Dosariyah: an Arabian Neolithic Coastal Community in the Central Gulf [1 ed.] 9781784919634

Citation preview

Dosariyah Reinvestigating a Neolithic coastal community in eastern Arabia edited by

Philipp Drechsler

British Foundation for the Study of Arabia Monographs No. 19 Series editors: D. Kennet & St J. Simpson

Dosariyah Reinvestigating a Neolithic coastal community in eastern Arabia edited by

Philipp Drechsler

British Foundation for the Study of Arabia Monographs No. 19 Series editors: D. Kennet & St J. Simpson

Archaeopress Publishing Ltd

Summertown Pavilion 18-24 Middle Way Summertown Oxford OX2 7LG

www.archaeopress.com

ISBN 978 1 78491 962 7 ISBN 978 1 78491 963 4 (e-Pdf)

© Archaeopress and the authors 2018 Cover illustration: Trench E1 during excavation. Spring 2012.

All rights reserved. No part of this book may be reproduced, or transmitted, in any form or by any means, electronic, mechanical, photocopying or otherwise, without the prior written permission of the copyright owners. Printed in England by The Holywell Press, Oxford This book is available direct from Archaeopress or from our website www.archaeopress.com

Contents List of Contributors................................................................................................................................................................... iii Abstract..........................................................................................................................................................................................v Zusammenfassung..................................................................................................................................................................... vi Acknowledgements................................................................................................................................................................. viii Preface............................................................................................................................................................................................x Chapter 1 The Site and its Context.........................................................................................................................................1 Philipp Drechsler Chapter 2 Geomorphology, Geoarcheology and Paleoenvironments..........................................................................21 Adrian G. Parker, Simon J. Armitage, Max Engel, Mike W. Morley, Ash Parton, Gareth W. Preston and Hannah Russ Chapter 3 Defining the Archeological Setting: the Dosariyah Survey................................................................................................................................................................56 Shumon T. Hussain and Felix Levenson Chapter 4 Geophysical Survey..............................................................................................................................................134 Martin Posselt Chapter 5 Archeological Surface Collections and Excavations...................................................................................144 Philipp Drechsler Chapter 6 Dating Dosariyah..................................................................................................................................................172 Philipp Drechsler Chapter 7 The Pottery ...........................................................................................................................................................183 Christine Kainert Chapter 8 Geochemical Analysis of Putative Local and Ubaid Ceramics from Dosariyah...................................199 Peter Magee and Steven Karacic Chapter 9 Reworked Pottery................................................................................................................................................209 Christine Kainert and Philipp Drechsler Chapter 10 Fired Clay Objects..............................................................................................................................................219 Christine Kainert Chapter 11 Lithic Industry....................................................................................................................................................227 Philipp Drechsler Chapter 12 Variability of Arrowhead Shapes..................................................................................................................291 Philipp Drechsler Chapter 13 Hematite Objects and the Use of Red Pigments........................................................................................302 Philipp Drechsler, Christoph Berthold and Christine Kainert Chapter 14 Bone Industry......................................................................................................................................................310 Philipp Drechsler Chapter 15 Personal Adornment.........................................................................................................................................319 Philipp Drechsler Chapter 16 Bitumen Objects.................................................................................................................................................332 Philipp Drechsler Chapter 17 Chemical Composition of Bitumen................................................................................................................342 Thomas Van de Velde Chapter 18 Plaster Morphology...........................................................................................................................................350 Philipp Drechsler

i

Chapter 19 Morphological and Geochemical Analysis of Plaster Samples..............................................................358 Susan M. Mentzer, Markus Seil, Hilmar Adler, Thomas Chassé, Bertrand Ligouis, Christoph Berthold and Christopher E. Miller Chapter 20 Faunal Remains and Subsistence Strategies..............................................................................................384 Margarethe Uerpmann and Hans-Peter Uerpmann Chapter 21 Isotopic Analyses of Cattle Teeth..................................................................................................................423 Corina Knipper and Michael Maus Chapter 22 Archeomalacology of Dosariyah: Diversity, Taphonomy and Distribution of Gastropods and Bivalves........................................................................................................................................................436 James H. Nebelsick, Philipp Drechsler and Paolo G. Albano Chapter 23 Exploitation of the Marine Snail Hexaplex kuesterianus...........................................................................455 Georg Häussler, James H. Nebelsick and Philipp Drechsler Conclusions................................................................................................................................................................................462 Philipp Drechsler Epilogue......................................................................................................................................................................................471 References..................................................................................................................................................................................473

ii

List of Contributors Hilmar Adler Institute of Physical and Theoretical Chemistry, Division of Condensed Matter, University of Tübingen, Tübingen, Germany Paolo G. Albano Department of Paleontology, University of Vienna, Vienna, Austria Simon Armitage Department of Geography, Royal Holloway, University of London, Egham, UK; SSF Centre for Early Sapiens Behavior (SapienCe), University of Bergen, Norway Christoph Berthold Applied Mineralogy, Department of Geosciences, University of Tübingen, Tübingen, Germany; Competence Centre Archeometry, Baden-Württemberg (CCA-BW), Germany Thomas Chassé Institute of Physical and Theoretical Chemistry, Division of Condensed Matter, University of Tübingen, Tübingen, Germany Philipp Drechsler Institute of Pre- & Protohistory and Medieval Archeology, Early Prehistory and Quaternary Ecology, Department of Geosciences, University of Tübingen, Tübingen, Germany Max Engel Institute of Geography, University of Cologne, Cologne, Germany Georg Häussler Institute of Pre- & Protohistory and Medieval Archeology, Early Prehistory and Quaternary Ecology, Department of Geosciences, University of Tübingen, Tübingen, Germany; ArchaeoTask GdR, Engen-Welschingen, Germany Shumon Hussein Department of World Archeology, Leiden University, Leiden, Netherlands Christine Kainert Institute of Ancient Near Eastern Archeology, Free University of Berlin, Berlin, Germany Steven Karacic Department of Classics, Florida State University, Tallahassee, FL, USA Corina Knipper Curt Engelhorn Centre Archeometry gGmbH, Mannheim, Germany Felix Levenson Institute of Ancient Near Eastern Archeology, Free University of Berlin, Berlin, Germany Bertrand Ligouis Institute for Archeological Sciences, Department of Geosciences, University of Tübingen, Tübingen, Germany Peter Magee Department of Classical and Near Eastern Archeology, Bryn Mawr College, Bryn Mawr, PA, USA Michael Maus Department of Applied and Analytical Paleontology, Institute of Geosciences, University of Mainz, Mainz, Germany

iii

Susan M. Mentzer Institute for Archeological Sciences, Department of Geosciences, University of Tübingen, Tübingen, Germany; Competence Centre Archeometry, Baden-Württemberg (CCA-BW), Germany; The School of Anthropology, University of Arizona, Tucson, AZ, USA Christopher E. Miller Institute for Archeological Sciences, Department of Geosciences, University of Tübingen, Tübingen, Germany; Senckenberg Centre for Human Evolution and Paleoenvironment, Frankfurt a. M. and Tübingen, Germany Mike W. Morley Centre for Archaeological Science, University of Wollongong, Wollongong, Australia James Nebelsick Department of Geosciences, University of Tübingen, Tübingen, Germany Adrian G. Parker Human Origins and Palaeoenvironments Research Group, Department of Social Sciences, Oxford Brookes University, Oxford, UK Ash Parton Human Origins and Palaeoenvironments Research Group, Department of Social Sciences, Oxford Brookes University, Oxford, UK; Mansfield College, University of Oxford, Oxford, UK Martin Posselt Posselt & Zickgraf GdR, Mühltal/Traisa, Germany Gareth W. Preston Human Origins and Palaeoenvironments Research Group, Department of Social Sciences, Oxford Brookes University, UK Hannah Russ Northern Archaeological Associates, Barnard Castle, County Durham, UK; Faculty of Humanities and Performing Arts, University of Wales Trinity Saint David, Ceredigion, UK Markus Seil Institute for Archeological Sciences, Department of Geosciences, University of Tübingen, Tübingen, Germany Hans-Peter Uerpmann Institute of Pre- & Protohistory and Medieval Archeology, Early Prehistory and Quaternary Ecology, Department of Geosciences, University of Tübingen, Tübingen, Germany Margarethe Uerpmann Institute of Pre- & Protohistory and Medieval Archeology, Early Prehistory and Quaternary Ecology, Department of Geosciences, University of Tübingen, Tübingen, Germany Thomas Van de Velde Department of Archeology, Ghent University, Ghent, Belgium

iv

Abstract Re-investigations at Dosariyah, located in the Eastern Province of Saudi Arabia, took place between 2010 and 2014. They were carried out by the joint German-Saudi Dosariyah Archeological Research Project (DARP). During excavations at different parts of the shallow hill that forms the site, a wealth of material remains was found within almost three metres of anthropogenic deposits. Both radiocarbon dates and comparative studies of artefacts securely date the occupation to the first centuries of the fifth millennium BC. Investigations at the site were supplemented by geomorphological studies and an archeological survey. The spectrum of material remains excavated at Dosariyah shows unique characteristics. The lithic industry is technologically- and typologically-rooted in the local Arabian Middle Neolithic. Reminiscent of the Arabian Bifacial Tradition (ABT) are bifacially-shaped winged and stemmed arrowheads and bifacial foliates. The poorly-developed primary production is oriented towards the manufacture of flakes although natural pieces of shatter were often used as tool blanks. Production waste proves flint-knapping activities at the site. Likewise, most personal adornment was manufactured in the local tradition. In contrast, large amounts of Black-on-Buff pottery find their closest parallels at Ubaid period settlements in southern Mesopotamia. Other elements of material culture that originate in this cultural sphere are fired-clay objects, shaped bitumen and obsidian artefacts. The absence of related production waste, however, qualifies these objects as imports. Unexplained remains the origin of the Coarse Ware. While evidence for local production at Dosariyah itself is absent, the incised decoration found on one single Coarse Ware vessel is clearly distinct from decorative styles known from southern Mesopotamia. The mechanisms behind the distribution of foreign materials along the Arabian Gulf coast are still poorly understood. The assumption that Arabian societies were less developed in comparison to societies in Mesopotamia led to the unjustifiable view that foreign traders coming from the upper Gulf were responsible for the appearance of foreign objects in the central Gulf area and beyond. However, the exploitation of marine resources always played a major role for local populations settled along the coasts of the Arabian peninsula. Further, the spectrum of fish remains discovered at Dosariyah includes species that demonstrate the capability for off-shore fishing, while excavations failed to unearth convincing evidence for the presence of Mesopotamian seafarers. It is therefore the central proposition of this book that the local societies living along the shores of the Arabian Gulf coast played an active role in the acquisition of Ubaid pottery and other objects. In addition to the hunting of a broad spectrum of wild animals and the herding of sheep, goat and cattle, fishing and the gathering of shellfish formed the basis of the diet of the population settled at Dosariyah. Such a broad subsistence base can serve as a protection against environmental crisis, but it also helps to satisfy higher demands for food during short periods of time. As indications for fluctuating occupation densities were likewise observed during excavations, gatherings from time to time of larger groups of people at Dosariyah are plausible. A predominance of imported objects, considered as ‘exotic items’, can therewith be understood as integral components of rituals that were part of these gatherings. Based on the material evidence from Dosariyah, such collective social events were embedded in everyday life during the fifth millennium BC.

v

Zusammenfassung Zwischen 2010 und 2014 wurden in Dosariyah, einer an der Küste des Arabischen Golfes in der Eastern Province in Saudi-Arabien gelegenen Fundstelle, neue Forschungen durch das deutsch-saudische Dosariyah Archeological Research Project (DARP) durchgeführt. Während der Ausgrabungen konnte in unterschiedlichen Bereichen der Fundstelle ein breites Spektrum an Funden innerhalb einer bis zu drei Meter mächtigen stratigraphischen Abfolge anthropogener Sedimente dokumentiert werden. Das Alter der Besiedlung wurde sowohl radiometrisch als auch durch vergleichende Artefaktstudien in die ersten Jahrhunderte des 5. Jahrtausends v. Chr. datiert. Ergänzt wurden die Ausgrabungen durch lokale geomorphologische Untersuchungen und einen archäologischen Survey. Das Spektrum der materiellen Hinterlassenschaften in Dosariyah weist eine Reihe von Besonderheiten auf. Technologische und typologische Charakteristika des Steinartefaktinventars finden ihre Parallelen im Arabischen Mittelneolithikum. Insbesondere bifazial geschlagene, gestielte und geflügelte Pfeilspitzen sowie bifaziale Blattspitzen assoziieren das Inventar mit der Arabian Bifacial Tradition. Die auf die Herstellung von Abschlägen ausgerichtete Grundformproduktion ist wenig elaboriert, gleichzeitig finden als Ausgangsform für Werkzeuge häufig natürliche Silex-Trümmer Verwendung. Präparationsabschläge und Kerne geben deutliche Hinweise auf die lokale Herstellung von Steinartefakten an der Fundstelle. Auch Muschel- und Scheibenperlen, welche als persönlicher Schmuck interpretiert werden, zeigen deutliche Bezüge zum Arabischen Neolithikum. Demgegenüber steht die Herkunft großer Mengen von bemalter Black-on-Buff Keramik, welche sowohl in Gefäßform und Verzierung als auch in der chemischen Zusammensetzung des Tons große Ähnlichkeit mit zeitgleichen Tongefäßen der Ubaid-Kultur Südmesopotamiens aufweist. Andere Aspekte materieller Kultur, welche ebenfalls mit Funden aus Mesopotamien assoziiert werden können, sind Ton- und Bitumenobjekte sowie Obsidianartefakte. Da in Dosariyah Hinweise auf ihre lokale Herstellung fehlen, muss bei diesen Objekten von Importen ausgegangen werden. Ungeklärt dagegen bleibt der Ursprung der sogenannten Coarse-Ware, welche in Dosariyah mit einem Anteil von ungefähr 20 Prozent gefunden wird. Während auch hier Hinweise für eine lokale Herstellung der Keramik an der Fundstelle fehlen, unterscheidet sich das eingeritzte Dekor eines einzelnen Coarse-Ware Gefäßes sehr deutlich von allen aus Mesopotamien bekannten Verzierungsstilen. Bis heute können diejenigen kulturellen und sozialen Mechanismen, welche hinter der Verbreitung importierter Objekte entlang der Küste des Arabischen Golfes liegen, noch nicht vollständig nachvollzogen werden. Die Vermutung, dass Arabische Gesellschaften während des Neolithikums im Vergleich zu den Gesellschaften Mesopotamiens nur wenig entwickelt waren, führte lange Zeit zu der unbegründeten Annahme, dass nur Seefahrer aus Mesopotamien für die Verbreitung der Importe im zentralen Golf verantwortlich sein könnten. Forschungen an Küstensiedlungen auf der Arabischen Halbinsel konnten hingegen nachweisen, dass insbesondere die Nutzung mariner Ressourcen eine große Rolle für die lokalen Gesellschaften spielte. Der Nachweis von Fischknochen von Meeresfischen in Dosariyah, welche ausschließlich jenseits der Küste zu finden sind, belegt die Befahrung des offenen Meeres. Demgegenüber gelang es in Dosariyah nicht, die Anwesenheit von Personen aus Mesopotamien eindeutig zu belegen. Als Konsequenz wird im vorliegenden Buch argumentiert, dass nicht mesopotamischen Seefahrern, sondern den Vertretern lokaler Gesellschaften, welche an den Küsten des Arabischen Golfes lebten, eine zentrale Rolle bei der Erlangung und Verbreitung der Ubaid-Keramik und anderer Importe zukam. Zusätzlich zur Fischerei sowie dem Sammeln von Muscheln, stellte die Jagd auf ein breites Spektrum von Wildtieren sowie die Haltung von Ziegen, Schafen und Rindern einen wichtigen Beitrag zur Ernährung der Einwohner von Dosariyah dar. Eine derartig breite Subsistenzbasis kann als Schutz und Risikominimierung unter schwierigen Umweltbedingungen verstanden werden. Gleichzeitig hilft sie jedoch auch, einen kurzzeitig erhöhten Bedarf an Nahrung zu decken. Da während der Ausgrabungen an der Fundstelle deutliche Hinweise auf eine unterschiedliche Intensität der Besiedlung dokumentiert werden konnten, sind zeitlich begrenzte Zusammenkünfte größerer Bevölkerungsgruppen in Dosariyah denkbar. Die große Zahl importierter Gegenstände, welche als 'exotische Objekte' angesehen werden, können damit als zentrale Komponenten von Ritualen verstanden werden, die Bestandteile dieser Zusammenkünfte waren. Basierend auf den aktuellen Erkenntnissen aus den Forschungen in Dosariyah fanden derartige gemeinschaftliche soziale Ereignisse eingebettet in das tägliche Leben in einer Küstensiedlung des 5.Jt. v. Chr. statt.

vi

‫ال ُملخص‬ ‫ٌ‬ ‫بحث جديدٌ خالل الفترة بين عامي ‪ 2010‬و‪ 2014‬م ضمن إطار مشروع البحث األثري األلماني السعودي (‪ )DARP‬في الدّوسرية وهي منطقة أثرية‬ ‫أُجري‬ ‫تقع على شاطئ الخليج العربي في المنطقة الشرقية من المملكة العربية السعودية‪ .‬أمكن خالل أعمال التنقيب توثيق تسلسل طبقي للترسبات والبقايا البشرية‬ ‫والعثور على مجموعة واسعة من اللّقى في أماكن مختلفة من هذه المنطقة األثرية على عمق يتراوح من من متر إلى ثالثة أمتار‪ .‬تم تأريخ اإلستيطان في‬ ‫القرون األولى من األلف الخامس قبل الميالد باإلعتماد على المواد المشعة وعلى دراسة تعتمد على مقارنة القطع األثرية‪ ،‬هذا وقد استكملت أعمال التنقيب‬ ‫بتطبيق دراسة جيومورفولوجية محليّة ومسح أثري‪.‬‬ ‫تمتلك اآلثار المادية التي عُثر عليها في الدوسرية العديد من السمات الخاصة‪ ،‬وقد أظهر جرد القطع األثرية الحجرية خصائصا ً تقنية وطبولوجية متشابهة‬ ‫مع مثيالتها من العصر الحجري الوسيط في المنطقة العربية‪ ،‬مثل العثور على نصل حجري محفور لرأس سهم له جناحين وكذلك أنصال لها شكل ورقة‬ ‫الشجر المسنّنة التي أوحت جميعها بتقليد صناعة النّصل العربي‪.‬كانت أشكال المادة الخام األساسية المستخدمة في اإلنتاج غير متطورة‪ ،‬وبنفس الوقت لوحظ‬ ‫استخدام كسرات السليكس الطبيعية كمادة خام لصناعة أدوات العمل‪ .‬تعطي األجزاء المشغولة والحبوب التي عُثر عليها مؤشرات واضحة على االنتاج‬ ‫المحلي لقطع الحجر ضمن المنطقة األثرية‪ ،‬وكذلك المحار والصدف وقطع اللؤلؤ التي كانت على األغلب تستخدم كحلي زينة شخصية تملك دالئل واضحة‬ ‫على العصر الحجري الحديث في المنطقة العربية‪.‬‬ ‫ومن ناحية أخرى يعود أصل عدد كبير جدا ً من رسومات الفخار ‪ -‬المرسومة باللون األسود على خلفية من اللون البنّي المصفر‪ -‬إلى ثقافة عصر العُبيد التي‬ ‫كانت منتشرة جنوبي بالد الرافدين نظرا ً لوجود تشابه بأشكال الجرار أو التزيينات أو عن طريق تشابه المكونات الكيميائية للصلصال المستخدم‪ .‬كذلك األمر‬ ‫صلصال والقير وكذلك حجر السبج‪ ،‬و لعدم توفر الدّليل الكافي في الدّوسرية‬ ‫فقد عُثر على مواد أخرى توحي بلُقى من بالد الرافدين بعضها مصنوعة من ال ّ‬ ‫على أنّها منتجات محلية فيعتقد باستيرادهم من مناطق أخرى‪.‬‬ ‫أما أصل المنتجات الف ّخارية الخشنة والتي عُثر عليها في الدّوسرية بنسبة ‪ 20%‬تقريبا ً بقي غير معروفاً‪ ،‬وعلى الرغم من عدم توافر دالئل على اإلنتاج‬ ‫المحلي للف ّخار هنا فإن الفرق واضح بين المنتجات المزينة بالتحزيز والقطع المزينة من بالد الرافدين‪.‬‬ ‫ال يمكن حتى اليوم فهم اآللية الثقافية واإلجتماعية التي تقف خلف انتشار استيراد العناصر على طول ساحل الخليج العربي‪ .‬هذا ويعتقد أن المجتمعات‬ ‫العربية خالل العصر الحجري الحديث لم تكن متطورة بشكل كافِ بالمقارنة مع بالد الرافدين‪ ،‬مما قاد إلى أن البحّارة من بالد الرافدين فقط هم من كانوا‬ ‫مسؤولين عن التوريد إلى وسط الخليج‪ ،‬وعلى العكس من ذلك فقد أظهرت البحوث المتعلقة بالمستوطنات الساحلية في شبه الجزيرة العربية أن الموارد‬ ‫البحرية كانت تلعب دورا ً رئيسيا ً في حياة السكان المحليين‪ .‬كما أن العثور على عظام األسماك البحرية في الدّوسرية والتي ال يمكن العثور عليها إال في‬ ‫الجانب اآلخر من السّاحل‪ ،‬لهو دلي ٌل آخر على استكشاف البحر والوصول إلى هناك‪ ،‬وفي المقابل لم يتسنى إثبات تواجد واضح ألشخاص من بالد الرافدين‬ ‫الرئيسي الستيراد ونشر ف ّخار عصر العُبيد وغيرها من الواردات لم يكن لبحارة بالد الرافدين‬ ‫في الدّوسرية‪ ،‬وكنتيجة لذلك يبرهن هذا الكتاب أن الدّور ّ‬ ‫وانّما قام به مندوبين من سكان محليين مقيمين على ساحل الخليج العربي‪.‬‬ ‫صيد وجمع المحّار‪ ،‬ويعدّ ذلك من‬ ‫هذا وقد ساهمت تربية الماعز واألعنام والبقر بشكل رئيسي في تأمين المواد الغذائية لشعب الدّوسرية باإلضافة إلى ال ّ‬ ‫عوامل اإلكتفاء الذّاتي التي تؤ ّمن الحماية وتقلّل من المخاطر أثناء الظروف البيئية القاسية‪ ،‬وفي نفس الوقت تساعد على تغطية الحاجات المتزايدة بشكل‬ ‫مفاجئ على مدى قصير‪ ،‬وبما أنّه أمكن توثيق أدلّة هامة اثناء التنقيبات تثبت إمكانية وجود كثافة استيطانية فقد قاد ذلك إلى التفكير بتجمعات بشرية كبيرة‬ ‫بأوقات متقاربة في الدّوسرية يمكن أن يعود إليها عدد كبير من األشياء المستوردة على أنها «عناصر غريبة نادرة» استخدمت في ممارسة طقوس محددة‪.‬‬ ‫استنادا ً إلى نتائج البحوث الحالية في الدّوسرية فقد كانت هناك فعاليات اجتماعية مشتركة تنتمي إلى الحياة اليومية في مستوطنة ساحلية تعود إلى األلف‬ ‫الخامس قبل الميالد‪.‬‬

‫‪vii‬‬

Acknowledgements The Dosariyah Archeological Research Project (DARP) and the results presented in this monograph could not have been achieved without the collaboration of institutions, colleagues, students and friends. DARP was established as a joint archeological research project between the Saudi Commission of Tourism and National Heritage (SCTH) and the University of Tübingen. Financial support for the project was kindly provided by the German Research Foundation (DFG), research grants DR824/1-1 and DR824/1-2, the University of Tübingen (Förderung für Nachwuchswissenschaftler) and SCTH. Additional institutional support came from the German Embassy in Riyadh. Hans-Peter Uerpmann, Margarethe Uerpmann (University of Tübingen), Ricardo Eichmann (German Archeological Institute) and Abdulhamid M. al-Hashash (Dammam Archeological Museum) were crucial for the success of the project. While Margarethe and Hans-Peter Uerpmann provided help and advice during the whole process of research and funding application, fieldwork and the study and interpretation of finds, Ricardo Eichmann initially opened the door for my work in Saudi Arabia. Abdulhamid al-Hashash welcomed me with great openness in Dammam. Without his enthusiasm in the field of archeology, work at Dosariyah would not have been possible at all. The research application for reinvestigations at Dosariyah was well received by the scientific research committee of the Saudi Commission for Tourism and Antiquities (SCTA), now the Saudi Commission for Tourism and National Heritage (SCTH). I would like to express my sincere gratitude to Ali I. al-Ghabban, Vice-President of SCTH and Supervisor of the Custodian’s Project for Caring of Kingdom’s Cultural Heritage for his ongoing interest and support of the project. My special thanks to Daifallah al-Thali, Khalid M. Escoubi, Abdulaziz S. al-Ghazzi and Jamal S. Omar, Directors for Survey and Excavation during different stages of the project for their great help and unequalled cooperation. In addition, I would like to thank sincerely Abdullah H. Masry, whom I met in Riyadh in 2012. He kindly allowed me to study and copy all his field notes and photographs from his investigations at Dosaryiah that remained in his private library. I was very impressed by the accuracy and thoroughness of his documentation. During the planning phase of the project I benefited from the broad expertise of Nicholas Conard (University of Tübingen), Mirko Nowak (University of Bern), Peter Jablonka (University of Tübingen) and Simone Riehl (University of Tübingen). Collaborating researchers from the outset of the project were Canan Çakırlar (University of Groningen), Katleen Deckers (University of Tübingen), Corina Knipper (CEZ Mannheim), Simone Riehl (University of Tübingen) and Hans-Peter Uerpmann (University of Tübingen). Additional colleagues became more deeply involved as the data and research questions grew: Christoph Berthold (University of Tübingen), Shumon Hussein (University of Leiden), Christine Kainert (Free University of Berlin), Felix Levenson (Free University of Berlin), Peter Magee (Bryn Mawr College), Susan Mentzer (University of Tübingen), James Nebelsick (University of Tübingen), Adrian Parker (Oxford Brookes University), Martin Posselt (Posselt and Zickgraf GbR), Margarethe Uerpmann (University of Tübingen) and Thomas van de Velde (Ghent University). I would like to thank them all for their efforts in answering the many questions resulting from the fieldwork at Dosariyah as well as their encouragement and professionalism. Over the whole course of the project, Susan Pollock (Free University of Berlin) and Rob Carter (UCLQ) have closely followed the research at Dosariyah and I would like to thank them very much for all their thought-provoking input. The fieldwork at Dosariyah would not have been possible without the kind support from colleagues working at Dammam Archeological Museum. In particular I would like to thank Azzhr al-Toobi and Hassan Hamdoun for their willingness to participate in the fieldwork. I very much appreciated their true enthusiasm in the field. During my first stay at Dammam Archeological Museum in 2008 I was introduced to Mahmoud al-Hijri who became a trusted friend during subsequent field campaigns. Finally, I have to thank Nabil al-Shaikh. His serenity was of great help especially during the more hectic phases of fieldwork. My grateful thanks also to all students and postgraduates who participated in the field campaigns between 2010 and 2013: Sirwan Ali, Gregor Bader, Julian Bega, Chris Britsch, Frank Brodbeck, Andrej Girod, Tobias Helms, Simon Herdt, Michael Herles, Georg Häussler, Benni Höpfer, Shumon Hussein, Christine Kainert, Felix Levenson, Stefan Piefel, Mathias Probst, Martin Riesenberg and Alexander Städler. It must have been hard for them to stay close together in the outskirts of Jubail in a fancy hotel for weeks, but I hope that they were able to take home something beyond profession: life experience.

viii

The quality of the individual chapters of the final book was considerably improved by reviewers who were not directly involved in the Dosariyah Archeological Research Project. I fully acknowledge the detailed comments and suggestions of Rob Carter, Rémy Crassard, Susanne Lindauer, Alicia Ventresca Miller, Leah Minc, Kris Poduska, Susan Pollock, Seth Priestman and Vince Van Thienen. Of course, all remaining errors are solely my and the authors’ responsibilities. My special thanks go to the BFSA Monograph Series editors St John Simpson and Derek Kennet. Their thorough assistance and verve safely guided me during the publication process. Lastly, I would like to thank Barbara and Anna. They not only tolerated my frequent absences from home during various field seasons and additional periods in Saudi Arabia, but too many times also accepted my distractions from everyday life during the different phases of the project. Philipp Drechsler May, 2018

ix

Preface The present monograph on reinvestigations at Dosariyah follows two research objectives. First, it outlines the results of fieldwork at this site and its immediate surroundings in the Eastern Province of the Kingdom of Saudi Arabia and comprehensively documents the range of material culture found during the archeological surveys and excavations. The time span of occupation is closely dated by an extensive series of radiocarbon dates and artefactual analyses which securely place the Neolithic settlement into the first half of the fifth millennium BC. The finds from Dosariyah presented in this book also serve as an important benchmark for less well-dated archeological contexts in eastern Arabia and therefore represent a substantial basis for future research in the region. Second, this monograph challenges several basic axioms of archeological research in the Arabian Gulf.1 For decades, the proverbial ‘seafaring merchants of Ur’ — or people from southern Mesopotamia in general — were seen as the driving force behind the distribution of Ubaid pottery and other ‘foreign’ elements of material culture in the Arabian Gulf during the fifth millennium BC. Dosariyah has the distinction of being the site with the highest number of imports, both in terms of counts and diversity of finds, but the presence of southern Mesopotamian people here, or at any other site in this region, remains unproven. Instead arguments are presented which suggest that an indigenous Arabian Neolithic population living along the shores of the central Gulf played an active role in the acquisition of goods from southern Mesopotamia. Being familiar with the exploitation of marine resources from the open sea, these people living at Dosariyah and other contemporaneous settlements along the shores of the Arabian Gulf were plausibly the agents that determined the (re)distribution of material elements of southern Mesopotamian material culture in the central Gulf. Moreover, aspects of environmental conditions, everyday life and material culture are also presented. The first part of the monograph introduces the site and its immediate surroundings. Chapter 1 reveals the history of research and contextualizes Dosariyah in terms of its environment and archeology. A more detailed analysis of the site’s environmental history and its change through time is presented in Chapter 2. Geomorphological and paleoenvironmental studies focus especially on the fluctuating sea level of the Arabian Gulf during the Holocene, but also consider the development of wide sabkha flats that form a prominent feature in the landscape around Dosariyah. The Dosariyah survey, covering the direct vicinity of the site, is outlined in Chapter  3. It takes a diachronic perspective and considers human activities in the wider surroundings from the Neolithic to modern times. Geophysical surveys carried out at Dosariyah are presented in Chapter 4. Although not successful in detecting architectural remains, these investigations clearly suggest the presence of anthropogenic accumulations beyond the central fenced area of the site. Chapter 5 introduces the archeological methodology and excavation techniques, interprets the distribution of finds collected from the surface of the site prior to excavations and describes the excavated stratigraphy and deposits. Chapter  6 completes this first part of the monograph and compiles all radiometric and archeological data which help to date the occupation at Dosariyah. The second part of the monograph focuses on the varied material culture recovered from the excavation. Most prominent are finds of thousands of sherds of (partly painted) Ubaid pottery, deriving from vessels imported from southern Mesopotamia. Supplemented by coarse ware of unknown origin, the pottery assemblage from Dosariyah, presented in Chapter 7, is the largest excavated in the Arabian Gulf so far. Geochemical analyses of selected pieces of pottery are discussed in Chapter 8. Chapters 9 and 10 describe objects that also fall into the realms of ceramics: reworked pottery and fired clay objects. The spectrum of artefacts made from stone, bone and shell are presented in subsequent Chapters 11 to 15. The lithic industry outlined in Chapter  11 provides conclusive evidence for the presence of indigenous Arabian populations inhabiting Dosariyah, while a foreign, probably Mesopotamian lithic technology is exclusively restricted to artifacts made from imported raw material. A morphometric study of arrowhead shapes, presented in Chapter  12, convincingly documents the morphological homogeneity of arrowheads and suggests occupation by a single, culturally homogeneous, group of people. A group of groundstone artifacts rarely observed in Neolithic contexts in eastern Arabia are objects made from hematite discussed in Chapter 13. While this generally represents a suitable raw material for stone tools, the blood red color that can be obtained from ground hematite might have been used for decorative or ceremonial purposes as well. A rich bone 1 I am fully aware of the political dispute about the naming of the Persian — or Arabian, depending on one’s point of view — Gulf. According to international organizations including the United Nations and the International Hydrographic Organization and historical sources, the term Persian Gulf is generally preferred. The use of this name, however, would be rejected by the majority of our colleagues from Saudi Arabia. Having taken advantage of their invitation to carry out research in Saudi Arabia and therefore feeling personally committed, I have decided to follow the tradition of archeologists working in the Arabian Peninsula and term the waters ‘Arabian Gulf ’.

x

industry described in Chapter 14 represents a little-known aspect of Middle Neolithic material culture in Arabia. Repeated finds of production waste prove the on-site production of bone tools while a restricted spectrum of bone tool types suggests a considerable degree of tool standardization. Important indicators for the cultural affiliation of the people settling at Dosariyah are pieces of personal adornment which are presented in Chapter 15. Chapters 16 to 19 consider two unusual groups of artefacts: bitumen objects and pieces of plaster. Detailed morphological studies and a broad range of analytical methods have been employed to analyze their origin and use at Dosariyah. Chapters 20 to 23 focus on subsistence strategies. The analysis of the zooarcheological assemblage (Chapter 20) allows detailed insights into both the terrestrial and marine component of everyday diet. It emphasizes the broad spectrum of exploited animals which include both domesticated and wild species. Additional isotopic analyses of cattle teeth (Chapter  21) help to understand patterns of animal husbandry and mobility. These studies are supplemented by analyses of marine shell (Chapters 22 and 23) which highlight the importance of this particular marine resource for the inhabitants of the site. Finally, the concluding Chapter  24 subsumes the results from individual studies with the aim of drawing a comprehensive — though fragmentary — picture of the community living at Dosariyah during the first half of the fifth millennium BC.

xi

xii

Chapter 1 The Site and its Context Philipp Drechsler

1 Location

archeologist and enthusiastic ‘pot picker’.1 While scouting for pottery fragments in the windswept salt flats south of Jubail she came to the site which she named ‘site 1’ in subsequent publications and maps, stating it was ‘carpeted with debris’.2 Not being familiar with the kind of decorated pottery that was frequent at the site, Burkholder contacted Geoffrey Bibby of the Danish archeological expedition to the Gulf, which was then visiting archeological sites in the Eastern Province of Saudi Arabia.3 Bibby identified the pottery as Ubaid pottery known from archeological sites in southern Mesopotamia. He further concluded from a short reconnaissance study of the area: ‘Here (…) was a strip of low sandy hills fencing the sabkha off from the sea. It must have been a string of islands, I thought, six and seven thousand years ago, when the sabkha was sea. There were no traces of buildings, Grace went on, but there were pieces of plaster showing a smooth face on one side and the impress of bound bundles of reeds on the other’.4 With an approximate dating to the fifth millennium BC, Burkholder’s discovery of Dosariyah extended the ‘range of the Ubaid culture by nearly a thousand kilometers’, but also ‘added a millennium to Saudi Arabia’s prehistory’.5 By 1972, Burkholder had identified as many as 39 locations with fragments of Ubaid pottery, including the sites of Abu Khamis in the Ras al Ghair area (quoted as site 11), and Khursaniyah (quoted as site 2).6 While most sites with Ubaid pottery that were found by Burkholder north of Jubail are located close to the present-day coast, inland sites also occur towards the south and these are often situated along the edges of sabkhas. In addition, sites identified in al-Hasa oasis and around Ain Dar and Abqaiq also provided pieces of Ubaid pottery. Although Burkholder did not provide a detailed list of objects she collected from the surface of Dosariyah, most of the pieces of jewelry, flint and obsidian implements and of painted Ubaid pottery illustrated in her initial publication are now (2014) on display in the National Museum in Riyadh.7

Dosariyah is located 10  km south of the old town of Jubail in the Eastern Province of the Kingdom of Saudi Arabia, approximately 1  km inland from the present shores of the Arabian Gulf (Figure 1.1). Surrounded by salt flats (sabkhas), the archeological site is situated within a shallow trough or valley close to the southern edge of a raised area surmounting the sabkhas by up to 14 m. The name of the site refers to the local family of al-Dossary who in the past used the area as a preferred grazing ground. The immediate surroundings of the site were still wilderness as late as 2008, when only a few tents and small wooden houses used for leisure activities at weekends as well as corrals for sheltering camel, sheep and goat, existed in the area. During archeological fieldwork at the site between 2010 and 2013, however, extensive construction activities were begun to the north of Dosariyah as new residential neighborhoods were built for the growing population of Jubail. The infill of sediment and rubble into the salt flats as well as earth-moving in connection with these construction activities represented major obstacles for both archeological and geomorphological surveys in the area. In 2015, construction work began directly west of the site with heavy bulldozing, which reached the fenced area of the site. It is therefore foreseeable that at least the archeological and geomorphological context of the site will be completely destroyed during the next few years. 2 History of Research The history of research at Dosariyah falls into three different phases: its initial discovery in 1968, archeological excavations and scientific studies during the early 1970s, and the joint Saudi-German Dosariyah Archeological Research Project (DARP) between 2010 and 2014. 2.1 Initial discovery

1 2

Dosariyah was initially discovered in 1968 by Grace Burkholder, a school teacher for the Arabian American Oil Company (ARAMCO) in Dhahran, amateur

3 4 5 6 7

1

Burkholder 1972: 264. Burkholder 1972: 264; see also 1984. Bibby 1970: 376. Bibby 1970: 376. Burkholder 1972: 264. Burkholder 1984: 17. Cf. Burkholder 1972.

Dosariyah

Figure 1.1. Location of Dosariyah south of Jubail Old Town.

2.2 First excavations and scientific studies

were called ‘Al Hasa Expedition – The Ubaid Sites Project. Winter/Spring 1972’ and consisted of five main parts: 1) test excavations at Ain Qannas in the Al Hasa area; 2) excavations at Dosariyah and surveys in the surrounding coastal zones; 3) test excavations and surveys around Abu Khamis; 4) excavations at several locations on Tarut island; and 5) surveys in the Jabrin oasis.9 Over the course of these investigations Masry’s field investigations considered all areas and the most substantial sites identified by Burkholder.

After examining the pieces of painted Black-on-Buff pottery that Burkholder had collected at Dosariyah and other sites in the Eastern Province, Bibby called for immediate investigations of those findings and of the underlying mechanisms of their distribution. These field investigations were subsequently carried out in spring 1972 by Abdullah Masry, then a PhD student at the Oriental Institute in Chicago and a member of the Saudi Arabian scholarship program for higher studies abroad. Based on this fieldwork, Masry presented his dissertation, which was supervised by Robert McC. Adams and Robert J. Braidwood and entitled Prehistory in Northeastern Arabia: The Problem of Interregional Interaction, to the Department of Anthropology at the University of Chicago in June 1973.8

The field expedition began on March 1, 1972 when Masry left Riyadh together with four assistants: two field surveyors, a driver and a cook and after nearly two months of ‘trials and tribulations involving financial, logistic, personnel and a host of other matters and complications’.10 After investigations at Ain Qanas (site 18 according to Burkholder’s enumeration), Masry and his field team reached Dosariyah on March 23.

Originally planned as a collaborative field project involving four students from Chicago, only Juris Zarins was finally able to take part in the work in Saudi Arabia. The field investigations carried out by this expedition 8

Masry 1974: v. Field notes by A. Masry. The original field notebook as well as all other field documents and photographs have been made accessible to the author by courtesy of A. Masry and are henceforth referred to as Masry 1972. 9

10

Masry 1974.

2

Philipp Drechsler: The Site and its Context

Figure 1.2. a Location and orientation of Masry’s surface survey grid in relation to the grid used by the Dosariyah Archeological Research Project. b Field notes from surface sampling (courtesy A. Masry 2013).

Inspections of the site area identified two different parts which were subsequently labelled site 1a (the main site) and a smaller site (site 1b) located approximately 500 m to the north.11 Also noteworthy was a thick vegetation cover at the site, in sharp contrast to the dune-covered surroundings. As no workmen were available in Jubail, fieldwork at Dosariyah started on March 25 with topographic surveys and collection from the surface of the site. To do this systematically, a square grid measuring 10  x 10  m was established (Figure 1.2a) and each square collected separately during the next few days until March 28. Based on the color of paste and treatment, Masry established a ‘rough typology of the pottery’, which was used for the classification of ceramics (Figure 1.2b).12

this area, dense and thick layers of shell were unearthed immediately below the surface. According to Masry’s field notes, excavations were carried out in spits measuring between 20 and 40 cm in depth. Down to a depth of 1 m Masry noted: ‘What is remarkable about the finds until now is that they duplicate 1:1 those lying scattered on the surface, in terms of variety’.13 On the next day, and at a depth of 1.6  m below the surface, the ‘same variety of artefacts and finds’ was noted. At that point, ‘the trench sides fell down before we dug any deeper below [level] 4. The dryness and looseness of sand plus wind blow [one word unreadable] to make of the trench walls a very fragile easily crumbling sides: A level 5 was partially exposed before the trench collapsed’.14

Excavations at Dosariyah started on March 28 with sondage I, located in square 89 at the western steep slope of the site and measuring 2 x 2 m (Figure 1.3). In

Until April 8, a total of seven soundings were excavated in squares 83, 88, 89, 90 and 92, with each sounding measuring 2  x 2  m (Figure 1.4). During the excavation of sondages I to IV in squares 88 and 89 at

Site 1b is located only 330  m to the north, measured from the highest point of Dosariyah. 12 Masry 1972: March 26. 11

13 14

3

Masry 1972: March 29. Masry 1972: March 30.

Dosariyah

Figure 1.3. a Location of soudings in squares 83, 88, 89, 90 and 92. b The beginning of excavations at Dosariyah. c Field notes, March 28, 1972 (courtesy A. Masry 2013).

the westernmost part of the site, Masry noted that the succession of sediment layers did not only dip to the east, but also came increasingly closer to the surface — suggesting a west–east dipping of the underlying paleotopography of the site.15 Further on, at this early stage of excavations and on the basis of massive layers of shell and the absence of plaster pieces, Masry proposed that the western part of the site once represented a dumping area of a settlement further east.16

investigations at Dosariyah, Masry stopped excavations at the site and returned to Dammam the next day. The results of Masry’s investigations at Dosariyah were published in his dissertation submitted in June 1973.19 A republication in 1997 uses the original manuscript with some minor changes in both the text and figures.20 2.3 Dosariyah Archeological Research Project (2010– 2014)

On April 4 investigations moved temporarily towards site 1b. After surface collection, one sounding (A) was excavated that indicated only a very shallow occupation in this area.17 As site 1b did not show potential for further work, Masry returned to site 1a and began excavating sondage VII, located in square 92. Excavating this sounding down to a depth of 2.4 m, higher densities of pottery and flint but less dense shell accumulations led him to conclude that ‘It seems that, when in the east side of the site there must have existed more dense occupation than near the shell mound’.18 In the afternoon of April 8, 1972, after 10 days of field 15 16 17 18

The Dosariyah Archeological Project arose from my first visit to the site in October 2008. At that time, I was kindly invited by the Saudi Commission for Tourism and National Heritage (SCTH) to visit Neolithic sites in the Eastern Province for potential joint excavation and research projects. Enthusiastically guided by Abdulhamid al-Hashash, then head of the Dammam Archeological Museum, and Mahmoud al-Hijri, one of his colleagues, we went northwards from Dammam to visit the sites of Dosariyah and Abu Khamis. Although Abu Khamis was much more impressive both in its size and scenery, Dosariyah looked more promising for the study of an Arabian Neolithic coastal community in the

Masry 1972: March 30; 1974: 121. Masry 1972: March 30. Masry 1972: April 5. Masry 1972: April 5.

19 20

4

Masry 1974. Masry 1997.

Philipp Drechsler: The Site and its Context

Figure 1.4. Excavations at Dosariyah, 1972. a Sondage II. b Sondage IV, level 2. c Sondage I: survey engineer Salah al Helwa leaves the trench (courtesy A. Masry 2013).

11 in the Abu Dhabi Emirate, excavated by Katelin Flavin, Elizabeth Shepherd and Mark Beech.22 A first proposal for the establishment of a joint Saudi-German research project focusing on excavations at Dosariyah was kindly approved by SCTH (formerly SCTA) in 2009 and extended in 2013. A research grant from the Deutsche Forschungsgemeinschaft (DFG) provided the financial background for all investigations.

central Gulf area. The surface of the site was littered with archeological remains, and animal burrows suggested the presence of ashy material indicative of substantial activities all over the well-fenced site. Further on there was no evidence of a later settlement at the site after the Neolithic, but limited amounts of modern rubbish suggested some kind of occupation during the short period of time between Masry’s excavations in 1972 and the fencing of the site in the early 1980s.

Investigations of the Dosariyah Archeological Research Project (DARP) were intended to center on four main research questions:

Although my initial visit did not last longer than 30 minutes, the potential of the site for reinvestigations was clearly evident given its size, preservation conditions and densities of material remains. Masry’s previous excavations had furthermore proved the deep stratification of the site that led to my expectation that the occupation at Dosariyah covered at least several centuries. In addition, fieldwork at Dosariyah would provide the opportunity to fill the geographical gap between two other Middle Neolithic sites investigated at that time in the Gulf: H3/As-Sabiyah in Kuwait, excavated by Rob Carter and Harriet Crawford between 1998 and 2004 and finally published in 2010,21 and Dalma 21

•• what are the economic and cultural roots of the Neolithic in the eastern part of Arabia? •• what characteristics of the Arabian Neolithic are the results of economic and social adaptations to local environmental conditions? •• what kinds of resources were used during the Middle Neolithic occupation of eastern Arabia? •• which kinds of relationships existed between the Arabian Neolithic and the Ubaid culture of Mesopotamia? 22 Flavin and Shepherd 1994; Beech and Elders 1999; Beech et al. 2000; 2016; Beech and Glover 2005.

Carter and Crawford 2010a.

5

Dosariyah To approach these questions, the project was at first planned for three years, with field seasons taking place both in the spring and fall. Nevertheless, uncomfortable climatic conditions with heat and heavy winds during the first fall season forced a change of plan, with only one excavation season per year. It therefore became necessary to extend the project until 2014.

during previous excavation seasons. In addition to laboratory work, one week was spent at the site to test one area (trench N2) suggested by georadar investigations and to trial a new excavation technique that would prevent the collapse of large and deep trenches. In February and March 2012, excavations took place in both the northern and eastern part of the site and these were actively supported by colleagues from Dammam, Riyadh and Doha. The goal to document the full stratigraphic sequences in these parts of the site necessitated the establishment of two large trenches, each measuring 8  x 8  m (E1.1, N3). An additional extension by 2  m to the west became necessary for trench E1.1 due to the presence of a single installation. Towards the end of the season, Robert Carter and Susan Pollock visited the excavations at Dosariyah and during this time, an excursion was organized which led us to Tarut, Al Hasa and Abu Khamis.

A first field season took place for five weeks in February and March 2010. Together with two archeologists from Dammam Regional Museum, Azzhr al-Toobi and Hassan Hamdoun, three test trenches, measuring 2 x 2 m and 2 x 3 m, were excavated by the German team in the central, southern and eastern part of the site. Although one of the major goals of the campaign, the documentation of the site’s complete stratigraphy, was only reached in trench S1 due to the massive collapse of trenches, results from this campaign clearly demonstrated the potential of Dosariyah, especially the high density of finds in all parts of the site and the unexpectedly excellent preservation of fish and mammal remains. In contrast, the absence of any evidence of architectural remains was somewhat disappointing. Datable material was retrieved during excavation and provided the basis for a first series of four AMS radiocarbon dates from Pinctada radiata (pearl oyster) shell that all fell within the fifth millennium cal. BC.23

A second season that was dedicated to the analysis and documentation of finds took place in October 2012. During a week in the storerooms of the old National Museum in Riyadh, the inventory of objects from Masry’s excavations at Dosariyah in 1972 was documented and the finds partly redrawn. During a short meeting with A. Masry permission was kindly granted to study the field notes from his investigations during the next months. Afterwards, the mission moved to Dammam. Kindly hosted by A. al-Hashash at the provisional offices of the Dammam Regional Museum, almost all the objects recovered during the previous field seasons were studied in detail.

The second field season in October and November 2010 suffered from extreme heat. Nevertheless, the start of larger-scale excavations in the southern part of the site was promising as we were able to document a total of 10 natural pearls among many other finds from trench S2. Again, evidence of architecture remained elusive. As in the previous season, Azzhr al-Toobi and Hassan Hamdoun from Dammam Regional Museum joined the excavations.

A short visit in Riyadh in February 2013 was reserved for the study of Masry’s field notebooks and photographs. As the amount of written and photographic record that was available clearly exceeded all my expectations, I obtained permission to make copies of most of the material for subsequent detailed evaluation. In November 2013, intensive foot surveys were carried out in the vicinity of Dosariyah with the aim of contextualizing the site. Short visits in search of flint raw material sources and additional Middle Neolithic sites led the team to Thaj and Khursaniyah. In addition, geomorphological studies were carried out by Adrian Parker and Mike Morley over 10 days.

As a consequence of the first two seasons, the consecutive third field season in spring 2011 started with geophysical investigations aimed at locating architectural remains. Both geomagnetic and georadar measurements were carried out within the fenced area of the site and within a restricted area further south by Martin Posselt from Posselt and Zickgraf GdR. While geomagnetic measurements did not work particularly well due to severe disturbance caused by the metal fence, results from georadar survey indicated several hot spots for potential architecture. Archeological excavations during that season focused on the complete excavation of trench S2 down to the natural soil.

In parallel with the writing of this monograph, a twoday symposium was held at the University of Tübingen in June 2014 during which contributors of this book and external specialists met for discussions about the essential results of the Dosariyah Archeological Research Project. One major goal of this workshop was to address the research questions formulated at the very beginning of the project in light of the fieldwork results and the study of finds.

The large number of archeological remains prevented the study of finds during the field seasons. It was therefore decided to reserve the fall 2011 season for analysis and documentation of the objects retrieved 23

Drechsler 2011: 74.

6

Philipp Drechsler: The Site and its Context 3 Environmental Context

to 100 times the local mean annual rainfall, resulting in a very high water deficit for plants.32 Dew is an important source of moisture in coastal areas where it occurs at night with up to 0.04 mm per night recorded in the winter months.33

Dosariyah is located in the central coastal lowlands of the Eastern Province of Saudi Arabia, about 800 m inland from the present shores of the Arabian Gulf. The central coastal lowlands stretch along the coast approximately between As Saffaniyah in the north to Dhahran in the south, while they are bordered by the northern Summan plateau in the west.24 They are characterized by moderate relief. The surface is generally covered by sands although limestone exposures are frequent. In the coastal area, wide sabkhas predominate while in the southeast up to Jubail, barren and unstable sand dunes represent the northward extension of the alJafurah desert.25

Wind patterns play a major role in the natural surface landscape development. Strong shamal winds affect the northern Gulf during the summer months with winds blowing at speeds of 40–50 km/hr, while northwesterlies originating from depressions in the Mediterranean occur from November to February.34 These winds are primarily responsible for the inland movement of sand across the central Gulf area,35 but they also transport vast quantities of dust into the lower atmosphere and severely reduce visibility.36 Wind data for the region indicate that the most important direction is from the north and northwest, with a second southeasterly peak originating offshore.37 Barth measured mean wind velocities between 12.6 and 24.8  km/hr for the region with gust values exceeding 43.2 km/hr.38 These winds have a powerful effect on the topography and in particular, the dunes.

3.1 Climate Today, most parts of the Arabian Peninsula form part of the Old World Dry belt, spanning Mauritania in the west, across North Africa and Arabia and into central Asia. According to the Köppen-Geiger climate classification system, BWh climatic conditions prevail. Precipitation is less than 50% of potential evaporation, with an average annual temperature above 18°C.26 With its geographical position around 27° northern latitude, Dosariyah is located within the sphere of the northeasterly trade winds, clearly north of the northernmost extension of the Inter Tropical Convergence Zone (ITCZ) during the Boreal summer.

Daily wind patterns follow a characteristic time schedule along the coast. In the morning, relatively dry winds blow from inland in the west towards the sea in the east. Heat radiation is higher over land during the night, inducing a gradient in local barometric pressure. Increasing insolation during the day reverses this pattern. An increase in temperature over land benefits the development of low barometric pressure and, as a consequence, humid wind moving from the sea towards the land normally begins blowing before noon and continues until sunset.39 These coastal convection winds tend to move sand inland before it is transported further south, currently preventing the development of large coastal dunes.40

The region has a desert and semi-desert climate, characterized by high summer temperatures and aridity throughout the year (Figure 1.5a–d). The mean annual temperatures recorded from the region are 26.5°C (Dhahran) to the south of Jubail,27 and 26.5°C (Abu Kharuf) and 26.5°C (Abu Ali) to the north.28 The mean July temperature at Dhahran is 35.8°C and 15.8°C for December. Extreme temperature values range between 49.2°C and 4.1°C at Abu Kharuf and 44.3°C and 12.1°C at Abu Ali. The average annual rainfall for the region is 85  mm with high inter-annual variability (Figure 1.5e). Mandaville reported annual rainfall values between 5 and 277  mm at Dhahran which is confined to the winter months from October to early May.29 Predominant sources of rainfall are Mediterranean depressions30 supplemented by local convection cells, cyclonic depressions that develop in front of the Zagros Mountains above Iraq and eastern Iran and currents from equatorial areas in Sudan and Ethiopia.31 Due to high temperatures, evaporation rates range from 35 up 24 25 26 27 28 29 30 31

3.2 Geology and geomorphology Eastern Arabia is part of the Arabian platform, built up by sediments of Paleogene and Neogene age. The strata dip gently away from the edge of the Arabian Shield in the west and form zones of cuesta topography with scarps following the curvature of the eastern edge of the Shield. Towards the Gulf these cuestas become barely perceptible41 and are often covered by Quaternary deposits. 32 33

See Mandaville 1990: 8. Fryberger et al. 1984. Peel et al. 2007: 1641. Vincent 2008: 74. Barth 2001. Mandaville 1990. Mandaville 1990: 11. Barth and Steinkohl 2004: 101.

34 35 36 37 38 39 40 41

7

Mandaville 1990: 16. Barth 2002. Loughland et al. 2012. Loughland et al. 2012: 194. Vincent 2008. Barth 2002. Barth 2001. Personal observation P. Drechsler; see also Vincent 2008: 147. Loughland et al. 2012: 194. Vincent 2008: 48.

Dosariyah

Figure 1.5. Climatic charts from a Qatif; b Dhahran; c Hofuf; and d Yabrin indicating average monthly temperature and precipitation (data source: Vincent 2008: tables 4.1, 4.3). e Climatic chart from Dhahran showing the inter-annual variability of the average monthly temperature and precipitation between January 1949 and March 1963 (data source: www.ncdc.noaa.gov; recalculated).

3.2.1 Geology

comprise hard limey sandstone, marl, soft sandstone and beach rock.

The relief between Dammam and north of Jubail is generally low with a few minor outcrops of Miocene and Pliocene bedrock belonging to the Hadrukh and Dam formations. Such outcrops usually occur as minor escarpments (5–20  m high). The Hadrukh formation that comprises sandy limestone, marl, gypsum, and beach rock underlies the whole of the Jubail area. One major outcrop occurs at Jebel Barri, 9 km northeast of Dosariyah (N26°53’54” E049°37’52”). In the limestone of the Hadrukh formations chert and gypsum layers are prominent. Occasional small outcrops can be found on the surface to the west of Dosariyah.42 Rock outcrops north of Jubail belong to the Dam formation and 42

A number of major structural geological features are found in the region. These include a series of north– south trending anticlinal structures that are oblique to the axis of the Arabian Gulf. Tectonic uplift and folding from the Middle Eocene has led to the development of several large structural anticlines. These include the Qatar and Dukhan anticlines located on the Qatar peninsula and the Bahrain anticline.43 To the west and southwest of Jubail, the Ghawar anticline is draped over a basement horst structure. It grew initially during the Carboniferous (Hercynian) deformation and was reactivated episodically during the Late Cretaceous.44 43

A. Parker, personal observation, November 2013.

44

8

Kassler 1973. Saner et al. 2005.

Philipp Drechsler: The Site and its Context 3.2.1 The central coastal lowlands

Kassler postulated that a northwest extension of the Bahrain Ridge passes through the Al Jubail area, parallel to the coastline, and that this tectonic mechanism accounts for the raised beaches observed near Jubail.45 According to Kassler, the Gulf of Salwah, to the southeast, is believed to be a Late Quaternary structural feature and represents an area of synclinal subsidence.46 Nevertheless, the raised beaches reported by Holm as far south as Salwah do not conform to this interpretation of the Gulf of Salwah being an area of subsidence.47

The central coastal lowlands are part of the Al-Jafurah desert that extends along the Arabian Gulf from Kuwait in the north to the Rub’ al-Khali in the south, over a distance of almost 800 km.54 The Al-Jafurah desert itself can be subdivided into three broad zones according to sand budget and wind speed. Within a northern zone, wind velocities are high and deflation predominates. The central zone, where Dosariyah is located, is characterized by dune transport while wind velocities are lower. In the first two zones sand dunes are more or less isolated while extensive sabkha plains exist. The southern zone stretches into the Rub’ al-Khali, where wind energies are lower and continuous dune fields predominate. Based on Landsat images, Loughland et al. outlined the actual spatial configuration of geomorphologic units of the Eastern Province that provides a general overview of the landscape features in the central coastal lowlands (Figure 1.6).55

The anticlines of the Gulf region are of great hydraulic importance. Due to the fractional deformation of the bedded rocks, water flows are generated even through the impervious layers. Artesian springs occur along the anticlines, forming a natural outflow.48 Hoetzl studied the regional hydrogeology and demonstrated that groundwater sourced from the Umm er Radhuma aquifer feeds spring systems along the Gulf.49 Aquifer recharge occurred during former humid periods with only a minor recent groundwater recharge component.

Studies by Anton indicate that the dunes of the AlJafurah desert south of Dhahran are not older than 4000 years.56 They may be the result of a major change in the sand dynamics of the northern regions, triggered by large-scale aridification which started around 5500 BP,57 followed by degradation of the northern areas.58 Anton and Vincent suggest that some of the sand in the Al-Jafurah sand sea originates from the major Quaternary fans developed by wadis transporting sands and gravels from the Arabian Shield towards the Gulf.59 One possible major source is the Ad-Dibdibah plain, a vast fan that was developed by the Wadi al-Batin-Wadi ar-Rumah system in northeastern Saudi Arabia. The Arabian Gulf Basin, which was exposed during periods of low sea level during the Late Pleistocene,60 represents another sediment source.

In addition to the major anticlinal structures, salt diapirism in the region has led to localized uplift. The Dammam dome, located 70  km to the southeast of Dosariyah, is a salt dome where Eocene age rocks crop out through the surrounding Quaternary sand and sabkha deposits.50 Modern uplift of the dome occurs at a rate of 0.56 to 0.75 mm per century.51 Deflated longitudinal dunes with their axis in a northnorthwest–south-southeast direction dominate the Quaternary sediments in the region. In places, secondary dunes reworked by eolian action to form smaller dunes during the Holocene overlie these longitudinal dunes. Sand sheets and barchans cover much of the region. Many barchans are found traversing the sabkhas. At many locations dome dunes likewise occur, which evolve into barchans and vice versa.52 Sand sheets form flat sandy plains, mostly covered by scattered perennial grasses and other seed-bearing plants. The coastal plain is low lying and in many places the water table is high enough to give rise to extensive sabkhas.53 In recent times, much of the natural landscape and vegetation has been destroyed by infrastructure developments associated with the oil and gas industry and the rapid population growth of the Eastern Province.

45 46 47 48 49 50 51 52 53

3.2.2 The landscape around Dosariyah The wider landscape around Dosariyah is characterized by a narrow coastal strip, dunes, interdune areas, sand sheets and siliciclastic sabkha terrains.61 Dosariyah itself is located on a raised area build up by eolianites. These fossilized dunes break off rather steeply towards the sea but are increasingly covered by sand sheets as they move inland. Both Pleistocene deflated longitudinal dunes and smaller Holocene dunes overlay the eolianites (see Chapter 2). Over wide areas, these

Kassler 1973. Kassler 1973. Holm 1960. Johnson et al. 1978. Hoetzl 1995. Weijermars 1999. Hariri 2014. Fryberger et al. 1984. Anton and Vincent 1986.

54 55 56 57 58 59 60 61

9

Fryberger et al. 1984: 413. Loughland et al. 2012. Anton 1983. Sirocko 1996. Barth 2001: 399. Anton and Vincent 1986. Lambeck 1996. Fryberger et al. 1984: 413.

Dosariyah

Figure 1.6. General geomorphological configuration of the central coastal lowlands. Dosariyah is located within a zone dominated by sabkhas (after Loughland et al. 2012: 196).

dunes are covered by scattered perennial grasses and other small plants.

scabridum are found on the surface of these sabkhas and indicate that they were once flooded during times of higher sea levels. A series of both active and inactive sand spits along the coast suggest that in the past the sabkhas were connected to the open sea. While clearly of marine origin, the surface morphology and height of the sabkhas is at least partly controlled by the local groundwater level.63 Strips of wind-blown sand cover part of the sabkha surfaces to the southeast of the

The landscape to the west, east and south of Dosariyah is dominated by wide sabkhas of the Sabkhat asSumm, which constitutes one northern part of the larger Sabkhat ar-Riyas.62 These sabkhas form a wide arc around the raised area where Dosariyah is located (Figure 1.7a). Shells from the marine snail Cerithium 62

Edgell 2006: 396.

63

10

Barth 1998.

Philipp Drechsler: The Site and its Context

Figure 1.7. Landscape around Dosariyah indicating the anthropogenic impact on the landscape. a 1968: declassified CORONA satellite image; b 2006: Google Earth.

raised area. They are both indicative of current sand transport induced by winds from the north-northwest and a comparatively older date of the sabkha itself.

connecting the oil and gas extracting areas north of Jubail with the processing and shipping facilities at Ras Tanura.

South and west of Sabkhat as-Summ, there are extensive dune fields interrupted by smaller sabkhas (Sabkhat al-Fasl, Sabkhat as-Safi, Sabkhat Mahbulah) and hard rock outcrops (Jebel Barri, Jebel Abu Sharif).64 Along the western borders of Sabkhat as-Summ and Sabkhat alFasl, there are extended palm groves that are no longer under cultivation but which indicate near-surface occurrences of sweet water.

3.2.3 Hydrology Today, perennial rivers do not exist in eastern Arabia and occurrences of sweet water are therefore restricted to springs and wells fed by (fossil) groundwater, and ponds receiving surface runoff from rainfall. Relict palm stands along the western edges of Sabkhat asSumm are indicative of extensive subsurface sweet water sources as are several shallow hand-dug wells in the region that once provide relatively good, if brackish, water.65 Sporadic occurrences of individual date-palm trees in the vicinity of Dosariyah and especially along the edges of the eolianites close to the seashore suggest the presence of sweet water closer to the site as well.

Today, most features of the landscape are widely blurred by infrastructure developments (Figure 1.7b). North of Dosariyah, the construction of residential neighborhoods has led to the infill of sabkhas. Further north along the coast, the King Abdul Aziz Naval Base occupies wide areas. Along the southern edge of the sabkha south of the site, an extensive housing complex was built for the Jubail seawater desalination plant run by SWCC (Sea Water Conversion Cooperation). Other major interference is caused by several pipelines crossing the area from northwest to southeast, 64

Extensive occurrences of artesian water were observed both within the dune fields and in the sabkhas during geomorphological studies carried out during the planning of the Industrial City north of Jubail and

Defense Mapping Agency 1982.

65

11

Mandaville 1990: 8.

Dosariyah approximately 15  km north of Dosariyah.66 With sodium chloride (NaCl) concentrations down to 0.2%, and gypsum (CaSO4) concentrations of 0.07% this water is almost potable.67 According to James and Little, two different sources for this artesian water exist:68 extended aquifers which conduct water from distant sources in the west to the springs at Bahrain and Qatif, as well as to submarine springs along the southern shores of the Gulf; it is suggested that a second source of sweet water exists in the dune fields surrounding the sabkhas of the central coastal lowlands themselves. These dunes contain high groundwater tables up to 20 m above the sabkha surfaces. As some pinnacles of the Hadrukh formation reach high levels in the dunes, it is possible that water stored at high levels in the dunes enters the Hadrukh formation through sub-vertical fissures contributing to artesian pressure beneath the sabkhas.69

While generally considered to be of limited value for grazing, rimth provides good firewood which is often carried some distance by Bedouins.74 Thumam grass shrubland is found on relatively well-drained sand substratum, although it also occurs on elevated ground near saline terrain. It is widely distributed in the central coastal lowlands where it represents a very important grazing association. Panicum turgidum is often found associated with shrubs of Calligonum comosum that produces excellent firewood.75 In the past, the coastal areas of the central coastal lowlands were comparatively densely vegetated although major grazing activity during the last few decades, and a continuously increasing animal population accompanied by supplemental feeding and trucking of water, have changed the vegetation significantly.76 Former extensive shrubland communities covering 15–20% of the ground have become degraded thumam grasslands that seldom exceed a ground cover of more than 5%.77

3.2.4 Flora The flora of the Arabian Gulf coast belongs to the NuboSindian center of endemism,70 generally characterized by scattered trees of species such as Aerva javanica (Arabic: ra’), Acacia tortilis (Arabic: samur), the saltloving shrub Halopeplis perfoliata (Arabic: khurrayz) and the grass Panicum turgidum (Arabic: thumam).71

3.2.5 Fauna The wild fauna of eastern Arabia is heavily affected by human action today. Millennia of intensive hunting as well as recent industrialization and urbanization have had a very strong impact both on the faunal composition and frequencies of wild animals.

While the wide coastal sabkhas of the central coastal lowlands are devoid of higher plant life, areas covered with fixed sand show denser vegetation and provide the habitat preferred by wild or naturalized date palms and Tamarix shrubs.72 Two trends in the local vegetation can be observed along the central Gulf area: the overall proportion of therophyte elements (annual plants) declines from north to south and this is paralleled by decreasing total species diversity.73

The Arabian Peninsula shows a four-fold faunal division between the Afro-tropical, Saharo-Sindian, Palearctic and Oriental Region. Eastern Arabia can be assigned to the Saharo-Sindian region.78 It is part of a vast region that links the western Sahara to the deserts of Afghanistan and central Asia. Animal species in this region are particularly adapted to arid, dry and desiccated landscapes. Characteristic herbivores of this faunal community in eastern Arabia were the Arabian Oryx (Oryx leucoryx), gazelle (Gazella gazella) and wild goat (Capra aegagrus). Carnivores include the striped hyena (Hyaena hyaena), golden jackal (Canis aureus), red fox (Vulpes vulpes) and caracal (Caracal caracal).79 Lagomorphs and rodents are frequent even today in coastal deserts and plains. Desert reptiles are also frequent, and lizards (Uromastyx sp.) in particular were the most commonly eaten desert animals until historic times.80 The bird fauna of eastern Arabia is diverse. In addition to migrant desert and marine birds, partridges, ducks and geese inhabit the reed beds of marshy depressions. Although now extinct, ostrich

Characteristic plant communities for the central coastal lowlands are the rimth (dominated by Haloxylum salicornicum) saltbush shrubland and the thumam (Panicum turgidum) grass shrubland. Open rimth saltbush shrubland covers major parts of northeast Arabia where it ranges from Iraq in the northeast down into the northern edge of the Rub’ al Khali. It is found on sands in wide topographic lows with groundwater not far below the surface. Less salt tolerant then other halophytes, it is not generally found on the highly saline sabkha margins, but dominates areas surrounding such salt-rich lows at only marginally higher elevations. James and Little 1994: 88. James and Little 1994: 90, table 2. The limit of NaCl for potable water is 1000 mg/l by WHO standards. World Health Organization 1984. 68 James and Little 1994: 92. 69 James and Little 1994: 92. 70 White and Leonhard 1991. 71 Vincent 2008: 174. 72 Mandaville 1990: 8. 73 Mandaville 1990: 23. 66 67

74 75 76 77 78 79 80

12

Mandaville 1990: 28. Mandaville 1990: 107. Barth 1998; 1999. Barth 1998. Harrison 1964: 10. Masry 1997: 29. Masry 1997: 30.

Philipp Drechsler: The Site and its Context (Struthio camelus syriacus) used to be common along the coastal plains and further inland.81 Along the coast and offshore, a great variety of marine species are still of major economic importance.

the development of ephemeral lakes in interdunes91 and endorheic depressions92 suggest higher amounts of (irregular) precipitation in northern and central Arabia contemporaneous with the Arabian Middle Neolithic.93

4 Cultural Context

4.1.2 Settlement and mobility pattern

A series of radiocarbon dates place Dosariyah at the beginning of the fifth millennium cal. BC. The spectrum of material culture discovered at the site shows affinities with the Arabian Middle Neolithic,82 but elements of southern Mesopotamian material culture also occur.

Neolithic groups whose subsistence was primarily based on domesticated sheep, goat and cattle as well as on hunting a broad range of wild animals benefited from these favorable environmental conditions. The distribution pattern of stylistically similar arrowheads94 demonstrates the great mobility of these populations. The presence of similar point shapes in the desert interior, coastal regions and mountain ranges suggests the exploitation of all these regions by the same or closely related groups.95 Further evidence of the mobility of these communities is provided by personal adornments made of marine shell that have been found up to 180  km away from the closest seashore, demonstrating the mobility of the pastoral communities who must regularly have come into contact with coastal communities as part of their annual cycle of migration.96 Additional evidence of mobility comes in the form of exotic goods such as obsidian in southwest Arabia.97

4.1 The Arabian Middle Neolithic The fifth millennium BC represents a flourishing period in prehistoric Arabia. High numbers of archeological sites suggest a rapid growth of human populations, potentially enabled by ameliorating climatic conditions.83 4.1.1 Paleoclimatic background In the southern part of the Arabian Peninsula, a northward shift of the Indian Ocean Monsoon provided higher amounts of precipitation during the midHolocene.84 As a result, dune fields became stabilized and vegetated with C3 grasslands and scatters of woody vegetation.85 The lowlands and desert areas were covered by grassland vegetation with woody elements,86 while grass and scrub vegetation with a higher density of trees predominated in the Yemen highlands.87

The predominance of a highly mobile way of life in Arabia during the Middle Neolithic constrained the development of settlements with fixed dwellings. Places that were occupied permanently or at least for longer periods of time developed exclusively in those regions where diverse or especially resource-rich environments supported a greater degree of sedentism. This process is documented along the shores of the Peninsula, where such settlements relied on marine resources. Along the coast of the Gulf of Oman, settlements affirm this date to the sixth millennium BC,98 therefore suggesting a long tradition of marine exploitation in Arabia. Based on the environmental characteristics of the site locations that provided a year-round food supply, it has been suggested that at least semi-sedentary communities settled in this area.99

In the Levant, an increase in precipitation is evident in a dripstone sequence from Soreq Cave located in the Judean Hills, which suggests higher precipitation with rainfall of 675–950 mm/yr between 8000 and 6000 BC. During the fifth millennium BC, rainfall did not exceed present values.88 While mechanisms of changing climatic conditions during the Late Pleistocene and Holocene are comparatively well understood both in the southern and northern part of Arabia, poorly developed climatic archives in central and eastern Arabia obscure the pattern for these regions.89 The general absence of dripstones in the central and eastern parts of Arabia that date to the Holocene contradicts the notion of a welldeveloped mid-Holocene period of moister climatic conditions induced by a northward shift of the Indian Ocean Monsoon beyond 23°–24° north.90 Nevertheless, 81 82 83 84 85 86 87 88 89 90

In contrast, comparable settlements along the southern shores of the Arabian Gulf barely predate the fifth millennium BC. Rising sea levels during the Holocene led to the flooding of the Arabian Gulf Schulz and Whitney 1986: 181. Garrard et al. 1981. 93 Schulz and Whitney 1986. 94 Drechsler 2009. 95 Spoor 1997. 96 Uerpmann M. et al. 2000; 2006; Beech et al. 2006; Uerpmann M. and Uerpmann H.-P. 2008. 97 Edens and Wilkinson 1998. 98 Uerpmann H.-P. and Uerpmann M. 2003; Biagi and Nisbet 2006; Charpentier 2008. 99 Biagi and Nisbet 2006. 91

Potts 2001a. Drechsler 2007; 2009: 22; Charpentier 2008: 108. Parker and Goudie 2008: 467. Lezine et al. 1998; Neff et al. 2001; Fleitmann et al. 2003. Parker et al. 2004. Schulz and Whitney 1986; Lézine et al. 1998; 2007; Parker et al. 2004. Wilkinson 1997. Bar-Matthews et al. 2003: 3195. Fleitmann et al. 2004: 1. Fleitmann et al. 2004: 20.

92

13

Dosariyah basin,100 preventing the preservation of older sites. In the northern Arabian Gulf, sea level was possibly up to 2.5  m higher during the fifth to the third millennium BC, while thereafter it was relatively stable, close to present-day levels.101 Coastal settlements along the Gulf are often associated with dense accumulations of shell, suggesting that the collection and consumption of shellfish played a major role in subsistence in addition to the catching of fish.102 Frequent finds of both wild and domesticated mammals at these coastal sites indicate the exploitation of a broad spectrum of resources for the daily diet. These data suggest a mixed foraging and food-producing strategy prevailing in Arabia during the Middle Neolithic. Significant populations lived across the Arabian Peninsula, practicing a diverse range of subsistence strategies tied to a variety of well-watered or marine habitats (lakes, lagoons, springs, creeks).

intensive use of the rich oasis environment near the site, potentially linked to cultural contacts with southern Mesopotamia, as indicated by the Ubaid pottery found at Ain Qannas. Besides Ain Qannas, evidence of continuously inhabited settlements in inland eastern Arabia during the fifth millennium BC is scarce. 4.1.3 Subsistence Because conditions for the preservation of animal bones are generally poor in Arabia, the number of archeological sites that provide information on prevailing economic activities is restricted. Until recently archeologists have favored the view that it was predominantly groups of hunter-gatherers who roamed the wide landmass of the Peninsula during the early and mid-Holocene.108 Although rarely explicitly noted, the assumption of hunting and gathering as the predominant way of life is based on the overwhelming number of arrowheads found at countless locations in the desert interior of Arabia.109 The comparatively small number of archeological sites with clear evidence of the presence of domesticated animals supported the view that these bones represent an exotic complement to hunting. Nevertheless, a compilation of archeological sites with bone preservation clearly suggests that domesticated animals represent a common element of the zooarcheological assemblages in Arabia during the mid-Holocene.110 The presence of domesticated sheep, goat and cattle in sixth-millennium BC111 archeological contexts in southern Arabia indicates that the herding of domesticated animals had a long tradition in Arabia, clearly predating the fifth millennium BC. One can plausibly argue that the question of whether the herding of domesticated animals or the hunting of wild animals was of greater importance finds its answer in the diverse environments of Arabia: whenever possible, herding was supplemented by hunting wild animals and vice versa. Along the coasts, both fishing and the collection of shellfish played a major role in the economy.

Evidence of nucleated, long-term settlements associated with the Arabian Middle Neolithic also comes from the Yemen Plateau, where environmental reconstruction suggests woodland vegetation, generally higher water tables and scattered ponds in many upland basins as a result of higher precipitation during the midHolocene.103 Similarly, rich habitats that allowed for a year-round settlement included oases along the coast and in the interior of the Arabian landmass. Oases in the interior that are based on fossil groundwater are situated in three major basins in the cuesta (ridge) landscape of the Arabian Shield in central and eastern Arabia, and in the sediment bodies of the northern Arabian Peninsula.104 The outcrops of the water-bearing formations along the edges of these basins are zones of groundwater recharge during periods of high precipitation, while artesian springs can be found within the centers of the basins. Long intervals between water recharge and discharge which can range up to tens of thousands of years,105 as well as the distance between recharging and discharging areas, make these water sources independent of prevailing climatic conditions. Accordingly, springs are the origins of oases within the flat desert, as is the case with the Al-Qatif and Al-Hasa oases in the Eastern Province of Saudi Arabia.106

4.1.4 Material culture In contrast to the restricted evidence of places that were permanently occupied, there are numerous sites all across the Arabian Peninsula with evidence of an ephemeral occupation that can be related to the Arabian Middle Neolithic. Characterized by scatters of flint artifacts and in some cases spatially associated with fireplaces, such localities can be interpreted as the remains of short-term occupations of highly mobile societies. The ephemeral character of most

The existence of a curving, packed stone and mud wall associated with the Ubaid-related, Middle Neolithic levels 1–4 at Ain Qannas indicates the presence of permanent architecture there.107 This suggests 100 101 102 103 104 105 106 107

Lambeck 1996. Al-Asfour 1978; Sanlaville 1989; Engel and Brückner 2014. Beech 2004. Fedele and Zaccara 2005; Fedele 2008. Burdon 1977. Wushiki 1997. Hötzl and Zötl 1984. Masry 1974.

108 Tosi 1986; Potts 1993: 168; Cleuziou and Tosi 1998: 123; Edens and Wilkinson 1998: 68; Cleuziou et al. 2002: 20; McCorriston et al. 2002: 83. 109 Uerpmann M. et al. 2000: 321. 110 Drechsler 2007. 111 Martin et al. 2009.

14

Philipp Drechsler: The Site and its Context Province of Saudi Arabia, hunting and gathering is attested in sites in the Rub’ al Khali. Although distinguished by these environmental adaptations, the different lithic facies can be seen as formalized due to inter-regional exchange121 or a common cultural background as the result of the development of the Arabian Middle Neolithic out of refugia during a period of climatic deterioration.122

Arabian Middle Neolithic sites had a major effect on the preservation of material culture. Due to the restricted anthropogenic accumulation of settlement debris and prevailing natural erosion of sediment, flint artifacts constitute the most frequent category of finds. Less resistant material such as pottery often shows heavy traces of weathering, while artifacts made from organic material are preserved only at a very few sites. The defining element for Middle Neolithic artifact assemblages is therefore flint artifacts, supplemented by objects made from other stone.

Both technological and typological characteristics of the Middle Neolithic flint industries as well as the spectrum of flint raw material at many sites is characteristic of mobile societies: raw material is often diverse, and even at sites close to raw material outcrops, imported flint occurs regularly.123 Apparently, pieces of flint were carried along for some time until they were discarded or replaced by a better piece. Despite the presence of bifacial chipping and pressure flaking for the production of bifacial implements, the Arabian Middle Neolithic flint artifact production has been described as opportunistic.124

Characteristic elements of the flint artifact assemblages that date approximately to the fifth millennium BC and therefore fall into the later phase of the Arabian Middle Neolithic,112 are bifacial forms with an emphasis on stemmed and usually shouldered points and on narrow foliates, supplemented by scrapers and diverse light and heavy-duty tools.113 While originally defined for the Rub’ al Khali, comparable tool types occur in a much broader area defining an ‘Arabian bifacial lithic tradition’, sensu Edens,114 often referred to as Arabian Bifacial Tradition (ABT). The corresponding primary production is poorly developed, not standardized and highly heterogeneous. It is oriented towards the production of flakes and can best be described as expedient.115 Despite basic technological and typological similarities, Arabian Middle Neolithic flint artifact assemblages reminiscent of the Arabian Bifacial Tradition show some degree of regional variability.116 In the western Rub’ al Khali, the Eastern Province of Saudi Arabia and Qatar, the assemblages are dominated by stemmed bifacial points.117 In contrast, sites in eastern Oman show a clear bifacial component while stemmed points are under-represented.118 Stemmed points also occur in southern Oman, but they are not barbed and tend to show a pronounced triangular cross section. Greater similarities to the Rub’ al Khali assemblages exist in the areas west of the Rub’ al Khali, in Wadi Dawasir and the Asir highlands as well as in the southern Hejaz and central Nejd in the north.119

4.2 Beyond the Ubaid phenomenon: life in the central Gulf area during the Middle Neolithic Evidence of persistent settlement activities comes from a number of Middle Neolithic sites along the shores of the upper and central Gulf, many of which are characterized by architectural remains and a broad spectrum of artifacts (H3/As-Sabiyah,125 Bahra 1/As-Sabiyah,126 Marawah 11127). Although generally associated with artifact assemblages that show strong affinities with the Arabian Middle Neolithic, the occurrence of Ubaid pottery at these places indicates cultural ties to southern Mesopotamia.128 In the lower Gulf towards the Strait of Hormuz, contemporaneous sites are often associated with substantial shell middens, but lack any traces of architecture and are characterized by a more restricted spectrum of material remains. They can therefore be considered to be the remains of ephemeral camp sites129 that were temporarily used by mobile societies during a year-round migration cycle that was well established in southeast Arabia on the Oman peninsula.

Regional variability has been explained by the presence of local populations adapted to different regional environments.120 Beyond differences in the lithic assemblages, these adaptations are also mirrored in differences in the structuring of space on sites, as well as differences in the prevailing economy. While pastoralism predominates in the southern Arabian highlands, the coastal sites of Oman and the Eastern 112 113 114 115 116 117 118 119 120

4.2.1 Archeological evidence The Arabian Middle Neolithic is mostly aceramic. With the exception of about 40 archeological sites that are predominantly located along the shores of the Arabian

Charpentier 2008. Kapel 1967; Tixier 1980; Edens 1982. Edens 1982: 120. Nelson 1991: 64. Spoor 1997. Edens 1988: 33, Bergne and Copeland 1976; Uerpmann H.-P. et al. 2013. Edens 1988: 34. Edens 1988: 35; see Spoor 1997.

121 122 123 124 125 126 127 128 129

15

Edens 1988: 37. Drechsler 2009; Preston et al. 2015. Uerpmann M. et al. 2008; Drechsler 2010. Kallweit and Davies 2010: 105. Carter and Crawford 2010a. Bieliński 2011; 2012. Beech et al. 2005. Oates J. et al. 1977; Roaf and Galbraith 1994. Uerpmann M. and Uerpmann H.-P. 1996.

Dosariyah

Figure 1.8. Sites with Ubaid pottery along the shores of the Arabian Gulf (open circles) and major contemporary sites in southern Mesopotamia and Iran. Shoreline reconstruction for the upper Gulf between c.5300 and 4550 BC (after Pournelle 2003: 123, fig. 44).

Gulf (Figure 1.8), pottery vessels or fired clay objects are known neither from occupation nor from burial or ritual contexts. Only at sites close to the Arabian coast of the Gulf and the Al-Hasa oasis, does pottery occur in the form of painted buff ware and coarse ware. Both vessel shape and decorative style and the chemical composition of the clay suggest an origin of the painted Black-on-Buff ware in southern Mesopotamia and possibly Iran.130 While the underlying mechanisms of this distribution are unknown, trade or exchange are often invoked as driving forces.131 The frequency of Ubaid pottery at eastern Arabian sites seems to decrease as the distance from southern Mesopotamia increases.132

It occurs in high quantities at Middle Neolithic sites in Kuwait (H3/As Sabiyah, Bahra 1/As-Sabiyah) and in the central Gulf (Abu Khamis, Dosariyah), but was found in significantly lower numbers at several other coastal sites between Ras al-Khaimah and the Qatar peninsula.133 Stylistically, ceramics from these sites fall within the Ubaid 3 to Ubaid 5 phases,134 with most assemblages falling in the Ubaid 3 period. Prominent exceptions are H3/As-Sabiyah and Bahra 1/As-Sabiyah, both located in the As-Sabiyah region of present-day Kuwait, where an Ubaid 2 or early Ubaid 3 (Ubaid 2/3) component was documented.135 Therefore one can argue that the sites with Ubaid pottery and therefore the contacts with southern Mesopotamia did not last

Bibby 1970; Oates J. et al. 1977; Roaf and Galbraith 1994; Beech et al. 2005: 46; Carter and Crawford 2010b: 206. 131 Oates J. et al. 1977; Uerpmann M. and Uerpmann H.-P. 1996; Carter and Crawford 2010b; Drechsler 2011. 132 A small quantity of painted buff ware from sites along the coast of eastern Arabia shows painted decorations that are not commonly associated with Mesopotamian decorative styles. The term ‘Ubaid pottery’, however, is well established for finds of painted buff ware in the Arabian Gulf. In this volume ‘Ubaid pottery’ is used in the latter 130

sense as a general term for painted (and unpainted) buff wares found at fifth-millennium BC Arabian Middle Neolithic sites, to avoid the somewhat cumbersome and technical terms ‘painted buff ware’ or ‘Black-on-Buff ware’. 133 Boucharlat et al. 1991; Vogt 1994; Uerpmann M. and Uerpmann H.P. 1996; Beech et al. 2000; Phillips 2002. 134 Oates J. 1993. 135 Carter 2010b; Smogorzewska 2013: 556.

16

Philipp Drechsler: The Site and its Context longer than a millennium.136 The location of most sites with Ubaid pottery along the raised shoreline of the Gulf during the fifth millennium BC suggests that contact between southern Mesopotamia and the Gulf was established by seafaring. Direct evidence of boat building was discovered at H3/As-Sabiyah in Kuwait where finds of bitumen fragments with reed impressions and barnacles were interpreted as boat remains.137 The painting of a boat on a clay disc also discovered at H3/As-Sabiyah proposes the use of the sail in the Gulf at this time.138

investigations at Ain Qannas, Dosariyah and Abu Khamis, Masry suggests a common cultural sphere with intense overland contacts or even an origin of the Ubaid ‘culture’ in eastern Arabia.141 Excavations at Oueili and the discovery of predecessors of the Ubaid pottery from the central Gulf seized this interpretation.142 Early levels, designated as Ubaid  0,143 contain pottery and bricks reminiscent of those from cultural traditions in northern and central Mesopotamia. Subsequent research therefore favored the view of maritime interaction144 and looked for causes and potential goods for exchange145 without investigating the fundamental social and economic systems. But is it plausible to argue that communities in southern Mesopotamia, still in the process of state formation, sent out trade expeditions down the Gulf to obtain prestigious goods (i.e. natural pearls)?146 Or is it likewise reasonable to assume that indigenous eastern Arabian societies who were intimate with seafaring and the exploitation of marine resources sailed in their vessels towards the northwest to obtain pottery?

The role of Ubaid pottery and also of Coarse Ware within the Neolithic societies is still not clear. The spectrum of vessel forms suggests both daily domestic activities (storage jars, cooking pots) and representative items (highly decorated open bowls). Nevertheless, the supply of pottery was marginal at most sites, almost excluding a regular use in daily life. The fact that Ubaid pottery was valuable goods is confirmed by the frequent mending of broken vessels and the reuse of sherds of Ubaid pottery as pendants (see Chapters 7, 9).

States with well-developed urban centers and administrative hierarchies potentially able to organize and sustain a trading network did not appear in southern Mesopotamia before the Late Uruk period in c.3000 BC.147 Although similarities of architecture and painted pottery styles in southern Mesopotamia from the Early Ubaid through the Late Uruk periods suggest a certain degree of demographic and cultural continuity, it is debatable whether later mechanisms of formalized exchange (or trade) can be extrapolated from the more distant past. By the beginning of the Ubaid period (Ubaid 0–1), small villages and towns were common across Mesopotamia. Much of this region was linked through social networks, and similarities in artifacts indicate a widespread exchange of goods and knowledge.148 At this time, even the larger communities such as Eridu, Ur or Tell al-Ubaid were small, averaging about 1  ha, with estimated populations not exceeding 1000 people. As these small communities were widely dispersed and lacked the linear distribution typical of settlements dependent on irrigation canals,149 the practice of irrigation agriculture in southern Mesopotamia at that time is questionable. During the Middle Ubaid (Ubaid 2–3), a two-tiered settlement system appeared that often marks the emergence of hierarchically organized, non-state societies.150 Important centers such as Eridu

With the exception of Ain Qannas, located in the northern part of the Al-Hasa oasis in the Eastern Province of Saudi Arabia about 60  km inland, Ubaid pottery was scarcely taken into the interior as the majority of ‘inland’ sites in the Eastern Province are located at the margins of sabkhas, most likely flooded during the mid-Holocene.139 In contrast to Ubaid pottery, the occurrence of Arabian Coarse Ware is restricted to the upper and central Gulf where it occurs in lower quantities in comparison to — and always together with — Ubaid pottery. Originally designated as ‘simple cooking ware’ the broad spectrum of vessel shapes, as well as the presence of decorated vessels,140 suggests a more diverse use. Although the places of production of this pottery are unknown, the spatial distribution pattern of this kind of pottery suggests a local, Arabian origin. This kind of pottery therefore represents a characteristic element of the Middle Neolithic coastal sites in eastern Arabia. 4.2.2 Trading networks, exchange patterns or kin relationships? The presence of foreign pottery at Middle Holocene sites along the southern shores of the Gulf which originates from different manufacturing locations in southern Mesopotamia suggests a well-established and stable social and/or exchange network during the fifth millennium BC. The underlying mechanisms are, however, poorly understood. Based on his 136 137 138 139 140

Masry 1997: 123. Huot 1989. 143 Oates J. 2004. 144 Oates J. et al. 1977; Carter 2006; Carter and Crawford 2010b. 145 Uerpmann M. and Uerpmann H.-P. 1996. 146 Oates J. et al. 1977: 233; Uerpmann M. and Uerpmann H.-P. 1996: 135. 147 Adams 1981; Nissen 1988; 2001; Pollock 1999; Yoffee 2005. 148 Kennet D.J. and Kennet J.P. 2006. 149 Adams 1981: 59. 150 Wright 1981; Stein 1994. 141 142

Rose 2010: 864. Carter 2006; Carter and Crawford 2010b. Carter 2010d: 91. R. Carter, personal communication. Kainert and Drechsler 2014.

17

Dosariyah grew up to 10  ha, with estimated populations of 2000 to 3000.151 During this period, Ubaid pottery appears in Middle Neolithic settlements along the southern coast of the Arabian Peninsula for the first time. Although an economic and political differentiation is indicated by this hierarchical distribution of settlements in Mesopotamia, little evidence exists to support elite control of long-distance exchange systems and centralized control of high-status craft production.152 Instead, Ubaid-period societies were centered on the temple complex, and ideology appears to have played an important role in organizing these communities.153 Only by the end of the Ubaid period were some communities substantially larger than their neighbors, ruled by hereditary leaders and administered by institutionalized administrative organizations.154 But it was not before the late Uruk period that ‘colonists’ and/or ‘merchants’ from southern Mesopotamia were able to gain access to critical material and wealth objects and that the movement of goods to southern cities was facilitated by outposts strategically located in areas containing valuable resources, affording control over the distribution of these materials.155

livestock,160 dried fish,161 leather and textiles, access to water and a peaceful passage,162 or marriage partners to encourage a multi-directional network of family connections.163 Although plausible to some degree, all these suggestions are difficult to verify based on the archeological record. The question remains open as to what goods or commodities were so important in southern Mesopotamian societies that Mesopotamian seafarers sailed down the Gulf. While the problem of goods provided for exchange remains, at least the question of impetus for this exchange can be broached if one considers the local Arabian populations living along the shores of the Gulf as the active agents in this kind of exchange; it is not only from mending holes that we know that Ubaid pottery was considered valuable. At Dalma 11, located in the lower Gulf and well beyond the area of the central Gulf that had easier access to southern Mesopotamian pottery, Ubaid pottery was imitated in painted plaster vessels.164 Clearly, there was a certain need for painted pottery along the Arabian shores of the Gulf but again, the specific character of this need is difficult to demonstrate. Carter suggests that the Ubaid pottery found in Arabian Neolithic contexts was oriented ‘towards serving and display, both of the ceramics themselves and of the food served.’ He further proposes that ‘Ubaid pottery was not only used to present food, but was also redistributed in acts of ceremonial giftgiving or exchange at communal events, perhaps in feasting contexts’.165

Even if the fifth-millennium BC southern Mesopotamian socio-economic environment did not support regular trade it remains a valid hypothesis that southern Mesopotamian seafarers sporadically traveled down the Gulf to obtain some kind of local products not available in their homeland in exchange for pottery. But what would be appropriate and beneficial targets of this exchange? Copper, the most important export of eastern Arabia during the third millennium BC, can be excluded, as there is no evidence of copper mining during this period. In addition, eastern Arabia in the central Gulf region lacks any copper sources. One possibility that has been repeatedly mentioned is that Ubaid pottery was exchanged for pearls.156 Although pearl fishing in the Gulf clearly dates back to the fifth or even sixth millennium BC,157 the use of pearls is most convincingly attested at that time for southeast Arabian societies who used them to adorn their dead.158 In addition, mother-of-pearl was used for the manufacture of adornment and fishing hooks by local coastal populations.159 In addition to pearls and shell, the spectrum of potential goods for exchange that has been suggested by different researchers includes

It is the idea of this feasting context that is further developed by Peter Magee.166 Following Ibn Khaldun, Magee introduces the idea of assabiya, social mechanisms that derived from shared experiences of desert life, as one of the formative elements of Arabian (pre)history.167 According to Ibn Khaldun, assabiya is antithetical to an urban existence based on royalty, wealth and luxury, but emphasizes kin relationships that allow the overcoming of social or environmental challenges. It can be defined as ‘a corporate spirit oriented towards obtaining and keeping power’ which was ‘held together by the sense of common ancestry, whether real or fictitious’.168 Was the presence of Ubaid pottery in eastern Arabia an integral part of ceremonies that center on social cohesion in this region? During the subsequent fourth millennium Kallweit 2003: 63. Uerpmann M. and Uerpmann H.-P. 1996: 139, n. 70. Carter and Crawford 2010b: 209. 163 Carter and Crawford 2010b: 210. 164 Beech et al. 2000; http://www.adias-uae.com/plaster.html, accessed 17/09/2014. 165 Carter 2006: 60. 166 Magee 2014. 167 Magee 2014: 11. 168 Hourani 1992: 2 after Magee 2014: 12.

Adams 1981. Stein 1994. 153 Hole 1994: 139; Stein 1994. 154 Kennet and Kennet 2006: 83. 155 Algaze 2001; Rothman 2001. 156 Oates J. et al. 1977; Uerpmann M. and Uerpmann H.-P. 1996; Carter and Crawford 2001; Phillips 2002; Carter 2006; Charpentier et al. 2012. 157 Charpentier et al. 2012. 158 Kiesewetter et al. 2000; Beauclair 2005; 2008. 159 Beech and Elders 1999; Méry et al. 2009.

160

151

161

152

162

18

Philipp Drechsler: The Site and its Context precipitation was reduced in the interior. According to Rose, arid conditions during MIS 2 forced nomadic groups to contract toward the Arabian Gulf basin.178 In contrast, ameliorating climatic conditions in Arabia from the onset of the Holocene onwards, together with the contemporaneous postglacial flooding of the Gulf, drew populations into the hinterland and thus led to a permanent inland shift of the settlements in accordance with the transgressive sea. It has already been argued that the formation of rich coastal zones along the northern shorelines of the Gulf — in southern Mesopotamia — promoted the development of Ubaid communities and played a critical role in the process of state formation.179

BC, such social strategies are well attested. While a mound of dugong bones at Akab, Umm al-Quwain, indicates mass slaughter and consumption, potentially in the context of totemism,169 an oval arrangement of forty cattle skulls that were pushed into the sediment at Kheshiya, Wadi Sana, Yemen,170 was associated with a collective event that ‘commemorated the convergence of a social group or several social groups whose practice of ritual sacrifice emphasized their community ties’.171 Another fundamental characteristic of Arabian societies is the high degree of mobility in view of a diverse spatial distribution of environmental resources. This high degree of mobility was often accompanied — and is traceable through — the displacement of material culture. In addition to an exchange of material goods and ideas, ‘regularized interaction between groups would also have provided an opportunity for the alignment of ideological and social systems across the entire peninsula’.172 The distribution of soft stone, obsidian, worked stone and shell173 that were found in distances up to 180 km from their original sources174 suggests that this kind of mobility was already established in the fifth millennium BC, be it in the form of intense interaction between different population groups or actual movements of people.175 It is a fascinating and peculiar characteristic of the spatial distribution of Ubaid pottery that — in contrast to many other aspects of material culture — it scarcely found its way inland, but it almost exclusively occurs along the (former) coastline. Although considered a valuable item at coastal sites, it was not distributed further inland. No explanations can be given at present, but it is reasonable to assume that some kind of social mechanisms/restrictions hindered an effective redistribution.176

But what were the consequences of these changing environmental conditions for the populations in eastern Arabia? First of all, an almost dramatic increase in the number of archeological sites can be observed. With the exception of Qatar, sites that are (often only tentatively) dated to the seventh and sixth millennia BC are very rare in eastern Arabia. In contrast, more than 60 sites are known today along the shores of the Gulf that date to the fifth millennium BC, thus coinciding with the final phase of the marine incursion. Further on, the settlements along the shores of the central and lower Gulf seem to lack local ancestors predating the Arabian Middle Neolithic/Ubaid 3 period. The only exception might be Ain Qannas, where Ubaid pottery was found that can be associated with Ubaid 2. The fact that Ain Qannas is located about 60  km inland and on the edge of a well-watered oasis suggests a different mechanism for the establishment of this site. It is therefore conceivable that the present picture of the spatial distribution of sites along the shores of the Gulf is the result of the history of its flooding, with the majority of Middle Neolithic sites located along or close to the shoreline of the mid-Holocene high stand. Contemporaneous with the incursion of the Gulf, populations and their settlements were pushed outwards onto dry land. Such a mechanism would have one further important implication. During the time when the Arabian Gulf basin served as a refugium it might have been occupied by a single population with a common cultural background, consecutively driven apart by the transgression. In that case, did the imports of Ubaid pottery function as a symbol for a common ancestry, as an assurance of shared identity? The early Ubaid settlements in southern Mesopotamia were located on slight rises (‘turtle backs’) within aquatic habitats in the wetlands at the head of the Arabian Gulf.180 Such locations were at the interface between fresh- and saltwater and were optimal for freshwater

One interesting aspect in the discussion about the distribution of Ubaid pottery along the Gulf coast comes from Rose’s hypothesis of a population refuge in the Arabian Gulf basin (‘Arabo-Persian Gulf Oasis’) that became submerged by rising sea levels until the mid-Holocene.177 Fed by both allochthonous rivers and groundwater aquifers, favorable environmental conditions are suggested to have existed at times when Méry et al. 2009. McCorriston and Martin 2009; McCorriston 2011: 100. 171 McCorriston and Martin 2009: 246. 172 Magee 2014: 62. 173 Edens 1988. 174 Beech et al. 2006: 24. 175 See Uerpmann M. et al. 2000 for year-round migration cycles in southeastern Arabia. 176 The use of pottery vessels is widely restricted to sedentary communities, while mobile societies often replace pottery by other, less breakable, materials. It is therefore plausible that at first sight the pots did not travel further inland due to the risk of breakage during transport. Nevertheless, if we consider Ubaid pottery in an Arabian Neolithic social context as a valuable or exotic item rather than a functional one, even single — especially painted — potsherds would have kept their essential value. 177 Rose 2010. 169 170

178 179 180

19

Rose 2010: 854. Kennet D.J. and Kennet J.P. 2006. Pournelle 2003.

Dosariyah accessibility, hunting and fishing, and transportation.181 The settlements in southern Mesopotamia therefore closely mirror the characteristics of the majority of Middle Neolithic settlements along the southern shores of the Gulf, both in their locations on islands or very close to the former sea lines and in characteristic adaptations to marine resources. Are these similarities simply the result of congeneric adaptations to similar environmental conditions? Or are we faced with

181

developments that find their basis in a common cultural context deeply rooted in both southern Mesopotamian and coastal Arabian societies? The study of material remains excavated at Dosariyah unveiled in the subsequent chapters will try to find answers to these questions, ultimately contributing to a growing body of knowledge about a fascinating facet of the prehistory of the Arabian Peninsula.

Oates J. 1960.

20

Chapter 2 Geomorphology, Geoarcheology and Paleoenvironments Adrian G. Parker, Simon J. Armitage, Max Engel, Mike W. Morley, Ash Parton, Gareth W. Preston and Hannah Russ 1 Introduction

which are orientated parallel to the modern coast and are up to 15 m asl (above mean sea level). The coastal facing side of the ridges is exposed to intensive eolian erosion with exposed eolianite at the surface, while on the landward side younger, secondary dunes have formed on the leeward side of these ridges from the reworking of the primary eolian bedforms (Figure 2.1).

Environmental change during the Holocene in eastern Saudi Arabia is poorly understood. Few detailed records have been examined to date, with limited evidence available from dunes, lakes and sea-level records. While the geomorphological setting of the Jubail region has been described in detail by Barth,1 the chronology for the development of this landscape is largely unknown. Lacustrine deposits from the Al Sulb Plateau,2 Al-Hasa,3 Nafud As Sirr4 and the southern playa of Bahrain5 record wetter conditions during the Early to midHolocene. These sites have yielded radiocarbon dates between 7900 and 4700 cal.  BC (8300–5800 14C  BP) but no detailed physical, chemical or biological analyses have been conducted. Therefore, the environmental backdrop against which the occupation at Dosariyah can be set is limited.

To the south of Dosariyah, Hussain refers to deposits of Dammam Eolianite.6 This eolianite is dominated by carbonate material, comprising mostly ooliths, rounded skeletal fragments (coral, shell debris) and peloids, cemented by calcite, aragonite and gypsum. The cemented dunes comprise pale brown-buff calcareous grainstone with over 85% carbonate grains. The Dammam Eolianite is characterized by well-sorted, fine to medium sand-size particles. Glennie noted that carbonate sand grains are eroded by wind action 2–4 times as fast as quartz grains and thus carbonate allochems in eolianite are usually fairly well rounded.7 Outcrops of Dammam Eolianite are exposed at Dosariyah at the surface along the coastal strip (Figure 2.1). They form a resistant cap (Figure 2.2), which in places has been faceted by wind action to form yardangs (Figure 2.3) or is covered by younger dunes. Goudie notes that yardangs are developed by unidirectional or narrow bimodal wind directions and are often found in areas associated with barchans.8 In exposed sections, foreset strata dip at varying angles between 9° and 21° while individual laminae are typically 1–3  mm thick, giving rise to well-developed pin-stripe lamination. These laminae are commonly formed by coarser grain flows of larger, well-rounded carbonate grains (ooliths, skeletal material and peloids) interbedded with finer-grained carbonate material. Further inland from the coast, the material contains an increasing amount of quartz and other siliciclastic material.

An area approximately 3 x 2 km was mapped in detail (Figure 2.1) to record the geomorphology of the site and its surroundings through field observations, logging of natural and artificial quarry sections, and test pit evaluation to record sub-surface stratigraphy and collect samples for paleoenvironmental analyses. Geomorphological and paleoenvironmental research in the Dosariyah area was conducted over 10 days of field investigation, during which parts of the wider region were also visited. The topography in and around the site is generally low lying, comprising small bedrock outcrops up to 20 m high, eolianites, dune sands, sand sheets, sabkhas and coastal spits, bars and barrier complexes. 2 Geomorphology Along much of the southern coast of the Arabian Gulf extensive outcrops of cemented eolian sands occur. These are the remnants of deflated longitudinal dunes with their axis trending in a north-northwest–southsoutheast direction. At Dosariyah two remnant ridges of these deflated eolian sands were mapped, both of 1 2 3 4 5

Sections through these cemented dune ridges show they are formed from more than one generation of eolian deposits. Underlying the cemented Damman Eolianite cap rock several exposures revealed soft, uncemented, well-sorted medium sands comprising ~95% quartz, and less than 1% ooids. Key exposures

Barth 1998; 2002. Jado and Zötl 1984. Larsen 1983. Schulz and Whitney 1986. Doornkamp et al. 1980.

6 7 8

21

Hussain 2006. Glennie 1970. Goudie 2007.

Dosariyah

Figure 2.1. Geomorphological map of the Dosariyah area (geomorphological mapping: A. Parker and M. Morley).

Figure 2.2. Carbonaterich Dammam Eolianite capping quartz-rich uncemented sand in coastal cliff exposure, Dosariyah. The lower weakly cemented sand unit underlying the carbonate-rich cemented cap unit was OSL dated to 6.8 ± 0.5 ka (4.8 ± 0.5 ka BC) (photograph: A. Parker).

22

Adrian G. Parker et al.: Geomorphology, Geoarcheology and Paleoenvironments were noted in a disused quarry section, where up to 8 m of the deposits were exposed beneath a 1 m thick Damman Eolianite cap (N26°54’52.1” E049°44’21.5”), and also in a coastal cliff section (Figure 2.2) under a 0.5 m cap (N26°55’34.3” E049°44’30.8”).

low-lying hills and ridges of the Upper Tertiary Rus, Dammam and Hofuf formations, and Quaternary carbonate eolianite formations. At Dosariyah mesoscale yardangs were mapped adjacent to the coast, developed in carbonate-rich eolianites. The yardangs here are up to 1.5 m high, 3 m wide and up to 15 m in length (Figure 2.3).

In the United Arab Emirates carbonate cemented sands, mapped as ‘miliolite’ by Huntings9 as they contained many miliolid foraminifera, equate to the Damman Eolianite observed in the eastern coastal region of Saudi Arabia. More recently, however, they have been termed the Ghayathi Formation10 after a type section at Jabal Marban in Abu Dhabi. The Ghayathi Formation consists of carbonate-dominated paleo-dune sandstones, which form a series of semi-continuous outcrops across the region. These generally occur as extensive deflated sheets occupying low-lying areas. Small isolated outcrops of miliolite draping the older orange quartzrich sands, which are synonymous with the Madinat Zayed Formation of the UAE,11 occur locally.

The dominant orientation of the yardangs is to the south-southeast and roughly parallel to the presentday prevailing wind patterns along the Arabian Gulf coastline. Yardangs faceted into Hadrukh bedrock were also noted in the vicinity of Jebel Bari (N26°54’26.4” E049°29’59.2”). To the west and northwest of the study site, barchanoid ridges and barchan dunes overlying sabkha deposits are also found. These dunes have migrated across the sabkhas in the same prevailing wind direction.14 Elsewhere in the Arabian Gulf region meso-scale yardangs up to 10 m high have developed on resistant dolomitic Eocene limestones (Rus Formation) in the southwest corner of the main island of Bahrain’s central depression, while yardangs up to 4–6 m high are also found on eroded eolianites on Jiddah Island, off the west coast of Bahrain.15

Low dunes, reworked and emplaced during the Holocene, are found superimposed on the main eolianite dunes. They are most conspicuous on the north-northwest– south-southeast trended deflated linear dune closest to the coast. Here they top the summits of the high ground and are draped over the eolianite surface on the landward side. In places these dunes are up to 3 m above the eolianite surface and are whitish in color, having largely been reworked from the Dammam Formation. They trend north-northeast–south-southwest, with several wind gaps between the white dune areas. These dunes are covered in scrubby vegetation. Areas of cross-bedded and planar-bedded grainstone occur locally between areas of low dune or sand veneer. These are exposed as corrugated pavements, which are flat in nature due to deflation by wind, which has exposed the bedding planes. Good examples occur at N26°55’11.0” E049°44’42.0” and N26°55’13.6” E049°44’35.5”. Within the wind gaps and on the dune flanks of the sabkha areas at Dosariyah extensive areas of low, small scrubby nebkha dunes up to 1 m high occur.

3 Coastal Processes The coastline running south of Jubail to the desalination plant, south of Dosariyah, is dominated by a series of complex coastal barriers, spits, lagoons and sabkhas. The coastal configuration comprises several headlands and embayments, which are heavily influenced by coastal processes, especially tidal patterns, wave refraction and longshore drift. Tidal patterns within the Arabian Gulf are complex with regional variations in tidal range,16 including two amphidromic points that have been noted where there is no tidal range. The first is located off the northern Saudi Arabian coast to the north of Dosariyah, and the second is located off the western Abu Dhabi coast (UAE). Tidal ranges increase away from these nodes. Over most of the Gulf the tidal range offshore is  2 cm. Despite the absence of clear contexts such as architectural remains or other installations, this form of recording allows a detailed sherd-bysherd analysis and interpretation of the assemblage recovered from the site. Corresponding to this indepth field documentation it was decided to document every fragment separately. To this purpose, I created a Microsoft Access database, which gave the opportunity to record a variety of (identifiable) characteristics of each piece, including weight, thickness, diameter, fabric, form, color, general remarks, etc. Moreover, refitting studies were carried out and this was considered especially useful since the entire assemblage is highly fragmented with only a handful of preserved complete profiles and without any complete or near-complete vessels.

Table 7.1. Sherd count of all documented pottery sherds recovered from the site which form the basis for further data analysis. DARP Excavation, piece-plotted

DARP Excavation, from collapsed profiles

DARP Excavation, collected from overburden

5171 844

13

DARP Systematic Surface Survey

6025

Masry’s Collection, partly re-examined

2733

DARP Surface, unsystematically collected Burkholder’s Collection, partly re-examined Total

245

3 Classification of Macroscopic Wares

379

In fall 2011, I was able to examine the first part of the pottery. At this point, the goal was to establish a preliminary macroscopic classification of present wares obtained from excavation. For this purpose, the material excavated in trench S1 was chosen. With a count of 238 sherds from well-stratified contexts, the collection seemed suitable for this task. Over the course of the following three seasons, the classification was continuously refined, finally leading to the identification of five main ware groups divided into 16 different sub-ware groups for all 6028 fragments from excavated contexts. Due to the exclusion of sherds collected from collapsed profiles and the overburden, only 5171 piece-plotted sherds are included in the

15410

2 Methodology The data presented here were collected during five field seasons between spring 2010 and spring 2012 and documented during four stays in Jubail and Dammam between fall 2011 and fall 2013. The documentation was carried out in the Dammam Regional Museum, in the Gulf Flower Hotel in Dammam and in the al-Farhan Hotel in Jubail.

183

Dosariyah following statistical evaluations (Table 7.2), unless otherwise stated.

heterogeneous. Overall, seven different subgroups were distinguished although the transition between each is not well defined. A common characteristic is reddish to brownish color, sometimes with a very dark brownish to blackish core. The subgroups mainly differ in terms of presence/absence, quantity and size of organic and mineral inclusions. Most common are wares 1a–d that account for 92.6% of the piece-plotted CW collection. Much less common but still clearly distinguishable from other subgroups are wares 1e–g which make up the remaining 7.4%.

The differentiation of main wares and sub-wares is primarily based on the description of the sherd fabrics. The main criteria for this were temper and texture.1 Additionally, in order to provide a comprehensive description of main wares and sub-wares, color, surface treatment and decoration were added. The most common type of pottery in Dosariyah is Black-on-Buff Ware (BoBW) which accounts for 81.7% (N  =  4223) of the piece-plotted assemblage. A total of four subgroups were defined, which basically differ from each other in terms of presence/absence, quantity and size of organic and mineral inclusions. These four subgroups reflect the high degree of homogeneity of the Black-on-Buff Ware, which is in clear contrast to the much smaller Coarse Ware (CW) assemblage that shows a considerably higher macroscopic diversity. Although the color of Black-on-Buff Ware is usually, as the name suggests, some form of buff, the actual range varies from light reddish or light brownish to different nuances of green. This type of pottery is widely known in the southern regions of Mesopotamia (here also referred to as Ubaid Ware) and southwestern Iran (here also referred to as Susiana Ware within the Susiana Plain in Khuzestan or Bakun Ware in Fars).

Coarse Ware is identical with what was first identified and described by Burkholder in 1968.5 She listed seven sites in the Eastern Province of Saudi Arabia where she was able to document this particular type of ware.6 Here, Coarse Ware exclusively occurs in combination with Black-on-Buff Ware, which functions as a relative chronological marker due to its striking formal characteristics. The rather simple appearance and unspecific characteristics of Coarse Ware, however, do not allow it to be closely dated when found out of context or without any other indicative markers such as BoBW or Middle Neolithic flint artifacts. Sites only with Coarse Ware have not yet been identified. A third type of main ware has been entitled Coarse Mineral Ware (CMW). Its proportion only amounts to 0.5% (N  =  26) of the whole excavated and pieceplotted ceramic assemblage. Owing to its rather rough appearance and a predominance of simple vessel forms (which are identifiable only in a very few cases), it cannot be completely excluded that it actually represents a form of Coarse Ware. One clear difference from CW, however, is the large amount of coarse-sized mineral inclusions. Although only represented by 26 sherds, three subgroups have been defined which show the heterogeneity of this kind of pottery. Subgroup 3c has been associated with just one sherd which comes from a collapsed profile in area E1.

Since its first identification in the Eastern Province of Saudi Arabia by Grace Burkholder in 1968,2 this type of pottery has now been identified unambiguously in at least 38 sites in Kuwait, Bahrain, Qatar and the United Arab Emirates (Figure 1.9).3 Chemical studies conducted in 19774 with sherds of BoBW as well as of CW from Saudi Arabia, Bahrain and Qatar supported the assumption that the origin of the analyzed BoBW material is located somewhere in southern Mesopotamia. Further chemical analyses conducted by Peter Magee in 2011 with 34 BoBW samples and nine CW samples from Dosariyah confirmed those results (see Chapter 8). That study also substantiates the presumption that the chemical composition of BoBW and CW differs drastically, strongly indicating that the origin of the clays used for manufacture should be sought in separate regions.

A fourth main ware group which was defined for a handful of potsherds is characterized by remains of small but macroscopically visible bone fragments within the fabric; this group has therefore been termed Soft Bone Tempered Ware (SBTW). It accounts for less than 0.1% (N = 3) of the piece-plotted sherds.

Totaling about 17.8% (N = 918), the Coarse Ware (CW) represents the second largest group of the excavated and piece-plotted assemblage and is also the most

The fifth main ware is the rarest with a percentage of less than 0.1% (N = 1) and is termed Organic Buff Ware (OBW). The single sherd looks very similar to Blackon-Buff Ware at first sight as it has a buff color with remains of dark monochrome painted decoration, yet it possesses a high amount of organic temper and no visible mineral inclusions. The macroscopic fabric is

The criteria for the definition of wares are based on Schneider et al. 1989 but were adjusted to fit the conditions of the pottery recovered from the site. 2 Burkholder and Golding 1971; Burkholder 1972; Bibby 1980. 3 Roaf 1976; Oates J. 1978; Haerinck 1991; Hermansen 1993; Jasim 1996; Uerpmann M. and Uerpmann H.-P. 1996; Masry 1997; Beech 2005; Beech et al. 2005; Charpentier and Méry 2008; Carter 2010b; alNaimi et al. 2011; Smogorzewska 2013; 2016; Drechsler 2014. 4 Oates J. et al. 1977; Roaf and Galbraith 1994. 1

5 6

184

Burkholder 1972. Potts 1990: 59–61.

Christine Kainert: The Pottery therefore clearly different, justifying its definition as a separate ware.

sherd count as well as their general appearance, it seems plausible that CMW and SBTW both represent forms of individual and local products which were manufactured either at the site or close by and according to need. Due to some shared characteristics with Coarse Ware, both are considered as belonging to the extended family of Coarse Ware. In contrast, the origin of OBW is more puzzling. Some formal similarities, namely the use of a dark monochrome paint, make an origin in Mesopotamia or adjacent regions the most likely possibility; the high amount of organic temper and lack of visible mineral inclusions are a strong contrast

The occurrence of Coarse Mineral Ware, Soft Bone Tempered Ware, and Organic Buff Ware at Dosariyah alongside the Black-on-Buff Ware and Coarse Ware (which together constitute the vast majority of the collection), signals the former heterogeneity of the site’s pottery, even though those three wares only occur in tiny quantities. None of these three other wares has yet been identified at other contemporary sites in the Gulf. This is not surprising since, based on the very low

Table 7.2. List of all macroscopically defined main ware groups and sub-ware groups and their respective sherd count and percentage of the piece-plotted collection. Name

Code

Description

1a

This handmade and very coarse chaff-tempered ware shows no visible mineral inclusions. The numerous chaff inclusions (medium to coarse size in medium to high quantities) are clearly visible in the form of relatively large elongated N = 179 or rounded pores. The texture is hard to slightly porous. The inner and outer (3.4 %) surfaces usually appear in reddish to brownish color with a dark grayish to dark brownish core.

1b

1c Coarse Ware, CW 1d

1e

1f

1g

Black-on-Buff (Standard) Ware, BoBW

2a

The characteristics of this handmade ware correspond to 1a. But mineral inclusions (fine to coarse size in low to medium quantities) in the form of N = 231 small angular creamish to reddish to darkish inclusions are visible as well. The (4.5 %) texture is hard to slightly porous. The inner and outer surfaces usually appear in reddish to brownish color with a dark grayish to dark brownish core. Compared to Ware 1a/b this handmade ware is a relatively fine chaff-tempered ware without any visible mineral inclusions. The chaff inclusions (fine to medium size in low to medium quantities) are quite small and in very low cases hardly visible. The texture is hard to very hard and, in some cases, slightly porous. The inner and outer surface usually appears in reddish to brownish color with a usually dark grayish to dark brownish core. The characteristics of this handmade ware correspond to 1c. But mineral inclusions (fine to medium size in low to medium quantities) in the form of angular lightish to grayish inclusions are visible as well. The texture is medium to hard and in some cases slightly porous. The inner and outer surface usually appears in reddish to brownish color with a usually dark grayish to dark brownish core.

N = 180 (3.5 %)

N = 260 (5.0 %)

The characteristics of this handmade ware correspond to 1c. But inclusions of reddish clay lumps (medium to coarse size in low to medium quantities) are N = 14 visible as well. The texture is hard and the color appears relatively homogene- (0.3 %) ous in light brownish to creamy.

The characteristics of this handmade ware correspond to 1a. But inclusions of white lime (medium to coarse size in medium quantities) are visible as well. N = 15 The texture is rather tight. The color appears reddish-brown on the inner and (0.3 %) outer surface with a dark grayish core.

This handmade ware shows only very low fine inclusions of chaff temper and N = 39 no mineral temper. The texture is hard and the color appears relatively homo(0.8 %) geneous from light reddish brown to light creamish brown. This partly handmade, partly slow-wheel-made and mineral-tempered ware (medium to coarse size in medium quantities) shows no visible chaff inclusions. The abundant angular mineral inclusions occur in reddish to buffish to darkish color. The texture is hard and quite compact. The color ranges from light creamy to buffish to light reddish to greenish. Only some pieces of this ware show remains of a monochrome dark painting as well as of traces burnishing or polishing.

185

N = 662 (12.8 %)

Dosariyah

Name

Code

Description

2b

This partly handmade, partly slow-wheel-made ware shows a fine mineral temper (fine to medium size in low to medium quantities) and no visible chaff inclusions. The minerals are partly angular and partly rounded in shape. The N = 1451 texture is hard and the color ranges from light creamy to buffish to light red- (28.1 %) dish to light greenish. Numerous fragments show remains of a monochrome dark painting as well as traces of burnishing or polishing.

Black-on-Buff (Fine) Ware, BoBW 2c

Black-on-Buff (Eggshell) Ware, BoBW

2d

3a

Coarse Mineral Ware, CMW

3b

3c

Soft Bone Tempered Ware, SBTW

Organic Buff Ware OBW

4

5

The characteristics of this ware correspond to 2b. But chaff inclusions (fine to medium size in low to medium quantities) are visible as well. The texture is N = 1923 hard and the color ranges from light creamy to buffish to light reddish to light (37.1 %) greenish. Numerous fragments show remains of a monochrome dark painting as well as traces of burnishing or polishing. This partly handmade, partly slow-wheel-made ware shows no signs of any visible mineral or chaff inclusions. The texture is hard and the color ranges from light creamy to light greenish. Most of the fragments show remains of a monochrome dark painting as well as traces of burnishing or polishing. This handmade and coarse tempered ware shows a lot of rounded yellowish white to light gray mineral inclusions (medium to coarse size in medium to high quantities). Chaff temper occurs as well (fine to medium size in low to medium quantities). The texture is very porous and the sherds appear light yellowish in color.

N = 187 (3.6 %)

N = 10 (0.2 %)

This handmade and coarse tempered ware shows a lot of rounded light grayish mineral inclusions (coarse size in high quantities). Abundant chaff temper N = 16 (fine to medium size in medium quantities) occurs as well. The texture is (0.3 %) slightly porous and the sherds appear pinkish in color.

This handmade and coarse tempered ware shows a lot of rounded slightly translucent mineral inclusions (fine to coarse size in medium to high quanonly from collapsed tities). Some chaff temper occurs as well (fine size in low to medium quantities). The texture is hard fired but slightly porous and the sherds appear light profiles grayish to bluish in color. This partly handmade, partly slow-wheel-made ware shows mineral inclusions (fine size in low to medium quantities) and chaff inclusions (fine to medium size in low to medium quantities). Furthermore, pieces of bone (medium sized N = 3 in low quantities) occur as temper. The texture is soft and slightly porous, and (> 0.1 %) the sherds appear light yellowish in color. One of the sherds bears remains of a monochrome dark painting.

This partly handmade, partly slow-wheel made ware shows a high amount of chaff inclusions (medium to coarse size in medium quantities) and no visible N = 1 mineral inclusions. The texture is hard fired and the sherds appear yellowish– (> 0.1 %) buff in color. One of the sherds bears remains of a monochrome dark painting.

southern Mesopotamia based on the comparison of specific forms and motifs.7 In addition, a study of the geochemical composition of potsherds in 1977 revealed that some of the clay from tested fragments from Saudi Arabia, Bahrain and Qatar originated somewhere in southern Mesopotamia.8

to commonly known BoBW, but it seems reasonable to assume that OBW is a variant of BoBW. 3.1. Black-on-Buff Ware (BoBW) This type of pottery was widely distributed over much of the Near East during the sixth, fifth and early fourth millennia BC, a fact which adds a multitude of facets to the study of single assemblages. Shared similarities include the use of monochrome dark paint on a usually buff surface as well as aspects of the manufacturing process. Based on formal characteristics such as vessel form, shape and painted motifs, Black-on-Buff Ware is commonly classified into different regional variants such as Ubaid Ware, Susiana Ware or Bakun Ware. Similar pottery found on the Arabian Peninsula was initially referred to as Ubaid pottery imported from

The great majority of Black-on-Buff Ware documented at Dosariyah is likely to fit into this scheme. The results of the analysis of 34 BoBW sherds from trench S1 match the supposition of a Mesopotamian origin (see Chapter 8). An examination of the range of painted motifs on all recovered pieces, however, suggests a more diverse provenance which includes present-day Iran. 7 8

186

Burkholder 1972; Oates J. 1976; 1978; Bibby 1980. Oates J. et al. 1977; Roaf and Galbraith 1994.

Christine Kainert: The Pottery

Figure 7.1. Rim fragments of open bowls from Dosariyah, Ware 2d (a DOS2012-27297; b DOS2012-22721) (drawings: C. Kainert).

number of excavated sites in Mesopotamia including Abu Shahrain/Eridu and Tell el-Oueili.10

A total of 4223 sherds (81.7% of the excavated and pieceplotted assemblage) can be assigned to Black-on-Buff Ware. This shows a rather homogeneous distribution across the site with slightly higher percentages in the northern and eastern part of the site and a comparatively lower percentage in the south. Most vessel fragments belong to the macroscopic sub-ware groups 2b and 2c, group 2a with a higher proportion of coarser temper has been identified less frequently. The scarcest yet most distinct type is ware 2d. It is characterized by the absence of macroscopically visible organic or mineral inclusions. Moreover, this type shows very thin walls continuously from top to bottom of the vessel. Within Mesopotamia this is commonly referred to as Eggshell Ware ever since Lloyd and Safar first described this type as ‘bowls of almost egg-shell fragility’.9 Interestingly, this type of pottery seems to be linked to a simple open vessel form with a mostly rounded base (Figure 7.1) which could be identified at Dosariyah throughout all excavated layers in varying quantities and with differences only in the height and diameter of the vessels. Almost no sherd assigned to ware 2d lacks remains of paint. Hence, it can be assumed that all Eggshell Ware vessels showed some form of painted decoration consisting of simple geometric motifs, mostly close to the rim or close to the bottom. Comparable vessels are known from a 9

Of the piece-plotted assemblage of BoBW, 19.3% could be assigned either to an open or a closed vessel form; of those assigned, a clear majority (66.2%) belongs to open vessels. These numbers however, must not be equated with the proportions of open and closed vessels present at the site as the rim of an open bowl is likely to break into more fragments than the rim of a closed jar and forthcoming statistical analysis will provide more detailed data on this topic. The broad range of vessel forms strongly suggests that their functions appeared to be quite versatile, both within daily utilitarian and commensal activities ranging from the storage and transport of goods in closed jars to the processing, display and consumption of goods and foods in open bowls. The fragmented state of the pottery from Dosariyah, as well as the absence of more or less completely preserved vessels, require a formal typology of fragments which focuses on details rather than on, for example, complete vessel shapes. The spectrum of rim forms is much more heterogeneous than initially suspected. In a few cases, specific rims were identified, which in combination with remains of painted decoration can function as a distinct chronological marker and may

Lloyd and Safar 1948: 123; see also Oates J. 1960: 36.

10

187

Safar et al. 1981; Lebeau 1991.

Dosariyah

Figure 7.2. Rim fragments of closed jars with a split rim from Dosariyah (a DOS2012-30247.2; b DOS2012-25732; c DOS2011-26626) (drawings: C. Kainert).

also refer to certain inter-regional relations. A good example is a slightly globular jar with a distinctive split rim, whose inner rim has a series of evenly placed holes (Figure 7.2). This specific type of jar has been continuously attested for the levels XVIII to VIII of the Temple Sounding at Abu Shahrain/Eridu and is therefore referred to as a rare connecting link between phases from late Ubaid 1 to Ubaid 3. Lloyd and Safar further point to the decoration of those jars from the upper levels, stating that the quality of paint is typical of Ubaid 2 but the painted design resembles more what is known from Ubaid 3.11

The presence of secondary perforations clearly shows that at least some vessels were considered valuable enough for someone to try to mend them. This could mean that an adequate supply, especially of BoBW, was not constant. As a result, the pots that were in the current possession of the inhabitants were cherished and repaired for as long as possible as it was not known if or when new vessels would arrive. Another possibility is that the vessels were considered a valuable item in general, regardless of supply, and were thus treated with great care, although in this case one might expect more than just five perforated fragments of CW within the piece-plotted collection, unless there was either a considerable difference in valuing vessels of BoBW and CW, or vessels made of CW simply did not break as easily as those made of BoBW. Furthermore, a small number of micro-borers suitable for drilling small holes into pottery was documented at the site, suggesting that this task was carried out at Dosariyah.13 Perforated fragments appear throughout the excavated sequence in varying quantities and were found in the north, east and south of the site. This suggests that the task of mending broken vessels was not an unusual one but something which was done fairly regularly.

In contrast to the rim assemblage at Dosariyah, the spectrum of base forms identified could not be less diverse. Evidence of base rings is completely lacking, although attested from contemporaneous Mesopotamian and Iranian sites.12 Instead, all base fragments documented at Dosariyah belong to flat or rounded bottoms which can only be differentiated on the basis of the transition angle to the vessel wall where preserved. This made it especially difficult to distinguish them from body fragments, which is why the actual percentage of base fragments at Dosariyah is very low. Indeed, it is very likely that an unknown number of base fragments were documented as body fragments owing to their close morphological resemblance.

More than 17% of the piece-plotted BoBW fragments show remains of a dark monochrome paint on the exterior and sometimes on the interior as well. The color of the paint ranges from dark reddish or brownish to blackish and sometimes even slightly bluish, and the thicker the pigment, the darker it appeared. The reddish/brownish color visible on some of the painted sherds can be explained by the presence of a certain amount of hematite, which was a component

An interesting aspect noted during the process of documentation of the pottery is the appearance of secondary perforations which represent attempts at mending broken pots. Of the piece-plotted BoBW assemblage, more than 2% show traces of such repairs. 11 12

Lloyd and Safar 1948: 124; see also Oates J. 1960: 37–38. Safar et al. 1981; Alizadeh 2006; 2008.

13

188

Kainert and Drechsler 2014.

Christine Kainert: The Pottery

Figure 7.3. Distribution of Coarse Ware in the Gulf. Shoreline reconstruction for the upper Gulf between c.5300–4550 BC; after Pournelle 2003: 123, fig. 44 (cartography: P. Drechsler).

of the mixture used to paint the vessels.14 The range of painted motifs is dominated by simple geometric designs, for example horizontal running straight lines, wavy and zigzag lines, and rows of triangles, dots, bows or grid patterns. Horizontal lines often appear alone, but in some cases two or more patterns were combined into more complex motifs. A very few stylized floral depictions were identified and most likely resemble leaves in a radial arrangement. The great majority of open vessels were painted close to the rim. In contrast, closed jars unsurprisingly show paint remains on the exterior, invariably along the neck and shoulder areas.

a common phenomenon in the central Gulf region with at least ten further sites identified in Saudi Arabia (Sites 2, 9, 10, 11, 12, 16, 22, 29, 34,15 Ain as-Sayh),16 one site in Qatar (al-Da’asa)17 and another in the United Arab Emirates (Jazirat al Hamra 4).18 Hitherto, no detailed studies of Coarse Ware have been carried out at any of these sites, mainly because of the small number of potsherds but also due to the absence of a copious stratified collection which would have allowed such an analysis. Recent research in the Kuwait Bay region has considerably extended the distribution of Coarse Ware and helped broaden our knowledge on this matter as the excavations at H3/As-Sabiyah and Bahra 1/AsSabiyah present the first stratified and comprehensively analyzed ceramic collections with Coarse Ware of Neolithic sites in the upper Gulf.19 Consequently,

3.2. Coarse Ware (CW) Coarse Ware occurs at Dosariyah alongside Black-onBuff Ware from the beginning of occupation onward. The simultaneous existence of both types appears to be

15

Many thanks to Christoph Berthold from the University of Tübingen for the analysis of the paint remains of some selected sherds from Dosariyah. See Simpson 1997 for similar findings from Mesopotamia.

14

16 17 18 19

189

Potts 1990: 59–61; Masry 1997. McClure and al-Shaikh 1993. de Cardi 1986. Vogt 1994. Carter and Crawford 2010a; Smogorzewska 2013.

Dosariyah only those assemblages provide a suitable basis for comparison of the pottery from Dosariyah. From the excavated and piece-plotted assemblage from Dosariyah, a total of 918 pieces (17.8%) can be assigned to Coarse Ware with the highest percentages in the southern part of the site. The high number of identified macroscopic subware groups suggests an unstandardized preparation process of the clay and this underlines the assumption that the manufacture of Coarse Ware vessels was an individual task which was carried out according to necessity.20 These vessels were made by hand and although for the majority of fragments, it is not possible to determine exactly how they were made, some sherds show evidence of the use of either coiling or slab-building, both commonly used techniques. Of all 154 piece-plotted CW base fragments, 23 pieces (14.9%) show impressions of circular plaited mats on the outer surface. This suggests that some pots were made on such a surface and its texture was accidentally impressed on the undersides during manufacture, either during forming or during drying prior to firing, although it cannot be totally excluded that these impressions were left on purpose for reasons of style. The texture of the Coarse Ware fragments is mostly porous and friable and suggests a rather low firing temperature, probably in an open bonfire. This was previously observed in relation to the Coarse Ware assemblage recovered from H3/As-Sabiyah.21 Despite the diverse range of macroscopic wares, the range of vessel forms is relatively limited. Most common are simple open bowls with a diameter in the range of 8 to 29  cm and with straight or slightly rounded walls, although several variations have been attested as well. Other fragments derive from closed pots and jars, and some pieces belong to small cup-like shapes (Figure 7.4). The presence of numerous knobs suggests the need for secure transportation, which might indicate the handling of heated or heavy vessels. Their possible use as cooking pots has been discussed before,22 but there was insufficient evidence of charring which might prove this as all of the the sherds were too friable and fragmented. Apart from this suspected (but unproven) function, the predominance of open forms and generally thick walls seem well suited to common domestic tasks like the temporary storage, processing or even display of goods. Furthermore, the smaller cuplike vessels found in trench E1 might have been used as drinking vessels. One fragment from trench S2 belongs to a flat plate-like vessel that, as a part of commensal activities, might have been suitable for the display or the consumption of food or other goods.

20 21 22

Figure 7.4. Fragments of Coarse Ware from Dosariyah (a DOS2012-19801; b DOS2012-15629; c DOS2012-23541+23542+23543; d DOS2010-2863+2864+2865+2866+2867+2868+2872) drawings: C. Kainert).

The Coarse Ware from Dosariyah shows no signs of painted decoration. Only two rejoined rim fragments of a small cup-like shape from trench E1 show a complex geometric incised decoration on the outer surface (DOS2011-21696+22069; Figure 7.5).23 The horizontal and diagonal running lines were incised while the clay

Carter 2010b: 36. Carter 2010b. Carter 2010b: 33.

23

190

Kainert and Drechsler 2014: 220.

Christine Kainert: The Pottery

Figure 7.5. Incised Coarse Ware from Trench E1 (DOS2012-21696+22069) (photograph: B. Kiepenheuer-Drechsler).

At present, the value of diagnostic Coarse Ware pieces as a chronological marker is minimal. When found on the surface without other Neolithic remains, it is almost impossible to ascertain a clear connection to the Neolithic period since the simplicity of vessel forms plus the absence of decoration do not allow an unambiguous distinction from similar Coarse Ware fragments of later periods.

was still wet and cover the whole preserved surface. So far, no comparable pieces are known from other sites in the Gulf region. What could be considered a form of decoration are small flat knobs with a diameter of c.1  cm which were applied to the exterior of several fragments found in the east of the site and belong to cup-like vessels. Even though these might have had a practical function like securing the grip of the cup, they might also have been applied for decorative purposes.

Table 7.3. Sherd count of all excavated and piece-plotted (in brackets) pottery sherds from each trench and each field season from 2010 until 2012. An additional 13 sherds from the overburden are excluded here. Spring 2010 N1

N2

N3

E1

E1.1

E1.2

E2

S1

S2

S3

261 (211)

-

-

225 (217)

-

-

-

238 (189)

-

-

Total 724 (617)

Fall 2010 -

-

-

-

-

-

-

-

Spring 2011 -

-

-

-

-

-

9 (9)

-

291 (287) 703 (701)

Fall 2011 -

117 (71)

-

-

-

-

-

-

-

Spring 2012 -

Total 261 (211)

-

117 (71)

-

225 (217)

1755 (1380) 1755 (1380)

1763 (1520) 1763 (1520)

505 (438)

-

-

-

505 (438)

9 (9)

238 (189)

994 (988)

148 (148) -

-

-

439 (435) 712 (710)

117 (71)

4023 (3338) 6015 (5171)

191

148 (148)

Dosariyah Table 7.4. Stratigraphic units of trench N1 with a sherd count of piece-plotted sherds for each main ware group. Pieces that originate from collapsed profiles are excluded from this table. BoBW N1-I

N1-II

N1-III

82

23

CW

CMW

16

-

4

18

2

-

OBW

Total

-

98

-

-

-

66

-

211

-

-

N1-IV

59

7

-

-

Total

182

29

-

-

BoBW Stratigraphic Unit

BTW

-

27 20

CW

N1-I N1-II N1-III N1-IV 0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Figure 7.6. Stratigraphic units of trench N1 with percentages of piece-plotted sherds for each main ware group.

4 The Excavated Collection

Trench N2

Over the course of two years from spring 2010 until spring 2012, five field seasons were conducted with a total of eight trenches excavated (Table 7.3). While the extension of the trenches E1, S1, S3 and N1 was kept small in order to investigate the stratigraphic sequence at different parts of the site, trenches S2, N2, N3 and E1 covered larger areas and were extended to identify intra-site spatial patterns and specific zones of activity within the settlement. As explained in Chapter 5, all trenches were excavated in single spits. Afterwards, stratigraphic units were defined based on common characteristics of the individual sediment layers observed in the profiles. Accordingly, these units correspond more or less to the natural process of sedimentation and represent the units for further analysis, description and interpretation.

Trench N2 measured 8  x  8  m and was excavated to a depth of 2.95  m. In total, 117 sherds of pottery were recovered from this trench within approximately 53.5 m3 of excavated sediment, with 71 of them being piece-plotted. All sherds (and all artifacts in general) were discovered in the first four spits, with the first being virtually free of finds (Table 7.5, Figure 7.6). Four macroscopic main ware groups have been identified here, with CMW and OBW being represented by just one sherd each. Due to the low count of recovered pottery in this trench, clear trends within their stratigraphic dispersion can barely be determined. The overall low count of finds suggests a minimal utilization of that part of the site or severe post-depositional displacements. Trench N3

4.1 The trenches

In trench N3, 1753 pottery sherds were documented within 76.3  m3 of excavated sediment, spread over six stratigraphic units with 1380 of them being pieceplotted. Three different types of main ware groups were identified here, with CMW being represented by just two fragments. The vast majority (87.5%) of sherds were assigned to BoBW, while 12.4% of the pieceplotted sherds belong to CW (Table 7.6, Figure 7.8). A clear trend is discernable for the vertical distribution of both, CW and BoBW: from the lowest to the uppermost unit the percentage of CW constantly decreases with only a minimal increase in stratigraphic unit N3-I, representing the deflation horizon.

Trench N1 Trench N1 measured 3  x  2  m and was excavated to a depth of 1.7 m. A total of 261 sherds were documented here within 9.5 m3 of excavated sediment, with 211 of them found in situ and therefore being piece-plotted (Table 7.4). The material from this trench appears rather homogeneous with only two macroscopic main wares present. In total, 86.3% belong to Black-on-Buff Ware and 13.7% to Coarse Ware with a decreasing proportion of BoBW following the stratigraphy upwards (Figure 7.6). 192

Christine Kainert: The Pottery Table 7.5. Stratigraphic units of trench N2 with a sherd count of piece-plotted sherds for each main ware group. Pieces that originate from collapsed profiles are excluded from this table. BoBW

CW

CMW

BTW

OBW

Total

N2-I

-

-

-

-

-

-

N2-III

11

1

-

-

-

12

N2-II

34

N2-IV

11

10

Total

1

2

55

-

14

-

-

1

BoBW Stratigraphic Unit

-

1

-

CW

CMW

46 13

1

71

OBW

N2-I N2-II N2-III N2-IV 0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Figure 7.7. Stratigraphic units of trench N2 with percentages of piece-plotted sherds for each main ware group.

Trench E1 (E1/E1.1/E1.2)

excavation, the whole trench measured 8  x  10  m and reached a depth of 3.3 m below the surface, with a total of 93.3 m3 of excavated sediment. An additional sondage down to a total depth of 3.7 m below the surface proofed

Trench E1 was first excavated in spring 2010 and subsequently extended in spring 2012. At the end of the

Table 7.6. Stratigraphic units of trench N3 with a sherd count of piece-plotted sherds for each main ware group. Pieces that originate from collapsed profiles are excluded from this table. BoBW CW

CMW

BTW OBW

Total

N3-I

108

7

-

-

-

115

N3-III

73

9

-

-

-

82

N3-II

267

N3-IV

664

N3-V

92

N3-VI

3

Total

1207

16 111 23 5

171

BoBW

-

-

2

-

-

-

-

-

2

-

CW

-

283 777 115 8

1380

CMW

Stratigraphic Unit

N3-I N3-II N3-III N3-IV N3-V N3-VI 0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

Figure 7.8. Stratigraphic units of trench N3 with percentages of piece-plotted sherds for each main ware group.

193

100%

Dosariyah Table 7.7. Stratigraphic units of trench E1 with a sherd count of piece-plotted sherds for each main ware group. Pieces that originate from collapsed profiles are excluded from this table. BoBW CW

CMW

BTW OBW

Total

E1-I

78

24

1

-

-

103

E1-III

1097

100

10

3

-

1210

E1-II

E1-IV E1-V

E1-VI

287 289 27 20

E1-VII 16 Total

1814

170

7

22

-

4

17

-

-

1

336

BoBW

315

-

-

-

464

-

-

-

2

-

44

-

-

21

-

18

22

3

-

2175

CW

CMW

BTW

Stratigraphic Unit

E1-I E1-II E1-III E1-IV E1-V E1-VI E1-VII 0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Figure 7.9. Stratigraphic units of trench E1 with percentages of piece-plotted sherds for each main ware group.

to be free of artifacts. Although trenches E1, E1.1 and E1.2 were excavated and documented separately, the following description of the recovered pottery includes the documented material from all three sub-trenches and is just labeled E1. Altogether, 2506 sherds were removed from the area with 2175 of them being pieceplotted in situ. They were assigned to four different main ware groups and represent the most heterogeneous collection from a trench at the site.

A total of 83.4% was assigned to BoBW and 15.5% of the assemblage belongs to CW. A minority of 1.1% belongs to CMW and BTW (Table 7.7, Figure 7.9). The fifth main ware group SBTW was identified in E1 as well, but only for sherds from collapsed profiles, which is the reason why SBTW is excluded from further statistical evaluation presented here. The stratigraphic distribution shows no clear trends over time but varies in relation to the percentage of the main wares. In particular, frequency shifts between CW and BoBW

Table 7.8. Stratigraphic units of trench E2 with sherd count of piece-plotted sherds for each main ware group. Pieces that originate from collapsed profiles are excluded from this table. BoBW CW

Stratigraphic Unit

E2-I 4

5

CMW

BTW OBW

Total

-

-

9

BoBW

CW

-

E2-I 0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

Figure 7.10. Stratigraphic units of trench E2 with percentages of piece-plotted sherds for each main ware group.

194

100%

Christine Kainert: The Pottery Trench S1

are clear: after a decrease of the CW-percentage from stratigraphic unit E1-VII to E1-VI, it suddenly increases to almost 40% in unit E1-V and then rapidly decreases again in unit E1-IV from where it starts to increase again up to unit E1-II, and finally to decrease in unit E1I. The minor proportions of CMW and BTW only appear in the upper four stratigraphic units E1-I to E1-IV in a more or less steady proportion.

Trench S1, measuring 2 x 2 m, was excavated to a depth of 1.1 m below the surface, where the natural soil was reached. Overall, 238 sherds were documented from a total of 7.5  m3 of excavated sediment, with 189 being piece plotted. Before the actual documentation process of the pottery began in the fall of 2011, 43 pieces had already been removed in spring 2010 for chemical analyses (see Chapter 8). For those sherds, only basic characteristics were documented while a more detailed study was not possible.

Trench E2 Trench E2 was excavated to a depth of 1.75 m below the surface and represents the trench with the lowest number of recovered artifacts. Within 7.0  m3 of excavated sediment, only nine pottery sherds were documented and piece-plotted here. They were found within the upper part of the stratigraphic sequence (Table 7.8, Figure 7.10). Due to the homogeneity of the sediments, stratigraphic units could not be further differentiated. All identified artifacts come from the uppermost part of the stratigraphic sequence E2-I that represents a deflation and trampling horizon (see Chapter 5).

In total, 65.6% belong to BoBW while the remaining pieces with a percentage of 34.4% of the recovered assemblage can be assigned to CW (Table 7.9, Figure 7.11). In the lowest stratigraphic unit S1-V, the amount of CW accounts for almost 50% but its proportion decreases rapidly towards the upper part of the stratigraphic sequence. Trench S2 Trench S2 measured 8  x  4  m in the upper part and 8 x 3 m in the lower part. It was excavated to a depth of 1.7 m, resulting in 46.4 m3 of excavated sediment. A total of 994 sherds were documented with 988 of them being piece-plotted.

Besides the fact that five of the nine documented sherds belong to CW and four to BoBW, the small assemblage does not provide much more information since all were non-diagnostic and no vessel forms could be identified.

Despite the comparatively high number of documented sherds, the material itself is rather homogeneous with

Table 7.9. Stratigraphic units of trench S1 with sherd count of piece-plotted sherds for each main ware group. Pieces that originate from collapsed profiles are excluded from this table. BoBW CW

CMW

BTW OBW

Total

22

1

-

-

-

23

S1-III 23

4

-

-

-

27

S1-I

S1-II

18

S1-IV 53 S1-V

8

Total 124

2

57 1

65

-

-

-

-

-

-

-

-

-

-

-

20

110 9

189

Figure 7.11. Stratigraphic units of trench S1 with percentages of piece-plotted sherds for each main ware group.

195

Dosariyah Table 7.10. Stratigraphic units of trench S2 with sherd count of piece-plotted sherds for each main ware group. Pieces that originate from collapsed profiles are excluded from this table.

S2-I

S2-II

BoBW CW

CMW

BTW OBW

Total

14

-

-

-

-

14

91

-

-

-

328

167

S2-III 237

S2-IV 299 Total 717

6

174 271

-

-

-

-

-

-

Stratigraphic Unit

BoBW

-

173

-

473

-

988

CW

S2-I S2-II S2-III S2-IV 0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Figure 7.12. Stratigraphic units of trench S2 with percentages of piece-plotted sherds for each main ware group.

Table 7.11. Stratigraphic units of trench S3 with sherd count of piece-plotted sherds for each main ware group. Pieces that originate from collapsed profiles are excluded from this table. BoBW CW

CMW

BTW OBW

Total

15

-

1

-

-

16

S3-III 56

7

-

-

-

63

S3-I

S3-II

26

S3-IV 22 S3-V

1

Total 120

7

13 -

27

-

-

-

-

-

-

1

-

-

-

-

33

35 1

148

Figure 7.13. Stratigraphic units of trench S3 with percentages of piece-plotted sherds for each main ware group.

only two main ware groups being identified here. A proportion of 72.6% belong to BoBW and 27.4% belong to CW (Table 7.10, Figure 7.12). A clear trend is detectable following the stratigraphy upward, with the proportion of CW constantly decreasing and completely disappearing in the uppermost unit S2-I.

Trench S3 This trench measuring 2 x 2 m was excavated to a total depth of 1.5  m, with 4.9  m3 of sediment explored. A total of 148 pottery fragments were recovered, all of which were piece-plotted. Three macroscopic main 196

Christine Kainert: The Pottery assemblages are too small to confirm a reliable dating, but Ubaid 5 has been tentatively suggested for their respective assemblages.32

ware groups were identified here: 81.1% of the sherds were assigned to BoBW and 18.2% to CW, while CMW is represented by only one sherd from stratigraphic unit S3-I, which makes up 0.7% of the whole collection (Table 7.11). The proportion of CW strongly decreases from the lower unit S3-IV to the overlying unit S3-III but then slightly increases again in unit S3-II (Figure 7.13).

Based on a provisional examination of formal aspects of the pottery of Dosariyah, the dating basically fits within the previously suggested time frame of Ubaid 3 to early Ubaid 4. Considering the 14C assays of the site (all from marine shell), this equates to an approximate date range of 5100–4600 BC (see Chapter 6).

5 Chronological Aspects

Nevertheless, two sherds from Dosariyah contradict this general dating of the site. One piece is a pendant made from a reworked painted potsherd (DOS2011-12573, see Chapters 11, 15) and the other is a small sherd (DOS2010-8141.1). Both were found in the southern area of the site, the latter in a collapsed profile of trench S2. Both pieces share the same type of decoration in dark monochrome paint depicting a simple geometric motif with a later applied fine incised wavy or zigzag line. This specific type of decoration is described as diagnostic of the early Ubaid 2 phase.33 An explanation for their appearance within an apparently younger context cannot be provided, yet perhaps this type of decoration simply lasted longer than previously assumed. Another possibility is that people took especial care of particular vessels and these had extended curation lifespans. Such vessels might have gained a certain value over time and were therefore seen as high-value heirlooms or gifts.

There has been a broad consensus over the relative chronology of Black-on-Buff Ware pottery found along the southern coastline of the Gulf. No material is known from the oldest phase (Ubaid 0), only very little material is associated with older (Ubaid 1 and 2) and younger (Ubaid 5) phases, whereas the majority of sites with pottery are assumed to date to the range of Ubaid 2/3, 3 and 4 phases.24 Thus far, the earliest sites according to this framework are Ain Qannas, Marawah 11/MR11, Umm al-Qaiwain 2/ UAQ2 and al-Ayram 6/AR6, and especially noteworthy is the high-necked jar decorated with triangles, chevrons and dotted lines found at MR11. Based on stylistic comparisons, a chronological classification anywhere between the late Ubaid 1 and early Ubaid 3 is possible.25 Noticeably, Ain Qannas is located inland in the central Gulf region in Saudi Arabia, while the other three are located along the lower region of the Gulf coast in the UAE. At Ain Qannas, a few sherds were discovered of a distinct bowl type with a dense grid pattern which is associated with Ubaid phase 2 but is also known from the subsequent phase Ubaid 2/3.26 From a third site, UAQ2, a group of c.250 plain and painted fragments was collected which reportedly dates to the second half of the sixth millennium BC based on stylistic characteristics of the material as well as on 14C dates obtained from the same excavated level.27 The fourth site, AR6, only had a single sherd reported from the surface and its painted decoration was described as characteristic for Ubaid 2.28

6 Conclusions The pottery recovered from Dosariyah is quite remarkable in many aspects. To begin with, the sheer number of 15,410 documented potsherds from the surface, from excavated contexts as well as from re-examined collections is exceptional. This is partly due to the fact that few other sites have been properly excavated as yet and apart from the recently investigated sites of H3/As-Sabiyah and Bahra 1/AsSabiyah in Kuwait, all other pottery-yielding Neolithic sites on the Arabian Peninsula have only yielded considerably smaller amount of sherds of no more than a few hundred at most.

Pottery from other sites in the upper Gulf such as H3/ As-Sabiyah and Bahra 1/As-Sabiyah in Kuwait belongs mainly within Ubaid 2/3 and Ubaid 3.29 Dosariyah and Abu Khamis in Saudi Arabia are both commonly associated with Ubaid 3 and 4,30 although Abu Khamis is considered to be slightly younger.31 At two sites, Ras Abaruk (Qatar) and al-Markh (Bahrain), the recovered 24 25 26 27 28 29 30 31

The lack of contexts such as architectural remains or installations make it much more difficult to extract all the information necessary to fully understand the variety of aspects that led to the deposition of pottery at the site. Nevertheless, with a stratigraphy of up to 3.3  m and excavated areas measuring up to 8  x  10  m, combined with an extensive approach of piece-plotting, the site and its pottery offer rare insights into a rarely explored period, as well as the opportunity to follow

Frifelt 1989; Carter 2010b. Beech et al. 2005: 46–47; Méry et al. 2016: 164. Masry 1997: 75; Oates J. 1976: 24–25. Charpentier et al. 2012: 3; Méry and Charpentier 2013: 76. Anon. 2001. Masry 1997: 114; Carter 2010b: 63–64; Smogorzewska 2013: 548. Oates J. 1976: 25; Carter 2010b. Oates J. 1976: 26.

32 33

197

Oates J. 1983: 255. Oates J. 1960: 35–36; Crawford 2010: 163.

Dosariyah the depositional sequence of the site throughout the late sixth and early fifth millennia BC.

of specialized craftsman or major workshops. The predominance of simple open forms plus numerous large knobs for secure handling further supports the idea that the ware was manufactured primarily for daily utilitarian purposes. Nevertheless, the spectrum of recovered CW forms also includes small cups, holemouth jars and plates. At the same time, numerous fragments of BoBW represent a much more diverse set of open and closed vessel forms in different sizes. This might represent a similarly diverse set of possible functions but could also indicate certain social aspects such as commensal activities. The relative dating of Dosariyah based on formal aspects of BoBW generally fits into the already established phases of Ubaid 3 and early Ubaid 4, and equals the absolute dating based on 14C assays within the time span of c.5100–4600 BC. A forthcoming detailed analysis of vessel forms and painted motifs will further lead to a more multifarious relative chronological analysis with comparisons to other contemporaneous sites in the Gulf as well as in adjacent regions.

The fact that the eight excavated areas in the north, east and south of the site show different stratigraphic sequences makes it very difficult to understand details of the settlement dynamics as a whole. Of course, comparative analysis can still be carried out but only the three southern trenches allow a matching of their respective stratigraphic units and future analysis will therefore focus on changes between each of the three investigated areas of the site. Three possibilities will be tested in this manner, namely that possible differences may be due to chronological, functional or social causes. This analysis will be carried out primarily based on a detailed examination of the recorded characteristics of the collection such as shape or decoration. Aspects regarding patterns of distribution will be considered as well. In contrast to previous claims,34 it is now clear that the percentages of two of the main ware groups account for 81.6% (BoBW) and 17.8% (CW). It remains unclear how former investigations by Burkholder in 1968 and Masry in 1972 ended up with a percentage of CW in the range of 45–50%, particularly as the amount of documented CW is nowhere as high either on the surface or in any of the excavated trenches.

Acknowledgements The detailed analysis and interpretation of the Dosariyah pottery is the topic of my PhD thesis, which is planned as a succeeding volume to the Dosariyah publications. I would like to express my gratitude to the Saudi Commission for Tourism and National Heritage (SCTH), Ali al-Ghabban, Jamal Omar and Abdulhamid al-Hashhash for their kind permission to study the pottery excavated in Dosariyah and to re-examine the collection assembled by Masry in 1972 within the facilities of the National Museum in Riyadh. I would also like to thank Susan Pollock and Philipp Drechsler for their helpful comments on this chapter.

The heterogeneous fabric of CW stands in contrast to the rather homogeneous fabric of BoBW, suggesting a comparatively unsystematic approach to its manufacture. It is therefore plausible that CW vessels were more likely produced according to needs than as a stock. The heterogeneous appearance of CW might also represent a relatively high number of makers instead

34

Frifelt 1989: 408; Masry 1997: 80–81.

198

Chapter 8 Geochemical Analysis of Putative Local and Ubaid Ceramics from Dosariyah Peter Magee and Steven Karacic 1 Introduction

strata excavated at trench S1, located at the southern part of the site.

In this report we present the results and implications of geochemical analysis of pottery from the Neolithic site of Dosariyah in Saudi Arabia. This site is critically important for our understanding of Ubaid-period trade in the Arabian Gulf since it has provided large amounts of Ubaid ceramics and a Coarse Red Ware which has for a long time been considered to be locally produced.1 The excavations at Dosariyah have provided samples of both of these wares in carefully excavated stratigraphic contexts. Analysis of material from this site thus permits a detailed study of the nature and impact of interaction between Neolithic nomadic pastoralists of Arabia and early sedentary farming cultures of southern Mesopotamia. In this first round of analysis, we aim to contribute to this issue by addressing two inter-related questions and our results indicate meaningful answers can be provided to both:

The chosen method of analysis was Inductively Coupled Plasma (ICP) at ACTLABS in Canada. The use of ICP in compositional studies is now widespread. Comparative studies between ICP and instrumental neutron activation analysis (INAA), the standard analytical method for archeological ceramics, have indicated that it is a robust analytical tool with widespread application.2 A key issue remains the method of digestion for the sample before the resultant vapor is injected into an argon flame. The use of strong acids, such as hydrochloric, is viewed as potentially problematic because it may create calcium fluorides in digestion which will precipitate other elements. Microwave digestion of the sample with hydrochloric, hydrofluoric and nitric acids has proved successful when measured against known standards.3

•• does geochemical analysis support the long-held belief that Coarse Red Ware and Ubaid pottery have different origins? •• does geochemical analysis indicate any diachronic or synchronic differences in the sources of Ubaid pottery found at Dosariyah?

A proprietary method of digestion was employed at ACTLABS which guarantees total sample digestion. Samples are prepared and analyzed in a batch system.4 Each batch contains a method reagent blank, certified reference material and 17% replicates. Samples are mixed with a flux of lithium metaborate and lithium tetraborate and fused in an induction furnace. The molten melt is immediately poured into a solution of 5% nitric acid containing an internal standard and mixed continuously until completely dissolved (c.30 minutes). The samples are run for major oxides and selected trace elements on a combination simultaneous/sequential Thermo Jarrell-Ash ENVIRO II ICP or a Spectro Cirros ICP. Calibration is performed using seven prepared USGS and CANMET certified reference materials.

2 Methodology A total of 43 samples of Coarse Red and Ubaid wares were analyzed from Dosariyah. The samples were chosen so as to represent the diversity of both assumed local and imported pottery. Selection of the samples was made prior to the typology of wares presented in Kainert’s Chapter 7, and therefore different terminologies were employed. For the geochemical study, the samples were divided into four macroscopic wares (Table 8.1) with the Coarse Red Ware labelled as Ware 1. The Ubaid wares were split into Fine Ware, sometimes painted (Ware 4) and two coarser standard wares, one with a high concentration of medium to fine mineral inclusions (Ware 2) and the other combined medium to fine mineral inclusions with traces of vegetal inclusions (Ware 3). Samples of these wares were chosen from all

1

3 Analytical Results Elemental abundance for the following elements were obtained: Al, Ba, Be, Ca, Ce, Co, Cr, Cs, Cu, Dy, Er, Eu, Fe, Ga, Gd, Ge, Hf, Ho, K, La, Lu, Mg, Mn, Na, Nb, Nd, Ni, Pr, Rb, Sc, Si, Sm, Sr, Ta, Tb, Th, Ti, Tm, U, V, Y Yb, Zn and Zr 2 Moens and Dams 1995; Hein et al. 2002; Little et al. 2004; Tsukada et al. 2005. 3 Kennett et al. 2002. 4 The text describing the method of analysis was provided by ACTLABS.

Masry 1974; Oates J. et al. 1977; Drechsler 2011.

199

Dosariyah Table 8.1. The stratigraphic origin and macroscopic ware attribution for the analyzed samples. Stratigraphic unit

Ware 1 Unidentified (Coarse Red Ware)

S1-I Surface/deflation horizon

329

S1-II Above shell-midden deposits layer S1-008 S1-III Intermediate horizon between shell-midden 2310 deposit layer S1-008 and ‘settlement horizons’ layer S1-016 and S1-018

682 1306 1803

S1-IV ‘Settlement horizon’ layer S1-016 and ‘settlement horizon’ layer S1-018/basis of the sequence

3079 3208

2873 2902 3004 3084 3099

Total

3

9

Statistical analyses identified three geochemical groups labeled alpha (N = 10), beta (N = 13) and gamma (N = 15). A plot of principal components 1–3, which accounts for 77% of the total variation within the data set, best illustrates the differences between the groups (Figure 8.1). A jack-knifed Mahalanobis distance test supports the identification of these three groups. principal components 1–8, accounting for 93.3% of the cumulative variation, were used in the Mahalanobis distance test, a necessary procedure given the small group size6 (Appendix Table 8.2). Group alpha separates clearly from both beta and gamma. This is, in large part, due to the low levels of Ca, Mn, Cr and Ni in alpha and elevated levels of K and Rb. The averages for these elements are much closer for beta and gamma, resulting in more overlap between the latter two groups. Additional samples could not be placed in one of these groups and were classified as no group (N = 5). It is not uncommon for statistical analysis to leave as much as 20% of a data set unclassified.7 There are multiple explanations 6 7

Ware 3 Ubaid Standard Ware

Ware 4 Ubaid Fine Ware

1802 2311

2014 2399

2079 2625

2531 2553 2603 2604 2666 2880 2978 3085 3100 3244 3275

2370 2765 2878 3078 3193 3254

2761 3091 3205

18

8

5

386 390

920 1063 1307

for the ungrouped ceramics: (1) ungrouped samples may be the result of raw materials under-represented in the analysis and so could not form a statistically robust geochemical group; (2) they may be statistical anomalies; (3) or there may have been contamination. As future analysis increases the data set it may be possible to define additional geochemical groups and, by doing so, classify these ungrouped samples.

(Appendix Table 8.1). Multivariate statistical analyses were conducted on these results following established analytical methods which include the use of hierarchical clustering and principal component analysis (PCA) to classify geochemical groups and Mahalanobis distance to evaluate group membership.5

5

Ware 2 Ubaid Standard Ware

4 Data Interpretation The three geochemical groups correlate well with the macroscopically defined wares. The Coarse Red Ware samples are in geochemical group alpha; the only exception is DOS2010-682 which is an outlier. Group alpha includes two sherds which are not described as Coarse Red Ware. DOS2010-2370 was formally identified as Ware 3, one of the Ubaid standard wares. Its presence in alpha is puzzling but possibly due to misidentification since it contains a heavy patina. DOS2010-3079 was not attributed to any of the recognized wares but exhibits the same coarse and poorly fired appearance typical of Coarse Red Ware. There is a fairly basic geochemical difference between alpha, on the one hand, and groups gamma and beta on the other, which is consistent with previously observed differences between Mesopotamian and nonMesopotamian pottery production locales. J. Oates et al. for example, note mineralogical differences, that is, the

Glascock and Neff 2003; Glascock et al. 2004; Glascock, in press. Glascock, in press. Wallis et al. 2010.

200

Peter Magee and Steven Karacic: Geochemical Analysis of Putative Local and Ubaid Ceramics

Figure 8.1. Variance-covariance matrix PCA plot of principal components 1–3 for all samples analyzed with ICP. Ellipses are drawn to 90% confidence intervals. Red = alpha, green = beta, blue = gamma. Arrows identify samples that do not belong to a group or are addressed in the text.

There are also differences in geochemical internal variation between alpha and groups gamma and beta. For nearly every element in the analysis, the coefficient of variation (standard deviation/average*100) is significantly higher in alpha than in beta and gamma (Figure 8.2). The relative homogeneity of beta and gamma compared to alpha may be interpreted as reflecting either a more limited geological catchment area and/or a more established ‘recipe’ that is intensively produced over time.

presence of Ca plagioclase and absence of K feldspars in Ubaid ceramics and the opposite in Coarse Red Ware.8 More recently, the comparison of INAA data from over 600 ceramic vessels from Mesopotamia, the Susiana Plain and the Jazira found that high levels of Cr and Mn distinguish the pottery of southern Mesopotamia from its neighbours.9 Thus in answer to the first research question: different origins are suggested for the Ubaid pottery and Coarse Red Ware. The sheer abundance of Coarse Red Ware at Dosariyah circumstantially supports an Arabian origin while the discovery of kilns, wasters and Ubaid pottery at sites in southern Mesopotamia is compelling evidence for locating the source for Wares 2–4 at the head of the Gulf. 8 9

The separation of the Ubaid ware into two geochemical groups is of particular importance. Beta is comprised entirely of Wares 2 and 3, the standard Ubaid wares found at the site. Wares 2 and 3 are also found in group gamma along with four of the five painted and unpainted samples of Ware 4 (Ubaid Fine Ware). If

J. Oates et al. 1977. Emberling and Minc 2016: fig. 2.

201

Dosariyah

Figure 8.2. Coefficients of variation in elemental abundances between groups alpha, beta and gamma.

It is clear that there is very little evidence which would support the existence of long-term centralized production systems in the Ubaid period and instead pottery production seems very decentralized. Mediumsize settlements such as Tell Abada, for example, contained multiple kilns ranging from simple pit kilns to double-chambered up-draught kilns.11 Kilns are also known from Tell Songor and Tell el-‘Oueili.12 Settlements like Tell al-‘Ubaid and Eridu also had their own production systems, for which extensive remains have been documented.13 Furthermore, J. Oates’s analysis of material from Ur and Eridu noted formal (shape and decoration) differences in the pottery repertoire between these two sites.14 Even though all these production centers are located in the southern Mesopotamian alluvium, in theory each could produce a slightly different geochemistry which reflected variation in the cultural recipe passed down as part of the workshop tradition and the effects of differing transportation of sediment in the river system. While efforts are currently being undertaken to fix geochemical characterizations to geographic locations within Mesopotamia,15 it remains the case that the extent to which the geochemical differences between groups beta and gamma reflect spatial differences on the ground is unclear. The two groups could be from two very proximal locations which were in continual contact with each other or two quite distant settlements: the geochemical data does not permit us to address this issue.

gamma consisted solely of fine ware samples, then its existence might be attributed to the anthropogenic alteration of clay necessary to produce fine wares. This might include levigation or the reduction in the temper used for coarser thicker wares. The fact that only four samples in gamma are fine ware and the rest are the typical Ubaid coarser wares (2 and 3), however, raises the possibility that gamma may represent a different production locale. PCA was conducted on the beta and gamma samples alone in order to explore this potential difference (Figure 8.3; Appendix Table 8.3). The results reinforced the differences between these groups. The averages for each geochemical group indicate differing abundances of Ba, K, Na, Rb, Sc, Sr and V. In order to test whether or not the gamma and beta groups reflect diachronic changes in the origin of the Ubaid pottery at Dosariyah we examined the stratigraphic location of samples from each group. Both groups are distributed throughout the stratigraphic sequence: they are found in each of the layers from which material was analyzed from the very earliest to the surface. Therefore, there seems to be no difference in the origin(s) of Ubaid pottery through the occupation of Dosariyah. 5 Archeological Interpretation The analysis confirms the long-held belief that Coarse Red Ware and Ubaid ceramics come from different production locales, most likely Mesopotamia and eastern Arabia respectively.10 Perhaps more importantly, it indicates the existence of two different Ubaid geochemical groups. The assessment of two potentially different Ubaid production locales must be contextualized within our understanding of Ubaid ceramic production which, unfortunately, is still scant.

Nevertheless, it is possible that the existence of two geochemical Ubaid groups from Dosariyah reflects the engagement of two different southern Mesopotamian 11 12 13

10

14

Masry 1974; Oates J. et al. 1977; Drechsler 2011.

15

202

Jasim 1985: 208; Stein 1996: 28. Hansen-Streily 2000. Moore 2002. Oates 1960. Minc 2016.

Peter Magee and Steven Karacic: Geochemical Analysis of Putative Local and Ubaid Ceramics

Figure 8.3. Variance-covariance matrix PCA plot of principal components 1–3 for samples belonging to beta and gamma. Ellipses are drawn to 90% confidence intervals. Red = beta, blue = gamma.

production locales in Arabian Gulf exchange. Given that the two chemical groups are found throughout the sequence at Dosariyah, these two production locales would have to have been active throughout the Ubaid 2/3–4 periods when contact with Dosariyah was active. There are several candidates for such places, for example, Ur, Tell al-‘Ubaid and Eridu. Pottery from these sites was not included in this analysis and existing data from these sites cannot be brought into our dataset in a meaningful fashion.

produced in southern Mesopotamia.16 Furthermore, she argued that at least some, and possibly most, of the pottery from the Arabian Gulf was actually coming from Ur and thus the beginning of Oppenheim’s famous ‘seafaring merchants of Ur’ could be stretched back into prehistory.17 Roaf and Galbraith disagreed with Oates’s analytical protocols and interpretation and suggested that it was not possible to differentiate between Ubaid pottery made within southern Mesopotamia with the data available to Oates.18

If, as we assume, all of the pottery from groups beta and gamma is Ubaid pottery from southern Mesopotamia, then our data has important implications for a longrunning discussion about the origins of Ubaid pottery in the Gulf. In a seminal paper published in 1977, J. Oates et al. suggested that Ubaid pottery from Ur had a geochemically distinct signature from other pottery

The results of our analysis support the assertion that geochemical differences can be discerned between contemporary Ubaid pottery production centers in southern Mesopotamia. The fact that four of the five Ubaid Fine Ware sherds are included in gamma, 16 17 18

203

Oates et al. 1977. See Oppenheim 1954. Roaf and Galbraith 1994.

Dosariyah but entered the (re)distribution system closer to Kuwait. As cautioned, this idea must remain in the realm of high conjecture, especially since we know so little about the geochemistry of Ubaid pottery from southern Mesopotamia. Furthermore, projections about increased sea levels in the Arabian Gulf and its effect upon occupation in the very southern reaches of Mesopotamia have serious implications for our understanding of settlement location and exchange networks at the northern head of the Arabian Gulf.23

however, complicates the matter since all Ubaid sites produced some form of fine ware. The fact that these sherds are limited to one geochemical group could simply be a matter of sampling that would be resolved by further analysis. It is also possible that the differences between groups beta and gamma reflect the path by which Ubaid ceramics entered the Arabian Gulf. Since the time Dosariyah was first discovered, and subsequent to Oates’s claim that the pottery was coming from Ur, major discoveries have been made in Kuwait. The two most important published sites, Bahra 1/As-Sabiyah and H3/As-Sabiyah have revealed extensive evidence for Ubaid material culture. That these sites played a key role in the maritime trade to places like Dosariyah is indicated by the presence of enormous amounts of Coarse Red Ware. At H3/As-Sabiyah it accounts for 21.1% of the sherds by count19 and at Bahra 1/As-Sabiyah it accounts for 47% of the pottery.20 In addition a model boat made in Coarse Red Ware and bituminous boat remains are known from H3/As-Sabiyah.21 Thus, these sites probably operated as staging points or perhaps even entrepôts for the trip further down the Arabian Gulf. At each site, the excavators note that there are large amounts of painted fine wares22 which would be compatible with Ware 4 as identified at Dosariyah and which reflect the larger proportion of finer wares at Ubaid-related sites in the Arabian Gulf. Thicker-walled Ubaid common wares are also known.

6 Conclusions The analysis confirms that Ubaid ware and Coarse Red Ware come from different production locales. There are no indications of a diachronic shift in the origin of either type of pottery through the Dosariyah sequence but the Ubaid pottery can be divided into two geochemical groups: if these represent individual production locales, each supplied Dosariyah with Ubaid ceramics throughout the period of occupation. The most obvious location for these is southern Mesopotamia, as indicated by previous analyses and strong stylistic parallels.24 Geochemical research is ongoing on the ceramics from Dosariyah. Our future analysis focuses on the origins and technology of Coarse Red Ware. The research presented herein supports the conclusion that this ware is quite heterogeneous in terms of composition, raising the possibility of multiple production locales operating simultaneously or through time. Clay sources from eastern Saudi Arabia are included in future analysis so as to examine possible production location. In addition, archeomagnetic research on these sherds will help assess their thermal profile and in so doing illuminate the technology behind their production. When complete, the analysis will contribute to our understanding of early ceramic technology in Arabia (assuming Coarse Red Ware is local) and the myriad of cultural and economic factors that accompanied contact with Ubaid-period Mesopotamia.

It is conceivable but must still dwell in the realm of conjecture that because of their fragility the finer wares were imported into Kuwait from a relatively close Ubaid production locale. Ubaid standard wares would also have been produced in such an area and been transported to Kuwait. Settlements such as Ur, Tell al-‘Ubaid and Eridu to the north were ultimately the major players in this trade and regular traffic between these and the settlements in Kuwait must have been responsible for bringing the bulk of the pottery. In contrast, the finer wares would not have come from these more distant sources because of their fragility,

19 20 21 22

Carter 2010b: 37. Smogorzewska 2016: 602. Carter 2010d: 89–92. Carter 2010b: 53; Smogorzewska 2013: 556.

23 24

204

See Hritz et al. 2012. Oates J. et al. 1977; Roaf and Galbraith 1994.

 

0.049  0.068    0.0563 

6.59  4.24    6.286 

2.49  1.05    1.43 

%  1.14  0.98  1.02  1.48  2.55  1.06  1.08  1.45 

Na2O 

3.83  2.82    3.621 

%  4.96  3.86  3.21  2.78  4.03  3.41  3.66  3.65 

K2O 

0.841  0.575    0.7506 

%  0.861  0.775  0.705  0.686  0.88  0.745  0.783  0.655 

TiO2 

Sc 

14  9    13.2 

ppm  14  16  15  11  14  14  13  12 

Be 

2  1    2.1 

ppm  2  3  2  2  2  3  2  2 

V

133 119   145.8

ppm 99 197 137 107  131 189 135  211

Ba

447 355   442.3

ppm 543 634 302 467  522 349 480  324

Sr

318 1106   802.3

ppm 670 836 980 805  378 608 1700  622

Y

21 18   22

ppm 20 27 19 22  25 27 25  16

Zr

212 174   186

ppm 263 143 120 192  213 189 191  163

Cr

130 90   124

ppm 110 130 140 90  150 170 100  130

%  1.78  1.96  1.9  1.74  2.09  1.84  1.73  1.89  1.76  2  1.96  1.72 

%  0.56  0.673  0.604  0.601  0.688  0.686  0.671  0.688  0.679  0.765  0.736  0.723 

ppm  12  15  15  14  17  16  16  16  16  17  17  17 

ppm  1  1  1  1  2  1  1  1  1  2  1  2 

ppm 86  103 100 89 113 119 111 111  101 112 112  110

ppm 390  335 367 304 356 313 306 337  328 389 365  326

ppm 1163  1163 920 1144 867 869 1125 1062  1191 1117 1188  978

ppm 17  20 18 18 20 21 20 20  20 21 20  21

ppm 101  113 103 104 119 116 111 113  120 137 123  140

ppm 330  330 280 260 380 330 280 320  330 370 360  420

20 20   22

Zn

60 70   82

ppm 70 110 100 80  60 90 90  90

Ga

17 14   18.4

ppm 20 24 21 15  19 19 18  17

Ge

1 1   1.2

ppm 1 1 1 1  2 1 1  2

Rb

77 52   73.6

ppm 104 87 74 39  83 73 69  78

Nb

4 9   10.6

ppm 12 13 12 10  13 11 12  10

ppm 160  200 200 180 230 230 210 230  210 240 240  220

Ni

Co ppm 19  21 22 20 23 24 22 23  22 26 27  24

 

 

ppm 40  30 40 30 40 30 40 30  30 40 40  40

Cu

 

ppm 70  90 90 70 90 110 90 90  90 110 110  110

Zn

 

ppm 11  14 14 13 15 15 13 15  14 17 16  15

Ga

 

ppm 1  1 1 1 2 2 1 2  2 2 2  2

Ge

 

ppm 45  57 57 52 58 54 49 59  54 60 65  58

Rb

 

ppm 10  11 11 10 12 12 11 12  11 14 12  13

Nb

 

Cs 

ppm  2.4  3  3.3  2.8  3.2  2.9  2.9  3.6  3.2  3.5  4  3.6 

129.50 8645        Cs 

3.2  2.1    5.41 

ppm  25.2  4.4  4.2  1.6  3.5  3.3  2.9  3.7 

La 

Ce 

53.2  43.6    55.12 

ppm  59.2  71.8  58.4  44.1  58  64.6  54  44.3 

Pr

6.6 5.06   6.649

ppm 6.91 8.79 6.94 5.44  7.18 7.56 6.84  5.17

Nd

25.3 20.7   26.48

ppm 26.3 35.5 27.9 22.6  29 30 27.3  20.2

Sm

5.4 4.3   5.64

ppm 5.3 7.6 5.7 5  6.4 6.5 6  4.2

Eu

1.05 0.86   1.108

ppm 1 1.44 1.08 1.07  1.21 1.32 1.21  0.84

ppm  17.6  20.3  20.3  19.2  21.5  21.5  19  21.6  21.1  22.9  21.9  23.6 

ppm  34.4  40.2  40  37.5  41.5  43  37.4  42.9  41.1  47.3  44.6  46 

ppm 4.08  4.75 4.62 4.48 4.98 4.99 4.45 4.98  4.89 5.33 5.08  5.35

ppm 16.4  19.2 18.8 18.1 19.8 20.3 17.8 20  19.8 21.4 20.5  21.3

ppm 3.5  4 4 3.8 4.2 4.3 3.9 4.3  4.1 4.5 4.3  4.5

ppm 0.79  0.97 0.94 0.89 1 0.98 0.91 0.97  0.91 1 0.96  1.04

17.390 16.889 17.410 17.447 18.448 17.191 36756  02537  96702  87464  80663  7655                      La  Ce  Pr Nd Sm Eu

27.4  20.5    27.14 

ppm  30.5  34.8  29.2  21.3  29.2  30.4  26.7  21.4 

Gd

Tb

0.7 0.5   0.69

ppm 0.6 0.8 0.7 0.7  0.8 0.8 0.8  0.5

ppm 3  3.5 3.4 3.2 3.6 3.7 3.5 3.6  3.5 3.8 3.7  3.9

Gd

 

ppm 0.5  0.6 0.5 0.5 0.6 0.6 0.6 0.6  0.6 0.6 0.6  0.6

Tb

 

18.269 17.351 05149 

4.4 3.4   4.39

ppm 3.8 5.7 4.4 4  5.1 5.1 4.8  3.2

Dy

Ho

0.8 0.7   0.79

ppm 0.7 0.9 0.7 0.8  0.9 0.9 0.9  0.6

Er

2.5 2   2.37

ppm 2.2 2.7 2.2 2.4  2.8 2.5 2.7  1.7

Tm

0.38 0.3   0.364

ppm 0.34 0.4 0.34 0.37  0.44 0.38 0.42  0.27

Yb

2.5 2.1   2.46

ppm 2.4 2.5 2.3 2.6  3 2.5 2.8  1.9

Lu

0.42 0.33   0.397

ppm 0.4 0.41 0.36 0.4  0.5 0.39 0.46  0.3

205

5.67  5.3  5.64 

0.122  0.134  0.128 

15.72  17.93  15.74 

% 

1.78  1.35  1.88 

% 

1.29  1.37  1.01 

% 

% 

3.14  5.01  6.06  6.7  6.01 

% 

0.062  0.116  0.121  0.145  0.109 

19.27  15.53  17.9  14.84  14.54 

% 

0.74  1.1  1.39  2.14  0.99 

% 

2.7  1.96  1.41  1.2  1.92 

% 

0.135  6.12  13.47  1.62  1.66  0.127  5.57  16.93  1.53  1.1  0.118  5.24  19.81  1.29  1.03  0.135  6.13  18  1.94  0.72  0.135  5.48  15.92  1.74  1.04  0.112  5.72  17.25  1.45  1.02  0.138  5.72  12.77  1.92  1.81  0.119  5.51  15.66  1.55  1.33  0.13  6.2  17.92  1.77  0.63  0.124  6.09  17.54  1.79  0.98  0.133  5.48  16.15  1.43  1.22  0.137  6.4  12.14  1.72  1.67            0.1284 5.7513 16.196 1.6506 1.192  66667  33333  66667  66667  6.1705 6.1307 13.010 12.615 28.313 09858  49988  7172  49607  3367                                MnO  MgO  CaO  Na2O  K2O 

% 

% 

17  17  17 

ppm 

1  1  2 

ppm  123 124 126 

ppm 297 330 286 

ppm 670 719 726 

ppm 19 20 20 

ppm 112 108 113 

ppm 320 260 300 

ppm 27 26 25 

ppm 250 220 220 

ppm 30 70 40 

ppm 90 110 70 

ppm 15 15 16 

ppm 2 2 2 

ppm 27 38 21 

ppm 12 11 12 

ppm 3.1  4.1  3.7 

ppm  21.6  21.2  21.7 

ppm  42.4  41.5  43.4 

ppm 

4.98 4.87 5.01 

ppm

20.3 19.3 19.7 

ppm

4.4 4.2 4.2 

ppm

1.01 0.92 0.97 

ppm

3.8 3.6 3.6 

ppm

0.6 0.6 0.6 

ppm

4.7 4.6   4.78

ppm 2.9  3.3 3.3 3.1 3.5 3.6 3.3 3.6  3.4 3.7 3.5  3.8

Dy

 

ppm 0.6  0.7 0.7 0.6 0.7 0.7 0.7 0.7  0.7 0.7 0.7  0.8

Ho

 

ppm 1.7  2 1.9 1.9 2.2 2.1 1.9 2.2  2.1 2.2 2.1  2.3

Er

 

ppm 0.27  0.3 0.3 0.29 0.33 0.32 0.29 0.33  0.32 0.35 0.33  0.35

Tm

 

Yb

 

Lu

 

3.6 3.4 3.5 

ppm

0.7 0.7 0.7 

ppm

2.1 2.1 2.1 

ppm

0.33 0.31 0.32 

ppm

2.2 2.1 2.2 

ppm

0.9

ppm 0.7  0.8 0.7 0.7 0.8 0.8 0.7 0.8  0.8 1 0.9  0.9

Ta

 

25

0.3 0.7   0.8

Th

U

2.5 1.9   2.9

ppm 7.6 2.5 2.1 2.3  2.7 2.1 2.8  2.5

6.4

ppm 5.1  5.9 5.8 5.7 6.3 6.4 5.7 6.5  5.9 7 6.5  6.7

Th

 

1.8

ppm 1.8  2.2 1.8 1.9 2.1 2 1.8 2  1.8 2.2 1.9  2

U

 

14.357 57.769 07866  34195 

9.3 5.9   8.3

ppm 9.5 9.3 7.8 7.4  9.7 7.9 8.5  7.7

0.35 0.34 0.36 

ppm

  Lu

2.9 2.6 2.9 

ppm

  Hf

0.8 0.8 0.8 

ppm

  Ta

6.6 6.2 6.7 

ppm

  Th

2 1.9 2 

ppm

  U

0.3423 3.0769 0.8076 6.1461 1.9461 07692  23077  92308  53846  53846  8.3615 16.258 11.812 8.3655 7.7403 438  3312  33965  32718  77877 

4.2

ppm 2.5  2.7 2.8 2.7 2.9 3.1 2.5 3  3.1 3.6 3.2  3.7

Hf

 

Ta ppm 0.9 0.9 0.9 0.8  1 0.9 0.9  0.7

0.517  0.706  0.693  0.937  0.75 

% 

8  16  18  21  19 

ppm 

1  1  2  2  2 

ppm  92 112  130 161 149

ppm 327 368  289 298 327

ppm 1061 1230  660 869 459

ppm 16 20  20 23 21

ppm 119 124  121 159 127

ppm 90 300  300 860 310

ppm 11 35  25 31 30

ppm 40 230  210 280 270

ppm

20 40  30 40 40

ppm

70 100  80 120 90

ppm

11 15  17 17 17

ppm

1 2  1 2 2

ppm

51 50  36 31 70

ppm

8 12  5 14 14

ppm

2.4  2.9  3  3.7  4.9 

ppm 

18.9  22.9  24.1  22.5  24.9 

ppm 

33.2  44.5  46.1  44.5  48.3 

ppm 

4.29 5.23  5.49 5.22 5.57

ppm

17.4 21.1  21.4 21 23

ppm

3.5 4.5  4.6 4.6 4.8

ppm

0.68 1.04  1.05 1.07 1.12

ppm

2.8 3.9  4 3.9 4.2

ppm

0.5 0.6  0.6 0.6 0.7

ppm

2.8 3.8  3.8 3.9 4.2

ppm

0.6 0.8  0.8 0.8 0.8

ppm

1.8 2.3  2.2 2.4 2.5

ppm

0.28 0.34  0.32 0.37 0.38

ppm

1.9 2.3  2.2 2.5 2.5

ppm

0.31 0.36  0.35 0.41 0.42

ppm

3.2 3.3  2.7 3.8 3.5

ppm

0.6 0.9  0.2 0.9 1

ppm

6 6.4  7.5 6.7 7.9

ppm

1.9 2.2  2.2 2.3 2.3

ppm

0.795  19  1  130 334 997 21 134 460 28 270 30 90 16 2 40 13 3.2  22.1  43.6  5.11 20.5 4.4 1.06 3.8 0.6 3.8 0.8 2.3 0.35 2.4 0.4 3.5 0.9 6.5 2 0.695  17  2  130 330 529 21 117 340 27 230 40 100 16 2 28 12 4.4  22.7  44.4  5.16 21.1 4.3 1.01 3.7 0.6 3.7 0.7 2.2 0.34 2.3 0.38 3.1 0.9 6.7 2.1 0.682  17  1  124 269 1016 21 111 260 23 220 30 90 15 2 24 12 3.6  22.2  44.2  5.18 21.3 4.6 1.03 4.1 0.6 3.8 0.8 2.3 0.34 2.3 0.35 3 0.8 6.8 2.3 0.716  18  2  132 233 460 20 122 240 58 220 50 100 16 2 14 12 3.5  20.7  42.3  4.78 19 4.1 0.87 3.4 0.5 3.3 0.6 1.9 0.3 2 0.34 2.9 0.9 6.6 1.8 0.708  17  1  126 324 794 21 116 300 27 250 30 90 16 2 24 12 3.7  22.5  43.5  5.11 20.8 4.4 1.06 4 0.6 3.8 0.8 2.2 0.34 2.2 0.36 3.1 0.8 6.5 2.1 0.668  17  2  140 275 733 20 111 230 24 230 50 130 16 2 39 12 4.5  20.8  40.4  4.7 18.7 4 0.9 3.3 0.5 3.4 0.7 2 0.31 2.1 0.33 2.7 0.8 6 2.2 0.795  20  1  134  304  990  21  124  360  30  270  40  100  16  2  51  12  3.7  21  40.6  4.76  19.3  4.2  1.02  3.8  0.6  3.8  0.8  2.3  0.34  2.3  0.38  3.1  0.9  6.3  2  0.704  16  1  111 312 1075 21 116 340 24 220 40 90 15 2 31 12 3.5  22.2  44.2  5.15 21.1 4.6 1.04 4 0.6 3.8 0.8 2.2 0.34 2.3 0.35 3.1 0.9 6.8 2.4 0.749  17  2  136 268 570 21 128 400 27 240 30 90 15 2 13 12 3.5  23.1  45.2  5.34 21.8 4.6 1.11 4 0.7 3.9 0.8 2.3 0.34 2.3 0.36 3.3 0.9 6.6 3 0.709  17  1  137  280  489  20  121  290  26  210  30  100  16  2  18  13  3.9  23  44.9  5.27  21.2  4.4  1.03  3.9  0.6  3.7  0.7  2.2  0.33  2.2  0.36  3.1  0.9  6.9  1.9  0.693  16  2  115 306 651 20 112 300 26 220 30 80 15 1 30 12 3.6  21.4  42  4.91 19.9 4.2 0.97 3.6 0.6 3.6 0.7 2.1 0.33 2.2 0.36 3 0.8 6.5 2 0.774  19  2  131 309 891 22 122 370 28 260 50 100 16 2 50 13 3.4  22  43  5.08 20.6 4.6 1.08 3.9 0.6 3.9 0.7 2.2 0.33 2.3 0.37 3 0.8 6.4 1.9                                                                         0.7168 17.4  1.4666 127.93 297.13 754 20.533 117.8 318 28.4 235.33 39.333 95.333 15.6 1.9333 29.866 12.133 3.6933 21.88  43.04  5.0273 20.306 4.3466 1.0053 3.7666 0.5933 3.6666 0.7333 2.1666 0.33 2.2266 0.3593 3.02 0.8466 6.54 2.1066 66667  66667  33333  33333  33333  33333  33333  33333  33333  66667  33333  33333  33333  66667  66667  33333  66667  33333  66667  33333  66667  66667  33333  66667  66667  5.8034 6.4438 35.208 6.2004 9.5993 26.715 3.6195 6.1566 19.645 29.513 8.4783 29.565 14.221 3.2505 13.355 39.431 4.2560 10.658 3.4986 3.4381 3.7961 4.6512 4.5071 6.7449 6.1584 7.7146 5.2223 8.4165 5.4237 4.2854 4.6383 4.9856 7.2112 6.0991 3.6915 13.951 97957  15064  93951  58259  40241  30963  94251  94181  55071  92043  247  74113  29766  93287  11499  56556  25655  12038  57991  23424  02822  20951  45071  52548  63696  80476  29679  46362  19859  95644  03409  61568  91958  86372  03744  25158                                      TiO2  Sc  Be  V Ba Sr Y Zr Cr Co Ni Cu Zn Ga Ge Rb Nb Cs  La  Ce  Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Hf Ta Th U

0.682  0.68  0.703 

% 

Hf ppm 6.5 3.8 3.2 5.3  5.6 4.7 5.3  4.1

15.848 13.930 14.623 14.432 12.883 14.689 19.890 44609  44221  22783  88521  32176  93599  98632 

4.2 3.3   4.1

ppm 3.6 4.9 3.9 4.1  4.8 4.7 4.5  3

0.36

%  1.12  1.33  1.37  1  1.4  1.29  1.17  1.31  1.27  1.32  1.17  1.3 

40 40   47

Cu ppm 30 20 30 10  20 30 20  20

18.885 28.457 28.747 20.567 15.834 35.136 24.424 25.235 25746  30872  97873  65958  61839  41845  63031  7266 

14 9   13.5

Ni ppm 30 50 80 40  40 50 50  50

0.118  5.08  14.1  1.39  1.79  0.742  17  1  116 345 1503 20 176 320 22 200 30 100 15 2 56 13 2.9  23.4  45.3  5.3 21 4.5 1 3.9 0.6 3.7 0.8 2.2 0.33 2.3                       0.1156 4.9784 15.093 1.2646 1.8584 0.6781 15.769 1.2307 106.38 343.15 1099.2 19.692 121.23 331.53 22.692 211.53 35.384 93.846 14.384 1.6153 55.692 11.692 3.1769 21.069 41.630 4.8676 19.569 4.1461 0.9507 3.5615 0.5769 3.4384 0.7  2.0615 0.3161 2.1  92308  61538  07692  15385  61538  53846  23077  69231  46154  38462  30769  30769  07692  84615  30769  84615  61538  15385  61538  84615  30769  30769  23077  23077  76923  92308  23077  53846  69231  38462  23077  61538  38462  53846  13.010 8.3798 11.548 9.2814 6.3584 8.7984 9.3894 35.630 9.4509 8.3962 15.438 6.3509 16.696 13.253 9.7522 11.227 14.663 14.776 10.442 31.346 9.1081 10.110 13.364 8.3933 9.0210 7.8214 7.5431 7.2767 6.7818 7.2982 7.6011 7.4656 8.2478 8.2933 7.6979 8.0154 4125  96573  54521  89034  7455  33819  51978  48203  33769  8299  40871  10624  13871  52088  69997  67144  84518  84949  65135  6947  70903  33887  06747  74992  33991  96096  42378  33338  9489  77419  69501  92316  60988  24915  52321  8011                                                                                                                      MnO  MgO  CaO  Na2O  K2O  TiO2  Sc  Be  V Ba Sr Y Zr Cr Co Ni Cu Zn Ga Ge Rb Nb Cs  La  Ce  Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb

%  15.29  15.98  14.12  19.26  13.3  14.54  17.47  13.22  15.97  13.82  13.97  15.17 

Co ppm 12 16 17 12  17 13 13  12

ppm 0.29  0.33 0.33 0.3 0.35 0.35 0.31 0.36  0.35 0.38 0.36  0.38

%  4.75  4.51  4.64  4.21  4.9  5.89  5.28  5.08  4.86  5.25  5.25  5.02 

26.833 49.650 41.954 17.528 12.947 15.484 27.030 26.788 24.455 49.738 17.539 21.417 21.234 90295  94566  15677  53236  95276  55527  77201  60901  48036  11331  14682  08279  8577                                                        MgO  CaO  Na2O  K2O  TiO2  Sc  Be  V Ba Sr Y Zr Cr

4.23  7.38    5.232 

%  2.39  2.27  6.01  10.65  4.75  4.36  7.05  3.23 

CaO 

ppm 1.8  2 2 1.9 2.1 2.2 1.9 2.2  2.1 2.3 2.2  2.3

20.002 31.741 44164  0499              Fe2O3( MnO  T)  %  %  5.07  0.096  6.4  0.105  5.91  0.1  5.69  0.112  6.89  0.113  6.55  0.111  6.66  0.115  6.79  0.115  6.3  0.117  7.08  0.158  7.31  0.127  6.72  0.117 

%  2  3.42  3.54  6.14  3.23  4.24  4.56  4.26 

%  0.031  0.045  0.047  0.091  0.048  0.055  0.079  0.05 

4.13  4.37    3.989 

MgO 

MnO 

Fe2O3( T)  %  5.28  8.31  6.83  4.7  6.7  7.28  5.83  7.1 

Beta    %  %  386  2  47.7  9.43  390  2  49.23  11.91  1802  2  49.87  12.22  2014  3  46.37  10.65  2399  3  52.6  13.15  2553  2  48.87  11.67  2603  2  46.17  11.8  2878  3  49.37  12.31  2880  2  49.26  11.29  3078  3  50.63  12.97  3193  3  50.45  12.95  3208  Ind 48.89  12.06  et.  3254  3  50.27  12.17  6.6              49.206 11.890 6.4592 Avera ge  15385  76923  30769  CoVar    3.5413 8.5061 9.3668 80828  49996  58678                                    SiO2  Al2O3  Fe2O3( T)  Gamm   %  %  %  a  1307  2  50.87  12.33  6.88  2079  4  47.59  12.68  6.93  2310  Ind 51.7  13.39  7.27  et.  2311  2  50.06  12.84  7.64  2531  2  51.49  12.73  7.65  2604  2  45.97  11.9  6.87  2625  4  49.53  13.15  9.07  2666  2  51.18  12.36  6.88  2761  4  46.67  12.88  7.45  2765  3  50.75  13.21  7.82  2978  2  50.34  12.07  6.62  3085  2  48.59  11.89  7.68  3091  4  50.75  13.16  7.35  3244  2  53.2  12.3  7.01  3275  2  50.35  12.45  7.48            Avera   49.936  12.622 7.3733 ge  66667  33333  CoVar    3.9384 3.8357 8.0925 71589  21786  30904                                    SiO2  Al2O3  Fe2O3( T)  Unassi   %  %  %  gned  682  1  38.79  8.6  4.27  920  2  48.15  12.18  6.71  1063  2  49.84  12.99  7.2  3100  2  50.21  13.07  8.25  3205  4  50.97  14.23  7.86 

  Wa SiO2  Al2O3  re  Alpha    %  %  329  1  61.1  15.02  1306  1  59.17  17.56  1803  1  53.03  14.09  2370  3  56.28  11.67  2873  1  61.19  13.89  2902  1  60.72  15.15  3004  1  56.12  13.26  3079  ind 60.14  13.4  et.  3084  1  59.84  13.58  3099  1  58.48  9.82          Avera   58.607  13.744  ge  CoVar    4.5592 15.026 51671  42433                              SiO2  Al2O3 

Appendix Table 8.1. Average and coefficient of variation for elemental abundances in groups alpha, beta and gamma.

Peter Magee and Steven Karacic: Geochemical Analysis of Putative Local and Ubaid Ceramics

Appendix

Dosariyah Appendix Table 8.2. PCA conducted for all samples included in the analysis. PC1

PC2

PC3

PC4

PC5

PC6

PC7

PC8

54.16891

14.2037

8.62278

5.100834

4.167631

2.930833

2.38299

1.675437

54.16891

68.37262

76.9954

82.09623

86.26386

89.1947

91.57769

93.25312

Eigenvalues

0.349917

0.091752

0.055701

0.03295

0.026922

0.018932

0.015393

0.010823

Ni

-0.51463

0.154417

0.246632

-0.01965

-0.18027

0.030446

-0.06116

-0.12122

Variance

Cumulative Variance

Loadings Ca

Cr

Mn

Co

Cu

Ge

Mg

Sc

Zn

Nb

Fe

Na

Sr

Ta

Ti

Y

Ho

Eu

Er

Al

Tm

Lu

Yb

Dy

V

Tb

Gd

Si

Ga

Cs

La

Ce

Sm

Th

Nd

Pr

U

Ba

Be

Zr

-0.38015

-0.31558

-0.27314

-0.22207

-0.17044

-0.14322

-0.13384

-0.08707

-0.05858

-0.05276

-0.05061

-0.04675

-0.04041

-0.0254

0.012248

0.016384

0.019475

0.023973

0.024413

0.02542

0.026715

0.029039

0.029147

0.030224

0.033126

0.039627

0.041007

0.04659

0.049055

0.056706

0.064575

0.070859

0.074166

0.07556

0.077015

0.081503

0.092815

0.111111

0.115954

0.129101

-0.27253

0.212245

-0.01141

0.198229

0.127662

0.215603

0.030957

0.190177

0.089496

0.132019

0.181859

0.176135

-0.25686

0.132207

0.13749

0.101313

0.079805

0.130507

0.099322

0.148002

0.096511

0.102335

0.091406

0.116349

0.166816

0.099804

0.12485

0.060866

0.161442

0.300197

0.131158

0.141903

0.131045

0.123807

0.124039

0.129523

0.133575

0.030175

0.323231

0.093705

-0.09129

0.227462

-0.00177

0.008865

0.199348

0.03989

-0.1056

0.058065

0.107135

0.177192

-0.01054

-0.2184

0.422075

0.188722

0.01206

-0.00144

-0.00223

0.004335

-0.00634

0.016327

-0.00611

-0.00755

-0.01414

-0.01681

-0.07704

-0.00897

-0.00859

-0.01683

0.006103

0.170399

0.007698

0.006593

-0.0088

-0.04288

-0.00417

-0.00363

0.00785

0.095887

-0.1894

-0.03768

-0.06177

-0.07655

-0.20304

0.029884

0.309912

-0.04862

-0.22217

-0.01848

-0.0685

-0.06291

-0.0233

-0.05312

-0.24173

-0.08155

-0.05119

-0.13459

-0.15525

-0.1489

-0.13878

-0.00734

-0.12982

-0.10684

-0.12027

-0.15315

-0.09537

-0.15685

-0.1539

-0.02839

-0.03179

0.614601

-0.03661

-0.0624

-0.13038

-0.02316

-0.099

-0.07975

0.223314

-0.08438

-0.05643

-0.06946

206

0.051964

-0.22028

0.0091

-0.05663

-0.10423

0.240808

-0.05962

-0.09006

0.067921

0.414846

-0.07821

-0.08496

0.231834

0.582126

-0.01085

0.013889

-0.00038

-0.0092

0.008896

-0.03856

0.019607

0.018785

0.034672

-0.00622

-0.03686

-0.00087

-0.00861

-0.01096

-0.0081

0.143603

-0.01067

0.002612

0.008475

-0.0211

0.015868

-0.00217

0.130285

0.007988

-0.17343

0.065055

0.054361

-0.19514

-0.01566

0.008311

0.175637

-0.41641

0.007308

-0.00511

0.17856

0.099984

0.063537

-0.4539

0.200924

0.057578

-0.09081

0.050283

-0.0129

0.060002

-0.01878

0.058184

-0.02289

-0.06429

-0.04812

0.021217

0.110213

0.03991

0.047683

-0.04124

0.077394

-0.01031

0.06255

0.078673

0.075958

-0.02678

0.082738

0.077195

-0.0822

-0.07595

0.503457 -0.20853

-0.23423

-0.13745

-0.0875

0.026788

0.159086

0.391527

0.085039

-0.07248

0.146431

0.126692

0.115874

0.038135

-0.25086

0.215431

-0.09608

-0.09142

-0.14931

-0.11657

-0.1315

0.00817

-0.10924

-0.10675

-0.10867

-0.10197

0.270774

-0.12115

-0.10552

0.018966

-0.00794

-0.29052

-0.09053

-0.06279

-0.0796

-0.0712

-0.07808

-0.0813

-0.35302

-0.14623

0.121273 -0.1245

0.267643

0.208316

0.179964

-0.10279

0.044842

0.03883

0.049272

-0.09545

-0.06923

0.078555

-0.13192

0.044152

-0.04073

0.114071

0.035261

-0.02032

-0.07121

-0.10786

-0.02318

-0.10367

0.009137

0.053534

0.029619

-0.06656

-0.18034

-0.06376

-0.15605

-0.00596

-0.12545

-0.0108

-0.11589

-0.12783

-0.13467

-0.0388

-0.13165

-0.12574

0.144032

0.048492

0.636669

0.235483

Peter Magee and Steven Karacic: Geochemical Analysis of Putative Local and Ubaid Ceramics

Hf

Rb

K

PC1

PC2

PC3

PC4

PC5

PC6

PC7

PC8

0.132399

0.087199

-0.02461

-0.10645

0.151933

-0.1897

-0.12344

0.228328

0.242286

0.316985

-0.02313

-0.05548

0.577194

0.316002

-0.06611

-0.11632

-0.34653

-0.168

-0.10525

-0.14939

0.161662

0.084732

0.168129

0.075449

Appendix Table 8.3. PCA conducted for all samples belonging to beta and gamma.

Variance

Cumulative Variance

Eigenvalues Loadings Rb

K

Sr

Ba

Cr

Zr

Si

Hf

Ho

Tb

Zn

Y

Tm

Er

Cu

Lu

Eu

Yb

Pr

Sm

Nd

Gd

Nb

Ti

La

Dy

Ce

Al

Ta

Th

Ni

Sc

U

PC1

PC2

PC3

PC4

PC5

PC6

PC7

PC8

47.24653

19.43648

10.75763

6.39915

3.531913

2.474232

2.096087

1.878295

47.24653

66.68301

77.44064

83.83979

87.37171

89.84594

91.94203

93.82032

0.093995

0.038668

0.021402

0.012731

0.007027

0.004922

0.00417

0.003737

-0.60975

0.33732

-0.25281

0.104325

-0.02101

-0.21056

0.152112

0.07269

-0.43034

-0.36937

-0.12436

-0.03738

0.01103

0.011904

0.016209

0.01772

0.025514

0.02736

0.03156

0.031904

0.034168

0.035144

0.036651

0.037018

0.037713

0.038599

0.039091

0.039252

0.039891

0.040589

0.041684

0.042184

0.043888

0.044304

0.050585

0.056781

0.060557

0.063273

0.066477

0.066485

0.201977

0.109701

0.059954

0.224169

0.167851

0.034858

0.182299

0.10465

0.086448

0.193865

0.088286

0.113121

0.124215

0.052739

0.13222

0.100812

0.122704

0.097713

0.098648

0.094985

0.09851

0.134343

0.140358

0.104204

0.11385

0.108313

0.094904

0.144831

0.093731

0.180508

0.152005

0.035022

-0.06476

0.191433

-0.02512

0.075099

0.064488

-0.00407

0.110542

0.13364

0.09791

-0.07281

0.020978

0.060523

0.083246

-0.45014

0.03936

0.088135

0.069593

0.057038

0.070232

0.071731

0.111851

0.017235

0.040246

0.053648

0.076236

0.048432

-0.02539

0.037387

0.046922

0.039691

0.018431

0.057596

0.03078

-0.09823

-0.05833

-0.30967

-0.0966

-0.03848

-0.15109

-0.05615

-0.09957

0.247689

-0.01066

-0.05779

-0.05457

0.469675

-0.03955

-0.09061

-0.04766

-0.05678

-0.03646

-0.06401

-0.03907

-0.04476

-0.00875

-0.05053

-0.05208

-0.04458

0.069645

-0.02477

-0.02791

0.054169

0.082427

-0.18396

207

0.285002

-0.25934

0.032232

0.246125

-0.00903

0.101359

-0.03073

-0.08044

-0.01337

-0.19472

-0.02258

-0.02069

-0.0457

-0.13342

0.061156

0.018357

-0.02154

-0.06321

-0.04482

-0.06078

-0.03099

0.021448

0.083315

-0.06557

-0.00897

-0.06086

0.060392

-0.00449

-0.01998

0.145068

0.094125

-0.20332

-0.09227

0.559237

-0.05031

0.053974

0.254476

-0.03793

0.153668

-0.05959

-0.09301

-0.17711

-0.02963

-0.07235

-0.09038

0.227394

-0.06209

-0.12447

-0.05297

-0.04307

-0.02816

-0.06757

-0.05042

-0.00161

0.04641

-0.05616

-0.07751

-0.02883

-0.108

0.086676

-0.04783

-0.14669

-0.01895

-0.11249

-0.08543

0.100417

-0.21235

-0.38959

0.229325

-0.03229

0.20475

-0.02865

-0.09689

0.254997

0.036579

0.004605

-0.01798

-0.36439

0.015209

-0.03516

-0.00113

0.04391

0.01724

0.040468

0.004584

0.096539

0.079646

0.047461

0.001384

0.078308

0.083514

0.102694

0.069097

-0.02685

0.090295

-0.13396

0.120527

-0.40961

0.055626

-0.03169

0.048989

0.045062

0.118226

-0.25777

-0.12839

0.038676

-0.11188

-0.02733

-0.06583

-0.37485

-0.01073

-0.13246

-0.05603

0.023836

-0.0529

0.012298

-0.12137

0.013602

-0.1393

0.030517

-0.10717

0.034581

-0.06044

0.025443

0.020201

-0.17721

-0.14731

-0.24952

Dosariyah

Ga

Ca

Mn

Mg

Cs

Fe

V

Na

Co

Ge

Be

PC1

PC2

PC3

PC4

PC5

PC6

PC7

PC8

0.068811

0.138245

0.001548

0.046131

0.015111

-0.09406

0.122998

-0.02285

0.072466

0.078077

0.10703

0.110733

0.115673

0.137218

0.191568

0.192203

0.196732

0.238904

-0.18428

0.12814

0.078676

0.115553

0.113487

0.114918

0.094413

0.083861

0.452143

0.192145

0.029008

-0.00522

0.021821

-0.07673

-0.03468

0.024862

0.058868

-0.09795

0.180295

-0.71353

0.025224

0.020303

0.057126

0.15989

0.08274

0.108015

0.054021

0.243019

0.259325

-0.53201

208

-0.30515

0.105519

0.126293

-0.21489

0.109006

0.024775

0.366136

0.405223

-0.33813

-0.12277

-0.02925

0.069414

0.002816

-0.34792

0.008573

-0.13528

0.024966

0.333515

0.233072

0.097019

0.201919

0.10665

-0.05842

0.061677

0.173151

0.040319

-0.17941

0.336212

-0.35419

0.045493

-0.02377

-0.18334

-0.06694

-0.13385

-0.08992

-0.13087

-0.12771

0.069584

0.498509

0.055473

Chapter 9 Reworked Pottery Christine Kainert and Philipp Drechsler 1 Introduction

The function (or range of functions) of these perforated pottery discs is not clear and the contextual information does not permit a clear answer. Owing to morphological differences it seems reasonable to assume that they were used for more than just one specific task and two options are considered the most likely. The discs might have been used as net weights used in fishing, but they may also have been used for the processing of organic fibers or wool obtained from husbandry, although the use as a spindle whorl was presumably limited to rounded and centrally perforated discs as asymmetrical or rectangular shapes would have been less effective for the spinning of fibers. The processing of wool or fibers at Dosariyah may also be suggested by another artifact identified as a spindle whorl although it is unclear at this point what kind of fibers could have been processed at the site as the existence of the woolproducing sheep in Arabia during the sixth and fifth millennia BC has not yet been substantiated (DOS201231258, see Chapter 10). It is possible that some kind of hairy sheep were herded in eastern Arabia at this period but their fibers are hardly suitable for processing with spindles. Furthermore, in contrast to vegetal fibers, faunal fibers are less useful for the manufacture of nets or ropes, which are assumed to have been present at Dosariyah as part of fishing equipment. Faunal fibers become extremely flexible and stretchy when wet and can easily be ripped apart in the process and they would therefore be more suitable for dry tasks. In contrast, vegetal fibers become even more resistant when wet.2

During the analysis of the pottery of Dosariyah, some fragments drew attention because of visible traces of deliberate reworking in antiquity. A total of 23 items were identified as reworked vessels, four originally painted on the exterior and 16 pieces with secondary perforations. All belonged to the main ware group Black-on-Buff Ware (see Chapter 7). The documented artifacts fall into four formal categories and are described below: ceramic discs (N = 12), scraping tools (N  =  7), pendants (N  =  3) and miscellaneous objects (N = 1). Most find their closest formal parallels at other sites in the Gulf region but also in Mesopotamia and Iran. 2 Reworked Objects 2.1 Pottery discs Twelve reworked potsherds were identified as sherd discs (Table 9.1) of which ten were found broken. Although referred to here as discs, five appear to have been shaped into a rhomboid form, five into a rounded form, one into a rectangular shape; the last was too damaged to be certain of its original shape. Nine discs were found in trenches N1, N3 and E1, another was collected from the surface and two were identified during the study of Masry’s old collection (of which at least one was collected from the surface).1 All of the discs were shaped from broken pottery vessels by grinding or chipping. At least one half-finished disc allows for the identification of the production sequence (chaîne opératoire): initially a rough shaping of the outline was made by breaking off smaller pieces along the edges of the sherd; the final shape was then obtained by grinding and a slightly invasive retouch, and either before or after this stage a large secondary perforation was drilled through the object (DOS201231045, Figure 9.1). It can be assumed that in all cases the secondary perforation was originally located in the center of the sherd. On some pieces, the surface appears rather weathered which makes it difficult to ascertain actual traces of reworking and suggests that the discs were exposed to the atmosphere before they became embedded within the sediments they were found in. 1

One pottery disc combines the characteristics of a shaped disc with a secondary perforation in the center and a scraper with an invasively chipped edge (DOS2012-31375.19). Another round disc with chipped edges shows no signs of a secondary perforation at all (DOS2012-31045). The latter either represents a semifinished ceramic disc, or a ceramic scraper with a convex working edge and invasive retouch. Few pottery discs, either with or without secondary perforations, are known from contemporary sites in the Gulf region: one of the exceptions is H3/As Sabiyah where nine are published, of which three were shaped as such from wet clay.3 The other three discs were made from broken pieces of pottery, reworked into round 2

Masry 1997: pl. XXIV.

3

209

C. Schoch, personal communication, 2017. Crawford and Carter 2010: fig. 4.6: 9–11.

Dosariyah

Figure 9.1. Completely preserved and unfinished pottery discs manufactured by chipping and drilling (a DOS2012-31375.19; b DOS2012-31045) (drawings: F. Brodbeck, B. Höpfer).

210

Christine Kainert and Philipp Drechsler: Reworked Pottery

Figure 9.2. Broken pottery discs predominantly shaped by chipping (a DOS2012-27976; b DOS2012-21978; c DOS2012-25528) (drawings: F. Brodbeck, B. Höpfer).

211

Dosariyah

Figure 9.3. Broken pottery discs predominantly shaped by grinding (a DOS2010-1086; b DOS2010-1904; c Masry Box 3, Bag 174; d Masry Box 3, Bag 171; e DOS2012-24071) (drawings: F. Brodbeck, G. Häussler, B. Höpfer).

212

Christine Kainert and Philipp Drechsler: Reworked Pottery Table 9.1. Pottery discs reshaped from vessel sherds from Dosariyah. Find-ID

Context

Description

DOS2010-1086

N1-II

Fragment of a round-shaped disc showing a secondary perforation in the center and three notches along the edge at regular intervals; remains of a monochrome dark paint on the outer surface

DOS2010-1904

N1-III

DOS2012-21978

N3-II

DOS2012-25528

N3-inIV

DOS2012-24071

E1-III

DOS2012-27976

E1-IV

DOS2012-31045

E1-IV

DOS2012-31375.19

E1

1946

SSU m

Masry 1972, Box 3, Bag 171

-

Masry 1972, Box 3, Bag 174 Masry 1972, Box 5, Bag 223

Masry 1997, Plate XXIV -

Fragment of a round-shaped disc showing a secondary perforation in the center; surface is weathered with only minimal signs of abrasion left along the edge Fragment of a rhomboid-shaped disc with a secondary perforation in the center

Fragment of a round-shaped disc with a secondary perforation in the center and with visible traces of chipping along the edges of the outer surface Fragment of a rhomboid-shaped disc with a secondary perforation in the center; surface appears weathered Fragment of a rhomboid-shaped disc with a secondary perforation in the center; traces of chipping at one edge on the outer surface

Completely preserved round-shaped disc without a secondary perforation; traces of chipping along the edge

Completely preserved irregular rectangular-shaped disc showing a secondary perforation at the center; traces of chipping from both sides at one end; possibly used as a scraper Fragment of round-shaped disc with a secondary perforation in the center; surface appears weathered

Fragment of a rhomboid-shaped disc showing a secondary perforation in the center; traces of chipping along the edges from both sides Fragment of a rhomboid-shaped disc showing a secondary perforation in the center; possibly pentagonal; surface appears weathered

Fragment of a disc with secondary perforation; no shape identifiable; surface appears weathered

pottery discs with or without perforations have been noted at several contemporary sites in Iraq (e.g. Tell Abada,9 Ur10) and southern Iran (e.g. Tall-i Bakun in Fars province11).

discs with two having a secondary perforation in the center;4 an unperforated example retains monochrome dark paint on one side. At the nearby site of Bahra 1/ As-Sabiyah, reworked pottery discs were also recorded although they were much smaller in size.5 While only some show a secondary perforation, all have a more or less rounded shape and at least five retain remains of the original monochrome painted decoration while their size varies from almost 5 cm to less than 1 cm in diameter. Their purpose remains unclear although it is assumed that the smaller ones might have functioned as beads. The re-use of sherds in this manner is also reported from the late Ubaid midden site of al-Markh on Bahrain.6 Two sites with corresponding surface finds are also known from Qatar, namely Ras Abaruk7 and Groon Baida.8 At Ras Abaruk, two fragmentary and one complete disc were collected in Area C, with the latter showing unfinished secondary perforations from both sites which fail to meet. Similar finds of trimmed

An extraordinary piece was documented at Abu Khamis, located c.100  km to the north of Dosariyah. Masry identified a round perforated disc from an excavated context but did not provide additional information.12 During a visit to the storage rooms of the National Museum in Riyadh in 2012, however, it was possible to identify this very piece. It turned out to be a, thus far, unique type of reworked pottery fragment with no known parallels: along the reshaped edge a regular groove, running completely around the disc, was carefully carved into the material. 2.2 Scraping tools A total of seven reworked potsherds were formally classified as scraping tools (Table 9.2, Figure 9.4). All

Crawford and Carter 2010: fig. 4.6:12–14. Reiche 2010: 69, figs. 57.13–19, 64; 2011: 78, fig. 7.2–7; 2012: 70, figs. 5–6. 6 Roaf 1976: 151. 7 Oates J. 1978: pl. A. 8 Drechsler, personal observation, 2014. 4 5

Jasim 1985: 60, fig. 54. Woolley 1955: pl. 16 = U 17689a. 11 Langsdorff and McCown 1942: 69; Alizadeh 2006: 78. 12 Masry 1997: pl. XXX9.2. 9

10

213

Dosariyah Table 9.2. Scraping tools reshaped from ceramic sherds. Find-ID

Context

Description

DOS2012-22263

E1

Body fragment with traces of lateral bifacial chipping along the longest edge

DOS2012-21811

E1-II

Irregularly shaped body fragment with traces of bifacial invasive chipping along one of the shorter straight edges

DOS2012-29033.1

E1-VI

DOS2012-29985

E1-II

DOS2011-9385

S2-III

DOS2011-12183

S2-III

DOS2012-26728

E1

Trapezoid-shaped body fragment with bifacial chipping at one end

Small quadrangular body fragment with traces of bifacial chipping along one edge; shows traces of secondary burning along the chipped edge Large and thick elongated triangular body fragment with irregular chippings at the short edge from the exterior toward the interior of the vessel Triangular-shaped body fragment with traces of bifacial chipping at one tip

Large irregular semi-circular body fragment with traces of an invasive bifacial retouch along two opposite edges, but most prominently along the outer convex edge above an unretouched triangular handle. Fragile and thin working edge more suitable for cutting than for scraping

Figure 9.4. Scrapers made from broken pieces of pottery by bifacial knapping (a DOS2012-29033.1; b DOS2011-12183; c DOS2012-21811) (drawings: F. Brodbeck, B. Höpfer).

214

Christine Kainert and Philipp Drechsler: Reworked Pottery were found in trenches E1 and S2. Although of different sizes and shapes, they are characterized by a straight or convex retouched working edge. One piece also shows traces of secondary grinding along the reworked edge (DOS2012-29033).

Mesopotamia where they are commonly referred to as Hajji Mohammad style and defined by Joan Oates as a chronological marker for the Ubaid 2 phase.14 This is considerably earlier than the occupation layers at Dosariyah (see Chapters 6–7). There is only one other sherd from Dosariyah that bears that specific type of decoration (DOS2010-8141.1). The third pendant bears a simple motif of two straight lines (DOS2012-31244).

The original purpose of these items is unclear but on the basis of morphological characteristics it is reasonable to assume that they were used for some kind of scraping or cutting. The retouched edge is either chipped from the inner toward the outer face of the potsherd, resulting in a rather steep-angled retouch very similar to the working edge of a scraper. In contrast, a retouch obtained by bi-directional flaking results in a flat angle, broadly resembling the working edge of bifacial knives or tile knives made from flint. Similar finds are mentioned occasionally for sites in Mesopotamia and Iran, but rarely discussed in detail.13

The disparity in dating can be explained in several ways. It seems unlikely that old pots were brought into the central Gulf but it is possible that some were curated for very long periods outside their places of manufacture. Alternatively, this type of decoration might represent a tradition which lasted beyond Ubaid 2, or it is possible that certain sherds of older pots, like this pendant, were transported into the central Gulf area in an already re‑worked state and served here as an item of personal adornment (see Chapter 15).

The reasons for the reuse of these potsherds as raw material are also unclear: in general pottery is softer than flint but high-fired Black-on-Buff Ware can reach a similar hardness. In addition, pieces of flint raw material suitable especially for the manufacture of large tool blanks are rare at Dosariyah and pieces of highly fired pottery might have represented an appropriate substitute for stone.

Comparable pottery pendants from contemporary sites are rarely described but a few are reported from excavations in Mesopotamia, including Tell Abada15and Tell el-‘Oueili.16 Other pottery objects described as pendants are known from several Mesopotamian sites, but those pieces were actually made from scratch and not reworked from broken vessel pieces. Most recently, three similar objects reworked from a potsherd were documented in H3/As-Sabiyah.17 Two of these show remains of a painted decoration and all have traces of secondary perforations.

2.3 Pottery pendants Three pieces of reworked pottery were classified as pendants (Table 9.3, Figure 9.5). They were found in trenches E1 and S2 and share several characteristics. All were made from fine and thin-walled Black-onBuff Ware vessels (Ware 2d; see Chapter 7) originally decorated with a monochrome dark paint with simple geometric patterns on the exterior. Two pieces each show two secondary perforations, of which one is fractured in both cases. It can be assumed therefore that these perforations represent mending holes for the pots when they broke in antiquity. For some reason it was decided to carry out further reshaping, either because of repeated breakage of the vessel or perhaps because the original repair was unsuccessful.

2.4 Miscellaneous One pottery fragment was identified as being reworked but could not be classified any further: the body sherd shows clear signs of grinding on the exterior and along the edges (DOS2012-20678, Figure 9.6). Most plausibly, it was used as an abrasive tool (e.g. a file or polisher, see Chapter 11 for a large limestone block with similar wear pattern) used for shaping, sharpening or polishing. 3 Spatial Distribution of Reworked Pottery A total of 19 pieces of reworked pottery were documented from stratified contexts, allowing for a study of the spatial distribution of those objects. While reworked pottery occurs in almost all excavated areas, pottery discs were not found in the trenches located in the southern part of the site. By contrast, scrapers made from pieces of pottery were recovered both in the south and in the east and the latter distribution suggests an association between this group of tools and some kind of domestic activities as the best evidence of

All three finds show traces of careful but extensive grinding along the edges to bring the items into a specific rounded shape. One shows a wavy double-S shape along the right edge whereas its left side remained unworked, giving the impression that the process of reworking was not finished (DOS2012-26529). Another shows a combination of a simple geometric painted design and an incised wavy line (DOS2011-12573). This type of decoration is known from sites in southern

14

See e.g. Langsdorff and McCown 1942: 70; Delougaz et al. 1996: 250; Wilkinson et al. 1996: fig. 13.7; Fazeli Nashali et al. 2010: figs 30–31; Sudo 2010: 172.

13

15 16 17

215

Oates J. 1960: 35–36; Crawford 2010: 163. Jasim 1985: 67, fig. 64.a–f. Breniquet 1987: 143, pl. III.2. Crawford and Carter 2010: fig. 4.6:15–17.

Dosariyah Table 9.3. Pottery pendants reshaped from pottery sherds. Find-ID

Context

DOS2012-26529 E1-III

DOS2012-31244 E1-III

DOS2011-12573 S2-IV

Description Rim fragment of a BoBW vessel; remains of dark monochrome painted decoration on the outer surface with two thick horizontal bands and two narrow horizontal bands, with nine short diagonal slightly irregular lines in between connected by an irregular horizontal line in the center; two small secondary perforations in the upper right corner, the upper perforation is broken; the right side of the fragment shows clear signs of grinding from the rim to the lowest part of the piece in an irregularly shaped double-S form; possibly unfinished reworking? Ware 2d Body fragment of a thin-walled BoBW vessel; monochrome dark painted decoration on the outer surface with two horizontal bands; two secondary perforations, one in the center; the other, at the edge, is fragmented; pendant shows clear traces of grinding along the edge with a smooth but irregular ellipsoidal shape Ware 2d Body fragment of thin-walled BoBW vessel showing a teardrop shape; remains of dark monochrome painted decoration on the outer surface with a wavy incision on the paint; traces of grinding along the edge; secondary perforation close to the narrower top Ware 2d

Figure 9.5. Ceramic pendants reshaped from broken pieces of pottery (a DOS2011-12573; b DOS2012-26529; c DOS2012-31244) (drawings: F. Brodbeck, B. Höpfer).

216

Christine Kainert and Philipp Drechsler: Reworked Pottery

Figure 9.6. Piece of pottery reworked by grinding (DOS2012-20678) (drawing: B. Höpfer).

Figure 9.7. Spatial distribution of reworked pottery.

217

Dosariyah

Figure 9.8. Stratigraphic distribution of reworked pottery in trench E1. Trench E1: north section.

distinct occupation horizons comes from trenches E1, S1, S2 and S3 (Figure 9.7). A more specific stratigraphic pattern was not observed, however, and these pieces of reworked pottery occur throughout the stratigraphy (Figure 9.8).

affected by the general role those objects might have played in everyday life at Dosariyah: some were intended for specific domestic tasks and crafts (ceramic discs, scraping tools) whereas others can be considered as a form of non-utilitarian personal adornment (pendants). While the actual process of reworking was similar for both categories, their subsequent role in people’s lives was very different.

4 Conclusions Considering the total amount of pottery sherds from Dosariyah (N = 15,410), the number of 23 (0.2%) reworked vessel fragments appears very small. Nevertheless, the identification of similar finds from other contemporary sites of the Gulf region suggests that the reworking of pottery from broken vessels was not completely unusual in that area during the late sixth and early fifth millennia BC.

It seems plausible that the reuse of broken sherds was part of the widespread local tendency to make ad hoc tools, in this case from already broken sherds of hardfired pottery when access to large pieces of flint raw material may have been limited or the relative hardness of the pottery recognized. On the other hand, it is not certain that this reworking took place at Dosariyah, leaving open the possibility that at least some of these artifacts might have been transported to the site in an already reworked state.

Although the number of reworked pieces of pottery is comparatively low, two different spheres are

218

Chapter 10 Fired Clay Objects Christine Kainert

1 Introduction

2.1.1 Boat model

Some artifacts recovered from Dosariyah were fashioned out of fired clay but were not identified as parts of vessels or else they fall into the category of reworked pottery; they were therefore classified as fired clay objects. Some appeared to be similar to typical pottery sherds as they had identical fabrics but during the documentation process it turned out that they were not. Other pieces caught our attention due to their unusual form and similarity to finds known from other sites of the same period.

One fired clay fragment turned out to be part of a boat model (DOS2011-11654, Figure 10.1).1 It is assumed to represent the front part of a nutshell-like shaped object of BoBW with a slightly weathered surface. Its pale greenish color and compact dense texture point to a very high firing temperature. It bears the remains of a dark monochrome paint along both sides of the bow in form of a simple and slightly irregular straight line on top of the edge. At the peak of the bow there are two perforations on top of each other; the upper perforation is broken. They were made intentionally while the clay was still wet and pierced from the outside toward the inside and might have functioned to hold a rope or something similar to keep up a mast or a sail.2

All fired clay objects can be assigned to one of the two main types of pottery fabric known from Dosariyah owing to their general appearance and characteristics of color, temper and texture: seven pieces resemble Black-on-Buff Ware (BoBW) and twelve pieces were made from Coarse Ware (CW). No additional macroscopic subdivision was possible as this would have necessitated making fresh breaks to these finds. The artifacts described below originate both from the new fieldwork and the re-analysis of Masry’s old collection. Only some finds could be compared to specific types of objects that are already known from contemporaneous sites in neighboring regions. Whenever possible, those similarities are mentioned and conclusions concerning their potential use were drawn.

So far, 14 similar boat models are known from contemporaneous sites. They show a wide geographic distribution and range chronologically from Ubaid 1 to Ubaid 5. The northernmost known fragments come from Tell Mashnaqa3 and Tell Zeidan,4 both in Syria. Two others were identified in Tell ‘Abada in the Hamrin basin of east central Iraq,5 as well as one each from Tell al-‘Ubaid,6 Uruk/Warka7 and Tell Uqair8 in southern Iraq. Three are recorded from Abu Shahrain/Eridu9 as well as another three from Tell el-‘Oueili.10 Most recently a complete boat model was recovered from H3/As-Sabiyah,11 which at the same time represents the only one that does not belong to Black-on-Buff Ware, but was instead made of Coarse Ware. Unfortunately, none has been found in contexts that might explain their intended use. Carter speculates that they may simply have been toys but they might also have had some kind of symbolic function, protecting those who traveled by boat.12

2 Fired Clay Objects 2.1 Black-on-Buff Ware (BoBW) The broad morphological spectrum of the seven BoBW artifacts listed in Table 10.1 suggests a wider range of possible functions although it is not clear whether those tasks were carried out at Dosariyah or whether the objects were brought to the site as objects per se. In spite of the fact that the BoBW objects show a high morphological diversity, the actual count of artifacts is rather low. This indicates that fired clay objects — other than usual pottery — were uncommon and probably did not play a significant role in everyday life. Apart for the fired clay sickle fragment described below which showed a small fresh break probably caused during excavation, it was not possible to determine a specific fabric of Black-on-Buff Ware (2a, 2b, 2c or 2d; see Chapter 7).

1 Thanks to Robert Carter who identified the boat model fragment during a visit in Jubail in spring 2012. 2 Carter 2012: 350. 3 Thuesen 2000: fig. 5. 4 Stein 2011: fig. 5. 5 Jasim 1985: 66, fig. 63a–b. 6 Hall and Woolley 1927: pl. XLVIII. 7 Lenzen 1968: pl. 23h. 8 Lloyd et al. 1943: pl. XVIII.13. 9 Safar et al. 1981: 230, fig. 111; Carter 2012: 349. 10 Breniquet 1996: 166, pl. XXVIII.2; 1987: pl. III. 11 Carter 2010d: 89–91. 12 Carter 2010d: 90.

219

Dosariyah

Figure 10.1. Boat model fragment (drawing: F. Brodbeck).

2.1.2 Fired clay sickle

bitumen, both are attested on contemporary fired clay sickles from southern Mesopotamia.14

One of the most unexpected finds at Dosariyah was the fragment of a fired clay sickle (DOS2012-30807, Figure 10.3a). They are described as a typical part of the material culture for the Ubaid and subsequent Uruk periods in southern Mesopotamia with a wide distribution from sites of different sizes.13 The discovery of this fragment at Dosariyah in 2012 is the first from the Arabian Peninsula and belongs to part of the handle. It shows very fine traces of shaping on one side, whereas the other side is slightly weathered. It is of a greenish olive color and has a very dense texture, both signs of very high firing. It was found in trench E1 (stratigraphic unit E1-III), in close proximity to large unpainted and highly fragmented BoBW storage jars, and suggests this context had a domestic function.

It is not clear what the fired clay sickles found at Dosariyah and Abu Khamis were used for and only one is known from each site, despite thorough surface survey and large-scale excavations at Dosariyah. In Mesopotamia it is assumed that fired clay sickles were used for multiple tasks, including harvesting grain within fields, cutting thick aquatic plants along marshes and water channels and processing plants for secondary uses within settlements.15 After studying a sample of fired clay sickles from the Uruk mound at Abu Salabikh, Benco concluded that the use of sickles as a tool to harvest and process aquatic plants was the most likely, but does not exclude their use for other domestic tasks.16 The results of a use-wear analysis, carried out by Anderson-Gerfaud on a small sample of fired clay sickles found in Ubaid contexts at Tell el-‘Oueili, were interpreted to conclude that different forms of sickles were manufactured for different tasks as the microwear on one was thought to indicate it had been used to cut aquatic plants and another was thought to cut grain.

A second fragment of a fired clay sickle handle was found on the surface of Abu Khamis during a visit to that site by the Dosariyah team in spring 2012 (Figure 10.2). This piece is light yellowish in color, considerably larger than that from Dosariyah and shows the join of the handle and blade. It also has remains of a dark monochrome paint on both sides close to the join area: this does not resemble a specific motif and instead looks like random brushwork, but it is unclear whether this is the remains of the usual Ubaid painted decoration or of

Judging by the studies carried out by Benco and Anderson-Gerfaud and given the location of Dosariyah and Abu Khamis, it seems most likely that fired clay Thompson 1920: 135; Adams and Nissen 1972: 208; Hall and Woolley 1972: 48, pl. XV.4; Benco 1992: 123. 15 Benco 1992: 120–21. 16 Benco 1992: 131–32. 14

13

Benco 1992: 119.

220

Christine Kainert: Fired Clay Objects

Figure 10.2. Painted fired clay sickle fragment from Abu Khamis, found during a visit to the site kindly guided by A. al-Hashash, head of the Dammam Archeological Museum in spring 2012 (drawing: B. Höpfer).

yellowish clay which was modeled into a smooth cone. Most likely, this particular shape represents a common type of spindle whorl that was made from scratch and not shaped from a reworked vessel fragment (see Chapter 9). In the 1970s, two identical objects were excavated at Abu Khamis by Masry.22 Further north in Kuwait, 11 similar finds were documented at H3/AsSabiyah and considered to be spindle whorls.23 Similar finds remain absent from the lower Gulf. The pieces from Abu Khamis and H3/As-Sabiyah all show a similar form to the one from Dosariyah, and one from H3/ As-Sabiyah is also decorated with grooves.24 Another example with narrow rectangular incisions is known from Bahra 1/As-Sabiyah.25 In addition, numerous pierced fired clay objects which are interpreted as spindle whorls are known from contemporaneous sites in Mesopotamia and southern Iran.26

sickles were used for cutting thick-stemmed aquatic plants growing along the coast. As only one was found at each site, however, it is safe to assume that these tools were not an integral element in the daily routines of the people living at either place. 2.1.3 Fired clay plug A so-called fired clay plug made from BoBW was recovered in trench E1 (stratigraphic unit E1-III) (DOS2010-4800, Figure 10.3b). It is cylindrical and tapers slightly at one end where parts are missing. Its function is unclear. The closest parallels are from Bahra 1/As-Sabiyah and H3/As-Sabiyah in Kuwait Bay where nine nails/plugs of different shapes and sizes were found in assorted contexts and phases at H317 and two fine buff fired clay plugs are reported from Bahra 1/AsSabiyah.18 Furthermore, a total of 22 so-called flanged discs were recovered at H3,19 as well as several at Bahra 1/As-Sabiyah,20 mostly made from low-fired coarse red clay. Both classes of object were interpreted as labrets used as personal adornments inserted in the ears, nostrils or lips.21

The find from Dosariyah is most likely a spindle whorl, judging by its shape, yet its small size and light weight should be noted. The environmental analyses suggest that processing of fibers took place at Dosariyah or in the near vicinity but the identification of only one object as a spindle whorl suggests that spinning was not carried out on a regular basis. Furthermore, objects that would have been helpful to fulfill the purpose of spinning might have been made spontaneously, for example out of broken potsherds or organic material,

2.1.4 Pierced fired clay object A completly pierced fired clay object was recovered from trench E1 (stratigraphic unit E1-III) (DOS2012-31258, Figure 10.3c). It is 2.4 cm across, 1.2 cm long, weighs approximately 4.9  g and is made of finely tempered 17 18 19 20 21

Masry 1997: pl. XXXIX.3. Crawford and Carter 2010: 79, fig. 4.6:1–5. 24 Crawford and Carter 2010: fig. 4.6:1. 25 Reiche 2010: 78, figs 7.1, 14.a–b. 26 See e.g. Stronach 1961; Safar et al. 1981: fig.117.1–6; Lebeau 1983; Jasim 1985: figs. 50–53; Breniquet 1987: pl. IV.6–9; Sudo 2010; Schoch 2018. 22 23

Crawford and Carter 2010: fig. 4.2. Reiche 2010: figs 57.21, 63; 2011: figs 1.6, 7.8. Crawford and Carter 2010: fig. 4.1. Reiche 2010: figs 61–62; 2011: fig. 4.11–12; 2012: 70, fig. 7.1. Crawford and Carter 2010: 70,

221

Dosariyah according to need. Therefore, another possibility is that a symmetrically formed and pierced fired clay object might represent some kind of exotic item. In that case, the use for completely different purposes cannot be ruled out, for example a pierced clay object of small size and low weight might have functioned as a bead. This would correspond to a small number of other beads found at Dosariyah which were made of calcite and rock crystal (see Chapter 15). Furthermore, the manufacture of beads or pendants from a variety of raw materials and the subsequent use as a personal adornment is well attested for contemporary sites in the northern Gulf such as H3/As-Sabiyah and Bahra 1/ As-Sabiyah.27

of the BoBW models from Mesopotamia.30 On the other hand, its indentation might not be deep enough for a boat model. Furthermore, the puncture on the fractured site also points to an object that was composed of more than one piece, perhaps of different materials, or that sticks had been applied to it for some unknown reason. 2.2.2 Fired clay plug One artifact was recovered from trench S2 (stratigraphic unit S2-IV) which can best be described as representing a Coarse Ware fired clay plug (DOS2011-12384). Its high degree of fragmentation and poor preservation are the results of a relatively low firing temperature and coarse organic temper. It also shows a dark blackish core with a red surface. It is the only CW piece of this particular shape known from Dosariyah. Due to its poor state of preservation no comparative finds have been identified so far.

2.1.5 Fired clay ring During our re-examination of Masry’s collection in 2012, the fragment of a fired clay ring was identified (Figure 10.3d). No other objects of this type have been recovered during the DARP excavation at Dosariyah and its purpose is unknown. A similar find was documented at H3/As-Sabiyah but was made of stone: it was assumed to represent a large stone bead or perhaps used in some activity like fishing.28 The example from Dosariyah might have been worn as a personal adornment but owing to its size more likely functioned as a spindle whorl or a weight for domestic tasks.

2.2.3 Unidentified objects Nine shaped fired clay objects made from Coarse Ware, although rather small fragmented, are almost identical in shape (DOS2010-2188; DOS2010-2817; DOS20107120.1; DOS2012-30044.2; DOS2012-31332.42; Masry Box 3 Bag 175, Masry Box 5 Bag 192, Masry Box 5 Bag 219, plus one unnumbered example; Figure 10.5). They all show a rounded or angular corner with a rim on both sides, usually dispersing at an angle of 90° or more. Furthermore, most of the pieces are flat on one side as if they had been placed on an even surface while the clay was still wet, but one fragment appears slightly curved as well (DOS2012-31332.42). Only two originate from stratigraphic contexts, three were collected during the systematic surface survey, one was a random find and three were identified during our reexamination of Masry’s finds in 2012. One of the nine pieces shows traces of 11 small punctures on one side that were made while the clay was still wet (DOS201230044.2). All punctures have a diameter of about 1 mm and might represent a form of decoration rather than being functional.

2.2 Coarse Ware (CW) In contrast to the small but heterogeneous BoBW group discussed above, a group of 12 CW objects listed in Table 10.2 appears more homogeneous. Although often of similar morphology, their original shape and function cannot be determined as all were highly fragmented and only two objects can be more precisely identified. 2.2.1 Boat model fragment The most unexpected CW object from Dosariyah was recovered from the overburden (DOS2010-11228, Figure 10.4). It is a slightly irregular conical piece with a concave indentation on one side. The tip is slightly curved toward this indentation and is a little crooked. In the middle of the fracture, a rectangular puncture (5 x 3 mm, c.1 cm depth) is visible. The piece has a dark greyish core and a reddish abraded surface. Unfortunately, it is too small to ascertain its original form. The most likely suggestion is that it represents the bow of a CW boat model and when compared to the CW boat model from H3/As-Sabiyah,29 some similarities such as the curved tip are recognizable, a physical characteristic which has been noted previously on most 27 28 29

Another CW fragment was recovered from trench E1 (stratigraphic unit E1-II) (DOS2012-21864). This piece has some kind of spike on the top with a rim on both sides slightly bending down, and two punctures on the left, below which are six fine linear radial incised lines. The fragment itself is not straight but slightly bent, resembling DOS2012-31332.42. Therefore, it cannot be fully excluded that it represents the fragment of a vessel-like object.

Crawford and Carter 2010; Reiche 2010; 2011; 2012. Crawford and Carter 2010: fig. 4.9:14. Carter 2010d: 89–91, fig. 5.1.

30

222

Carter 2012: 350.

Christine Kainert: Fired Clay Objects Table 10.1. List of handmade fired clay objects that resemble Black-on-Buff Ware. Find-ID

Context

DOS201111654

Surface/ Over- Boat model burden fragment

DOS201230807

E1-III

Clay sickle fragment

E1-III

Plug

Clay object of cylindrical shape with slight tapering at one end

E1-III

Pierced object

Completely preserved conically formed pierced clay object

DOS20104800 DOS201231258

Box5, Bag 220

Type

Masry Collec- Clay ring tion fragment

Description Fired clay fragment identified as the bow of a boat model; with two perforations at the top very close to one another and made while the clay was still wet with one of them being fragmented; remains of a dark monochrome paint along both corners; surface appears weathered

Fragment of a clay sickle; very high fired; traces of fine shaping on one side; ware 2a

Fragment of clay ring, circular shaped

Figure 10.3. Fired clay objects made from Black-on-Buff Ware. a Sickle fragment (DOS2012-30807); b Plug (DOS2010-4800); c Spindle whorl (DOS2012-31258); d Ring (Masry Box 5, bag 220) (drawings: C. Kainert, F. Brodbeck, B. Höpfer).

223

Dosariyah

Figure 10.4. Unidentified Coarse Ware object, potentially part of a boat model (DOS2010-11228) (drawing: C. Kainert).

2.2.4 Coarse Ware objects from Abu Khamis

3 Conclusions

During a visit of our team to the site of Abu Khamis in spring 2012, a small amount of CW fragments were collected from the surface which confirmed the general picture obtained from the analysis of the pottery from Dosariyah. The pieces found at Abu Khamis confirm our initial impression that objects other than ceramic vessels were also manufactured from Coarse Ware. One particular fragment (not illustrated) found on the surface of Abu Khamis resembles an artifact from H3/As-Sabiyah:31 it is flat and slightly curved with an elongated tip but unfortunately it is quite weathered and its original function cannot be determined. Another CW piece from Abu Khamis is a rim fragment measuring 5.8 (height) x 10.0 (length) x 0.7 cm (width), which is completely straight without any kind of curvature expected for a normal pottery vessel (not illustrated). Its function is therefore unclear although it might have belonged to a square vessel resembling one made of Coarse Green Ware found at H3/As-Sabiyah.32

The general lack of contextual information makes it difficult to draw definite conclusions about the function and significance of the fired clay objects found at Dosariyah. Nevertheless, the fact that they actually existed within an Arabian Neolithic assemblage is intriguing. Utilitarian items made of Black-on-Buff Ware were very common in Mesopotamia and some types are diagnostic of the long period from the sixth to fourth millennia BC. It has been assumed by some authors that in southern Mesopotamia the inhabitants tried to compensate the insufficient access to certain raw materials such as wood, metal or stone by making copies in clay, which they had in abundance.33 Despite the growing list of objects found in Mesopotamia and now the Gulf, there are still some classes such as figurines which did not travel and we are left with the conclusion that those that did were associated with two basic activities: utilitarian tools (spindle whorls, sickles, weights) and personal adornments (beads, pendants, flanged discs, nails, plugs). On the one hand, the finds from Dosariyah described here suggest activities connected with animal husbandry and/or the processing of primary and

31 32

Crawford and Carter 2010: fig. 4.7:2. Carter 2010b: fig. 3.18:25.

33

224

Benco 1992: 119; Gibson 2010: 88.

Christine Kainert: Fired Clay Objects Table 10.2. Modelled fired clay objects made from Coarse Ware. Find-ID

Context

Type

Description

DOS2011-11228 Surface/overburden

Boat model fragment

Fragment of a clay object probably representing a CW boat model bow

DOS2011-12384 S2-IV

Clay plug

DOS2012-21864 E1-II

Clay object

DOS201230044.2

E1

Clay object

DOS201231332.42

E1 PV

Clay object

Small CW fragment; strongly weathered; possibly a plug or part of a handle

Fragment of unknown CW clay object; with a prominent spike and corners on both sides of it; with small punctures stung into the still wet clay on the left corner; four short linear incisions under the punctures in a radial arrangement; with slight bending; Ware 1a Small corner fragment of unknown CW clay object with punctures on one side; shows very slight bending Ware 1c Corner fragment of unknown CW clay object with a slight bending; Ware 1d

Corner fragment of unknown fine tempered flat CW clay object; with two corners; surface appears weathered

DOS2010-7120.1 Surface, SSU i

Clay object

DOS2010-2188

Surface, SSU i

Clay object

DOS2010-2817

Surface, SSU r

Clay object

-

Collected unsystematically from the surface during an initial visit at the site in 2008

Clay object

Masry’s Collection

Clay object

Masry’s Collection

Clay object

Corner fragment of unknown flat CW clay object; with pointed corner at 90º angle

Masry’s Collection

Clay object

Corner fragment of unknown flat CW clay object

Masry Box 3, Bag 175 Masry Box 5, Bag 192 Masry Box 5, Bag 219

Corner fragment of unknown flat CW clay object

Corner fragment of unknown flat CW clay object; thicker on one side Corner fragment of unknown flat CW clay object

Corner fragment of unknown flat CW clay object; surface appears weathered

secondary products at Dosariyah. On the other hand, their actual number is so small that this collection of objects may have been brought to the site without serving a specific domestic purpose but instead as ‘exotic’ curiosities. The boat model fragment, like the boat-related finds from H3/As-Sabiyah in Kuwait,34 are clearly connected with seafaring which was essential for bringing many of these items to Dosariyah and endorse the idea of a far-reaching maritime exchange network.35

from Dosariyah is from H3/As-Sabiyah in Kuwait. Additional finds of Coarse Ware fragments from the surface of Abu Khamis, however, cannot be associated with commonly known vessel shapes, suggesting that it was not uncommon to manufacture a variety of Coarse Ware objects. In conclusion, it is evident that a wide range of fired clay objects commonly used in Mesopotamia is also found at sites along the Gulf coast but it is uncertain whether their function was the same. The total number of these objects is so small that it does not seem plausible that they would have been used for their original purpose on a daily or regular basis, especially as these objects are durable but not unbreakable. Their function at Dosariyah therefore most likely depended on who brought them to the site and who ultimately used them. If transported and used by Mesopotamian people, they would have known their intended function, so why would they not use them for what they were initially made for? Indeed, why would they not bring even larger quantities if they were aware how helpful they might be for a stay so far from home, whether short- or long-term?

In contrast to the Black-on-Buff Ware artifacts, the fired clay objects made from Coarse Ware show a rather restricted morphological range. They prove that this fabric of clay was used not only to make pottery vessels but also other types of object, although the functions of these remain obscure. Unfortunately, the lack of research on this specific type of object as well as its friable nature makes it very difficult to provide a more comprehensive review of their use and meaning, and at present the only other published assemblage of Coarse Ware fired clay objects comparable to those 34 35

Carter 2010d. Carter 2012: 351–52.

225

Dosariyah

Figure 10.5. Unidentified Coarse Ware fragments. a Flat fragment (DOS2010-2188 SSU i); b flat fragment with lip (DOS2010-2817, SSU r); c slightly bent fragment (DOS2012-31332.42); d slightly bent fragment with star-like incision (DOS2012-21864); e flat fragment with V-shaped incision and a group of impressed dots (DOS2012-30044.2) (drawings: C. Kainert).

Another picture arises if we consider the inhabitants of Dosariyah to be people of local origin, who, at some point, had contact with people from Mesopotamia. In that case, the fired clay objects could have been items of curiosity or prestige, obtained by exchange or as gifts. Such a view would also help to explain the low number

of fired clay objects made of Black-on-Buff Ware. Within a system of sporadic or irregular contact, a constant supply was not a given and if local inhabitants were not entirely familiar with the original intended use of these objects, they may have employed them for other purposes instead.

226

Chapter 11 Lithic Industry Philipp Drechsler

1 Introduction

bifacially pressure flaked points, especially stemmed and barbed but also other forms, other bifacially worked forms (large points, ‘knives’, axe/adzes and other heavy equipment), and opportunistic flake production adapted to the vagaries of local raw material (…)’.7

The lithic industry from Dosariyah documented during the fieldwork of the Dosariyah Archeological Research Project (DARP) consists of a total of 3235 pieces. Among them, 1548 pieces were collected from the surface during systematic surface sampling,1 while 1687 stone artifacts were documented during excavations.2 The surface finds were only cursorily studied but the excavated lithic finds were studied in detail and this was undertaken during two study trips in Dammam Archeological Museum during fall 2011 and fall 2012. The remaining pieces from bulk samples and overburden were analyzed in fall 2013.

The presence of (micro-)debitage, cores and hammer stones in Dosariyah indicates flint knapping activities at the site itself. The flint technology is strongly oriented towards the production of flakes and thereby suggests an indigenous Arabian technological tradition. Blades in a technological sense are represented by only seven pieces. Their raw material — obsidian imported from Armenia and Anatolia — qualifies these pieces as imports.

With a total number of 3214 pieces, the overall frequency of flint artifacts at Dosariyah is remarkably low.3 At other Middle Neolithic sites along the shores of the Arabian Gulf, the number of recorded flint artifacts reaches up to 10,000 pieces and more (e.g. H3/As-Sabiyah, Kuwait: 12,618;4 Al-Markh, Bahrain: 6896;5 Bahra 1/As-Sabiyah, Kuwait: 48646). The paucity of flint artifacts at Dosariyah can be plausibly linked to the absence of suitable raw material sources in the wider vicinity of the site. Although limited in the overall number of artifacts, both the high proportion of characteristic flint tools and the well-dated stratified context qualify the stone tool assemblage from Dosariyah as an important reference for the central Gulf area. Major emphasis has therefore been placed on the detailed description and illustration of the artifacts.

Another characteristic element of the flint industry at Dosariyah is the use of natural blanks as pre-forms for the production of tools.8 Its repeated documentation suggests a distinct and characteristic manipulation scheme for the flint raw material: the starting point for the preparation of tools, for example modified blanks, was not necessarily a piece of debitage (flake or blade). More often, the desired tool was manufactured from shattered natural flint that already showed a suitable morphology and dimension. In accordance with this specific handling of flint, naturally broken pieces of flint that do not show traces of modification (i.e. cannot be classified as primary or secondary production) are much more frequent at the site than one would expect in a place where flint raw material is not available.

Both technological and typological characteristics of the lithic industry from Dosariyah are reminiscent of the Arabian Bifacial Tradition as defined by Christopher Edens on the basis of broadly contemporary flint artifacts collected from surface sites in the Rub alKhali that are ‘(…) polythetically characterized by small

Along the stratigraphic sequence excavated at Dosariyah, no significant technological or typological changes could be observed. This might reflect the comparatively short occupation at Dosariyah not exceeding a few centuries, but might also suggest a conservative lithic technology.9 Exceptions are the

Of which 1544 pieces are knapped stone artifacts and four are ground-stone tools. 2 Of the 1687 pieces collected during excavations, a total of 1538 pieces — including 11 ground-stone tools — could be assigned to a stratigraphic context, while 149 pieces — including six ground-stone tools — were selectively collected during the cleaning of sections or from sediment collapse during excavation. For excavation and recovery techniques see Chapter 5. 3 See Drechsler 2011. 4 Kallweit and Davies 2010. 5 Roaf 1976: 151 6 Białowarczuk 2012: 58.

Edens 1988: 38; see also Edens 1982. Natural blanks, also described as Thermoscherben or thermoclastic pieces (Kindermann 2010: 59), occur due to the physical weathering of flint nodules exposed on the surface in arid areas (McFadden et al. 2005; Eppes et al. 2010). Their use as blanks for the production of stone tools has also been observed in mid-Holocene flint artifact assemblages in the Western Desert in Egypt; see Kindermann 2010: 86. 9 The term ‘conservative’ is used here to describe a lithic industry that widely retains both typological and technological traditions over a prolonged period of time; see Kallweit and Davies 2010: 105 for a

1

7 8

227

Dosariyah

Figure 11.1. Occurrences of desert pavements potentially used as raw material sources for flint artifacts at Dosariyah.

2.1 Raw material outcrops

shapes of arrowheads that show subtle changes of shape and raw material (see Chapter 12). Likewise, there is only limited evidence of the spatial differentiation of artifacts at the site: the spatial distribution of cores and (micro-)debitage suggests that flint knapping activities occurred almost everywhere all over the site, although no distinct flint workshops were documented during excavation. Evidence of area-specific handling of flint comes from small rubbish pits excavated at the basal part of the stratigraphic sequence in trench S2 as well as from a dense concentration of natural blanks and scraping tools in trench E1. With the exception of grindstones, ground-stone tools are generally very rare at Dosariyah.

Natural outcrops of flint are repeatedly mentioned in the wider area around Dosariyah. Masry describes ‘tabular flint, associated with thick beds of greenishyellow clay as close as 15  km to the west of the site’ although without specifying the exact location of these outcrops.10 Jebel Barri is revealed as another potential source of flint, located approximately 11 km west of Dosariyah.11 Repeated visits to Jebel Barri during the field seasons at Dosariyah between 2010 and 2013, however, did not result in the identification of suitable sources of flint raw material. Nevertheless, flint-bearing sediments of the Hadrukh formation completely underlie the Jubail area and it is plausible that at least some of the flint raw material found at Dosariyah derives from local sources.

2 Raw Materials The range of raw materials used for the lithic industry of Dosariyah is extremely varied and consists of different types of flint as well as quartz and obsidian as minor components of the knapped assemblage, and metamorphic and sedimentary rocks as sources for the ground-stone tools. As there are no outcrops of such hard stones in the direct vicinity of the site, these must have been brought to the site from some distance.

A high proportion (56.6%, N  =  193) of flint tools from Dosariyah is made from natural blanks which show traces of physical weathering — so-called thermoclastic pieces.12 These pieces of flint indicate the systematic exploitation of locations where flint formed extended Masry 1997: 80. Masry 1997: 80. 12 A total of 341 flint artifacts from stratified contexts were classified as tools, see below. 10 11

similar observation at H3/As-Sabiyah.

228

Philipp Drechsler: Lithic Industry

300

1000

250 100 Log(N)

N

200 150 100

10

50 1 Type 1 Type 3 Type 2 Type 8 Type 4 Type 17 Type 5 Type 10 Type 15 Type 6 Type 9 Type 20 Type 7 Type 18 Type 16 Type 11 Type 12 Type 13 Type 19 Type 21 Type 22 Type 14 Type 23 indet.

0

Raw Material Type Figure 11.2. Frequency distribution of raw materials used for the production of knapped stone artifacts (primary and secondary production without unmodified pieces of flint shatter, N = 1028). The line shows the log-normal distribution.

desert pavements. Such occurrences were documented on the way to Thaj (at N26.965° E48.961°), about 75 km west of Dosariyah, and around Ain Dar (at N26.052° E49.518°), approximately 100 km to the south/southwest of the site (Figure 11.1), but similar phenomena might also occur in closer proximity to Dosariyah.

frequency distribution of flint raw materials used for the production of artifacts (i.e. without pieces of unmodified shatter) shows a log-normal distribution. A few flint varieties are very common, while many varieties are represented by a restricted number of pieces only (Figure 11.2).

Quarz, quartzite and other metamorphic rocks are not endemic in eastern Arabia, but have found their way from the Arabian shield by river transportation during the Late Tertiary and Quaternary (Wadi ar-Rimah/Wadi al-Batin system, Wadi Atk, Wadi al Sahba system).13 Such secondary raw material sources were extensively used by the inhabitants at H3/As Sabiya.14

About 50% of all flint artifacts were made from raw material types 1–3: a dark brown translucent flint with few inhomogeneities which occurs both in tabular and nodular form (Type 1), whitish-yellow (Type 2) and greyish–brownish (Type 3) opaque flint with numerous inclusions (Table 11.1). The dominance of these three flint types suggests a regular supply from a few raw material sources that were frequently exploited. On the other hand, a total of 12 raw material types occur as 10 or fewer pieces (with a total of 51 artifacts, accounting for 5.0% of the whole artifact assemblage), indicating that these flint raw materials were only used sporadically. As an alternative explanation, frequencies of raw material types merely reflect the general occurrence of flint types in the landscape with a natural predominance of several types which appear more frequently than others.

2.2 Raw materials for knapped stone artifacts A total of 23 flint raw material varieties were distinguished during the detailed analysis of 1527 knapped stone artifacts from excavations at Dosariyah (Table 11.1). In addition, obsidian was identified as the raw material for nine artifacts. All 23 flint varieties have been differentiated by the naked eye on the basis of color, transparency, luster and inclusions; additional criteria were the characteristics of the cortex and the overall appearance of the pieces.15 The

Although it cannot be excluded that a single flint raw material source holds more than one characteristic raw material, the high number of different raw material types suggests the exploitation of different sources. Both the absence of one dominating raw material

See Edgell 1989; 2006: 94. Kallweit and Davies 2010. 15 It should be noted, however, that this macroscopic approach for the differentiation of flint raw material is clearly biased. While some flint varieties show very distinct characteristics and therefore are well discernible, other varieties tend to be unspecific, resulting in rather broad categories. Accordingly, disproportionate numbers of pieces are assigned to the latter groups, resulting in much larger counts. 13 14

This tendency can be observed for raw material varieties type 1 to type 3 that are not only the most frequent flint types but are also rather poorly defined.

229

Dosariyah Table 11.1. Characteristics of flint raw material types. Flint raw material type

Photograph

Description

Type 1

Dark brown translucent flint with few macroscopic inclusions. Sometimes with greyish patches. Cortex chalky whitish and well-separated from flint. Pieces occur both as nodular and tabular flint

Type 2

Whitish slightly translucent flint with many yellowish or whitish and cloudy inclusions, inhomogeneous, sometimes porous. Rather poor knapping characteristics

Type 3

Wanly greyish or brownish translucent flint with numerous whitish and well-defined inclusions

Type 4

Whitish homogenous but coarse-grained, slightly translucent or translucent flint

Type 5

Greyish-brownish translucent flint with numerous, partly streaky inclusions. Inclusions consist of homogenously distributed small particles that weather into small cavities

Type 6

Wanly light brownish-greyish yellow transparent or translucent flint, homogenous with almost no inclusions

Type 7

Whitish opaque coarse grained flint raw material with many cavernous inclusions

Type 8

Greyish-brown opaque vitreous flint. Numerous globular chalky inclusions

Type 9

Brownish-yellowish translucent flint, alternating brownish and wanly greyish layers that consist of numerous globular chalky inclusions. Almost continuous transition between cortex and dense flint. Thickness of cortex up to 5 mm

Type 10

Yellowish-brownish homogenous and vitreous flint, translucent with few light inclusions

230

Philipp Drechsler: Lithic Industry

Flint raw material type

Photograph

Description

Type 11

Dark brown translucent vitreous flint with small black globular inclusions. Inclusions predominantly arranged in layers. Cortex yellowish, massive

Type 12

Greyish-brown translucent to opaque flint with many inclusions and small caverns. Raw material dull and chalky

Type 13

Opaque whitish flint with black streaks. Chalzedony?

Type 14

Slightly translucent dense flint. Major portion is dense bright brownish or yellowish striped with dark brown or blackish inclusions. Near the cortex homogenous yellowish-brown layer without inclusions

Type 15

Reddish-yellowish, highly translucent flint. Yellowish matrix with reddish streaks, cortex porous, small pits closely packed with quartz crystals

Type 16

Yellowish pale transparent – translucent flint with few reddish inclusions. Inclusions blear transparent flint matrix

Type 17

Greyish translucent or opaque flint, in parts highly homogenous, other parts of the raw material with numerous chalky inclusions

231

Dosariyah

Flint raw material type

Photograph

Description

Typ 18

Dark brownish or greyish, slightly translucent dense and homogenous flint without inclusions. Thin and well-separated cortex

Type 19

Reddish brown and grey mottled translucent flint. Greyish parts with reticular inclusions, reddish parts with small whitish and greyish spots. Cortex well-separated

Type 20

Dark brown or olive grey colored dull translucent and partly opaque flint with some bright and cloudy inclusions

Type 21

Transparent or semitransparent flint with dense, mottled inclusions. Close to the cortex many brownish inclusions that cause a darker red-brownish color

Type 22

Whitish, slightly bluish transparent flint with white globular inclusions. Close to the cortex numerous reddish streaks. Cortex porous and heavily silicified

Type 23

Dull greyish or whitish, coarse-grained flint. Breakage occurs along grain boundaries while grains itself remain predominantly intact

600 Unworked Pieces

500

Flint Artefacts

N

400 300 200 100

indet.

Type 23

Type 14

Type 22

Type 21

Type 19

Type 13

Type 12

Type 11

Type 16

Type 18

Type 7

Type 20

Type 9

Type 6

Type 15

Type 10

Type 5

Type 17

Type 4

Type 8

Type 2

Type 3

Type 1

0

Raw Material Type Figure 11.3. Frequency distribution of flint raw materials used for the production of knapped stone artifacts (N = 1028) and unworked flint pieces (N = 501).

type and the diversity of flint raw materials in general suggest that the collection of suitable pieces of raw material was embedded into a pattern of extended movements across the landscape as is characteristic of mobile societies.16

flint (Figure 11.3). Unworked pieces of raw material types 1 and 3 occur much more frequently than any other raw material types and therefore must have been brought to the site more often as raw material slabs or thermoclastic pieces for further reworking into tools.

The assumption that the most frequent raw material types 1–3 represent frequently visited raw material sources is further substantiated if one considers the frequency distribution of both worked and unworked

2.3 Sourcing the obsidian from Dosariyah17

16

A total of 10 obsidian artifacts were found at Dosariyah: nine are fragments of blades or flakes and were produced 17 Analyses carried out by Ernst Pernicka, Bernard Gratuze and Sophie Boucetta.

Binford 1979; 1980; 1982.

232

Philipp Drechsler: Lithic Industry 2.4 Raw materials for ground-stone artifacts

by knapping processes and one is the fragment of an object produced by grinding (see Chapter 15).

With a total number of 21 pieces, ground-stone artifacts are a regular part of the material culture at Dosariyah. As there are neither outcrops of hard rocks nor alluvial deposits in the direct vicinity of the site, all of the pieces were either manufactured from raw materials brought from some distance or imported as finished objects.

All nine knapped obsidian artifacts were geochemically characterized in order to determine their provenance. Studies were carried out at two independent laboratories: four artifacts retrieved during the first field season were sent to the Curt-Engelhorn-Zentrum für Archäometrie in Mannheim (Germany), where they were analyzed by Ernst Pernicka using a neutron activation analysis (Table 11.2).18 Six samples were analyzed by Bernard Gratuze and Sophie Boucetta at IRAMAT, Institut de Recherche sur les Archéomatériaux in Orléans (France) using Laser Ablation High Resolution Inductively Coupled Plasma Mass Spectrometry (Table 11.3).

The commonest type of raw material for ground-stone artifacts was siliceous sandstone (N = 5) and calcareous sandstone/conglomerate (N  =  2) (Figure 11.5) and these were mainly used for the manufacture of grinding plates and rubbing stones. The fact that all were manufactured from slabs instead of pebbles, and that they do not show traces of desert vanish indicates that these materials were obtained from primary deposits of sedimentary origin. Another common raw material was a dark greenish fine-grained volcanic rock with few macroscopic mineral inclusions (N = 6). It was used for the manufacture of an axe/adz head, but also served as the raw material for rubbing stones and a doublenotched net sinker. Some of these artifacts are made from naturally rounded pebbles, suggesting that the raw material was collected from a secondary context: potential sources are fluviatile deposits from the Late Miocene/Pliocene Hofuf formation or Pleistocene wadis draining the Arabian Shield toward the west. A similar origin can be expected for a single quartz pebble that was used as an anvil stone.

Of the four samples sent to Mannheim, three pieces of obsidian can be associated with per-alkaline obsidian sources typical of eastern Anatolia (DOS2010-257.4; DOS2010-4998; DOS2010-4999). The composition of trace elements indicates that these three samples come from Nemrut Dag and fit Blackman’s Nemrut Dag IV source.19 One sample derives from a chalk-alkaline source (DOS2010-3234). Such sources frequently occur in Middle Anatolia and Armenia but according to the distribution of trace elements, it most likely originates in Central Anatolia. The six samples studied in Orléans yield comparable results: four samples derive from peralkaline sources and can be associated with obsidian outcrops from Nemrut Dag (DOS2010-257.4; DOS201010272; DOS2011-11239; DOS2011-20478.1).20 One sample derives from Pasinler (Pasinler 1) in northeast Turkey (DOS2011-20478.2), while the sixth sample refers to obsidian sources from the Gegham Mountains in Armenia (DOS2012-29370) (Figure 11.4).

Three ground-stone artifacts were produced from hematite/magnetite (see Chapter 13). These include a small axe head and two ‘palettes’ and were fashioned from small pieces of hematite which show distinct grinding facets. A similar-shaped ‘palette’ consists of a reddish-greenish mottled fine-grained stone. Other raw materials were only used occasionally for the production of ground-stone artifacts: in two cases, limestone blocks were used as anvils and a small stone vessel was made from a reddish/pinkish magmatic rock.

The geochemical analyses of the obsidian artifacts from Dosariyah exclude an Arabian origin and instead point to their import from the north.21 It is therefore plausible that these obsidian artifacts found their way to Dosariyah together with other goods and commodities from southern Mesopotamia. The fact that all knapped obsidian artifacts show a knapping method incompatible with all other flint artifacts (see below) clearly supports this view and indicates that obsidian artifacts were imported as finished items from southern Mesopotamia.

3 Primary Production and the Study of Flint Technology A total of 1529 knapped stone tools were individually documented during excavation. According to field protocol, only pieces of flint > 2 cm or artifacts which showed clear traces of secondary modifications (i.e. tools) were measured individually and studied, whereas smaller pieces of flint were collected and documented separately as bulk samples. The following study of the primary production and flint technology refers exclusively to the sample of 1529 individually documented artifacts.

Pernicka 1992. Blackman 1984. 20 The blade fragment DOS2010-257.4, collected from the surface of the site, was studied both in Mannheim and Orléans with identical results. 21 See Khalidi et al. 2016, 752 for a more detailed description of the obsidian analyses from Dosariyah and a comparison with contemporaneous sites both in the Gulf and in Mesopotamia. 18 19

Primary production consisting of cores and debitage dominates the knapped stone artifact assemblage with 233

MA-102366

DOS2010-4998

162

DOS2010-4998

234

La

78.4

DOS2010-257.4

34.3

4.30

20

129

154

110

Nd

4.10

4.00

4.60

K% 3.65

3.78

20.2

25.0

21.1

Sm

0.59

0.54

0.66

Sc 45.0

0.55

0.74

1.09

0.98

Eu

33.1

22.4

52.0

Cr 0.44

0.71

3.80

4.10

3.70

Tb

2.38

3.55

2.42

Fe % 2.24

4.0

14.6

15

15

Yb

1.51

3.93

6.30

Co 35

0.32

2.24

2.27

2.13

Lu

178

204

187

Zn 8.1

3.18

30.6

30.2

30.6

Hf

35

39

35

As 231

4.90

4.10

4.12

4.30

Ta

224

231

234

Rb 90

30.7

28.5

35.2

29.3

Th

1130

1240

1150

Zr 0.96

16.4

9.60

11.9

10.2

U

1.70

0.84

0.72

Sb

27.7

169.5

75.5

164.6

171.4

Ce

43

28

93

95

98

93

B

2.68

6.79

18.37

16.37

16.88

17.82

Pr

29359

29893

34194

36968

34874

34562

Na

9.71

23.6

74.81

66.34

66.26

72.33

Nd

234

246

7

7

8

6

Mg

2.47

4.06

16.33

14.38

14.82

15.72

Sm

83456

81430

73076

67340

69506

72698

Al

0.15

0.08

0.39

0.36

0.35

0.4

Eu

2.33

4.38

16.35

14.29

14.44

15.84

Gd

350445 33178

348110 36167

345657 31810

346827 34675

347631 32470

K

344942 32363

Si

0.44

0.68

2.92

2.49

2.55

2.78

Tb

4140

2847

1914

1921

1932

1952

Ca

2.85

4.29

18.08

15.62

15.83

17.14

Dy

12

12

13

13

13

13

Sc

0.63

0.91

3.82

3.23

3.37

3.65

Ho

368

528

952

928

965

962

Ti

1.87

2.74

11.22

9.45

9.59

10.51

Er

586

333

422

431

432

436

Mn

0.27

0.42

1.58

1.39

1.37

1.5

Tm

3276

7626

21087

22053

21727

21848

Fe

1.96

3.15

11.36

9.94

10.38

11.1

Yb

27

35

159

186

181

165

Zn

0.29

0.49

1.67

1.41

1.44

1.57

Lu

187

171

189

203

192

193

Rb

2.5

4.8

23.07

19.69

20.42

21.74

Hf

7.75

1.99

0.1

0.21

0.1

0.09

Sr

3.87

1.69

3.4

3

3.07

3.26

Ta

18.1

26.2

101.5

86.2

90.9

96

Y

8.70

55.2

56.4

57.1

Nb

24.3

32.1

23.4

20.1

20.9

22.1

Th

53.4

143.5

13.48

10.13

7.61

7.98

7.86

7.89

U

46.6

25.3

< 310

< 200

< 170

< 110

Ba

9.38

13.88

0.62

0.7

0.65

0.72

Gegham

Pasinler

Nemrut Dag

Nemrut Dag

Nemrut Dag

Nemrut Dag

Proposed Source

7.02

5.77

6.65

6.99

6.68

6.69

Cs

Nemrut Dag

Ba

Middle Anatolia Nemrut Dag

Nemrut Dag

Proposed source

8.60

13.7

9.20

Cs

1064.8 57

911.7

946.8

998.3

Zr

Table 11.3. Element composition of six obsidian artefacts according to LAHR-ICP-MS. Values in ppm (Data source: IRAMAT Orléans).

210

307

DOS2011-20478.2 44.6

14.2

4.14

215

162.1

DOS2012-29370

3.47

4.65

4.29

Na %

Ce

71.9

73.9

DOS2011-20478.1 82

DOS2011-11239

DOS2011-10272

56

Find-ID

DOS2012-29370

35

59

62

DOS2011-20478.2

DOS2011-20478.1

DOS2011-11239

63

61

DOS2010-257.4

DOS2011-10272

Li

108

Find-ID

DOS2010-4999

19.1

109

DOS2010-257.4

DOS2010-3234

La

MA-102367

Find-ID

DOS2010-4999

MA-102365

MA-102364

DOS2010-257.4

DOS2010-3234

Lab-ID

Find-ID

Table 11.2. Element composition of four obsidian artefacts according to neutron activation analysis. Values in ppm. Na. K and Fe in wt.% (Data source: CEZ Mannheim).

Dosariyah

Philipp Drechsler: Lithic Industry

Figure 11.4. Binary diagram Ba/Zr for six obsidian artifacts from Dosariyah in comparison to known raw material sources (graph provided by B. Gratuze and S. Boucetta).

687 pieces. A high number of natural blanks (N = 501) which do not show any traces of working but had been intentionally brought to the site complement this primary production. Finally, a total of 341 pieces of flint show traces of secondary modifications. These pieces are described separately below.

Other 14% Vulcanite 29% Haematite/ magnetite 14% Limestone 10%

Calcareous sandstone/ conglomerate 9%

3.1. Assemblage characterization The knapped stone artifact assemblage consists of 576 pieces of debitage and 111 pieces classified as cores.22 The presence of both debitage and cores at Dosariyah proves the production of flint artifacts at the site. This finding is supported by numerous pieces of (micro-)debitage   2x  blank width) do occur, these pieces are not significantly different from all other blanks in their characteristic technological features.

3.3 Core characteristics A total of 111 flint cores were documented during excavation and subsequently studied in detail. While 87 of these cores are completely preserved, a total of 21 pieces were recorded as broken; in three cases, no conclusions about the status of preservation could be drawn.

The main knapping technique at Dosariyah was direct hard hammer percussion. Prominent bulbs on the ventral surface, often occurring together with impact scars on the striking platform, predominate while convincing evidence of direct soft hammer percussion using a soft hammer stone made either from

28

238

See Pelegrin 1991; 2000.

Philipp Drechsler: Lithic Industry

140 120 100 N

80 60 40 20 Cortex & Other

Asymmetric Right

Asymmetric Left

Straight

BetweenRidge

BehindRidge

0

Location of Impact Point Figure 11.10. Location of impact point in relation to the exterior shape of striking platforms (N = 379).

80 70

Width (mm)

60 50 40 30 20 10 0

0

10

20

30

40

50

60

70

80

90

100

Length (mm) Figure 11.11. Length/width scatterplot of completely preserved blanks from Dosariyah (N = 314).

239

Dosariyah

Figure 11.12. Combination of features related to the flint knapping technique (N = 366): impact point, cone, bulb, bulb scar and lipping. The opposite position of lipping and all other features in a correspondence analysis suggests a second knapping technique, most plausibly pressure retouch, as the resultant blanks are smaller than the whole assemblage (mean length lipping: 19.5 mm, mean length all other blanks: 24.0 mm. Unequal variance t-test: 3.6107, hypothesis of same mean is rejected on a 95% confidence level [p(same) mean 0.001]; analysis carried out using the software PAST 2.10, Øyvind Hammer, Natural History Museum, University of Oslo).

60 Hard Hammer Percussion

50

Pressure Retouch

N

40 30 20 10 0

0%

1-10 %

11-40 %

41-60 %

61-90 %

91-99 %

100 %

Dorsal Cortex Coverage Figure 11.13. Comparison of the dorsal cortex coverage between pieces of debitage showing indications of hard hammer percussion (N = 343), and pieces of debitage resulting from pressure retouch (N = 23). Pieces removed by hard hammer percussion generally show a higher degree of cortex coverage which suggests that this technique was used during the initial steps of core preparation.

240

Debitage

Unknown

Tabular Piece

Pebble

90 80 70 60 50 40 30 20 10 0 Nodule

N

Philipp Drechsler: Lithic Industry

Core Preform Figure 11.14. Core preforms (N = 111).

30 % Debitage 25

% Cores

%

20 15 10 5

indet.

Type 23

Type 14

Type 21

Type 11

Type 22

Type 19

Type 13

Type 12

Type 16

Type 20

Type 7

Type 18

Type 10

Type 9

Type 15

Type 6

Type 5

Type 4

Type 17

Type 8

Type 2

Type 1

Type 3

0

Raw Material Type Figure 11.15. Frequencies of flint raw material types recorded for cores (N = 111) and debitage (N = 576). Widely comparable frequencies clearly indicate flint knapping at Dosariyah.

Preforms for the majority of cores were nodules (N = 82) — voluminous pieces of flint in various shapes. In contrast, tabular pieces of flint were used less often as raw material for primary production (N  =  15). In seven cases, large blanks were transformed into cores by detaching small blanks and in one case a flint pebble from a secondary geological context was used as a core preform (Figure 11.14).

The range of flint raw material types observed for the cores closely mirrors the frequencies of flint raw material types of debitage (Figure 11.15). Types 1–3 predominate while all other types occur more rarely. A substantial cortex coverage of cores is typical (Figure 11.16). More than 50% of all completely preserved cores show retention of cortex over at least 40% of

241

Dosariyah

25 Complete

20

Incomplete Undetermined

N

15 10 5 0

0%

1-20 %

21-40 % 41-60 % 61-80 % 81-99 %

100 %

Cortex Coverage Figure 11.16. Cortex coverage of cores (N = 111).

90 80 70 60 N

50 40 30 20 10 0

Platform

Inclined

Parallel

Indet.

Core Type (after Conard et al. 2004) Figure 11.17. Core types (N = 111).

their surface.29 The finding that high proportions of the core surfaces still show cortex again suggests a low degree of core reduction and the poor knapping qualities of the flint: After the removal of a few blanks, cores were quickly discarded and in only 25.3% of cases was more than 80% of the cortex removed during blank production and/or core preparation.

only three cases. The presence of inclined and parallel cores has to be considered to be the result of highly variable, one might even argue ad hoc, core preparation and reduction processes rather than the presence of different reduction schemes. The number of striking platforms is closely related to the core type. Most platform cores have a single striking platform and constitute single platform cores (Figure 11.19). Less frequent are cores with two striking platforms which are represented by platform cores, inclined cores and parallel cores. Very few platform and inclined cores show more than two striking platforms, underlining the limited degree of

The majority of the cores found at Dosariyah are platform cores (Figures 11.17–18).30 Less common are inclined cores, while parallel cores were observed in 29 Cortex coverage is slightly reduced for incomplete cores as the breaking surfaces are by definition not covered with cortex. 30 For the classification of core types see Conard et al. 2004.

242

Philipp Drechsler: Lithic Industry

Figure 11.18. Platform cores. a Wedge-shaped core with two opposed striking platforms and one removal surface resulting in a bidirectional flake removal (DOS2010-5488); b-d single platform cores (DOS2010-9952; DOS2010-7261; DOS2012-21439) (drawings: F. Brodbeck and B. Höpfer).

option was the removal of debitage based on two striking platforms but from a single removal surface (N = 12). All other combinations of up to four striking platforms and four independent removal surfaces are also recorded, but their low numbers imply only sporadic application (Table 11.5; Figure 11.20).

core conceptualization and limited complexity of core reduction techniques employed. If more than one striking platform is present, the platform orientation can generally provide insights into distinct core reduction sequences. The sample of cores studied from Dosariyah, however, does not provide a conclusive picture: 44.2% (N = 23) of the cores show adjacent striking platforms whereas 32% (N = 17) of the striking platforms are opposed. The remaining 12 cores show either an irregular (N = 6; 11.5%) or an undetermined (N = 6; 11.5%) orientation of the striking platforms.

Table 11.5. Combination of striking platforms and removal surfaces (N = 109). Number of removal surfaces

Number of striking platforms

Equally variable is the number of removal surfaces. On cores with one striking platform, a single removal surface predominates (N  =  48), while in some cases (N  =  9) pieces of debitage were also removed from a second (spatially independent) removal surface. Comparatively frequent was the establishment of two spatially independent removal surfaces in conjunction with two striking platforms (N = 25). Another frequent

1

2

3

12

25

3

0

1

1

48

3

3

2 4

9

2

0 3

0

4 0 0

1

2

In contrast to the pronounced variability of core shapes, unidirectional core reduction clearly predominates. About 75.8% of all removal surfaces show evidence of 243

Dosariyah

60 50

Platform Cores Inclined Cores

N

40

Parallel Cores

30 20 10 0

1

2

3

4

Number of Striking Platforms Figure 11.19. Number of striking platforms (N = 109).

Figure 11.20. Combination of striking platforms and removal surfaces (N = 109).

unidirectional debitage removal, while 4.7% indicate bidirectional and 6.7% crossed debitage removal. More frequent is the multidirectional removal of debitage (12.8%). The shape and dimension of removal scars observed on all removal surfaces (N = 149) exclusively suggest the removal of flakes and no indications of blade production were observed. The average number of removal scars per removal surface is very low: about 70% of all removal surfaces show five or less removal scars, while three or four removal scars were found most frequently (Figure 11.21).

All characteristics observed on flint cores suggest ad hoc production of flakes from flake cores. The majority of the cores are platform cores with a single striking platform and one removal surface. Evidence of a specific core preparation strategy is almost completely absent. The number of striking platforms ranges between one and four. No clear spatial arrangement of striking platforms could be observed, although adjacent striking platforms predominate. Low numbers of flakes were removed from the removal surface(s) before the core was discarded. 244

Philipp Drechsler: Lithic Industry

30 25

N

20 15 10 5 0

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

Counts of Removal Scars per Removal Surface Figure 11.21. Number of removal scars counted per removal surface.

It is therefore possible to argue that most pieces of unworked flint at Dosariyah represent tool blanks which were deliberately brought to the site. These pieces were therefore classified as natural blanks. They originate from areas where flint nodules weathered out from solid bedrock and subsequent physical weathering of the surfaces through heating and cooling and dissolution by water eventually caused the nodules to disintegrate into angular debris or flint sherds.31 As a result of internal stresses occurring during these weathering process, the majority of thermoclastic pieces show many internal fractures which must have led to breakage during the manufacture of tools, and this uncontrolled breakage was responsible for the high amounts of shatter documented at the site. As a result, we conclude that although initially suitable for the production of tools, many thermoclastic pieces most probably broken during tool manufacture.

Prima facie, this situational handling of flint for the production of blanks seems to contradict the absence of flint raw material sources in the direct vicinity of the site as only sufficient supply of raw material makes such a handling reasonable. One explanation that is at least partly justifiable is the poor knapping quality of most of the available flint raw materials. The material shows many inhomogeneities and inclusions and also frequent breakages. On the other hand, many tools were not manufactured on pieces of debitage but on suitable pieces of flint raw material broken by physical weathering (thermoclastic pieces). This strategy of blank-generation simply through collection obviously rendered any investment in primary production unnecessary. 3.4 Natural blanks In addition to the debitage and cores, a total of 341 tools and 501 pieces of shatter form a major component of the flint artifact assemblage from Dosariyah. The number of pieces of shatter is remarkably high considering the absence of raw material sources within the site’s surroundings.

The use of natural blanks for the production of flint tools is fundamentally different from the manufacture of blanks based on primary production, but in a region where suitable pieces of angular debris are readily accessible on the surface and can be easily collected while moving across the landscape, their use for the production of flint tools represents a highly economic strategy. It is well attested in the Arabian Middle Neolithic for the manufacture of tabular tile knives32 but little attention has been paid until now to other tool categories. The flint artifact assemblage from Dosariyah indicates that the same general blank-tool concept was applied by different groups to a much broader spectrum of tools (see below) and the use of

The fact that these pieces of shatter were not accidentally brought to the site became obvious during excavation of trench E1 as within stratigraphic unit E1-II, more than 100 flint pieces were found within an area of c.0.7 m2. While the majority of these pieces were initially classified as pieces of shatter, namely unworked pieces of flint broken by physical weathering, it became clear during the study of the flint assemblage that most pieces were modified into scraping tools.

31 32

245

McFadden et al. 2005; Eppes et al. 2010. Kapel 1967: 21.

Other

Sicle Blade

Preform Foliate

Denticulate

Hammerstone

Borer

Splintered Piece

Projectile Point

Tile Knife

Foliate

Retouched Flake

180 160 140 120 100 80 60 40 20 0 Scraping Tool

N

Dosariyah

Formal Flint Tool Types Figure 11.22. Spectrum of formal flint tools (N = 341).

natural blanks for the manufacture of tools seems conceptually much closer to the idea of artifact shaping — the concept of façonnage — than to a well-defined primary production (concept of débitage).

number of tools in comparison to pieces of debitage, cores and natural blanks clearly illustrates the absence of local flint raw material sources around Dosariyah. Although flint knapping took place at the site, a large proportion of debitage natural blanks were turned into tools. All pieces of flint showing secondary modification were assigned to predefined and formal categories of tools. Considering the morphological variability of the tools, rather broad tool categories were defined on the basis of well-established tool typologies for the Arabian Neolithic.35

The conclusion that the use of natural blanks for the manufacture of stone tools was not the exception but rather the rule, and therefore a technological characteristic for the manufacture of flint tools in the region, might well explain the character of many Middle Neolithic flint artifact assemblages in eastern Arabia which are often cursorily described as opportunistic or ad hoc industries.33 The greater morphological diversity of natural blanks in comparison to pieces of debitage results in less homogeneous artifact assemblages. The shape and size of tools is highly variable and depends on the weathering processes affecting the flint raw materials. As a result, many flint tools appear to be less standardized. Similarly, the opportunity to use natural blanks as raw material for the manufacture of tools may result in the impoverishment of the primary production, another reason which might have led to the opportunistic appearance of many of these lithic assemblages.

The tool spectrum is dominated by different scraping tools which account for as much as 47.2% of the flint tool assemblage. Other characteristic elements are retouched flakes, bifacials, tile knives and projectile points (Figure 11.22). With these characteristics, the tool assemblage from Dosariyah clearly falls into the definitional range of the Arabian Bifacial Tradition.36 4.1 Scraping tools Scraping tools represent the main flint tool form at Dosariyah. In most cases, they are entirely preserved (N  =  132; 82%), while fragments are rare. Considering their shape, size and tool blanks, this category of tools was highly diverse and possibly served multiple tasks. A characteristic attribute of the scraping tools is a regular working edge which was mainly formed through steep retouch.37 This working edge is usually convex although

4 Secondary Production — Toward a Typology of Flint Tools With a total of 341 objects, flint tools represent 22.3% of the entire lithic assemblage.34 This exceptionally high Kallweit and Davies 2010: 123. All pieces of debitage as well as natural blanks (thermoclastic pieces) that show secondary modifications are classified as tools.

33

35

34

36 37

246

Edens 1982; 1988; Uerpmann M. 1992; Davies and Kallweit 2010. Edens 1982: 120. Hahn 1993: 223.

Philipp Drechsler: Lithic Industry as tile knives might also represent early stages in the production of other tools.

distinct scraping tools with a straight (Plate 11.5c–d) or concave working edge (Plates 11.4g, 11.5a–b) also exist. Scraping tools with a less regular working edge form part of a continuum towards denticulates,38 although distinct denticulates only rarely occur at Dosariyah (see below).

Tile knives are a common, if not typical, element of Middle Neolithic assemblages in the coastal areas of eastern Arabia but are conspicuously absent in the interior.39 They were found for example at Ain Qannas,40 Abu Khamis41 and Khursaniyah42 in the Eastern Province of Saudi Arabia, al-Markh in Bahrain,43 and Ras Abaruk, Site 4b,44 Groon Beidha45 and numerous other sites in Qatar.46

The majority of tool blanks used for the manufacture of scraping tools are natural blanks (N = 112; 96.6%; Plates 11.6–8). Less frequently, flakes were transformed into scraping tools (N  =  44; 27.3%), while in five cases the tool blank could not be reliably classified. When pieces of debitage are used as tool blanks, steep retouch regularly occurs along one lateral edge of the dorsal surface (e.g. Plate 11.1), but the ventral surfaces often also show a marginal flat retouch. More rarely, the retouch of the ventral surface covers a large part (Plate 11.2) or covers it almost completely (Plate 11.3). In very rare cases, steep retouch was applied to the short edges of elongated thick tool blanks (Plate 11.4a–b).

4.4 Projectile points With a total of 36 pieces, projectile points, commonly termed arrowheads, were frequently found at Dosariyah: 16 pieces were documented from stratigraphic contexts and the remaining 20 pieces were collected from the surface and overburden (Plates 11.13–14). The majority of these pieces from Dosariyah are stemmed bifacial points.47 They have triangular to subtriangular bodies and well-marked stems. Barbs are distinctly formed, while a few pieces are shouldered. Only one projectile point lacks a distinct stem but is nearly lozenge-shaped and double-pointed (DOS201011109). Cross sections are lenticular and presumably finished by pressure flaking across both faces. Blanks for the production of arrowheads are either flakes or thin pieces of tabular flint. Arrowheads made from tabular flint are sometimes only retouched along the edges. With these characteristics, the arrowheads from Dosariyah fall within the category of ‘bifacial short winged and tanged arrowheads type I’.48 Their main distribution is within the central Gulf, while similar, although shouldered, winged and tanged arrowheads predominate in the Rub al-Khali.49 Stemmed triangular bifacial points with barbs are also present in the upper Gulf, where they were found together with lozengeshaped, double pointed projectile points (H3/AsSabiyah)50. The latter shape occurs both in the lower Gulf in Qatar51 and the Rub al-Khali.52

4.2 Retouched flakes Retouched flakes are the second most common category of flint tools at Dosariyah (N  =  38). They are characterized by a partial or complete edge or by invasive/facial retouch which often follows the natural edge of the used blanks (Plates 11.9–10). The retouch can be applied to the dorsal or ventral surface of the flake or on both surfaces. With these poorly defined criteria, retouched flakes represent a highly heterogeneous group of flint tools. The acute-angled retouch nevertheless suggests that cutting was a major task for this group of flint tools. More than one third (N = 13; 34.2%) of all retouched flakes are preserved as fragments only, suggesting a use which often resulted in their breakage. 4.3 Tile knives A total of 20 flint tools which were manufactured from thin pieces of tabular flint (thickness  4.69 g were taken into account corresponding to the 10% percentage of the weight of tools made from natural blanks. All natural pieces of shatter below this weight are considered to be too small for a general use as tool blanks. Taking into account all pieces of shatter, the trend towards an increase of natural pieces of shatter along the stratigraphic sequence remains but is less pronounced.

and tool blanks should be considered to be a deliberate cache (Figure 11.34, marked in red in profile projection Figure 11.36).

any clear trends, yet in trench S2 scraping tools made from natural blanks were only recorded in the uppermost part of the stratigraphic sequence (Figure 11.35). This peculiarity was not observed in the other trenches where scraping tools made both from pieces of natural shatter and flakes were found within the same stratigraphic units. Tile knives were found much

Other tools and objects show different patterns. The stratigraphic distribution of scraping tools made on flakes or natural blanks generally does not show 258

Philipp Drechsler: Lithic Industry

Figure 11.33. Spatial distribution of flint tools. a Scraping tools made on flakes (dots) and natural blanks (open circles); b tile knives; c projectile points; d bifacials (diamonds) and bifacial preforms (open diamonds).

more frequently in the eastern and northern areas of the site but are less frequent in the southern trenches. They appear exclusively in the upper parts of the stratigraphic sequences, suggesting that they were introduced into the tool spectrum during the later stages of the occupation of the site. The horizontal distribution of projectile points is most likely related to different excavation techniques. These points, exclusively classified as stemmed arrowheads, were documented in much higher frequencies in those trenches which were excavated in square meter units (i.e. trenches N1, S1, S2, S3). Considering the rather

small size of these arrowheads, it seems plausible that they were overlooked during the excavation of the larger step trenches N2, N3 and E1. In contrast to this, bifacials and bifacial preforms were found in all excavated areas (Figure 11.35–37) yet their absence from trench E1 suggests that the preparation of these artifacts was reserved for selected areas. The suggestion that trench E1 cuts through a part of the site which had been used mainly as an occupation area finds support in the spatial distribution of ground-stone tools (Figure 11.38). Although grinding stones were 259

Dosariyah

Figure 11.34. Dense concentration of natural blanks and scraping tools prepared on natural blanks. Trench E1, spit 6.

Figure 11.35. Trench S2: stratigraphic distribution of a primary production (square: debitage; pentagon: core), natural blanks (circle) and tools (triangle) and b flint tool types (circle: scraping tool on flake; open circle: scraping tool on natural blank; triangle: projectile point; diamond: bifacial; open diamond: bifacial preform; square: splintered piece; semi-circle: tile knife). Trench S2, east section.

260

Philipp Drechsler: Lithic Industry

Figure 11.36. Trench E1: stratigraphic distribution of a primary production (square: debitage; pentagon: core), natural blanks (circle) and tools (triangle) and b flint tool types (circle: scraping tool on flake; open circle: scraping tool on natural blank; triangle: projectile point; diamond: bifacial; open diamond: bifacial preform; square: splintered piece; semi-circle: tile knife). Cache of scraping tools within stratigraphic unit E1-II marked by red rectangles. Trench E1, north section.

diagnostic, such as fluting or the presence of trihedral arrowheads in southern and southeastern Arabia.108 A second obstacle for the establishment of comprehensive artifact studies in Arabia is the homogeneous appearance of most lithic assemblages. Specific tool types such as arrowheads, foliates, scraping tools and splintered pieces occur in similar shapes in different parts of the Peninsula but it is unclear whether this reflects social interaction, the existence of a common cultural substratum or convergent developments resulting from technological adaptations to comparable but specific economies and environments. Finally, the situation is complicated by the fact that most ‘Neolithic’ sites in Arabia are surface scatters without stratigraphy or secure dating.

also found in trenches N1 and S1, all other elements of the ground-stone assemblage comprising hand stones, pounders, polishers, hematite billets as well as a single stone axe were recorded during excavation of trench E1. 7 Contextualization of the Lithic Industry from Dosariyah The virtual lack of detailed studies of Neolithic flint and ground-stone tool assemblages in eastern Arabia makes it difficult archeologically to contextualize the assemblage from Dosariyah. So far, relatively few attempts have been made to investigate diachronic trends and developments in lithic technology and typology at a regional level.107 The majority of studies instead limit themselves to a description of individual artifact assemblages and/or synchronous comparison and consequently the contextualization of these assemblages is vague and restricted to a few tool types or technological characteristics considered to be 107

Characteristic elements of the lithic industry from Dosariyah, however, are found at numerous sites in eastern and southeast Arabia which are associated with

108 Charpentier and Inizan 2002; Charpentier 2003; 2004; 2008; Crassard et al. 2006; Charpentier and Crassard 2013.

E.g. Kapel 1967; Uerpmann M. 1992; Spoor 1998; Charpentier 2008.

261

Dosariyah

Figure 11.37. Trench N3: stratigraphic distribution of a primary production (square: debitage; pentagon: core), natural blanks (circle) and tools (triangle) and b flint tool types (circle: scraping tool on flake; open circle: scraping tool on natural blank; triangle: projectile point; diamond: bifacial; open diamond: bifacial preform; square: splintered piece; semi-circle: tile knife). Trench N3, south section.

Figure 11.38. Spatial distribution of ground-stone tools. a Grind stones (square) and hand stones (circle); b pounders (triangle), polisher (star), stone axe (diamond) and billets (asterisk).

262

Philipp Drechsler: Lithic Industry Chris Edens’ Arabian Bifacial Tradition,109 most notably the bifacially tanged and barbed arrowheads, the bifacial points and their preforms. Other tool categories present at Dosariyah are reminiscent of lithic facies known from the shores of southeast Arabia: splintered pieces, net sinkers and hammer stones have been extensively reported at Middle Neolithic sites along the Omani coast of the Indian Ocean.110 Nevertheless, other tool types such as scrapers and borers are ubiquitous and cannot be used as chronological or geographical markers.

In general, the flint artifact assemblage from Dosariyah conforms to artifact assemblages documented from other coastal Middle Neolithic sites in the central Gulf area, both in terms of lithic technology and tool typology. It closely matches the flint assemblages from Abu Khamis114 and Ain Qannas (levels 3 and 4),115 although tool-type frequencies vary.116 The assemblage is also comparable to al-Markh (Bahrain) where the tool spectrum was dominated by various scraper forms plus tile knives, a barbed and tanged arrowhead and a bifacially flaked ovoid/foliate projectile point.117 In accordance with observations made at Dosariyah, ‘any difference between the industries of the earlier and later levels seems quantitative rather than qualitative’,118 suggesting a consistent flint technology and tool typology at al-Markh over a prolonged period of time. A similar spectrum of flint tools likewise dominated by diverse scraper forms but supplemented by (bifacial) points, various cutting and piercing tools have been recorded at several sites on the western coast of Qatar, namely al-Da’asa site 46,119 Ras Abaruk site 4120 and Groon Beidha.121

Similarly, technological traits accompanying primary production at Dosariyah have been repeatedly described from Neolithic assemblages found in different regions of the Arabian Peninsula. At the majority of these, both façonnage and débitage concepts co-occur. Primary production is generally flake-based with cores which present little, if any, evidence of formal core preparation, hence a relatively low degree of blank-shape control. Consequently the respective lithic technologies have been described either as opportunistic in the case of southeast and eastern Arabia111 or reflecting an ‘opportunistic approach to the problem of obtaining utile [advantageous] lithic blanks’ for northeastern Arabia.112 In contrast, elaborate bifacial working (Formüberarbeitung) of specific tool types (arrowheads, bifacial foliates, bifacial knives) documented at most Middle Neolithic sites in Arabia implies high flint knapping abilities.

At present, it seems that technological and typological inter-site differences increase towards the east coast of Qatar (al-Khor122) and the lower Gulf (Dalma 11,123 alMadar/site 69,124 al-Zuhra,125 Hamriyah126). In addition, primary production becomes more important in the lower Gulf,127 with a ‘slightly stronger leptolithic tendency’.128 The overall spectrum of flint tool types, however, remains constant with a predominance of scraping tools, cutting tools and piercing tools which occur together with bifacially chipped foliates and arrowheads.

At Dosariyah, arrowheads and a few bifacials are carefully shaped but the evidence for pressure flaking (retouching) remains weak. Tile knives and the majority of scraping tools were made from natural blanks. While the overall shape was provided by (suitable) pieces of raw material, only limited shaping was necessary to establish the working edges. While scraping tools should be considered as an almost universal tool type hardly suitable for chronological or cultural determinations, tile knives with an invasive oscillating retouch along the convex working edge and made on thin tabular pieces of flint represent a common and characteristic — if not diagnostic — tool category at Ubaid-related sites along the eastern Arabian shores. As similar shaped pieces also occur at sites on the coast of the Indian Ocean in southeast Arabia,113 this particular tool type is potentially related to processing of marine resources. Similarly, splintered pieces were probably used to process marine snails and/or shell.

In contrast to evident technological and typological parallels with Arabian Middle Neolithic assemblages, the flint artifact assemblage from Dosariyah shares almost no similarities with contemporary lithic assemblages from Mesopotamia or the Iranian Plateau. Masry 1997: 90. Masry 1997: 70. Worth noting is the higher frequency of small flint borers at Abu Khamis that were only rarely found at Dosariyah, indicating extensive bead production at the site. 117 Roaf 1976. 118 Roaf 1976: 151. 119 Smith 1978a: 64. 120 Smith 1978b: 86. 121 Drechsler 2014. 122 Inizan 1988. 123 Flavin and Shepherd 1994; Beech and Elders 1999; Beech et al. 2000. 124 Boucharlat et al. 1991. 125 Haerinck 1991. 126 Jasim 1996. 127 A preference for primary production at the expense of the use of natural blanks in the lower Gulf should be closely related to the differences in raw material availability (S. Hussein, personal communication), a suggestion that is outside the scope of this book but should be considered in future research. 128 Uerpmann M. and Uerpmann H.-P. 1996: 132; ‘leptolithic’ refers to the production of blades as blanks. 114 115 116

Edens 1982. Uerpmann M. 1992; Uerpmann H.-P. and Uerpmann M. 2003; Charpentier 2008. 111 Edens 1988; Uerpmann M. 1992: 77. 112 Kallweit and Davies 2010: 105. 113 See Uerpmann M. 1992: 79. 109 110

263

Dosariyah This concerns mainly the structure of primary production which is generally blade-based in the latter regions and characterized by tool blanks removed from bullet-shaped platform cores, a knapping method which developed during the Levantine and IranoTuranian Neolithic and virtually absent from Arabia.129 The only knapped stone artifacts known from Dosariyah whose technological traits are consistent with such a reduction method are a total of nine obsidian blade fragments (Figure 11.39). With their thin triangular or trapezoidal cross sections, regular parallel edges and parallel-unidirectional dorsal scars without substantial traces of cortex, these blades were clearly struck from unidirectional platform cores carefully prepared for the production of regular blades of a certain size and shape. The necessary core reduction strategies to generate such blanks are not known from the Middle Neolithic of the Arabian peninsula. This technology instead finds its closest parallels in Mesopotamia, the Zagros region and the Iranian Plateau. This finding is consistent with the identification of the raw material sources for these artifacts in eastern Anatolia and Armenia (see section 2.3). These artifacts, therefore, were either produced by non-locals visiting Dosariyah or represent proper imports of already finished artifacts. At the present state of research, the latter conclusion is more likely as neither production waste from obsidian knapping nor any other artifacts with similar technological characteristics produced from local raw materials have been found at Dosariyah.

tool types at that site, with at least some of the scrapers made of ‘natural flat pebbles and rocks’.133 The picture which emerges from the contextualization of the ground-stone tool assemblage from Dosariyah is different. Ground-stone tools, especially grindstones, hand stones, pestles and axes, have been documented at Neolithic sites in the interior of Arabia.134 They seem to occur in much higher frequencies, however, and in broader typological spectra at coastal sites where items identified as net sinkers and anvil/hammer stones can be directly related to the catching and processing of marine resources. Moreover, polishers and polishing plates might have been frequently used for the manufacture of shell beads, as indeed their appearance in special bead workshops at H3/As-Sabiyah and Bahra 1/As-Sabiyah suggests.135 The frequency of grindstones and hand stones, as well as the predominant use of local stone for their manufacture, suggest utilization embedded in other domestic activities.136 On the other hand, more elaborate ground-stone tools such as axe heads are often made from non-local raw material. Hence, at least parts of the ground-stone tool assemblages from the coastal sites in eastern Arabia should be considered as exotic imports which potentially carried non-utilitarian values.137 This conclusion is supported especially at Dosariyah, where these pieces were found in association with other imported items such as Ubaid pottery and obsidian artifacts.

Yet although the lithic assemblage from Dosariyah does not show substantial parallels with contemporaneous Mesopotamian flint technologies, similarities with flint assemblages documented from coastal Ubaid sites in the upper Gulf clearly exist. The artifact assemblage from site H3/As-Sabiyah has been related to the Arabian Bifacial Tradition, although cutting, piercing and drilling tools represent the commonest tool types found. In accordance with the observations at Dosariyah, the lithic technology at H3/As-Sabiyah does not change dramatically through time and was considered to be conservative.130 Its primary production was flake-based and described as opportunistic, lacking convincing indications of standardized core reduction schemes, and flakes were frequently used without secondary modification.131 Comparable conclusions have been drawn from the lithic assemblage at Bahra 1/As-Sabiyah where no standardized flint knapping strategies have been identified, although a tendency toward the production of small blades has been noted.132 Drilling, cutting and scraping tools are the commonest

8 Conclusions With a total of 3235 pieces, stone artifacts represent the second most frequently documented category of material culture at Dosariyah (the commonest being the pottery, see Chapter 7). Nevertheless, the overall frequency of stone artifacts is remarkably low in comparison to other broadly contemporaneous sites along the shores of the Arabian Gulf and can be plausibly related to site function, and also to the absence of raw material sources in the direct vicinity of the site. The relative scarcity of suitable flint raw material is expressed in the high proportion of flint tools in the assemblage (N = 341, 22.3%) relative to cores (N = 111, 7.3%) and pieces of debitage (N  =  576, 37.7%). This indicates extensive exploitation of the available raw material. The contrasting high number of unworked natural flint pieces (N  =  501, 32.7% of all pieces of flint) can be interpreted as the result of rather poor raw material knapping qualities and a specialized Białowarczuk 2012: 61. Sordinas 1978: 24, 25, pl. 16; Kiesewetter 2006; Guagnin et al. 2017. Kallweit and Davies 2010; Białowarczuk 2012. 136 See Inizan 1988. 137 See Drechsler et al. 2013 for a similar conclusion for Neolithic hematite axes from Qatar. 133

129 130 131 132

134

Uerpmann M. and Uerpmann H.-P. 1996: 131. Kallweit and Davies 2010: 105. Kallweit and Davies 2010: 105. Białowarczuk 2012.

135

264

Philipp Drechsler: Lithic Industry

Figure 11.39. Obsidian blades and blade fragments (a DOS2010-4998; b DOS2011-20478.1; c DOS2012-29370; d DOS201120478.2; e DOS2010-10872; f DOS2010-3234; g DOS2010-4999; h DOS2010257.4; i DOS2011-11239) (drawings: F. Brodbeck and B. Höpfer)

technological strategy. The majority of flint pieces arrived at the site as collected natural blanks destined to be transformed into various tool forms.

bifacial foliates and bifacial knives. This spectrum of tools is supplemented by scraping tools, borers, denticulates, splintered pieces, and laterally retouched flakes and blades. The diverse set of flint tools supports the view that Dosariyah represents a settlement inhabited over prolonged periods of time where a broad spectrum of activities was carried out.

The presence of both cores and debitage within the assemblage indicates that flint knapping occurred at the site. Primary production was orientated towards the generation of flake blanks. Clear evidence of the standardized production of blanks is absent and welldefined core maintenance strategies are difficult to reconstruct. Nevertheless, blanks repeatedly show evidence of dorsal reduction and dorsal chipping, indicating a certain degree of technical anticipation and full control over the knapping process. Blank production technology was generally characterized by direct hard hammer percussion and examples of corresponding hammer stones made from flint were frequently found. In contrast, indications for soft hammer percussion are rare although a few blanks which show distinct technological features such as lipping and a circular or semi-circular arrangement of blank negatives on their dorsal surfaces might suggest limited use of (organic) soft hammers during the shaping of some stone tools. Therefore, both débitage and façonnage concepts were used together at the site. The high level of knapping skills is exemplified by the ability to shape even small flint tools such as projectile points. Other facially retouched tool types include

The lithic technology and tool typology at Dosariyah demonstrate substantial technological and typological parallels with other coastal Middle Neolithic sites in the central Gulf and there are also significant similarities with assemblages from the upper and lower Gulf. The flint technology and tool typology at Dosariyah therefore reflect the socio-economic context of an indigenous Arabian population and it is most likely that flint knapping at Dosariyah was (exclusively) carried out by a local population. Stone artifacts are almost ubiquitous at Dosariyah yet spatially distinct workshops could not be identified during the excavation. Flint knapping was most likely carried out in a rather ad hoc fashion, and was thus embedded in daily economic and domestic activities. One exception represents a singular dense accumulation of natural pieces of flint in trench E1, half of which had been modified into scraping tools by a steep partial lateral retouch. Although this specific tool type was 265

Dosariyah previously identified in Dosariyah, this particular clustering reveals a distinct and characteristic manipulation scheme of the raw material. The starting point for the preparation of tools was not necessarily a piece of debitage, that is, a pre-manufactured blank, but more often the desired tool was made from pieces of natural flint which already had suitable morphologies and/or dimensions. This explains the comparatively high number of pieces of shatter at the site which are then interpreted as natural blanks for tool production.

from the sediment, these nodules are subject to severe chemical and physical weathering. Extreme fluctuations of temperature over long periods of time initiate the cracking of these nodules into rather thin and highly angular pieces which are well suited to the preparation of tools with a steep lateral retouch, such as scrapers. A similar tool preparation strategy allowed people to skip the debitage stage of the operational sequence observed in the case of arrowheads and tile knives. Most often these tools were manufactured from naturally occurring tabular flint pieces, hence the cortex remaining on both faces. Here again, suitable natural raw materials allowed operational short cuts to save time and effort during the production of flint tools.

A necessary precondition for this rather uncommon handling of flint is the occurrence of suitable raw material slabs. Natural flint occurs in its primary context, for instance in the sediment where it has been formed, in the form of nodules. Once exposed

266

Philipp Drechsler: Lithic Industry Plates

Plate 11.1. Convex scraping tools made on flakes. While the main steepening retouch is applied on the dorsal surface, a secondary flat retouch can often be found on the ventral surface as well (a DOS2010-7121.3; b DOS2010-4965; c DOS2011-16555; d DOS2012-22722; e DOS2010-9009) (drawings: F. Brodbeck and B. Höpfer).

267

Dosariyah

Plate 11.2. Convex scraping tools made on flakes with a prominent retouch on the ventral surface (a DOS2010-2534; b DOS2012-24064; c DOS2010-6642.2; d DOS2012-26205) (drawings: F. Brodbeck and B. Höpfer).

268

Philipp Drechsler: Lithic Industry

Plate 11.3. Circular scraping tools with a facial retouch (a DOS2010-7121.1; b DOS2011-14609; c DOS2011-18650; d DOS2011-819) (drawings: F. Brodbeck and B. Höpfer).

269

Dosariyah

Plate 11.4. Scraping tools made on flakes. a End-scraper (DOS2012-21886); b extreme form of a side scraper morphologically resembling an end-scraper (DOS2012-21225); c–e broad end-scraper (c DOS2012-29935; d DOS2012-22912; e DOS2012-257.3); f–g side-scraper made on thick blade-like flakes with a concave (f DOS2012-26465) and convex (g DOS2011-19871) working edge (drawings: F. Brodbeck and B. Höpfer).

270

Philipp Drechsler: Lithic Industry

Plate 11.5. Scraping tools with concave working edges. a–b Concave scrapers made from natural blanks (a DOS2012-21747; b DOS2012-21642); c–d concave scrapers made on flakes (c DOS2012-30877; d DOS2012-30867) (drawings: F. Brodbeck and B. Höpfer).

271

Dosariyah

Plate 11.6. Convex scraping tools made on natural blanks (a DOS2012-30012; b DOS2012-29560 c DOS2012-31292; d DOS201222822; e DOS2012-29903) (drawings: F. Brodbeck and B. Höpfer).

272

Philipp Drechsler: Lithic Industry

Plate 11.7. Scraping tools made from natural blanks with only marginal retouch (a DOS2012-23308; b DOS2012-29930; c DOS2012-23338; d DOS20012-22042; e DOS2012-22828) (drawings: F. Brodbeck and B. Höpfer).

273

Dosariyah

Plate 11.8. Convex scraping tools made on natural blanks (a DOS2012-22269; b DOS2010-3514; c DOS2012-22029; d DOS2010257.20) (drawings: F. Brodbeck and B. Höpfer).

274

Philipp Drechsler: Lithic Industry

Plate 11.9. Marginally retouched flakes (a DOS2010-6539; b DOS2012-21508; c DOS2010-3242; d DOS2010-5687.1; e DOS201120333; f DOS2010-758; g DOS2012-29882) (drawings: F. Brodbeck and B. Höpfer).

275

Dosariyah

Plate 11.10. Laterally retouched flakes (a DOS2010-1097.3; b DOS2011-19474; c DOS2012-26322; d DOS2010-6124; e DOS20103365; f DOS2010-2875) (drawings: F. Brodbeck and B. Höpfer).

276

Philipp Drechsler: Lithic Industry

Plate 11.11. Tile knives (a DOS2010-9804; b DOS2010-8471; c DOS2012-30927; d DOS2012-21639; e DOS2012-29568) (drawings: F. Brodbeck and B. Höpfer).

277

Dosariyah

Plate 11.12. Tile knives (a DOS2010-6887; b DOS2012-21672; c DOS2012-21184; d DOS2012-26547; e DOS2012-21789) (drawings: F. Brodbeck and B. Höpfer).

278

Philipp Drechsler: Lithic Industry

Plate 11.13. Arrowheads (a DOS2012-29952; b DOS2012-21396; c DOS2011-12510; d DOS2013-31385; e DOS2013-31382; f DOS2011-12512; g DOS2013-32655; h DOS2011-15714; i DOS2010-257.1; k DOS2011-12134; l DOS2012-29257; m DOS2012-21505; n DOS2011-20472; o DOS2012-21354; p DOS2010-6021.3; q DOS2012-22461; r DOS2010-987) (drawings: F. Brodbeck and B. Höpfer).

279

Dosariyah

Plate 11.14. Arrowheads (a DOS2010-11109; b DOS2011-19675; c DOS20107121.2; d DOS2011-12133; e DOS2011-15862; f DOS201112852; g DOS2011-15835; h DOS2012-30051.1; i DOS2011-13163; k DOS2010-5687.4; l DOS2011-18763; m DOS2010-11219.1; n DOS2010-7872; o DOS2010-10314; p DOS2010-3076; q DOS2011-14610; r DOS2010-6646.1; s DOS2012-21402) (drawings: F. Brodbeck and B. Höpfer).

280

Philipp Drechsler: Lithic Industry

Plate 11.15. a Bifacial foliate (DOS2010-257.2); b bifacial knife (DOS2010-9100.9); c–d broad bifacial foliates (c DOS2012-21552; d DOS2012-31084) (drawings: F. Brodbeck and B. Höpfer).

281

Dosariyah

Plate 11.16. Roughly shaped bifacial foliates (a DOS2012-22271; b DOS2012-21168; c DOS2011-20744) (drawings: F. Brodbeck and B. Höpfer).

282

Philipp Drechsler: Lithic Industry

Plate 11.17. Roughly shaped bifacial foliates (a DOS2011-18328; b DOS2010-1777; c DOS2011-18498+18503) (drawings: F. Brodbeck and B. Höpfer).

283

Dosariyah

Plate 11.18. Possible bifacial preforms (a DOS2010-2374; b DOS2010-6646.3; c DOS2012-21785; d DOS2010-9795; e DOS201112613) (drawings: F. Brodbeck and B. Höpfer).

284

Philipp Drechsler: Lithic Industry

Plate 11.19. a–f Splintered pieces (a DOS2010-7270; b DOS2010-9401.7; c DOS2010-11219.6; d DOS2012-29717; e DOS2012-29342; f DOS2010-9100.6); g–i; m–p borers (g DOS2010-5689.1; h DOS2010-1430.9; i DOS2010-697; m DOS2012-29412; n DOS2010-1097.1; o DOS2012-21340; p DOS2012-29258); k–l micro-borers (k DOS2010-6805; l DOS2010-6705) (drawings: F. Brodbeck and B. Höpfer).

285

Dosariyah

Plate 11.20. Flakes with sickle sheen (a DOS2010-11231.1; b DOS2010-1737; c DOS2012-21934) (drawings: F. Brodbeck and B. Höpfer).

286

Philipp Drechsler: Lithic Industry

Plate 11.21. a–b Denticulates (a DOS2010-738; b DOS2010-10844); c hammer stone made from nodular flint raw material (DOS2012-26540) (drawings: F. Brodbeck and B. Höpfer).

287

Dosariyah

Plate 11.22. a–c Stone axes (a DOS2011-11335; b DOS2012-22115; c 5/235 (Masry 1972); d stone ring fragment (5/222; Masry 1972); e–f hematite billets (e DOS2010-4952; f DOS2012-30967); g net sinker (DOS2011-14372) (drawings: F. Brodbeck and B. Höpfer).

288

Philipp Drechsler: Lithic Industry

Plate 11.23. Querns (a DOS2010-746; b DOS2010-1656) (drawings: F. Brodbeck and B. Höpfer).

289

Dosariyah

Plate 11.24. a Potential pestle (DOS2012-21802); b hand stone (DOS2012-21798) (drawings: F. Brodbeck and B. Höpfer).

290

Chapter 12 Variability of Arrowhead Shapes Philipp Drechsler 1 Introduction

fields of raw material procurement, raw material pretreatment, knapping technology and formal variation of flint tools. As a result of the Bordes-Binford academic debate about the very nature of flint tools,5 there has been a tendency towards the study of raw material acquisition and treatment as well as technological studies. Formal variation in flint tools, however, is similarly affected by their morphological appearance which can be understood as an ethnic idiom imparted in all material culture.6 Although the actual form of any object should be considered as entirely suited to its function, there exists a great range of alternatives that would be more or less equally appropriate. These functional equivalents constitute a spectrum of what has been termed isochrestic or equivalent in use. From this spectrum, a craftsman must choose between options which are theoretically available and as a large number of options can be assumed as potentially available for the same functional task, the probability that two independent groups or societies would select the same combination of such choices by chance is unlikely. Since material culture is a product of learned behaviors that are socially transmitted, a direct correlation between specific choices and their position in cultural history exists.7

Chipped stone arrowheads represent a conspicuous and frequent group of artifacts at Dosariyah. Based on her early visits at the site, Burkholder mentions ‘a wealth of colorful flint implements ... The crudely worked arrowheads, tanged and barbed, are the most interesting of these, and range in length from two to more than three centimeters. Many appear to have been chipped from thin pieces of tabular flint’.1 While these attractive implements were extensively collected from here (sadly like many other sites across Arabia) during subsequent years,2 a total of 35 arrowheads were documented during recent fieldwork at the site. This material provides the basis for a detailed morphological study investigating the outlines of individual arrowheads using Elliptic Fourier descriptors to establish distinct morphological groups. Eye-catching bifacial pressure-flaked projectile points and bifacial foliates were often collected by geologists, petroleum engineers and amateur archeologists working in Arabia particularly during the 1950s and 1960s. At that time arrowheads provided a basis for dating these assemblages to the Neolithic period.3 Later, the wide distribution of bifacially shaped implements as part of rather homogeneous tool assemblages and an associated flake-based lithic technology led to the definition of the Arabian Bifacial Tradition as a formative element of the Neolithic in eastern and central Arabia.4

Arrowheads are a class of artifacts which is especially suited for the study of formal stylistic variation.8 Although their function as the cutting tips of accelerated projectiles is self-evident, they show a high degree of morphological variability. Owing to the fact that the possible range of physical characteristics of projectile tips for this kind of accelerated system is relatively wide,9 a broad spectrum of isocrestic shapes sensu Sackett have been invented that are independent of function. Likewise, a second characteristic of arrows supports the study of stylistic variation: due to the low velocity of such projectiles, arrows travel in an arched trajectory. In order to achieve an accurate shot, an archer must be able to estimate the distance to the target and know the properties of their equipment

In more recent decades there has been a partial shift in interest from lithic artifacts as chronological fossiles directeurs towards understanding their implications for the study of social, economic, and political life. Both approaches assume a priori that remains of material culture contain information which is sensitive to answering such questions. This assumption is plausible as long as the characteristics of an artifact result from cultural rules rather than from functional determinations, and variability in flint artifacts which can be related to cultural constraints comes from the 1 2 3 4

Bordes 1961; 1972; 1973; 1978; Binford L.R. and Binford S.R. 1966; Binford L.R. 1968; 1973; 1983; 1989; Bordes and Sonneville-Bordes 1970. 6 Sackett 1982. 7 Sackett 1982: 72. 8 Wiessner 1983. 9 Hughes 1998. 5

Burkholder 1972: 266. Masry 1997: 78. Zeuner 1954; Field 1955; 1958; 1960a; 1960b; Gramly 1971. Edens 1982; 1988.

291

Dosariyah in order to choose the right trajectory.10 Due to this dependency between equipment and accuracy, one can infer that the archer would try to reduce variability within his own equipment and this would lead to rather standardized personal projectile points.

(EFA) is most commonly applied to complex shapes17 but has been severely criticized.18 Other comparable methodologies which have been successfully applied to biological shapes are Eigenshape analyses19 and outline analyses using B-splines.20 Most shape-based methods need at least one point which has to be considered as homologous, but some methods calculate a point from the overall geometry of the outlines which are considered as homologous during subsequent analyses.

The potential of arrowheads as a source of cultural information about prehistoric Arabia has been demonstrated both for the distinction of population groups11 and for the study of the dispersal of Neolithic populations across Arabia.12 Nevertheless, such studies are faced with the problem of comprehensive descriptions of arrowhead morphology. Their physical appearance seems to represent a continuum between extreme forms, a problem which most typological studies are faced with. Possible solutions are the definition of broad categories of arrowhead types on the basis of fundamental characteristics or a detailed but subjective distinction between arrowhead types based on overall similarities and dissimilarities.

Although these outline-based analyses of morphological variation may have drawbacks compared to classic landmark-based approaches,21 they appear be more suitable for the study of specific aspects of material culture: for example, most flint artifacts and especially flint tools lack common features which can be considered as homologous. Impact points on blanks as a result of a universal blank production process are an obvious exception, but this feature is often removed during subsequent modification of blanks into tools. Furthermore, flint tools often show a functional part whose shape is largely predetermined by function, and a non-functional part which is influenced by the shape of the initial blank. Although some methods are able to deal with open forms as well (Eigenshape analysis) and can analyze functional parts, the majority of shapebased methods require closed curves.

A precondition for both an objective and detailed classification of arrowhead morphology is a precise and replicable description and analysis of shape. As several disciplines are faced with this challenge, a wide range of methodologies have been introduced during the last decades and the field of geometric morphometrics developed.13 The application and adaptation of these methodologies originally intended for the analysis of biological shape is disputable for the study of material culture in general, and stone artifacts in particular. Several attempts have been made, but have not replaced the stage of pre-studies14 or focus on data acquisition.15 A major drawback of many methods of shape analysis is the need for landmarks, in other words, well-defined and retrievable points occurring on all objects, which can be considered as homologous.16 Such points clearly exist in biological shapes where their existence is predetermined by genetic code. In contrast, humanmade objects often lack such clearly defined points but exceptions might be found where functional constraints exist which lead to equal expressions in shape, or where the production process of objects results in comparable features.

Aside from these general constraints, projectile points, and especially arrowheads, are well suited for outlinebased morphometric studies. Firstly, they are essentially two-dimensional objects and their morphological variability is mostly determined by their outline. In contrast, cross sections of arrowheads are often less varied, although exceptions exist.22 Furthermore, they show at least one well-defined and de facto homologous point, namely the tip which is intended to penetrate its target. Finally, projectile points often have more than one functional part — tip, wings, barbs and tang/ socket — and therefore many parts of the object are the result of intentional shaping and should be considered comprehensively.

In contrast to landmark-based approaches, few methods are based on outlines which require only a few or even no landmarks; instead, a two- or three-dimensional outline of an object may be applied. Elliptical Fourier Analysis

Traditionally, variation in form is considered as the result of variation in shape and size (Figure 12.1).23 Accordingly, morphometric analyses intend to reduce the impact of size in their results but size itself cannot be excluded a priori as the characteristic element of stylistic variation for the study of material culture, and

Hughes 1998. 11 Spoor 1997. 12 Drechsler 2009. 13 Blackith and Reyment 1971; Bookstein 1989; 1991; Dryden and Mardia 1998; Elewa 2004; 2010; Zelditsch et al. 2004. 14 Cardillo 2010. 15 Lenardi and Merwin 2010. 16 Homologous points — also called landmarks — are corresponding, definable points, boundaries or parts of objects; see Bookstein 1991.

17 Kuhl and Giardina 1982; Rohlf and Archie 1984; Ferson et al. 1985; White et al. 1988; Liu et al. 1996; Iwata et al. 1998; Haines and Crampton 2000. 18 Bookstein et al. 1982. 19 Lohmann 1983; Lohmann and Schweizer 1990; McLeod 1999. 20 de Boor 1978; Neubauer 2007; Neubauer and Linhart 2008. 21 Full and Ehrlich 1986. 22 Trihedral points from Neolithic contexts in southern Arabia: Charpentier 2004. 23 See Richtsmeier et al. 2002.

10

292

Philipp Drechsler: Variability of Arrowhead Shapes perpendicular to a plane created by length and width. All measurements were taken with a precision of 1/10 mm. Exploratory analyses explain the overall variability of these measurements using statistical descriptions. While these unrelated linear measurements themselves refer to size, they have also been used to calculate the shape-related parameter-free length/width index that places each arrowhead along a scale ranging from ‘short and broad’ to ‘long and narrow’. For the study of arrowhead shape, an outline-based approach was followed which uses Elliptic Fourier Descriptors. These can delineate any type of shape with a closed two-dimensional contour25 and have been applied effectively to the analysis of various biological shapes in plants and animals. In recent years, Elliptic Fourier Descriptors have also been applied to the study of lithic artifacts.26 Figure 12.1. Relationship between form, shape and size (after Richtsmeier et al. 2002).

In the following study, analyses have been carried out using the software package SHAPE 1.3.27 As a first step, drawings of flint arrowheads were scanned with a resolution of 150 dpi and silhouettes subsequently filled with black color. In addition, a black square measuring 10 x 10 mm was added to each file for the subsequent calculation of size. Contour extraction was carried out using ChainCoder, which is part of the SHAPE software package. Chain code is a coding system for the description of geometric information of contours.28 The ChainCoder software first converts the scanned image to a binary image, reduces noise, traces the contour of the object and finally describes the contour information as chain code.29 As a next step, SHAPE calculates normalized Elliptic Fourier descriptors (EFD) from the chain code information according to the procedure suggested by Kuhl and Giardina.30 Normalization and orientation was based on the point of the contour farthest from the center (i.e. the longest radius) which was manually corrected to the tip of each arrowhead point. In a later step, a principal component analysis of the normalized EFDs was performed to efficiently summarize the information contained in these coefficients.31 The principal component analysis conducted by SHAPE is based on the variance/covariance matrix of both symmetric and asymmetric coefficients. As the final step, principal component scores were plotted to visually inspect the outcome of the analyses, and clustered using the statistical software JMP10 to obtain an objective classification of shape.

the variability of size will therefore also be considered in the following study of arrowhead morphology. 2 Methodology The morphological appearance of an arrowhead is determined by its size and shape. Apart from the shape of the arrowhead tip, the occurrence of specific functional traits such as wings, tangs/socket and barbs affects the outline of an arrowhead. These physical features can be abstract to categories and their presence or absence is noted. The general terminology of lithic arrowheads established for eastern and southeastern Arabia differentiates between: 1) short bifacial winged and tanged arrowheads; 2) short winged and tanged arrowheads without bifacial retouch; 3) bifacial leafshaped arrowheads without a tang; 4) bifacial leafshaped arrowheads with a clear tang; and 5) denticulated arrowheads with a triangular cross section.24 The first of these groups can be further differentiated between shouldered and barbed pieces. As a first step toward an assessment of stylistic variation, each arrowhead from Dosariyah was assigned to one of these categories. The size of arrowheads was documented by three linear measurements: length, width and thickness. The length of each arrowhead was measured as the maximum distance between the tip and the base of the implement along its major symmetrical axis. The width was measured perpendicular to the length at the broadest point while the thickness was measured

Kuhl and Giardina 1982. Gero and Mazullo 1984; Saragusti et al. 2005; Cardillo and Charlin 2009; Iovita 2009; 2010; 2011; Iovita and McPherron 2011; Serwatka 2015; Borel et al. 2016. 27 Iwata and Ukai 2002; Iwata 2006. 28 Freeman 1974. 29 Iwata 2006. 30 Kuhl and Giardina 1982. 31 Rohlf and Archie 1984. 25 26

Spoor 1997; but see Crassard 2008 for a different classification schema for southern Arabia that places more emphasis on technological aspects.

24

293

Dosariyah 3 Material

surface and five from profile collapse. A total of 26 arrowheads were entirely preserved and allowed for the full set of analyses. In contrast, nine pieces are broken — they lack some part of the original projectile point. In eight cases, however, the missing parts were confidently reconstructed for the outline analyses.

Excavations and surface collections at Dosariyah resulted in the recovery of 35 arrowheads (Table 12.1). A total of 13 have been documented from stratigraphic contexts, while 17 pieces were collected from the

Table 12.1. Dimensions, weight, preservation and stratigraphic position of arrowheads from Dosariyah. Find ID

Lengh Width Thick-ness Weight Preservation (mm) (mm) (mm) (g)

DOS201024.7 275.1 DOS201018.7 987 DOS201017.0 3076 DOS201030.5 5687.4 DOS201027.4 6021.3 DOS201016.5 6646.1 DOS201030.2 7121.2 DOS201015.1 7872 DOS201018.4 10314 DOS201030.4 11109 DOS201025.6 11219.1 DOS201111240

DOS201124.8 12133 DOS201124.0 12134 DOS201125.3 12510 DOS201131.5 12512 DOS201129.5 12852 DOS201128.3 13163 DOS201119.2 14610 DOS201115714

DOS201126.5 15835 DOS201126.6 15862

Metrical analysis

Morpho-metric Context analysis

Strat. Unit

17.1

3.8

1.28

Complete

yes

yes

Surface

 

15.5

2.6

0.80

Tip missing

no

yes

Upper

N1-I

15.4

3.0

0.90

Minor part of tip and no stem missing

yes

Lower

S1-IV

14.6

3.1

1.38

Complete

yes

yes

Surface

 

16.9

4.4

1.44

Complete

yes

yes

Surface

 

12.8

4.0

0.80

Minor part of tip missing

no

yes

Surface

 

11.1

2.6

0.96

Complete

yes

yes

Surface

 

16.6

3.7

0.80

Minor part of tip missing

no

yes

Overburden

 

14.4

3.6

0.87

Complete

yes

yes

Overburden

 

10.6

3.7

0.82

Complete

yes

yes

Lower

S3-IV

14.4

2.6

0.86

Complete

yes

yes

Surface

 

-

-

-

Tit, stem and wings/ no barbs missing

no

Surface

10.4

2.9

0.56

Complete

yes

yes

Surface

 

23.9

4.2

1.99

Minor part of tip missing

no

yes

Surface

 

15.1

3.6

0.76

Complete

yes

yes

Surface

 

19.4

4.9

2.60

Complete

yes

yes

Surface

 

14.6

4.0

1.28

Complete

yes

yes

Lower

S2-IV

14.6

4.0

1.58

Complete

yes

yes

Lower

S2-IV

15.2

3.6

0.56

Complete

yes

yes

Lower

S2-IV

19.0

3.4

0.81

Tip missing

no

yes

Surface

 

13.0

2.7

0.67

Complete

yes

yes

Lower

S2-IV

12.4

4.0

0.95

Complete

yes

yes

Lower

S2-IV

294

Philipp Drechsler: Variability of Arrowhead Shapes

Find ID

Lengh Width Thick-ness Weight Preservation (mm) (mm) (mm) (g)

DOS201127.4 18763 DOS201136.7 19675 DOS201122.3 20472 DOS201224.6 21396 DOS201218.1 21402 DOS201221505

DOS201228.5 22461 DOS201229257

DOS201223.0 30051.1 DOS201224.1 31381 DOS201341.1 31382 DOS201333.2 31385 DOS201321.7 32655

Strat. Unit

3.5

1.00

Complete

yes

yes

Surface

 

12.1

3.5

1.25

Complete

yes

yes

Lower

S2-IV

18.3

3.7

1.42

Complete

yes

yes

Surface

 

14.8

3.6

0.89

Complete

yes

yes

Upper

E1-I

10.5

2.3

0.41

Complete

yes

yes

Upper

E1-I

15.5

4.5

1.38

Tip missing

no

yes

Upper

E1-I

13.1

3.4

0.91

Complete

yes

yes

Upper

N3-II

17.4

2.4

0.89

Stem missing

no

yes

Surface

 

12.7

3.2

0.83

Complete

yes

yes

Overburden

 

13.7

1.8

0.57

Complete

yes

yes

Surface

E1-II

20.3

4.6

3.38

Complete

yes

no

Surface

18.0

3.6

1.74

Complete

yes

no

Overburden

14.4

3.1

0.67

Complete

yes

no

Overburden

distribution, and the average thickness of arrowheads is 3.4 mm, with a standard deviation of only 0.7 mm.34

All chipped stone arrowheads from Dosariyah can be confidently identified as short bifacial winged and tanged arrowheads.32 Distinct barbs were formed on 16 implements, while all other pieces show more or less prominent shoulders. With these general morphological characteristics, the arrowheads from Dosariyah confirm the previously noted predominant occurrence of arrowheads of this type in the central Gulf, while other types are almost absent.33

The arrowheads from Dosariyah are elongated and show an average length/width index of 1.9, thus on average their length is twice that of their width. With the exception of three pieces, the maximum length does not exceed the width by a factor of 2.5. The graphical representation of arrowhead length and width, however, emphasizes the generally homogeneous character of size and gross shape in the arrowhead assemblage (Figure 12.3).

The average size of flint arrowheads is conspicuously homogeneous with a mean length of 26.4 mm and a standard deviation of only 4.4  mm. (Table 12.1). A dot plot suggests a rather homogeneous length distribution, with one point substantially longer, and three points shorter than all other pieces (Figure 12.2). The average width of arrowheads is 14.8  mm, with a standard deviation of 2.9 mm: again, the data set is quite homogeneous with only one outlier. The least variation can be observed in the thickness: the dot plot does not indicate any outliers, suggesting a very homogeneous

33

Morpho-metric Context analysis

14.2

4 Results

32

Metrical analysis

For the analysis of shape using Elliptic Fourier Descriptors, the outlines of 31 arrowheads were recorded. During this analysis, 20 harmonics were considered, resulting in 80 analyzed coefficients and 80 principal components. Nevertheless, only seven principal component account for more than 1.25% overall variability (1/Number of Analyzed Components) and these should be considered as significant. The first principal component (PC1) represents 48.28% of total data variability, PC2 23.28% and PC3 13.28%. Together, Corresponding to size variability, arrowhead weights range between 0.41 and 3.38 g with an average weight of 1.12 g (see Table 12.1).

34

Spoor 1997: Group 1. Spoor 1997.

295

Dosariyah

Figure 12.2. Dot plot for a maximum length (N = 26); b width (N = 34) and c thickness (N = 34) of arrowheads from Dosariyah.

25

Width (mm)

20 15 10 5 0

0

10

20

30

40

Length (mm) Figure 12.3. Length/width plot of entirely preserved arrowheads (N = 26).

296

50

Philipp Drechsler: Variability of Arrowhead Shapes One major advantage of an analysis using Elliptic Fourier Descriptors is the possibility of reconstructing tendencies of shape represented by each principal component. An outline reconstruction of the first principal component suggests that the major variability of arrowhead shape occurs in the slenderness of the silhouette. Long and narrow points are in opposition to short and compact points. The second principal component considers the distinctiveness of barbs as a second characteristic element of the arrowhead assemblage. PC 3 describes the location of the shoulder, thus the portions of tip and tang (Figure 12.4). Trends in shape for the remaining principal components are more difficult to specify, but minor deviations in axis symmetry and small-scale variation in shape account for the remaining 15.16% of variability.

the first three PCs account for 84.84% of total variability (Table 12.2). Table 12.2. Results of Elliptic Fourier analysis. Eigenvalue and proportion of the first seven significant Principal Components that account for 84.84% of shape variability. Eigenvalue Proportion (%) Cumulative (%) PC1

1.53E-02

48.2786

48.2786

PC3

4.22E-03

13.282

84.8358

PC2 PC4 PC5 PC6 PC7

7.40E-03 1.22E-03 9.39E-04 7.60E-04 4.44E-04

23.2752 3.8459 2.954 2.389

1.3972

71.5538 88.6817 91.6357 94.0246 95.4218

Figure 12.4. Outline reconstruction of the first three principal components.

297

Dosariyah

Figure 12.5. Plot of the first two PCs in a two-dimensional space (x = PC2, y = PC1, N = 31). Although the arrangement of implements does not show an obvious structuring of the data, a trend from compact triangular shapes in the lower right over slender points toward elongated arrowheads with prominent barbs in the left can be observed.

are isolated but further subdivided into short and elongated forms. Again, arrowheads with prominent barbs form a distinctive group. Finally, the remaining pieces are further differentiated into a shorter form with prominent barbs, and a slender shape where barbs are also present but less prominent.

Plotting the first two PCs into a single diagram does not reveal an obvious structuring of the data, confirming the overall homogeneity in shape of the arrowhead sample (Figure 12.5). On the left side of the scatter plot, specimens with barbs predominate and become increasingly slender along the y-axis. On the right side, arrowheads with more or less developed shoulders prevail. The arrangement of arrowheads in the twodimensional space suggests that three different shapes, or types, can be separated: compact triangular arrowheads with short wings and tangs in the lower right of the diagram, slightly elongated arrowheads with prominent barbs in the far left, and an almost continuum of elongated winged and tanged arrowheads with and without barbs centering in the middle.

5 Discussion Formal variation in arrowhead morphology is suspiciously low at Dosariyah. All of the 35 arrowheads can be characterized as short bifacial winged and tanged arrowheads. The greatest variation in form is in the shape of the barbs: 16 arrowheads show prominent barbs while the remainder are shouldered. Nevertheless, it is difficult to draw a distinct line between barbed and shouldered pieces. This morphological continuum is remarkable as shoulders and barbs represent different technological traits. Arrowheads with very narrow shoulders would fit into the shaft of the arrow continuously without any protrusions and could

To further separate this group a cluster analysis of the first seven significant PCs has been carried out using Ward’s method. It resulted in five distinct clusters that further differentiate the groupings (Figure 12.6). Compact triangular winged and tanged arrowheads 298

Philipp Drechsler: Variability of Arrowhead Shapes

Figure 12.6. Results of the cluster analysis of the first seven significant PCs against the background of the plot of the first two PCs in a two-dimensional space (N = 31).

in general were a common characteristic of the group living at Dosariyah. Although an additional distinct ‘meaning’ of shape clusters cannot be excluded it is also plausible that these observed differences in shape were the result of differences in raw material use, arrowhead preparation and reshaping: short triangular forms are more often made from thin pieces of tabular flint, while elongated arrowheads have been predominantly made from elongated flakes.

be easily removed from carcasses, whereas barbed arrowheads are made to lodge securely. The overall consistency of the arrowhead sizes suggest standardized bow-and-arrow equipment. According to the stratigraphic observations, the arrowheads found at Dosariyah do not belong to any single individual but to different people over a range of time. Owing to their homogeneity, arrowheads could have been easily interchanged between users without major discrepancies in shooting trajectories. This functional homogeneity lasted from a minimum of several decades to a period of some centuries. It indicates that not only arrowhead style but shooting technology itself was consistent and characteristic for the specific group settling at Dosariyah.

In accordance with size, arrowhead shape does not show clear development over time. Nevertheless, there exist weak correlations between the first principal components and stratigraphic positions (Figure 12.7), but the absence of a clear chronological stratification of the settlement questions observed trends: a tendency towards shorter and compact arrowheads is discernible towards the upper parts of the stratigraphy, while more slender silhouettes occur at the beginning of the occupation at Dosariyah and prominent barbs predominate in the lower part of the sequence. The broadest variations in shape nevertheless show the finds from the surface, indicating that this material

Arrowhead shape variation is also low. Although differences in form are noticeable and well represented in the shape analysis using Elliptic Fourier Descriptors, it is questionable if these differences represent significant stylistic variations and it is more likely that winged and barbed arrowheads with a tang for shafting 299

Dosariyah

Figure 12.7. Stratigraphic trends of a PC1; b PC2 and c PC3 (N(surface) = 15, N(upper) = 5, N(lower) = 8).

represents an actual mix of objects both from the younger and older periods of the settlement.

At present it is difficult to explain the functional, chronological or cultural meaning behind these groupings. As demonstrated above, clear stratigraphic trends cannot be observed which are in accordance with the radiocarbon dates and results from the lithic study (see Chapter 11). The occupation at the site did not exceed a few centuries, a period of time which is probably too limited to show any marked changes in the shape and style of arrowheads. The same degree of conservatism pervades the entire flint assemblage. On the other hand, the overall homogeneity of the arrowhead assemblage points towards a high degree of cultural consistency. The hypothesis that Dosariyah was a meeting point of (culturally or technologically) different population groups is not supported by the stylistic variability of arrowheads and instead the shapes of arrowheads suggest a high degree of stylistic — and therewith plausibly social — proximity.

The application of Elliptic Fourier Descriptors for the classification of arrowhead shapes may well confirm the homogeneity of the arrowheads that has already been observed by more conservative approaches. Future research should therefore include arrowheads from a range of other contemporaneous sites to assess whether observed groupings obtained by this method in combination with statistical clustering have an intrinsic, group-specific meaning or not. The study of arrowhead shape variability at Dosariyah demonstrates the potentials of morphometric analyses for the study of material culture. It confirms the overall homogeneity of the arrowhead assemblage with all individual pieces falling into the category of short bifacial winged and tanged arrowheads according to Spoor’s classification.35 The outline analysis using Elliptic Fourier Descriptors, however, suggests at least a tripartite subdivision of the arrowhead assemblage into compact triangular shapes with small barbs or shoulders, slender points with shoulders and triangular arrowheads with prominent barbs.

The observed variability in the shape of arrowheads therefore most likely reflects function. The dichotomy between elongated and triangular bodies influences the penetration of the projectile marginally but differs in the size of the actual wound, in other words, broader points (Clusters 2 to 5) generally result in more intensive bleeding. In contrast, prominent barbs (Clusters 2 and 5) should result in their secure lodging within the body, while shoulders (Clusters 3 and 4) do not fulfill this task reliably. The imperfect resolution of arrowheads assigned to Cluster 1 which are all characterized by a slender body, with or without tang, indicates either the relative insignificance of these traits (at least for this group of points) or the limits of automated shape classification.

Using cluster analysis, this subdivision of the assemblage can be refined further. The largest group of arrowheads (Cluster 1, N  =  14) is dominated by elongated shapes with rather broad tangs in relation to the width of the body, while barbs or shoulder are not considered. A second cluster (Cluster 2, N  =  6) comprises more compact arrowheads with a triangular body and clearly developed broad tangs. Cluster 5, represented by two arrowheads only, represents an extreme expression of arrowheads assigned to Cluster 2, with very prominent barbs which are almost as long as their tangs. On the other side, Clusters 3 and Cluster 4 refer to arrowheads with either elongated triangular (Cluster 3, N  =  4) or short triangular (Cluster 4, N = 5) bodies with a shoulder or only weakly developed tangs.

35

6 Conclusions The present study of arrowhead shape variability suggests a general homogeneity of the arrowhead assemblage from Dosariyah. All those that are sufficiently preserved fall within the range of short bifacial winged and tanged arrowheads and both barbs and shoulders occur in almost equal numbers. The arrowheads from Dosariyah therefore fall within the

Spoor 1997.

300

Philipp Drechsler: Variability of Arrowhead Shapes different arrowhead (outline) shapes present at the site. The boundaries between individual groups are often blurred, however, at first reflecting the fluent transition between these groups and underlining the homogeneity of the assemblage.

known spectrum of Middle Neolithic arrowheads from the central Gulf region.36 Conceding arrowhead shape and cultural meaning, this high degree of morphological similarity likewise suggests close cultural proximity. The absence of morphologically different arrowheads therewith challenges the hypothesis that Dosariyah served as a gathering place for people from different cultural spheres.

Reasons behind the distinct tendencies in arrowhead shapes are difficult to explain. Chronological developments are weakly expressed in the data. Similarly, differences in shape seem to be too small to relate to (archeologically ascertainable) cultural change. What remains, therefore, are functional differences that potentially relate to different use strategies including hunting or warfare, but might also fall within the realms of flint raw material properties, shafting methods, reshaping and individual variation.

Automated outline classification of 31 arrowheads indicates three distinct trends of arrowhead shapes within the assemblage from Dosariyah: compact triangular arrowheads with small barbs or shoulders cooccur with slender points with shoulders and triangular arrowheads with prominent barbs. Subsequent cluster analysis further refines this picture and identifies five

36

Spoor 1997.

301

Chapter 13 Hematite Objects and the Use of Red Pigments Philipp Drechsler, Christoph Berthold and Christine Kainert 1 Introduction

resemble flat bars with an approximately quadrilateral outline (DOS2010-4952: 25.4  x 20.6  x 4.5  mm, 5.83  g, Figure 13.1a; DOS2012-30967: 20.9  x 27.6  x 5.5  mm, 9.69  g, Figure 30.1b). On both flat surfaces, but also along their outer edges, there are clear traces of grinding which have left parallel striations. In addition, individual grinding facets created by grinding at different angles are visible along the edges. Both pieces show one unworked edge with a surface characteristic of a fracture. During the excavation of one of the pieces (DOS2010-4952), the surrounding sediment layers were found to have turned red, suggesting that these items represented pieces of raw material used for the production of red pigment by grinding or scraping and they therefore represent fragments of larger blocks of hematite that had been deliberately discarded.

The use of hematite, the mineral form of iron(III)oxide (Fe2O3),1 as raw material for the manufacture of axe heads was published from the Middle Neolithic sites of Groon Baida 1/SQS12-17 and SQS12-4,2 both located on the Qatar peninsula and broadly contemporary with the settlement at Dosariyah.3 From these sites, three hematite objects were found that morphologically resemble small axe or adz heads. Objects of similar shape and made from a wide variety of solid rocks are known from other Middle Neolithic sites along the shores of the Gulf, representing part of a material assemblage which is characterized by bifacially chipped flint artefacts and often associated with Ubaid pottery imported from southern Mesopotamia. Although the physical qualities of hematite (hardness, toughness) qualify this mineral as raw material for the production of axe or adze heads, another unique characteristic should also be considered: during the grinding of hematite, a bright red powder — or in the case of wet grinding, a blood-red liquid — appears. This is highly significant as the use of hematite powder as a red pigment, both for the decoration of objects as well as the human body, is well-known from the African Middle Stone Age (MSA)4 and the European Middle Paleolithic5 onwards.

A third object resembles a broken axe head (DOS201222115, Figure 13.2). The cutting edge is symmetrical and elaborate but the body is only roughly worked; the whole piece is highly patinated and most of the surface has a purple weathering layer covering a dark grayish core. Measuring c.5.4 x 3.2 x 1.1 cm, this axe head falls well within the size spectrum of the hematite axes from Qatar, although the physical characteristics of the raw material are different.

2 Material and Methods

In addition to these three objects, traces of red pigment were identified on two grinders/hand stones, again both excavated in trench E1. One hand stone made from a greenish metamorphous solid rock shows a rather intense coloring on the grinding surface, suggesting that it had been used for the preparation of red pigment (DOS2012-30460, Figure 13.3a). A second wedge-shaped hand stone made from siliceously cemented sandstone or quartzite also shows traces of red pigment, although not over its entire surface (DOS2012-31302, Figure 13.3b).

During the excavations at Dosariyah, three objects found in trench E1 shared similar physical characteristics of a grayish to reddish and greenish dull color and high specific weight which implied some sort of raw material containing a high iron content. Two of these A distinction must be made between hematite on the one hand and ochre and red ochre, used as pigments for both the decoration of objects and the human body, on the other hand. Hematite is pure iron(III) oxide (Fe2O3) and crystallizes in the rhombohedral lattice system. It is colored black to steel or silver-gray, brown to reddish brown, or red. Characteristic is a rust-red streak, referring to the red color of the pulverized mineral. In contrast, ochre is a natural earth pigment predominantly containing hydrated iron oxide (limonite, FeO(OH)·nH2O), partly hydrated iron oxide (goethite, FeO(OH)) and anhydrous iron oxide (hematite, Fe2O3), and also clay minerals and quartz. In dependence of the proportions of iron oxides, the color spectrum of ochre ranges from yellow (predominance of limonite) over brown (predominance of goethite) to red (predominance of hematite). 2 Drechsler et al. 2013. 3 Drechsler 2014. 4 Watts 2009; 2010; Henshilwood et al. 2011. 5 Roebroeks et al. 2012. 1

Apart from these two grinders, a total of 61 pieces of Black-on-Buff Ware pottery were found that show traces of red pigment on their surface.6 The majority of these was found in trench N3 (N = 40) and S2 (N = 13), while a smaller number was documented from trenches Remains of red pigment were not observed on vessel fragments of Coarse Ware. The reddish color of many Coarse Ware sherds as well as their rather friable state, however, might have prevented a reliable identification, thus the presence of red pigments on Coarse Ware sherds cannot be excluded.

6

302

Philipp Drechsler et al.: Hematite Objects and the Use of Red Pigments

Figure 13.1. Polished hematite pieces used for the production of red pigment (a DOS2010-4952, SU E1-III; b DOS2012-30967, SU E1-III) (drawings: B. Hoepfer, F. Brodbeck).

Figure 13.2. Broken axe head made from hematite (DOS2012-22115, SU E1-II) (drawing: B. Hoepfer, F. Brodbeck).

N1 (N = 2), N2 (N = 1) and E1 (N = 5). Finally, a single bead made from a piece of shell with a concentric sculpture on the exterior surface showed traces of red pigment between the ridges of the shell (DOS2010-5270, SU E1III).

powder diffraction which needs certain amounts of separated material, X-ray micro-diffraction gives the opportunity to identify mineral phases directly on the sample surface without removing material from it for subsequent analysis. Applying a large 2-dimensional detector (µ-XRD2) in such a micro-diffractometer set-up, the detector image provides supplemental information, such as texture and crystallite sizes of the mineral phases in the investigated area.8

To characterize the mineral composition of the two polished pieces (Figure 13.1a–b), the axe head (Figure 13.2) and the red pigment on two potsherds (DOS201114185, DOS2011-14733), were analyzed using a nondestructive, locally resolved X-ray micro-diffraction technique.7 In contrast to classical destructive X-ray 7

The µ-XRD2 set-up used is a BRUKER D8 Discover θ/θ GADDS micro-diffractometer equipped with a Co-

Berthold et al. 2009.

8

303

Berthold et al. 2015.

Dosariyah

Figure 13.3. Hand stones with traces of red pigment (a DOS2012-30460, SU E1-VII); b DOS2012-31302, SU E1-III) (photographs: B. Kiepenheuer-Drechsler).

than 1 wt%) of quartz. The hematite is well crystalline with some coarser single crystals, due to the sharp reflections with visible single crystal spots in the diffraction image (Figure 13.4).

sealed tube running at 30 kV/30 mA, a HOPG-primary monochromator, a 500 µm monocapillary optic with a 300 µm pinhole at the exit, resulting in a beam diameter of 300  µm and a large VÅNTEC-500 2-dimensional detector covering 40° in the 2θ and Chi-range. Due to this large detector only two detector positions are necessary, slightly overlapping, to cover a 2θ-range from app. 7° till 70°. The measurement was performed with a fixed incident angle of 10°, the measurement time of each detector image was 1 minute. The samples were not rotated during the measurement to avoid the smearing of single crystal spots causing the loss of information on the differences in crystallinity of the different samples. Therefore, the measured intensities are typically not fulfilling the requirement for perfect crystallite statistics in such an unground sample due to the small measurement volume. For phase identification the PDF-2 database was used, while the semi-quantitative estimation of the mineral content was done with the Rietveld software package SIROQUANT (Sietronics).

In contrast, polished piece DOS2012-30967 consists of a slightly finer crystalline hematite due to the significantly broader reflections compared to the X-ray pattern of DOS2010-4952, but coarser hematite crystals are also present in this material, also visible from the single crystal spots in the corresponding detector image. The quartz content is slightly increased by c.2  wt%, as is magnetite by c.4  wt%, which is clearly visible in the X-ray pattern (Figure 13.5). The magnetite content in this piece was also verified with a simple test of the magnetic property using a Neodym magnet. Nevertheless, the quantification done by the Rietveld analysis gives only a rough estimation of the mineral proportions, due to the single crystal spots in the hematite reflections, which falsify the measured hematite intensities, thus influencing the calculation of the mineral content.

3 Results

The axe head (DOS2012-22115) mainly consists of hematite and no quartz could be detected by X-ray micro-diffraction; again, magnetite is clearly visible in the diffraction pattern (Figure 13.6). A semiquantitative estimation of the magnetite concentration by the Rietveld analysis suggests amounts of c.7  wt%. This significantly higher magnetite content is also compatible with the much stronger magnetic behavior compared to polished piece DOS2012-30967. In contrast to the two polished pieces, however, the hematite used for the axe head is fine crystalline, clearly visible

3.1 Measurements of the polished pieces and axe head The supposed mineral content by color and density of the two polished pieces (DOS2010-4952, DOS201230967) and the axe head (DOS2012-22115) could be verified by µ-XRD2 mainly as hematite, although significant differences were identified through a detailed examination of the diffraction patterns and the detector image: the polished piece DOS2010-4952 consists only of hematite with minor amounts (less 304

Philipp Drechsler et al.: Hematite Objects and the Use of Red Pigments

Figure 13.4. µ-XRD2 detector image of the first frame and corresponding pattern from both measured images of polished piece DOS 2010-4952. The spotty diffraction rings of the hematite intensities due to some coarse hematite single crystals in the measured area are clearly visible in the detector image.

Figure 13.5. µ-XRD2 detector image of the first frame and corresponding pattern from both measured images of polished piece DOS 2012-30967. The spotty diffraction rings of the hematite intensities due to some coarse hematite single crystals in the measured area are clearly visible in the detector image. In addition, the piece is shown during measurement in the microdiffractometer.

3.2 Measurement of the potsherds

through the diffraction rings in the detector image that do not show any single crystal spots. According to the micro-diffraction study, the hematite crystals themselves are possibly slightly textured, as the visible differences of measured peak intensities of certain reflections compared to the theoretical intensities of the PDF-card suggest.

µ-XRD2 measurements of two potsherds with traces of a red pigment clearly indicate that the red pigment mainly consists of hematite. No other Fe-containing pigment minerals could be identified by X-ray diffraction. The red pigment on DOS 2011-14733 had been applied to the outer (convex) surface of the potsherd and consists of 305

Dosariyah

Figure 13.6. µ-XRD2 detector image of the first frame and corresponding pattern from both measured images of axe head (DOS 2012-22115). The diffraction rings of the hematite are without any single crystal spots in contrast to both the polished pieces. Only magnetite as a second mineral is detectable, quartz is below the detection limit (< 0.5 wt%).

Figure 13.7. Typical µ-XRD2 patterns of the red painted and unpainted area on the outer surface of potsherd DOS2011-14733. Mainly hematite and Mg-free calcite is visible in the pattern of the red-painted area.

hematite with additional amounts of +/- Mg free calcite (Figure 13.7) which could not be detected on the uncolored surface (matrix) of the potsherd. Therefore, the calcite should be considered as a component of the pigment material.

pigment. Despite the fact that the red pigment cover is very thin, comparing the µ-XRD2-patterns of colored and uncolored areas of different locations on this potsherd, one additional reflection at 38.6°2θ is visible in all the patterns of the colored areas. This is exactly the position of the main 104-reflection of hematite. No traces of other Fe-containing pigments were detectable in any measurement. Typical patterns of a red-colored

The concave (inner) surface of potsherd DOS 2011-14185 is almost completely covered with a reddish but flimsy 306

Philipp Drechsler et al.: Hematite Objects and the Use of Red Pigments

Figure 13.8. Typical µ-XRD2 patterns of a red-colored and uncolored area on the concave inner surface of potsherd DOS 2011-14185. The position of the hematite 104 reflection in the pattern of the red-colored area is marked.

and an uncolored area are shown in Figure 13.8 and therefore, this pigment cover should also be mainly composed of hematite.

itself (see Chapter 11).9 Two complete and two semifinished axe heads are reported from coastal H3/AsSabiyah, Kuwait.10 Similar polished axes also occur at broadly contemporaneous inland sites on the Arabian Peninsula11 and in Mesopotamia.12 Their production from solid rock, together with substantial traces of wear along the cutting edges visible on several items, suggest their use as working tools. Similar traces of wear along the cutting edge can be observed on the hematite axe head from Dosariyah. In addition, the physical characteristics of fine crystalline hematite that is characteristic for this particular item, a Mohs hardness between 5 and 6 that is only slightly softer than quartz,13 its toughness and specific weight would make it an excellent raw material for axe heads.

4 Discussion Hematite objects represent an unusual but repeatedly occurring category of material culture in the central Gulf area during the Arabian Middle Neolithic. In addition to the three hematite objects identified from Dosariyah, three axe or adz heads made from the same material were also found in Qatar. The hematite objects from Dosariyah fall into two categories: (potentially) utilitarian and (potentially) non-utilitarian. The axe head (DOS2012-22115) morphologically resembles a tool type which occurs at Dosariyah and other Arabian Middle Neolithic sites. The fact that these axe heads were made from a wide variety of different solid rocks leaves open the possibility that hematite represents one potential choice for the manufacture of axe heads as working tools. In contrast, the use of hematite as a red pigment suggests a non-utilitarian application. This particular practice is documented at Dosariyah both from the two fragments of polished pieces but also from traces of red pigment on Black-on-Buff pottery and a reddish sediment color observed in trench E1 during excavation.

Nevertheless, the exclusively utilitarian character of similar hematite axes was already questioned in the publication of the three axe heads found in Qatar.14 Although fine crystalline hematite fulfills the requirements of hardness and toughness, the raw material is not commonly available in eastern Arabia where most parts of the landscape are covered by Tertiary marine and Pleistocene and Holocene terrestrial sediments which can almost be excluded as natural deposits for hematite. Exceptions are alluvial deposits of the Pleistocene age that can be found in the

Stone axes are repeatedly documented at Middle Neolithic sites in the eastern part of Arabia: two examples of ground axes of similar shape and size but made from solid rock are known from Dosariyah

Masry 1974: fig. 49. Kallweit and Davies 2010: 119. Sordinas 1978: pl. 16:1–4; Kiesewetter 2006: fig. 2.8f. 12 Kallweit and Davies 2010: 127. 13 Deer et al. 2011. 14 Drechsler et al. 2013. 9

10 11

307

Dosariyah The first µ-XRD2 measurements on two potsherds with traces of red color are further proof of the use of hematite as a red pigment. No other minerals which are possibly responsible for the reddish color could be detected by X-ray diffraction. This pigment can therefore be directly related to the two polished pieces. The Mg free calcite content of the color powder also identified on potsherd DOS 2011-14733 might be a relict of the production of the pigment: it could derive from a limestone mortar that was used for the preparation of the red powder.

southern part of Qatar. These deposits bear at least ironrich quartzite pebbles.15 Similar deposits can be found on the eastern Arabian mainland as well, where parts of the Arabian Platform are covered by alluvial deposits originating in the Arabian Shield. One definite source of fine crystalline hematite is the island of Sir Ban Yas, off the coast of Abu Dhabi.16 Crystalline hematite also occurs on other islands within the lower Gulf: Arzaneh island, Abu Mousa island, Hormuz island and Hengam island.17 Other sources are known from the northern and western part of the Kingdom of Saudi Arabia, as well as Oman. Differences in the mineral composition between the pieces from Qatar and the hematite/ magnetite axe head from Dosariyah suggest various deposits or the exploitation of different raw material sources. Differences regarding crystallinity, texture and also the magnetite content of the hematite objects from Dosariyah point into the same direction and could therefore be important for the estimation of possible raw material sources.

A similar red pigment, most likely also hematite, was found on an additional 59 pieces of Black-on-Buff pottery. In most cases (N = 51), traces of red color were observed on vessel fragments that represent parts of the vessel body. Base fragments with red pigment rarely occur (N = 3) while it is present on shoulder (N = 2) and rim fragments (N  =  5). In 47 cases, the pigment was found exclusively on the interior of the fragments. Only three pieces show red color on the exterior, while in three cases pigment was found both on the exterior and interior face of the former vessel. In one case, red-colored areas were found on the interior face and on one broken edge, while the exterior remained free of red color. The fact that red pigment was mainly found on the interior face of vessel fragments might imply that the vessels once contained some substance which was stained with hematite, but the fact that in several cases the color application was incomplete and patchy (Figure 13.8) contradicts this assumption. At least for part of the potsherds it can be assumed that they served as palettes used in the processing of the hematite powder.

The shape of all hematite axe heads from the central Gulf area further convincingly indicates that the raw material itself was rather rare. Both the hematite axe head from Dosariyah and all three hematite axes from Qatar are not completely finished by grinding, but still show their original surfaces. It can therefore be argued that the raw material was too valuable for much to be removed during the process of manufacture. The second argument for a non-utilitarian character of the hematite axe head comes from the raw material itself. During the manufacturing process, an intensively dyed red powder, or, considering the use of water during the grinding process, a blood-red colored liquid appears. Finds of two polished pieces, two grinding stones and several potsherds with traces of a red color confirm the use of hematite as a red pigment. Direct evidence of the extraction of the red pigment from the two polished pieces is lacking. Considering their rather small size and extent of grinding facets, only minor quantities of powdered hematite could have been extracted from them. Their exact use remains unknown as no small — and potentially grooved — grinding stones or ‘files’ were found.

Although the archeological evidence from Dosariyah is hardly sufficient to unambiguously explain the function of red pigment within the society, it does suggest some kind of non-utilitarian use. In the central Gulf, the use of hematite for the production of a blood-red color is not restricted to Dosariyah. At Shagra, located in the southeastern part of the Qatar peninsula, a ‘molette’, or hand stone, was found that shows traces of coloring with a red mineral,18 most likely also hematite. In southeast Arabia, shell beads with traces of red color, possibly deriving from a pigmented thread or sinew that was used to sew the beads onto a piece of garment, were documented from the Middle Neolithic graveyard al-Buhais 18.19 Finds of ‘red ochre’ are likewise documented in Mesopotamia. At Tell Abada, located in the Hamrin region of Iraq, both ‘large quantities of red ochre’ and ‘grind-stones still bearing its traces’ were found in Ubaid contexts.20 ‘Red ochre’ was also

Nevertheless, an insight into the processing of hematite on a larger scale provides a grinder that shows traces of red pigment (DOS2012-30460). Its large size suggests that remarkable quantities of color powder must have been prepared at one and the same time. This observation finds its confirmation in the reddish color of sediment layers in trench E1.

15 16 17

P. Drechsler, personal observation, 03/2014. Aston 1985; Edgell 2006: 419. Edgell 2006: 419; www.mindat.org.

18 19 20

308

Inizan 1988: 119. de Beauclair 2005: 24 Jasim 1985: 18, 75, fig. 72.

Philipp Drechsler et al.: Hematite Objects and the Use of Red Pigments mentioned in prehistoric and later funeral contexts throughout Mesopotamia.21

production of axe heads and for the preparation of a red pigment. The size, shape and traces of wear along the cutting edge of the hematite axe head from Dosariyah suggest that it had been used as a working tool. Nevertheless, the formal determination of this object as an axe head derives first and foremost from its observable morphology and it cannot be excluded that it was also used as a polisher or a source for producing hematite powder. The latter function can be convincingly assigned to two polished pieces that show clear grinding facets along their edges. With the help of such pieces, a deep red pigment could be produced. Remains of powdered hematite were identified on two potsherds using µ-XRD2-techniques. Traces of red color were found on a further 59 sherds of Black-onBuff pottery, two hand stones and a shell bead (see Chapter 15). The visual similarity of the red pigment observed on these objects with the securely identified hematite suggests that all traces of red color observed at Dosariyah derive from the latter material. This broad spectrum of objects and their spatial distribution at the site indicates a comparatively frequent use of red pigment at Dosariyah, possibly in different aspects of daily life. Nevertheless, the scarcity of natural deposits indicates that hematite must have been regarded as a precious material. Differences in the mineralogical composition may help identify these natural deposits and in future, potentially reconstruct exchange and communication routes.

The function of the red color that can be produced from hematite, often inaccurately referred to as ‘red ochre’, during prehistory is highly debated. Extensive use of ‘red ochre’ in burial contexts22 and in association with figural art and personal adornment23 suggests some kind of symbolic use as early as the European and Near Eastern Middle Paleolithic and African MSA. Similarly, it was, and still is, commonly used to decorate the body in some African and Australian societies24 but the application of red ochre on the human body also detains UV light from solar radiation and is both antimicrobial and anti-parasitic.25 Moreover, powdered hematite might have been used for the tanning of leather,26 the preservation of food or as a component in hafting materials.27 5 Conclusions There is increasing evidence that hematite played a similar role in eastern Arabia as it did in other parts of the world during the Paleolithic and Neolithic. Although the finds of hematite objects and red pigment that has been produced from hematite are not conclusive enough to explain its actual application, both utilitarian and non-utilitarian functions can be assumed. Hematite served both as a suitable raw material for the

E.g. Eridu (Safar et al. 1981: 133), Tell es Sawwan (El-Wailly and esSoof 1965: 23) and Tell Abu Dhahir (Simpson 2007). Many thanks to St John Simpson for this note. 22 Vandermeersch 1969; Aldhouse-Green and Pettitt 1998; RielSalvatore and Clark 2001. 23 Zilhão et al. 2010. 24 Basedow 1925; Knight et al. 1995. 25 Velo 1984. 26 Audouin and Plisson 1982. 27 Wadley et al. 2004. 21

309

Chapter 14 Bone Industry Philipp Drechsler 1 Introduction

The overall length of the completely preserved type 1 awls are highly standardized and range between 100 and 151 mm, probably representing the minimal length suitable for use (Figure 14.2a). The length of the tips, measured from the point where the bone shaft has been ground down to the tip of the awl, ranges between 32 and 63 mm (Figure 14.2b). In five cases, the tips of the awls show old fractures, suggesting their discard after breakage. Six piercing awls are completely preserved while in one case the tip was broken during excavation.

The state of preservation of the bones at Dosariyah was outstanding. This is in contrast to most other prehistoric sites in Arabia and is a direct result of the fact that the dense accumulations of mollusk shells and fish bones provided a localized chemical environment in the soil with a high concentration of phosphate (high pH-value) which prevented the decomposition of bones, but although the mineral part of the bones is well preserved, all of the organic components have completely disintegrated.

One completely preserved bone tool morphologically resembles a piercing awl but had been manufactured from a different source of solid bone (DOS2010-3725). It has a slender, flat and asymmetric tip and produced elongated rather than round piercings (Figure 14.1d). Due to these different characteristics it has been classified as awl type 2.

Because of these exceptional preservation conditions, it was no surprise to find pieces of worked bone, both during surface sampling and during excavation and a total of 65 pieces were recorded from these investigations.1 Moreover, the presence of waste pieces clearly indicates the production of bone tools at the site. The bone industry from Dosariyah is now among the largest assemblages known from Neolithic Arabia.

Another eight bone tools most plausibly represent smaller fragments of piercing awls. They all show traces of workmanship but are either broken at the tip or shaft. While seven plausibly represent fragments of piercing awls type 1, this cannot be verified for one other pointed bone tool which was unfortunately broken during excavation (DOS2012-29586).

2 Bone Tools A total of 52 pieces of worked bone fall formally within the category of bone tools, thus representing objects that could have been used to manufacture other items. Most of these tools are pointed and most plausibly were used for piercing. As there is no standardized classification for Neolithic bone tools in Arabia, a new classification was established for this category of finds from Dosariyah. This is based exclusively on the morphological characteristics of the worked bone and does not consider use-ware studies.

Pointed bone tools commonly classified as awls repeatedly occur in Neolithic contexts on the Arabian Peninsula, although in low numbers. From settlement contexts at Ra’s al-Hamra 6 (RH6), located at the southern end of the Batinah coast in Oman, two (broken) bone points were found which closely resemble the broken tips of piercing awls from Dosariyah.2 Similar pointed bone tools were found at Akab Island near the Strait of Hormuz,3 where at least one of the bone tools4 again closely resembles the bone points from Dosariyah (see below). More extensive is the bone tool assemblage from Ra’s al-Khabbah (KHB-1), a fishing settlement on the Indian Ocean located in the Ja’alan area of the Sultanate of Oman. Dating to the late fifth and fourth millennia BC, 22 bone tools were found there and were classified into large points, awls and needles with eyes.5

2.1 Piercing awls The majority of the pointed bone tools (N  =  13) can be described as piercing awls. With one exception, all were made from the metatarsals and metacarpals of domesticated sheep (Ovis aries) and hunted gazelle (Gazella sp.) (Table 14.1) with the distal end of the bone forming the handle of the awl (piercing awl type 1). The medial part of the bone shaft forms the shaft of the awl, while the proximal part is reworked into a sharp point (Figure 14.1a–c).

2

Two additional pieces of worked bone were identified during the study of finds excavated by A.  Masry in 1972. The total number of pieces of worked bone known from the site is therefore 67. 1

3 4 5

310

Marcucci et al. 2014: 239, fig. 3.13–14. Charpentier and Méry 2008: 127, 124, fig. 9.10–11. Charpentier and Méry 2008: 124, fig. 9.11. Cavulli et al. 2009: 73, 74, fig. 1.1–13.

Philipp Drechsler: Bone Industry

Figure 14.1. Piercing tools, punches and chisels from Dosariyah. a-c Type 1 piercing awls made from metatarsal/ metacarpal bones (a DOS2012-24116; b DOS2012-31044; c DOS2012-25713); d Type 2 piercing awl made from compact bone (DOS2010-3725); e punch (DOS2012-22276); f chisel or smoothing tool (DOS2012- 30109.48) (drawings: B. Hoepfer, F. Brodbeck).

similarity to the piercing awl type 2.9 Other pointed bone tools found at H3/As-Sabiyah are described as fish gorges,10 a category of bone tools which was absent at Dosariyah. Nine bone objects have been reported from Tell el-Oueili, among them one object which might be part of a piercing awl type 1,11 while a second broadly resembles a type 2 piercing awl.12 A few bone artifacts, at least two of them pointed, are present at Eridu.13 A more extensive assemblage of bone tools was excavated

Two bone awls were also found at Khor Milk 1, located in the Muscat area of Oman.6 In the upper Gulf area and southern Mesopotamia, comparable finds are less frequently documented.7 At least one worked bone object from H3/As-Sabiyah resembles a broken tip from a piercing awl type 1,8 while a second bone point found there shows some degree of

Uerpmann H.-P. and Uerpmann M. 2003: 111, fig. 7.4.6, 10. The relatively low numbers reported in the older literature from Mesopotamia probably reflects a combination of factors, including more selective retention and publication. 8 Crawford and Carter 2010: 80, fig. 4.8:1.

Crawford and Carter 2010: 80, fig. 4.8:11. Crawford and Carter 2010: 82. 11 Breniquet 1996: 160, pl. IV.2. 12 Breniquet 1996: 158, pl. II.1. 13 Campbell Thompson 1920: pl. IX.

6

9

7

10

311

Dosariyah Table 14.1. Determination of bones used for the production of piercing awls type 1.1 Find ID

Bone

Preservation

Determination

DOS2010-8413

Metatarsal bone

Distal and shaft

Sheep (Ovis aries)

DOS2011-20480.2

Metapodial bone

Shaft

not determined

DOS2011-20480.1 DOS2012-24116

DOS2012-24842

DOS2012-25713 DOS2012-25714 DOS2012-25937 DOS2012-26155 DOS2012-27261 DOS2012-31044

DOS2012-31332.65

Metapodial bone Metatarsal bone

Metatarsal bone

Metatarsal bone Metatarsal bone

Metacarpal bone

Metapodial bone Metatarsal bone Metatarsal bone

Metapodial bone

Shaft

Distal and shaft

Distal and shaft

Distal and shaft Distal and shaft Distal and shaft Shaft

Distal and shaft Distal and shaft Shaft

not determined

Sheep (Ovis aries)

Sheep (Ovis aries) Gazella spec.

Gazella spec.? Gazella spec.

Small ruminant Gazella spec. Gazella spec.

Sheep (Ovis aries)

The identification of the bones was kindly provided by M. Uerpmann and H.-P. Uerpmann and carried out on the basis of photographs.

1

Figure 14.2. Measurements of type 1 piercing awls. a Length of completely preserved pieces (N = 8). Similarities in size indicate a highly standardized tool; b length of tips (N = 9). Note the higher variability in comparison to the overall length of the completely preserved tools.

at Tell Abada, Hamrin region of Iraq. It consists of 49 bone artifacts, with the majority of them awls (N = 41) followed by perforators, scrapers and spatulas.14 The ‘metapodial awls’ from Tell Abada broadly resemble the piercing awls type 1 from Dosaryiah, although they are generally shorter and more compact.15 In contrast, the ‘flat splinter awls’ and ‘small spear-like awls’ from Tell Abada show morphological similarities to the piercing awls type 2 from Dosariyah.16 Also frequently found were bone artifacts in Ubaid/Late Halaf levels at Tepe Gawra in northern Mesopotamia. Here, bone implements that were made from metapodial bones retaining the jointed end to serve as the butt of the tool, closely resembling the type 1 piercing awls from Dosariyah, were found in 14 15 16

a well in stratum XIII;17 additional piercing awls came from strata XVI and XVII.18 2.2 Bone needles The second frequent category of bone tools in Dosariyah is needles. Seven slender pieces of worked bone were made from compact bone and show a characteristic eye (Figure 14.3), qualifying these as needles. The sizes of the bone needle fragments range up to 127  mm in length for an almost completely preserved example, suggesting impressively large tools. Diameters are generally oval or round, with a maximum width between 4.5 and 7.4 mm and a maximum thickness between 3.1 and 5.3  mm. Another eight slender pieces of worked

Jasim 1985: 85. Jasim 1985: figs 88.b–c, e, 89.a–c. Jasim 1985: 85; fig. 90.a–d.

17 18

312

Tobler 1950: 214, pl. XCIX.a. Tobler 1950: pl. XCVIII.b.

Philipp Drechsler: Bone Industry

Figure 14.3. Bone needle fragments from Dosariyah. a, g Needle shafts (a DOS2010-3618; g DOS2011-18182); b–f, h needle shafts with eyes (b DOS2010-10076; c DOS2011-17375; d DOS2011-14726; e DOS2011-14779; f DOS2010-2540; h DOS201112882) (drawings: B. Hoepfer, F. Brodbeck).

8

Length (mm)

6

4

2

2

4

6

8

10

Width (mm)

Figure 14.4. Maximum width and thickness of bone needles with preserved eye (diamond) and medial fragments without eyes (square). The overlapping/slightly smaller sizes of medial fragments suggest that these pieces also belong to bone needles.

compact bone, broken at either end, show a comparable cross section, width and thickness and most plausibly represent fragments of broken bone needles (Figure 14.4). While sometimes broken during excavation, five show sintered sand on the breaks, which indicates breakage and discard shortly after use. Likewise, four

out of eight potential bone needle fragments without eyes show old breaks. In general, needles with eyes are suitable for the repair or manufacture of textiles, using yarn that was potentially prepared with the help of spindle whorls 313

Dosariyah

Figure 14.5. Bone points made from compact bone; note the carefully shaped bases (a DOS2011-19883; b DOS2010-8745; c DOS2010-7809) (drawings: F. Brodbeck).

2.4 Punches and chisels

(see Chapter 10). The small size of the eyes mirrors the availability of thin threads. Nevertheless, the large size of the needles might also point to alternative functions. The occurrence of morphologically identical bone needles at the fishing settlement at Ra’s al-Khabbah (KHB-1), located in the Gulf of Oman and dating to the late fifth and fourth millennia BC, suggests that these tools could also have been used for the preparation of fishing equipment.19 Conspicuously, no bone needles with eyes were described from Bahra 1/As-Sabiyah and H3/As-Sabiyah. They are present, however, at the Ras al ‘Amiyah in central Iraq. A single bone point and four bone needles were found at this site, at least one of the latter showing an eye.20 The bone needles have been described as ‘remarkable for their wide, flattened heads and often considerable size’.21

Four large pointed bone tools cannot be assigned to one of the three categories of tools described above. Based on their larger dimensions and a rather massive general appearance they are best described as punches (with a pointed tool tip) and chisels (with a rounded or angular tool tip), although the latter might have been used as some form of ‘smoothing tool’ owing to their large and slightly convex smooth surfaces. One completely preserved punch is made from a piece of flat, compact bone (DOS2012-22276, Figure 14.1e). While the pointed end shows clear traces of workmanship, the basis remains unworked. A second, irregularly shaped and broad bone point (DOS201221855) is broken off from the shaft and might also represent the working tip of a broad punch.

2.3 Bone points

Two other pieces can best be described as chisels. Both are made of flat compact bones and show a chiselshaped working end. One is completely preserved (DOS2012-30109.48, Figure 14.1f) and shows no traces of workmanship opposite the working edge. It finds its closest parallel in the bone tool assemblage from Ra’s al-Khabbah (KHB-1), where it was described as a large point.23 Similar bone tools were found at Tell Abada and other Mesopotamian sites where they have been described as ‘spatulae’.24 A second piece from Dosariyah was broken during excavation and only the working end is preserved (DOS2012-31311.1): its chisel-shaped tip shows conspicuous traces of workmanship — or tool use — on the surface, running at an angle of about 45° to the axis of the original tool.

Three pieces of worked bone most likely represent tips of arrows or spears (Figure 14.5). Broadly comparable in size, all three show common morphological characteristics. They are made from pieces of compact bone and have a slender pointed tip. The bases of the bone points are clearly reduced in width and thickness by grinding and/or gnawing to facilitate their hafting in a shaft. The absence of a pointed base makes it highly unlikely that these bone points represent fishing gorges, although Crawford and Carter interpret similar pieces of worked bone as such.22

19 20 21 22

Cavulli et al. 2009: 73, 74, fig. 1.10–12. Stronach 1961: pl. 43.7. Stronach 1961: 106. Crawford and Carter 2010: 80, 82, fig. 4.8:9–10.

23 24

314

Cavulli et al. 2009: 74, fig. 1.4. Jasim 1985: 86 and references; fig. 90.g–i.

Philipp Drechsler: Bone Industry

Figure 14.7. A piece of worked bone with carefully shaped and smoothed notches of unknown function, or production waste (DOS2012-24418) (photograph: B. KiepenheuerDrechsler).

Figure 14.6. The medial fragment of a ray spine with traces of wo1rking on both sides, suggesting its use as a tool (DOS2012-26506) (photograph: B. Kiepenheuer-Drechsler).

2.5 Other pieces of worked bone

were interpreted as armatures.25 Likewise, the function of object DOS2012-24418 remains unclear. It consists of a broken piece of long bone with one end of the bone shaft preserved, while the joint itself has been removed along the epiphysis (Figure 14.7). With the exception of four large, carefully smoothed notches, this piece of bone does not show additional tool marks and it cannot be excluded that this piece represents production waste.

Nine bone objects show clear traces of working, but could not be classified reliably. Most are only partially preserved and some or all might have been discarded during production. Two pieces most likely represent the remains of pointed bone tools: one of these is completely preserved and shows a clear point at one end, while the other end is formed by the joint; two irregular, circular cuts close to the joint suggest that it was discarded during production (DOS2010-405). A second pointed bone object may represent a piercing awl but is very badly preserved as it was a surface find (DOS2011-11241). Two partially preserved small pieces of worked compact bone might represent the remains of needles or small bone rods (DOS2011-20480.3; DOS2012-28568.1). Similar in width and thickness, a third has an asymmetric pointed end and is made from compact bone; it is highly polished with clear traces of grinding from its manufacture or use (DOS201115225). Another, slightly larger, bone rod consists of the poorly preserved medial fragment of a slender piece of compact bone with some remains of spongy bone preserved (DOS2012-27480). Another bone object can best be described as an (informal) chisel and is a short broad piece of compact bone with clear traces of chipping at both ends which resemble the ‘splintered pieces’ made from flint (DOS2012-21899).

3 Production Waste A total of 13 pieces represent waste pieces from the production of bone objects. Eleven of these are morphologically identical and consist of joints sawn from equid metapodial bones; while the compact bone is clearly cut through, the remaining spongeous bone was finally broken off (Figure 14.8).26 The fact that as much as 85% of the waste pieces are morphologically identical suggests a rather standardized production of bone artifacts at Dosariyah. The two exceptions are a joint sawn off from a small long bone (DOS2012-28521) and a long bone fragment at the transition towards the joint which shows clear traces of breakage and impact fractures/hammer marks, as well as a row of small incisions (either intentionally made or gnawing marks) (DOS2012-29979).

The medial fragment of a ray spine shows traces of working along both faces (DOS2012-26506, Figure 14.6). Its use as an arrow or spear tip can be hypothesized, but not proven as both the tip and base are broken and missing. Similar, but more completely preserved pieces, found at Ras al-Hadd 6, located on the Indian Ocean coast of Oman and dating between 3100 and 2700 BC,

Charpentier et al. 2009: 14, figs. 8–9. The joints were determined as parts of the metapodial bones of Equus sp., most likely wild ass. The identification of the bones was kindly carried out by M. Uerpmann and H.-P. Uerpmann on the basis of photographs. 25 26

315

Dosariyah

Figure 14.8. Waste pieces from the production of bone tools: joints carefully sawn off from metapodial bones (Equus sp., most likely wild ass). After the solid part of the bone was cut the spongious part was broken off (a DOS2012-25743; b DOS2012-26955; c DOS2012-29835)1 (photographs: B. Kiepenheuer-Drechsler). 1

The identification of the bones was kindly provided by M. Uerpmann and H.-P. Uerpmann on the basis of photographs.

their use and/or discard associated with some kind of domestic activities here (Figure 14.10). One piercing awl was found in both trenches N3 and S3 but were absent from trench S2. The spatial distribution of needles with eyes shows a reverse picture: they only occur in trench S2 where they were mainly found associated with the settlement horizons in the lower part of the stratigraphic sequence (stratigraphic unit S2-IV, Figure 14.11). Nevertheless, this distribution of needles and needle fragments is also (at least in part) the result of a sampling bias as, in contrast to all other trenches, the sediment excavated in trench S2 was completely sieved. Likewise, all three bone points were found outside shell deposits in the southern area of the site in trench S2 and S3. In contrast, punches and chisels occur exclusively in the northern and western part of the site, predominantly in trench E1.

4 Spatial Distribution of Worked Bone With the exception of trench N2, worked bone was found in all trenches excavated at Dosariyah, providing the basis for the study of the spatial distribution of worked bone. In contrast, finds of worked bone on the surface of the site are rare (N = 8) and those pieces are therefore excluded from the respective analyses. The spatial analysis of the worked bone reveals clear differences between areas and trenches (Figure 14.9). An accumulation of waste pieces can be observed in trench E1, suggesting the production of bone tools in an area characterized by traces of occupation (post-holes, combustion features, shallow pits). Waste pieces occur less frequently in trenches N1, N3, S1 and S2. Both in trenches S1 and S2, waste pieces occur in the lowermost part of the stratigraphic sequence, characterized by layers of occupation debris.

5 Conclusions

Piercing awls, the most frequent category of bone tools, shows a similar distribution. The majority of piercing awls (complete: N  =  9, fragments: N  =  4) were found in trench E1 (stratigraphic unit E1-III), suggesting

The bone industry from Dosariyah provides insights into a little-known aspect of Middle Neolithic material culture on the Arabian Peninsula. It is dominated by two characteristic forms: piercing awls and needles with 316

Philipp Drechsler: Bone Industry

Figure 14.9. Spatial distribution of worked bone. a Production waste: joints sawn off from bone shafts; b completely preserved (cross) piercing awls and fragments (open cross); c needles with eyes (dot) and needle fragments (open circle); d bone points (asterisk), punches (dot) and chisels (triangle).

eyes. Together, these represent 69% of the whole bone tool assemblage which appears highly standardized. This conclusion finds support in the selective use of specific bones as raw material for making bone tools, as the majority of production waste (joints sawn off from bone shafts) is represented by the distal ends of metapodial bones from equids, whereas piercing awls were made from the metapodial bones of small ruminants, clearly indicating a general preference for

metapodial bones for the production of bone tools at Dosariyah. The bone tool assemblage does not indicate a functional specification. Awls can be used to pierce a wide variety of organic materials, including leather and hide. Needles with eyes are suitable for the repair or manufacture of textiles but could also have been used for the preparation of fishing nets. The fact that both 317

Dosariyah

Figure 14.10. Stratigraphic distribution of worked bone in trench E1. Complete (cross) and broken (open cross) piercing awls are predominantly associated with the occupation layers of stratigraphic unit E1-III, while production waste (open diamonds) was found both in stratigraphic units E1-II and E1-IV. Two punches come from the upper stratigraphic unit E1-II (dots). Noteworthy is the absence of bone artifacts in the lower part of the stratigraphic sequence (stratigraphic units E1-V–E1-VII). Trench E1, north section.

Figure 14.11. Stratigraphic distribution of worked bone in trench S2. Needles with eyes (dot) needle fragments (open circle) and production waste (diamond) are almost exclusively associated with stratigraphic unit S2-IV, characterized by settlement deposits. Likewise, bone points (asterisk) were found outside shell accumulations. Trench S2, west section.

less well clear as these pieces are either too fragmented or unspecific in their shape.

piercing awls and bone needles were found in spatial association with occupation debris suggests their use in domestic contexts.

The spatial distribution of pieces of worked bone at the site shows a clear patterning. Production waste occurs in all excavated areas, which suggests that the manufacture of bone tools was not restricted to specific workshop areas. Instead, bone tools were made within the context of other domestic activities. Likewise, the majority of bone tools were found associated with occupation layers. Differences in the frequency of specific bone tool forms, however, suggest some kind of spatial differentiation of tool use and associated activities at the site.

Other types of fishing equipment which could have been made from bone, such as fishing gorges with their typical triangular shape, do not occur at Dosariyah. Only three bone points with a carefully thinned base can be considered as projectile points and thus potentially fall into the area of hunting equipment. The same may be true of the fragment of a worked ray spine which could potentially have been used as the tip of an arrow or spear but even this function is unproven. In contrast, the function of other pieces of worked bone is

318

Chapter 15 Personal Adornment Philipp Drechsler 1 Introduction

29701.1–3, Figure 15.1h–k) are very fragile and show an irregular outline. With diameters between 6.9 and 8.0  mm they are very thin (0.1 to 0.9  mm) and show small piercings (1.0 to 1.6 mm) only.

Objects found during excavations at Dosariyah which have an exclusively decorative or ornamental character and are small enough to be worn are considered to be items of personal adornment. This group of artifacts is analyzed in terms of their typology and production technique and the provenience of the raw materials. Once embedded in social contexts, personal adornments were objects of communication, prestige and identity. They therefore provide rare insights into the community occupying Dosariyah beyond the spheres of subsistence and production.

A total of four pierced vertebrae of cartilaginous fishes were identified as a particular form of disc bead (DOS2011-13121; DOS2011-15359; DOS2011-16049, Figure 15.1y–aa). With sizes in the range of 6.2 to 11.4 mm in diameter and 1.4 to 2.6 mm in length, all of these beads fall into the shape range of disc beads. All four pieces are vertebrae split close to the center and carefully smoothed at the fractured surface, resulting in a thinner bead.

2 Categories of Adornment

Worth mentioning is one large disc bead made from mollusk shell (DOS2010-5270, Figures 15.1ab, 4a). Although both surfaces were carefully ground, riblets were still visible on the outer surface of the shell and within the intervening gaps substantial remains of red pigment, most likely powdered hematite, were found. At the Neolithic graveyard at al-Buhais 18, similar traces of red color was recorded for 18 ‘big disc beads’ with a diameter between 8.0 and 11.6 mm, where it was thought to reflect the traces of ‘a pigmented thread or sinew by which the disc beads were sewn flat onto some surface, presumably a piece of garment’.3 The overall distribution of the red color on the surface of the disc bead from Dosariyah nevertheless suggests a deliberate coloration of this particular bead itself (see Chapter 13).

A total of 82 objects were identified as personal adornment. They can be subdivided into beads (N = 74), pendants (N = 5) and other ornamental objects (N = 3), with beads representing the majority of personal adornment at the site. In addition, an unknown number of comparable objects were found on the surface during previous investigations at the site as both Burkholder1 and Masry2 mentioned finding beads. 2.1 Disc beads Disc beads (N = 39; Figure 15.1a–t; ab), characterized by their flat round shape with a central perforation, represent the most common form. Their diameters range from 4.0 to 11.3  mm, but diameters between 4.5 and 7.0  mm are the most common (Figure 15.2). Heights range between 0.7 and 2.7  mm. Piercings are conical or hourglass-shaped, round and smooth, suggesting that they were made individually from one or both sides of the bead with the help of borers (see Chapter 11.4.7). Piercings are located close to or at the center of the round disc and range in diameter between 1.0 and 2.9 mm. Outlines of the disc beads are generally well shaped by grinding, while the two surfaces are not always worked. Raw materials include whitish or orange mollusk shell (N = 32), limestone (N = 2), greenish translucent metamorphic rock (N = 1) and bone (N = 5).

2.2 Tubular beads In comparison to disc beads, tubular beads (N  =  22; Figures 15.1u–x, ac–af) are less frequent and show a greater diversity both in shape, size and raw material. As with other studies of Neolithic beads in Arabia,4 a difficulty arises in distinguishing disc beads from tubular beads. In this study, all round beads with a central perforation and a diameter/length ratio below 2.5 were considered to be tubular beads regardless of size (Figure 15.3). Diameters of tubular beads range between 1.7 and 13.1  mm, with lengths between 1.6 and 25.8 mm. The shape of the holes suggests that the piercings were generally made from both sides.

Three disc beads made from mollusk shell collected from a collapsed profile in trench E1 (DOS20121 2

Burkholder 1972: 266. Masry 1997: 80.

3 4

319

de Beauclair 2005: 24. Uerpmann H.-P. and Uerpmann M. 2003: 103; de Beauclair 2005: 22.

Dosariyah

Figure 15.1. Range of personal adornment from Dosariyah: a–t, ab disc beads (a DOS2011-18391.2; b DOS2011-18391.1; c DOS2011-18391.3; d DOS2010-8313; e DOS2011-16073; f DOS2011-15638; g DOS2012-29701.1; h DOS2012-29701.2; i DOS201229701.3; k DOS2011-12956; l DOS2011-15859; m DOS2011-12850; n DOS2011-15335; o DOS2011-13440; p DOS2011-13440; q DOS2011-16281; r DOS2011-14611; s DOS2011-15611; t DOS2011-13060; ab DOS2010-5270), y–aa disc beads made from split cartilaginous fish vertebrae (y DOS2011-13121; z DOS2011-15359; aa DOS2011-16049); u–x, ac–af tubular beads (u DOS201112993.1; v DOS2011-12993.2; w DOS2010-9801; x DOS2010-8314; ac DOS2011-14839; ad DOS2011-20477.1; ae DOS2011-17772; af DOS2010-6023), ag Strombus shell ring fragment (DOS2010-8756), ah–ai fragments of plaques (ah DOS2012-21202; ai DOS20208739), ak T-shaped obsidian object (DOS2010-5580), al pottery pendant (DOS2011-12573), am fired clay plug (DOS2010-4800) (drawings: B. Höpfer and F. Brodbeck).

320

Philipp Drechsler: Personal Adornment

12 11 10 9 8

mm

7 6 5 4 3 2 1 0

Figure 15.2. Diameter of disc beads from Dosariyah (N = 35).

30 Disc Beads: Diameter/Length>2.5 25

Tubular beads: Diameter/Length  4  mm) or medium-sized (2–3  mm) pearl was found per c.4200 oysters.18 According to these numbers, more than 35,000 oysters must have been opened to obtain the nine pearls > 2 mm found in trench S2 at Dosariyah. As this number roughly corresponds to calculations based on the shell samples collected during the excavation of this trench, it is open to question whether any pearls were deliberately collected after opening the pearl oysters to extract the meat, and the fact that none

of the pearls found at Dosariyah shows any traces of working further questions the intentional collection of pearls for the production of personal adornment. Finds of pearls are not uncommon in the Gulf area during the Middle Neolithic. Nevertheless it is not always clear whether they show traces of working or not. They are commonly found in burial contexts but rarely occur within settlements. A total of 62 pearls, all perforated, were recorded at the Neolithic graveyard al-Buhais 18 (Sharjah Emirate, UAE), which dates between 5200 and 4500 BC.19 At this site, pearls played a particular role in funeral rites. They were repeatedly found in front of the face of the deceased, below the nose and above the upper lip and while semi-perforated pearls were found with male burials, completely perforated pearls were associated with females.20 Three pearls were found

16 All sediments from trench S2 were sieved with a mesh size of 10 and 3  mm, while excavated sediments from other trenches were only partially sieved. 17 Charpentier et al. 2012: 3. 18 Carter 2012: 186.

19 20

325

Kiesewetter et al. 2000; de Beauclair 2005; 2008. de Beauclair 2008.

Dosariyah

Tubular Beads

Disc Beads S2-I Stratigraphic Unit

Stratigraphic Unit

S2-I S2-II S2-III S2-IV

S2-II S2-III S2-IV S2-V

S2-V 0

2

4

6

8

10

12

0

14

1

2

Snail Beads

4

5

6

Pendants

S2-I

S2-I Stratigraphic Unit

Stratigraphic Unit

3 N

N

S2-II S2-III S2-IV S2-V

S2-II S2-III S2-IV S2-V

0

1

2

3

4

5

6

7

8

0

1

N

2

3

4

N

Figure 15.9. Stratigraphic distribution of disc beads, tubular beads, snail beads and pendants in trench S2: disc beads and snail beads occur predominantly in the lower part of the stratigraphic sequence, while pendants and tubular beads are ubiquitous.

at the broadly contemporaneous site of Faya NE15, another Neolithic graveyard located about 14 km north of al-Buhais 18.21 One perforated pearl also comes from a funerary context at Umm al-Quwain UAQ2, which dates to the second half of the sixth millennium BC.22 At the Ras al-Hamra headland, located at the southern end of the Batinah coast, pierced pearls were found in graveyards at Ras al-Hamra 523 and Ras al-Hamra 10.24

unlikely that the pearls found at Dosariyah can be considered as personal adornment, at least in local society, but it cannot be excluded that pearls were items of exchange with other communities on the Arabian Peninsula or in Mesopotamia.27 4 Stratigraphic Distribution of Personal Adornment Most finds of personal adornment at Dosariyah come from trench S2 (N = 53), probably due to the complete sieving of all removed sediment. Only larger pieces were documented from other areas of the site and studies of the spatial and stratigraphic distribution of personal adornment therefore focus on trench S2.

Apart from Dosariyah, only one pearl was documented in a settlement context at H3/As-Sabiyah.25 Although not well preserved, it has a complete perforation and an irregular shape. It was found within a building structure (Chamber 15) at the time when that room was used to manufacture shell beads and other artifacts.26

Beads (N = 49) and pendants (N = 4) occur throughout the stratigraphic sequence of trench S2, but different types of beads show distinct spatial distributions (Figure 15.9). Disc beads (N = 21) are most frequent in the lower stratigraphic units, especially in stratigraphic unit S2-IV which is considered to represent a settlement horizon, but they are almost absent towards the upper part of the stratigraphy. Closely associated with the spatial distribution of disc beads are finds of bead preforms, small broken pieces of Spondylus shell (N = 26 from eight different contexts, see below). Those pieces

Considering the occurrences and contexts of natural pearls found at Middle Neolithic sites in the Gulf area, it is notable that the majority of pearls have been found in burial contexts and only one pearl comes from a bead workshop area within a settlement. Furthermore, most of the pearls are perforated. It is therefore highly 21 22 23 24 25 26

Kutterer and de Beauclair 2008. Charpentier et al. 2012. Salvatori 2007. Santini 1987: 184. Crawford and Carter 2010. Crawford and Carter 2010: 77.

27

326

Phillips 2002: 181; Carter 2012.

Philipp Drechsler: Personal Adornment exclusively occurred in stratigraphic unit S2-III where they were found scattered within a small area in the southern part of the trench.

at least eight of the disc beads found at Dosariyah were made from, suggests local bead production or at least the production of preforms (Figure 15.10). All Spondylus shell fragments were found within a square meter in stratigraphic unit S2-III, indicating a spatially focused human activity (Figure 15.11). Natural agency or accidental breakage can almost be excluded for this occurrence as Spondylus shell is too robust to get broken unintentionally.

Shell beads (N  =  11) widely mirror the distribution of disc beads and occur exclusively in stratigraphic units S2-IV and S2-III. The fact that four of the five pierced shells of Nassarius marmoreus were found close together suggests that they were once part of the same piece of jewelry. Although lower in numbers, beads made from split and pierced vertebrae of cartilaginous fishes (N  =  3) were also found at the very beginning of the stratigraphic sequence in stratigraphic unit S2-IV. In contrast, tubular beads (N = 14) occur in almost equal numbers across the stratigraphy of trench S2. Four pendants were found during excavations of trench S2. While one pendant made from a piece of Ubaid fine ware comes from stratigraphic unit S2-IV, the remaining three pieces — two of shell and the other obsidian — were found in stratigraphic unit S2-II.

Although less frequent compared to other sites where the production of beads has been considered a major task (Abu Khamis,28 H3/As-Sabiyah,29 Bahra 1/AsSabiyah30), some of the flint tools found at Dosariyah could have been used as drills for the perforation of beads. As these borers could also have been used to make repair holes in pottery (see Chapter 11), these flint tools cannot be considered as proof of the production of beads at Dosariyah. 6 Personal Adornment as an Indicator of an Arabian or Mesopotamian Occupation at Dosariyah

According to the distribution of different types of adornment, raw materials show a significant pattern. Beads made from (fish) bone exclusively occur during the early occupation phase at Dosariyah, but fell out of use later. The same holds true for snail shells whereas bivalve shells remained in use continuously. A slight increase of stone as a raw material can be observed although it is present from the beginning of occupation in the southern part of the site. Remarkable is the distribution of exotic raw materials such as rock crystal and obsidian: each of the beads made from these raw materials derives from the uppermost stratigraphic units while limestone and ‘greenstone’ do not show any spatial patterning. In general, a trend towards a higher degree of inhomogeneity and a lower degree of standardization of bead shapes and raw materials can be observed during the occupation of Dosariyah. The spatial distribution of natural pearls is noteworthy. They occur exclusively in the upper part of the stratigraphy in trench S2 associated with shell midden deposits and predominantly eolian deposits, which suggests reduced settlement activities in the southern part of the site. According to this interpretation, the pearls were found in an area mainly used for dumping and waste disposal, supporting the hypothesis that pearls were not considered as valuable items.

Beads and pendants were common elements of personal adornment in Arabian Middle Neolithic societies. They have been repeatedly found in funerary contexts, especially in southeast Arabia. At the alBuhais 18 graveyard, more than 24,000 ornamental objects (beads and pendants) were found spatially associated with more than 420 primary and secondary burials.31 At that site disc beads outnumbered all other pieces of personal adornment while tubular beads of different shapes, sizes and raw material also occur. In addition, snail beads are common: while Ancilla (Sparella) cf. farsiana was the commonest snail, Engina mendicaria was the second frequent gastropod species used for snail beads at al-Buhais 18. Pierced Nerita (N. adenensis) and Mitrella blanda shells were also found but were much less frequent at that site. A similar spectrum of personal adornment comes from the FAY-NE 15 graveyard where three individuals were excavated:32 disc beads dominated the assemblage, while snail beads and tubular beads were less common. In comparison to BHS-18, minor differences in size classes were observed.33 Remains of personal adornment were also found in a funerary context at site UAQ-2, located on the shores of the lower Gulf and broadly contemporaneous with Dosariyah. Associated with at least 42 buried individuals were several types of shell beads and a discshaped soft-stone bead, a number of bitumen beads were found and part of a shell bracelet which resembles

5 Bead Production at Dosariyah? Direct evidence for bead production at Dosariyah is weak and no distinct bead workshops were found. Only one disc bead among 74 pieces (disc beads, tubular beads, snail beads) was documented which is not completely perforated and thus cannot be considered as a finished item of personal adornment (DOS201118391.2). Nevertheless, the discovery of 26 small fragments of Spondylus shell, a raw material of which

28 29 30 31 32 33

327

Masry 1997: 90. Kallweit and Davies 2010: 117. Reiche 2011: 78, Bialowarczuk 2012: 60. de Beauclair 2008. Uerpmann M. et al. 2012: 395, fig.10. Kutterer and de Beauclair 2008: 138; Uerpmann M. et al. 2012: 395.

Dosariyah

Figure 15.10. Intentionally broken Spondylus shell fragments (photographs: B. Kiepenheuer-Drechsler).

Figure 15.11. Stratigraphic distribution of intentionally broken Spondylus shell fragments in trench S2.

a type which occurs on the eastern side of the Oman peninsula.34

Pinctada margaritifera shell.36 These shell pendants occur exclusively in grave contexts and are absent from the occupational areas of RH-5 and RH-10. It has therefore been suggested that these pendants were regarded as funerary ornaments.37 Other types of personal ornaments at RH-10 include double-perforated shark’s teeth and cylindrical stone beads made from steatite and limestone. Disc-shaped shell beads and rectangular, slightly bent, small shell plates with perforations at the four corners once formed bracelets.38 Similar

Broadly comparable, although with different shapes and characteristics, are the assemblages of personal adornment found in graveyards at Ras al-Hamra 5 (RH5) and the slightly earlier site of Ras al-Hamra 10 (RH10), both at the southern end of the Batinah coast.35 At RH-10, shell pendants form the most characteristic and widespread class of ornaments. They are laurelshaped, drop-shaped or eye-shaped and made from 34 35

36

Phillips 2002: 176. Coppa et al. 1985: 99; Santini 1987: 194.

37 38

328

Santini 1987: 194, fig. 8.1–3. Santini 1987: 183. Santini 1987: 195, fig. 9.3.

Philipp Drechsler: Personal Adornment types. The commonest ornaments at H3/As-Sabiyah were disc beads, of which 1,494 pieces were found. 99% of these were made from shell, while a few pieces were made from stone and ceramic. Both finished and unfinished pieces occur. The sizes of the disc beads at H3/As-Sabiyah vary but most measure between 4 and 10 mm in diameter and 1–2 mm in thickness. Specific shell species were selected for these disc beads at H3/AsSabiyah and the cockle Acrosterigma lacunose and spiny oyster Spondylus marisrubi dominate this assemblage.49 Noteworthy is the identification of a disc bead made from obsidian. The spatial distribution of disc beads at H3/AsSabiyah suggests the presence of distinct and spatially well-defined bead workshops within the settlement, where high numbers of unfinished disc beads together with characteristic drilling tools were found.50 A total of 1243 borers represent the majority of stone tools at H3/As-Sabiyah.51 Snail beads are probably the second most common type of personal adornment at H3/AsSabiyah.52 Gastropod shells of Planaxis sulcatus were pierced with a single hole in the body of the shell. Two different techniques were used, either drilling/piercing or abrading the shell surface until a hole appeared, and occasionally a combination of both techniques was used.53 A single Columbella sp. shell was pierced in the same way. Other species including Cerithium caeruleum, Umbonium vestiarium and Clyperomorus bifasciatus were found pierced with single holes although the artificial character of these piercings is uncertain. Other bead types identified at H3/As-Sabiyah include snail beads, tubular beads, whorl beads, otolith beads, bivalve beads and mother-of-pearl plaques. The latter are described as a highly distinctive feature of the H3/As-Sabiyah assemblage: they have four holes and were made of mother-of-pearl from the pearl oyster Pinctada.54 The spatial distribution of these objects suggests that they were locally made at the site during later phases of occupation. The range of personal ornaments found at H3/As-Sabiyah also includes four pendants (one made from a reworked painted Ubaid potsherd),55 shell rings made from Strombus shell and a single pierced pearl.56

types of personal adornment, including disc beads, tubular beads and snail shell beads, were found at numerous sites along the Omani coast, including Wadi Wutayya (predominantly fifth millennium BC),39 Khor Milk 1 (fourth millennium BC),40 Khor Milk 2 (fourth millennium BC),41 and al-Haddah (BJD1) in the Ja’alan (fourth millennium BC).42 Personal ornaments from settlement contexts are known from a number of sites in eastern and southeast Arabia, although in much smaller numbers. In the lower Gulf, assemblages of personal adornment were found on the islands of Dalma and Marawah, off the Abu Dhabi coast of the United Arab Emirates. At Dalma 11, ‘a small number of beads and other personal ornaments were (…) recovered’.43 At Marawah 11, a total of 139 beads were recovered during excavation of MR 11.6/ Area 1 and MR 11.4/Area 2. They were mainly made from marine shell and extremely small, although a few larger and longer tubular beads made of limestone were also found. Overall similarities with beads from other late sixth-/early fifth-millennium BC sites in the Gulf region (Dalma 11, BHS-18) were observed.44 A particular kind of ornament found at Marawah 11 are two doubleperforated ‘buttons’ made from pearl oyster shell which resemble pieces found in a burial at al-Khor in Qatar45 and in settlement contexts at H3/As-Sabiyah in Kuwait.46 Higher numbers of personal ornaments at settlement contexts are only known where excavations have revealed bead workshops at H3/As-Sabiyah and Bahra 1/As-Sabiyah, both located in the Bay of Kuwait and therefore close to southern Mesopotamia. At H3/AsSabiyah, a broad variety of personal ornaments was recorded, among them flanged discs or ‘labrets’ and (fired clay) ‘nails’ or ‘plugs’, forms which are unknown in other Arabian Neolithic contexts. The closest parallels for these objects can instead be found in southern Mesopotamia: flanged discs are known from the site of Uruk/Warka (from unstratified and Seleucid levels) and from Ubaid levels at Tell el-‘Oueili, Ras al-‘Amiya and Ur, and also from sites in southwest Iran.47 Likewise, fired clay nails occur in southern Mesopotamia and Iran, although often these are longer and made from different materials including stone, bitumen and shell.48

The broadly contemporary site of Bahra 1/As-Sabiyah is located about 7.3  km to the west-northwest of H3/ As-Sabiyah but shows a somewhat different range of personal ornaments. Tubular beads made from Conomurex persicus shell were the commonest and were found at different stages of production.57 Flint drills represent the largest group of flint tools at the site and further attest to the presence of at least one bead

Other groups of personal adornment from H3/AsSabiyah closely resemble Arabian Middle Neolithic 39 40 41 42 43 44 45 46 47 48

Uerpmann H.-P. and Uerpmann M. 2003: 45–46, fig. 4.2. Uerpmann H.-P. and Uerpmann M. 2003: 103, 107, fig. 7.3. Uerpmann H.-P. and Uerpmann M. 2003: 133, fig. 7.21, 134. Charpentier et al. 1997: 105–107. Beech and Elders 1999: 19; see also Beech et al. 2016: 17, fig. 6. Beech et al. 2005. Midant-Reynes 1985: pl. II.6. Crawford and Carter 2010: 75. After Crawford and Carter 2010: 67. After Crawford and Carter 2010: 70.

49 50 51 52 53 54 55 56 57

329

Crawford and Carter 2010: 70. Crawford and Carter 2010: 71. Kallweit and Davies 2010: 117. Glover 2010: 183. Glover 2010: 183. Crawford and Carter 2010: 75. Crawford and Carter 2010: 76, fig. 6.4:17. Crawford and Carter 2010: 77. Reiche 2011: 78; 2012: 70.

Dosariyah workshop at this site.58 In addition, circular beads and fragments of pierced plaques, all made from shell, and several snail beads also occur in low numbers at the site.59 One characteristic type of object which might also fall into the category of personal adornment is conical rings made from fired clay, although their function is open to question, as are small irregular pottery discs cut from fragments of Ubaid pottery and pierced in the center. Flanged discs/labrets and a fired clay nail/peg also found at Bahra 1/As-Sabiyah are more definitely identifiable as body ornaments.60

is conspicuous by its absence’.74 Small beads are rarely comprehensively described from Mesopotamia, but often cursorily reported. Bead necklaces were found ‘in a fairly large number’ within burial contexts at Eridu, the vast majority of them appearing to be made of obsidian.75 Others are made from shell and freshwater shells. Notable are repeated occurrences of ‘frit beads’ that were found in the graves. Necklaces made of fired clay beads were a popular form of personal adornment in southern Mesopotamia. As early as the Samarran period, thousands of fired clay beads in a variety of different shapes and sizes were recovered at Choga Mami.76 Clay beads were also common in northern Mesopotamia where a single string of beads made from fired clay was found within occupational debris in the Ubaid level XVI at Tepe Gawra.77 As an exceptional case, beads made from wood were discovered at the Ubaid period grave 1 at Tell Songor A. With a diameter of up to 4 mm, they were found associated with many other tiny beads made of stone as well as larger beads made of obsidian, carnelian and chalk.78

Personal adornment from Ubaid funerary or domestic contexts in southern Mesopotamia is often only cursorily described in the literature. From his excavations at Ur, Woolley mentions ‘beads of shell and steatite and two of amazonite’ from the Ubaid 1 occupation level in ‘Pit F’.61 ‘Studs’ or labrets were common at Ur and were ‘sometimes made of baked pottery, more often of steatite or other stone, including obsidian, wellshaped and finely polished’.62 From graves dug into the so-called ‘flood stratum’ at Ur, a ‘string of small shell and steatite ring beads’ is mentioned from grave T, but other graves lack evidence of personal adornment.63 The spectrum of personal adornment recorded at Tell el-‘Oueili in the Ubaid 0 and Ubaid 1 levels appears more limited.64 Apart from small ‘clous’ (nails) made of stone, several fired clay labrets were found. Two fired clay beads show a characteristic shape: they are flat biconical and have a central perforation; one is decorated with a series of short lines,65 the other is undecorated.66 Likewise, stone labrets and nails/studs found at Uruk/Warka can be associated with the Ubaid occupation.67 They were also found at Eridu68 and Tell al-‘Ubaid.69 Flanged discs/labrets were documented from Ras al-‘Amiyah where they were made from stone and fired clay.70 Plugs/pegs, often found together with flanged discs/labrets, were exclusively made from fired clay at this site.71 The only other ornaments made from stone are two pendants, a small disc-shaped bead of white limestone and three small balls of rock crystal.72 Worth noting is a ‘fish-tailed’ small broken object made from ground obsidian that finds its counterpart in a similar object made from fired clay,73 while ‘shell alone 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73

Considering the range of personal ornaments used across Mesopotamia during the Ubaid period, the different character of beads and pendants occurring at sites in southeast Arabia is evident. While labrets/ flanged discs and plugs/pegs — if they represent ornaments at all — are common objects in southern Mesopotamia, they do not appear in Arabian Neolithic contexts in southeast Arabia. On the other hand, snail beads and disc beads made from shell predominate in Arabian Neolithic contexts but are less frequently found in Mesopotamia. Only (tubular) stone beads might occur in both areas. Dosariyah, like H3/As-Sabiyah and Bahra 1/As-Sabiyah, occupies an intermediate position. The latter two sites, located closer to the major Ubaid settlements of southern Mesopotamia, provide evidence of ornaments characteristic of both the southern Mesopotamian Ubaid and the Arabian Neolithic. At Dosariyah, however, ornaments characteristic of the Arabian Middle Neolithic predominate and the few ornamental objects known from Dosariyah which resemble southern Mesopotamian types should therefore be considered as exotic items rather than personal ornaments. In contrast, disc beads made from Spondylus shell, Engina mendicaria snail beads, tubular stone beads and rectangular plaques are common at Dosariyah and these represent ornaments characteristic of the Arabian Neolithic: they suggest an occupation of the site by local populations. The fact that other kinds of personal ornament such as partially pierced pearls,

Bialowarczuk 2012: 60. Reiche 2011: 79; 2012: 71. Reiche 2011; 2012. Woolley 1955: 8. Woolley 1955: 13, pl. 15. Woolley 1955: 20. Breniquet 1996. Breniquet 1996: 152, pl. I.5. Breniquet 1996: 154, pl. V.2. Becker 1993: 12, pl. 7. Thompson 1920: pl. IX. Hall and Woolley 1927: pl. XIII.6–7. Stronach 1961: pl. XLIII.10–12. Stronach 1961: 107. Stronach 1961: 106. Stronach 1961: 107, pl. XLIII.13–14.

74 75 76 77 78

330

Stronach 1961: 105. Safar et al. 1981: 123., see Thompson 1920: pl. IX. Oates J. 1969: 130; Oates D. and Oates J. 1976: 68. Tobler 1950: 192, pl. XC.a.1. Fujii 1981: 171.

Philipp Drechsler: Personal Adornment stone balls, laurel leaf-shaped pendants made from mother-of-pearl and elongated tubular beads with angled double distal perforations were not found at Dosariyah may be explained by the fact that these objects were (predominantly) reserved for funerary contexts.

In general, the whole assemblage of personal adornment from Dosariyah shows close similarities to finds from other Neolithic sites in eastern and southeast Arabia. This is particularly the case for snail and disc beads and the rectangular shell plaques, which are not only found in the central Gulf area but also in southeast Arabia, and suggest some kind of common cultural identity and far-reaching — direct or indirect — social contacts between these regions. In contrast, only a few finds point to connections with southern Mesopotamia. These are first and foremost a single ‘clay plug’ and a small T-shaped obsidian fragment, both made from raw materials which were not locally processed. As a consequence, these objects should be considered to be imported exotic items and do not indicate an occupation by southern Mesopotamian population at Dosariyah.

7 Conclusions The total number of objects at Dosariyah which can be assigned to personal adornment is remarkably low (N = 82). Nevertheless, this matches observations from contemporaneous occupational contexts in the Gulf where personal adornment was generally less frequently documented in comparison to funerary sites. The range of items from Dosariyah which were potentially used for personal ornament is dominated by different forms of bead, whereas other kinds of ornaments only occur as individual pieces. Bead production at Dosariyah cannot be excluded owing to finds of an incompletely pierced disc bead as well as several disc bead blanks made from Spondylus shell, but no distinct workshop areas were found. The discovery of ten natural pearls in trench S2 leaves open the possibility that pearls were deliberately collected at the site: while none had any traces of working, pierced pearls have been frequently found in funerary contexts in southeast Arabia during the Middle Neolithic, as well as in the upper Gulf area.

Acknowledgement I would like to thank the colleagues who helped with the taxonomic identification of mollusk shell beads: Sirwan Ali (University of Tübingen) who made a preliminary examination in the field, which was later refined and corrected by Paolo G. Albano (Department of Paleontology, University of Vienna) and Marco Passamonti (University of Bologna).

331

Chapter 16 Bitumen Objects Philipp Drechsler 1 Introduction

Iran around the Deh Luran and Susiana plains were major suppliers for the ancient settlements.11

A total of 321 pieces of bitumen have been documented from the investigations at Dosariyah. Of these, 244 were discovered and individually measured during excavation and the remaining 66 pieces were collected from overburden. In addition, six sherds of pottery, four stones and a Pinctada shell show partial covering with bitumen. The majority of bitumen pieces from Dosariyah are shapeless lumps, but 60 pieces (18.8%) show distinct forms or impressions which allow morphological classification and these shaped pieces are described and grouped according to macroscopically observable characteristics. A functional interpretation of these pieces raises the question whether bitumen — a nonlocal raw material — was regularly in use at the site or represents some kind of exotic item.

2 Description of Shaped Bitumen Objects The description of shaped bitumen objects solely reflects the overall morphology of the objects, while differences in the temper — mineral or vegetal matter added to the bitumen to form a usable mixture — were not taken into consideration. These components can account for up to 80% of the total volume of the bitumen mixtures.12 In general, the temper is highly variable in quality and quantity. It is also difficult to describe any possible vegetal matter more comprehensively as this component may have changed through weathering.13 On the other hand, the internal structure of bitumen objects, for example the layered occurrence of higher quantities of vegetal matter, can provide insights into technological aspects of bitumen coatings.14 In contrast, studies of the mineral component of bitumen objects suggest that no specific materials were used or were necessary for the creation of bituminous mixtures, and ordinary sand was simply added to the heated bitumen until the desired viscosity was reached.15

According to the geochemical studies, the bitumen found at Dosariyah predominantly derives from sources located in northern Iraq, while the origin of several samples is suspected to be found at the Burgan field in Kuwait (see Chapter 17).1 This stands in sharp contrast to the origin of bitumen at the approximately contemporaneous site of H3/As-Sabiyah, which used bitumen exclusively from the Burgan field,2 potentially indicating different (re)distribution networks in the upper and central Gulf.

Among the 60 pieces of shaped bitumen objects from Dosariyah, four morphological groups can be distinguished: cylindrical objects with one end topped with a disc-shaped, egg-shaped or conical/pyramidal piece — most likely plugs or stoppers (N  =  18), small cup-shaped objects (N  =  13), bitumen pieces with impressions of ropes and/or mats (N = 20) and shaped pieces lacking a specific morphology (N = 9).

Finds of bitumen are repeatedly reported from Neolithic (H3/As-Sabiyah,3 Ain as-Sayh site C;4 UAQ25) and Bronze Age contexts (Al-Khidr/Failaka;6 Saar;7 Qala’at al-Bahrain;8 Umm an-Nar;9 Ra’s al-Junaz, RJ210) in eastern and southeast Arabia. These finds fall into the categories of either waterproofing (coating of vessels, architecture and boats) or sealing (stoppers/ plugs). The use of bitumen was far more common in Mesopotamia, where the bitumen seepages in the Hit and the Mosul area in Iraq and outflows in southwest

2.1 Stoppers and plugs A total of 15 pieces of shaped bitumen are characterized by a conical or tapered shaft with a diameter between 6.0 and 13.9 mm (median value 9.1 mm). Attached to this shaft is a thickened area: the shape of this varies and may be cylindrical (N = 1), hemispherical (N = 6), ogive/ egg-shaped (N = 2) or conical (N = 6) (Figure 16.1). On the basis of this thickening, a circular imprint alongside the shaft can often be found. The overall morphology

Van de Velde et al. 2015. Carter 2010d: 101; Connan 2010: 287. Carter 2010d. 4 McClure and Al-Shaikh 1993. 5 Phillips 2002. 6 Barta et al. 2008. 7 Moon 2005. 8 Højlund and Andersen 1994. 9 Frifelt 1995. 10 Cleuziou and Tosi 1994; 2000. 1 2 3

11 12 13 14 15

332

Van de Velde 2015: 12. See Van de Velde 2015: 15. Connan 2010: 269. See Carter 2010d: 99. Connan and Van de Velde 2010.

Philipp Drechsler: Bitumen Objects

Figure 16.1. Stoppers from Dosariyah; note the great morphological variability of the thickenings on top of the shaft (a DOS2012-26814; b DOS2012-27081; c DOS2012-26452; d DOS2012-28471; e DOS2012-26361; f DOS2012-31260; g DOS2010-2164) (drawings: C. Kainert).

of these objects suggests that they represent lumps of bitumen which were attached on top of some type of thin-walled hollow cylinder. The conical or tapered shaft once extended into the hole itself, while part of the thickening was smeared over the outer surface of the cylinder. The circular imprint thus represents the cross section of the cylinder itself. According to a total

of 13 measurements, the wall thickness of this hollow cylinder ranges between 1.7 to 5.3  mm (median value of 2.4 mm). Three additional bitumen objects probably represent fragments of the same type of artifact but missing the shaft, and in each case only part of the thickening is preserved — these are cylindrical (N = 1), hemispherical (N = 1) or ogival (N = 1). 333

Dosariyah This group of finds closely resembles bitumen objects described as stoppers or plugs which occur at Middle Neolithic sites such as H3/As-Sabiyah in Kuwait16 and Bronze Age sites in Bahrain (Qala’at al-Bahrain17, Saar,18) and the lower Gulf (Umm an-Nar19). In contrast to the pieces from Dosariyah and H3/As-Sabiyah, the Bronze Age plugs show a much broader conical shaft and a flat mushroom-shaped thickening (Qala’at al-Bahrain, Saar) or a less distinct transition between shaft and thickening (Umm an-Nar). They are interpreted as lids and stoppers for pottery vessels,20 although none has been found in situ in the mouths of jars.21

thereof (N  =  10). The three completely preserved examples have a simple rim and shallow cavity; their outer rim diameter ranges from 38 to 41  mm, while the inner diameter of the cups was found between 28.5 and 38 mm. Both the interior and exterior is carefully smoothed. In each case, the base is comparatively thick and has a prominent hollow with a diameter of between 9.5 and 11  mm but which does not cut through the whole wall of the object (Figure 16.2). The function of these small cup-like objects remains unclear. Given their low capacity well below 20 cm3, they are hardly usable as drinking cups. More likely, they were containers used to display valuable liquids or powdered solid substances.

At the Early Dilmun settlement at Saar, three different types were distinguished. The first were described as ‘stoppers’ designed to be pushed into a narrow neck, and are simple cylinders with a slight taper but without a thickening.22 The second resemble ‘lids’ and are more elaborate, consisting of short cylinders which directly fitted within the jar neck with a wide flange extending over the top.23 Both of these types are considerably larger than the stoppers from Dosariyah, although the ‘lids’ are said to vary. The third are mushroom-shaped pieces of bitumen which were employed as bungs for broken pottery vessels;24 smaller examples from Saar could have been either repair bungs or small lids.25

Comparable pieces are unknown from Neolithic or Bronze Age contexts in eastern Arabia but if they are considered to be imports, their frequency at Dosariyah suggests that such objects must have been common in areas that were part of a wider exchange network. Alternatively, they may have been made in Dosariyah itself. No evidence was found of large-scale processing of bitumen at Dosariyah,26 although bitumen mixtures can easily be produced and/or recycled in simple fire pits or pottery vessels.27 The discovery of pieces of pottery with a bitumen coating on the inside may be considered as evidence of bitumen processing (Figure 16.3). Likewise, the presence of bitumen fragments with impressions of reed and cords — interpreted as the potential remains of architecture or boat building — represent at least indirect evidence of the reshaping of bitumen at the site, which would provide the opportunity for the production of such small objects.

The latter broadly resemble the shape of the objects from Dosariyah but cannot have been intended as repair bungs for pots as the bases are not flat (as they should be if applied to the inner or outer surface of a vessel), but cylindrical. Moreover, the thickening is more pronounced, voluminous and variable in shape while the cylindrical portion is generally longer and often thinner. It is more plausible instead to interpret these objects as stoppers or plugs. As pottery vessels with correspondingly narrow necks were not found at Dosariyah, the stoppers might have been used to seal some other kind of small container or flask made of a perishable organic material, such as a hollow wooden tube or reed stem. Considering the narrow diameter of the conical shafts, one has to assume that the container was rather small and contained small amounts of liquid or a solid, granular material.

2.3 Bitumen pieces with impressions The largest group of shaped bitumen objects from Dosariyah are pieces of bitumen with impressions of reed28 and/or rope (N  =  20). The majority of these are flat (N  =  15), with one smoothed face and the impressions on the reverse (Figure 16.4). In one case, badly preserved impressions of reed or rope were found on one lateral edge but not touching the flat face (DOS2012-25719.2). The remaining four pieces were irregular with impressions of thin rope (N  =  2), approximately prismatic with impressions of rope and reed on one side, while the opposite side shows impressions of reed only (N = 1), and triangular in cross section (N = 1). The latter piece is important as it closely

2.2 Small cup-shaped objects A total of 13 shaped bitumen objects are almost completely preserved small cups (N = 3) or fragments 16 17 18 19 20 21 22 23 24 25

Carter 2010d: 100, fig. 5.9a. Højlund and Andersen 1994: 408, fig. 2046. Moon 2005: 193. Frifelt 1995: 120, fig. 344. Moon 2005: 193; Højlund and Andersen 1994: 408. Moon 2005:193. Moon 2005: fig. 5.15c. Moon 2005: figs 5.15a, 5.14j. Moon 2005: fig. 5.14k–l. Moon 2005: fig. 5.14i.

Bitumen processing facilities are only required if larger amounts of bitumen are needed. An example is an Early Dynastic bitumen oven from Nippur; McCown et al. 1978: 16. 27 Van de Velde 2015: 16. 28 The term reed is used here as a general name for tall grass-like plants of wetlands which include, among others, the Poaceae family, but does not refer specifically to the common reed (Phragmites australis) which formed the major species in the marshes of southern Mesopotamia. 26

334

Philipp Drechsler: Bitumen Objects

Figure 16.2. Cup-like bitumen objects (a DOS2012-28700; b DOS2012-31349.58; c DOS2012-31349.54; d DOS2012-31106) (drawings: C. Kainert).

resembles the pieces of reed-impressed plaster found at Dosariyah: one face is carefully smoothed, while the opposite two sides show impressions of reed and rope; the reed stems run parallel to the ridge of the triangle

whereas the cord impression runs perpendicular to it (DOS2012-30275). In addition to the shapeless lumps, pieces of bitumen with imprints of reed or rope are among the most 335

Dosariyah

Figure 16.3. Ubaid ware potsherd with a partial bitumen coating on the interior surface (DOS2012-29164) (photograph: B. Kiepenheuer-Drechsler).

frequent finds of bitumen in eastern and southeast Arabia in Neolithic and Bronze Age contexts. Finds of reed- and rope-impressed bitumen fragments at the Middle Neolithic site of H3/As-Sabiyah have been interpreted as the remains of boat building or repair (i.e. bitumen stripped from vessels).29 This argument is supported by the identification of barnacles being attached exclusively to the outer/flat surface of the bitumen pieces, indicating a prolonged period of submersion.30

as the remains of a roof coating.34 Similarly, a piece of bitumen with a triangular cross section from Qala’at al-Bahrain lacks the characteristic growth of barnacles and the idea that this piece derives from a seagoing vessel is questionable.35 In addition to the pieces with impressions of reed and ropes, pieces of bitumen with impressions of woven palm leaves repeatedly occur at Bronze Age sites.36 While one side shows the woven leaf imprints, the other side is smoothed. According to the excavators, these bitumen fragments represent remains of bitumen-coated mats or vessels made from woven date-palm leaves.37

Similar pieces of bitumen with impressions of reeds and ropes have also been reported from Bronze Age sites in the Gulf. Bitumen was extensively used at the third-millennium BC settlement at Umm an-Nar where several pieces show impressions of reeds and ropes and possible uses either in house construction or ship building were proposed.31 Over 300 bitumen pieces were also found at the third-millennium BC coastal settlement of RJ-2 (Ras al-Jinz) in Oman. Most of these bore impressions of reed, rope or woven mats on one side, with barnacles often found attached on the other, and were interpreted as the remains of coatings of boats constructed from reed bundles and wood.32 Impressions of reed mats were most frequent among the pieces with impressions at RJ-2, suggesting that most of the outside of the hull was once covered with matting which was fixed with ropes before the application of the bitumen coating.33 In contrast, four bitumen fragments from the Early Dilmun settlement at Saar show impressions of parallel plant stalks interpreted 29 30 31 32 33

The function or origin of the reed- and rope-impressed bitumen fragments from Dosariyah cannot be positively identified. The absence of barnacles on the well-smoothed exterior, non-impressed surface does not per se exclude its use for the coating of boats, but final proof for this interpretation is lacking. Alternative explanations are that specific parts of architecture/ installations at the site were once (partly) covered with bitumen or that the fragments of impressed bitumen were stored raw material obtained as recycled material, similar to the remaining 260 pieces of unshaped bitumen. 2.4 Miscellaneous objects Nine bitumen objects show clear traces of shaping although their original morphology remains obscure due to their fragmentary state. Among these, four

Carter 2010d: 101. Carter 2010d: 98. Frifelt 1995: 226. Cleuziou and Tosi 2000: 63; Vosmer 2000: 235. Vosmer 2003a: 52; 2003b: 155.

Moon 2005: 198. Højlund and Andersen 1994: 409. 36 Højlund and Andersen 1994: 409; Frifelt 1995: fig. 341; Moon 2005: fig. 5.15e. 37 Højlund and Andersen 1994: 409; Frifelt 1995: 226; Moon 2005: 196. 34 35

336

Philipp Drechsler: Bitumen Objects

Figure 16.4. Bitumen pieces with imprints of reed and ropes (a DOS2012-31262.14; b DOS2012-26771; c DOS2012-26450; d DOS2012-30860; e DOS2012-26463; f DOS2012-30275; g DOS2012-31332.66; h DOS2012-25164) (drawings: C. Kainert).

337

Dosariyah pieces are essentially flat and in two instances show a straight rim (i.e. parts of an original edge are preserved), suggesting that these fragments were parts of flat tablets. Other bulkier pieces show convex exterior faces and should be considered as the remains of larger objects.

weakly developed. Noticeable is the distribution of stoppers that were not found in the southern part of the site, which potentially is the result of a functional differentiation (see differences in the occurrences of stone and bone tools: Chapters 11, 14). Most conspicuous, however, is the exclusive occurrence of bitumen pieces with imprints of reed and rope that were found only in trench E1. The fact that these pieces occur within a restricted part of the stratigraphic sequence only (Figure 16.6a) suggests either that they were collected and subsequently stored within one event and later redistributed by trampling, or that they derive from a specific, short-lived architectural feature whose traces could not be detected during excavation. The spatially restricted occurrence of shaped bitumen objects in comparison to the much wider distribution of bitumen remains (Figure 16.6b) is difficult to explain functionally, but can be potentially related to taphonomic processes: most bitumen lumps are small and might represent highly eroded fragments of larger pieces that once might have represented shaped bitumen objects or larger pieces with imprints. This suggestion finds its preliminary confirmation in the occurrence of ‘dark brownish colored’ sediments that show a higher density of bitumen finds. Most likely, at least part of the sediment color is the result of finely dispersed bitumen that has been degraded by wind action (Figure 16.7).

3 Spatial Distribution of Bitumen Objects With the exception of trenches N2 and E2, pieces of bitumen were found in all excavated areas (Table 16.1). The highest number of bitumen pieces was documented in trench E1, although sediments from this trench were only partially screened. Finds of bitumen in the southern part of the site were less frequent, while they were almost absent in the central and northern part. The spatial distribution of shaped bitumen objects closely follows this overall frequency. They occur in the highest numbers in trench E1 and trench S2. If considering the individual morphological groups, some spatial patterns are noticeable (Figure 16.5): fragments of cup-like bitumen objects occur both in the eastern and southern parts of the site, but they are completely absent in the central northern trenches. The reverse is true for the spatial distribution of stoppers. They do not occur in the south but are present in the northern and — in high numbers — the eastern part of the site. Finally, bitumen pieces with imprints of reed and/or rope are exclusively restricted to trench E1.

4 Significance of the Bitumen Objects at Dosariyah

In all trenches, pieces of bitumen are predominantly found spatially associated with occupation layers, but almost never occur together with dense accumulations of Pinctada shell interpreted as task-specific dumps. This general conclusion explains the sparse occurrence of bitumen in the central northern part of the site where distinct occupation horizons are absent or only

Even if most pieces of bitumen found at Dosariyah represent shapeless lumps, the shaped bitumen objects can be well differentiated according to their morphological characteristics. The repeated occurrence of morphologically similar objects also points to some degree of functional standardization, namely, the use of bitumen for a specific set of tasks:

Table 16.1. Number of shaped bitumen objects and bitumen lumps excavated at Dosariyah. Trench

Stoppers/ potential Cup-like Imprints stoppers objects

Other

Bitumen-covered stone or pottery

Bitumen lumps

Total1

N1

2

-

-

-

-

4

6

N3

-

-

-

-

1

6

7

N2 E1

E2 S1

S2

S3

-

15 -

-

-

-

-

1-

1

1 -

-

2-

-

-

-

-

7 -

-

2 -

-

1-

-

-

-

136 -

3

28 4

-

198 -

4

31 4

1 The sum of total count (N = 250) deviates from the total number of bitumen pieces given at the beginning of this chapter as finds of bitumen from the surface of the site and from overburden are not considered here.

338

Philipp Drechsler: Bitumen Objects

Figure 16.5. Spatial distribution of shaped bitumen objects and bitumen-covered pottery and stone: a stoppers (dots) and potential stoppers (open circles); b cup-like bitumen objects; c bitumen pieces with imprints of reed or ropes; d pottery (dot), stone (square) and shell (triangle) with a partial cover with bitumen.

stoppers were probably used for the sealing of small vessels made from perishable organic material. Small cup-like objects were presumably fixed on top of poles and can be associated with the display of liquids or powders. Alternatively, they may have been mounted upside down on top of sticks or poles. In the marshes of southern Iraq between 1968 and 1990, Ochsenschlager observed the application of bitumen to form mace heads on long sticks used to keep dogs away or to seal boat poles.38 38

The highest degrees of morphological variability show pieces of bitumen with impressions of reed and ropes. While it cannot be excluded that these pieces represent bitumen coatings stripped off boats, it is also possible that they derive from ephemeral architecture built from reeds stitched together with ropes and partly coated with bitumen. Although the morphological homogeneity of bitumen objects legitimates the establishment of distinct functional groups, it does not specify the actual

Ochsenschlager 2004: 185–86.

339

Dosariyah

Figure 16.6. Stratigraphic distribution of bitumen in trench E1. Note the generally wider distribution of bitumen lumps (b) in comparison to the occurrence of shaped bitumen objects (a) (circle: stopper; open circle: potential stopper; semi-circle: small cup/cup fragment; cross: reed/rope impressed bitumen; triangle: bitumen lump). Trench E1, north section.

Figure 16.7. Highly degraded cup-like bitumen object (DOS2010-3504) during excavation (a) and after recovery and cleaning (b). Without better-preserved finds of such objects, the original shape could not have been determined (photographs: P. Drechsler, B. Kiepenheuer-Drechsler)

340

Philipp Drechsler: Bitumen Objects function of the objects for the inhabitants of Dosariyah. This seems especially true for the small cup-like objects: it can well be imagined that they were purposefully manufactured at Dosariyah for a specific use, but they may also represent (parts of) imported objects — deriving from southern Mesopotamia — considered as exotic items without a functional meaning.

recycling. But if bitumen was considered a (re)usable material in the local Arabian Middle Neolithic cultural context, it remains unclear why both bitumen lumps and shaped bitumen objects were found in such high numbers. If the processing of bitumen took place at Dosariyah then it would not have been discarded but carefully stored for future use: bitumen is one of the few substances that can be almost indefinitely recycled by reheating until its major characteristics — plasticity and stickiness when heated — disappear.

Such a scenario might also be true for stoppers, but there is another argument: these pieces were possibly used for sealing small vessels made from perishable materials. Considering the origin of the bitumen in Mesopotamia, it seems at least probable that the associated vessels as well as their content also derived from Mesopotamia. Once in Dosariyah, the stoppers were removed from the vessels and either stored (as exotic items, again without functional meaning or as usable raw material) or discarded (if considered useless by the inhabitants of the site).

5 Conclusions Bitumen objects are a fascinating aspect of the material culture at Dosariyah. Its wide stratigraphic distribution suggests a regular supply of bitumen from the beginning of the settlement at Dosariyah and throughout the period of occupation. The number and variety of shaped objects is limited and shapeless lumps generally predominate. The predominance of shapeless pieces is most likely the result of taphonomic processes. There is strong evidence for heavy wind action leading to a mobilization of sand and, as a consequence, the rapid disintegration of bitumen objects.

Although not final proof, finds of bitumen pieces with impressions of reed and ropes are one argument for a purposeful reshaping and recycling of bitumen at the site, as these objects seem to be too unattractive to represent items beyond a functional meaning. They therefore derive either from mobile objects (i.e. bitumen-covered containers and chests, and also boats) or fixed installations.

The range of shaped bitumen objects found at Dosariyah suggests a highly specialized application of this raw material which clearly refers to its sealing capacities: it was used for the sealing of vessels, the production of liquid-proof small cups and the coating of reeds stitched together with ropes. The value afforded to bitumen and bitumen objects at Dosariyah, however, remains ambiguous. Considering the long distance from the original sources, many pieces of both shaped and unshaped bitumen were discarded, signaling that bitumen as a raw material was potentially of little use to the inhabitants of the site.

The frequency of bitumen at the site stands in clear contrast to the fact that bitumen does not occur locally, but was brought from sources located in Mesopotamia. In accordance with imported Ubaid pottery, finds of bitumen extend across the whole stratigraphic sequence of the site, suggesting a regular and stable supply. Finds of potsherds and stones that show a (partial) bitumen cover are probably arguments for a local reshaping/

341

Chapter 17 Chemical Composition of Bitumen Thomas Van de Velde 1 Introduction

that it had been used for the caulking of boats.8 Twelve samples from H3/As-Sabiyah were analyzed and the origin shown to be the Burgan Hill in Kuwait.9

Bitumen is a natural resource formed by the thermal degradation of organic content within sediment. It is the heaviest fraction of crude oil but unlike oil it does not migrate through sediments.1 The adhesive and waterproofing properties of bitumen enabled this material to be used for multiple applications. Among its earliest uses was the adhesion of flint fragments to handles, creating composite tools as early as the Middle Paleolithic period.2 Other popular uses of the material were as a mortar or water-proofing agent in architecture, either directly applied on brick walls or as bitumen-coated matting, which has been attested for example from the fifth millennium BC and later periods at Tell al-‘Ubaid (southern Iraq),3 Chagha Sefid (southwest Iran),4 Saar (Bahrain)5 and Umm anNar island (lower Gulf).6 One of the most important traditional uses of the material, however, was in naval architecture as boats were caulked with bitumen to waterproof and protect the hulls and this practice was common up until the twentieth century in central and southern Iraq.7

2 The Dosariyah Bitumen Samples A total of 321 bitumen objects and lumps were recovered from six of the eight trenches excavated by the Dosariyah Archeological Research Project (DARP) (Table 17.1). Twenty samples were selected for further research, all being shapeless lumps as these analyses were destructive. Due to the viscosity of natural bitumen, a mixture needs to be made in order to be able to work properly with the material. This was commonly done by melting the bitumen and adding temper. Generally, these mixtures were consistent throughout antiquity and consisted of ca. 20–30% natural bitumen with the addition of mainly mineral matter, although remains of vegetal matter are also often observed.10 This additional material in the mixtures was not studied in detail for the Dosariyah samples and only macroscopically observable specifics were noted. At this level the Dosariyah bitumen mixtures appear very similar to many other archeological bitumen samples with most samples having the typical brownish-black color and usually tempered with sand, but tiny pebbles were noted in many samples as were crushed shells. It is generally believed that the composition of the mineral matter is of no great importance when making these mixtures and that ordinary sand was added to the heated bitumen until the desired viscosity was reached.11 Given the coastal nature of Dosariyah, these observations should come as no surprise.

In the Middle East, an area famous for its oil fields and deposits, bitumen surfaces at several locations (Figure 17.1). In this chapter we will present the study of several bitumen pieces which aimed to determine the geological origin of the samples, namely to identify the seepage where the material came from. Generally, bitumen identified at Mesopotamian and Arabian Gulf sites comes from one of three major extraction zones: the area around Hit in central Iraq, the zone near the upper Tigris river in northern Iraq and the Deh Luran and Susiana plains of southwest Iran. Dosariyah is the second archeological site in the Gulf where bitumen has been found in levels dating to the Arabian Neolithic, the first being H3/As-Sabiyah. At the latter site, most of the bitumen had reed impressions with several showing traces of barnacles, indicating

1 2 3 4 5 6 7

Most of the studied samples (N  =  16) were excavated from trench E1, an excavation unit with a complex stratigraphy containing several features including fireplaces, pits, post-holes and an unidentified panshaped structure (Figure 17.2) (see Chapter 5). One sample from this trench bears crushed shell inclusions: such inclusions have been attested previously from

Peters et al. 2005: 360. Boeda et al. 1996; 2008. Forbes 1964: 71. Hole 1977: 225. Moon 2005: 198. Frifelt 1995: 226. Connan et al. 2005: 22; Thesiger 1964; Ochsenschlager 1992.

Carter 2010d. Connan and Carter 2007; Connan 2010. 10 Forbes 1964: 56; Connan and Deschesne 1996: 117; Connan et al. 2005: 43; Connan 2012: 141. 11 Connan and Van de Velde 2010. 8 9

342

Thomas Van de Velde: Chemical Composition of Bitumen

Figure 17.1. Major oil seepages in the Near East that are presumed to have supplied bitumen in antiquity (after Connan 2012).

studies. Stable Carbon Isotope Analysis measures the relation between the isotopes 12C and 13C and is commonly expressed as δ13C whereas GC-MS is used to separate and quantify different molecular compounds. These compounds are generally referred to as biomarkers and may contain source-specific compounds. The outcome of such separation of a sample is a chromatogram in which a specific fingerprint (m/z 191) is used for compound determination and quantification (Figure 17.5). If source-specific molecules are absent, diagnostic molecular ratios are used to determine the origin of the samples. Just as δ13C values (expressed in ‰), these ratios by themselves reveal nothing and should be compared with comparative data sets and in general, matching with other archeological samples works better than comparing data from current seepages as molecular changes may occur when the bitumen was worked.12 According to the primary findings and observations, reference data sets were chosen to

archeological bitumen samples although only very sporadically. Marine sediments often contain similar shell fragments and provide a hint of the source of the temper in this case; unfortunately, the shell fragments are too tiny for any identification of species. A total of 31 bitumen objects were found in trench S2 of which four were used for geochemical research; four bitumen artefacts were recovered in trench S1, of which one was selected for analysis (Figures 17.3–17.4). 3 Analytical Techniques Two techniques were used to characterize the bitumen samples: Gas Chromatography – Mass Spectrometry (GC-MS) and Stable Isotope Analysis respectively for the Saturated Hydrocarbon and Asphaltenes fractions. The first technique was developed specifically for identification and characterization of oil spills by the petroleum industry/chemistry, while the second technique — based on the natural isotope 13C — is primarily used for environmental and climatological

12

343

Connan and Carter 2007: 65.

Dosariyah

Figure 17.2. Stratigraphic distribution of bitumen remains found in trench E1. Samples used for fingerprinting are marked with white triangles. The two lowest white triangles are samples DOS2012-29197 and DOS2012-29239. Trench E1, north section.

Figure 17.3. Stratigraphic distribution of bitumen found in trench S2. White triangles mark finds used for analysis. Trench S2, east section.

further identify the nature of the Dosariyah bitumen samples. For this research, reference samples from the archeological sites of H3/As-Sabiyah, Kosak Shamali, Tell el-‘Oueili and Ra’s al Jinz were used (Table 17.2).13 These references were chosen because not only do they cover all of the known major bitumen extraction sites from which the Dosariyah samples may have come, but also because several samples from both Kosak Shamali and Tell el-‘Oueili overlap chronologically with the occupation at Dosariyah during the first half of the fifth millennium BC. In contrast, the bitumens from H3/AsSabiyah should be slightly earlier as they date to the final phase of the sixth millennium BC and those from RJ/Ra’s al-Jinz are younger as they come from a site dating to the second half of the third millennium BC.

Figure 17.4. Stratigraphic distribution of bitumen found in trench S1. White triangles mark sample used for analysis. Trench S1, west section.

4 Methods Prior to analysis, all archeological samples need to undergo extensive sample preparation in order to obtain reliable and qualitative measurements. To Connan and Nishiaki 2003; Connan et al. 2005; Connan 2010; Connan and Oates J., in preparation.

13

344

Thomas Van de Velde: Chemical Composition of Bitumen begin with, the bitumen needs to be separated from the archeological samples. Bitumen found at ancient sites is commonly a mixture of pure bitumen with added mineral (and often also organic) matter. This temper needs to be removed from the sample prior to analysis. About 1.5  g of archeological samples was taken (or less, when the desired quantity was not available) and manually crushed by mortar for solvent extraction using CH2Cl2:Methanol (2:1 ratio). The solution was treated ultrasonically and was centrifuged prior to recuperation of the liquid fraction (holding the bitumen). This part of the sample preparation was executed twice in order to isolate as much of the bitumen as possible, once prior to GC-MS analysis and once for the Carbon Isotope Analysis. In the latter case, an Accelerated Solvent Extractor (Dionex ASE 350) was used to automate the separation of the organic from the mineral matter.

Table 17.1. Bitumen samples used for sourcing. Sample ID

Trench; Stratigraphic Remarks Unit

DOS2010-2613

S1, S1-IV

DOS2011-14668

S2; S2-III

DOS2011-14330 DOS2011-16190 DOS2011-19554 DOS2012-25364 DOS2012-26038 DOS2012-26341 DOS2012-26531 DOS2012-26551 DOS2012-26607 DOS2012-27083 DOS2012-28266 DOS2012-28344 DOS2012-28814 DOS2012-28912 DOS2012-29197 DOS2012-29239 DOS2012-30852 DOS2012-30994

S2; S2-IV

Results not included

S2; S2-IV S2; S2-IV E1; E1-III E1; E1-III E1; E1-III E1; E1-III E1; E1-III E1; E1-III

E1; E1-IV E1; E1-IV E1; E1-IV E1; E1-IV

Results not included

E1; E1-VI

Results not included

E1; E1-VII

Crushed shell inclusions

E1; E1-VII E1; E1-III

For asphaltenes isolation, the obtained solution was dried using a rotovapor and then taken back into a hexane solution. The residue is the asphaltene fraction. This operation was repeated several times to be sure that no co-precipitated molecules were trapped in the asphaltenes matrix.

Results not included

Standard column chromatography was used for the separation of fractions prior to GC-MS analysis, as

E1; E1-III

Table 17.2. Molecular and δ13C values of the used reference samples and sites. Site

Sample ID Dating

δ13C

Ts1/Tm2

GCRN3/C304

Origin

H3/As-Sabiyah

1644

-27.4

0.24

0.13

Burgan Hill (Kuwait)

1645

1646

1647

1648a 1649a

1649b

1649c 1650

1771

1772

Oueili

5300-4900

5300-4900

-27.1

0.50

0.11

5300-4900

-26.9

0.50

0.11

5300-4900

-27.0

0.24

0.18

5300-4900

5300-4900

5300-4900

5300-4900

5300-4900

5300-4900

5300-4900

0.19

-27.1

0.29 0.28

0.29

-27.0

-27.2

-27.1

0.41

Burgan Hill (Kuwait)

Burgan Hill (Kuwait)

0.20

Burgan Hill (Kuwait)

0.17

Burgan Hill (Kuwait)

0.17 0.13

0.13

Burgan Hill (Kuwait)

Burgan Hill (Kuwait) Burgan Hill (Kuwait)

Burgan Hill (Kuwait)

Burgan Hill (Kuwait)

Burgan Hill (Kuwait)

1773

5300-4900

-27.2

0.36

0.14

Burgan Hill (Kuwait)

22

5000-4500

-27.7

0.17

0.21

Northern Iraq

21

23

28

29

32

34

5000-4500

5000-4500

5000-4500

5000-4500

5000-4500

5000-4500

-27.7

0.17

-27.7

0.11

-28.2

0.23

-28.3

0.20

-28.1

0.17

-28.0

0.10

345

0.26

0.20

0.27

0.24

0.26

0.17

Northern Iraq

Northern Iraq

Northern Iraq

Northern Iraq

Northern Iraq

Northern Iraq

Dosariyah

Site

Sample ID Dating 37

40

Ra’s al-Jinz

42

757

759

762

770

Kosak Shamali

771

1375

1376 1377

1378

1379

1380

1381

1383

1384

1385

1386

1395 1 2 3 4

5000-4500

5000-4500

5000-4500 2500-2300

2500-2300

2500-2300

2300-2100

2300-2100 4940-4550

5100-4700

5100-4700

5100-4700

5100-4700

5190-4800

5190-4800

5300-4900

5300-4900

4550-4260

4550-4260

5300-4900

δ13C

Ts1/Tm2

-27.7

0.16

-27.7

0.17

-28.0

0.19

-27.6

0.10

-27.6

0.12

-27.6

0.12

-27.5

0.12

-27.5

0.12

-27.7

0.11

0.10

-28.3

0.09

-28.3

0.16

-28.2

0.15

-28.1

0.12

-28.0

0.13

-27.7

0.13

-27.8

0.12

-28.2

0.25

-28.0

0.12

-27.9

0.13

0.25

0.23

0.23 0.22

0.25

0.22

0.20

0.23 0.24

0.17

0.13

0.53

0.59

0.29

0.33

0.21

0.22

0.63

0.32

0.22

Origin Northern Iraq

Northern Iraq

Northern Iraq Northern Iraq

Northern Iraq

Northern Iraq

Northern Iraq

Northern Iraq Northern Iraq Hit

Hit

Samsat? Northern Iraq?

Samsat? Northern Iraq?

Hit

Hit

Northern Iraq

Northern Iraq

Samsat? Northern Iraq?

Hit

Northern Iraq

Ts = 18α(H)-22,29,30-trisnorneohopane Tm = 17α(H)-22,29,30-trisnorhopane GCRN = Gammacerane C30 = 17α(H), 21β(H)-hopane = C30αβ-Hopane

described by Peters et al.14 A column was packed with silica and pentane: CH₂Cl₂ (3:1) was used as solvent. The eluted fraction, the so-called saturates, is recovered and ready for GC-MS analysis.

For C the international reference is Vienna Pee Dee Belemnite (VPDB), which has a carbon isotopic ratio of 0.0111802 (± 0.0000028). For both techniques several replicate analyses of samples were conducted to cross-check the consistency of the measurements. No discrepancies were identified. If a sample was measured more than once, the average was calculated and used for further analyses and interpretation.

For the GC-MS analysis, a Hewlett-Packard 6890-5973 GC-MS system equipped with an Agilent Technologies HP-5MS column (30  m  x  0.25  mm ID, 0.25  µm) was used. Helium was used as a carrier gas, with a gas flow of 1.5  ml/min. One microlitre of the saturated hydrocarbon fraction dissolved in CH₂Cl₂ was injected (in splitless mode). The oven temperature increased step-wise from 40 to 250°C at 6°C per minute, and from 250 to 300°C at 2°C per minute. The temperature was then held at 300°C for 30 minutes.

5 Results As the first step of the study, GC-MS analysis was conducted on the 20 samples from Dosariyah. The results of this research have been published in detail before but are summarized below.15 Out of the entire data set, four samples showed a chromatogram of extreme low quality with lots of background noise, which means unreliable data for further processing and these samples were discarded from the data set. Three samples were selected for Stable Carbon Isotope Analysis but it was unfortunately not possible to do this for all samples. This analysis excluded the seepages in

The carbon isotopic composition of the asphaltene fraction was determined using an elemental analyzer (ANCA-SL, PDZ Europa, UK), coupled to an isotope ratio mass spectrometer (IRMS) (20-20, SerCon, UK). The isotopic composition of natural samples (i.e. not synthetic isotopic enrichment) is reported relative to an international reference, using the so called ‘δ’ scale and is typically expressed in ‰. 14

GCRN3/C304

Peters et al. 2005: 200.

15

346

Van de Velde et al. 2015.

Thomas Van de Velde: Chemical Composition of Bitumen

Figure 17.5. Examples of chromatograms (m/z 191) of samples DOS2012-19554 and DOS2012-29239.

Figure 17.6. Cross-plots of Ts/Tm vs. Gammacerane/C30αβ-Hopane.

the Hit area as a possible supplier, as their δ13C value was significantly lower than −28‰. Two samples (DOS2012-29239; DOS2012-26038) fall within a batch of bitumen samples from Ra’s al Jinz (RJ2) and several samples from both Tell el-‘Oueili and Kosak Shamali (see Table 17.2). The matching samples from the latter two sites have been identified as coming from northern Iraqi seepages.16 The third sample (DOS2012-27083) does not fall within the same range, but is still within 16

the general range of northern Iraqi seepages and the molecular values retrieved from GC-MS analysis also support a northern Iraqi origin (Figure 17.6). As far as molecular ratios are concerned, Ts/Tm vs. Gammacerane/C30αβ-hopane were plotted to further explore the data set (see Figure 17.6). These ratios were also used to conduct Hierarchical clustering (Furthest neighbor, Squared Euclidian) on the data set. The main body of the samples correlates with the archeological samples from which a northern Iraqi

Connan et al. 1996; Connan and Nishiaki 2003.

347

Dosariyah Table 17.3. Molecular and δ13C values of the Dosariyah bitumen samples, including their origin. Sample ID

δ13C

Ts/Tm

GCRN/C30

GCRN/31R1

Origin

DOS2011-14330

0.20

0.19

0.52

Northern Iraq

DOS2011-16190

0.23

0.20

0.57

Northern Iraq

DOS2011-14668

0.27

DOS2011-19554 DOS2012-25364 DOS2012-26038

DOS2012-26341

0.22 -27.64

DOS2012-26531

DOS2012-27083

DOS2012-28814 DOS2012-29197 DOS2012-29239

DOS2012-30852

-27.19

-27.75

DOS2012-30994 1

0.61

0.23

0.54

0.33

0.52

Northern Iraq

0.64 0.60

0.26

0.67

0.19

0.22

0.57

0.43

0.16

0.50

0.19 0.22

Northern Iraq

Northern Iraq

0.36

0.48

Northern Iraq

0.65

0.24

0.25

Northern Iraq

0.22

0.21

0.19

0.55

0.23

0.20 0.20

DOS2012-26551 DOS2012-26607

0.23

0.21

0.20

0.21 0.20

0.18

Northern Iraq Northern Iraq

Northern Iraq

0.54

Northern Iraq

0.45

Burgan Hill?

0.60

0.24

Northern Iraq

0.68

Burgan Hill?

Northern Iraq Northern Iraq

31R = 17α(H), 21β(H), 22(R)-homohopane

was established that 14 of the samples have a northern Iraqi origin; the origin of two samples is harder to confirm but a Burgan Hill origin in Kuwait is possible.

origin was determined, with two exceptions (DOS201229197, DOS2012-29239). The latter show a molecular pattern closer to the bitumen from Burgan Hill and the site of H3/As-Sabiyah in Kuwait, yet DOS201229239 has a measured δ13C value of −27.75‰ which is reminiscent of bitumen from northern Iraq rather than Burgan Hill. The two analytical techniques therefore are in partial contradiction for this sample making it difficult to say anything conclusive. It is remarkable, however, that these two samples are deviant as both were excavated from the oldest deposits in the site and clearly separated from other samples by bitumen-free layers (see the lowest 2 white triangles in Figure 17.2). Samples DOS2012-26531 and DOS2012-26551 show a remarkable higher Gammacerane/C30αβ-Hopane value than any other sample in the data set, but this is most likely to be attributed to differential alteration of the hopane compounds through biodegradation. The Gammacerane to 31R ratio of both samples show no extraordinary value and we are therefore inclined to believe that these samples also have a northern Iraqi origin.

The results of this study differ from those on the bitumen remains from H3/As-Sabiyah. This site is situated in the same cultural complex as Dosariyah — the Arabian Neolithic in the Gulf — but the source of its bitumen was Burgan Hill rather than northern Iraq. This difference is most likely to do with the difference in dating between the sites, as Dosariyah dates to the so-called Ubaid Expansion whereas H3/As-Sabiyah predates this phenomenon.17 The analyses of the bitumen fragments from Dosariyah show that the Ubaid expansion in the Gulf embodies more than just the import of Mesopotamian pottery. Considering the frequency and abundance of the material, it seems unlikely that all of the excavated bitumen was residual or the result of small-scale recycling and it is more likely that it had either been deliberately traded or scraped off boats of Mesopotamian origin and reused locally, although that would have had consequences for the seaworthiness of the vessels concerned, with implications for the crew and their return to the head of the Gulf.

6 Conclusions Despite the fact that four samples had to be removed from the data set due to heavy weathering, it was possible to conduct analysis on a total of 16 bitumen samples from Dosariyah. Each sample underwent GCMS analysis in order to identify and quantify biomarkers and stable isotope analyses were performed on three. It

The sites of H3/As-Sabiyah and Dosariyah are the only Arabian Neolithic sites to offer archeological evidence 17

348

Van de Velde et al. 2015.

Thomas Van de Velde: Chemical Composition of Bitumen of the use of bitumen in the later sixth and fifth millennia BC respectively and the results are important to begin to understand the exploitation and circulation of bitumen in these early periods. Unlike those from H3/As-Sabiyah, which were local, most of the bitumen remains from Dosariyah originate from Iraqi sources, as does the majority of the pottery that was excavated (see Chapters 7–8). The spread of the typical Ubaid Black-on-Buff pottery can be explained by the complex cultural framework of the expansion of the Ubaid cultural and commercial sphere. This expansion not

only introduced pottery to the Arabian Gulf but bitumen as well. It would seem logical that bitumen traveled to and through the Gulf employing the same network(s) as the pottery. During the fourth millennium BC there is virtually no archeological evidence of systematic maritime trade between the populations of southern Iraq and the Gulf, but in later periods there is evidence of a resumption and expansion in trade and contact with specialized long-distance export of bitumen across the Gulf.

349

Chapter 18 Plaster Morphology Philipp Drechsler

1 Introduction

micromorphological analyses were conducted on 11 pieces excavated at Dosariyah in 2010 and 2012 and these now permit a more comprehensive study of the material (see Chapter 19).

A total of 235 pieces of plaster were found during field investigations by the Dosariyah Archeological Research Project (DARP) between 2010 and 2013: 219 pieces come from excavations and the remaining 16 were collected from the surface. Of these, 185 were individually documented during excavation, while 34 pieces derive from collapsed profiles.

2 Plaster Morphology Only general observations about their morphology can be provided as detailed and precise morphometrical studies of the 235 newly discovered plaster pieces were not undertaken. A total of 175 pieces were cursorily described during the field campaigns and the information provided derives from this sub-sample. As a consequence, explanations are limited and should be considered as preliminary.

Pieces of plaster, often with impressions of reeds, were previously recovered as surface finds during the initial visits of the site by Grace Burkholder and Geoffrey Bibby in 1968.1 The pieces immediately attracted interest and were interpreted as the remains of architecture: flat, smooth and sometimes with a thick black iron oxide paint on one side but bearing impressions of reeds on the other, the plaster was, according to Burkholder, ‘used to strengthen reed buildings. The impressions of the reed remain showing how the buildings were formed’.2 Further similarities were noted with pieces of reed-impressed mud and clay found in Ubaid contexts at the sites of Ur and Eridu in southern Mesopotamia.3 At the same time, finds of similar plaster fragments with barnacles attached to the smooth side of the pieces were seen as evidence of changing sea levels since the Ubaid period.4 Later and more extensive investigations at Dosariyah by A. Masry documented ‘[…] large amounts of lime plaster […] scattered all over the site area, particularly near and at the edges’, and were interpreted as evidence of a plastered reed wall enclosing the settlement.5 Information about finds of plaster in Masry’s soundings is lacking in his final publication,6 but field notes from April 2nd and 3rd, 1972 mention finds of plaster in sounding V.7

The majority of plaster pieces exhibit a triangular cross section (N = 131, Figure 18.1c–f) with one flat and often well-smoothed side (N  =  108). In several cases, this side is either slightly concave (N = 9), concave (N = 4), slightly convex (N  =  7) or irregular in shape (N  =  3). The opposite two faces are semi-circular concave and show imprints of twisted ropes and parallel imprints of plant stalks running perpendicular to the ropes (N = 70). The concavity of these two faces suggests that the plaster was applied on a substrate consisting of individual and closely arranged round bundles of plant stalks which had been wrapped with cord. Although a detailed examination of the individual imprints of plant stalks was not carried out, it is plausible to assume that they derive from the widespread brackish water reed Phragmites communis.9 Similarly frequent are pieces with a comparable triangular cross section but only with rope impressions (N = 60). The fact that many of the latter pieces show a rounded or blurred surface suggests that at least in some cases the reed impressions have disappeared as a result of weathering. No attempt was made systematically and reliably to reconstruct the diameter of the wrapped stalk bundles, but an average diameter between 100 and 160  mm is realistic (see Chapter 19).

Investigations on the chemical composition of one plaster piece were carried out by Malinowski and Frifelt to indicate the presence of a lime plaster rather than gypsum plaster.8 New geochemical and Bibby 1970: 376; Burkholder 1972: 266. Burkholder 1972: 266. 3 Burkholder 1972: 266. 4 Bibby 1970: 377; Burkholder 1972: 267; Vita-Finzi and McClure 1991: 198, 199; McClure and Al-Shaikh 1993: 110. 5 Masry 1997: 80. 6 Masry 1974; 1997. 7 Masry 1972. 8 Malinowski and Frifelt 1993. 1 2

In some cases, the outer flat surface of the described pieces show narrow parallel stripes with a width 9 See McClure and Al-Shaikh 1993: 110 for the determination of reed imprints in morphologically identical pieces of plaster from Ain asSayh site A.

350

Philipp Drechsler: Plaster Morphology

Figure 18.1. Pieces of plaster. In general, one surface is flat and well-smoothed while the other shows impressions of reed and ropes. a–b Flat pieces (a DOS2012-31261.11; b DOS2012-26190); c–f pieces with a triangular cross section (c DOS2012-22398; d DOS2012-21881; e DOS2012-22832; f DOS2012-22249) (photographs: B. Kiepenheuer-Drechsler).

351

Dosariyah

Figure 18.2. Pieces of plaster with traces of consecutive coating on the flat surface (a DOS2010-3704; b DOS2012-210943; c DOS2012-22351) (photographs: B. Kiepenheuer-Drechsler).

between 1 and 3 mm which resulted from the flattening of the plaster (N = 7). No finger imprints were observed. Other pieces show traces of spalling along the flat surface, suggesting a coating of the wrapped bundles in up to three consecutive steps or as a single recoating (N = 8, Figure 18.2). In contrast, a repeated recoating of the outer face with thin layers of plaster, characteristic of the renewal of wall surfaces, was not observed. Likewise, the traces of a thick black iron oxide paint as described by Burkholder were not documented10 but in three cases there were traces of a faint dark grayish or blackish residue on the flat surface.

The encrustation of several pieces of plaster with barnacles was previously noted by Burkholder and Masry.11 The new samples provide more detailed insights into this phenomenon (Figure 18.3). A total of 31 (17.7%) pieces of plaster show some kind of growth which, with one exception, is located exclusively on the flat surface: in 23 cases, barnacles were observed, either co-occurring with remains of oysters (N = 4), tube worms (Serpulides) (N = 4), oysters and tube worms (N = 1) or algae incrustations (N = 1). Tube worms alone occur in five cases, while in two cases remains were identified as barnacles or oysters. Whitish residues were observed in one case and interpreted as deriving from the growth of algae. In only one case were incrustations observed on a surface which also exhibited impressions: two barnacles were found attached to an imprint of rope on a plaster piece while the flat surface was also covered with tube worms (DOS2012-27679; see Figure 18.3a).

In addition to pieces of plaster with a triangular cross section, others had a rectangular cross section (N = 12) or were essentially flat (N  =  8; Figure 18.1a-b). As the majority of these pieces are small and show one smooth side while the opposite sides have impressions of reed and/or rope, they most likely originate from the coating of the outer surface of one reed bundle. Finally, 21 plaster pieces show an irregular cross section but still bear imprints of reed and rope. 10

Barnacles and algae can occur both in inter-tidal and sub-tidal marine environments. In contrast, oysters and tube worms require constant coverage with seawater

Burkholder 1972: 266.

11

352

Burkholder 1972; Masry 1974; 1997.

Philipp Drechsler: Plaster Morphology

Figure 18.3. Pieces of plaster showing incrustations with barnacles, oysters and tube worms on the a inner (DOS201227679) and b–d outer, flat surface (b DOS2012-31218; c DOS2010-904; d DOS2012-23409) (photographs: B. Kiepenheuer-Drechsler).

and they therefore prefer a sub-tidal environment. As a result, the co-occurrence of barnacles, algae incrustations, oysters and tube worms on plaster remains from Dosariyah suggests a prolonged and uninterrupted submergence (of the flat surface) of the plaster-covered structures in the sea.

the surface of a dense series of reed bundles stitched together with strands of rope. The outer faces of the coating, while sometimes slightly concave in individual pieces, might have been either flat or even convex. The concavity is most likely the result of an application technique which did not completely fill the interstices between the reed bundles. An intentional coating with black paint could not be verified, nor were any traces of repeated (re)coating to regenerate the outer surfaces

The general morphology of the plaster pieces from Dosariyah suggests that plaster was used to cover 353

Dosariyah

Figure 18.4. Spatial distribution of plaster in trenches excavated at Dosariyah.

observed. Most puzzling are the incrustations with barnacles, oysters and tube worms. These organisms have been observed on fragments from multiple sites and have repeatedly been interpreted as evidence of an inundation of Ubaid coastal settlements by rising sea levels.12 This model does not apply to Dosariyah, however, owing to the stratigraphic situation of the plaster pieces which were found up to 6.3 m above the present sea level and therefore too high to be covered by the sea (see below and Chapter 2). Moreover, the sedimentological record excludes the possibility of the site having been flooded and the reed bundles covered with plaster must therefore once have been in contact with seawater beyond the locations where they were found at the site.

were generally scarce on the surface with no evidence of higher densities along the edges. While pieces of plaster were found in lower numbers in the southern trenches S2 and S3 as well as in the northwestern trench N2, most pieces were documented in trenches N3 and E1 (Figure 18.4). The stratigraphical distribution of plaster pieces in both trenches does not show any significant pattern. In trench E1 they occur almost ubiquitously in the upper part of the stratigraphic sequence. Only in the lowermost stratigraphic units E1-VI, E1-VII and E1-VIII were pieces of plaster completely absent. Plaster pieces with incrustations were recovered in close proximity to pieces without incrustations (Figure 18.5b). As the site itself was never affected by marine transgression during the Holocene, which did not exceed c.2 m above the present sea level in the Dosariyah area (see Chapter 2), the pieces of plaster must have already been incrusted when they were brought to the site. Likewise, the stratigraphic distribution of freshly preserved, weathered and heavily weathered pieces does not show any patterning, suggesting that all stages of weathering occurred synchronously (Figure 18.5a).

3 Spatial Distribution of the Plaster Pieces According to Masry, pieces of plaster were found scattered all over the site, with higher densities near and at the edges.13 The latter observation could not be confirmed during recent fieldwork: finds of plaster 12 Bibby 1970: 377; Burkholder 1972: 267; McClure and Al-Shaikh 1993: 110. 13 Masry 1997: 80.

354

Philipp Drechsler: Plaster Morphology

Figure 18.5. Stratigraphic distribution of plaster in trench E1. a Preservation conditions; green: fresh, yellow: slightly weathered, red: heavily weathered; b incrustations; green: no incrustations, red: incrustations. Trench E1, north section.

excavation clearly indicates that all the pieces of plaster derive from the covering of structures or objects which consisted of individual, closely arranged round bundles of plant stalks with an assumed diameter between 100 and 160 mm. The bundles most likely consisted of the brackish water reed Phragmites communis, wrapped with twisted cord.

In trench N3, the majority of plaster pieces occur in stratigraphic unit N3-II, resulting from the disposal of waste and characterized by shell accumulations and articulated fish bones. Only very few plaster pieces were found in stratigraphic units N3-III, N3VI and N3-V which were predominantly accumulated through natural agents (see Chapter 5). Conspicuous is the absence of plaster remains in the lower part of the stratigraphy including stratigraphic unit N3-VI, interpreted as the remains of an occupation horizon. Consequently, the presence of reed bundles covered with plaster is temporally and spatially restricted and not directly related to distinct occupation horizons. As in trench E1, pieces of plaster at different stages of weathering and both pieces with and without incrustations occur within the same stratigraphic units in trench N-3 (Figure 18.6a–b).

The morphological parallels of these pieces of reedimpressed plaster with similar pieces of mud with reed impressions from the Ubaid settlement sites in southern Mesopotamia14 led Burkholder to the conclusion that the pieces from Dosariyah were used to strengthen reed buildings,15 but the spatial distribution of the pieces of plaster documented at Dosariyah seems to contradict this interpretation. Plaster pieces occur without significant spatial patterns, predominantly in the upper stratigraphic units of trenches N3 and E1. They are also present in much lower numbers in trenches S2 and S3. Contrary to this field observation,

4 Conclusions Pieces of plaster with impressions of reed and ropes are a common yet puzzling element of material culture at Dosariyah. The great morphological homogeneity of the pieces documented during surface sampling and

Woolley 1955: 7, fig. 2. Burkholder 1972: 266; see McClure and Al-Shaikh 1993: 110, 112 for a similar interpretation of plaster pieces at the sites Ain as-Sayh A and F.

14 15

355

Dosariyah

Figure 18.6. Stratigraphic distribution plaster trench N3. a Preservation conditions; green: fresh, yellow: slightly weathered, red: heavily weathered; b incrustations; green: no incrustations, red: incrustations. Trench N3, south section.

the presence of distinct, plaster-covered reed buildings should have resulted in the identification of a spatially well-defined occurrence of plaster pieces. Likewise, the occurrence of higher amounts of plaster pieces along the outer edge of the site could not be confirmed.

considering the absence of any traces of representative — or at least substantial — architecture at Dosariyah. The third observation which contradicts such a use of plaster to strengthen reed architecture is the frequency and spatial distribution of pieces of plaster with incrustations of barnacles, oysters and tube worms. To begin with, no other category of material culture excavated at Dosariyah shows any traces of incrustations, with the exception of five sherds of broken Ubaid pottery (trench N3: DOS2012-23784, trench E1: DOS2012-24322; DOS2012-25328; DOS201229439, trench S1: DOS2010-2984). In general, hard ground such as beach rock or debris dumped at the water’s edge can be occupied by barnacles in a very short time and no more than a month.17 The majority of plaster pieces shows incrustations exclusively on the exterior flat surfaces. Such an occurrence can only be explained if these were standing and in contact with seawater, yet neither the stratigraphic positions of pieces of plaster with incrustations (which occur in contexts up to 5.5  m above the present mean sea

While the use of mud for the covering of the walls of reed huts and mud-brick houses in southern Mesopotamia is well justified, a covering of comparable architecture with plaster is not. Mud is an easily available material. If it is found in the right consistency, it can be applied to the walls of houses without pre-preparation.16 In contrast, the preparation of plaster is a multi-stage process which includes the procurement of the raw material (limestone), heat treatment/burning, mixing with additional inorganic components (quartz sand), slaking and finally its application. As yet, however, no features such as kilns have been identified at Dosariyah which could be associated with the processing of lime plaster, which means that the burned lime would have been brought to the site from outside. Such a labor-intensive scenario is even less plausible when 16

Ochsenschlager 2004: 98.

17

356

McClure and Al-Shaikh 1993: 124.

Philipp Drechsler: Plaster Morphology level) nor the characteristics of sediments excavated at Dosariyah indicate a flooding of the site at any point during or after its occupation. Therefore, the only possible — and plausible — explanation is that the plaster came in contact with fresh seawater at a place outside the settlement, and individual pieces were brought to the site for an unknown reason.

that the plaster derives from (parts of) plaster-covered reed structures submerged in the sea that were used for the exploitation of marine resources. They might therefore represent the remains of sections from offshore fish traps (haddrah) that were brought back to the settlement for repairs.18 A general spatial cooccurrence of pieces of plaster and bitumen lumps can similarly be observed at Dosariyah: bitumen lumps that potentially represent bitumen stripped off reed boats with a bitumen caulking.19 The interpretation of pieces of plaster as fragments deriving from sea-going vessels rather than from reed buildings might therefore be another possibility to be further investigated.

Considering all of the available information, the character of the reed-built structures covered by plaster remains speculative. The majority of plaster pieces from Dosariyah were found within stratigraphic units that are dominated by refuse from shellfish gathering and the processing of fish. It is therefore conceivable

18 19

357

Many thanks to St John Simpson for suggesting this idea. See Carter 2010d: 101.

Chapter 19 Morphological and Geochemical Analysis of Plaster Samples Susan M. Mentzer, Markus Seil, Hilmar Adler, Thomas Chassé, Bertrand Ligouis, Christoph Berthold and Christopher E. Miller stratigraphic levels. In addition, the fragments are morphologically similar to pieces of bitumen recovered from the site (see Chapter 17).9 Based on these two observations, it is possible that the material was used in a submerged or semi-submerged marine context. In this study, further geochemical, petrographic and morphological analyses were undertaken in order to address three broad research questions related to the composition and function of the pieces:

1 Introduction This study addresses the composition and function of eleven pieces of a construction material suspected to be lime plaster or mortar recovered during the 2010 and 2012 excavations at the site of Dosariyah. The specimens examined in this study were recovered throughout the archeological sequence of trench E1 and were identified as plaster in the field, based on their white to gray color and high apparent density (Figure 19.1). Previous analyses by Malinowski and Frifelt suggested that similar materials had been prepared from a mixture of slaked lime and diatomite, yielding a hydraulic plaster.1 If the materials are indeed hydraulic plaster, they would be some of the oldest examples in the world. According to Regev et al. hydraulic plaster was deliberately produced at the site of Tell es-Safi/Gath in Israel during the Iron Age2 but this technology did not become prevalent until Roman times. Other researchers have noted that hydraulic plasters may occasionally have been produced accidentally in Neolithic periods.3

•• what is the composition of the plaster fragments and, more specifically, is there any evidence that these plasters contain both fired lime and hydraulic compounds? •• how were the plasters manufactured? •• can the compositional and morphological characteristics of the plaster fragments be used to understand their possible functions? 2 Background 2.1 Archeological context

Plaster samples from the first excavations at Dosariyah were visually identified by Burkholder4 and Masry,5 and studied by Malinowski and Frifelt.6 They concluded that these were remains of hydraulic plaster and proposed, on the basis of preserved impressions of plants, that the material had been used in the construction of a typical traditional Gulf form of palm-branch architecture known as barasti or arish.7 Given the lack of any other elements related to architecture at the site (see Chapters 4–5); however, it is unclear whether the recovered plaster fragments did derive from a structure or from some other type of feature. Burkholder and Masry also reported barnacles on the flat surfaces of some plaster fragments.8

The majority (N = 10) of the plaster samples recovered in 2012 were from trench E1, located along the eastern edge of the site (Figure 19.1a), and came from different depths below the surface (Table 19.1). Other features observed within the excavated area included combustion features, pits and a pan-shaped structure of unknown function (see Chapter 5). One additional sample collected in 2010 from the surface of the site was also analyzed. 2.2 Physical attributes — high-resolution 3D scanning Nine samples were scanned at high resolution in 3D in order to characterize their shapes prior to partial destruction. These scans were conducted in the Paleoanthropology Imaging Laboratory at the University of Tübingen. Each object exhibits a triangular morphology (Figure 19.2) with a flat to slightly concave basal surface, and two concave surfaces with a series of surface impressions. The orientations of the impressions range from random to strongly parallel. The impressions are themselves semicircular in cross section (Figure 19.2e), with internal

During the recent excavations, barnacles and other marine encrustations were discovered adhering to 23 fragments of the material (Figure 19.1d). Encrusted samples were distributed evenly throughout the 1 2 3 4 5 6 7 8

Malinowski and Frifelt 1993. Regev et al. 2010b. Affonso and Pernicka 2001: 12–13. Burkholder 1972. Masry 1974. Malinowski and Frifelt 1993. Malinowski and Frifelt 1993: 5. Burkholder1972: 267; Masry 1974.

9

358

Van de Velde et al. 2015: 250.

Susan M. Mentzer et al.: Morphological and Geochemical Analysis of Plaster Samples

Figure 19.1. Overview of the study. a Stratigraphic section showing the locations of the samples from trench E1; b photograph of a suspected plaster object (DOS2010-904) with visible impressions on two sides; c photograph of the same fragment in cross section, illustrating the typical triangular morphology and approximately flat face; d photograph of the same fragment with visible barnacle bases. The inset shows the morphology of a cluster of intact barnacles on a different fragment (DOS201223173).

longitudinal impressions that over some length have a twisted arrangement.

Table 19.1. List of samples. Sample number

Trench; Stratigraphic z (m a.m.s.l.) Unit

DOS2012-29789

E1; E1-II

5.667

DOS2012-22249

E1; E1-II

5.33

DOS2012-21881 DOS2012-23459 DOS2012-24079 DOS2012-23826 DOS2012-31042 DOS2012-26573 DOS2012-27697 DOS2012-28283 DOS2010-11221

E1; E1-II E1; E1-II E1; E1-II E1; E1-II

E1; profile E1; E1-II

E1; E1-IV E1; E1-IV

Surface find

The impressions in the plaster objects are strongly reminiscent of parallel wrapped bundles containing multiple strands of twisted plant material, such as reeds or cords. The radii or diameter of the wrapped bundles (assuming circular segments) can be calculated using the formula R = (4h² + s²) / 8h where s = section length and h  =  section height. For two artifacts with wellpreserved impressions, the diameters of the bundles are calculated to be 15.36 and 15.02  cm (DOS201228283; DOS2012-22249). This combination of earthen construction material and bundled plant material is typical of an object or architecture composed of an underlying structural lattice or skeleton which has been coated and sealed (e.g. wattle and daub, lath and plaster, or jacal architecture). Although daub is commonly associated with structures, it can be applied to a variety of woven plant skeletons for a variety of purposes. When boats and containers are coated in similar

5.533 5.204 5.066 4.95

4.647 4.568 4.235 4.03

359

Dosariyah

Figure 19.2. Examples of 3D scanning. a DOS2012-22249 scan in cross section; b–d same piece in different orientations; e image of the surface of DOS2012-31042 with two different types of impressions. In green, the long axis of twisted fibers with semicircular shape. In red, the long axis of parallel impressions.

360

Susan M. Mentzer et al.: Morphological and Geochemical Analysis of Plaster Samples materials, especially if they are composed of fired lime and are fine in texture, they are typically referred to as ‘plastered’10 and the material termed ‘non-architectural plaster’.11 Depending on their function, the materials described in this study could be termed either daub or plaster. Regardless of nomenclature, it is clear that these materials were originally associated with organic skeletons that have been lost to taphonomic processes.

a slurry. The mixing stage is termed slaking and initiates an exothermic reaction resulting in the formation of calcium hydroxide (Ca(OH)2). For use in construction purposes, the slaked lime may be mixed with other materials, such as clay and silt to add volume, sand to serve as aggregate or temper. As the water evaporates and the material sets, the Ca(OH)2 reacts with CO2 in the surrounding atmosphere to form solid CaCO3.

2.3 Plaster and daub composition

2.5 Analytical approaches to the identification of fired lime in plaster and daub

Most plasters and daubs (as well as mortars) consist of a mixture of two or three phases: the binder and aggregate phases are present in all types, and the temper phase present in some. The binder phase is the matrix which is typically composed of sediment with a fine texture (either clay or silty clay) mixed with water to form a mud. The aggregate phase typically contains sand-sized materials and serves to provide bulk and structure. The temper phase is composed of plant materials, such as chaff or dung, which prevents cracking and improves the tensile strength of the final product.

When studying calcareous plasters and daubs it is important to determine whether the source of the fine carbonate is fired lime or another type of material. The presence or absence of fired lime can be important for understanding the production sequence, not only in terms of the time and energy invested in the construction material,13 but also the chemical and mechanical properties of the finished product. As the parent material and the final product are both composed of CaCO3, it is almost impossible to identify fired lime using chemical composition alone,14 although a mixed mineralogy of calcite and aragonite (both polymorphs of CaCO3) may help to confirm a pyrogenic origin in well-preserved samples.15 Furthermore, certain types of unheated soft carbonates such as marl and chalk were used in Neolithic times as ingredients in plaster or mortar.16 Tests of slaking are used to determine the resistivity of a material to water by measuring or observing the extent to which a material disaggregates when submerged. Unlike binders produced from calcareous muds, binders derived from lime will not disaggregate when exposed to water owing to their low solubility.17 Lime binders are also typically harder than other types of materials. In prehistoric archeological contexts, however, water resistance and hardness may be impacted both positively and negatively through diagenesis, as well as dilution of the lime with other materials. Owing to the difficulty of identifying fired lime in the field, researchers have developed a series of additional laboratory analyses.18

Calcareous construction materials contain minerals such as calcite or aragonite which are present either in the binder phase or as coarse aggregate inclusions. Coarse calcareous inclusions can be fragments of limestone, shell or other recycled and crushed anthropogenic materials. Sources of fine calcareous material in daub can vary and include diffuse soil carbonates, calcareous muds from lakes or lagoons, soft limestone such as chalk, wood ashes or fired lime. These components may be naturally present within source materials used in the production of the daub or plaster, or may be intentionally selected in order to influence the final properties of the material including the color, mechanical strength and resistance to chemical weathering. Of the potential sources of fine calcareous material, fired lime produces a strong final product which is resistant to water. 2.4 Fired lime Fired lime or quicklime is a calcareous substance which can be used in the binder phase of plaster, mortar or daub to improve both its mechanical strength and resistance to water. Its production process is termed the lime cycle.12 During the first step of the cycle, calcite or aragonite (CaCO3), typically in the form of limestone, is heated to temperatures of 800–900°C, driving off carbon dioxide (CO2) and leaving behind calcium oxide (CaO). After the CaO cools, it may be stored dry, mixed with water to form a putty and sealed for future use or immediately crushed and mixed with water to produce

Thin-section petrography is one approach to the description and identification of fired lime plaster. Some researchers strongly advocate that this method be applied to all suspected lime mortars when information about the production sequence, including the source of raw materials, is desired;19 although common in the field of conservation, the technique has been applied to

13 14 15

10 11 12

16

E.g. Ward 2006: 123. E.g. Kingery et al. 1988: 226–27. Leslie and Hughes 2002: 257.

17 18 19

361

E.g. Goren and Goring-Morris 2008. Gourdin and Kingery 1975: 137; Kingery et al. 1988: 221. Toffolo and Boaretto 2014. Goren and Goldberg 1991: 135; Matthews 2005b: 365–66, 388. Kingery et al. 1988: 222. Weiner 2010: 189–90, 284–86. E.g. Leslie and Hughes 2002: 257.

Dosariyah a variety of Neolithic archeological and experimental samples.20

of discrete plaster layers with thicknesses in the centimeter to sub-millimeter range.29

In thin section, the binder phase of pure lime plasters and other materials is dark gray under plane-polarized light, and exhibits high-order interference colors and crystallitic birefringence fabrics under cross-polarized light owing to the small crystal size of the calcium carbonate.21 The aggregate and temper phases can be identified as inclusions of sand-sized minerals and rock fragments, and vegetal material in the form of phytoliths or humified organic matter.22 Lime plasters may also contain ash particles derived from the firing stage.23 Although certain morphological characteristics may be used to distinguish lime from ashes and other fine-grained carbonates,24 it may be impossible to distinguish between plaster and diffuse microcrystalline soil carbonates, groundwater carbonates or marls, especially when it is suspected that lime was mixed with fine sediment to increase volume. Petrography is especially useful for identifying raw material sources and production pathways, when plasters are poorly manufactured and contain inclusions of unslaked or partially slaked lime or fragments of incompletely combusted limestone.25 Some researchers have also utilized observations of microscopic structures at high resolution using scanning electron microscopy (SEM) to identify fired lime,26 but some of their results have been called into question by later re-analysis of the same materials.27 According to Weiner, diagenesis leads to changes in lime binder crystal morphologies over time.28

Fourier transform infrared spectroscopy (FTIR) and X-ray diffraction (XRD) are two other approaches which can be used in the documentation of the mineralogical composition of daub and the identification of fired lime. FTIR is a commonly used technique for the rapid identification of minerals based on characteristic peaks in infra-red spectra, which correspond to vibrational energies of specific bonds in molecules or solids.30 Chu et al. and Regev et al. report a technique for distinguishing between calcites of geogenic and anthropogenic origin using ratios of IR peak heights, and changes in these ratios with successive grinding of powder samples.31 According to Chu et al., different types of calcite have varying degrees of structural order of the calcite crystal lattice based on their formation history.32 Geogenic calcites such as spar and limestone have higher crystal order relative to pyrogenic calcites, such as wood ashes and plaster. Regev et al. used these trends to identify lime plaster from archeological sites, but caution that interpretation of the results should take re-crystallization or weathering of the suspected lime samples into consideration and the analyses must be based on materials local to the area.33 Both FTIR and XRD can also help in the identification of lime when poor production technique leads to the presence of unslaked or partially slaked lime. These materials contain calcium oxide or the mineral portlandite (Ca(OH)2). In addition, experimental studies have demonstrated that some fired lime plasters contain aragonite in addition to calcite.34 FTIR and XRD can be used to identify these minerals. Finally, like petrography, both FTIR and XRD can be used to determine the mineral composition of the aggregate phase.

Quantitative and semi-quantitative elemental analyses can be employed in the study of suspected lime plasters in order to document the overall chemical composition of the bulk sample. When visual inspection or petrography indicates the presence of multiple layers of plaster, careful sub-sampling or in situ elemental measurements are possible. Microscopic X-ray fluorescence (µ-XRF) is an ideal approach in these cases as measurements can be conducted on loose samples of plaster, resin-impregnated plaster fragments and petrographic thin sections. This approach allows for the rapid and minimally destructive characterization

Stable isotopic analyses have also been reported as approaches to the identification of fired lime.35 Measurements conducted along transects reveal a characteristic linear relationship between carbon and oxygen isotopic ratios, which Kosednar-Legenstein et al. attribute to the fractionation during the process of recarbonation which proceeds from the outside of the material towards its core.36 This approach is especially useful for relatively pure lime samples without inclusions of other types of carbonate and with little post-depositional diagenesis.

20 Goren and Goldberg 1991; Arpin 2004; Matthews 2005b: 371, 381–88; Karkanas 2007; Goren and Goring-Morris 2008; Karkanas and Efstratiou 2009; Mentzer and Quade 2013. 21 Courty et al. 1989: 121; Leslie and Hughes 2002: 261. 22 Goren and Goldberg 1991: 135–37. 23 Goren and Goldberg 1991: 137; Leslie and Hughes 2002: 259; Karkanas 2007: 790. 24 Mentzer and Quade 2013: 92–93. 25 See Courty et al. 1989: 121; Leslie and Hughes 2002: 259; Karkanas 2007: 780–90; Goren and Goring-Morris 2008: 792–93. 26 Gourdin and Kingery 1975; Kingery et al. 1988: 222–23. 27 Matthews 2005b: 388. 28 Weiner 2010: 190–91.

Mentzer and Quade 2013. Weiner et al. 1993; Weiner 2010: 71–98, 275–316. 31 Chu et al. 2008; Regev et al. 2010a. 32 Chu et al. 2008: 906. 33 Regev et al. 2010a; 2010b. 34 Toffolo and Boaretto 2014 35 Rafai et al. 1998; Kosednar-Legenstein et al. 2008; Mentzer and Quade 2013: 92–93. 36 Kosednar-Legenstein et al. 2008. 29 30

362

Susan M. Mentzer et al.: Morphological and Geochemical Analysis of Plaster Samples 2.7 Analytical approaches to the identification of hydraulic plasters

Finally, positive identification of lime in construction materials may of course be supported by the archeological identification of lime plaster production areas containing combustion features, such as open hearths containing fragments of burnt limestone37 or the remains of lime kilns.38

According to Charola and Henriques, hydraulic plasters and mortars can be difficult to identify.41 One main approach is to document the presence of amorphous silicate phases and hydraulic compounds using SEM, XRD or thermo-gravimetric analysis (TG-DTA).42

2.6 Hydraulic plasters

The first approach employs a scanning electron microscope, first to observe the microstructure and fabric of the plaster at high magnification using the secondary electron detector, and second to document the elemental composition of the materials using the secondary X-ray detector and energy dispersive spectrometer (SEM-EDS). Gels and minerals containing calcium silicate hydrates (CSH) have characteristic micro-structures. Needle-shaped crystals, typically less than 10 µm in length, have been observed in hydraulic plasters or mortars at high magnification.43 Furthermore, the presence of calcium, silicon and (sometimes) aluminum in these needles can be confirmed using elemental measurements with the X-ray detector (SEM-EDS). This combined morphological and chemical analysis has been used to identify several types of hydraulic compounds in both archeological samples and experimental plasters.44 Thus, SEM-based identifications of hydraulic lime are more convincing than those of more basic fired lime because in addition to structural changes, multiple chemical phases are present in the final product which differ from the original parent material.

A sub-set of fired lime plasters (hydraulic plasters) also contain, in addition to the lime, siliceous components which may be present either naturally within the limestone used to produce the lime, or were added. These compounds react with the lime during the slaking and recarbonation processes to yield silicate hydrate compounds (typically amorphous calcium silicate hydrates) which serve to further increase the mechanical strength of the plaster and also render it fully waterproof. These hydraulic plasters were produced intentionally on an industrial scale from Roman times when siliceous additives in the form of volcanic ash were mined and added to slaked lime.39 These types of plasters are termed ‘pozzolanic’ in reference to the regional source of the volcanic ash. In Neolithic contexts, three types of hydraulic plasters or mortars could feasibly be produced in open fires or simple kilns: natural hydraulic plaster (also known as calcined hydraulic lime), artificial hydraulic plaster and pozzolanic hydraulic plaster. Natural hydraulic plasters are produced when limestone rich in siliceous compounds, such as clay or diatoms, is heated to produce lime. Artificial hydraulic plasters are produced when limestone and clay from another source are first mixed, then fired together. Pozzolanic hydraulic plasters are produced when lime is first slaked in water, then mixed with siliceous additives. Non-anthropogenic forms of siliceous additives are not limited to volcanic ash and can include diatomite, crushed pumice or volcanic tuff. Anthropogenic forms include fly ash and fragments of brick. During the setting phase, the siliceous additives react with the calcium hydroxide to form poorly crystalline calcium silicate hydrates and calcium aluminum hydrates.

XRD measurements can also help the identification of hydraulic plasters depending on whether the compounds are crystalline. Several researchers have reported the identification of the mineral gehlenite, or gehlenite hydrate, in certain types of natural hydraulic mortars.45 Gehlenite (Ca2Al[AlSiO7]) forms when limestone containing silicates is heated to temperatures between 900 and 1150°C.46 Other compounds and minerals documented in hydraulic lime plasters using XRD include calcium aluminum hydrate (Ca4Al2O7.xH2O) and calcium silicate hydrate ((5Ca2Si)4.6H2O), tobermorite (Ca5Si6O16(OH)2·4H2O) and afwillite (Ca3(SiO3OH)2·2H2O).47 In contrast, amorphous calcium silicate hydrates (gels) are difficult to identify with XRD.48 Böke et al. also report the use of differential thermal gravimetric analysis (TG-DTA) to identify the presence of hydraulic compounds, which lose

Much of the research on hydraulic plasters and mortars focuses on materials from historical contexts owing to the fact that hydraulic plasters were not routinely produced prior to Roman times. Older examples of hydraulic plasters do exist,40 but the technology appears to have been localized, in some cases through the result of fortuitous use of raw materials.

Charola and Henriques 2000. Callebaut et al. 2001; Böke et al. 2008; Regev et al. 2010b. 43 Callebaut et al. 2001: fig. 5; Lanas et al. 2006: fig. 12; Böke et al. 2008: fig. 11. 44 Lewin 1982; Affonso and Pernicka 2001: 11–12; Böke et al. 2008. 45 Callebaut et al. 2001. 46 Callebaut et al. 2001: 400. 47 Lewin 1982. 48 Middendorf et al. 2005: 762. 41 42

Kingery et al. 1988: 223; Matthews 2005b: 371, 381–88. Arpin 2004; Goren and Goring-Morris 2008. 39 Ward-Perkins 1981: 97–120. 40 Affonso and Pernicka 2001; Regev et al. 2010b; Theodoridou et al. 2013. 37 38

363

Dosariyah 6  x  9  cm size, glued to a glass slide and ground to a thickness of 30  µm. Three of the samples were sliced along two perpendicular axes to obtain different cross sections. The thin sections were studied using standard petrographic and stereoscopic microscopes at a variety of magnifications and under plane-polarized (PPL), cross-polarized (XPL), reflected (RL), and oblique incident (OIL) light. Descriptions of the void shapes and fabric followed terminology developed for the field of soil micromorphology.56 All analyses were conducted in the Geoarcheology Laboratory at the University of Tübingen.

structurally bound water at temperatures between 200 and 600°C.49 As many researchers have noted, it is possible to produce hydraulic plasters accidentally when the limestone used in the production of fired lime contains sufficient quantities of clay or other siliceous materials.50 Inclusions of materials related to the firing process, including ash, may also inadvertently yield hydraulic compounds.51 The analytical approaches described above allow one to identify hydraulic compounds but not to determine the degree of intentionality in their production. For this, it is necessary to look at other lines of evidence, such as function and raw material sources.

Thin-section petrography was the first analytical technique applied to the Dosariyah specimens and some of the initial observations were used to select additional approaches. In two cases, possible analyses were discounted due to findings from the petrography. The presence of at least two types of carbonate (sand-sized fragments of limestone and fragments of calcified plant tissues) in addition to the potential fired lime binder, precluded the use of stable isotopic measurements to identify recarbonation fronts, and this technique was therefore not applied to the Dosariyah specimens. Furthermore, all samples contained fragments of organic plant tissue: as organic materials break down at temperatures between 200 and 600°C, it was determined that it would not be possible to use TG-DTA to identify hydraulic compounds within the samples.

3 Analytical Methods In addition to the visual identification, photography and 3D scanning reported above, eleven analytical methods were applied to the Dosariyah plaster samples. Our choice of methods derived from previously published studies of fired lime and hydraulic mortars and plasters summarized above (Sections 2.5, 2.7), as well as the main research questions listed in Section 1. Moreover, we have adopted an integrative approach, balancing quantitative chemical data with qualitative microanalyses, such as FTIR, µ-XRF, petrography and SEM. Combinations of multiple analytical methods are typical for the study of historical mortars and plasters for conservation purposes,52 but are less common in prehistoric archeological contexts. By using this approach, we avoid issues of scale in interpretation of the results, such as those reported by Goren and Goldberg.53 We should note that additional international standardized measurements of the physical properties of the materials, such as compressive strength and porosity, were not conducted owing to the small size of the plaster fragments. These values have been reported previously for the Dosariyah samples analyzed by Malinowski and Frifelt54 but both Karkanas and Leslie and Hughes have suggested that these measures may be inaccurate when applied to very old archeological samples because of diagenesis.55

3.2 Organic petrology Following thin-section description, one sample was selected for further analysis using organic petrology. The study was conducted on a resin-impregnated piece of sample DOS2012-31042 left over from the thinsection production process. Dry fine polishing (without lubricants) of the analytical face was completed in several steps. The petrographic analysis was performed with a Leica DMRX/MPV-SP microscope photometer in reflected white-light (RLo) and incident ultra-violet light (RVLo), and in plane-polarized (RPPLo) and crosspolarized light (RXPLo), under oil immersion.57 The magnifications ranged from 200 to 500x.

3.1. Thin-section petrography

Description and classification of the organic microcomponents (macerals) was based on the nomenclature of macerals in brown coal and coal.58 This initial phase of analysis was used to identify the types of organic tissues present in the sample. The second step of analysis was the measurement of reflectance of the plant tissues, which is an established means of measuring the maturation degree of organic matter in

Petrographic thin sections were prepared from five samples. The fragments were embedded in polyester resin diluted with styrene, and catalyzed with MEKP. Once hardened and cured in an oven overnight, the embedded samples were sliced with a rock saw to Boke et al. 2008: 871–73. E.g. Leslie and Hughes 2002: 257; Böke et al. 2008: 866–67; Regev et al. 2010b: 3005. 51 Leslie and Hughes 2002: 259. 52 E.g. Middendorf et al. 2005. 53 Goren and Goldberg 1991: 132. 54 Malinowski and Frifelt 1993: table 1. 55 Karkanas 2007; Leslie and Hughes 2002: 260. 49 50

56 57 58

364

Stoops 2003. Taylor et al. 1998. Taylor et al. 1998; Sýkorová et al. 2005.

Susan M. Mentzer et al.: Morphological and Geochemical Analysis of Plaster Samples peats, brown coals, coals and sedimentary rocks.59 In soil and paleoenvironmental studies, measurements of reflectance are also used to characterize the humification process60 and to characterize charcoal particles.61 Random reflectance in oil (mean % Rr) of several organic particles was measured according to standard procedure.62 All organic petrology analyses were conducted in the Laboratories for Applied Organic Petrology.

were used to calculate first, the fraction (by mass) of insoluble material and second, the ratio of binder to aggregate. The insoluble fraction was then powdered and analyzed with FTIR following procedures described below. Finally, six samples were subjected to typical phytolith extraction procedures65 in order to obtain the portion of the insoluble fraction that, if present, would contain phytoliths and diatoms. The extracted materials were processed into permanent grain mounts using Entellan New (Merck) mounting medium on glass slides with coverslips, and were studied at 20x magnification using PPL in order to document the presence of phytoliths and diatoms.

3.3 Calcium carbonate abundance Calcium carbonate abundance was calculated based on the volume of evolved CO2 upon reaction with 10% hydrochloric acid (HCl) following the Chittick method.63 Simple calculations are based on the assumption that 100% of the evolved gas is derived from CaCO3. The presence of dolomite (CaMg(CO3)2) which would be indicated by the mineralogical or elemental analyses (see below) requires additional correction to the results.

3.5 Density The dry bulk densities were calculated for five samples using the volume of each sample recorded during the 3D scanning procedure, and the mass in grams. Equation: p = m/v; where p is the density, m is the mass, and v is the volume. These calculations are equivalent to the ‘apparent density’ measured by Malinowsky and Frifelt.66

3.4 Slaking, dissolution, extraction of insoluble materials and calculation of binder/aggregate ratio In order to test the resistance of the materials to water, sub-samples of two specimens (c.1  cm3 each) were placed in covered beakers and submerged in 25  ml of de-ionized water for one week. For comparison, two samples of modern lime plaster with and without an aggregate phase were also tested. The samples were monitored daily to provide visual assessment of their decomposition. In addition, after the first 24 hours, each sample was sonicated for 15 minutes per day to accelerate any breakdown which was occurring. Throughout this week-long period, the coherence of the samples was observed and the pH of the water measured in order to document dissolution of calcium carbonate.

3.6 Microscopic X-ray fluorescence (µ-XRF) In situ measurements of the relative abundances of major elements (atomic mass of 12 and above) were conducted on four of the five petrographic samples using a dual-detector Bruker M4 Tornado benchtop µ-XRF analyzer equipped with a Rh X-ray tube and two silicon drift detectors. Elemental distribution maps were generated under full vacuum at 50 kV and 600 µA, with a spot size of 20 µm and a dwell time of 10–20 ms per pixel. Spacing between measurements ranged from 5 to 70 µm, depending on the desired resolution of the feature. Spectral and image processing was conducted using the Bruker Esprit software. Test measurements were conducted on 1) a fragment of plaster sawn in half with the second half of the fragment embedded in polyester resin; and 2) a fragment of plaster embedded in resin and its corresponding petrographic thin section. These tests were intended to document the impacts of resin and thin sectioning on the chemical composition of the samples. The remaining mapping measurements were conducted on plaster fragments embedded in resin. All measurements were conducted on flat surfaces.

Following Regev et al., the calcareous component of six samples was removed using acetic acid in order to obtain the insoluble fraction without altering the composition of potential hydraulic compounds.64 The dry samples were weighed, then submerged in 99% acetic acid until bubbles were no longer observed. The samples were centrifuged to separate the insoluble fraction from the acid, and the acid removed by pipette. De-ionized water was then repeatedly added to the samples to dilute them and the mixtures centrifuged with the supernatant removed each time until the pH reached 7. The samples were then oven-dried at 60°C and weighed. The masses before and after dissolution 59 60 61 62 63 64

3.7 Scanning electron microscopy (SEM) SEM analyses were conducted in two phases. During the first phase of analysis, aggregates of three samples

Borrego et al. 2006. Jacob 1974; 1980; Schwaar et al. 1991. Jones et al. 1991; Guo and Bustin 1998; Bustin and Guo 1999. Taylor et al. 1998. Dreimanis 1962. Regev et al. 2010b: 3003–3004.

65 66

365

Piperno 2006: 90–96. Malinowsky and Frifelt 1993: table 1.

Dosariyah . The original powder sample was then ground again and sub-sampled three times for a total of five spectra produced on powder ranging from lightly to intensively ground. Heights of the calcite ν2, ν3 and ν4 peaks were measured according to methods described by Chu et al.69 Reference samples of modern lime plaster, archeological lime plaster, fresh wood ash, marine and freshwater limestone and local calcareous sediments were also obtained and processed. Spectra generated from both phases of loose sediment analysis were also used to document the presence of different types of siliceous compounds. Finally, the acid insoluble fraction (see Section 3.4) was also analyzed.

of plaster were adhered to carbon tape and imaged using the secondary electron detector of a LEO 1450 VP SEM housed in the Department of Geosciences at the University of Tübingen. Compositional analyses were conducted over points and areas using the EDS detector. The analyses were conducted at low magnifications, and therefore the analyses yielded general information about the structure and chemical composition.

1

During the second phase of analysis a set of five loose samples was gently crushed and mounted on carbon tape. The mounted samples were then sputter-coated with a thin layer of platinum (Pt) to minimize charging. The samples were first imaged, again using the secondary electron detector of the LEO 1450 VP SEM instrument housed in the Department of Geosciences, at up to c.35,000x magnifications with an accelerating voltage of 15 keV. Based on the positive results of these analyses, a second set of measurements was conducted at higher magnifications (e.g. 90,000x) using a Hitachi SU8030 field emission SEM (FESEM) instrument housed in the Institute of Physical and Theoretical Chemistry at the University of Tübingen. These measurements included images collected using the secondary electron detector, as well as elemental data collected using the EDS detector. Elemental measurements were collected as points, areas and line scans, typically using an accelerating voltage of 8–10 keV.

Microscopic FTIR measurements (µ-FTIR) were conducted on the samples in thin section using an FTIR microscope attached to the laboratory bench (Cary 670; Agilent Technologies). Spectra were collected over a range of 4000–550  cm-1 with 32 co-added scans at resolutions of 4  cm-1 and, in some cases 1  cm-1, using a germanium crystal ATR objective with an effective spot size of 50–70  µm. A total of 4280 spectra were collected as automated grids distributed on each of the thin sections, in order to observe spatial variability in carbonate mineralogy and to detect any spatially discrete concentrations of amorphous silicates or calcium hydroxide. Where present, multiple layers of plaster were targeted within the same sample in order to investigate potential variations in mineralogy of the matrix and aggregates. Infra-red peaks associated with the polyester resin were removed from the sample spectra using the Resolutions Pro software package (Agilent Technologies) when necessary. Mineral identification proceeded as above. All FTIR analyses were conducted in the Geoarcheology Laboratory at the University of Tübingen.

3.8 Fourier transform infrared spectroscopy (FTIR) FTIR analyses were conducted on fragments of plaster using two different methods. First, basic spectra in the mid-IR region (4000–400  cm-1) were generated on powdered samples using a laboratory grade bench instrument (Cary 660; Agilent Technologies) equipped with a diamond crystal attenuated total reflectance (ATR) accessory (GladiATR; Pike Technologies). Spectra were produced from 32 co-added scans at a resolution of 4 cm-1. The peaks were matched to minerals in a digital database composed of ATR spectra made available by the RRUFF project67 and a personal mineral collection. This first phase of analysis was used to determine whether the plasters were composed of calcite or gypsum plaster.

3.9 X-ray diffraction (XRD) XRD analyses were conducted on six loose samples in the Applied Mineralogy X-ray Analysis Laboratory at the University of Tübingen. Measurements were conducted over 1–6 hours on several grams of prepared powder at 40  kV, 20  mA, a step size of 0.008 degrees, and 600–700 seconds per step using a Bruker AXS D8 AdvanceTM diffractometer with a Vantec 1 detector, parallel beam optic and primary divergence slit of 0.2 mm. Mineral identifications were performed using the International Centre for Diffraction Data (ICDD) PDF4 database. Quartz, which is present in each sample, was used as the internal standard for calibration. Peak fitting was then conducted on peaks associated with the mineral calcite, and measurements of the full width at half maximum obtained to estimate differences in crystallite domain size.

A second phase of analysis was conducted using a successive grinding technique described by Regev et al.68 Several milligrams of sample were lightly ground using an agate mortar and pestle, and a subsample of