at KG5a in comparison with the 4 level group at KG23c which reinforces the serration shown in Figure 4.10.

Five Butana Group types make up 97 percent of the sherd sample at KG1a. The largest type is Bahri Smoothed (n=499) which constitutes 49 percent of the sherd sample. In level

56

Definition of the Malawiya/Butana Transition and Butana Group

Figure 4.10. Seriation of the five levels at KG5a based on the relative frequencies of the five major ceramic types.

Table 4.17. Relative Frequencies and 2 Standard Error Margins of the Five Major Butana Group Types Among the 4 Level Groups at KG23c and the Collapsed Deposit at KG5a

KG5a Level Group 1 Level Group 2 Level Group 3 Level Group 4

B. R-tppd. % (± 2 s.e.)

B. Smthd. % (± 2 s.e.)

S. Wpd. % (± 2 s.e.)

D. Brnshd. % (± 2 s.e.)

B. Cmbd. % (± 2 s.e.)

6.4 (±1.4) 14.0 (±1.6) 22.1 (±2.5) 32.5 (±2.7) 35.8 (±2.1)

54.6 (±2.8) 34.9 (±1.5) 26.9 (±2.7) 16.0 (±2.1) 11.1 (±1.3)

13.6 (±2.0) 20.2 (±1.4) 21.1 (±2.5) 26.1 (±2.5) 18.8 (±1.7)

21.5 (±2.4) 23.7 (±1.3) 18.5 (±2.4) 10.8 (±1.8) 9.9 (±1.3)

3.9 (±1.1) 7.2 (±1.0) 11.4 (±1.9) 14.7 (±2.0) 24.5 (±1.8)

57

total % 100 100 100 100 100

The Butana Group Ceramics

Table 4.18. Frequencies of the Ceramic Types Through the Levels at KG1a level 1 2 3 4 total

B. R-tpped. freq. 68 48 24 5 145

B. Smthd. freq. 344 78 63 14 499

S. D. Wpd. Brnshd. freq. freq. 65 13 10 5 93

76 25 23 6 130

B. Cmbd. freq.

A. Pncttd. freq.

82 24 19 1 126

0 0 0 0 0

K. Pln. freq. 0 0 0 0 0

D.C. Z. Mrkd. Brnshd. freq. freq. 14 4 0 0 18

0 0 0 0 0

S. Dnttd. freq. 4 3 0 0 7

S.Rp. Mrkd. freq. 0 1 0 0 1

total freq. 653 196 139 31 1019

The fifth largest type at KG1a is represented by Socodad Wiped (n = 93) which constitutes 9 percent of the sherd sample. The relative frequencies of this type decline from 16 percent at level four, to 7 percent in level three. From level two to one, the relative frequencies increase from 7 to 10 percent. The standard errors in relation to the relative frequencies are very large, thus, no monotonic pattern through the levels can be discerned. When the Socodad relative frequencies at KG5a are compared with KG23c, no conclusive positioning of the KG5a Socodad Wiped relative frequencies can be made with the level groups at KG23c. There is a lower incidence of this type at KGla than at KG23c.

four and three, the relative frequency for this type is 45 percent. In level two, the relative frequency declines to 40 percent and then increases to 53 percent at level one. The 2 standard errors margins indicate a wide variance in the relative frequencies of this type, and as a result, does not show any temporal patterning. In comparison to KG23c, the Bahri Smoothed relative frequencies at KGla correspond best with level group 1 at the former site where the relative frequency of the same type is at 34 percent. The second most numerous type is Brindetta Red-topped (n = 145) which accounts for 14 percent of the sherd sample. From levels four to three, the relative frequencies of this type increase slightly from 16 to 17 percent. In level two, the relative frequency value increases to 24 percent and then drops down to 10 percent at level one. This type may show a monotonic pattern through the levels at KGla, however, when the relative frequencies are bracketed within their respective 2 standard error margins, the pattern dissipates. In comparison to KG23c, the relative frequencies of this type at KG5a correspond best with level group 1 were the relative frequency of the same type is at 14 percent.

The remaining 3 minor types (Dratmuir Chatter Marked, Sirjino Simple Dentated, and Sarb’ut Ripple Marked) consist of only 3 percent of the sherd sample at KG1a. Dratmuira Chatter Marked (n = 18) is the most numerous of the minor types and constitutes a little less than 2 percent of the sherd sample. There are no Dratmuira sherds present in levels four or three at KG1a. In levels two and one, the relative frequencies of this type are both at 2 percent. In comparison with KG23c, the relative frequencies of this type at KGla fit best with level group 1 of the former site where the relative frequencies are at 2 %.

After Brindetta Red-topped, the third largest type at KGla is Dratmuir Burnished (n = 130) which constitutes 13 percent of the sherd sample. Overall, the relative frequencies of this type decrease from 19 percent in level four, to 12 percent in level one. However, within 2 standard error margins, the probable range of the relative frequencies for this type is very high, nullifying any monotonic patterning seen through the levels at KG1a. When the Dratmuir Burnished relative frequencies are compared between KG23c and KG1a, the relative frequencies at KGla fit best within level groups 2 and 3 where the relative frequencies of the same type range between 11 and 19 percent.

After Dratmuira Chatter Marked, is Sirjino Simple Dentated (n = 7) which accounts for a little over 1/2 of 1 percent of the sherd sample at KG1a. Like Dratmuira Chatter Marked, this type is not present in level four or three at KG1a. In levels two and one, the relative frequencies are low being 1 and 1/2 of 1 percent respectively. This type is very rare at KG1a as it is at KG23c. There was one Sarb’ut Ripple sherd at KG1a which was located in level two. When the proportions of the five major Butana Group types at KGla are seriated using the principal component application, no horizontal ordering of the levels is apparent, as can be seen in Figure 4.11. All four levels are situated between .9 and 1 on the horizontal scale of Factor 1. As at KG1a, the lack of horizontal separation of the 4 levels at KG1a suggests little time depth at this site.

The fourth largest type at KG1a is Butana Combed (n = 126) which accounts for 12 percent of the sherd sample. The relative frequency for this type in level four is 3 percent. In level three, it jumps to 13 percent, declines slightly to 12 percent in level two, then increases back to 13 percent in level one. When the 2 standard errors are taken into consideration, it appears that this type stays pretty much at a constant - with the exception of level four. In comparison to KG23c, the relative frequencies of this type at KG1a correspond with level group 3 at the former site where the relative frequencies are 15 percent.

When the levels at KGla are combined into one unit and the proportions of the 5 principal ceramic types are seriated with level groups 1 through 4 at KG23c (using the principal component application) the former site is located between level group 1 and 2 as shown in Figure 4.12. As it can be seen in table 4.20, when the relative frequencies of the five major Butana Group types at KGla are compared with the four

58

Definition of the Malawiya/Butana Transition and Butana Group

Figure 4.11. Seriation of the 4 levels at KG1a based on the relative frequencies of the five major ceramic types.

types found at KG23c. The types Alabaska Punctated and Sarb’ut Ripple Marked are not present. KG29N consists of a 50 cm deposit which was condensed into five, 10 cm levels. The frequencies, relative frequencies and corresponding 2 standard errors for the 8 types through levels 1 through 5 are provided in Tables 4.21 and 4.22.

level groups at KG23c, their placement between level group 1 and 2, is the best fit. This is consistent with the individual matching of most of the type relative frequencies through the levels at KGla in respect to the correspond level groups at KG23c. However, as it can be seen, some of the type relative frequencies at KGla do not fall within the overall monotonic patterns produced by the four level groups at KG23c. Even when the 2 standard error margins are applied, the Bahri Smoothed type relative frequency value are too high, while the Socodad Wiped and Dratmuir Burnished type relative frequency values are too low.

KG29N

Three of the five main Butana Group types (Bahri Smoothed, Socodad Wiped, and Dratmuir Burnished) comprise 88 percent of the sherd sample at KG29N. The most numerous type at KG29N consists of the Bahri Smoothed type (n = 225) which accounts for 43 percent of the sherd sample. The relative frequencies of this type are very erratic through the levels where they oscillate between 21 and 57 percent in levels five and four, 33 and 50 percent in levels three and two, and 39 percent at level one. Overall, there seems to be no patterning of this type through the levels. When the Bahri Smoothed relative frequencies are compared to those at KG23c, they correlate best with level group 1 at KG23c where the relative frequency for the same type is 34 percent.

KG29N contains the smallest number of sherds (n = 517) within the Butana Group sites and possesses 9 of the 10 Butana Group

The second most popular type at KG29N is Socodad Wiped (n= 133) which accounts for 26 percent of the sherd sample.

The anomalous relative frequencies associated with the Bahri Smoothed, Socodad Wiped, and Dratmuir Burnished types at KG1a do not appear to be a result of differential vessel portion deposition. Since three other common Butana Group ceramic types are missing from this site, this probably indicates that KGla is different in some way (culturally or functionally) from other Butana Group sites such as KG7 and KG23.

59

The Butana Group Ceramics

Figure 4.12. Seriation of the 4 level groups at KG23c and the collapsed deposit at KG1a based on the relative frequencies of the five major ceramic types.

The Socodad Wiped relative frequencies begin at 33 percent at level five and decline to 25 percent in level four. From level four to one, the relative frequencies fluctuating between 27 and 22 percent. When the relative frequencies of this type are bracketed within their 2 standard error margin, the probable range of variability is quite wide. The large standard errors would nullify any pattern produced by the relative frequencies for this type through the levels at KG29N. Given the high relative frequencies of the Bahri Smoothed type and its probable correspondence to the upper levels at KG23c, the best correlation of the Socodad Wiped relative frequencies at KG29N would also be with level group 1 at KG23c.

frequencies fit best with level group 2 at KG23c where the relative frequency of the same type is at 24 percent. The remaining six types; Dratmuira Chatter Marked, Butana Combed, Sirjino Simple Dentated, Zirhalla Patterned Burnished, Brindetta Red-topped, and Khordhag Plain, constitute 12 percent of the sherd sample at KG29N. Of these six types, Dratmuira Chatter Marked (n=21) is the most numerous and comprises 4 percent of the sherd sample at KG29N. There are no Dratmuira Chatter Marked sherds present in level five at KG29N. In level four, the relative proportions of this type begin at 1 percent and increase to 20 percent in level three. From level three, the relative proportions decline to 2 percent in level one. Arguably, this particular type displays a good monotonic pattern through the levels at KG29N, however, when the 2 standard error margins are taken into consideration, the monotonic patterning of this type is unconvincing. In comparison to KG23c, the relative frequencies of the Dratmuira type at KG29N are higher and no comparable correlation can be made between the two sites. In general, KG29N registers the highest relative frequencies for the Dratmuira Chatter Marked type among the seven Butana Group sites.

The third most numerous type at KG29N is Dratmuir Burnished (n = 98) which comprises 19 percent of the sherd sample. The relative frequencies of the Dratmuir Burnished type decline sharply from 45 percent at level five to 16 percent at level four. From level four on, the relative frequencies fluctuate from 13 to 19 percent. The 2 standard error margins are very wide and this type does not display any kind of convincing monotonic pattern. When the relative frequencies of this type at KG29N are compared with the same type at KG23c, the former relative

60

Definition of the Malawiya/Butana Transition and Butana Group

Table 4.19. Relative Frequencies and 2 Standard Error Margins of the Ceramic Types Through the Levels at KG1a

level

1 2 3 4

B. R-tpped. % (±2s.e.)

B. Smthd. % (±2s.e.).

S. Wpd. % (±2s.e.)

D. Brnshd. % (±2s.e.)

B. Cmbd. % (±2s.e.)

10.4 (±2.4) 24.5 (±6.1) 17.3 (±6.4) 16.1 (±13.2)

52.7 (±3.9) 39.8 (±7.0) 45.3 (±8.4) 45.2 (±17.9)

10.0 (±2.3) 6.6 (±3.6) 7.2 (±4.4) 16.1 (±13.2)

11.6 (±2.5) 12.8 (±4.8) 16.5 (±6.3) 19.4 (±14.2)

12.6 (±2.6) 12.2 (±4.7) 13.7 (±5.8) 3.2 (±6.3)

A. Pncttd. % (±2s.e.)

K. Pln. % (±2s.e.)

0.0

0.0

0.0

0.0

0.0 0.0

D.C. Mrkd. % (±2s.e.)

0.0

2.1 (±1.1) 2.0 (±1.1) 0.0

0.0

0.0

Z. Brnshd. % (±2s.e.)

S. Dnttd. % (±2s.e.)

0.0

S. Rp. Mrkd. % (±2s.e.)

0.0

0.6 (±0.6) 1.5 (±1.8) 0.0

0.5 (±1.0) 0.0

0.0

0.0

0.0

0.0

0.0

total % 100 100 100 100

Table 4.20. Relative Frequencies and 2 Standard Error Margins of the Five Major Butana Group Types Among the 4 Level Groups at KG23c and the Collapsed Deposit at KG1a

Level Group 1 KG1a Level Group 2 Level Group 3 Level Group 4

B. R-tppd. % (± 2 s.e.)

B. Smthd. % (± 2 s.e.)

S. Wpd. % (± 2 s.e.)

D. Brnshd. % (± 2 s.e.)

B. Cmbd. % (± 2 s.e.)

14.0 (±1.6) 14.6 (±2.2) 22.1 (±2.5) 32.5 (±2.7) 35.8 (±2.1)

34.9 (±1.9) 50.3 (±3.2) 26.9 (±2.7) 16.0 (±2.1) 11.1 (±1.3)

20.2 (±1.6) 9.4 (±1.8) 21.1 (±2.5) 26.1 (±2.5) 18.8 (±1.7)

23.7 (±1.7) 13.1 (±2.1) 18.5 (±2.4) 10.8 (±1.8) 9.9 (±1.3)

7.2 (±1.0) 12.7 (±2.1) 11.4 (±1.9) 14.7 (±2.0) 24.5 (±1.8)

total % 100 100 100 100 100

Table 4.21. Frequencies of the Ceramic Types Through the Levels at KG29N level 1 2 3 4 5 total

B. R-tpped. freq. 0 1 0 0 0 1

B. Smthd. freq. 63 94 13 43 12 225

S. D. Wpd. Brnshd. freq. freq. 36 50 10 19 18 133

31 24 6 12 25 98

B. Cmbd. freq.

A. Pncttd. freq.

14 1 1 1 0 17

0 0 0 0 0 0

K. Pln. freq. 0 1 0 0 0 1

D.C. Z. Mrkd. Brnshd. freq. freq. 4 8 8 1 0 21

2 4 1 0 0 7

S. Dnttd. freq. 12 2 0 0 0 14

S. Rp. Mrkd. freq. 0 0 0 0 0 0

total freq. 162 185 39 76 55 517

Table 4.22. Relative Frequencies and 2 Standard Error Margins of the Ceramic Types Through the Levels at KG29N

level

1 2 3 4 5

B. R-tpped. % (±2s.e.) 0.0 0.5 (±1.1) 0.0 0.0 0.0

B. Smthd. % (±2s.e.).

S. Wpd. % (±2s.e.)

D. Brnshd. % (±2s.e.)

B. Cmbd. % (±2s.e.)

38.9 (±7.7) 50.8 (±7.4) 33.3 (±15.1) 56.6 (±11.4) 21.8 (±11.1)

22.2 (±6.5) 27.0 (±6.5) 25.6 (±14.0) 25.0 (±9.9) 32.7 (±12.7)

19.1 (±6.2) 13.0 (±4.9) 15.4 (±15.8) 15.8 (±8.4) 45.5 (±13.4)

8.6 (±4.4) 0.5 (±1.1) 2.6 (±5.1) 1.3 (±2.6) 0.0

A. Pncttd. % (±2s.e.)

K. Pln. % (±2s.e.)

D.C. Mrkd. % (±2s.e.)

Z. Brnshd. % (±2s.e.)

S. Dnttd. % (±2s.e.)

0.0

0.0 (±) 0.5 (±1.1) 0.0

0.0

0.0

1.2 (±1.7) 2.2 (±2.1) 2.6 (±5.1) 0.0

7.4 (±4.1) 1.1 (±1.5) 0.0

0.0

0.0

0.0

0.0

0.0

2.5 (±2.4) 4.3 (±3.0) 20.5 (±12.9) 1.3 (±2.6) 0.0

0.0

0.0

0.0

0.0 0.0

61

S. Rp. Mrkd. % (±2s.e.) 0.0 0.0

total % 100 100 100 100 100

The Butana Group Ceramics

Figure 4.13. Seriation of the 5 levels at KG29N based on the relative frequencies of the five major ceramic types.

After Sirjino Simple Dentated, is Zirhalla Pattern Burnished (n = 7) which accounts for 2 percent of the total sherd sample. There are no Zirhalla Pattern Burnished sherds in either levels four or five at KG29N. In level three, the relative frequency for this type is 3 percent. From level two to one, the relative frequencies decline from 2 to 1 percent. When the 2 standard error margins are taken into account, it is apparent that the probable range of variability is great, thus nullifying the pattern of this type through the levels at KG29N.

Butana Combed is the next largest type ( n = 17) and accounts for 3 percent of the sherd sample. Butana Combed sherds are not present in level five at KG29N. In level four, the relative frequency for this type is 1 percent and it stays more or less the same until level one where it jumps to 9 percent. Overall, there appears to be no definitive monotonic pattern for the Butana type at KG29N. In comparison to KG23c, the Butana Combed relative frequencies at KG29N fit best with level group 1 at the former site. In comparison to the other Butana Group sites, KG29N has the smallest number of Butana Combed sherds.

There is only one Brindetta Red-topped type sherd and one Khordhag Plain sherd at this site. Like KG29S and KG1a, KG29N has virtually no Brindetta Red-topped, Alabaska Punctated, or Khordhag Plain type sherds.

There are 14 Sirjino Simple Dentated sherds at KG29N which comprises 3 percent of the sherd sample. There are no sherds of this type present in levels three through five. In level two, the relative frequency for this type is 1 percent, then increases to 7 percent in level one. The monotonic patterning of this type through the levels is quite distinctive, however, outside the absence of this type in the lower levels, the large 2 standard errors in both levels one and two indicate that this type does not really increase through the levels at KG29N. In comparison to KG23c, the relative frequencies of this type are higher at KG29N than they are at the former site.

When the 5 levels at KG29N are seriated with the principal component application (using the relative frequencies of the five major Butana Group types) there is no convincing stratigraphic arrangement of the levels as can be observed in Figure 4.13. Like the deposits at KG96a, KG5a, and KG1a, the lack of a good seriation among the levels at KG29N, plus the

62

Definition of the Malawiya/Butana Transition and Butana Group

Figure 4.14. Seriation of the 4 level groups at KG23c and the collapsed deposit at KG29N based on the relative frequencies of the five major ceramic types.

Table 4.23. Relative Frequencies and 2 Standard Error Margins of the Five Major Butana Group Types Among the 4 Level Groups at KG23c and the Collapsed Deposit at KG29N

KG29N Level Group 1 Level Group 2 Level Group 3 Level Group 4

B. R-tppd. % (± 2 s.e.)

B. Smthd. % (± 2 s.e.)

S. Wpd. % (± 2 s.e.)

D. Brnshd. % (± 2 s.e.)

B. Cmbd. % (± 2 s.e.)

0.2 (±.05) 14.0 (±1.6) 22.1 (±2.5) 32.5 (±2.7) 35.8 (±2.1)

47.5 (±1.5) 34.9 (±1.5) 26.9 (±2.7) 16.0 (±2.1) 11.1 (±1.3)

28.1 (±1.8) 20.2 (±1.4) 21.1 (±2.5) 26.1 (±2.5) 18.8 (±1.7)

20.7 (±1.9) 23.7 (±1.3) 18.5 (±2.4) 10.8 (±1.8) 9.9 (±1.3)

3.6 (±1.4) 7.2 (±1.0) 11.4 (±1.9) 14.7 (±2.0) 24.5 (±1.8)

63

total % 100 100 100 100 100

The Butana Group Ceramics

Table 4.24. Frequencies of the Ceramic Types Through the Levels at KG28B level

B. Cmbd. freq.

Bdhh. Smthd. freq.

1 2 3 total

312 991 677 1980

280 835 569 1684

N. Brnshd. freq. 42 256 227 525

S. Rp. Mrkd. freq.

S. Wpd. freq.

45 131 168 344

3 2 1 6

D.C. Mrkd. freq. 4 18 34 56

S. Dnttd. freq.

T.R. Stmpd. freq.

0 1 0 1

30 72 58 160

total freq. 716 2306 1734 4756

Table 4.25. Relative Frequencies and 2 Standard Error Margins of the Ceramic Types Through the Levels at KG28B

level

1 2 3

B. Cmbd. % (± 2 s.e.)

Bdhh. Smthd. % (± 2 s.e.)

43.6 (±3.7) 43.0 (±2.1) 39.0 (±2.3)

39.1 (±3.6) 36.2 (±2.0) 32.8 (±2.3)

N. S. Rp. Brnshd. Mrkd. % % (± 2 s.e.) (± 2 s.e.) 5.9 (±1.8) 11.1 (±1.3) 13.1 (±1.6)

S. D.C. Wpd. Mrkd. % % (± 2 s.e.) (± 2 s.e.)

6.3 (±1.8) 5.7 (±5.7) 9.7 (±9.7)

0.4 (±0.5) 0.1 (±0.1) 0.1 (±0.1)

0.6 (±0.6) 0.8 (±0.4) 2.0 (±0.7)

S. Dnttd. % (± 2 s.e.) 0.0 0.1 (±0.1) 0.0

T.R. Stmpd. % (± 2 s.e.) 4.2 (±1.5) 3.1 (±0.7) 3.3 (±0.9)

total freq. 100 100 100

Five major types (Butana Combed, Budahah Smoothed, Nulmswi Burnished, Sarb’ut Ripple Marked, and Saroba Combed Impressed) constitute 99 percent of the total sherd sample at KG28B. Butana Combed is the most numerous type (n = 1980) accounting for 42 percent of the sherd sample. The relative frequency of this type is 39 percent in level three and increases to 43 percent in level two. At level one, the relative frequency value increases slightly to 44 percent. With the 2 standard errors, the relative frequencies of this type do not vary significantly through the levels at KG28B. The relative frequencies values associated with the Butana Combed type at KG28B are larger than at KG23c, where they never exceed over 34 percent.

narrow horizontal spread of the levels along the horizontal axis, indicates that the occupation was of a short duration, suggesting that this site was inhabited for only a short period. When the levels at KG29N are combined and seriated (using the principal component application) with the four level groups at KG23c, the following temporal positioning of the deposit at KG29N can be seen in Figure 4.14. Based on the principal component seriation, the temporal positioning of KG29N is located very near level group 1 at KG23c. When the relative frequencies of the 5 principal ceramic types are compared in Table 4.23, it appears that KG29N would be slightly latter than level group 1 at KG23c. This would also be consistent with the individual matching of many of the type relative frequencies through the levels at KG29N compared to level group 1 at KG23c. One should note, however, that the Socodad Wiped type relative frequency value associated with KG29N is too high and does not conform to the expected monotonic trend even within a 2 standard error margin. The Dratmuir Burnished relative frequency value at KG29N also appears to be too low; however, given the 2 standard error margin, it would fall into the general range given for level group 1 at KG23c.

The next most populous type at KG28B is Budahah Smoothed (n = 1684) which constitutes 36 percent of the sherd sample. The relative frequencies of this particular type begin at 33 percent in level three and increase steadily through level one where they end at 39 percent. When the 2 standard errors are considered, the relative frequencies do not appear to vary significantly. The Budahah Smoothed type is considered to belong to the Malawiya/Butana transition and is not associated with any of the seven Butana Group sites. After Budahah Smoothed, the next most numerous type is Nulmswi Burnished (n = 525) which accounts for 11 percent of the sherd sample. The relative frequency of this particular type starts at 13 percent in level three and steadily decreases to 6 percent in level one. Overall, the relative frequencies of this type decreases through the levels, however, when the 2 standard error margins are considered this pattern is not significant. Like Budahah Smoothed, Nulmswi Burnished is considered to be a Saroba/Butana transition type and is not associated with any of the Butana Group sites.

KG28B KG28B represents the assemblage of Malawiya/Butana transition ceramics used in this study. It has 4756 typed sherds out of a deposit of 30 cm and was condensed into three 10 cm levels. Tables 4.24 and Figure 4.25 show the frequencies, relative frequencies and the 2 standard errors of the types through the three levels of KG28B.

64

Definition of the Malawiya/Butana Transition and Butana Group

Figure 4.15. Seriation of the 3 levels at KG28B based on the relative frequencies of the three major ceramic types.

The next largest sample of sherds Sarb’ut Ripple Marked (n = 344) which constitutes 7 percent of the sherd sample at KG28B. The relative frequencies of this type begin in level three at 10 percent and decrease down to 6 percent in levels two and one. When the 2 standard errors are taken into account, there is a slight monotonic pattern. The Sarb’ut Ripple Marked type is considered to be a Saroba/ Butana transition type, however, it does occur in very small quantities (less than 1/2 of 1 percent ) at all but two (KG5a and KG29N) of the seven Butana Groups sites.

represent only 1 percent of the sherd sample at KG28B. The largest of these is the Dratmuira Chatter Marked type (n = 56) which consists of almost 1 percent of the total sherd sample. The relative frequencies of this type decrease from 2 percent to less than 1 percent from level three to two. In level one, the relative frequency decreases to 1/2 of 1 percent. When the 2 standard errors are taken into consideration, the decreasing trend of this type appears to be is artificial. The Dratmuira Chatter Marked type is considered to be associated with the Butana Group, and its presence at KG28B is somewhat problematic. Therefore, it is likely that the Dratmuira Chatter Marked sherds at KG28B are associated with a latter Butana Group occupation which was light. Since this type is present throughout the three levels at KG28B, it is likely that the deposit associated with the Malawiya/Butana transition is disturbed and possibly mixed with some latter Butana Group ceramics.

Saroba Comb Impressed is the fifth largest type (n = 160) accounting for 3 percent of the sherd sample at KG28B. In levels 3 and 2, the relative frequency of this type is 3 percent. In level one the relative frequency of this type increases to 4 percent. When the two standard errors are taken into account, the pattern of the relative frequencies through the levels is neutralized. Sherds belonging to the Saroba Comb Impressed type are not found at any of the Butana Group sites, and the type is considered to be associated only with the Malawiya/Butana transition.

There are 6 Socodad Wiped sherds (accounting for 1% of the total sherd sample) from KG28B. Socodad Wiped is also considered to belong to the Butana Group and their presence help to reinforce a later occupation (albeit slight) at KG28B. There are 3 sherds of this type from level one, 2 from level two, and 1 from level one.

The remaining three types (Dratmuira Chatter Marked, Socodad Wiped, and Sirjino Simple Dentated) at KG28B

65

The Butana Group Ceramics

Table 4.26. Relative Frequencies of the Butana Combed and Sarb’ut Ripple Marked Types in Relation to the Remaining Types at KG28B and the 4 Level Groups at KG23c B. Cmbd. %

S. Rp. Mrkd. %

remaining types %

total %

Level Group 1

6.7

0.0

93.3

100

Level Group 2

10.5

0.0

89.5

100

Level Group 3

13.5

0.0

86.5

100

Level Group 4

22.8

0.0

77.1

100

KG28B

41.6

7.2

51.1

100

Definition of the Malawiya/Butana Transition and Butana Group Based on the Ceramic Types and Their Temporal Significance

One Sirjino Simple Dentated sherd was also recovered at KG28B from level two. Like Dratmuira Chatter Marked and Socodad Wiped, the Sirjino Simple Dentated type is considered to belong to the Butana Group.

The Malawiya/Butana Transition Ceramics

When the three levels at KG28B are seriated (using the three most numerous types in a principal component application), there is no horizontal separation among them as can be seen in Figure 4.15. The lack of horizontal separation observed with the levels suggests that there is no stratigraphic integrity at this site.

Based on the data provided above, the Malawiya/Butana transition can be defined by five principal ceramics types: Butana Combed, Budahah Smoothed, Nulmswi Burnished, Sarb’ut Ripple Marked, and Saroba Comb Impressed. Since no Malawiya/Butana transition sites have been found other than KG28B, it is assumed that other sites or occupations associated with this particular transitional phase would always include these five principal types. Nevertheless, further archaeological investigations in the southern Atbai need to be done to test this assumption.

Since KG28B contains only one of the five predominant Butana Group ceramic types (Butana Combed) a seriation of this site in conjunction with the four level groups at KG23c was not implemented. However, the single radiocarbon date of 4018 BC ± 67 recovered from KG28B (cf. Marks and Sadr 1988: 73) indicates that this site is earlier than the other seven Butana Group sites. As shown in Table 4.26, the significantly higher relative frequencies of the Butana Combed and Sarb’ut Ripple Marked types at KG28B (in relation to the total number of typed sherds at each site) would also place this site in a temporal position prior to level group 4 at KG23c, which would in turn reinforce the early radiocarbon date from the former site.

As far as determining the temporal patterning of the five Malawiya/Butana Group types, nothing can be said about their relative stratigraphic positioning at KG28B. However, since both the Saroba Comb Impressed and Budahah Smoothed types are present in the Malawiya Group (cf. Hays 1971; Mohammed-Ali 1989) it is logical to assume that these are the earliest types associated with the Malawiya/ Butana transition. Butana Combed, Nulmswi Burnished, and Sarb’ut Ripple Marked are not present in the Malawiya Group ceramics and appear first in the Malawiya/Butana Transition. Of these three types, only Butana Combed persists later into the Butana Group. Thus, the types Nulmswi Burnished and Sarb’ut Ripple Marked appear to be temporally intermediate and are essentially only present during the Malawiya/Butana transition. It is important to note that both Nulmswi Burnished and Sarb’ut Ripple Marked types represent the first exterior burnished ceramics in the southern Atbai.

One should also note that the decrease in the relative frequencies of Butana Combed is quite pronounced from KG28B to level group 4 at KG23c. Given that there is an approximate 474 year gap between the earliest date associated with KG23c (3544 BC ± 154 yrs) and the date associated with KG28B (4018 BC ± 67 yrs) there may have been other earlier Butana Group or latter Malawiya/Butana transition sites in the southern Atbai which have yet to be found. These potential earlier sites would probably have proportions of Butana Combed sherds falling between 22.8 and 41.6 percent of the sherd total. Whether or not these particular sites would be considered as Malawiya/Butana or early Butana is addressed in the following discussion.

The Butana Group The Butana Group consists primarily of ten types, of which five are considered as principal (Butana Combed, Brindetta Red-topped, Bahri Smoothed, Socodad Wiped, and Dratmuir Burnished) and five as minor types (Dratmuira Chatter

66

Definition of the Malawiya/Butana Transition and Butana Group

Table 4.27. Percentage Breakdown of the 10 Primary Butana Group Types Among the Seven Butana Group Sites KG23c %

KG7a %

KG29S %

KG96a %

KG5a %

KG1a %

KG29N %

Butana Combed

13.5

9.8

12.9

5.8

3.8

12.4

3.3

Brindetta Red-topped

23.4

15.0

1.5

11.8

6.2

14.2

0.2

Bahri Smoothed

21.2

27.9

15.8

27.2

53.1

49.0

43.5

Socodad Wiped

19.3

24.2

27.9

18.9

13.2

9.1

25.7

Dratmuir Burnished

15.1

16.5

32.9

13.3

20.9

12.8

18.9

sub total

92.5

93.4

91.0

77.0

97.2

97.5

91.6

KG29S %

KG96a %

KG5a %

KG1a %

minor types KG23c %

KG7a %

KG29N %

Dratmuira Chatter Marked

1.1

1.7

5.0

1.8

1.1

1.8

4.1

Khordhag Plain

2.0

2.0

0.0

19.4

0.6

0.0

0.2

Zirhalla Pattern Burnished

0.5

1.1

1.9

0.5

0.1

0.0

1.3

Sirjino Simple Dentated

0.3

0.4

2.0

0.2

0.5

0.7

2.7

Alabaska Punctated

3.6

1.4

0.1

1.1

0.4

0.0

0.0

sub total

7.5

6.6

9.0

23.0

2.7

2.5

8.3

total

100

100

100

100

100

100

100

Marked, Khordhag Plain, Zirhalla Pattern Burnished, Sirjino Simple Dentate, Alabaska Punctated ). Table 4.27 gives a percentage-wise breakdown of the ten Butana Group types in the seven Butana Group sites used in this study.

also situated on the bank of the Atbara River not far from KG29N and KG29S. Whether or not there exists two minor sub-facies within the Butana Group is a question which cannot be addressed at this time. Nevertheless, one should note that the presence/absence of these minor types may signify these kinds of differences.

With the exception of KG96a, over 90 percent of the ceramics from the other six Butana Group sites are comprised of the five principal types. Based on these data, a Butana Group site or occupation should always contain these particular types, and when totaled should constitute the majority (at least 75 percent of the total number of typed sherds) of the sherds. The remaining five minor types are also present at most of the seven Butana Group sites. Based on the presence/absence of these particular minor types, one can tentatively notice infra-regional connections between sites. For example, sites KG23c, KG7a, KG96a, and KG5a all seem to be closely related to one another due to the fact that all five of the minor types are present at these particular sites. Indeed, three of these sites (KG23c, KG96a, and KG5a) are located in the interior steppe region of the southern Atbai. On the other hand, sites such as KG29S, KG29N, and KG1a seem to be less related due to the fact that these particular sites are lacking some of the minor types. In fact, these particular sites all lack the Alabaska Punctated type which may signify some kind of connection between them. Indeed, KG29S and KG29N are practically adjacent to one another and are the only Butana Group sites so far located on the west side of the Atbara River. KG1a is

The deposit at KG23c displays the best overall sequence of the Butana Group ceramics and a generalized sequence of the Butana Group ceramics can be achieved by using the four level groups at KG23c to determine the basic temporal patterning of the 10 Butana Group ceramic types. In reviewing Table 4.4, the percentages (within a 2 standard error range) of the 10 Butana Group types (including Sarb’ut Ripple Marked) can be observed among the four level groups at KG23c. Since Butana Combed is present in the Malawiya/Butana transition, it is reasonable to assume that it is the earliest type associated with the Butana Group. As discussed earlier, this type (through the four level groups at KG23c) shows a pattern of steady decrease through time. Another early type associated with the Butana Group is Brindetta Red-topped which also shows a pattern of steady decrease through time. However, it is important to note that this type does not appear in the Malawiya/Butana transition. Both Bahri Smoothed and Dratmuir Burnished can be considered as temporally late types within the Butana Group and

67

The Butana Group Ceramics

Figure 4.16. Seriation of the 4 level groups at KG23c and the remaining six Butana Group sites based on the relative frequencies of the five principal Butana Group ceramic types.

show patterns of continued increase through time. Socodad Wiped can be considered an intermediate type which peaks during the time when the Butana Combed and Brindetta Redtopped types are declining and when the Bahri Smoothed and Dratmuir Burnished types are increasing in popularity.

To test the integrity of the temporal patterning of the Butana Group types as indicated in Tables 4.4 and 4.27, one can again seriate the ceramic type percentages from the remaining six Butana Group sites using the principal component application. Since the five principal ceramic types make up the vast majority of the Butana Group ceramics and display the best monotonic patterns, the relative frequencies of these particular types will be used in the principal component application.

As discussed previously, the five minor Butana Group types are too small in number to substantiate whether or not they show true temporal patterning through the four level groups at KG23c. However, using just the percentages and ignoring the 2 standard error margins, Alabaska Punctated may be temporally an intermediary type while both Dratmuira Chatter Marked and Sirjino Simple Dentated are late types in the Butana Group. The types Khordhag Plain and Sirjino Pattern Burnished do not show any clear temporal patterning among the 4 level groups at KG23c.

The order of the seriated four level groups at KG23c in conjunction with the other 6 Butana Group sites will determine the final temporal organization of the Butana Group types based on their respective relative frequencies. Figure 4.16 shows the factored outcome of the four level groups of KG23c and the other 6 Butana Group sites.

68

Definition of the Malawiya/Butana Transition and Butana Group

Table 4.28. Relative Frequencies and 2 Standard Error Margins of the Ten Primary Butana Group Types Among the 4 Level Groups at KG23c and the Other Six Butana Group Sites

KG29N KG5a Level Group 1 KG96a KG7a KG1a Level Group 2 KG29S Level Group 3 Level Group 4

B. Cmbd. % (±2s.e.)

B. R-tpp. % (±2s.e.)

B. S. Smthd. Wpd. % % (±2s.e.) (±2s.e.)

D. Brnsh. % (±2s.e.)

D.C. K. Mrkd. Pln. % % (±2s.e.) (±2s.e.)

Z. P. Brnsh. % (±2s.e.)

S. Dnttd. % (±2s.e.)

3.3 (±1.6) 3.8 (±1.1) 6.7 (±1.0) 5.7 (±1.0) 9.8 (±1.0) 12.4 (±2.1) 10.5 (±1.8) 12.9 (±1.3) 13.5 (±1.9) 22.5 (±1.7)

0.2 (±.04) 6.2 (±1.4) 12.9 (±1.3) 11.8 (±1.4) 15.0 (±1.1) 14.2 (±2.2) 20.3 (±2.3) 1.6 (±0.5) 30.0 (±2.5) 33.3 (±2.0)

43.5 (±4.4) 53.1 (±2.8) 32.3 (±1.8) 27.2 (±2.0) 27.9 (±1.4) 49.0 (±3.1) 24.7 (±2.5) 15.8 (±1.4) 14.8 (±2.0) 10.3 (±1.3)

19.0 (±3.4) 20.9 (±2.3) 21.9 (±1.6) 13.4 (±1.5) 16.5 (±1.2) 12.8 (±2.1) 17.0 (±2.2) 32.9 (±1.8) 10.0 (±1.7) 0.9 (±1.2)

4.1 (±1.7) 1.1 (±0.6) 1.6 (±0.5) 1.8 (±0.6) 1.7 (±0.4) 1.8 (±0.8) 1.2 (±0.6) 5.0 (±0.8) 1.1 (±0.6) 0.6 (±0.3)

0.2 (±0.4) 0.6 (±0.4) 1.6 (±0.5) 19.4 (±1.7) 2.0 (±0.5) 0.0

1.4 (±1.0) 0.1 (±0.2) 0.3 (±0.2) 0.5 (±0.3) 1.1 (±0.3) 0.0

1.8 (±0.8) 0.0

0.8 (±0.5) 1.9 (±0.5) 0.6 (±0.4) 0.4 (±0.3)

2.7 (±1.4) 0.5 (±0.4) 0.6 (±0.3) 0.2 (±0.2) 0.4 (±0.2) 0.7 (±0.5) 0.2 (±0.2) 2.0 (±0.5) 0.3 (±0.3) 0.1 (±0.1)

25.7 (±3.8) 13.2 (±1.9) 18.7 (±1.5) 18.9 (±1.7) 24.2 (±1.4) 9.1 (±1.8) 19.4 (±2.3) 27.9 (±1.7) 24.1 (±2.4) 17.5 (±1.6)

1.2 (±0.7) 2.9 (±0.7)

A. Pnctd. % (±2s.e.) 0.0 0.4 (±0.4) 3.5 (±0.7) 1.1 (±0.5) 1.4 (±0.4) 0.0 4.3 (±1.2) 0.04 (±0.01) 4.4 (±1.1) 2.9 (±0.7)

total % 100 100 100 100 100 100 100 100 100 100

From early to late on the horizontal scale, level group 4 of KG23c appears first, and is followed by level group 3 of KG23c, then KG29S, level group 2 of KG23c, KG1a, KG7a, KG96a, KG5a, KG29N, and level group 1 of KG23c. Overall, level groups 3 and 4 of KG23c are located in the negative sector of Factor 1 and can be considered to belong to the earliest part of the Butana Group sequence. KG29S is the first or earliest site located in the positive sector of Factor 1, and is equidistant between level groups 2 and 3 of KG23c. KG29S can be considered to belong to the early/ middle part of the Butana Group sequence. One should also note that on the vertical scale (Factor 2) KG29S is quite far into the negative part, and as suggested above, indicates that some other difference besides time, sets this site off from the others. After KG29S is level group 2 of KG23c, followed by KG1a, KG7a, KG96a, KG5a, KG29N, and level group 1 of KG23c. All of these sites are considered to belong to the latter part of the Butana sequence. Based on their horizontal distribution, level group 2 of KG23c is chronologically the earliest, followed by KGla, KG7a, and then KG96a. Sites KG5a, KG29N, and level group 1 of KG23c are all the same age.

4.28, reinforces the monotonic patterns of the same types shown through the stratigraphic ordering of the four level groups associated with KG23c in Table 4.4. However, several anomalous relative frequencies among some of the principal types can be noted. For example, the relative frequency of the Brindetta Red-topped type at KG29S is too low. Likewise, the relative frequency of the Bahri Smoothed type at KGla is too high. The relative frequencies of this type associated with KG7a and KG29N is also too high, while the relative frequency at KGla is too low. The Dratmuir Burnished relative frequencies associated with KG96a and KGla are too low, while at KG29S they are too high. At this time, no explanation can be given why these particular relative frequencies are anomalous. However, 6 of the 8 anomalous readings come from KG29S, KG29N, and KG1a, which as mentioned above, may indicate that these particular sites are different from the remaining four Butana Group sites. Nevertheless, considering a total of 50 relative frequency readings, with only 8 being anomalous, one can conclude, with some degree of confidence, that the five principal Butana Group ceramic types show some temporal significance.

Based on the horizontal distribution of the four level groups at KG23c and the remaining Butana Group sites in Figure 4.16, the monotonic patterns (based on the relative frequencies) for the 10 Butana Group ceramic types can be arranged in following manner as shown in Table 4.28.

Unfortunately, this cannot be said for the remaining five minor Butana Group ceramic types. As Table 4.28 shows, the 2 standard error margins are too large to give any indication of whether any of these types are temporally sensitive. Furthermore, there is no convincing monotonic pattern among the five minor types irrespective of the 2 standard error margins. As a hypothesis for future research, the monotonic patterns displayed in Table 4.4 for the five minor types can be used as a basic temporal guide. For

The monotonic patterns of the five principal ceramic types, as arranged by the seriation of the four level groups of KG23c and the remaining six Butana Group sites in Table

69

The Butana Group Ceramics

example, it believed that both the Dratmuira Chatter Marked and Sirjino Simple Dentated types increase in popularity through time. Likewise, the Zirhalla Pattern Burnished type may also increase in popularity. As for the Khordhag Plain type, it is possible that it may increase through time as well. On the other hand, it is believed that the Alabaska Punctated type decreases in popularity over time.

within the early and late Butana Group partition can be seen below. B. Cmbd Late Butana Early Butana

Partitioning the Butana Group into an Early and Late Interval Before closing, a word should be added on whether the Butana Group as a whole can be partitioned into two temporal intervals based on the relative frequencies of the ceramic types through time. As indicated on the scattergram in Figure 4.16, level groups 3 and 4 of KG23c fall within the negative sector of the horizontal scale while level groups 1 and 2 of KG23c and the other Butana Group sites fall within the positive sector. If the horizontal scale is representative of time, then one can place level groups 3 and 4 of KG23c in the early part of the Butana Group (designated as the early Butana interval), while level groups 1 and 2 and the remaining Butana Group sites can be placed in the latter part (designated as the late Butana interval). In addition, one could argue that the site of KG29S is somewhat transitional between the late and early complexes of the Butana Group.

7.4 (±3.5)

B. R-tppd.

B. Smthd.

D. Brnshd.

11.5 (±6.5) 34.8 (±11.2) 19.1 (±2.3)

18.0 (±6.4) 31.6 (±2.3)

12.5 (±3.1)

9.5 (±0.7)

Using these mean percentages and their respective standard deviations as a basic guide, it is expected, that a ceramic assemblage dating to the early Butana Group interval would contain more than 10 percent Butana Combed and more than 20 percent Brindetta Red-topped, while containing less than 20 percent Bahri Smoothed and less than 10 percent Dratmuir Burnished. A ceramic assemblage dating to the early Butana interval should also contain a significant number (more than 4 %) of Alabaska Punctated sherds. Since a few sherds of the Sarb’ut Ripple Marked type were found in the lowermost levels at KG23c, it would not be surprising to find a small presence of this type associated with an early Butana Group component. Within a late Butana Group interval ceramic assemblage, it is expected that the amount of Butana Combed and Brindetta Red-topped would be less than 10 and 20 percent respectively, while the amount of Bahri Smoothed and Dratmuir Burnished should be greater than 20 and 10 percent respectively. No Sarb’ut Ripple Marked sherds should be associated with a late Butana Group interval ceramic assemblage. Furthermore, it is expected that a significant amount of Dratmuira Chatter Marked and Sirjino Simple Dentated (close to 2 %), should be present.

Based on the ceramic type relative frequencies, the early and late Butana intervals can be essentially demarcated by the Butana Combed, Brindetta Red-topped, Bahri Smoothed, and Dratmuir Burnished types. Referring back to Table 4.28, the relative frequencies of these four types can be averaged within both the late and early parts of the Butana Group. Excluding KG29S, and the anomalous relative frequencies (those which have been underlined), the mean percentage (within 1 standard deviation) of the four types

70

Chapter 5 Comparisons of Culture Groups of Northeast Africa During the Neolithic and Post-Neolithic Era

Introduction

fashion (or the availability of materials) will make themselves felt in local variations as in periodic innovations. As a result, prehistoric ‘cultures’ can be quite accurately and consistently differentiated by their pottery even when they are alike in all other respects (Adams 1979: 121).

As stated in the beginning of this study, the Neolithic and post-Neolithic eras in northeast Africa have been primarily viewed through the cultural developments which took place along the Nile Valley. Yet, the Butana Group represents another significant Neolithic development which took place in northeast Africa outside the Nile Valley in the eastern Sudan (Fattovich, Marks, and Mohammed-Ali 1984). The key question concerning the Butana Group is whether it should be conceived as an autonomous development or one which was spawned from an influx of ideas, innovations, or people from the Nile Valley which filtered into the hinterlands of the eastern Sudan.

Thus, the following chapters essentially will revolve around an in depth comparison of the Butana Group ceramics with those of the Neolithic and post-Neolithic cultures of the Nile Valley in northeast Africa. Other ceramic-bearing groups in the eastern Sudan such as those associated with Shaqadud Cave, Kassala (the Gash Group), Agordat, and Erkowit, will also be examined. Through these basic ceramic comparisons, a final consensus can be achieved on whether or not the Butana Group should be viewed as a true non-Nilotic cultural manifestation.

In order to answer whether or not the Butana was an extension of Nile Valley development into the eastern Sudan, a comparison between the cultures of the Nile Valley and the Butana Group is needed. Like the Butana Group, all of the Neolithic and post-Neolithic developments along the Nile Valley have been principally defined by the ceramics, and through this particular class of artifacts one can determine whether or not the Butana Group developed from or was independent of these cultures. As William Adams has so eloquently stated about ceramics and cultural identity in general:

Before comparing the ceramics of the Butana Group with the Neolithic and post-Neolithic cultures of the Nile, it is important to first get a general understanding of the region of northeast Africa and the significant Neolithic and postNeolithic cultural developments within it. Therefore, a brief geographic overview of northeast Africa will be provided followed by a survey and descriptive account of the important Nilotic ceramic-bearing cultures and groups which developed within the area.

The value of pottery as an index of cultural identity has been demonstrated through a century’s experience in the field. In any given site we can observe that ceramics are not only far more abundant than are other cultural remains; they are also more uniform, more consistent, and above all more distinctive than are the designs of houses, implements, and the other imperishable products which largely made up the archaeological record. They also change more consistently, both in time and space. The same decorative traditions will usually be shared by all of the members of a society or community, but they are seldom shared in their entirety with other societies or communities, nor do they persist over long periods of time. Even where there is a common general influence, the caprices of

The Geographic Setting of Northeast Africa The study area of northeastern Africa includes Egypt and the northern half of the Sudan, in addition to the adjacent foothills of the Ethiopian Highlands (see Figure 5.1). It is in this area where the important Neolithic and post-Neolithic developments took place. The primary physiographic feature within this region of course, is the Nile River which runs northward from the lower reaches of its two major tributaries (the Blue and White Nile) to the Egyptian Delta. In assessing the overall physiographic setting of northeastern Africa, it is apparent that the lands which are crossed by the Nile range from the almost vacant, arid Sahara Desert in the north of Egypt, to the more hospitable sub-tropical and tropical steppes of the Sahel in the Sudan

71

The Butana Group Ceramics

Figure 5.1. Geographic areas of northeast Africa.

the central Sudan (Barbour 1964; Williams 1982; Wickens 1982; Banks 1984). As a consequence of this increasing rainfall, the Sahara Desert in northeastern Africa gradually gives way to a semi-desert/steppe within the central part of the Sudan which, in turn, gives way to a low rainfall savanna/steppe farther south (Wickens 1982: 36; Banks 1984: 5).

(Hance 1975: 47). The Sahara and the Sahel run the entire length of the African continent from the Atlantic to the Red Sea and, as biogeographic areas, are governed essentially by the amount of annual rainfall (graphically displayed as isohyets running roughly east-west across the breadth of Africa) which increases as one goes south from almost absolute zero in northern Egypt to over 400 mm per year in

72

Comparisons of Culture Groups

The Nile and Its Journey Through the Sahel and the Sahara

an expansive stretch of plains where the Western Sudan meets the Butana. All along Shendi Reach on both sides of the river there are a series of shallow basins which are the remnants of old paleo-channels of the meandering Nile. During the late summer floods, these basins fill with water. Today, the Shendi Reach is extensively cultivated and is heavily populated (Barbour 1964: 135-136; Adams 1977: 32-33). This area is also frequented on a seasonal basis by many groups of nomads who come in from different areas of the Sahel (Adams 1977: 33). As it will be noted later, the Shendi Reach is another important archaeological zone (including the heartland of Meroe) where many groups of the Neolithic and post-Neolithic eras lived and interacted.

In the southern reaches of the Sahel in northeastern Africa, the Nile begins as two rivers; the White and Blue Nile. The White Nile flows out of Lake Victoria further south in the highland of Uganda, while the Blue Nile flows out of Lake Tana in the highlands of Ethiopia to the east. The White Nile derives its name from the light color of its water, (resulting from decaying organic matter) while the Blue Nile receives its name from the darker sediments (often containing dark, blue-gray clays) it carries down from the highlands of Ethiopia (Barbour 1964: 111). Of the two Niles, the Blue discharges the greatest amount of water (over 80 %) into the main channel of the Nile north of Khartoum. Up river from Khartoum, the White Nile flows gently down slope (at a rate of 1.5 cm per kilometer) within a broad floodplain between the Western Sudan and the Gezira Plain (Barbour 1964: 117, 181). During the late summer floods of the Blue Nile, waters of the White Nile are backed up causing a flood zone a kilometer wide. When the flood waters recede, a very good cover of dense grass (primarily Echinochloa pyramidalis) soon grows over the floodplain, allowing for excellent grazing of which the local nomadic Baggara tribes take advantage during the winter months (Barbour 1964: 182). Unlike the White Nile, the grade of the Blue Nile (which is perched 60 m above the White) is more steep and cuts more deeply between the Gezira and Butana plains (Barbour 1964: 118-119; Gunn 1980: 90). Due to the deeper cut banks of the Blue Nile, overbank flooding does not occur which, in turn, inhibits the growth of pastorale grass in the floodplain (Barbour 1964: 119). As a result, nomadic groups (such as the Rufa el Hoi and the Kenana) have to travel further south into the marshes for winter grazing (Barbour 1964: 183). In some parts of the Butana, however, on the eastern side of the Blue Nile there are pools and other types of back water sloughs which allow stands of grasses to grow once they drain. It should be noted that during the summer rainy season, modern nomads find plentiful amounts of grass for grazing in the interior regions of the Sahel away from the Blue and White Nile (Barbour 1964: 183). It is along this stretch of the White, Blue, and the main Nile just north of Khartoum where the core archaeological area of both the Nilotic-Early Khartoum and Khartoum Neolithic exists.

From the confluence of the Atbara River (where the Sahara begins) through the 5th Cataract and to the 4th Cataract, the Nile continues north and then turns to the southwest through a rugged granite zone called the Abu Hamed Reach. This area is the least inhabited stretch of the Nile from Khartoum to the Delta (Adams 1977: 31-32). Along this stretch, the Nile passes through barren granite which splits the river basin into shoals, rapids, cataracts, and islands. On either side of the Nile is the southern extent of the Sahara Desert. Nevertheless, a little grazing and some farming (but not much) is done along some parts of the river along this stretch (Adams 1977: 33). From the 4th Cataract on up the 3rd Cataract, is the other fertile zone of the Nile south of Aswan, called the Dongola Reach. The course of the Nile through this reach begins in a southwestern direction for approximately 80 km and then turns due north for another 160 km. Here, the Nile enters another broad floodplain which imperceptibly blends into the surrounding desert (Adams 1977: 30). Like the Shendi Reach, the Nile floodplain along this stretch is interspersed with older paleo-channels (remnant meander belts) which form shallow basins for natural water catchments. During the summer floods, the entire floodplain is covered annually with rich sediments similar to those of Lower Egypt (Adams 1977: 30). Unlike, the Shendi Reach, few nomads from the interior visit the Dongola stretch of the Nile due to the desert climate which exists there today (Barbour 1964: 139). The lands along this stretch of the Nile, however, are very fertile and crops today are easily cultivated relative to the availability of irrigated water. In the past, these lands were much more extensively cultivated farther from the river as traces of earlier irrigation ditches and field boundaries can be still be detected on aerial photographs (Adams 1977: 30). Indeed, the Dongola Reach is the heartland of two Kushitic civilizations (Kerma and Napata) which flourished during the 2nd and 1st millennium BC respectively. The Egyptians of the of the 18th, 19th, and 25th Dynasties of the New Kingdom were also present in this area intermittently during the last half of the 2nd millennium BC (Arkell 1955, Adams 1977: 30). Prior to the existence of the C-Group and establishment of Kerma, there is very little information (albeit some reconnaissance work by Marks and Ferring 1971) about Neolithic or post-Neolithic cultures along this particular stretch of the Nile.

From Khartoum (where the White and the Blue merge into the main Nile) north to Aswan, there are only two other unbroken stretches along the Nile where regular overbank flooding occurs (Barbour 1964: 119). The first stretch is called Shendi Reach and is situated between Sabaloka (just north of Khartoum) and the Atbara River. Sabaloka itself represents the 6th Cataract of the Nile River and is characterized by a steep, restricting gorge (with 150 m rock sides) where the Nile flows through a resistant mass of granite, sometimes capped with Nubian sandstone (Barbour 1964: 135). North of the 6th Cataract is the Shendi Reach which stretches for some 300 km north to the Atbara River. There, the Nile broadens into a wide floodplain and the Sahel is seen on both sides as

73

The Butana Group Ceramics

Cairo lies the Fayum depression, which in antiquity was an attractive lacustine environment for early Neolithic groups (Caton-Thompson and Gardner 1934; Wendorf and Schild 1976; Hassan 1988). Overall, the reliable summer floods of the Nile dispersed extremely fertile sediments practically along the entire Egyptian stretch of the river allowing for the easy cultivation of cereals and fodder for domesticated animals. Of course, the Egyptian Nile Valley was intensively populated during both the Neolithic and post-Neolithic eras.

North of the Dongola Reach, past the 3rd Cataract is the Abri-Delgo Reach, which extends northward for about 200 km until it reaches the Dal Cataract. Through this stretch of desert, the Nile flows through another series of more barren granite zones (some of which are a little mountainous) interspersed with more gentler sandstone areas which are reminiscent of the Dongola Reach but on a smaller scale (Adams 1977: 28). Along many of these sandstone areas, the Nile floodplain is again very broad and lands are easily cultivated especially on the east side of the river which is more clay based and free of sand (Adams 1977: 29). Due to the many barren expanses along this particular stretch, the Abri­Delgo Reach was probably never heavily occupied during the Neolithic or post-Neolithic eras; however, very little is known archaeologically about this area.

As it will be pointed out below, the general environmental character of the Nile, the Sahara, and the Sahel, were not always the same due to extreme climatic fluctuations which radically altered the landscape of northeast Africa since the end of the Pleistocene. Like the Paleolithic cultures before them, many Neolithic and post-Neolithic cultures were heavily influenced by these changing climatic fluctuations.

From the Dal Cataract to the 2nd Cataract, the Nile River runs through the formidable, broken, granitic expanse called the Batn el Hajar or “belly of the rock”. This is the most impenetrable expanse of the Nile where rapids, gullies, and barren rock islands make the valley and surrounding desert almost the same (Adams 1977: 26). Other than a few isolated pockets of floodplain, the lands within the Batn el Hajar are almost totally inhospitable for any kind of cultivation (ibid.). Nevertheless, some Neolithic and postNeolithic era settlements have been found in this region (cf. Shiner 1971 and Nordstrom 1972).

Climatic Changes and Past Environments of Northeast Africa since the Holocene As discussed above, the boundary between the Sahara and the Sahel is determined by the amount of annual rainfall. This boundary is essentially controlled by a summer monsoonal rain pattern known as the Intertropical Convergent Zone (ITCZ) which stretches across the central African continent. The ITCZ is a result of a collision between two great weather patterns (wind systems) one of which originates in the southern Pacific (a low pressure system which brings rain) and blows across the African continent northwestward; while the other (a high pressure system of hot air) originates in southwest Asia and pushes across the Mediterranean and Africa in a southwesterly direction (Banks 1984: 6 via Mainguet and Canon 1976; Rognon 1976; Flohn and Nicholson, Messerli, Winiger and Rognon 1980; Mainguet, Canon, and Chemin 1980). North of the ITCZ there is little to no rain which results in a desertic environment, while within and south of the ITCZ the environment is more mesic due to increased rainfall. Today, the boundaries between the Sahara and the Sahel fall between the 50 and 100 mm isohyets (Barbour 1964: 47; Williams et al. 1980: 308-309; Williams et al 1982; 130; Wickens 1982: 35). In prehistoric, as well as in historic times, the ITCZ has shifted the mid African isohyets both to the south and north, which, in turn, has radically shifted the boundaries of the southern Sahara and the Sahel (Warren 1970; Street and Grove 1976; Williams et al. 1980; Wickens 1982). It is not known precisely how or why these shifts occur, but it likely that they are contingent upon the changes of global weather patterns which produced similar fluctuations in rainfall across other parts of the world (Street and Grove 1976). Based on the movement of dune systems (which demarcate the southern edge of the desert), it has been estimated that the Sahara Desert shifted 450 to 500 km to the south near the end of the Pleistocene (Warren 1970; Williams et al. 1980; Mainguet, Canon, and Chenin 1980). On the other hand, high lake levels in the Chad Basin and others playa areas within the Sahara have confirmed

North of the 2nd Cataract to the 1st Cataract at Aswan, the Nile runs through the territory of Lower Nubia which lies almost entirely in modern day Egypt. Today, all of Lower Nubia lies underneath what is now Lake Nassar. When the Nile passed through Lower Nubia, it cut a shallow channel through the Nubian sandstone bedrock which is ubiquitous through this stretch (Adams 1977: 24). As a result, the Nile Valley in Lower Nubia was somewhat restricted between bluffs of sandstone (which rose from 100 to 300 m above the river) on either side of the valley (Barbour 1964: 120; Adams 1977: 26). Along this 330 km stretch of river, however, some parts of the valley were less restricted, allowing for some broad floodplain alluviation (especially along some of the wadi entrances) to occur and these particular spots were heavily cultivated by the Nubians and their post-Neolithic ancestors (Nordstrom 1972; Adams 1977). Beyond the 1st Cataract at Aswan, lies the historic Egyptian Nile Valley which runs its course to the Mediterranean. From Aswan to Cairo, the Nile runs between the limestone cliffs of the Western Desert and the sandstone and limestone cliffs of the Eastern Desert. In comparison with Lower Nubia, however, the Egyptian Nile floodplain (although restricted by the elevated deserts) is wider and flows much more regularly. At Aswan, the valley is 5 km across, while it broadens to 10 km at Luxor, and finally to 15 km through the northern most parts of Egypt (Barbour 1964: 120). The Nile Valley from Aswan to Cairo is called Upper Egypt, while north of Cairo to the Delta the Nile runs through what is called Lower Egypt. Just off to the west below

74

Comparisons of Culture Groups

have consisted of a semi-desert grassland mixed with an Acacia mellifera thorn Savanna (Wickens 1982: 40). The Butana and Atbai areas would have been more of an Acacia mellifera and seyal thorn Savanna (ibid.). However, the clay areas of both the Butana and Atbai would have been subject to periodic flooding (Wickens 1982: 44) and it is quite possible that riparian gallery forests would have lined both the Atbara and Gash rivers during this period of time, as well (ibid.). Further to the west, between the Blue and the White Nile, the Gezira Plain would have been extensively covered with swamps and lakes resembling the present day Sudd (Arkell 1949: 12, 109-110; Wickens 1982: 39). Overall, it has been estimated that the Blue Nile floodplain near Khartoum was greater than 4 m above its present position (Arkell 1949: 12; Williams and Adamson 1980:298). Along this part of the Nile, a gallery forest may have existed, as well, sometime after 6000 BC (Arkell 1949: 15-30, 109-110). Farther to the north in Lower Nubia, the early Holocene wet phase corresponded to a period of aggradation within the Nile Valley (Butzer 1975: 396­-387), depositing some 12 m of sediments along the valley during this time (Butzer 1975: 396). This would indicate overall that the Nile was discharging a higher volume of water during the early Holocene wet phase, perhaps flooding a much greater area of land not only in Lower Nubia, but near Dongola and Shendi, as well. Indeed, the flood levels of the White Nile were very high and it is suspected that a lake of some size existed along the Nile around Khartoum as a result (Williams and Adamson 1973). Based on data recovered from the Nile Valley in Lower Nubia, many of the wadis flowing into the Nile were probably active also (Butzer 1975: 396-397). It is quite likely that large wadi systems such as el Howar and el Milek (which connect the Dongola Reach with the Sahara and the Sahel to the west) were also active (cf. Wickens 1982: 25). The water level within the Fayum Depression also would have been at a maximum during this period (Wendorf and Schild 1976).

that the Sahelean environment moved hundreds of km north during the beginning of the Holocene (Servant 1973; Rognon 1976; Servant and Servant-Vildary 1980; Wendorf and Hassan 1980; Wendorf and Schild 1980). Lakes further to the south (such as Victoria and Tana) have also confirmed similar fluctuating patterns of wet and dry spells at the end of the Pleistocene and beginning of the Holocene (Butzer et al. 1972; Gasse et al. 1974; Livingstone 1980; Gasse, Rognon, and Street 1980). These shifting climatic patterns also have affected the flood patterns of the Nile as well (Bell 1970; Butzer 1976; Livingstone 1980). In a less than perfect scenario, contingent upon a number of different data sets (some conflicting cf. Wendorf and Schild 1980: 239), it appears that the general landscape of northeast Africa since the end of the Pleistocene (ca 10,000 BP) underwent several major cycles of climatic change (alternating between wet and dry periods) defined by the shifting of the rainfall isohyets (Arkell 1949, 1953; Faure et al 1963; Clark 1967 a; Warren 1970; Street and Grove 1976; Servant and Servant-Vildary 1980; Gasse, Rognon, and Street 1980; Wendorf and Schild 1980; Petit-Marie and Riser 1981; Wickens 1982; Muzzolini 1982, 1986; Banks 1984). On the most general level, two series of wet and dry phases affected the environment of northeast Africa between the 9th and 4th millennium BC (Warren 1970; Street and Grove 1976; Hassan 1981, 1987; Hassan and Stucki 1987; Wickens 1982; Muzzolini 1982; Banks 1984; Sadr 1991). These particular wet and dry and phases (cf. Warren 1970; Street and Grove 1976; Wickens 1982; Muzzolini 1982, 1986; Mohammed-Ali 1982; Sack 1988) are commonly referred to as; 1) the early Holocene wet phase (corresponding to Street and Grove’s 1st phase of Period 3, Warren’s phase II, and Wickens’ wet early Holocene), 2) the mid Holocene arid phase (corresponding to Street and Grove’s arid interval between phases 1&2 of Period 3, Warren’s phase III, and Wickens’ sub-humid period), 3) the Neolithic humid phase (corresponding to Street and Grove’s 2nd phase of Period 3 and Warren’s phase IV), and 4) the post Neolithic arid phase (corresponding to Street and Groves Period 4).

With the onset of the short lived mid Holocene arid phase, it appears that the monsoonal rainfall isohyets shifted 200 km south of their present position (Warren 1970; Wickens 1982: 24; Mohammed-Ali 1982: 4). This in effect would have driven the Saharan desert environment into the present northern Sahelean zone. Based on Wickens’ (1982) paleoenvironmental reconstruction, the mixed semi-desert grassland and thorn Savanna (covered primarily with Acacia mellifera and senegal ) of the western Sudan and the northern half of the Butana (west of the White Nile in Shendi Reach) would have reverted back to a semi-desert scrub land. Further south and east (including the south half of the Butana and Atbai region) the environment would have been slightly more moist, supporting a thorn savanna (Wickens 1982: 40; Sadr 1988: 81). Swamps along the White Nile would have been greatly reduced or have disappeared altogether (Wickens 1982: 44-45). The draining of the swamps during this time appears to have roughly corresponded to a general lowering of the water table in the Nile River further north between 5900 and 5500 BC (Wendorf and Schild 1976; Hassan 1988). Water levels in

During the early Holocene wet phase, it is believed that the rainfall isohyets in the central Sudan area shifted north some 250 km (Warren 1970; Wickens 1982: 24). This would have pushed the semi-desert scrub and grassland environment of the Sahel some 250 km into the present day Sahara Desert. In the Western Desert, such lowland basins as Nabta Playa would have been filled with water and the overall environment would have been a semidesert (Wendorf and Schild 1980: 236-237). To the south around the Dongola Reach, the area probably would have resembled the Sahelean zone, which is presently east and west of the Shendi Reach. The Shendi Reach would have been more moist and probably would have been within a low rainfall savanna zone, resembling the environment of present day Qoz (a moist dune area in the western Sudan). Based on Wickens’ paleo­environmental reconstruction, the Western Sudan (the Sahel west of Shendi) would

75

The Butana Group Ceramics

the Fayum were also extremely low at this time (Wendorf and Shield 1976; Butzer 1976: 13).

during the Pleistocene and Holocene epochs. Along the Nile, a direct correlation between the number of radiocarbon dates (i.e. increase in the number of occupations) and successive periods of riverine aggradation can be seen from 21,000 to 4500 BP (Conner and Marks 1986: 190). During the Holocene, there is a clear bimodal distribution of radiocarbon dates in the Nile Valley which date between 8500 - 7500 BP and 6000 - 5000 BP respectively (ibid.). Conversely, there are very few radiocarbon dates between 7500 and 6000 BP, and after 5000 BP (ibid.). Given that the paucity of radiocarbon dates in the Nile Valley is not a direct result of geological erosion (cf. Conner and Marks 1986: 189), one is compelled to conclude that there were indeed dramatic increases and decreases in the number of settlements which correlated with the wet and dry periods discussed above (cf. Conner and Marks 1986: 188-192). These same patterns can also be observed away from the Nile Valley in the Sahara and to a lesser extent in the eastern Sahel (cf. Wendorf and Schild 1980; Fattovich, Marks, and Mohammed-Ali 1984; Marks 1987; Marks et al. 1987; Marks and Sadr 1988). The extreme fluctuations between xeric and mesic conditions, however, were more pronounced along the Nile and adjacent parts of northeast Africa north of Khartoum where effects of shifting Saharan and Sahelean biomes were more dramatic. As it will be discussed shortly below, the Nile Valley itself also acted as a place of refuge during times of hyperaridity (especially to the north of Khartoum) which would have attracted populations from other parts of northern Africa and perhaps southwestern Asia.

During the Neolithic humid phase, the monsoonal rainfall isohyets shifted northward again, pushing the Sahelean zone once more into the southern portion of the Saharan Desert (Warren 1970). It has been estimated that the Sahelean vegetation belt may have moved northward about a 100 km into the desert region (Warren 1970; Wickens 1982: 25). In the Western desert of the Sahara, wadis appear to have been running with water and there, perhaps, was savannalike vegetation in the area (Butzer 1971: 449). Once again, the Dongola Reach area would have probably been within a semi-desert, thorn savanna environment. During this period, there is evidence that the majority of rivers within the boundaries of the present day Sahel were flowing year round (Talbot 1980: 50). Wadi Howar (which runs into Dongola Reach) also appears to have been environmentally more favorable for human habitation (Wickens 1982: 25). Areas further south, such as the Shendi Reach, would have experienced more moisture, while sections along parts of the White Nile would have been covered once more by swamps. In the Atbai region, there are data to suggest that some parts were also swampy (Fattovich, Marks, and Mohammed-Ali 1984: 178). In the Egyptian Nile Valley, there appears to have been a resurgence of high flooding activity around 4500 BC (Butzer 1976: 30-32). At this time a stable, reddish paleosol formed which is ubiquitous throughout the area today (Butzer 1975). It is also suspected that the Nile floodplain expanded laterally along the low lying areas in Lower Nubia during this period (Trigger 1965: 67).

An Overview of the Neolithic and post-Neolithic Cultures of Northeast Africa

During the post Neolithic arid phase, the monsoonal rainfall isohyets moved back more or less to their present position and the southern boundary of the Sahara expanded to the point where it is today (Butzer 1959, Butzer 1976; Warren 1970; Wickens 1982). Based on early flood records of the Nile taken by rulers of the Old Kingdom (during the late 1st and early 2nd Dynasty), there was a marked decrease in the flood levels between 3000 and 2800 BC (Butzer 1976: 28). During this time, it appears that the Nile in Lower Nubia was cutting into its floodplain more than 6 meters (Butzer and Hansen 1968: 276). Since then, the nature of the Nile has stayed pretty much same (Butzer 1976: 28). Present day dune formations on the margins of the Sahel also seem to have begun during this time (Sombroek and Zonneveld 1971; Talbot 1980: 43). Rivers in the Sahelean zone which were permanently flowing during the Neolithic wet phase appear to have reverted back to intermittent streams (Talbot 1980: 51). In the southern Atbai, there are data to suggest that rainfall dropped off and the central steppe area became drier (Marks and Sadr 1988: 79; 81). The level of water at Lake Rudolf also experienced a sharp decline at around 4400 BP (Livingstone 1980: 346).

The Neolithic era in northeast Africa represents the interval when particular groups within the region received or developed domesticated plants or animals (cf. Arkell 1953; Hays 1971:19; Mohammed-Ali 1982; Clark and Brandt 1984; Barich 1987: 331-340; Smith 1989: 73) prior to the advent of metal tools. The use of ceramics in northeast Africa had already been in place for at least a millennium prior to the introduction of domestic plants and animals (cf. Adamson, Clark, and Williams 1974) and all of the known Neolithic cultures in this area were ceramic-bearing. The Neolithic era in northeast Africa may have begun as early 9000 or 8000 BC outside the Nile Valley with the possible domestication of cattle in the eastern Sahara (Wendorf and Schild 1984; 1987; Gautier 1987). These first neolithic developments would have coincided with the Holocene wet phase when many parts of the Sahara were more hospitable for plant and animal domestication. In the Nile Valley, the first Neolithic cultures are found in both the lower and central parts of the river basin by about 5000 BC (Hassan 1986, 1988). These particular Nilotic Neolithic manifestations would have coincided with the mid-Holocene arid phase and it is quite possible that there was an influx of people and/or domesticates from the deteriorating areas of the Sahara, as well as from southwest Asia which, in turn, helped initiate the Neolithic era in the Nile Valley.

The overall effects of the shifting rainfall isohyets and resulting climatic fluctuations had a profound impact on the settlement patterns of human populations in northeast Africa

76

Comparisons of Culture Groups

The post-Neolithic era encompasses the period when copper artifacts were first introduced (via southwest Asia) and used among various Neolithic cultures (also referred to as Chalcolithic) in northeast Africa. This era in northeast Africa basically coincides with increased interactions between complex societies in southwest Asia and the Neolithic cultures in Egypt and begins with the emergence of the Predynastic cultures around 4000 BC (Hayes 1965; Adams 1983; Hassan 1988). It is important to note that during the post-Neolithic era, many of the cultures in northeast Africa to the south of Egypt and Nubia never became Chalcolithic until much later, and these particular cultures have been deemed as being late Neolithic in character (cf. Haaland 1987; Caneva 1988). Therefore, the term late Neolithic will be used for those non-metal using cultures in northeast Africa which were contemporary with other Chalcolithic groups (i.e. the Predynastic cultures) during the post-Neolithic era.

groups which will be discussed below. As it can be seen, there are many areas along the Nile and interior parts of northeast Africa void of any known ceramic-bearing developments, and for the most part, these vacant areas represent sections where little archaeological work has been conducted. Therefore, one can appreciate that much of northeast Africa today still remains virtually unknown archaeologically, and there are some rather large voids in the overall understanding of the Neolithic and postNeolithic eras in many parts of this region. Nevertheless, the descriptions below of the various known Neolithic and post-Neolithic cultures and groups in northeast Africa will give the reader a general perspective on what was taking place along the Nile Valley and in the eastern Sudan in relation to the Butana Group. A word at this time should be mentioned on the usage of the terms “development”, “culture”, and “group”, which are used throughout the discussions below. The word “development” is used to depict a general archaeological manifestation which incorporates one or more cultures or groups during a particular time within a particular area. The term group (and the archaeological usage of the term “culture” - cf. Hassan 1987) is used interchangeably as a subdivision within a particular development to define collections of sites or occupations which share common sets of material culture. As it will be demonstrated later, all of the groups and archaeological cultures discussed below have been primarily defined by their ceramics.

The end of the post-Neolithic era in northeast Africa is associated with the incorporation of autonomous culture groups within larger and more complex, state level types of social organizations. Throughout the Nile Valley and other parts of northeast Africa, this process started in different areas at different times, between the end of the 4th millennium BC and the end of the 1st millennium BC. In Egypt and in Lower Nubia, the end of the post-Neolithic era would have coincided with the 1st Dynasty of the Pharaohs which was established sometime around 3000 BC. Farther to the south in Upper Nubia (i.e. the Dongola Reach), the post-Neolithic era would have ended with the emergence of Kerma sometime in the beginning of the 2nd millennium BC. Still farther south, in the central Nile Valley and the adjacent eastern Sudan, the post-Neolithic era persisted until the rise of the Napatan and Meroitic states during the 1st millennium BC. The Malawiya/Butana transition and Butana Group (ca. 4000-2500 BC) essentially represent late Neolithic cultural manifestations in the eastern Sudan during the post-Neolithic era, and were adjacent to other Neolithic and post-Neolithic cultures which were situated along the Nile Valley.

The Khartoum Mesolithic (Early Khartoum) The first archaeological development to be discussed is the Khartoum Mesolithic (Early Khartoum) which has technically been referred to as “Mesolithic” in the Nile Valley (where it was first described) because there has not been any unequivocal proof of any associated domestic plants or domestic animals associated with this cultural manifestation there (Arkell 1949; Hays 1971; Clark 1984; Phillipson 1985; Caneva 1988). However, recent archaeological investigations in the Sahara may ultimately prove that the term “Mesolithic” is incorrect in describing similar, early ceramic-bearing cultures to the west of the Nile Valley (cf. Wendorf and Schild 1980, 1984; Banks 1984; Barich 1987).

Given that the Malawiya/Butana transition and Butana Group were located well to the east of the Nile Valley, it is evident that other Neolithic and post-Neolithic developments west of the Nile Valley (such as those in the Sahara Desert and in the western Sudan) would not have directly affected either the Malawiya/Butana transition or the Butana Group. Nevertheless, given the prospect that both the Malawiya/Butana transition and the Butana Group were affiliated with an indigenous, non-Nilotic ceramic tradition (i.e. the Atbai Tradition), some attention will be given to the first pre-Neolithic ceramic manifestation in northeast Africa (the Early Khartoum Horizon), which was widespread along the upper portion of the Nile, as well as to the east and west in other parts of the Sahara and the Sahel.

In terms of material culture, the Early Khartoum can probably be best described as a horizon (cf. Hays 1971) with a considerable time depth which was widespread in parts of northeast Africa, as well as other parts of northern Africa across the Sahara and the Sahel (Arkell 1949; Arkell and Ucko 1965; Clark 1973; Adamson, Clark, and Williams 1974; Sutton 1974; Wendorf and Schild 1980; Smith 1980). It is equally clear, however, that this vast cultural manifestation throughout the Sahara and Sahel was made up of many separate social groups which were fundamentally different from one another in terms of economy and adaptation (Barich 1987: 332-333; Gautier 1987; Mohammed-Ali 1987). In its simplest form, the Early

Figure 5.2 shows the known spatial distributions and temporal placement of the ceramic-bearing cultures and

77

The Butana Group Ceramics

Figure 5.2. General location of important Neolithic and post-Neolithic archeological groups and cultures in northeast Africa.

Khartoum can be defined primarily by a set of similarly decorated ceramics which were shared by different groups of people (see discussion below) over several millennia, from as early as 10,000 BC on up to 5000 BC (Arkell 1949; Arkell and Ucko 1965; Clark 1973; Adamson, Clark, and Williams 1974; Sutton 1974; Wendorf and Schild 1980; Smith 1980; Close 1987; Barich 1987). In northeast Africa, Early Khartoum settlements are located in all parts of the

Sahel and in the Sahara west of the Nile Valley, as well as in the Nile Valley itself from the 2nd Cataract in the north, to Kosti about 250 km south of Khartoum (Arkell 1949; Arkell and Ucko 1965; Wendorf 1968; Shiner 1971; Wendorf and Schild 1980; Mohammed-Ali 1982; Clark 1984; Fattovich, Marks, and Mohammed-Ali 1984; Haaland 1987; Schuck 1989; Richter 1989). Early Khartoum settlements, however, are not known in the Nile Valley north of the 2nd Cataract,

78

Comparisons of Culture Groups

nor do they exist in the Sahara east of the Nile above the 1st Cataract. People affiliated with the Early Khartoum horizon, for the most part, were hunter-fisher-gatherers (however, some may have been pastoralists, cf. Wendorf and Schild 1980; Banks 1984; Barich 1987) some of whom relied heavily on the exploitation of aquatic resources along rivers and lakes (Arkell 1949; Adamson, Clark, and Williams 1974; Clark 1984; Caneva 1988). The Early Khartoum essentially coincides with the early Holocene wet phase which existed roughly between the 9th and 7th millennium BC (Adamson, Clark, Williams 1974; Hassan 1984).

been lumped within the Wavy Line wares by Arkell) were also common to the Early Khartoum (Hays 1971: 32-33). Burnished surfaces were never present on vessels associated with the Early Khartoum. Pastes of the Early Khartoum ceramics contained a high content of mineral inclusions and were rarely tempered with organic material (Arkell 1949: 85). Overall, shapes of the Early Khartoum pottery consisted of simple contoured bowls or jars usually with conical bases (Arkell 1949: 86). Rims on the vessels were almost always thinned in profile and the lips were never decorated.

The Early Khartoum was first defined in the central Nile Valley at the well known Khartoum Hospital site (located on the east bank of the Blue Nile on the outskirts of the city of Khartoum), situated at the very northern tip of the Gezira Plain (Arkell 1949). This site consisted of a mound approximately 7,500 m2 in size with over 2 meters of cultural deposits. These deposits were considered unstratified and were disturbed with a number of burials, some of which were recent, while others were associated with the Early Khartoum occupation (Arkell 1949: 4-6).

The faunal remains associated with the material culture at the Early Khartoum site (Arkell 1949: 15-29) consisted of a large quantity of fish (mostly catfish and Nile perch), reptiles (primarily the river turtle), and mammals (mostly antelope). Mollusks of the Pila wernei species (snails) were also found in significant quantities (Bate in Arkell 1949: 28). Very few bird remains were recovered and no domestic animals were identified (Bate in Arkell 1949: 18, 27). The only significant remains of plant food recovered on the site were the seed remains from the wild fruit of the Celtis integrifolia tree (Arkell 1949: 110).

The material culture recovered at the Early Khartoum site was generally composed of lithics (both flaked and ground), carved bone and ostrich egg shell, clay objects (mostly daub), and ceramics (Arkell 1949: 37-95). The flaked lithic artifacts were manufactured mostly from quartz pebbles derived from Nile River sediments and were composed of various microlithic tools including lunates (which was the dominant tool type), backed blades, scaled pieces, points, and triangles, chisel-type arrowheads, scrapers, borers, and burins (Arkell 1949: 42-47). No bifacially worked, macrolithic tools were found. Ground stone was plentiful and was composed mostly of grinders (manos), lower grindstones (matates), fish line sinkers (and other similar kinds of grooved stone), and characteristic stone rings (doughnut-shaped) believed to have been fitted on the end of a stick or club (Arkell 1949: 50-70).

Arkell essentially believed that the people who occupied the Early Khartoum site practiced a “Mesolithic” lifestyle premised on the fact that no remains of domesticated plants or animals had been found (Arkell 1949: 15, 2627, 111). Based on the high number of bone harpoons, the large quantity of fish remains, and other riverine animals (such as the Hippopotamus and River Turtle), it was felt that the Early Khartoum economy was largely aquatic, supplemented by the hunting of large ungulates (such as antelopes) who grazed near the Nile River (Arkell 1949: 27). Arkell believed that the people of the Early Khartoum site could have subsisted off of the aquatic/riparian resources on a year round basis and, therefore, concluded that the inhabitants were somewhat sedentary (ibid.). He further assessed that the past environment associated with the Early Khartoum site was a great deal wetter and that the Nile river bottom in this area was very swampy (ibid.). This was based on the fact that the majority of animals exploited by the Early Khartoum people were swamp-loving creatures, such as the Nile Lechwe (a kind of antelope), Water Mongoose, and Reed Rat (ibid.). The geological data from the site also indicated that Nile flood levels during the time of occupation were much higher (Arkell 1949: 12). From the fauna and geological data, Arkell summarized that the amount of rainfall was from 2 to 3 times greater then than it is modern times and that the surrounding area away from the Nile must have been heavily covered with grass in order to support the herds of antelope which were hunted in large numbers by the Early Khartoum people (Arkell 1949: 110).

Carved bone artifacts consisted of awls and points and very characteristic bone harpoons or spear heads (Arkell 1949: 75-78). The bone harpoons were barbed on one side (usually consisting of 3 or more barbs) and grooved on their bases. Disk beads carved from ostrich egg shell were also common and were used as necklaces (Arkell 1949: 37). Clay objects in the form of small cylinders and fragments of daub were also recovered (Arkell 1949: 79). The clay daub was believed to be from former house structures which were probably constructed of wattle and daub (ibid.). Other than bone harpoons, the most characteristic artifacts associated with the Early Khartoum site were the ceramics (Arkell 1949: 81-95). These Early Khartoum ceramics consisted mostly of the now well known “ Wavy Line” and “Dotted Wavy Line” wares, which were all decorated on the exterior with a multi-toothed comb (Arkell 1949: 81). Later, it was revealed that dentated rocker stamped (evenly serrated edge) decorations (which had formerly

The basic material culture, economic, and ecological configuration of Arkell’s Khartoum Mesolithic (i.e. Early Khartoum) in the central Nile Valley has changed little since it was first defined in 1949 (Arkell and Ucko 1968: Clark

79

The Butana Group Ceramics

1970; Clark 1984; Hays 1971; Sutton 1974; Adams 1977; Phillipson 1985; Haaland 1987; Caneva 1988). What has changed is what it represents as a cultural entity and where it might have originated (cf. Hays 1971; Mohammed-Ali 1982; Barich 1987; Caneva 1988). Arkell had originally believed that the archaeological remains recovered from the Early Khartoum site represented a culture which he originally called the “Wavy Line Culture” (Arkell 1949: 111). Based on the surface finds of Wavy Line ceramics on a number of sites within the “vicinity of Khartoum” he believed that the Early Khartoum culture had originated there and had radiated outward from the Khartoum area across the Sudanic Belt as far west as the central and western Sahara and as far east as the foothill of the Ethiopia Highlands near Kassala (Arkell 1949: 115-118). These assumptions were based on contemporary finds (made by others in Wadi Howar; Tibesti, the Hoggar, Tassile N Ajar, Tenere, Wadi Azaouak, and Kassala) of very similar bone harpoons and ceramics (none of which was true Wavy Line, however) associated with sites found in the Sahara and in other parts of northeastern Africa away from the Nile Valley (Arkell 1949: 112-115). Later excavations in the central Sahara confirmed the idea that Early Khartoumlike cultures (cf. Neolithic of the Sudanese Tradition or the Sahara-Sudanese Neolithic) were widespread throughout much of the Sahara (Tixier 1962; Hugot 1963; Huard and Massip 1964; Camps-Fabrer 1966; Camps 1969; Arkell and Ucko 1965; Clark 1967b; Richter 1989; Schuck 1989).

linguistic fact that the many languages within this same area were all related to a Nilo-Saharan super family which had earlier been postulated by Greenberg (Sutton 1974: 535538; Greenberg 1963). The idea of the Early Khartoum as being a pan-regional cultural manifestation affiliated with a common language across northern Africa continued to be widely held by many researchers in northeastern Africa (cf. Ehret 1984: 30; Haaland 1981: 25-27; 1987: 9-15, 227-230; Hassan 1988: 145). In contradistinction to the idea of a widespread Early Khartoum culture across northern Africa was the reality that many archaeological sites and occupations affiliated with the Early Khartoum (i.e. Neolithic of Sudanese Tradition or Saharan-Sudanese Neolithic) were in some ways more different than they were the same (Shiner 1968; Marks, Shiner, and Hays 1968; Hays 1971; Shiner 1971; Mohammed-Ali 1982; Fattovich, Marks, and Mohammed-Ali 1984; Barich 1987). From a material culture perspective, Hays’ critical appraisal of the notion of a widespread Early Khartoum culture (Hays 1971), to date, is still one of the best and most comprehensive works on this matter (however, see Mohammed-Ali and Jaeger specific critique [1989] on Hay’s culture horizon model). Hays basically took the argument that the only thing held in common between all of these Early Khartoum-like sites across northern Africa were the “decorative techniques and motifs” found on the pottery (Hays 1971: 4). Hays looked at many different Early Khartoum-like sites across northern Africa and compared the ceramics and lithic artifacts to see how similar the sites were (Hays 1971). Upon looking at the ceramics, Hays chose seven different sites. Three of these sites (Amekni-330, Meniet, and Wanyanga) were located in the central Sahara, while the remaining four sites (Early Khartoum, Second Cataract [site 340], Dongola Reach [site 282], and Khashm el Girba [site 78]) were located along the Nile and Atbara Rivers in northeastern Africa. Hays first looked at the different ceramic decorative motifs (of which he defined 14 distinct decorative styles within the overall Early Khartoum assemblages) and found that there was indeed “a homogeneous design style” which “pervades the ceramics of all the sites studied” and concluded that these particular design motifs were “first recognized at the site of Early Khartoum” (Hays 1971: 88). He suggested that these particular “design styles” were held in common over a wide area of northern Africa, from as far west as the Hogar in the central Sahara to the southern Atbai in the eastern Sudan, and should be called the “Khartoum Style” (ibid.). It is important to note that the particular ceramics from the southern Atbai associated with this “Khartoum Style” belong to the Malawiya Group (then known as the Saroba Industry) of the Atbai Tradition. Essentially, the Khartoum Style was based on two common decorative motifs; “Woven Mat” (dentated rocker stamped; evenly serrated edge - cf. Caneva 1988) and “Dotted Straight Line” (simple dentated, i.e. non-rocker stamped) both of which were produced with a comb. In the final analysis, it was felt that the established Wavy Line and Dotted Wavy Line stylistic criteria for the Early Khartoum was not an overall

In an elegant but controversial discourse, Sutton (1974) used both the environmental and archaeological data from the Sahara and the Nile and joined it with linguistic evidence to make an argument that the Early Khartoum represented a widespread “aquatic civilization” which existed along the lakes, rivers, and swamps across much of Africa north of the equatorial zone (from the Niger bend in the far west, across Lake Chad to the Nile and down the East African rift valleys). In essence, he demonstrated that there was a correlation between the distribution of bone harpoons, Wavy Line and Dotted Wavy Line ceramics (affiliated with the Early Khartoum or what others referred to as the Neolithic of the Sudanese Tradition or the SaharaSudanese Neolithic) and the present day distribution of the Nilo-Saharan languages (Sutton 1974: 536-537). Sutton argued that during the early Holocene wet phase (which he placed between the 9th and 6th millennium BC) there was an ecological situation across the central part of northern Africa which facilitated a rapid and far reaching aquatic adaptation based on a high protein diet of fish and other animals associated with lacustrine or riverine environments (Sutton 1974: 528­-529). He believed that the bone harpoons, net sinkers (used to catch fish and other game), and ceramic vessels (used for boiling of foods) were the essential tools to extract and render the rich aquatic resources which in turn promoted a material culture-uniformity across a wide area (Sutton 1974: 529-531). Nevertheless, Sutton held on to the notion that through the very similar looking harpoons and ceramics, this aquatic archaeological manifestation was somewhat unified culturally which he attributed to the

80

Comparisons of Culture Groups

good cultural indicator, based on the general ambiguity of these particular styles (due to the realization that many other different styles had been lumped by Arkell under the Wavy Line and Dotted Wavy Line wares) and the fact that the later Early Khartoum sites (such as those associated with the 2nd Cataract and Khashm el Girba) did not always contain Wavy or Dotted Wavy Line ceramics (Hays 1971: 32-33; 108). Hays did believe, however, that ceramics from both northeast Africa and the Sahara could be considered to belong to a similar stylistic tradition, indicating that “at least one half of all the decorated pottery from all sites studied displayed Woven Mat and Linear Mat design motifs produced by the normal or joined impression of a comb” (Hays 1971: 113). All of the ceramics from these sites were also unburnished, of a similar shape and construction (coil made bowls or jars with rounded bases) and had decorations covering almost the whole vessel (Hays 1971: 116). In the end, Hays concluded that based on these similarities seen in the ceramics, it “could be interpreted as indicating cultural contacts” between the various Early Khartoum sites he investigated (Hays 1971: 116, 137; also see Caneva’s more recent assessment of the Early Khartoum ceramics from the Sahara and the central Nile Valley [1987] which reinforces this conclusion).

rapid spread (however, see Mohammed-Ali and Jaeger 1989: 477­-478) of a particular ceramic design style (a diffusion of ideas as opposed to spread of actual peoples) which was adopted by a number of different cultures or groups of people over a wide area (Hays 1971: 134-137). In contradistinction to the concept of an overreaching Khartoum Horizon-Style across large parts of northern Africa, Barich has recently emphasized that this kind of outlook may no longer be valid at all in terms of understanding the developments of local cultures from the terminal Paleolithic to the Neolithic (Barich 1987: 332-340). Nevertheless, in terms of typological perspectives, there can be little question that some ceramic decorative styles (i.e. dotted wavy line, and rocker combed impressions) were indeed quite widespread across northern Africa during this time (Caneva 1987: 248-252). The Neolithic and post-Neolithic Settlements of Lower Egypt The earliest Neolithic developments in northeast Africa outside the Sahara Desert are associated with two separate manifestations along the Nile Valley, one occurring in Lower Egypt and the other occurring farther south in the central Nile near the vicinity of Khartoum. Of these two areas, the very earliest Neolithic development probably occurred (perhaps by a margin of less than a hundred years) in the Fayum Depression and the Nile Delta of Lower Egypt (Junker 1928; Caton-Thompson and Gardner 1934; Wendorf and Schild 1976; Hayes 1965; Arkell 1975; Wenke 1980; Hassan 1988). Sites associated with the early Neolithic development in the Fayum Depression and the Nile Delta yield significant quantities of Emmer Wheat (Triticum diccocum) and the remains of domesticated animals including goat/sheep, pig, and cattle (Junker 1928; Caton-Thompson and Gardner 1934). These early Neolithic occupations from both the Fayum and the Delta area share a similar set of cultural traits, including chaff tempered pottery (often burnished), well made bifacially flaked tools (including hollow based projectile points, sickle blades, and axes), small oval wood framed houses made from wickerwork, and basket-lined pits used to store grain (Hayes 1965: 93-99, 103-116; Arkell 1975: 12-19: Hassan 1988: 148­149). Radiocarbon dates from the Fayum and the Delta suggest that the early Neolithic settlements in both areas were occupied shortly after 5000 BC (Wendorf and Schild 1976; Hassan 1985). It appears that the establishment of these Neolithic settlements in the Delta and the Fayum occurred with the lowering of the Nile floods (Hassan 1988: 146) which would have coincided with the mid Holocene arid phase (ca. 5500-4500 BC). It has been postulated that the introduction of the Neolithic into Lower Egypt was a result of an influx of outside people and/or material culture (including all of the domesticates) from either the Sahara or the Levant, or both (Arkell 1975: 12-19; Butzer 1976: 10-11), as a response to the overall drying trend (Hassan 1988: 146-147). Once these initial Neolithic societies were established in Lower Egypt, they thrived for several

Based on the differences observed with the lithics, Hays concluded, however, that there were strong differences between the various Early Khartoum-like sites in northeastern Africa and the Sahara (Hays 1971: 133). These differences were premised on the fact that some of the Early Khartoum-like groups used different raw materials and manufactured different tool types. Nevertheless, all of the lithic assemblages observed were essentially microlithic (Hays 1971: 117-123). On the other hand, there were enough differences among the various lithic assemblages (coupled with the great geographic distances between the various areas involved) to conclude that the overall distribution of the Early Khartoum in northern Africa could not possibly be represented by a single culture or group of people (Hays 1971: 123). It was felt, however, that the Early Khartoum groups within the central Nile Valley and the adjacent southern Atbai (including the 2nd Cataract Khartoum Variant of Wadi Halfa, the Early Khartoum Related Group of Dongola Reach, the Malawiya Group of the southern Atbai, and the original Early Khartoum of Gezira) showed enough similarity through the lithics to be lumped in one subdivision, while the Saharan Early Khartoum groups (including Tanezrouft of the western Sahara, Amekni and Meniet of the Hogar, and Tenerian sites of Tenere) could be placed in another subdivision (Hays 1971:133). Hays’ overall solution to the Early Khartoum manifestation in northern Africa was to suggest that it could be best described as a horizon-style (which was based on the sharing of very similar ceramic decorative motifs) which he called the “Khartoum Horizon-Style” (Hays 1971: 134). Thus, in place of a widespread cultural manifestation called the Early Khartoum as advocated by Arkell and others, the Khartoum Horizon-Style could be looked upon as a rather

81

The Butana Group Ceramics

millennium (represented by the later Neolithic and postNeolithic settlements at el Oman and Maadi) and, finally, merged with Upper Egypt by the end of the 4th millennium BC under the newly unified Egyptian state.

from cutting stalks of grain), and hollow based arrowheads; all of which are very characteristic of the Fayum A culture (Caton-Thompson and Gardner 1934). None of these bifacially work artifacts had been found in Lower Egypt prior to their introduction to the Fayum area. Indeed, there is some indication that the hollow based points from the Fayum (the first of their kind in all of Egypt) originated either in the Sahara or the Levant (Arkell 1975: 18; Hassan 1980: 416). Moreover, the overall lithic assemblage of the Fayum A Neolithic does not resemble any of those associated with the terminal Paleolithic within the same area, suggesting that the former culture did not develop out of a local Nilotic tradition (Wendorf and Schild 1976: 317319). Indeed, the terminal Paleolithic chipped stone industry within the Fayum is essentially microlithic, where a high number of tools were made from blades, while the Fayum A Neolithic chipped stone industry is more macrolithic in character, (i.e. consisting of a lot of bifacially worked tools) with many smaller tools made from flakes (ibid.). Overall, the lithic material used on the Fayum A sites ranges widely from tabular Egyptian flint (which most of the larger bifacially worked tools were made from) to quartz, dolerite, limestone, and volcanic ash (the later two used for milling stones). Amazonite beads were also found on Fayum A sites and were probably traded in from northern Tibesti (Arkell 1975).

The Fayum A Sites. The Neolithic settlements of the Fayum Depression were located on the northern edge of Moeris Lake and were deemed by Caton-Thompson and Gardner as the Fayum A communities (Caton-Thompson and Gardner 1934; Hayes 1965: 93; Wendorf and Schild 1976: 225-226). The site of Kom W (the largest Fayum A site) consisted of a deep cultural deposit about 2 m thick and contained many hearths and basket-lined granary pits (Caton-Thompson and Gardner 1934: 24). The site was also covered with massive amounts of pot sherds and lithic debris and was estimated to have covered an area of 35,000 m2 or more (Wendorf and Schild 1976: 211-12). At Kom K, the site was smaller with a thinner deposit between 15 and 30 cm deep (CatonThompson and Gardner 1934: 41). About a kilometer from this site, however, there was a concentration of 56 subsurface granaries within a space of approximately 7000 m2 (ibid.). Based on the number and volume of the granaries at Kom K, it has been estimated that 280 individuals may have lived there at one time (Hassan 1988: 149).

Other kinds of artifacts recovered from Fayum A sites include barbed or slotted fishbone harpoons (the slotted harpoons would hold small barbed points) with beveled ends (Hayes 1965: 95). Bone awls, and pins were also common. Given that many limestone “sinkers” were also recovered, and that the Fayum A people were adept in making baskets (many of which were used to line the granaries), it is probable that fish nets were also manufactured (ibid.).

The fauna and flora remains recovered from both Kom W and Kom K suggested that the diet of the Fayum A people was very diversified (Caton-Thompson and Gardner 1934; Gautier 1976; Brewer 1986). The remains included pigs, cattle, sheep/goat, as well as gazelle, hartbeast, hippopotamus, elephant, hare, catfish, crocodile, snakes, turtle, and waterfowl. Both wheat (Emmer) and barley (six-row, four-row, and two row) were also found. All of the cultigens located in the Fayum depression were already well established hybrids (almost identical to the varieties being grown in Egypt today) indicating that they had been developed for some time in areas outside the Nile Valley (Hayes 1965: 93). Based on the wide diversity of food overall, it was evident that the Fayum A economy was not heavily based on domesticated products and it is very probable that much of the population there alternated from the growing of crops and the tending of livestock to hunting and gathering (Brewer 1986; Wenke et al. 1987; Hassan 1988: 148). Much of the Fayum A population may have practiced a pattern of transhumance, where they would have sown crops at the edge of the lake and then moved their livestock away to graze (probably during and just after the rainy season), returning back to the lake when the crops were ready to be harvested.

The Fayum A ceramics in no way resemble anything like the ceramics of the Early Khartoum. Unlike Early Khartoum pottery, which almost always consists of a grit paste and is decorated with combed impressions and left unburnished, the Fayum pottery is usually tempered with chaff and is commonly burnished (some of which has a red slip or is washed with a purply-red pigment) and is never decorated with any kind of impressed design (Arkell 1975: 13; Hoffman 1979: 185). Furthermore, even though the Early Khartoum pottery is well made and fired, it is nonetheless all of the same form (simple contour, conical based vessels) which is indicative of a technologically early ceramic industry. On the other hand, even though the Fayum A ceramics are crudely made and poorly fired, there are a number of different vessel shapes (many of which are of a composite form and have flat bases), indicating that the ceramic tradition there was not newly developed (Hayes 1965: 92; Arkell 1975: 13-14). Indeed, some of the Fayum A vessels have foot rings (cf. Hayes 1965: 96: Arkell 1975: 15), a trait which is never associated with incipient ceramic forms. Indeed, it appears that the Fayum A pottery is very reminiscent of contemporary Palestinian (Levantine) pottery of the Syro-Cilicia tradition (Arkell 1975: 15 via

The material culture of the Fayum A settlements includes mostly ceramics, lithics (flaked and ground), and some worked bone. The lithic artifacts are composed of a high number of bifacially worked tools including lanceolate blades, celts (where the distal edges have been ground on both sites), hollow backed gouges (which have been ground on one side), sickles (with sickle sheen resulting

82

Comparisons of Culture Groups

Perrot 1962). Such traits as red-finishing and high luster burnishing are common traits shared between the ceramics of Palestine and the Fayum (ibid).

adult males, who must have been buried elsewhere (Hayes 1965: 112; Hoffman 1979: 173). As with the Fayum area, the subsistence patterns of the Merimdian villagers was initially based on a mixed economy where the inhabitants practiced agriculture (growing wheat and possibly barley) and animal husbandry (goats, sheep, pig, and cattle), supplemented with hunting (primarily hippopotamus followed by antelope), fishing (catfish, mussels, etc.) and gathering (Hayes 1965: 112, 115; Hoffman 1979: 170, 176­177; Hassan 1988: 152). The consumption of pork at Merimda seems to have been considerable, based on the large quantity of pig bones recovered there (Hayes 1965: 112).

Merimda Beni Salama. The large site of Merimda Beni Salama on the western edge of the Nile Delta is less than 100 km due north of the Fayum A sites and is believed to be very similar (Baumgartel 1981; Hayes 1965: 103; Arkell 1975: 18; Butzer 1976: 10; Hoffman 1979: 185; Hassan 1988: 147). It is quite likely that much of the upper Delta in Lower Egypt was extensively occupied during this time; however, because of the aggradation in the Nile Delta, many more sites like Merimda are probably buried underneath tons of alluvial sediments (Butzer 1976: 22-25). Contrary to earlier beliefs (cf. Baumgartel 1981), the lower part of the Nile would have been well drained and dominated by swamplands during the time of occupation at Merimda (Butzer 1976: 25).

However, unlike the Fayum area (which had not been substantially occupied until the Middle Kingdom), the occupation at Merimda seem to have intensified, as the number of houses and granaries increased in the upper levels (ibid). By the latter part of the occupation, Merimda became more of a village community, organized and divided along formal kinship lines perhaps something like pueblo communities in the southwestern United States. Some light industry (beyond manufacturing goods solely for survival) may have been in existence during the latter part of the occupation at Merimda where there is evidence for a high degree of activity concerning the manufacturing of lithics, ceramics, cloth, wood, and bone products (Hayes 1965: 116).

The initial occupation of Merimda (ca. 4800 BC) probably occurred about the same time (perhaps slightly earlier) when the Fayum was being occupied by similar groups. The overall occupation of Merimda, however, was quite substantial and was probably continuously occupied (at least 3 occupation phases can be delineated) for at least 400 years (Hassan 1988: 151). Merimda was approximately five times as large as Kom W (the largest Fayum A site), covering an area of about 180,000 m2, and it has been estimated that this site would have supported about 1000 people during the first part of its occupation (Hassan 1988: 152). Like the Fayum A sites, Merimda contained a number of semi- subterranean houses which were oval in shape and constructed of wood (wicker) supported with courses of mud in the lower portions (Hayes 1965: 104-105). Nevertheless, the number and concentration of houses at Merimda was far greater, of which some were arranged in row-like configurations (Hayes 1965: 105; Arkell 1975: 18; Hassan 1988: 152). The actual remains of housing were also much better preserved at Merimda (Hayes 1965: 104-106; Hoffman 1979: 175-176). The houses contained interior oval-shaped mud lined hearths which “had mud fire dogs” along the top sides to support cooking vessels (Hayes 1965: 104; Hoffman 1979; 176). Other smaller oval depression were also common on the house floor which were used to support other round bottomed vessels (ibid.). Larger ceramic vessels were also sunk into the house floor for storage (Arkell 1975: 18). Like the Fayum, subterranean granaries lined with basket work were common at Merimda and traces of Emmer wheat were found within some of them (Hayes 1965: 105; Arkell 1975: 18; Hassan 1988: 151). Large ground stone querns (used for the grinding of grain) were also frequently found. Burials (which were not found at the Fayum) were often found beneath the house floors at Merimda in a flexed position, knees drawn up, hands in front, usually lying on their right sides (Hayes 1965: 105106, 112-113; Hoffman 1979: 173). Interestingly, most of these burials contained either women or children but rarely

The material culture recovered from Merimda is essentially the same as the Fayum A sites, however, artifacts from the former, are as a rule, more developed and there is a higher diversity of raw material being used (Hayes 1965; Arkell 1975; Hoffman 1979; Hassan 1988). The lithic industry (based primarily on Egyptian tabular flint) at Merimda was essentially macrolithic with an emphasis on bifacially worked tools including lanceolate knives, sickles, adzes, celts, and distinctive hollow based arrowheads (Hayes 1965: 108; Hoffman 1979: 177-180; Baumgartel 1981: 15-17). For the first time, pear-shaped maceheads (which are characteristic of the Predynastic cultures further to the south) appear at Merimda (Hayes 1965: 109). It has been argued that these particular kinds of maceheads were also very common in Palestine, as well as in Anatolia (ibid.). Grinding stones (made from granite, basalt, and sandstone) are also numerous and are found all throughout the occupation levels at Merimda. Unlike the Fayum A sites, hundreds of artifacts made from ivory, bone, and horn were also recovered from Merimda and many of these artifacts were probably used in the processing of skins used for clothing (Hayes 1965: 110; Hoffman 1979: 180). Like Fayum A, the ceramics from Merimda are usually chaff tempered (although, some of the more finely made vessels were not) and were fashioned into similar shapes (Hayes 1965: 106-107; Hoffman 1979: 180-181). The ceramic assemblage at Merimda, however, represents a more complex array of pottery where there was a marked

83

The Butana Group Ceramics

increase in the number of different style and shapes produced (ibid.). Like Fayum A, vessels at Merimda were often red-finished and burnished on the exteriors. A particular burnished red-topped vessel with an unburnished, herringbone banded incised design directly below the rim was one of the more popular types (probably late in the sequence) of vessel produced at Merimda (Hayes 1965: 104; Baumgartel 1981: 17-18). Other kinds of ceramic decorations found at Merimda include black-slipped polished, and a characteristic red-black speckled design (Hayes 1965: 107).

quality and consist of a greater variety of different shapes and sizes. The ceramics are either smoothed or burnished and are often red-finished, however, brown and black wares are present as well (Hayes 1965: 118; Hoffman 1979: 179180). The lithics at el Omari are in some ways similar to Fayum A and Merimda assemblages with the usual bifacially worked tools such as sickles, hollow-based points, axes, and celts. However, there were significantly more flake and blade tools produced at the former site (Hayes 1965: 117-118; Trigger 1983: 24).

El Omari of the Lower Nile.

The Lower Egyptian Post-Neolithic at Maadi.

The occupations of el Omari represent the later development of the Lower Egyptian Neolithic and overlaps into the postNeolithic era. These Neolithic occupations probably began at the close of the 5th millennium BC (probably succeeding the occupation at Merimda) and lasted until the beginning of the 4th millennium BC (Hoffman 1979: 194). El Omari is actually a locality and consists of three sites (Omari A, B, & C) situated about 50 km up river from Merimda on the east side of the Nile near the mouth of Wadi Hof (Hoffman 1979: 192-194). The settlements of el Omari were quite extensive (resembling Merimda) of which the largest site (Omari A) contains over a hundred oval-shaped house features (Hoffman 1979: 195). Houses at el Omari were semi-subterranean with wood frame infrastructures lined with mats and then covered over with clay (Vandier 1952; Hayes 1965: 117; Trigger 1983: 24). Mat lined granaries, much like the ones at Fayum A and Merimda, were also a common feature associated with the households. Many of the houses at the settlements of el Omari were partitioned off with small yards demarcated by low walls of earth. Like Merimda, burials were often placed under the house floors and actual cemeteries appeared later at el Omari during the post-Neolithic era (Hoffman 1979: 195).

The site of Maadi is situated about 10 km north of el Omari and about 2 kilometers east of the Nile River near the mouth of Wadi el Tih. The occupation of Maadi can be placed roughly between 4000 and 3000 BC and was contemporary with the later settlements at el Omari (Hoffman 1979: 201). The local utilitarian ceramics at Maadi suggest that it belongs with the general Neolithic tradition of Lower Egypt. These ceramics are chaff-tempered, thick walled and undecorated, and are either smoothed or burnished with a red or black finish (Hayes 1965: 124; Caneva 1987: 107-109). However, unlike el Omari and other Lower Egyptian Neolithic settlements, the material culture of Maadi possesses a significant amount of copper artifacts and foreign made ceramics not present at the other sites. By the middle part of the 4th millennium BC (ca. 3600 BC), the inhabitants at Maadi began to actively engage in trading relationships with other groups in the Sinai and parts of southern Palestine (Hoffman 1979). The stepped up trading activity at Maadi seems to have been synchronous with increased trading practices of the Predynastic cultures (especially the Gerzean, see below) farther to the south in Upper Egypt. In fact, Predynastic ceramics occur in limited numbers at Maadi (Hoffman 1979: 204). The active trade involving Maadi appears to have been associated with what is believed to be an on site production of copper artifacts, of which many were cast (Hoffman 1979: 207: Trigger 1983: 25). Due to the close proximity of Maadi to the Sinai Peninsula, raw copper ore was probably transported in from the Sinai (where it occurs naturally) and procured at Maadi (Hayes 1965 122; Hoffman 1979: 201; Baumgartel 1981: 122; Trigger 1983: 26). It is for that reason that Maadi was probably located near the entrance of Wadi el Tih, which in historic times was used as a route to the copper mines in the Sinai (ibid.). The remains of domestic donkey (the earliest in Egypt) have also been recovered at Maadi and there can be little question that such animals would have been used in caravans to transport copper ore and finished goods (Hoffman 1979: 201; Caneva, Frangipane, and Plamieri 1987: 107).

The subsistence practices at el Oman seem to more geared towards agriculture and the raising of stock than of a mixed economy supplemented heavily with hunting and gathering. At el Oman there is a greater variety of cereals (and they appear to be more highly developed) including for the first time in Egypt, club wheat (Triticum compactum). However, the use of club wheat at such an early date has been recently disputed (Hoffman 1979: 196). Nevertheless, remains of cake made of crushed wheat have been recovered there (Hayes 1965: 119), proving that bread wheats were cultivated at this time in Lower Egypt. Figs and dates were also grown, as well as flax for weaving (ibid.). Like Fayum A and Merimda; pigs, sheep/goat, and cattle were also raised (Hoffman 1979: 196). On occasion, hippopotamus, antelope, crocodile, and ostriches were hunted. Like Fayum A and Merimda, the ceramics at el Omari are of the same basic kind, consisting of plain, thick-walled undecorated flower-pot shaped vessels with flat or footed bases (Hayes 1965: 118; Hoffman 1979: 197; Trigger 1983: 24). However, the ceramics at el Omari are of a better

It is peculiar that the contemporary settlements of el Omari (just 10 km to the south) were not participating in the

84

Comparisons of Culture Groups

trading and copper production taking place at Maadi. At this point, there are no good explanations for the disparity between the two communities, however, it may have been due to a monopolization of trade and production by Maadi over el Omari, or possibly a result of a conservative attitude that the inhabitants of el Omari had which prompted them not to interact with the more progressive Maadians.

developments at Maadi, the characteristic austere cultural pattern of traditional Lower Egypt was to persist at Maadi and other sites for some time on up through the beginning of the Dynastic period (Hoffman 1979: 212-214). The Egyptian Predynastic of Upper Egypt The Egyptian Predynastic is post-Neolithic in character (due to the presence of copper artifacts) and is essentially confined to the Nile Valley in Upper Egypt (Petrie 1901; 1920; Hayes 1965: 147-148; Arkell 1975: 30-52). It is clear that the earliest Predynastic sites were contemporary with some of the later Neolithic (in particular the latter occupations at Merimda and el Omari) developments in Lower Egypt (cf. Hassan 1988: 153). There was a geographic gap, however, of more than 200 km between the southernmost Lower Egyptian Neolithic settlements and those of Upper Egypt and there was only limited contact between the two groups until well into the second half of the 4th millennium BC (Trigger 1983: 13-40). Indeed, no Neolithic settlements such as those in Lower Egypt have ever been found in Upper Egypt (Hayes 1965: 147). Unlike Lower Egypt, the remains of pig are not commonly found at any of the Upper Egyptian Predynastic sites (Trigger 1983: 26). The penning of animals seems to be more prevalent among the Predynastic cultures, as well (Hassan 1988). Cemeteries and elaborate burials (which as a rule are not associated with the Lower Egyptian Neolithic) were from the start very common features with all of the Upper Egyptian settlements. Overall, the material culture of the Upper Egyptian Predynastic societies is more elaborate and conspicuous than their neighbors to the north (Hoffman 1979: 211-214).

The overall economy of the post-Neolithic settlement at Maadi was indeed fundamentally different from the earlier Neolithic economies of Lower Egypt as it was based on the production of copper artifacts and on an increased volume of trade with other post-Neolithic societies. Nevertheless, the host of plants and animals consumed by the inhabitants of Maadi remained the same. Emmer wheat and six row barley were the principal grains grown, while cattle, goat/ sheep, and pig were the common domestic animals (Hayes 1965: 130-131). However, at Maadi there was a marked decrease in the number of wild animals (such as fish, hippopotamus, mollusks, and turtle) hunted (Trigger 1983: 25). The site of Maadi is quite large encompassing an area of 180,000 m2 and has cultural deposits up to 2 m deep (Trigger 1983: 25). The living area appears to have been situated in the center of the site where the remains of oval-shape housing (constructed of a wood infrastructure and probably covered with matting or wattle and daub) have been found (ibid.). Like other Lower Egyptian settlements of the time, the small houses at Maadi were accompanied with subterranean storage pits (usually mat covered and mud lined) used as granaries. However, there were other subterranean structures (2 m deep and 3 to 5 m in diameter) at Maadi which were clearly used for housing but were unlike any other house type found in Lower Egypt (Hoffman 1979: 201-202). These particular structures were, however, very reminiscent of housing of a similar age found at several sites in Beersheba of southern Palestine (ibid.). Other unusual features found at Maadi, not typical of Lower Egypt, include large storage cellars which housed large ceramic pithoi (very similar to ones produced in Palestine) which contained grain and smaller stone vessels (Hoffman 1979: 203). There is little doubt that these kinds of storage facilities were used for commercial purposes, the items of which were intended for trade (ibid.). Cemeteries also appear at Maadi, and their presence signify an increasing amount of contact with the Predynastic groups in Upper Egypt (where cemeteries are also common). Individuals buried in these cemeteries were associated with graves goods (including ceramics, ground and polished stones and other objects made from bone) common in Predynastic burials (Hoffman 1979: 209-210). The presence of cemeteries and the elaboration of burial customs overall, indicate that some individuals at Maadi were becoming wealthy, probably through trade, and that some class differences were being formed. Wealth and class differences through conspicuous consumption was very uncommon in Lower Egypt until Maadi became established as a center. Despite the new post-Neolithic

The Egyptian Predynastic can be divided into three to four chronological units (Early, Middle, Late, and/or Terminal) of which each roughly corresponds with a particular culture ( defined primarily by the ceramics); including the Badarian (Early Predynastic), Amratian (Middle Predynastic or Naqada I), Gerzean (Late Predynastic or Naqada II), and late Gerzean/Semainean (Terminal Predynastic or Naqada III). Formerly, it was believed that each of these cultures more or less represented sequential stages (conceived as separate social groups) within the Predynastic period (cf. Petrie 1901; 1920; Brunton and Caton-Thompson 1928). Nevertheless, it was soon realized that some of these Predynastic cultures were actually coexisting with one another (Hayes 1965: 147; Baumgartel 1981: 24; Hassan 1988: 138-140). In a very general sense, a sequence (through the ceramics) from Early to Late-Terminal can still be maintained in association with the Badarian, Amratian, and the Gerzean cultures. However, other than the end of the terminal Predynastic (which coincides with the unification of Upper and Lower Egypt), the changes from Early to Middle to Late was not always uniform throughout all parts of Upper Egypt (Hassan 1988: 138).

85

The Butana Group Ceramics

The Badarian of the Early Predynastic.

1928: 75-76; Baumgartel 1981: 21-22; Trigger 1983: 23). In fact, almost all of the Badarian chipped stone tools were manufactured from small, weathered surface nodules of flint (recovered from the desert) rather than the better quality tabular flint which was available in the nearby limestone cliffs along the edge of the Nile Valley (Brunton and Caton-Thompson 1928: 75). The choice of using inferior raw material for the manufacture of lithic tools indicated to both Brunton and Caton-Thompson that the Badarians were not familiar with the better quality flint resources, which lead them to believed that the Badarians were not native to the area (ibid.). They noted that the Badarians never used bone harpoons (a common Nilotic trait) and that sites were always on the east side of the river. Furthermore, sea shells from the Red Sea were commonly found in Badarian burials and the ibex (an animal which is only known to exist in the Red Sea Hills) was sometimes used as a decorative motif (ibid.). Based on these data, Caton-Thompson believed that the Badarians originated somewhere to the east and south, perhaps coming from the Red Sea Hills (Baumgartel 1981: 22). A possible southern Red Sea Hills homeland or route of entry for the Badarians (who may have also navigated the Red Sea for the acquisition copper artifacts) is still entertained by some scholars (cf. Tutundzic 1989).

The Badarian culture is defined by some 80 cemeteries and habitation sites on the east side of the Nile in the Asyut area of Upper Egypt (Brunton and Caton-Thompson 1928; Arkell and Ucko 1965: 150-151; Arkell 1975: 31-34; Hayes 1965: 147; Hassan 1988: 153-154). Surprisingly, the total extent of the Badarian culture today encompasses only a small stretch along the east side of the Nile less than 35 km in length (Hayes 1965: 147; Trigger 1983: 28; Hassan 1988: 153). Very few radiocarbon dates have been recovered from Badarian sites; however, a date of 4100 BC with a range from 4400 to 3800 BC is considered reliable (Hassan 1988: 141). Well known Badarian sites include El Matmar, El Mostagedda, El Bahri, El Hemamieh, and El Qaw. The Badarian culture is primarily defined by its distinctive and finely made ceramics which consists mostly of burnished (usually red-finished or brown in color) blacktopped hemispherical and carinated bowls with round bases (Brunton and Caton-Thompson 1928: 20-26, 53-55; Baumgartel 1981: 22; Arkell 1975: 32-34). As it will be discussed in greater detail below, the bulk of the Badarian ceramics (which consist mostly of crude utilitarian wares) are composed of chaff tempered pastes, while the more finely made ceramics were produced from grit pastes. Most of the finely made vessels have rippled exterior surfaces (combed and then burnished) which is the ceramic hallmark of the Badarian culture (ibid.). On the interior of some of the open bowls, are pattern burnished lines usually representing branch-like designs (Baumgartel 1981: 22). Both the technique of black-topping and rippling may have originated with the Badarian culture (Brunton and Caton-Thompson 1928; Baumgartel 1981: 22). However, black-topped and rippled sherds are present at the Khartoum Neolithic site of Esh Shaheinab and other central Nile sites farther north which were contemporary with the Badarian culture (Arkell 1953: plate 33, no.10&11: Bietak 1986: 114; Hassan 1988: 145). The technique of black-topping was done by placing the rim of the vessel face down in a bed of hot ashes which essentially carbonized the top rim portion of the vessel, leaving a distinct black-top on the vessel (Petrie 1901: 13). The technique of rippling was usually done by combing the exterior of the vessel and then partially obliterating the combed marks by repeated burnishing which in effect left a polished rippled surface on the vessel (Brunton and Caton-Thompson 1928: 20-23).

Other than the ceramics, the Badarian culture is also noted for its extensive cemeteries and an elaboration in burial customs which became a hallmark for the later Predynastic and Dynastic cultures of Egypt (Brunton and Caton-Thompson 1928; Arkell and Ucko 1965: 150). These cemeteries were always located to the east of the settlements and graves consisted of oval pits containing individuals in a flexed position (heads usually to the south) lying on their sides with hands in front of face (Brunton and Caton-Thompson 1928: 18-20; Arkell and Ucko 1965: 150; Baumgartel 1981: 21-23). Bodies of the deceased were often placed on reed matting (Arkell and Ucko 1965: 150) which was also to become a common custom among all of the Predynastic cultures. Many of these individuals were also wrapped in either linen or the skins of animals (Brunton and Caton-Thompson 1928: 19; Arkell and Ucko 1965: 150). Grave goods consisted of ground and chipped stone artifacts (including distinctive flat topped maceheads and rectangular slate palettes), shells, amazonite beads (imported from Tibesti), ceramics, and carved elephant ivory (probably imported from Nubia) including bangles, combs, and figurines (Brunton and Caton-Thompson 1928: 2-18; Arkell and Ucko 1965: 150; Trigger 1983: 27). Faience covered steatite beads were also found in addition to small copper items such as awls and fish hooks (Brunton and Caton-Thompson 1928: 27; Arkell and Ucko 1965: 150). These latter items were probably imports either traded in from the Sinai or Palestine (where similar artifacts have been found) and were perhaps transported down along the Red Sea to the Badarian area (Arkell and Ucko 1965: 150; Tutundzic 1989: 255-259).

The chipped stone industry of the Badarian culture is essentially macrolithic (flakes are dominant) and bifacially worked artifacts are common, including well made lanceolate and hollow-based arrowheads, as well as sickle blades (similar to those found in northern Egypt) and distinctive end scrapers (called “push plains”) which may have been used for the processing of hides (Brunton and Caton-Thompson 1928: 35-38, 75-77; Baumgartel 1981: 22). The majority of other stone artifacts, however, are less well made, and it was noted that the Badarian lithics are rather crude overall (Brunton and Caton-Thompson

The subsistence practices of the people associated with the Badarian culture revolved around agriculture and the raising

86

Comparisons of Culture Groups

of live stock (Brunton and Caton-Thompson 1928). Both wheat (Triticum diccocum) and barley (Hordeum vulgare) were cultivated (Baumgartel 1981: 23). Castor oil (derived from the wild plant Ricinus communis) was also commonly used (probably for cooking) and flax (woven into linen) was grown for clothing. Unlike the Neolithic communities in Lower Egypt, the Badarians were probably more reliant on agricultural crops and did not practice a pattern of transhumance, and as a result, kept their animals penned (Hassan 1988: 154). At the site of Deir Tasa, layers of goat and sheep manure (some over 20 cm thick) were often encountered indicating livestock pens of this sort (Gabra 1930). On the other hand, it is apparent that the majority of Badarian sites were temporary encampments with thin cultural deposits, and not permanent year round settlements (Trigger 1983: 27; Hassan 1988: 153). An average site such as Deir Tasa was no larger than 5000 m2 (Gabra 1930) and it has been estimated that a village of this size would have only supported between 40 to 65 people at a time (Hassan 1988: 154). The lack of any large sized Badarian sites may have also been due to the restriction of the floodplain along this part of the Nile which would have prohibited larger settlements (ibid.). Nevertheless, Badarian sites tend to have large subterranean granaries, some of which were over 2.5 m in diameter and 3 m deep (Hassan 1988: 153). No remains of houses have been found. However, they were probably small oval structures (similar to those at the Fayum and Merimda) constructed of wood framing and covered with thatch (probably reed matting) or skins (Trigger 1983: 27).

The burial customs and material items of the Amratian culture share many common characteristics with the Badarian culture, suggesting that the two were closely related (Arkell and Ucko 1965: 151; Arkell 1975: 36; Trigger 1983: 31). Like the Badarian, Amratian graves are oval, (however, some of the latter Amratian graves have side chambers) containing flexed individuals on their sides (hands in front of the face), and posses a similar array of artifacts (Petrie 1901). For example, small copper items, such as fish hooks and pins, maceheads, slate palettes, carved ivory combs and large number of figurines, as well as red-slipped pottery are found in Amratian graves (Petrie 1901: 31-36, plates II-VII, ix-xiv). In the earlier Amratian graves, both black-topped and ripple wares (most often flat-based) are also present and are similar to their Badarian counterparts (Arkell and Ucko 1965: 151). In later Amratian burials, however, ripple wares drop out rather quickly, while black-topped wares persist longer (Brunton and Caton-Thompson 1928: 74-75; Arkell 1975: 36, 40). White cross-lined (white paint on a red-finished burnished surface) and incised black ware (actually whitefilled incised decorations on a black surface) also appear for the first time in the Amratian culture, however, plain redfinished wares are still produced (Petrie 1901: 13-17, plate II; 1920: 3-4, 14-16, 45, plates X-XIV; Arkell 1975: 40; Hoffman 1979: 120). Overall, most of the Amratian burial pottery was produced from a fine-grained paste (Petrie 1901: 13) which had natural quartz inclusions (derived from “Nile mud”, some of which was micaceous). As in the Badarian culture, the bulk of the Amratian ceramics is composed of a common, utilitarian ware produced from chaff tempered pastes.

The Amratian of the Middle Predynastic (Naqada I).

Flat-topped maceheads were also replaced quickly by the very characteristic disk-shaped (more conical-shaped) Amratian macehead (Petrie 1921: 22, plate XXC, Arkell 1975: 39; Baumgartel 1981: 34). These particular maceheads may have been fashioned from hand-sized grinding stones and were usually made from diorite or from some other hard black rock (Arkell and Ucko 1965: 152). Rectangular slate palettes also became less popular and were succeeded by more naturalistic, animal-shaped ones (Arkell and Ucko 1965: 153). The flaked stone industry of the Amratian culture is similar to the Badarian but is more developed (has more true blades) and contains a higher quantity of bifacial tools (Petrie 1901: 23, plates VII-VIII; Baumgartel 1981: 38; Trigger 1983: 31). A highly skilled stone tool technique was also perfected in the Amratian culture where flint blanks were ground down to a very thin width and then finely flaked (often called “serial flaking”) into characteristic rhomboidal knives and fish-tail points (Petrie 1901: 23; Arkell 1975: 39, 47; Trigger 1983: 31).

Based on the stratigraphic positioning of ceramics from both an Amratian and Badarian occupation at Hemamieh, it was established that the Amratian culture (defined primarily by Petrie and called Early Predynastic) superseded the Badarian (Brunton and Caton-Thompson 1928: 74-79). The main part of the Amratian culture, however, seems to have developed about 100 km south of the Badarian area along the great eastward loop of the Nile River, between Abydos and Naqada, which is considered to be its core area (Kaizer 1957: 74-75; Hayes 1965: 148; Trigger 1983: 30). It is likely that the earliest development of the Amratian culture in this area may have been contemporary with the some of the Badarian occupations to the north (Baumgartel 1981; Hassan 1988). Recent radiocarbon dates suggest that the Amratian culture began sometime around 3750 BC (Hassan 1988: 141). Shortly after the Amratian culture was established in its core area, it spread both north into the Badarian area (as demonstrated by the stratified deposits at Hemamieh and Deir Tasa), as well as, south into the very northern parts of Lower Nubia (Reisner 1910; Trigger 1983: 28). Sites associated with the Amratian culture in the core area include El-Mahasna, Abydos, Hu, Naqada, Armant, and Gebeliin. A possible Amratian site in lower Nubia is Khor Bahan; however, it is also considered to be an A-Group site, (see discussion below).

Overall, Amratian settlements are more numerous and tend to be larger on average (and perhaps more prosperous) than Badarian settlements (Arkell: 1975: 42; Trigger 1983: 30; Hassan 1988: 154). Amratian sites range in size from shallow deposits of several thousand square meters to sites of more than 30,000 m2 in size with deposits of a meter

87

The Butana Group Ceramics

in depth (Hassan 1988: 154-155). Based on the range of site sizes, the presence of large (probably more permanent villages) and small communities (temporary like the Badarian settlements) suggests that there was a settlement hierarchy of some sort. Population estimates would have been 50 individuals for the smaller settlements and 250 for the larger communities (Hassan 1988: 155). Generally, the character of Amratian sites are similar to Badarian sites, consisting of small flimsily built huts, probably constructed of wattle and daub, associated with cooking hearths and subterranean granaries (Hassan 1988: 155). House-like features which have been found at Amratian sites consist of oval-shaped structures about 2 m in diameter (Arkell 1975: 40; Trigger 1983: 30). It appears that some of these house structures may have had more permanent foundations built of stone and mud (ibid.). Of course, large cemeteries associated with the Amratian settlements are also common, of which the famous cemetery of Naqada (containing over 2000 Amratian and later Gerzean graves) is the largest (Baumgartel 1981: 26).

Unlike Amratian sites, however, no true Gerzean sites exist south of the 1st Cataract, probably indicating that a cultural boundary had formed between Egypt and Lower Nubia at this time (Nordstrom 1972; Trigger 1983). Well known Gerzean sites include Naqada and Hierakonpolis in the south, and El Haraga, El Malaq, Abusir, El Girza, and Tarkhan in the north. The Gerzean culture followed the same subsistence patterns of the former Predynastic cultures where the people were primarily agriculturalists, with penned herds of goat, sheep, and cattle. Nevertheless, Gerzean sites tend to be more substantial and are more akin to Egyptian villages of the later historic periods. For the first time, mud brick was used for housing (Trigger 1983: 36; Hassan 1988: 155-156). Gerzean houses in general changed shape from round to square, some of which consisted of several room blocks (ibid.). Towards the latter part of the Gerzean, brick-lined burial chambers also appear, representing a marked sophistication in construction techniques as well as a class distinction. For example, at Hierakonpolis there are the famous painted tombs which contained a series of annexes and chambers reminiscent of the later precinct tombs associated with the 1st Dynasty (Kemp 1973; Trigger 1983: 36).

The subsistence practices associated with the Amratian culture were much the same as the Badarian culture, centering around agriculture (cultivating wheat, barley, flax, etc.) and the raising of livestock (Baumgartel 1981: 25). Thick deposits of sheep and goat manure present on many sites also demonstrate that people associated with the Amratian culture were penning their animals in the same fashion as was practiced during Badarian times (Hassan 1988: 155). Looms may have been used during Amratian times, as evidenced by a depiction of a horizontal and vertical loom on a ceramic sherd (Baumgartel 1981: 30).

Like the other Predynastic cultures, the ceramics are the best markers of the Gerzean culture. The presence of polished, red-finished and black-topped vessels (albeit more rare), in addition to an increase in “rough ware” (ceramics with a rough, smoothed exterior surface) within the Gerzean gives a strong indication that it developed from the Amratian (Petrie 1901: 13-17, plate II; Arkell and Ucko 1965: 154; Baumgartel 1981: 71­72). However, during the end of the Amratian and beginning of the Gerzean, the white painted geometric decorations on the red-finished ceramics were replaced by more naturalistic and zoomorphic forms which were painted red on buff surfaces (Petrie 1921; Hoffman 1979: 121; Baumgartel 1981: 52-88). Boats, people, palm fronds, and different kinds of animals such as gazelles, birds, fish, and scorpions become common decorative motifs on Gerzean vessels (Petrie 1921: plates XXXIIIXXXVII; Trigger 1983: 33). It is evident that many of the animals and motifs on the vessels were native to Africa, but it is also clear that the particular fashion in which they were drawn (such as files of animals and the sacred boat motif) was very reminiscent of ceramic paintings common to both the Uruk and Jemdet Nasr periods of Mesopotamia (Arkell and Ucko 1965: 154; Baumgartel 1981: 71-88). Much of this painted pottery also appears to have been mass produced (some of which was slow turned), probably being made at only a few centers (Lucas and Harris 1962: 369; Baumgartel 1970: 488; Hoffman 1982: 142; Trigger 1983: 33). Many new vessel forms were also introduced, including the very characteristic wavy handled jars (usually left undecorated) which were probably derived from the Early Bronze I cultures of Palestine (Trigger 1983: 32). There is also, for the first time within the Gerzean, a change from using natural alluvial clays to using a finer mixture of

The Gerzean of the Late to Terminal Predynastic (Naqada II & III). Relative to the Badarian and Amratian cultures, the developments of the Gerzean culture were more pronounced, aided with an influx of new styles and technologies, many of which probably resulted from increased contacts with southwest Asia (Arkell and Ucko 1965; Baumgartel 1981; Trigger 1983). The Gerzean culture, as a whole, was more socially complex than the more egalitarian-agrarian societies of the Badarians and Amratians (Hoffman 1979, Trigger 1983: 32-40). Craft specialization coupled with a wider sphere of trade and an increase in the use of metals (including copper, gold, and silver), is also evident during this time (Arkell and Ucko 1965: 154; Baumgartel 1981: 42­43; Trigger 1983: 33-34). On the other hand, it is clear that, to some extent, the Gerzean developed out the earlier Amratian culture (Arkell and Ucko 1965: 154). The beginning of the Gerzean can probably be placed somewhere around 3450 BC (Hassan 1988: 141) and it is believed that it originated in and expanded from the former Amratian core area (south of Abydos) near Thebes (Kaiser 1956; Arkell and Ucko 1965: 153; Trigger 1983: 35). The distribution of Gerzean sites is widespread, reaching into Lower Egypt (during late Gerzean times) and spreading further into Upper Egypt near the Nubian border.

88

Comparisons of Culture Groups

clay mixed with calcium carbonate (Baumgartel 1981: 40 via O.H. Myers; Trigger 1983: 33).

the transition period between the end of the Predynastic and beginning of the 1st Dynasty (Arkell and Ucko 1965; Hoffman 1979; Trigger 1983; Hassan 1988). It is during this time that Gerzean societies of Upper Egypt began to make a concerted effort to occupy and colonize Lower Egypt, which finally resulted in the political unification of Egypt by Menes (via the Scorpion King ?) around 3100 BC (Emery 1961; Hoffman 1979; 1982; Hassan 1988).

However, finely flaked knives of the Amratian tradition continue to be produced within the Gerzean culture. A good example of this continuation are the fishtail points which are present in both the Amratian and Gerzean cultures. Nevertheless, the basal shapes (“tail fins”) of the points change from “U” to “V” shaped during Gerzean times (Arkell 1975: 46). There are also exquisitely made parallel flaked (“ripple”) knives which are unique to the Gerzean. Overall, its chipped stone industry is strikingly different from the Amratian (which was macrolithic and bifacial) and consists of a very sophisticated blade industry where the production of blade tools (basically microlithic) appears to have been done by specialists (Baumgartel 1981: 39). The mass mining of flint was also a common Gerzean practice which was probably done by special laborers (ibid.). Hollow-based arrow points (which were common in the early and middle Predynastic) were also replaced by the more common Gerzean chisel-ended form (manufactured from blades and already present in Mesopotamia) which persisted through the Dynastic period (Baumgartel 1981: 39, 42). Pear-shaped maceheads, which were common earlier throughout southwest Asia, also make their appearance during Gerzean times (Petrie 1921: 22; Arkell 1975).

The A-Group of Lower Nubia Like Upper Egypt, there does not appear to have been a true Neolithic development in Lower Nubia (between the 1st and 2nd Cataract) as there was in Lower Egypt. In the far south of Lower Nubia near the 2nd Cataract, there were ceramic-bearing cultures such as the Khartoum Variant (Shiner 1968: 769-790) which was affiliated with the Early Khartoum Horizon (Hays 1971: 131). However, the geographic distribution of the Khartoum Variant was very limited and probably extended no farther north than 50 km down river from the 2nd Cataract in the Ballana area (Nordstrom 1972: 10). The ceramic-bearing culture represented by the Abkan Group is unquestionably Neolithic or post-Neolithic in date (see below), however, this group was more or less restricted to the Barn el Hajar region and probably did not extend farther north than Wadi Haifa (Nordstrom 1972: 16).

Large, utilitarian types of copper artifacts such as axes, knives, chisels, celts, and adzes were also introduced into the Gerzean culture for the first time (Petrie 1921: 25-26; Trigger 1983: 33). Many of these copper artifacts were cast instead of cold hammered and were probably imported from the Sinai, where the copper was mined (Trigger 1983: 33­ 34). Gold and silver were also being worked within the Gerzean culture and many wood and bone luxury items (including knife handles) were sheathed with these metals. Gold may have been particularly important to the later Gerzean populations of Egypt because it was a natural resource in the Eastern Desert (probably near the Red Sea Hills), within easy access of the Nile through the Wadi Hanunamat (Trigger 1983: 39). Based on the availability of gold within Upper Egypt, this newly found resource may have prompted the Mesopotamians to make more meaningful contacts with the Predynastic Egyptians. The significantly higher frequency of Mesopotamian artifacts in Upper Egypt (particularly within burials at Naqada and Hierakonpolis), as opposed to the few found in Lower Egypt, probably indicates that the Mesopotamians made contact with the Gerzean Egyptians across the Red Sea and along Wadi Hammamat instead of through the Delta and up the Nile River (Trigger 1983: 39-40).

In the region north of Wadi Haifa to the 1st Cataract the first ceramic-bearing culture which is known is the A-Group (Reisner 1910; Nordstrom 1972). The A-Group is a postNeolithic development (copper artifacts are present) and is closely affiliated with the Upper Egyptian Predynastic in terms of customs, material culture, and economy (Reisner 1910; Nordstrom 1972; Arkell and Ucko 1965; Baumgartel 1981). The Nubian A-Group is principally defined by a number of cemeteries and associated habitation sites which occur along the Nile Valley as far north as Kubania (approximately 10 km north of Aswan) to as far south as Sawas, approximately 10 km south of the 2nd Cataract (Reisner 1910; Nordstrom 1966: 63; 1972: 17: Adams 1977: 120). Several A-Group-like burials have also been reported farther to the south near Khartoum (Arkell 1949: 99-106). Nevertheless, the vast majority of A-Group sites fall within the traditional boundaries of Lower Nubia which was restricted to the Nile Valley between the 1st and 2nd Cataract. Overall, the Nubian A-Group is best distinguished by its burial complex (due to excavation biases), and cemeteries are by far the most common A-Group site. A-Group cemeteries contain as few as 17 individuals to as many as 150 individuals at the largest A-Group cemetery - Cemetery 7 at Sheilal (Adams 1977: 127). Typical A-Group graves are oval (either dug straight into the ground or are “niched” at the bottom of the shaft with a side chamber) and contain one or more individuals who are usually flexed and lying on their sides with the hands close to the face. In the Egyptian Predynastic tradition, A-Group burials contain a wide array

The proliferation of material wealth, possibly as a response to an increase in trade with southwest Asia, and the construction of more elaborate burial tombs, signifying the emergence of an elite ruling class, marks the latter part of the Gerzean culture, often referred to as Semainean, Protodynastic, or Terminal Predynastic, which demarcates

89

The Butana Group Ceramics

of grave goods which includes pottery, ground and chipped stone implements, carved bone, and occasionally copper tools and ornaments, and in some instances, gold.

I & II) and is associated primarily with Reisner’s “Early Predynastic Period” burials at Cemetery 17 at Khor Bahan (Nordstrom 1972: 28; Reisner 1910: 114-141). Distinctive ceramics associated with the Early A-Group burials include mostly black-topped vessels (identical to Amratian types in the Egyptian Predynastic), in addition to black-brown or red-brown polished vessels which were decorated with white filled incised lines (Reisner 1910: 316-319, plate 60a; Nordstrom 1972: 28). Other diagnostic grave material associated with the Early A-Group include disk-shaped maceheads (of the Amratian type), slate pallets (oval, diamond-shaped, and crescent-shaped) alabaster plumb bob-shaped pendants, tanged arrowheads, clay figurines, ivory bracelets, and rings (Reisner 1910: 316-319, plate 62a-c). Other than the tanged arrowheads, these particular artifacts were also common in Amratian burials in Upper Egypt (Reisner 1910: 319). Indeed, the Amratian-like burials at Khor Bahan are indistinguishable from their northern counterparts in Upper Egypt (Reisner 1910: 319; Nordstrom 1972: 28).

The ceramics are the most distinctive and ubiquitous artifacts in A-Group cemeteries (as well as in the habitation sites) and consist of burnished red-finished, brown, and black vessels which have often been black-topped (Reisner 1910: 314-332; Nordstrom 1966: 63; Adams 1977: 119). Ripple wares are also very prominent. Other A-Group ceramics include vessels which have been decorated with incised geometric lines filled with a white pigment or clay which are very similar to Amratian incised black wares (cf. Petrie 1901: 14; Reisner 1910: 319). The chipped stone associated with the A-Group consist mostly of flake tools with a low index of lunates (Haaland 1972: 116). Blades and bifacially worked tools are almost totally absent (ibid.). There are, however, a high number of denticulates (ibid.). The majority of lithic tools were produced either from chert or quartzite derived from local Nile pebbles (Haaland 1972: 97, 109). Oval-shaped ground stone in the form of manos, metates, and tabular palettes are also common and were probably used for grinding grain and pigment (Nordstrom 1972). It is worth noting that the documented lithics from the earliest A-Group burials near the 1st Cataract consisted of large blades (some of which were retouched on the end), hollow-based, tanged, and chisel-shaped arrow points (Reisner 1910: plate 62). Other common artifacts associated with the A-Group include; alabaster vessels, copper tools (including awls, axes, adzes, and knives), bone and ivory implements (including spoons, awls, bangles, and combs), jewelry (made of ground and polished ostrich egg-shell, faience, carnelian, rock crystal, and chert), carved shell, and clay figurines (Reisner plates 63-67; Nordstrom 1972: plates 190-197).

The Early A-Group burial complex first appeared in the northern end of Lower Nubia and then spread southwards up the Nile Valley towards Wadi Allaqi (Nordstrom 1972: 28). Geographically, the distribution of the Early A-Group cemeteries ranged from Kubania to the Dakka-Sayala area at the confluence of Wadi Allaqi and the Nile (Trigger 1965: 69; Nordstrom 1972: 28). Based on the close similarity of burial goods shared with both the Early A-Group and the Predynastic Amratian, the origin of the Early A-Group may have been a result of a southern spread of the Amratian culture into Nubia either as a result of people moving south from Egypt, or an adoption of the Amratian culture by local Nubians (Reisner 1910: 319; Kaiser 1957; Fairservis 1962: 71; Trigger 1965; 67-70; Nordstrom 1972: 28). It is also important to note that the cranial indices and overall physical makeup of both A-Group and Naqada I-III Predynastic individuals are very similar (Reisner 1910; via Elliot Smith; Batrawi 1935: 165; Nordstrom 1972: 18; Adams 1977:94).

The A-Group essentially coincides with the Middle and Late Predynastic (Naqada I and II) of Upper Egypt and lasts on up to the beginning of the 1st Dynasty (Reisner 1910 314-332, 347; Kaizer 1957; Nordstrom 1972: 18). A chronological range for this group has been given between 3400 to 2900 BC (Nordstrom 1972: 30; Hassan 1986: 92). Climatically, the Nubian A-Group would have existed during the Neolithic humid phase (Trigger 1965: 67) and it is possible that the lack of any pre A-Group ceramic-bearing cultures in Lower Nubia north of Wadi Halfa may be due to the mid Holocene arid phase which occurred roughly from 5500 to 4500 BC.

The Classic phase represents the maximum extent of the A-Group throughout Lower Nubia and can be detected by numerous cemeteries throughout the region (cf. Trigger 1968: 66). The diagnostic hallmarks of the Classic A-Group are the black-topped, red-finished ripple wares which are ubiquitous throughout the Classic A-Group burials (Nordstrom 1972: 29). Based on the ceramics (through the overall continuation of black-topped wares) it appears that the Classic A-Group developed out of the Early A-Group and was contemporary with the late Gerzean (Naqada III phase) Predynastic culture of Upper Egypt (Nordstrom 1972: 29). In contrast to the Early phase, however, the Classic phase appears to represent a truer Nubian A-Group manifestation which had spread south from the Wadi Allaqi area into the Batn el Hajar region south of the 2nd Cataract (Trigger 1965: 70-73; Nordstrom 1972: 29). Thus, it appears that this phase represents a blending of traits from the Amratian and Gerzean cultures, represented

The A-Group has essentially been divided into three temporal phases (the Early [Trigger’s Bahan phase], the Classic [Trigger’s Nubian I phase], and the Terminal [Trigger’s Nubian II phase]), based on a generalized sequence of the locally made ceramics in connection with a procession of diagnostic Egyptian artifacts (Nordstrom 1972: 28-29; Trigger 1965: 68-78). The Early A-Group was contemporary with the Egyptian Predynastic cultures of the Amratian and Gerzean (Naqada

90

Comparisons of Culture Groups

through A-Group burial customs, with a more indigenous hunting and gathering complex, perhaps originating from the Abkan Group, which had been existing in Lower Nubia somewhere south of Wadi Allaqi towards the 2nd Cataract (Trigger 1965: 70; Nordstrom 1972: 29). This amalgam of traits can be best seen in the chipped stone lithic assemblage of the Classic A-Group which in some ways resembles the Abkan Group (Haaland 1972: 112). It is equally possible, however, that “cultural drift” may have occurred, at least north of Wadi Halfa, resulting in the development of a new “Nubian” society once populations of the A-Group established themselves more firmly in Lower Nubia. It is important to note that vessels associated with the Classic A-Group in the southern portion of lower Nubia near the 2nd Cataract were more often round-based (more Nubianlike) as opposed to a higher frequency of flat-based vessels (more Amratian-like) associated with the Early A-Group farther to the north near the 1st Cataract. Furthermore, the Early A-Group burial ceramics of the 1st Cataract were composed of a sandy clay paste (like typical Amratian burial wares) while the pastes associated with the Classic A-Group of the 2nd Cataract were always fiber tempered (Reisner 1910, Nordstrom 1972). One should also bear in mind that Abkan ceramics in the Batn el Hajar region were never fiber tempered.

The people associated with the A-Group in general were more than likely practicing a mixed economy involving agriculture and stock raising with some hunting and gathering (Firth 1912: 5-13; Lal 1963; Trigger 1965: 67; Nordstrom 1972: 23-24). The economy of the Nubian A-Group changed very little from the beginning of the Early A-Group phase through the end of the Terminal A-Group phase (Nordstrom 1972: 23-25). From the rather small number of habitation sites (less than 50 in all of Lower Nubia) which have been recorded, A-Group settlements appear to have been situated along the flat aridable stretches of the Nile River which would have been good for simple floodplain agriculture (seluka) where seasonal deposits of silt would have been sown with seed crops and then cultivated (Trigger 1965: 13, 72, 75; Nordstrom 1972: 23). From the very scant macrobotanical remains recovered from A-Group sites, both barley (six row) and wheat (Emmer) have been found, as well as lentils and peas (Lal 1967; Adams 1977: 125). The raising and tending of livestock may have been an important A-Group practice (cf. Firth 1912); however, surprisingly, only a few bones of goat/sheep, and cattle have been found (Reisner 1910: 115; Nordstrom 1972: 23-24). Nevertheless, the dominance of dung tempered ceramics throughout the Classic and Terminal A-Group sequence indirectly supports the assumption that cattle were raised in Lower Nubia during this time (Nordstrom 1972: 24). In addition, many of the A-Group burials were wrapped in ox hides (Trigger 1965: 68; Hoffman 1967: 84). In the 4th Dynasty, during King Sneferu’s reign, is the famous inscription on the Palermo stone which reads “Hacking up the Nhskw [Nubians], bringing [home] seven thousand living captive and 200,000 cattle (Smith 1966: 119 via Schafer n.d.). These particular Nubians of course, would be post A-Group in date, however, it is quite likely that pastoralism would have existed prior to King Sneferu’s raid into Lower Nubia (Trigger 1965: 68). It also appears that A-Group people fished and hunted wild game. Reisner (1910: 216) reported fish bone from the “The Archaic Camp” (the only A-Group habitation site he recorded) and copper fish hooks were imported by A-Group people from Egypt (Reisner 1910: Plate 65 a &b; Nordstrom 1972: 24). The remains of gazelles and crocodiles have also been found in A-Group graves, while the depictions of other savanna adapted animals such as giraffes, elephants, and antelopes have been drawn on pottery vessels (Hoffman 1967: 117). In the Batn el Hajar region, several hundred remains of gazelle were also recovered which date somewhere around 3000 BC (Adams 1977: 126).

The Terminal A-Group is a continuation of the Classic phase and corresponds to the period spanning the Naqada III phase (Late Gerzean) and the 1st Dynasty of Egypt (Nordstrom 1972: 29). There are essentially no changes in the spatial distribution between the Classic and Terminal A-Group. The distinguishing criteria for the Terminal A-Group are again seen in the ceramics where a very thin walled, highly burnished, red painted (red on buff) ceramic type (indigenous to Lower Nubia) is introduced (Nordstrom 1972: 29). As before, there is a continuation of Egyptian goods associated with the Terminal A-Group burials which date between the latter part of the Naqada III phase and the very beginning of the 1st Dynasty. There are, however, no Terminal A-Group burials which have any Egyptian goods post dating the beginning of the 1st Dynasty (Smith 1966: 118; Nordstrom 1972: 29). When the Nubian A-Group material was first being described, it was felt that after Naqada III times, the Nubians chose to retain the earlier styles of the Egyptian Predynastic period by rejecting the later Egyptian Dynastic styles (Reisner 1910: 320-332). Therefore, at the time, it was believed that the A-Group graves containing transitional Naqada III and 1st Dynasty artifacts were actually later and were essentially contemporary with the Dynastic period (ibid.). However, out of the many thousands of A-Group burials which had been excavated since Reisner’s original assessment, there was not a single Egyptian artifact dating after the end of the 1st Dynasty associated with these graves (Smith 1966: 118). Thus, it was finally concluded that the Terminal A-Group (as a material cultural entity) had actually ended by the beginning of the 1st Dynasty (Nordstrom 1972: 29).

Overall, A-Group habitation sites consisted of small ephemeral encampments which may have only been seasonally occupied (Adams 1977: 123). Of the only habitation site (“The Archaic Camp”) Reisner described, he said it was small “only large enough for a half-a-dozen families” containing “pottery and other objects...16 fireplaces or hearths-irregular round spots packed with ashes and wood coals, in some cases mixed with broken animal bones; but there were no traces of walls” (Reisner 1910: 215-216). Other habitation sites described latter by Trigger

91

The Butana Group Ceramics

were “composed of ash deposits, broken animal bones, and artifacts, which vary in thickness from superficial to over a meter thick” (Trigger 1965: 76). Of these sites he noted that they “seemed to be temporary” (ibid.). The habitation sites reported by Nordstrom also seem to be fairly small encampments (ranging from 3500 to 7000 m2) with thin, deflated scatters of lithics and ceramics (Nordstrom 1972: 20, 134­136; 140-144; 150-159). Post molds (as well as the remains of some framing) were common on many of the A-Group habitation sites suggesting that small, flimsy wood frame houses covered with grass or reeds were built (Adams 1977: 125; Nordstrom 1972: 20). A few larger sites (a total of four) have been found which may have represented more permanent villages such as Afyeh (Lal 1963; Trigger 1965: 77). This particular A-Group site was 12,000 m2 (small in comparison to the average sized, contemporary Predynastic site) and contained a number of stone wall and rectangular house foundations, one of which had 6 rooms (Trigger 1965: 77). Another rather anomalous A-Group site consisted of a cluster of some 578 storage pits (no nearby habitation site was found) located near Khor Daud at Wadi Allaqi (Piotrovsky 1967; Nordstrom 1972: 26). These storage pits contained a vast amount of pottery (mostly of Egyptian origin) which contained beer, wine, and oil. There is little question that this particular site represented some kind of trading post where finished Egyptian goods were probably exchanged for Nubian raw materials.

to import clothes, honey, ointment, and oil (Breasted 1906: 167: Trigger 1965: 70-71). These particular products of course, were luxury items which were highly desired by the Nubians; however, products such as grain (which may have been more vital) may have been imported, as well. In return for Egyptian goods, the Nubian A-Group people principally exported ebony and ivory and perhaps the skins of cattle and leopards; and possibly slaves (Save-Soderbergh 1941; Trigger 1965: 70-71; Nordstrom 1972: 25-26; Adams 1977). What is striking about the trade between the Egyptians and the A-Group is that the latter culture suddenly ceased to exist with the establishment of the 1st Egyptian Dynasty. Based on the lack of any good data to suggest that there was a deteriorating climate during this time, there is a strong indication that the demise of the A-Group was political probably resulting in a change of policy (i.e. shifting from peaceful trade relations to raiding) towards the Nubians after the unification of Lower and Upper Egypt (Nordstrom 1972: 29-32). Thus, the breaking off of trade ties between Egypt and Nubia could have been devastating for the latter group. This coupled with the apparent raiding by Egypt into Lower Nubia probably were the chief causes which ended the A-Group culture (Smith 1966; Nordstrom 1972: 25-26). It should be noted that the sudden ending of the A-Group probably did not mean a wholesale abandonment of populations from Lower Nubia. It is very likely that people associated with the A-Group continued to live in Lower Nubia, but became more nomadic and less materially oriented. As a result, the evidence for these kinds of occupations may be much harder to recognize in the archaeological record (cf. Nordstrom 1972: 32).

Trade between the A-Group people and Egyptians cannot be underestimated and was very substantial, if not critical, to the Nubians. Given that Lower Nubia had significantly less aridable land than Egypt, as a whole, it is possible that A-Group populations (estimated to be around 8000 people, cf Trigger 1965) were receiving food products from Egypt (in exchange for other goods, probably natural resources such as ivory, skins, wood, etc.) to supplement their harvests which may have been limited. Trigger states that Lower Nubia had a history of “not produc[ing] enough food for its inhabitants” (Trigger 1965: 19). He cites a traveler’s observation in 1737 about the scarcity of food in Nubia:

The Abkan and Karat Groups of the 2nd Cataract and the Dongola Reach The Abkan and Karat Groups represent ceramic-bearing cultures which developed during the Neolithic and postNeolithic eras in the Batn el Hajar (2nd Cataract area) and the Dongola Reach areas, respectively (Myers 1958, 1960; Shiner 1968, 1971; Marks et al. 1970; Nordstrom 1972). These particular groups are considered post-Early Khartoum in date and share some ceramic characteristics with the A-Group in the north and the Khartoum Neolithic in the south (ibid.). It is equally clear, however, that the Abkan and Karat Groups are separate cultural developments from either the Khartoum Neolithic or the A-Group and portray different subsistence and cultural patterns.

It was extremely difficult to obtain provisions anywhere above Aswan, for the people possessed very few sheep, goats or chickens: the narrow strips of cultivable land on the river bank did not produce grain sufficient for the needs of the people, and the natives never ground more dhurra than they needed for their immediate use (Trigger 1965: 19 via Budge 1907: 17).

The Abkan Group of Batn el Hajar.

It is intriguing to speculate that the overall scarcity of cultigens and animal bones on A-Group sites may have had more to do with the logistical problems of A-Group populations being able to raise large quantities of food as opposed to poor archaeological preservation. Indeed, it is known that the A-Group Nubians were receiving large quantities of beer and/or wine which was transported in Egyptian amphora (Nordstrom 1972: 25). In the 6th Dynasty, an Egyptian official listed that the Nubians liked

The Abkan Group was first defined by O.H. Myers in the late 1940’s when he surveyed a small area along the Nile Valley south of the 2nd Cataract around the town of Abka (Myers 1960: 174). The Abkan Group was principally defined by the excavation of one site (No. IX) a rockshelter, located on the eastern edge of the Nile which consisted of a “multiple pothole about 15 meters across” full of stratified archaeological deposits (Myers 1958: 131-135). Many rock drawings of humans, trees, animals, and other objects

92

Comparisons of Culture Groups

(including addax, rhinoceros, cattle, giraffe, elephants, ostrich, gnu, a baobab tree, and possible fish traps) of various ages had been drawn on the walls of the rockshelter (Myers 1960: 175). Unfortunately, Myers was never able to publish a definitive report on his survey of the Abkan area, and only a few cursory remarks about some of the Abkan Group sites and associated artifacts were printed (Myers 1958, 1960). However, through the later investigations of Shiner and Nordstrom, in the same area, more on the Abkan Group was elucidated (Shiner 1968; Nordstrom 1972). Based on these latter findings, Myer’s cursory descriptions of the Abkan were surprisingly accurate, given the brief time he had to investigate this group (Shiner 1968: 625).

bowls to cups to dishes (Nordstrom 1972: 15). The paste of Abkan pottery is sandy in texture and always contains mineral inclusions (composed primarily of quartz sands). It is believed that the paste was derived from alluvial sediments from the Nile River (Myers 1960: Carlson 1966: 56; Nordstrom 1972: 15, 49, 57). None of the Abkan ceramics have been tempered with dung. Abkan ceramics can basically be divided into a smoothed and a burnished ware; the first of which consists of thick walled vessels which have been smoothed or sometimes wiped on the exteriors, while the second ware consists of thin to medium walled vessels which have always been burnished on the exterior (Carlson 1966: 56-58; Nordstrom 1972: 14). A few of the of the burnished vessels have been decorated with parallel rows of impressed “vees” (triangular impressions) or zigzags (continuous lined rocker stamped impressions) patterns. Less frequent decorations consist of herringbone and short line incised patterns (Nordstrom 1972: 14-15). These particular impressed and incised ceramics are similar in style to comparable ceramics recovered from Khartoum Neolithic sites further to the south (see below). Ripple and black-topped red-finished wares are also present in some of the later Abkan ceramic assemblages and are very reminiscent of A-Group ceramics farther north (Carlson 1966: 58, 62; Nordstrom 1972: 17). Unlike the A-Group pottery, however, the Abkan ripple and black-topped wares have no dung temper (Nordstrom 1972: 17, 58-59).

Abkan Group sites are basically restricted to a 150 km stretch of the Nile Valley in the Batn el Hajar region between the 2nd Cataract to the north, and about 15 km north of the Dal Cataract in the south (Nordstrom 1972: 16). Abkan settlements tend to be more frequent on the eastern side of the Nile in this area where the otherwise barren rock out crops of the Batn el Hajar have been dissected heavily with small narrow wadis and sloughs of the Nile River (Shiner 1968: 541; Nordstrom 1972: 15). Some Abkan sites are also situated on the west side of the Nile, however, in similar micro-geographic settings (ibid.). Abkan Group sites range in size from 80,000 m2 to much smaller ones averaging only 750 m2 in size (Nordstrom 1972: 15). It is believed that the larger sites represent base camps, while the smaller sites reflect less occupied temporary settlements (ibid.). Cultural deposits on these sites are not deep, usually less than 50 cm (Carlson 1966: 54). No remains of housing has been found so far on any Abkan sites, however, it has been suggested that small reed houses (either circular or rectangular) were built (Nordstrom 1972: 15). Small hearths consisting of small cobbles and some ash have been recovered from some Abkan sites (Carlson 1966). Burials have not been found on any site associated with the Abkan Group.

The lithics are an important aspect to the Abkan Group, since it was suspected that they developed from an earlier pre-ceramic tradition (Shiner 1968). Indeed, if this was the case, the Abkan Group would have been the only known ceramic culture in all of the Nile Valley of northeast Africa which developed out of an indigenous terminal Paleolithic tradition. Recently, however, this contention has been disputed (cf. Connor and Marks 1986: 175176). Based on gradual proportional differences between particular tool types, Shiner had suspected that the Germain developed into the Qadan which in turn developed into the Abkan. Nonetheless, these particular changes in tool type frequencies appear to be more generic in nature and fail to link the Abkan genetically with earlier, Nilotic Paleolithic industries (Connor and Marks 1986: 176).

Radiocarbon dates for the Abkan are few, but, it has been estimated that the group existed sometime prior to 3000 BC perhaps as early as 3800 BC; or even earlier (Carlson 1966: 55; Nordstrom 1972: 17; Hassan 1988: 139). Due to the similar ceramic traits shared between the Abkan and the A-Group, the earliest A-Group settlements in the 2nd Cataract area may have been contemporary with the latest Abkan Group settlements (Shiner 1968: 626; Nordstrom 1972: 17). Nevertheless, it is quite clear that the Abkan predates the A-Group by the fact that the former developed out of an older, indigenous preceramic tradition (Final Qadan of the Cataract Tradition) and possesses ceramic traits also seen in the Khartoum Neolithic to the south (Myers 1960; Shiner 1968: 626; Nordstrom 1972).

The Abkan lithic industry is basically composed of a high number of denticulates, lightly retouched scrapers, groovers, and borers, the latter two of which are considered the hallmark of the industry (Myers 1960, di Cesnola 1960; Shiner 1968: 611-612, 625-626). There are very few lunates, backed flakes or blades in this industry, while at the same time there are many of unretouched flakes (Myers 1960: 175; Shiner 1968: 625). The majority of the Abkan chipped stone tools were produced from Nile quartz pebbles (Shiner 1968: 625). Other lithic tools associated with the Abkan Group include grinding stones (which are plentiful on Abkan sites) and a few polished stone celts (Nordstrom 1972: 14). Imported Egyptian flint (probably derived from some place northwest of the 2nd Cataract) has been found on some Abkan sites, as well (Nordstrom 1972: 13).

Abkan occupations are primarily recognized by the ceramics which tend to be plain (however, some vessels were decorated) and consist almost entirely of simple contour forms with round bases. Within Abkan assemblages, there are a variety of different shaped vessels, ranging from

93

The Butana Group Ceramics

Finely ground disk-shaped palettes were also recovered (Nordstrom 1972: 14).

spearing, or net sinkers were found in any of the Abkan sites relative to the high number of fish bone (Shiner 1968: 625; Nordstrom 1972: 15). Based on a single rock drawing of a fish trap at Myers’ “pothole” site, such items may have been used instead (Myers 1958; Shiner 1968: 625). The remains of gazelle, wild cattle, zebra or wild ass, have also been found on Abkan sites in varying amounts, which attests that some emphasis was placed on hunting larger game (Nordstrom 1968: 16). However, due to the lack of animal bone overall, and a striking decrease in the use of points, Shiner believed that there was a significant decline in hunting wild animals from the Qadan to the Abkan (Shiner 1968: 627). Nevertheless, it was noted that on other Abkan sites (not investigated by Shiner) there were significant quantities of bone (Nordstrom 1972: 16). For example, on one Abkan site (the Karagan site) a large quantity of gazelle remains was reported, with lesser amounts of wild ass, cattle (presumably wild) and hare (Carlson 1966: 61). At another site (AS 6-G-25), the remains of birds (Egyptian Goose) and one metapodial fragment of a domestic goat (Capra hircus) was found, indicating that domestic animals were being used in limited numbers by Abkan peoples (Perkins 1965: 58).

Other kinds of artifacts made from either shell or bone are rare in Abkan assemblages (Nordstrom 1968: 13-14). However, several fragments of pointed bone (probably awls) have been found on one Abkan site (Carlson 1966: 61). Ostrich eggshell is more common, some of which has been engraved (Nordstrom 1972: 14). On occasion, ringshaped beads manufactured from the same material have also been found (ibid.). Based primarily on Shiner’s analysis of the Abkan lithic industry, Nordstrom constructed three cultural stages of the Abkan Group (Nordstrom 1972: 17). The first stage is the Early Abkan which is believed to have directly developed out of the Final Qadan based on the changes in lithic tools which in turn may reflect a shift in lifestyle where the hunting of wild game was de-emphasized (Shiner 1968: 626). Of course, the presence of ceramics demarcates the beginning of the Early Abkan phase, however, they are rare at this stage (Nordstrom 1968: 17). It is important to note that burnished ceramics (signifying a post-Early Khartoum development) are present during this stage, and it is likely that the idea of ceramic manufacture was introduced into the Abkan Group from the outside.

The Karat Group of Dongola Reach. The Karat Group is represented by twenty-five sites along a 16.5 km stretch of the Dongola Reach area (Marks and Ferring 1971: 190). To date, the Karat Group represents the only known ceramic-bearing group in the Dongola Reach area which is Neolithic in age. This group was first defined by Marks and Hays in 1966 when a 80 km survey was conducted in the Dongola Reach area along the west bank of the Nile River between Wadi el Melik and Wadi Mugaddam (Shiner 1971: 8). The limited distribution of the Karat Group sites indicated that these sites were located only in a particular stretch where there was adequate vegetation and that much of the surveyed area up and down river from the Karat area was less desirable for settlement, due to rockier sediments which could not support plant life (Marks and Ferring 1971: 270-271). No radiocarbon dates were recovered from any of these sites; however, the ceramics indicate that the Karat Group was contemporaneous with both the Khartoum Neolithic and the A-Group (Marks and Ferring 1971: 190). Of course, with the recalibration of C14 dates in general, it is now known that the Khartoum Neolithic pre-dates the A-Group (Hassan 1986). As a result, the relative dating of the Karat Group still remains enigmatic.

The second stage is called the Developed Abkan and is recognized by a high number of denticulates, groovers, and borers (ibid.). There is also a lot of quartz debitage, and the number of ceramics increases. Among the decorated ceramics, the impressed zigzag pattern is predominant, followed by other impressed and incised patterns. The third stage is called the Terminal Abkan, which can be distinguished by the introduction of new ceramics styles including ripple wares, polished red wares, and black-topped wares (ibid.). Based on these new ceramics it is possible that the Terminal Abkan of the 2nd Cataract district was being influenced by Nubian A-Group cultures who were pushing southward into the Batn el Hajar region. There does not appear to have been a cultural displacement of the former by the latter group, however. This can be argued on the basis that the lithics of the Terminal Abkan changed little from the Developed Abkan and that the ceramic pastes does not change at all. Thus, it is possible that in some parts of the Batn el Hajar region, both the A-Group and the Abkan existed contemporaneously. The subsistence patterns of the Abkan inhabitants seems to have been riverine oriented, revolving around, fishing, hunting, and gathering, supplemented perhaps with a few domestic animals (Nordstrom 1972: 15-16). Fishing and shelling seem to have been predominant due to the large number of fish bone and shell found throughout Abkan middens (Shiner 1968: 625). Nile Perch (Lates niloticus) and Catfish (Clarias) were the most common fish found, alongside fresh water mollusks (Nordstrom 1972: 15). It is puzzling, however, that no fish hooks, bone implements for

Karat Group sites (all of which are surface sites) contain a sparse scattering of lithics and ceramics and are mainly characterized by piles of heat-treated chert pebbles and cores (Marks and Ferring 1971: 196, 203). These piles (ranging in size from 0.5 to 1.5 m) of heat-treated chert pebbles and cores are unique in all of the Nile Valley and probably represent activities related to the heat treatment of lithics for better flaking qualities (Marks and Ferring 1971: 204-209). The investigators also believed that the piles of

94

Comparisons of Culture Groups

heated chert may have also been used to cook gazelle and/ or goat (Marks and Ferring 1971: 208). This latter inference was based on ethnographic practices in the Hoggar, where similar piles of stones were used for the treatment of milk and the cooking of goat and sheep stomachs (Marks and Ferring 207-208). The size of these artifact scatters and associated piles of chert cobbles were very small, ranging between 65 and 310 m2 (Marks and Ferring 1971: 203).

In comparing the Karat Group with other known contemporaneous groups along the Nile, the marked differences in lithics strongly suggested that the Karat Group was independent from the Khartoum Neolithic which was situated 700 km farther to the south (Marks and Ferring 1971: 256-267). For example, the Khartoum Neolithic assemblage was dominated by geometries, while the Karat Group was dominated by end scrapers (Marks and Ferring 1971: 264). It was noted, however, that the Karat Group ceramics, both in paste and decoration, were similar to those of the Khartoum Neolithic (Marks and Ferring 1971: 265-266). Other lithic and ceramic comparisons were made with other groups farther north such as the A-Group. At the time, however, the A-Group was still in the process of being systematically defined, and as a result, little comparative work could be done (Marks and Ferring 1971: 265). What little data could be salvaged from the A-Group did indicate that all of the decorated Karat Group ceramics could be subsumed within the A-Group which possessed a larger array of ceramic styles (Nordstrom 1966: 66, figure 1, numbers 1-5, 12, & 14; Marks and Ferring 1971: 265). However, it was latter established that A-Group ceramics were dung tempered (Nordstrom 1972: 57), while the Karat ceramics were not. Due to the same lack of comparative data, the Karat Group was also not compared with the Abkan Group of the 2nd Cataract (however, the lithics of one non ceramic-bearing site of the 2nd Cataract were compared). However, when a fuller understanding of the Abkan was provided, it appeared that the Karat Group was significantly different from the Abkan in terms of the lithics (Nordstrom 1972: 16-17). For example, the Abkan Group was dominated by denticulates, groovers, and borers, while there were few such tools in the Karat Group (cf. Shiner 1968: 621 and Marks and Ferring 1971: 264). Nevertheless, when the ceramics of the Developed Abkan were compared with the ceramics of the Karat Group, clear similarities could be seen between the two (ibid.). The Abkan Group ceramics, however, do lack the dentated rocker stamped decorations which were dominant in the Karat Group (cf. Nordstrom 1972: plates 139 & 141), while decorations consisting of parallel rows of “vees” (common in the Abkan Group) were lacking in the Karat Group (Marks and Ferring 1971: 252-256). Based on the ceramic data, it seems that the Karat Group does share some characteristics with both the Abkan and the Khartoum Neolithic, but it shares less characteristics with the A-Group. Recently, it has been suggested that the presence of the Karat Group in the Dongola Reach area may represent temporary encampments of other Saharan groups farther to the west who periodically visited the Nile Valley (Marks and Caneva 1990a). In light of modern ethnographic observations on the seasonal movement of nomads from the western Sudan to the Nile Valley (cf Barbour 1964), this kind of scenario would fit well with the Karat Group.

Other than lithics and ceramics, no other kinds of cultural debris such as bone and shell were present on any of the Karat Group sites. Due to the light scattering of lithics and ceramics overall, it was believed that the occupations of these Karat Group camps were very temporary where the occupants may have stayed for only a couple of days to process workable rock for tool manufacture and the cooking of meat (Marks and Ferring 1971: 271). It is interesting to note that the burned rock piles on the Karat Group sites were always situated to the north of the artifact scatters and it was hypothesized that these rock piles were also used to blow smoke down wind to ward off gnats (Marks and Ferring 1971: 272-273). The burning of fires to ward off gnats in this manner was still being done by modern-day folk living in the region (ibid.). The ceramics of the Karat Group consist of a small assemblage of smoothed and burnished sherds, of which 40 percent have been decorated with impressed patterns consisting of dentated rocker stamped (para-wolves tooth and pars-saw tooth), continuous rocker-stamped (wolves tooth and saw tooth), twine impressed (linear) and some punctation decorations (Gilmore in Marks and Ferring 1971: 252-256). All sherds of the Karat Group had pastes with a high content of mineral inclusion consisting of fine quartz grains (Marks and Ferring 1971: 252). No sherds with organic tempering were found. Vessels within the Karat Group consist of small, thin-walled, simple contour bowls and jars with round bases. The lithics of the Karat Group consisted primarily of chipped stone, with a strong emphasis on the manufacturing of flake tools (Marks and Ferring 1971: 258). No bifacially worked tools were present and very few blades were recovered. The dominant chipped stone tool within the Karat Group were end scrapers, followed by retouched and notched pieces (Marks and Ferring 1971: 264). Microlithic tools such as lunates were rare (Marks and Ferring 1971: 240). Chert derived from Nile River pebbles was the predominant raw material used, while quartz Nile pebble was used infrequently (Marks and Ferring 1971: 214). Of course, the lithic raw material used in the manufacture of Karat Group tools were first heat-treated. Very little ground stone (most of which was made from sandstone) was recovered from any of the Karat Group sites (Marks and Ferring 1971: 248-251). The few ground stones recovered were round or ovoid querns (matates) and manos (ibid.). The lack of ground stone suggested that very little processing of plant foods was done at any Karat Group site.

The Khartoum Neolithic of the Central Nile Valley The Khartoum Neolithic represents another early Neolithic development which occurred along the central Nile Valley

95

The Butana Group Ceramics

of the Sudan. It was thought that the development of the Khartoum Neolithic were later than the earliest Neolithic development occurring in the Delta and Fayum area in Lower Egypt, and dates of the former were given between 4000 and 3000 BC (Arkell 1953: 102; Clark 1984: 118, 122). Recent calibrated radiocarbon dates, however, have pushed back the earliest occupations of the Khartoum Neolithic to the early part of the 5th millennium BC (49004500 BC), proving that it was indeed contemporary with the early Neolithic development in Lower Egypt (Haaland 1987; Caneva 1988; Hassan 1986: 91).

(Arkell 1953: 68-78). The most popular decorations consisted of impressed “vees” (triangular impressions made from a dentated implement which was rocker stamped) and “zigzags” (continuous rocker stamped impressions which created wolf tooth or saw tooth motifs). Other characteristic ceramics designs included horizontal and concentric, semicircle incised and impressed decorations. The vast majority of vessels at Esh Shaheinab consisted of simple contour, hemispherical bowls with either direct or expanding rims. Many of the rims were further decorated on the lips with impressed or incised patterns. The ceramics in general were made from a fine paste with mineral inclusions (Arkell 1953: 69). A few Dotted Wavy Line sherds (of which some were burnished) recovered from Esh Shaheinab lead Arkell to believe that the Khartoum Neolithic had developed from his Early Khartoum culture (Arkell 1953: 103). At the nearby site of El Qoz (situated just below Khartoum in the Gezira Plain) which contained an Early Khartoum and Khartoum Neolithic component, it was demonstrated that the majority of Khartoum Neolithic sherds were stratigraphically above Early Khartoum ceramics (Arkell 1953: 97; 101-102). Among the various other ceramics recovered at Esh Shaheinab were red-finished wares, and a few atypical black-topped, red-finished sherds and ripple-like sherds which Arkell believed to be similar if not related to some of the Predynastic ceramics (in particular the Badarian and A-Group) farther to the north (Arkell 1953: 73-77, 103).

The Khartoum Neolithic was first defined by Arkell at the site of Esh Shaheinab which is situated on the west bank of the Nile about 50 km north of Khartoum (Arkell 1953). Esh Shaheinab is relatively small, encompassing an area no greater than 6000 m2 and has cultural deposits less than 60 cm deep. The faunal remains at this site reflected an adaptation heavily reliant on hunting, fishing, and gathering (Bate in Arkell 1953: 102­103). However, a few remains of goat and sheep were also found indicating that at least some of the people’s diet at Esh Shaheinab was supplemented with products produced from domestic animals (ibid.). Nevertheless, no remains of cultigens were recovered, demonstrating that the inhabitants of Esh Shaheinab were not cultivating domesticated plants (ibid.). Like the people at the Early Khartoum site, however, fruits of the Celtis tree were still heavily exploited during the Khartoum Neolithic (Arkell 1953: 105).

The smaller chipped stone artifacts at Esh Shaheinab were basically similar to those of the Early Khartoum (Arkell 1953: 26-32, 55). The lithic industry was essentially microlithic with a predominance of lunates and blades, followed by scrapers and borers. Lunates were smaller at Esh Shaheinab, though. In the larger tool category (which consisted of chipped and ground stone tools), there was the dominance of a very characteristic gouge (polished on one side and all made from rhyolite) which had no counterpart in the Early Khartoum (Arkell 1949: 1953: 31-33). Other large stone tools not characteristic of the Early Khartoum found at Esh Shaheinab were polished celts of various sizes some of which were either used as axes or adzes (Arkell 1953: 32-39). One example of a flat-topped macehead, reminiscent of the Badarian type, was also recovered (Arkell 1953: 49). Disk-shaped ground stone (usually sandstone) such as manos and other larger flat basins, in addition to doughnut-shaped rings, were also present, all of which had similar counterparts in the Early Khartoum. However, Arkell noted that ground stone overall was not as frequent at Esh Shaheinab as it was at the Early Khartoum site. Other stone material at Esh Shaheinab included beads of amazonite (originating in Tibesti) and zeolite lip plugs (originating in Ethiopia). Bone artifacts at Esh Shaheinab included harpoons, awls, and well made celts (Arkell 1953: 56-63). The harpoons at Esh Shaheinab are similar to the ones recovered at the Early Khartoum site, except that the former were perforated at the butt end instead of being grooved. The bone celts (like their stone counterparts in size and shape) were not present in the Early Khartoum.

Unlike the Khartoum Hospital site, no burials or signs of housing were found at Esh Shaheinab (ibid.). The only outstanding features found on the former site were cooking hearths which ranged in size from .5 to 1.5 m in diameter; some of which contained small cobbles of sandstone at the base (Arkell 1953: 79-80). Arkell believed that the settlement at Esh Shaheinab was inhabited on a year round basis, since no comparable sites were found in the adjacent area (Arkell 1953: 102). It was suspected that the site was occupied during a moist period, due to the fact that the large land snail, Limicolaria flammata, (which needed at least 500 mm of annual rainfall) was recovered in the archaeological deposits (Arkell 1953: 105). However, the absence of swamp loving animals such as the reed rat, the water mongoose and the small land snail Trochonanina, coupled with the addition of new kinds of animals such as bush duiker, grivet monkey, and the more forest-loving snail Limicolaria kambeul, indicated that the swampy central Nile environment, contemporary with the Early Khartoum culture, was replaced by a more drier steppe environment interspersed with riparian forests by the time Esh Shaheinab was occupied (ibid.). The material culture of Esh Shaheinab consisted mainly of ceramics, chipped and ground stone, carved bone and shell. The ceramics at Esh Shaheinab were very characteristic and included a wide variety of impressed and incised vessels, the majority of which had been burnished on the exteriors

96

Comparisons of Culture Groups

Shell (Nile bivalves such as Ampullaria, Aetheria elliptica, and Aspatharia rubens) also was commonly used at Esh Shaheinab and finely made barbless fish hooks (notched on the end to secure fish line) were manufactured from them (Arkell 1953: 66). It was suspected that some of the shell was also used to thin, as well as to decorate (especially the curved “zigzag” patterns), some of the Khartoum Neolithic ceramics (Arkell 1953: 67).

Nevertheless, the ceramics at both sites were similar to one another, and the characteristic Esh Shaheinab celts and gouges made of rhyolite and bone were also found at Kadero (ibid.). The site of Shaqadud (situated about 50 km northeast of Esh Shaheinab), is even more different from the central Nile Valley sites and represents an inland Khartoum Neolithic manifestation in the Butana about 50 km east of the Nile River. Due to the inland proximity of Shaqadud, obvious differences in the fauna (such as no fish remains) and material culture (no shell hooks, or bone harpoons) were noted (Marks et al. 1985: 276). However, there were also no remains of any domestic animals nor were there any of the common Khartoum Neolithic bone and stone gouges and celts, amazonite beads, zeolite lip plugs found, and these particular differences suggested that the Khartoum Neolithic occupants at Shaqadud were separate from the groups living in the Nile Valley (Marks et al. 1985: 276). Nevertheless, the ceramics of the Khartoum Neolithic occupation at Shaqadud were similar to the ceramics found in the central Nile Valley (Marks et al. 1985: 271). Furthermore, within the entire midden deposit at Shaqadud there was a clear transition from an Early Khartoum occupation to a Khartoum Neolithic occupation which paralleled the same process which presumably (but until Shaqadud was never proven stratigraphically) took place in the central Nile Valley (Marks et al 1985: 276).

Based on a survey of other sites in the Khartoum area, Arkell felt that the distribution of Khartoum Neolithic sites was much more limited than the Early Khartoum and was generally restricted to the central Nile Valley between the 6th Cataract to the north and the Jebel Aulia dam on the White Nile about 30 km south of Khartoum (Arkell 1953: 107). Since Arkell’s excavations at Esh Shaheinab, some important modifications concerning the nature of the Khartoum Neolithic have taken place. First is the realization that, like the Early Khartoum, the Khartoum Neolithic should be viewed more as an archaeological development (which can be defined by a particular set of ceramics) as opposed of constituting a single culture (Krzyzaniak 1978: 163). This realization is based on the more recent work at other sites such as Kadero and Shaqadud which show that some of the Khartoum Neolithic settlements represented different adaptive strategies which, in turn, reflect striking differences seen in the non ceramic assemblages when compared to the type site of Esh Shaheinab (Krzyzaniak 1978: Marks et al. 1985).

Recently, it has been suggested that the differences noted in the material culture and subsistence strategies among the various Khartoum Neolithic sites could be due to a shifting economy based on the seasonal herding of cattle supplemented with the cultivation of wild plants (Haaland 1981, 1986, 1987). Haaland argues that the larger sites such as Kadero and Um Direiwa (which are 10,000 and 40,000 m2 respectively) were semi-sedentary base camps occupied more heavily during the rainy season when wild varieties of sorghum and millet were planted (Haaland 1986: 54). These sites were situated on naturally elevated areas at a distance from the river (6-7 km from the river’s edge) usually on the east side of the Nile (ibid). The basic characteristic of these base camps was a high concentration of lithics and ceramic debris accompanied with a high frequency of grinding stones and burials (Haaland 1987: 204). Smaller sites such as Zakiab (ca. 2000 m2) were located closer to the Nile River (3-4 km from the river’s edge) and was interpreted by Haaland as being temporary dry season camps for fishing and herding of animals with no emphasis on plant cultivation (Haaland 1987: 27-28). Sites such as Zakiab were characterized by thinner cultural deposits (lesser densities of lithics and ceramics) with fewer grinding stones and no burials (ibid.). It was also felt that inland sites in the Butana such as Shaqadud and another site called Hatab may represent other types of wet season camps used by the same Khartoum Neolithic groups who inhabited the Nile Valley during the drier parts of the year (Haaland 1987: 208-209) Haaland also believed that the Butana in general would have been especially attractive

At Kadero (situated on the east side of the Nile River about 10 km south of Esh Shaheinab) over 80 percent of the faunal remains consisted of domestic cattle (Bos primigenius), in addition to other domestic animals such as sheep and goat (Krzyzaniak 1978: 164-165). Based on these data, and in stark contrast to Esh Shaheinab, it was apparent that the economy at the Kadero settlement revolved around pastoralism with a much smaller emphasis on fishing, hunting, and gathering (Krzyzaniak 1978: 166). There were also notable differences in the material culture between the two sites. At Kadero there were very few crescents and backed blades, and no bone harpoons or shell fish hooks were found (Krzyzaniak 1978: 165). The lack of bone harpoons and shell fish hooks also corresponded to an overall paucity of fish bone at Kadero. On the other hand, Kadero contained a much larger quantity of ground stone than was found at Esh Shaheinab which indicated that the processing of plant foods (such as wild varieties of sorghum and millet) was more important at the former site (ibid.). Burials were also a common feature on the site of Kadero, of which none were found at Esh Shaheinab (Krzyzaniak 1978: 167-169). A few of these burials contained individuals (placed in a tightly flexed position) with associated grave goods such as pottery, necklaces of carnelian beads, and discoidal maceheads (made from porphyry) which demonstrated that these individuals had some status within the society (Krzyzaniak 1978: 169).

97

The Butana Group Ceramics

to the Nilotic based Khartoum Neolithic people during the extreme wet season for pasturing cattle (Haaland 1987: 208). As pointed out above, however, there were no remains of domestic animals at Shaqadud prior to 3600 BP - which is too late for the Khartoum Neolithic period (Marks et al. 1985: 274-275). Furthermore, the marked differences in tools between Shaqadud (such as the lack of gouges and celts) and the central Nile sites suggests that a different group of people occupied the former site (Marks et al. 1985: 276). On the other hand, the differences between such sites as Kadero and Esh Shaheinab within the central Nile Valley may be more compelling for Haaland’s seasonal settlement model. Haaland noted that both Zakiab and Esh Shaheinab were located in a similar riverine environment and were of similar age (Haaland 1987: 27-32, 62). Indeed, both sites are quite small in size (Zakiab=2000 m2, Esh Shaheinab=6000 m2) and lack a significant number of grinding stones, and contain no burials. Furthermore, Haaland’s excavations at Esh Shaheinab proved that domestic cattle were present on this site, suggesting that the economies between Zakiab and Esh Shaheinab were not as different as originally believed (Haaland 1987: 31-32).

least 100 km during the later part of the Khartoum Neolithic (corresponding to the Neolithic humid period which began around 4500 BC) there is little question that there would have been some trans-Sahelean contact between groups in the central Nile Valley and those farther to the west in the present day Sahara (cf. Marks and Caneva 1990a). In this light, the Karat Group of Dongola Reach may have some connection with Smith’s Tenerian complex. Indeed, Arkell had suspected that contacts between cultures in the Tibesti region had taken place between Esh Shaheinab and Fayum A (Arkell 1953: 104-105; Phillipson 1985: 116-117). This assumption was based on the presence of amazonite (which originates in northern Tibesti) at a number of Khartoum Neolithic sites in the central Nile Valley (as well as at Fayum A) and the presence of the characteristic Khartoum Neolithic gouge (i.e. Shaheinab adze; cf. Arkell 1975: 16, Fig. 3-2; Smith 1980: 457, Figure 18.5) which were found at the Fayum A sites and at several other sites in Tibesti (ibid.). Some Khartoum Neolithic ceramics were also found at the Tibesti sites (ibid.). The sites of Shaqadud and el Ghaba in the Butana and the Shendi Reach respectively, indicate that the spatial distribution of the Khartoum Neolithic may have extended farther to the north beyond the 6th Cataract (Geus 1979: 20; Marks et al. 1985; Haaland 1988: 23). The later assessment of the site of Rabak on the White Nile has also indicates that some form of the Khartoum Neolithic may have extended 230 km south of Khartoum (Haaland 1987: 45-47). Like the Khartoum Neolithic components at el Ghaba and Shaqadud, the early occupation at Rabak contained typical Khartoum Neolithic ceramics (but not as much of a variety) but lacked the characteristic gouges and adzes which were common at such sites as Esh Shaheinab and Kadero (Haaland 1987: 24).

Another survey, conducted on the east side of the Nile from Khartoum north to Geili (a distance of about 50 km) by Caneva and others, suggested that a complex settlement pattern based on the seasonal herding of animals and the cultivation of plant crops may have existed for the Khartoum Neolithic in that part of the central Nile, as well (Caneva 1988: 335). As in the Kadero vicinity, it was apparent that Khartoum Neolithic populations along the Geili-Khartoum stretch were mainly cattle pastoralists who probably subsisted on milk and blood more than the meat of their domesticated animals (Caneva 1988: 359-368). It was also clear that Khartoum Neolithic settlements in the Geili­ Khartoum area were not strictly riverine-adapted, due to the wide distribution of sites along the floodplain - which were always located at a distance from the river’s edge in order to take advantage of the inland vegetational resources within the Nile Valley (Caneva 1988: 334-335). It was also noted that there were more Khartoum Neolithic sites in contrast to the Early Khartoum sites along the Geili-Khartoum stretch (Caneva 1988: 327).

Other than the sites of Shaqadud, el Ghaba, and Rabak, it appears that practically all other confirmed Khartoum Neolithic sites were essentially restricted to the central Nile Valley between Sabaloka Gorge to the north and the Jebel Aulia Dam on the White Nile to the south (Arkell 1953; Haaland 1987: 25-47; Caneva 1988: 324, 369). All of the Khartoum Neolithic sites in this region (commonly referred to as the Khartoum province) seem to have been more or less contemporaneous with one another (at least within a span of 400 years) and share a very similar set of ceramics and other diagnostic artifacts such as gouges and adzes which would indicate that they belonged to a single cultural facies (Arkell 1953; Haaland 1981; Chlodniki 1984; Nowakowski 1984; Caneva 1988: 324). This area includes the well known sites of Esh Shaheinab, Geili, Islang, Kadero, Um Direiwa, Zakiab, Qoz, among others.

The basic tenet that the Khartoum Neolithic was restricted to the central Nile Valley is essentially correct (cf. Arkell 1953; Haaland 1987; Caneva 1988). Nevertheless, the findings at Shaqadud (and possibly an isolated find at Hatab) indicate that inland sites related to the Khartoum Neolithic may represent a separate group of people from the central Nile Valley who were orientated more towards the Butana steppe and relied more on hunting and gathering than on cattle herding (Marks et al. 1985). Arguments brought up by Smith have also indicated that the Khartoum Neolithic may be in part connected with Sahelean and Saharan groups associated with the so-called Tenerian complex (defined primarily by the presence of Shaheinab adzes) further to the east in Tenere and Tibesti (Smith 1980: 451-465). Given that the Sahel probably extended into the Sahara zone at

The Late Neolithic Development in the Central Nile Valley The late Neolithic development of the central Nile Valley represent a wide distribution of post-Khartoum Neolithic cultures (ca. 4000 BC) which extended from the Shendi

98

Comparisons of Culture Groups

Reach south of the Atbara River down to the southern Gezira Plain between the White and the Blue Nile. These cultures can be grouped into three distinct regional manifestations, the first of which existed in the Shendi Reach area and is associated with the site of Kadada (Geus 1984a & b), while the second existed in the Khartoum province and is primarily associated with post-Khartoum Neolithic burials at Esh Shaheinab, Omdurman Bridge, and el Geili, (Arkell 1949; 1953; Caneva 1988). The third regional manifestation of the late Neolithic existed in the southern Gezira Plain and is associated with ceramics from Jebel Moya and Jebel Tomat and the later occupation at Rabak (Clark 1974; Haaland 1987). The first two late Neolithic developments in Shendi and the Khartoum province appear to be very closely related to one another and, for all practical purposes, probably represent a single facies which contained two or more culture groups (Caneva 1988: 163-187). The third development, associated with the southern Gezira Plain, appears to represent another facies which does not share as many common elements as do the first two. It is very likely that this particular facies is later in time than the other developments in the Shendi Reach area and the Khartoum province. The common denominator among these late Neolithic developments within the central Nile Valley is that they were indigenous to the region and probably evolved out of an earlier Khartoum Neolithic base (Geus 1984a & b, Haaland 1987: Caneva 1988).

339, 370). These late Neolithic settlements (El Kenger Middle and East) consist of a very thin cultural deposit of ceramics (burnished with impressed decoration) and lithics (chipped and ground stone including a few polished axes) mixed in with a few bones of domesticated animals (Caneva 1988: 339). Based on the material culture, there is little doubt that the late Neolithic occupants of these sites were descendants of the Khartoum Neolithic (Caneva 1988: 334, 370). Other similar sites have also been located throughout the central Nile Valley north of the Khartoum province in the Shendi Reach and possibly as far north as Dongola Reach (Caneva 1988: 370 via Reinold personal communication). As opposed to the Khartoum Neolithic, cultural boundaries along the Nile River from Shendi to the Khartoum area were not as distinct, suggesting that these late Neolithic groups indeed were more mobile and traveled greater distances (Caneva 1988: 370). In addition to the very ephemeral camp sites, the presence of late Neolithic burials (which are much more visible) also occur at a number of earlier Khartoum Neolithic sites in the Khartoum province (Caneva 1988: 151). It appears that the nomadic late Neolithic groups in this area took advantage of the elevated Early Khartoum sites for the specific use of cemeteries (ibid.). The use of such sites for burial grounds in the central Nile was a common practice which lasted throughout the historic period (Arkell 1949: 1; 1953: 1; Caneva 1988: 151-226). At the site of Geili (a Khartoum Neolithic occupation site) it was estimated that about 10 percent of the 113 identified burials (which ranged from Khartoum Neolithic to Meroitic to Christian and Moslem in age) were late Neolithic in date (Caneva 1988: 152). The late Neolithic burials at Geili were considered to be “Protodynastic” in age and were very similar to other burials of the same age at Esh Shaheinab (Arkell 1953: 82­91) and Omdurman (Arkell 1949: 99-106). All but two of the late Neolithic burials at Geili were all tightly flexed (perhaps bound with string or placed in sacks) lying on their right sides (Caneva 1988: 163). This kind of tightly flexed “wrapped” burials were also common at Kadero (Dzierzykray-Rogalski 1984; Caneva 1988: 163). The two other late Neolithic burials at Geili were more or less flexed with straight backs and were similar to some of the burials farther north at Kadada (Reinold 1987: 26, 92-93; Caneva 1988: 163).

The Late Neolithic of the Khartoum Province. In the Khartoum province (as well as at Shaqadud), it appears that the Khartoum Neolithic was relatively short lived, lasting less than a thousand years (perhaps 600 years at most) between the 6th and 5th millennium BC (Marks et al 1985: 262; Haaland 1987: 59­62). By 4000 BC, all Khartoum Neolithic settlements within this area had been abandoned (ibid.). Like the A-Group, the Khartoum Neolithic vanished from the archaeological record, leaving almost nothing behind in terms of material culture (Haaland 1987: 224). It also appears that there was no sudden change in the climate which caused the disintegration of the Khartoum Neolithic, as a whole (ibid.). Due to the relatively high density of Khartoum Neolithic sites in the Khartoum province and an increasing reliance on animal husbandry towards the end, it is possible that the narrow fertile strip of land along the Nile in this area was overgrazed by the end of the 5th millennium BC (Caneva 1988: 370). This would have prompted the surviving Khartoum Neolithic groups to adopt to a more mobile existence with fewer domestic animals or to switch to a heavy reliance on hunting and gathering (ibid.). In either case, a more mobile existence for populations in this area would have meant a reduction in the material culture overall, and more short term occupations which would be less visible in the archaeological record. A few occupation sites (which are barely visible) dating to the post-Khartoum Neolithic period (or what is referred to in the central Nile Valley as late Neolithic) have been located within the Khartoum province which supports this kind of shift in settlement pattern (Caneva 1988: 334,

Grave goods associated with the burials included ceramic vessels, zeolite lip plugs, clay figurines, beads (made of ostrich egg shell, amazonite, and carnelian), flat disk-shaped maceheads, a polished palette and cylinder of quartz-diorite, polished celts and grinding stones (Caneva 1988: 166172). The late Neolithic ceramics found in the grave fill and associated with individuals, consisted mostly of “gray/ black burnished pottery” (Caneva 1988: 174). Most of this pottery was well fired and was composed of a compact fine grained paste (ibid.). The decorated ceramics included incised (linear, geometric, and concentric circles) simple dentate (both linear and geometric) and some ripple ware (Caneva 1988: 174-179). The late Neolithic vessel forms

99

The Butana Group Ceramics

with the A-Group) and show many affiliations with both A-Group and later C-Group pottery types (Geus 1980: 44; 1981: 37; 1984b: 368-369). Incised decorations, arranged in zoned, triangular and concentric square designs (usually executed on simple contour bowls with round bases), were most common among the ceramics at Kadada (Geus 1980: 47). There was also a considerable amount of ripple ware at this site (Caneva 1988: 174).

consisted of simple contour bowls or jars and on occasion a few bottles (ibid.). Based on the positioning of the individuals and the associated grave goods (including all of the ceramics) the late Neolithic burials at Geili were similar to the late Neolithic burials observed at Kadada to the north in the Shendi Reach area (Caneva 1988: 180). Other than the ceramics, the Khartoum Neolithic burials at Kadero also appeared to be ancestral to the late Neolithic burials at Geili (ibid.). Overall, it was suspected that the late Neolithic cemetery at Geili may represent a very small homogeneous group who buried their dead over a period of several generations (ibid.). Similar late Neolithic burials to those found at Geili were also recovered at Omdurman Bridge and at Esh Shaheinab (Arkell, 1949; 1953).

Utilitarian artifacts associated with the late Neolithic settlement at Kadada include lithics, bone, and shell. The chipped stone artifacts are basically microlithic but tend to be poorly made (Geus 1980: 44). Gouges are not present. Shell hooks and bone harpoons were also common (ibid.). Among the fauna recovered at Kadada there was a very high frequency of domesticated animals (cattle and ovicaprids) accompanied with a very small amount of wild animals (Geus 1981: 37).

Kadada of the Shendi Reach.

Jebel Moya, Jebel Tomat, and Rabak of the South Gezira Plain.

Unlike the Khartoum province, the Shendi Reach, around the site of Kadada, may have not experienced such a cultural deflation from the Khartoum Neolithic to the late Neolithic. Little is known about the extent of the Khartoum Neolithic settlements in the Shendi Reach, but there can be little question that some form of the Khartoum Neolithic did exist in this area (Geus 1982; 1984a). Based on the data recovered from El Ghaba (Khartoum Neolithic) and the earliest burials at Kadada (late Neolithic), there appears to have been a continuous development from the Khartoum Neolithic to the late Neolithic in the Shendi Reach area without any decline in population or decrease in cultural complexity (Geus 1981: 40-41; 1982: 33-34; 1984a).

In the Gezira Plain south of Jebel Aulia, are the sites of Jebel Moya, Jebel Tomat, and Rabak. Like Kadada to the north, these sites also appear to represent a vigorous late Neolithic development (Clark 1984: 122-124; Haaland 1987: 220-227). The late Neolithic component associated with these sites has been termed the Jebel Moya tradition or complex (Clark 1973: 60; 1984: 122-124; Haaland 1987: 220). The Jebel Moya complex is characterized by the ceramics which as a rule were decorated along the rim portion of the vessel (banded) within impressed or incised designs (Haaland 1987: 220). The exterior surfaces of these vessels have usually been wiped or smoothed, while burnishing is rarer. Many of the vessels within the Jebel Moya complex are necked and/or have pronounced everted rims. A characteristic vessel form associated with the Jebel Moya tradition is a long necked bottle which resembles a calabash.

The late Neolithic site of Kadada is situated midway between Atbara and Khartoum on the east side of the Nile River (near the river’s edge) in the Shendi Reach area. Kadada represents an extensive settlement with an estimated size of over 100,000 m2 (Geus 1980: 41-47; 1983: 23-25). Radiocarbon dates recovered there suggest that the site was occupied sometime after 4000 BC and, perhaps, intermittently on up to Meroitic times in the 1st century BC (Geus 1981; 37; 1982: 33-34). Much of Kadada consists of an extensive, multi-component cemetery (which covers at least several hectares) from which most of the reported information from the site is derived (Geus 1984a; Caneva 1988: 180). Graves associated with the late Neolithic occupation at Kadada include both pit and pot burials (Geus 1980: 43; Geus 1984a). The pit burials consist of flexed superimposed individuals (sometimes placed on mats) who were interred in oval graves. Associated grave goods include beads (made of bone, carnelian, and amazonite), nose and lip plugs, both decorated and undecorated pottery, ivory anklets, ostrich eggshell, mollusk shell, stone palettes, maceheads, and axes, and figurines made of sandstone and clay (Geus 1980: 43-44). The pot burials consist of small children or infants placed in a large ceramic pot. Grave goods were less, but included shell, beads, ostrich eggshell, and some ceramics. The ceramics associated with the burials overall indicate a Protodynastic date (contemporaneous

Jebel Moya. Jebel Moya is the type site of the Jebel Moya complex. Jebel Moya is about 200 km south of Khartoum and 40 km west of the Blue Nile. The site consists of an extensive village settlement (associated with thousands of graves) situated on top of a large granite massif (called Jebel Moya) which rises several hundred meters above the Gezira Plain. Jebel Moya was extensively excavated by Sir Henry Welcome between 1910 and 1914 and was finally reported by Frank Addison in 1949 (Addison 1949). The excavations at Jebel Moya concentrated mainly on the 2883 burials which were found there, and little information was gathered about the settlement associated with the cemetery (Addison 1949). An immense amount of archaeological data were recovered from the early excavations, but unfortunately, problems concerning provenience and the manner in which artifacts were recorded and labeled rendered much of this data to limited use (Addison 1949: 1-35). What could be

100

Comparisons of Culture Groups

assessed is that the majority of the burials recovered from Jebel Moya dated to Napatan and Meroitic times (ca. 600 BC to 400 AD), however, it was also realized that Jebel Moya may have been continuously occupied since the late Neolithic (Addison 1949: 249-260; Clark 1973: 59-60; 1984: 122-123). The late Neolithic component at Jebel Moya has been estimated to have begun sometime within the early part of the 4th millennium BC (Haaland 1987: 46) and, perhaps, lasted as late as 2500 or 2000 BC (Clark 1984: 122). However, it is still not clear what particular artifacts are associated with the late Neolithic occupation at this site. Nevertheless, there was the brief mention of early pottery which was found in the deepest levels at this site which might predate the Napatan-Meroitic occupation (Addison 1949: 202-203; Clark 1984: 122). Pottery of this sort consisted of open, simple contour bowls with impressed decorations, many of which had incurving rims which were thickened on the interiors (ibid.). Overall, the paste (which was pinkish) associated with the Jebel Moya ceramics had lots of mineral inclusions, consisting mainly of “crushed granite” (Addison 1949: 1949: 201). The lithic artifacts recovered from Jebel Moya were reported to also reinforce the presence of an “old Neolithic tradition” which included microliths, ground stone axes and maceheads (Clark 1984: 123).

Tomat were domestic consisting of cattle and goat/sheep (Clark 1984: 123). The remains of a few wild animals such as gazelle, reedbuck, oribi, hare, bushpig, Egyptian goose and pigeons were also found (ibid.). Overall, it appeared that the population at Jebel Tomat were semi-nomadic pastoralists who cultivated plants such as sorghum during the rainy season (ibid.). Rabak. The site of Rabak (also referred to as Kosti) is situated on the east side of the White Nile (ca. 3 km from the river’s edge) about 20 km upriver from Jebel Tomat and about 40 km to the east of Jebel Moya (Haaland 1987: 35, 45). Of the three sites, Rabak has the clearest evidence of a late Neolithic occupation in this part of the central Nile Valley. This site is about 16,000 m2 in size and has about 60 to 80 cm of midden deposits (Haaland 1987: 45-46). In the lower levels of the deposit is a Khartoum Neolithic occupation, while the upper levels contain a Late Neolithic occupation. The Khartoum Neolithic occupation dates roughly between 6000 and 5800 BP, while the Late Neolithic occupation dates to about 4500 BP (Haaland 1987: 61). Given that there is an approximate gap of 1500 years between the latest Khartoum Neolithic occupation and the Late Neolithic occupation, it is possible that there was a settlement hiatus between the Khartoum Neolithic and Late Neolithic at this site (Haaland 1987: 46). However, Haaland argues that the ceramics at Rabak represent a continuous sequence between the Khartoum Neolithic and the late Neolithic components there (Haaland 1989: 362). The lithics are also very reminiscent of the earlier Khartoum Neolithic material (however, no gouges are at Rabak) farther north (Haaland 1987: 35, 46).

Based on some later and very limited excavations, the remains of cattle and goat in addition to sorghum showed that the community of Jebel Moya was fully agricultural (Clark 1973: 60; 1984: 123). However, much of this probably dates to the Napatan­-Meroitic period. Jebel Tomat. The site of Jebel Tomat is situated about 40 km due west of Jebel Moya on the east side of the White Nile. Like Jebel Moya, Jebel Tomat consists of an extensive village occupation with graves, which probably began sometime during the late Neolithic and lasted well into Meroitic times (Clark 1973; 1984). The bulk of the occupation at Jebel Tomat consists of a midden deposit approximately 120 cm in depth. One radiocarbon date recovered at 80 cm below the surface placed the early occupation of the site at about 3000 BC (Clark 1973: 57). The ceramics from Jebel Tomat look very much like the ceramics at Jebel Moya and consist of small, thin burnished wares accompanied with some larger wares with thick rolled rims (ibid.). It was believed that the ceramics and other associated artifacts found at both sites constituted a “single culture area and cultural horizon” (Clark 1973: 59; 1984: 122-124). Lithic artifacts at Jebel Tomat include maceheads (mostly flattopped disks, but some pear-shaped forms are present, as well) made from igneous rocks, and grinding stones most of which were made from sandstone. The chipped stone tools were mainly derived from igneous rocks and gneiss and were composed mostly of nondescript modified flakes (Clark 1973: 58). A smaller portion of the chipped stone tools were made from quartz pebbles using a bipolar core method. Over 80 percent of the fauna recovered at Jebel

The ceramics of the late Neolithic occupation at Rabak consist of mostly simple contour vessels with either smoothed or wiped exteriors (burnishing is less common) with impressed decorations (dots, and simple dentations) many of which have been restricted to the rim portion of the vessel (Haaland 1989: 362-364). Overall, the late Neolithic ceramics at Rabak resemble those from Jebel Moya and Jebel Tomat and are felt be part of the Jebel Moya complex as proposed by Clark (Haaland 1989: 364, Clark 1984). Haaland also feels that some of the ceramics from the late Neolithic occupation at Rabak resemble those associated with the Butana Group (Haaland 1989: 364-365). Both Clark and Haaland believe that the Butana Group might be a part of the Jebel Moya Complex (Clark 1984; Haaland 1987; 1989). The chipped stone artifacts from Rabak were manufactured from flakes (mostly of quartz) consisting mostly of retouched flakes followed by convex and concave scrapers (Haaland 1989: 361-362). Some of the larger chipped stone artifacts (mostly convex scrapers) were made from rhyolite (which was a preferred raw material used in the Khartoum Neolithic in the Khartoum province) imported from the 6th Cataract some 300 km north of Rabak (Haaland 1989: 362).

101

The Butana Group Ceramics

The late Neolithic of Shaqadud Cave

very latest Khartoum Neolithic ceramics associated with the basin occupation (Marks et al. 1985: 272-274). Indeed, there appears to have been a hiatus of roughly a thousand years between the occupations associated with the midden and the cave, and based on the ceramics, it is suspected that the inhabitants of Shaqadud Cave may have been more closely associated with the cultural developments in the southern Atbai than they were with the central Nile Valley (Marks et al. 1985: 277). The Shaqadud Cave ceramics contain a high number of black and red-slipped burnished vessels which were left plain or further decorated with incised or impressed decorations (Marks et al. 1985: 272274; Robertson 1991). Vessels were made from temperless pastes which were derived from alluvial sediments or were further tempered with organic material (Robertson 1991).

The late Neolithic of Shaqadud Cave represents post Khartoum Neolithic occupations at the site of Shaqadud which dates roughly between 2900 and 1600 BC (Marks et al. 1985). Technically, the site of Shaqadud is located just to the east of the central Nile Valley, and as noted above, the Early Khartoum and Khartoum Neolithic aspect of the site falls within the general Nilotic developments observed farther to the west (Marks et al. 1985: 276). Both the Early Khartoum and Khartoum Neolithic occupations at Shaqadud are associated with a midden occupation situated in a small box canyon (Marks et al. 1985: 264). In the very back of the box canyon is a cave and pond which contain other cultural deposits associated with the late Neolithic occupations. The cave deposit is about 3.5 m in depth and contains a considerable amount of lithics, ceramics, bone, and macrobotanical material. From the faunal and plant remains, the late Neolithic inhabitants of Shaqadud Cave were in possession of both domestic plants and animals (consisting of donkey and cattle and millet) but were also subsisting off of wild animals such as antelopes and giraffes (Marks et al. 1985: 275).

The lithic material from the cave consists of both chipped and ground stone (Marks et al. 1985: 274). The chipped stone material is generally poorly made and consists mostly of flake tools with a low index of lunates (ibid.). It appears that at least some of the chipped stone tools (scaled pieces in particular) were manufactured using the bipolar method, which was not observed in the Khartoum Neolithic lithics associated with the midden occupation. The majority of ground stone consists of hand stones (manos) and elongated stones which may have been used as pestles (Marks 1985: 274).

As it will be discussed later in greater detail, the ceramics associated with Shaqadud Cave and the pond area do not resemble, or indicate any signs of developing from, the

102

Chapter 6 Comparisons of the Ceramics Associated with the Culture Group of Northeast Africa During the Neolithic and Post-Neolithic Era The Ceramics of the Badarian Culture. Khartoum Neolithic. Malawiya Group. and Malawiya/Butana Transition

Ceramics associated with the Badarian culture were recovered from several cemeteries between el Matmar and el Qaw and the habitation site of Hemamieh (Brunton and Caton-Thompson 1928). Overall, the Badarian ceramics were manufactured from either a chaff tempered (chopped straw) or an untempered, fine-grained (consisting of mineral inclusions) paste (Brunton and Caton-Thompson 1928: 21). Badarian vessels which were chaff tempered were almost always utilitarian (such as cooking pots and storage jars) and tended to be large and undecorated (ibid.). These pastes were loosely consolidated and quite friable. The vessels which were manufactured from the finer, untempered pastes were smaller, usually decorated, and better made (ibid.). This paste was compact, quite hard, and probably derived from clay-rich alluvial sediments (however this was never stated precisely in Brunton and Caton-Thompson 1928). It is important to note that outside the cemeteries, approximately 99 percent of all of the Badarian ceramics recovered consisted of coarse, brown-red (rough ware), undecorated chaff tempered pot sherds (Brunton and Caton-Thompson 1928: 71). As with all of the Predynastic archaeological projects conducted during the first half of the 20th century, these particular sherds (which were plentiful at all Predynastic sites) were discarded during the excavations and not analyzed further. As a rule, only decorated rim sherds were kept, and pastes associated with these kinds of sherds were usually finegrained and untempered. From the later excavations of a predominantly Amratian occupation (ca. 3650-3500 BC) at Hierakonpolis, all sherds were collected (over 350,000) and it was revealed that 96 percent were also chaff tempered, while the remainder were fine paste (Hoffman 1982: 72).

At the end of the 5th millennium BC (ca. 4000 BC), the first post-Neolithic cultures were being established in Lower and Upper Egypt, while in the central Nile Valley, the Khartoum Neolithic was disappearing as a cultural phenomenon. In the eastern Sudan, the first signs of the Butana Group was occurring through the Malawiya/Butana transition in the southern Atbai. From the perspective of the southern Atbai (some 1200 km south of the Nile Delta), the burgeoning Lower Egyptian trading societies associated with Maadi were a remote phenomenon and there are no traces of Maadian material culture (or other cultural vestiges associated with the Lower Egyptian Neolithic, post-Neolithic) in the former area to suggest that there was any contact between Lower Egypt and the eastern Sudan. Indeed, since there is virtually no evidence for any Lower Egyptian Neolithic or Chalcolithic occupation in Upper Egypt (cf. Trigger 1985), one can safely assume that the Neolithic and post-Neolithic manifestations of Lower Egypt had no direct involvement or contact with the Malawiya/ Butana transition much farther to the south in the eastern Sudan. However, the presence of ripple wares (which are most dominant in the Badarian ceramic inventory) in the Khartoum Neolithic (at the very end) and the Malawiya/ Butana transition may indicate some form of exchange and/ or contact between Upper Egypt, the central Nile Valley, and the eastern Sudan. Of particular interest is that during this time the Khartoum Neolithic declined very quickly while both the Badarian culture and the Malawiya/Butana transition began to take root and prosper as separate cultural manifestations in Upper Egypt and the eastern Sudan, respectively.

More than 80 percent of the decorated Badarian vessels (including whole vessels recovered from burials) were finished with characteristic ripple marks, usually done in a vertical or spiral-like fashion across the surface of the vessel (Brunton and Caton-Thompson 1928: 20, plate 1). On occasion, the ripple marks were done in a horizontal fashion. The rippled surface either covers the whole exterior of the vessel or the top portion near the rim. Rippling was rarely done on the interiors (Brunton and Caton-Thompson 1928: 21). Like the Sarb’ut Ripple type associated with the

The Ceramics of the Badarian Culture To date, the most comprehensive report on the Badarian ceramics is still Brunton and Caton-Thompson’s original work, the Badarian Civilization published in 1928.

103

The Butana Group Ceramics

Table 6.1. Comparative Table on Decorated Ceramics at a Midden Deposit at Hemamieh*

Gerzean 1 Amrantian-Early Gerzean 2 Transitional 3 Badarian 4

Badarian (ripple)

Total

no. (%)

WhiteCrossedLined & Redpolished no. (%)

no. (%)

no. (%)

5 (7) 51 (70) 18 (21) 27 (11)

0 (0) 4 (5) 1 (1) 0 (0)

0 (0) 6 (8) 67 (76) 188 (89)

67 (99) 73 (99) 87 (100) 215 (100)

Decorated (painted)

Wavy Handled

Blacktopped

no. (%)

no. (%)

59 (88) 7 (9) 1 (2) 0 (0)

3 (4) 5 (7) 0 (0) 0 (0)

Summarized occupation levels at Hemamieh, based on Caton-Thompson’s original stratigraphic assessment of “finer wares” within a five to six foot midden deposit (Bruton and Caton-Thompson 1928:79). All undecorated “rough village wares” were not considered (ibid.). The raw counts were condensed form the original excavated levels within each of the occupation zones. Relative frequencies were computed from Caton-Thompson’s original counts. 1 “Middle Predynastic Occupation” (0’ to 2’: 3 levels; 1st = 0-1’, 2nd = 1’-1’6”, 3rd = 1’6” –2’0”). 2 “Early Predynastic Occupation, passing up into Lower Middle Predynastic” (3’6’’-4’6’: 3 levels; 1st = 3’6”-4’0, 2nd = 2’6”-3’, 3rd = 3’- 3’6”). 3 “Transition or Disturbance” (3’6”-4’6”: 2 levels; 1st = 3’6”-4’0, 2nd = 4’0-4’6”). 4 “Badarian Occupation” (4’6”-6’: 4 levels; 1st = 4’6”-5’, 2nd = 5’-5’6”, 3rd = 5’6”-6’0 “or breccia level”, 4th = “beneath breccia” (Brunton and CatonThompson 1928:79). *

Malawiya/Butana transition, the rippling of the Badarian vessels was accomplished first by combing and then by burnishing which partially obliterated the combed striations, which in turn created the ripple marks (ibid.). The quality of the rippling among the Badarian ceramics varies from fine to coarse, based on the degree of obliteration of the combed striations. The very finely rippled pottery, which is most typical, is highly burnished and tends to have an even surface, while the more coarse rippled pottery has more of a combed, unburnished surface (cf. Brunton and CatonThompson 1928: plate 1). In fact, some of the more coarse rippled ceramics are referred to as “combed” (Brunton and Caton-Thompson 1928: 1, 70, 76). Unfortunately there are no descriptions, photographs, or illustrations of the “combed” Badarian ceramics and it cannot be determined how they would have compared with the Butana Combed type. It was felt, however, that the initial process of combing was also used by Badarian potters as an overall technique to thin vessel walls (Brunton and Caton-Thompson 1928: 20). The method of thinning vessel walls by combing is also a key element in the Atbai Ceramic Tradition (Fattovich, Marks, and Mohammed-Ali 1984: 178).

by Brunton, derived from Petrie’s original classification of the Egyptian Predynastic ceramics recovered at Naqada and Diospolis Parva (Petrie 1901; 1921; Brunton and Caton-Thompson 1928: 21). Petrie’s Egyptian Predynastic classification consisted of nine ceramic classes, of which six were based on exterior decorations or surface finish, while the remaining three were based essentially on vessel shape (Petrie 1901). Of course, Petrie’s ceramic classification was based on whole vessels recovered from burials and pot sherds were not classified. Nevertheless, Petrie’s general classification (cf. Petrie 1901, Hoffman 1982: 68) is still viable (albeit with some modification, cf. Kaiser 1956) and his basic Predynastic ceramic classes are still used today (Hoffman 1982: 66-85; Hassan and Matson 1989: 303-317). Petrie’s nine original ceramic types are as follows: 1. Black-topped (class b; associated with a fine-grained untempered paste ). 2. Polished Red (class p; associated with a fine-grained untempered paste ). 3. Fancy forms (class f; associated with a fine-grained untempered paste ). 4. White Cross-lined (class c; associated with a fine-grained untempered paste). 5. Incised Black (class n; ?). 6. Wavy-handled (class w; associated with a crushed stone tempered, hard orange paste). 7. Decorated, i.e. painted (class d; associated with a crushed stone tempered, hard orange paste). 8. Rough-faced (class r; associated with a chaff tempered paste). 9. Late Ware (class 1; ?).

Other than rippling, Badarian vessels are either burnished or smoothed, the latter of which has an even or rough finish (Brunton and Caton-Thompson 1928). Many of the better quality vessels are coated with either a red wash (it was never determined whether it was a true wash or a slip) or a brown slip (ibid.). On occasion, a few vessels have polished black (unslipped) exteriors. The black topping of vessels is also a very common trait and is present on vessels with either a red-finish or brown slip. The interiors of these black-topped vessels are always black. Of coarse, black-topping was done by turning the vessels upside down in a bed of ashes which carbonized and blackened both the rim portion and interior of the vessel.

Among the Badarian ceramics, seven Predynastic ceramic classes were defined by Brunton (Brunton and CatonThompson 1928: 21) and are as follows:

A preliminary classification of the Badarian pottery (consisting of ceramics found in the cemeteries) was created

1. Black-topped Polished Brown class (BB; probably associated with a fine-grained untempered paste).

104

Comparisons of the Ceramics

2. Black-topped Polished Red class (BR; probably associated with a fine-grained untempered paste). 3. Plain Polished Red class (PR; probably associated with a finegrained untempered paste). 4. Smoothed Brown class (SB; ?). 5. Rough Brown class (RB; mostly chaff tempered). 6. All Black class (AB; probably associated with a finegrained untempered paste). 7. Miscellaneous class (MS; ?).

Brown and Polished Red classes, there is a high incidence of shallow-walled carinated or keel-based bowls (Brunton and Caton-Thompson 1928: plates XII and XV). Vessels in the Black-topped Brown and Smoothed Brown classes consist mostly of simple contour (often referred to as “bagshaped”) bowls and jars with round bottoms (Brunton and Caton-Thompson 1928: plate XIII). Vessels in the Blacktopped Red class also contain a high number of simple contour bowls and jars and there is also a higher occurrence of keel-based vessels in this class than in either the Blacktopped Brown or Smoothed Brown classes (Brunton and Caton-Thompson 1928: plate XV). Almost all of the vessels in the Rough Brown class consist of larger, flaring or vertical-walled bowls with either round or flat bases (Brunton and Caton-Thompson 1928: XVII). Based on grave lot sequence dating, it appears that the Black-topped Polished Brown class was the earliest, while the Polished Red class was the latest (Brunton and Caton-Thompson 1928: 26).

Even though Brunton used Petrie’s basic classificatory scheme, he created several new ceramic classes (such as the Smoothed Brown, and All Black classes) which had not been previously defined by Petrie. Furthermore, rippling was not used as a distinguishing ceramic trait in Brunton’s typology and cross-cuts the ceramic classes above. This was probably because Brunton tried to stick to Petrie’s original Predynastic ceramic classification, which did not contain rippled wares, and by convention (even though he did create the Smoothed Brown and All Black classes) did not opt to make a new ceramic type based on the newly found ripple wares. Counts of vessels within each of the seven class were not compiled; however, it was noted that there were only a small number of vessels in the Polished Red (usually not rippled) and All Black classes (Brunton and Caton-Thompson 1928: 22).

Based on the stratified excavations at the multi-component site of Hemamieh, Caton-Thompson was able to establish another sequence using a different ceramic classification centered primarily on decorated pot sherds (Brunton and Caton-Thompson 1928: 79). She also separated sherds with rippled exterior surfaces and put them into a separate class. In using Caton-Thompson’s classification, a clear frequency change in the ceramics can be detected from the earlier Badarian levels through the latter Predynastic levels (Amratian and Gerzean) at Hemamieh. Table 6.1 gives a detailed look at the changes in frequency, as well as the proportional differences between the ceramic classes through the levels.

Both the Smoothed Brown and Rough Brown classes represent unslipped vessels (ibid.). Overall, vessels in the Smooth Brown class tend to be thicker and clumsier looking, while vessels in the Rough Brown class are mostly chaff tempered. These latter kinds of vessels were probably used for cooking (Brunton and Caton-Thompson 1928: 23). Large pithoi for the storage of grain are also lumped in this latter class (Brunton and Caton-Thompson 1928: 24).

One should note that Caton-Thompson adhered more precisely to Petrie’s taxonomic scheme for defining the ceramics at Hemamieh (probably due to the fact that there were also Amratian and Gerzean ceramics in the deposit, as well), but did not attempt to explain or integrate hers with Brunton’s ceramic classification, even though both were used in the same publication. Nevertheless, CatonThompson’s stratigraphic use of pot sherd frequencies to demonstrate temporal variability was quite innovative for the time and was the first of its kind used in the archaeology of northeast Africa. Indeed, the decorated sherds frequencies at the Hemamieh midden were the stratigraphic evidence on which Brunton and Caton-Thompson concluded that the Badarian culture predated Petrie’s “early” (Amratian) and “middle” (Gerzean) Predynastic cultures. As one can observe in Table 6.1, the changing relative frequencies through the various Predynastic occupation levels display perfect monotonic curves and show very good patterns of how the major ceramic classes changed from Badarian to Gerzean times. The two basic decorative types within the Badarian level are ripple and black-topped red-polished sherds. One should note that rippled sherds drop off rather quickly after the Badarian occupation, while black-topped sherds continue to increase on up to late Amratian/early Gerzean times. The proportion of rippled sherds relative

The thinnest and finest examples of rippled pottery are associated with the Black-topped Polished Brown class (Brunton and Caton-Thompson 1928: 26). The exterior surfaces of these vessels are all highly burnished which result in a very fine, but subtle rippled finish. In the Plain Polished Red and Black-topped classes, many of the vessels are not rippled, however, there are a number of open shallow bowls in these classes which contain pattern burnished decorations ranging from crossed lines, fern branch-lined, six-rayed stars, and parallel lines (Brunton and Caton-Thompson 1928: 22, plates XIV and XXV no. 5). It should be noted that pattern burnishing is a characteristic decorative trait in the Butana Group ceramics, but is not present in the Malawiya/Butana transition. Overall, the shapes of Badarian vessels consist mostly of medium sized, simple contour bowls and jars, with rounded or flat bases, which almost always have straight-sided walls with direct rims (Brunton and Caton-Thompson 1928). Carinated bowls are also common and a few bottle forms and some footed vessels are present (Brunton and CatonThompson 1928: 21-26). In the Black-topped Polished

105

The Butana Group Ceramics

to the other decorated sherds is also quite high during the Badarian and transitional occupation.

V. Various Atypical Sherds of Fine Ware. 1. Alternate hatched and plain chevrons. 2. Horizontal dotted lines. 3. Semicircular panels made with various forms of dotted lines.

The Ceramics of the Khartoum Neolithic Since the first excavations at Esh Shaheinab, the classification of Khartoum Neolithic ceramics has been based almost exclusively on pot sherds from habitation sites as opposed to burials (Arkell 1953: 68-78; Caneva 1988: 80). As a result, pot sherds, not whole vessels, are the basic analytical units defining the ceramics of the Khartoum Neolithic. The use of pot sherds in general does not pose too great a problem in classifying the Khartoum Neolithic ceramics since most vessels were finished with a single decoration which usually covered the entire surface (cf. Arkell 1953; Caneva 1988). Thus, if one can recognize a particular decoration on a Khartoum Neolithic pot sherd, one can usually associate the sherd with a vessel of the same decoration.

VI. Plain Ware. 1. Brown. 2. Blackish to black. VII. Black-Top Red Ware. Through later excavations on Khartoum Neolithic sites, the various sherd divisions and subdivisions of Arkell’s original ceramic classification were kept (cf. Chlodnicki 1984), but modified with a greater emphasis on the decorative motifs for comparative purposes (Haaland 1981; 1987; Mohammed-Ali 1982; Chlodnicki 1984; Caneva 1988). For example, Haaland (1987: 147) used 14 ceramic categories based on different decorative motifs to make general comparisons between a number of Khartoum Neolithic sites in the central Nile Valley. The 14 ceramic categories (with the corresponding wares in parentheses) are as follows:

In Arkell’s original typology, Khartoum Neolithic sherds were first divided into wares primarily defined by a particular decorative technique (such as impressed, combed, and incised) and then further subdivided by decorative motif (such as two lines of triangles, dots, and/or vees, multiple lines of closely spaced dots, zigzags, etc.). Arkell’s original ceramic typology of the Khartoum Neolithic ceramics is as follows:

1. Vees and dots (impressed ware). 2. Vees (impressed ware). 3. Dots (impressed ware). 4. Catfish-spines (impressed ware). 5. Dotted impressed lines (impressed ware). 6. Incised lines (incised ware). 7. Dotted sawtooth (impressed ware). 8. Sawtooth (impressed ware). 9. Combed (combed ware). 10. Nail impressed lines (impressed ware). 11. Undecorated (plain ware). 12. Red burnished (plain ware). 13. Black burnished (plain ware). 14. Curvilinear (incised ware).

I. Dotted Wavy Line Ware. II. Impressed Ware. 1. Burnished sherds with Dotted Wavy Line pattern. 2. Two lines of triangles alternating with two lines of dots. 3. Varieties of the characteristic pattern (2.) with additional lines of dots. 4. Varieties with only a single line of triangles and from one to seven lines of dots. 5. Multiple lines of closely spaced dots. 6. Two lines of triangles or vees with only a single line of dots between them. 7. Vees only. 8. Close zigzag of dotted lines. 9. Zigzag of well-spaced-out straight dotted lines. 10. Zigzag of curved dotted lines. 11. Zigzag of curved continuous lines. 12. Fish-scale.

In more recent years, Caneva has modified the general structure of Khartoum Neolithic ceramic typology into a hierarchical system (based on Nordstrom 1972) which was structured on how the different decorations were made on the ceramics. Caneva’s typological system is quite eloquent and self explanatory and can be reviewed in Table 6.2 concerning ceramics which have impressed decorations. Caneva’s hierarchical approach in classifying the Khartoum Neolithic ceramics integrates the conventionally used decorative motifs into a structure which explains how they were produced. Overall, this kind of analytical approach gives a better understanding about a particular decorative motif in a number of ways, such as the technique and implement used, the basic elements produced from an implement, the overall motif of the decoration and how it was applied on the surface of the vessel, all of which can be used for more accurate comparisons. Indeed, through

III. Combed Ware. IV. Incised Ware. 1. Black bowls decorated with closely spaced incised horizontal lines. 2. Fine red bowls decorated with semicircular panels of incised lines.

106

Comparisons of the Ceramics

Table 6.2. Hierarchical Structure of Prehistoric Impressed Ceramics of the Central Nile Valley (after Caneva 1988) Technique

rocker stamped---------

Alternately pivoting stamp----------

simple impression------

1

Implement

[plain edge-------------[ [ [ [ [ [ [ [ [ [ [ [evenly serrated edges[ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [unevenly serrated edges------------

[double pronged [ [[ [ [serrated edge-----------

[stylus-------------------[ [double pronged

Elements

Motifs

lines----------------------

[dots [ [ [ [ [ [ [dashes

[triangles + dots [ [ [ [ [ [ [ [ [ [complex pairs of dots [pairs of triangles [ [ [ [pairs of dashes [complete pairs of dashes----------[dotted lines [ [stippled lines

[single dots [single dashes [single curved lines---[pairs of dots [pairs of dashes

Lumped under Dotted Wavy Line in Arkell’s classification.

107

[straight zigzag--------[ [ [ [ [curved zigzag---------[ [ [ [packed zigzag----------

Structure [barbed [fish net [continuous [paneled [banded [continuous [straight banded [wavy banded1 [combined1 [continuous [parallel

[spaced zigzag---------[ [zigzag with spaced dots-[ [packed zigzag----------

continuous

[spaced zigzag---------[ [zigzag with spaced dashes---------

continuous

[packed zigzag---------[(vees and dots) [ [spaced zigzags--------[(vees and dots)

[continuous [banded

[paired lines -----------[ [pairs of single dots [ [ [smocking pattern-----[ [paired lines-------------

[straight banded [continuous straight [continuous wavy1 [paneled

[paired lines------------[ [smocking pattern-----[ [paired lines [paired lines -----------[individual lines [ [ [ [individual lines-------[punctate [punctate [punctate----------------[paired lines [paired lines

continuous [continuous [banded [wavy combed1

continuous

[continuous [banded

continuous straight banded [straight banded [continuous [paneled continuous [straight banded [straight banded [continuous [parallel lines [continuous [continuous [continuous [continuous [continuous [continuous

The Butana Group Ceramics

Table 6.3.Frequencies of Khartoum Neolithic Sherds Based on Exterior Surface Treatment decorative techniques impressed undecorated red-slipped combed black-polished incised ripple total 1 2

Zabiab1

Um Direiwa I 1

Kadero I 1

El Geili 2

weight (%) 5507 (80) 417 (6) 491 (7) 434 (6) 17 (>1) 5 (>1) nd

weight (%) 1635 (67) 420 (16) 150 (6) 60 (2) 140 (5) 90 (4) nd

weight (%) 1075 (57) 700 (37) 0 49 (3) 0 50 (3) nd

number (%) 3731 (67) 805 (14) 193 (3) nd 285 (5) 338 (6) 230 (4)

6871 (99)

2495 (99)

1874 (100)

5582 (100)

Based on total weight of sherds (in kg). Based on total count of sherds.

nd = no data.

this kind of hierarchical organization, Caneva has been able to elucidate many insights on the general development of certain styles through time from the Early Khartoum through the Khartoum Neolithic and Late Neolithic in the central Nile Valley (Caneva 1988: 80-114). Nevertheless, this approach has yet to be applied as a standard method of analysis among the Khartoum Neolithic ceramics, however, it promises to have great potential and more universal applicability in the future. In either case, given the clarity of such an approach, Caneva’s standard terminology will be used throughout the ceramic descriptions below.

polish (Haaland 1987: 147). In most cases, over 80 percent of the Khartoum Neolithic vessel exteriors are decorated, while the remaining vessels are either plain burnished or wiped. Decorations usually cover the entire surface of the pot and are most often applied in a series of horizontal rows across the face of the vessel (Haaland 1987; Caneva 1988). The decorations tend to be more carefully executed along the upper portion of the vessel and are done more carelessly as they get nearer to the base. Impressed designs account for over 70 percent of all decorations among the Khartoum Neolithic ceramics. The most popular kind of impressed decorations are those created by rocker stamping (Caneva 1988: 112). Rocker stamped decorations were done either by a serrated (dentated) or plain (non-serrated) edge implement and were rocked back and forth in a straight or curved fashion in a tight (packed) or loose (spaced) formation (Caneva 1988: 83). The very characteristic “dots and vees” decorative motif (the most diagnostic ceramic decoration associated with the Khartoum Neolithic) was produced by using an unevenly serrated edge implement which was rocked back and forth vertically in very tight, straight formations (Caneva 1988: 91-94). Another form of rocker stamping within the Khartoum Neolithic was produced by using a double pronged implement (however, multiple serrated edge implements were also used from time to time) which was “walked across the edge of the exterior surface” (Caneva 1988: 94). This particular technique has been referred to as an alternately pivoting stamping (Caneva 1988: 112). Simple impressed patterns (where the implement was impressed into the surface but not rocked back and forth) were also produced. These simple impressed patterns usually consist of zoned geometric decorations which are situated along the top of the vessel near the rim (Caneva 1988: 103). Some of the Khartoum Neolithic sherds which have simple impressed decorations (where an evenly serrated edge stylus was used) are very similar looking to the Sirjino Simple Dentated type associated with the Butana Group.

For the sake of expediency in the descriptions below, the various decorative mode sherd categories of the Khartoum Neolithic ceramics are condensed back into larger units which express the basic decorative technique of the sherds. These larger units are roughly equivalent to Arkell’s original ware groupings (cf. Arkell 1953: 68-76). It is important to note that since Arkell’s original ceramic descriptions at Esh Shaheinab (which were only meant to be preliminarily), none of the Khartoum Neolithic sherds have been formally placed into ceramic types and they continue to be described primarily by the particular decorative mode. The data provided in Table 6.3 is based on the ceramic data from the excavated sites of Zakiab, Urn Direiwa, Kadero I (Haaland 1987: 153, 155-156), and el Geili (Caneva 1988: 75). In general, the ceramics associated with the Khartoum Neolithic are composed of pastes with a heavy content of mineral inclusions (Haaland 1987: 241; Caneva 1988: 70, 72). The clays used in the manufacture of Khartoum Neolithic ceramics were derived from clay-rich, alluvial sediments which had a high content of fine grained sands or silts composed of quartz or feldspars (Francaviglia and Palmieri in Caneva 1988: 345-358). Overall, these pastes tend to be quite uniform in texture and are less coarse than the preceding Early Khartoum ceramic pastes. As a rule, fiber tempered pastes are rare in the Khartoum Neolithic. Other than sherds with wiped exteriors, the exterior surfaces of all other sherds associated with the Khartoum Neolithic (with the exception of Shaqadud) are always burnished to some degree, ranging from a dull mat finish to a very glossy

Many of the rocker stamped designs used in the Khartoum Neolithic can be arranged in a rough temporal sequence

108

Comparisons of the Ceramics

(Caneva 1988: 83, 111-114, also see Table 6.2). For example, one of the earliest rocker stamped designs consists of evenly serrated edge decorations (such as the characteristic “woven mat” design) which produced such decorative motifs as packed or spaced zigzags consisting of either dots (produced by a round toothed comb) or dashes (produced by a rectangular toothed comb). These particular decorative motifs were the most common designs used during the very end of the Early Khartoum in all parts of northern Africa (cf. the Tenerian complex in the Sahara and the Sudan, Smith 1980) and the central Nile Valley and were carried over into the Khartoum Neolithic. Latter on, the evenly serrated edge designs were replaced by the more complex, unevenly serrated edge designs which produced varying renditions (packed or spaced) of the veer and dots decorative motif. This particular decorative motif was unique to the Khartoum Neolithic in the central Nile Valley and did not carry over into the Sahara or the Sahel (Caneva 1988: 112). The plain edge stamped design, commonly referred to as wolf tooth (curved) and saw tooth (straight), also appeared during the Khartoum Neolithic and continued to increase in popularity into the succeeding late Neolithic cultures of the central Nile Valley. The wolf tooth and saw tooth decorations were not restricted to the central Nile Valley and can be seen in all parts of the Sudan as well being present in the Sahara. Towards the end of the Khartoum Neolithic, the alternately pivoting stamp technique (using a double pronged implement) came into vogue and was also carried over into the late Neolithic as well. The alternately pivoting stamp technique appears to be restricted more or less to Lower and Upper Nubia and the central Nile Valley.

combing as a decorative technique appears to have come late in the Khartoum Neolithic (Caneva 1987 personal communication). Overall, incised decorated ceramics tend to be less common among most of the Khartoum Neolithic sites but they appear to increase in popularity towards the end, and continue to become more popular during the late Neolithic in the same area (Caneva 1988: 114). Incised decorations associated with the Khartoum Neolithic were usually done with a two or three pronged instrument (probably the same implements used to make alternately pivoting stamped decorations). The incised lines consist of repeated horizontal linear or simple geometric patterns (Caneva 1988: 106). A very characteristic incised decoration of the Khartoum Neolithic is composed of concentric semicircular lines arranged in horizontal panels across the top part of the vessel. Black polished ceramics in the Khartoum Neolithic represent a wide array of burnished ceramics (of varying luster) which range in color from black to brown (Caneva 1988). Many of the black polished vessels were left undecorated, however, some were decorated with pseudo rippled marks which were produced solely by burnishing and not by a combination of combing and burnishing (Caneva 1988: 110). Vessels combed and burnished to produce a true rippled exterior surface are also present in the Khartoum Neolithic and are typologically similar to the Badarian rippled wares. The rippled striations were arranged on the vessel in a vertical fashion and burnished to a high luster (Arkell 1953: 73, 76, plates 33&34; Caneva 1988: 106-107). It is believed that the decorative technique of rippling appears at the tail end of the Khartoum Neolithic occupation in the central Nile Valley (Caneva 1988: 106).

Red-slipped vessels with a well polished surface are also common within the Khartoum Neolithic (Caneva 1988: 110). The slip is bright red and is occasionally present on the interior. Some of the red-slipped polished vessels are black-topped or decorated along the very top of the rim with small inverted, blackened triangles. At Esh Shaheinab, Arkell reported several hundred black-topped or blackened triangled (referred to as “black dog-tooth” decorations) “red ware” sherds of this sort (Arkell 1953: 75). He believed that the blackened triangle decorated “red ware” came first and then was followed by the black-topped “red ware” once the technique of black topping was perfected (ibid.). A few of the “red ware” and some similar brown polished sherds at Esh Shaheinab are interior blackened like the Badarian black-topped vessels.

Vessels associated with the Khartoum Neolithic are almost always simple contour bowls or jars with round bottoms. Basal sherds are rare due to the uniform curvature of the vessels overall (Caneva 1988: 76). Vessel walls are generally thin, ranging between 3 to 6 mm thick (Caneva 1988: 72). There also appears to be some correlation between vessel form and the manner of decoration (ibid.). For example, red-slipping is usually associated with thin walled bowls with flaring, thinned rims. Vessels which have been finished with continuous rocker stamped designs tend to be vertical-walled bowls with direct rims and flat lips. The unevenly serrated rocker stamped designs tend to be associated with vessels with expanding rims, while simple impressed designs and black burnishing are more associated with jars. Lip decorations on the Khartoum Neolithic vessels are common and consist of dentated and incised decorations usually arranged in cross-line or diagonal-parallel patterns (Caneva 1988: 85). Both of these kinds of lip decorations are also common in the Butana Group ceramics, but not in the Malawiya/Butana transition.

Exterior combed ceramics are also associated with the Khartoum Neolithic and are typologically similar to the Butana Combed type identified in the southern Atbai. The Khartoum Neolithic combed sherds have the same wide, shallow combed striations which are usually horizontal to the rim (Arkell 1953: 73, plate 33; Haaland 1987: 167; Caneva 1988: 104-105). Nevertheless, the proportion of combed sherds usually is around two percent of the total ceramic assemblage and does not exceed six percent at any of the Khartoum Neolithic sites. The presence of

109

The Butana Group Ceramics

Table 6.4. Breakdown of the Malawiya/Butana Transition Ceramics by Types, Exterior Surface Treatment, and Decoration ceramic type

exterior surface treatment

decoration

count

%

Butana Combed Budahah Smoothed Nulmswi Burnished Sarb’ut Ripple Marked Saroba Combed Impressed Dratmuira Chatter Marked1 Socodad Wiped1 Sirjino Dentated1

combed smoothed burnished burnished smoothed burnished wiped burnished

combed undecorated undecorated rippled impressed incised undecorated impressed

1980 1684 525 344 160 56 6 1

42 35 11 7 3 1 2

762 (50) 169 (11)

Ware Group M 2 Burnished Brown Abkan ware Ripple, brown, Abkan ware

0 0

403 (26) 5 (