Roman Period Statuettes in the Netherlands and beyond: Representation and Ritual Use in Context 9789048557004

Table of contents :
CONTENTS
ACKNOWLEDGEMENTS
1 A rare category of artefacts
2 Catalogue and description of find-spots
3 Spatial and iconographic analysis
4 Rituals with statuettes
5 Summary and conclusions
Abbreviations
Bibliography
List of maps, figures and drawings
Figures
Tables

Citation preview

Roman Period Statuettes in the Netherlands and beyond

A M STERDAM A RCHA EO LO GI CAL S T UD I E S 3 0 Editorial Board: Prof. dr. N. Roymans Prof. dr. P.A.J. Attema Prof. dr. E.M. Moormann

16. E.M. Moormann: Divine Interiors. Mural paintings in Greek and Roman sanctuaries ISBN 978 90 8964 261 5

Other titles in the AAS series:

18. N. Roymans / G. Creemers / S. Scheers: Late Iron Age Gold Hoards from the Low Countries and the Caesarian Conquest of Northern Gaul ISBN 978 90 8964 349 0

1.

N. Roymans (ed.): From the Sword to the Plough. Three Studies on the Earliest Romanisation of Northern Gaul Open Access edition: http://dare.uva.nl/record/19675

2.

T. Derks: Gods, Temples and Ritual Practices. The Transformation of Religious Ideas and Values in Roman Gaul Open Access edition: http://dare.uva.nl/aup/en/record/172370

3.

A. Verhoeven: Middeleeuws gebruiksaardewerk in Nederland (8e – 13e eeuw) Open Access edition: http://dare.uva.nl/aup/en/record/172373

4.

F. Theuws / N. Roymans (eds): Land and Ancestors. Cultural Dynamics in the Urnfield Period and the Middle Ages in the Southern Netherlands Open Access edition: http://dare.uva.nl/aup/en/record/172372

5.

6.

J. Bazelmans: By Weapons made Worthy. Lords, Retainers and their Relationship in Beowulf Open Access edition: http://dare.uva.nl/aup/en/record/172337 R. Corbey / W. Roebroeks (eds): Studying Human Origins. Disciplinary History and Epistemology Open Access edition: http://dare.uva.nl/aup/en/record/172272

7.

M. Diepeveen-Jansen: People, Ideas and Goods. New Perspectives on ‘Celtic barbarians’ in Western and Central Europe (500-250 BC) Open Access edition: http://dare.uva.nl/aup/en/record/172273

8.

G. J. van Wijngaarden: Use and Appreciation of Mycenean Pottery in the Levant, Cyprus and Italy (ca. 1600-1200 BC). The Significance of Context Open Access edition: http://dare.uva.nl/aup/en/record/172274

9.

F.A. Gerritsen: Local Identities. Landscape and community in the late prehistoric Meuse-Demer-Scheldt region Open Access edition: http://dare.uva.nl/aup/en/record/172820

10. N. Roymans: Ethnic Identity and Imperial Power. The Batavians in the Early Roman Empire Open Access edition: http://dare.uva.nl/aup/en/record/172930 11. J.A.W. Nicolay: Armed Batavians. Use and significance of weaponry and horse gear from non-military contexts in the Rhine delta (50 BC to AD 450) Open Access edition: http://dare.uva.nl/aup/nl/record/397232 12. M. Groot: Animals in ritual and economy in a Roman frontier community. Excavations in Tiel-Passewaaij Open Access edition: http://dare.uva.nl/aup/en/record/301888 13. T. Derks & N. Roymans (eds): Ethnic Constructs in Antiquity. The role of power and tradition Open Access edition: http://dare.uva.nl/aup/en/record/301890 14. T. D. Stek: Cult places and cultural change in Republican Italy. A contextual approach to religious aspects of rural society after the Roman conquest ISBN 978 90 8964 177 9 15. P. A.J. Attema / G.-J. L.M. Burgers / P. M. van Leusen: Regional Pathways to Complexity. Settlement and land-use dynamics in early italy from the bronze age to the republican period ISBN 978 90 8964 276 9

17. N. Roymans / T. Derks (eds): Villa Landscapes in the Roman North. Economy, Culture and Lifestyles ISBN 978 90 8964 348 3

19. D. S. Habermehl: Settling in a Changing World. Villa development in the northern provinces of the Roman Empire. ISBN 978 90 8964 506 7 20. D. G. Yntema: The Archaeology of South-East Italy in the first millenium BC. Greek and native societies of Apulia and Lucania between the 10th and the 1st century BC. ISBN 978 90 8964 579 1 21. M. Fernández-Götz: Identity and Power. The Transformation of Iron Age Societies in Northeast Gaul. ISBN 978 90 8964 597 5 22. N. Roymans / T. Derks / H. Hiddink (eds): The Roman Villa of Hoogeloon and the Archaeology of the Periphery. ISBN 978 90 8964 836 5 23. A. Van Oyen: How Things Make History. The Roman Empire and its Terra Sigillata Pottery. ISBN 978 94 6298 054 9 24. M. Groot: Livestock for Sale: Animal Husbandry in a Roman Frontier Zone. ISBN 978 94 6298 080 8 25. D. Lentjes: Landscape and Landuse in First Millennium BC Southeast Italy. Planting the Seeds of Change. ISBN 978 90 8964 794 8 26. N. Roymans / S. Heeren / W. De Clerq: Social Dynamics in the Northwest Frontiers of the Late Roman Empire. Beyond Decline or Transformation. ISBN 978 94 6298 360 1 27. M. Pitts: The Roman Object Revolution. Objectscapes and Intra-cultural Connectivity in Northwest Europe. ISBN 978 94 6372 820 1 28. J. de Bruin: Border Communities at the edge of the Roman Empire. Processes of Change in the Civitas Cananefatium ISBN 978 94 6372 810 2 29. M. A. Wijnhoven: European Mail Armour. Ringed Battle Shirts from the Iron Age, Roman Period and Early Middle Ages ISBN: 978 94 6372 126 4

Roman Period Statuettes in the Netherlands and beyond R E P R E S E N TAT I O N A N D R I T UA L U S E I N C O N T E X T

C HRISTEL V EEN

AMSTERDAM UNIVERSITY PRESS

This book meets the requirements of ISO 9706: 1994, Information and documentation – Paper for documents – Requirements for permanence. Cover Illustration: Image left: a rattle in the shape of a terracotta horned bird, found in Nijmegen in a children’s cremation grave dating to the Tiberian-Neronian period. Photo: Ronny Meijers. Collectie Museum Het Valkhof, Nijmegen. Image right: bronze statuette of Hercules holding a kantharos, found just north of the river Waal in Nijmegen. Photo: Ronny Meijers. Collectie Museum Het Valkhof, Nijmegen. Cover design: Kok Korpershoek, Amsterdam Lay-out: Bert Brouwenstijn, Vrije Universiteit Amsterdam

Photos front and back cover: Ronny Meijers, Museum Het Valkhof, Nijmegen

ISBN: 978 94 6372 938 3 e-ISBN: 978 90 4855 700 4 (pdf) NUR: 682 DOI: 10.5117/9789463729383

© C.W.A. Veen / Amsterdam University Press, Amsterdam 2023 All rights reserved. Without limiting the rights under copyright reserved above, no part of this book may be reproduced, stored in or introduced into a retrieval system, or transmitted, in any form or by any means (electronic, mechanical, photocopying, recording, or otherwise), without the written permission of both the copyright owner and the editors of this book.

For Stephan and Alba

CONTENTS ACKNOWLEDGEMENTS

xi

1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9

A R A R E C AT E G O RY O F A R T E FAC T S Introduction Theories on Romanization Lived religion The life path of statuettes Distribution and use The dataset The areas north and south of the Rhine and the Dutch part of the Roman limes zone Definition of terms Structure of the text

1 1 2 5 5 7 8 9 10 17

2 2.1 2.2

C ATA L O G U E A N D D E S C R I P T I O N O F F I N D - S P O T S Introduction The area north of the Rhine 2.2.1 Imported goods: evidence of trade or diplomatic gifts? 2.2.2 Friesland and Groningen: the northern coastal area 2.2.3 North Holland: the northwestern coastal area 2.2.4 Drenthe, Overijssel and the northern part of Gelderland: the northeast The area south of the Rhine: Gallia Belgica and Germania Inferior 2.3.1 Theories on post-conquest settlement development in Lower Germany 2.3.2 South Holland and Zeeland 2.3.3 Utrecht 2.3.4 Southern Gelderland 2.3.5 North Brabant and Limburg The Dutch part of the Roman limes zone 2.4.1 The forts and associated settlements on the left bank of the Rhine 2.4.2 The forts and associated settlements on the North Sea coast Conclusions

19 19 19 19 23 30 32 38 38 41 50 53 67 80 80 90 95

S PAT I A L A N D I C O N O G R A P H I C A N A LY S I S Introduction Written sources and the archaeological evidence Deities, human figures and animals The area north of the Rhine 3.4.1 Spatial analysis: a relation between size and function Iconographic analysis 3.5.1 The iconography of the Mars and Mercury statuettes 3.5.2 The Isis-Venus figurines from Houwerzijl 3.5.3 Jupiter-Sabazius? 3.5.4 A bronze mother goddess north of the Rhine? 3.5.5 Mounted warriors 3.5.6 The warrior from Ezinge 3.5.7 A male figurine from Wirdum

99 99 99 101 103 104 109 109 111 112 112 113 115 117

2.3

2.4

2.5 3 3.1 3.2 3.3 3.4 3.5

3.6 3.7

3.8

3.9 4 4.1 4.2

4.3

4.4

3.5.8 Domestic animals 3.5.9 The bear from Wirdum Conclusions on the area north of the Rhine The area south of the Rhine: spatial analysis 3.7.1 The rural settlements 3.7.2 The urban settlements 3.7.3 The limes zone 3.7.4 The cemeteries 3.7.5 Cemeteries: a relation between find-context and material The area south of the Rhine: iconographic analysis 3.8.1 A relation between material and subject 3.8.2 Interpretatio and creolisation 3.8.3 The iconography of the Hercules statuettes 3.8.4 The iconography of Hercules Magusanus 3.8.5 The veneration of Hercules Magusanus 3.8.6 Two controversial dedications 3.8.7 Venus and water nymphs 3.8.8 The iconography of mother goddesses 3.8.9 Venerated as mother goddesses 3.8.10 A horse-riding acrobat 3.8.11 Jupiter-Ammon in Ulpia Noviomagus 3.8.12 A mysterious lady from Vechten 3.8.13 Minerva or a gladiator? 3.8.14 The thorn-puller 3.8.15 The power of being different 3.8.16 ‘Exotic’ representations 3.8.17 Animals with a mission 3.8.18 Ithyphallic dogs 3.8.19 Triple-horned bulls 3.8.20 Horned birds Conclusions on the area south of the Rhine

117 119 120 121 121 123 126 128 130 131 131 133 136 138 139 140 141 142 144 145 147 149 150 150 151 153 153 154 154 155 156

R I T UA L S W I T H S TAT U E T T E S Introduction Ritual deposition 4.2.1 A long history 4.2.2 Categories of ritual deposits 4.2.3 A secondary use of appliques 4.2.4 Deposits related to crafts? 4.2.5 A comparison with British datasets 4.2.6 Placating the gods: in anticipation or afterwards? Deposition of statuette fragments and deliberate fragmentation 4.3.1 Broken statuettes in the Neolithic and Bronze Age 4.3.2 A healing ritual with female figurines in Egypt 4.3.3 The Roman period: deliberate fragmentation in Belgium 4.3.4 Deposition of statuette fragments in the Netherlands 4.3.5 Deposition of bronze caducei Magical practices with figurines

157 157 158 158 159 169 170 171 173 175 175 177 178 178 181 181

5

4.5 Conclusions

184

SUMMARY AND CONCLUSIONS Discussion and suggestions for further research

187 190

A B B R E V I AT I O N S

192

BIBLIOGRAPHY

193

L I S T O F M A P S , F I G U R E S A N D D R AW I N G S

217

FIGURES

229

TA B L E S

307

ix

ACKNOWLEDGEMENTS This is the reworked version of my PhD thesis, which I defended in June 2018 at Radboud University Nijmegen. The accomplishment would have been impossible without the support and advice of many people. First, I would like to express my gratitude to my supervisors Eric Moormann and Frits Nearabout. Before becoming a professor at Radboud University Nijmegen, Eric Moormann was one of my lecturers when I studied Classical Archaeology at the University of Amsterdam. He offered me the opportunity to obtain my PhD degree at Radboud University and has seen the project through to the end. I am thankful for his support and confidence. As an expert in ancient history and religion, Frits Naerebout has given me much to think about. His enthusiasm, encouragement and insightful comments motivated me to look at things from different perspectives and to refine my arguments. I owe particular debts to Rien Polak and Louis Swinkels for their advice and close reading of the entire text and for saving me from many errors and inconsistencies. I am also grateful to Wouter Vos for close reading the second chapter, to Wim van Es, for his help with the excavations at Dalfsen, Wijster and Ede, and to Greg Woolf, who put me on the right track at the beginning of my journey. Financial support for the publication of this manuscript was kindly provided by Radboud Institute of Culture and History and the dr. Hendrik Muller Fonds. I would also like to express my gratitude to my colleagues and many others who in one way or another have contributed to this book. My special thanks are extended to the staff of museums, archaeological depots and departments in the Netherlands who welcomed me, let me work in their facilities and provided the data and photos I needed. And lastly, I am very grateful to Stephan and Alba Mols for their continuous loving support and patience.

xi

“It is a capital mistake to theorize before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts.”

Sherlock Holmes in A Scandal in Bohemia, Sir Arthur Conan Doyle

xiii

1

A rare category of artefacts

1.1

Introduction

The subject of this book are bronze and terracotta statuettes that represent deities, human figures and animals. They were introduced in the northwestern provinces by Roman troops from the end of the 1st century BCE onwards.1 The statuettes have been recovered from military and non-military settlements, the surrounding landscape and, to a lesser extent, from sanctuaries and graves. The first chapter deals with the objective and approach of this study, followed by a description of the dataset and the research area, the definitions of terms and the structure of the complete text. This introduction comprises a summary of how the study area became politically divided in the 1st century. The first Roman troops arrived between 19 and 16 BCE in the easternmost part of the region nowadays known as the Netherlands. They built the first military base on the Hunerberg at Nijmegen. Later, other military camps would follow. When Drusus led his army through Frisian territory in 12 BCE, he did not meet with any resistance and placed a moderate tax on the Frisii in the form of cow-hides. In 28, the first documented fight in the area took place when the Frisii revolted against the Roman prefect Olennius. According to Tacitus, Olennius demanded larger hides, the size of the much larger aurochs or their equivalent in domestic cow-hides. The Frisii could not meet his demands and handed in their herds and their land, and sent their wives and children into slavery. Since the Romans remained deaf to their complaints, the Frisii sought remedy in war. They hanged the soldiers who came to collect the taxes and Olennius fled to a nearby fort, which Tacitus calls Flevum.2 When the provincial governor, Lucius Apronius, arrived at the fort with his troops, the Frisian insurgents who had besieged it had already left to protect their own possessions.3 Apronius ordered his men to go after the enemy and a battle followed. The Romans were defeated and suffered great losses, but for some reason did not seek revenge. They did not succeed in bringing the northern tribes back into line, which, from now on, no longer paid taxes. This situation became definitive in 47, when the emperor Claudius (41-54) ordered all Roman troops to retreat behind the Rhine. The political division between the regions north and south of the Rhine had far reaching consequences for the socio-economic and cultural development of their respective populations. Yet, the division did not bring an end to the contacts between Romans and northern tribes. Hoards of denarii dating to the 2nd and 3rd centuries testify to a continuing contact with the Romans. Epigraphic evidence demonstrates that men from the Frisii enlisted in the Roman army. Moreover, to prevent raids from the north, Roman authorities also tried to maintain friendly relationships with northern tribal leaders. Apart from Germanic men who joined the Roman army, contacts between northern tribes and the Romans were from now on limited to trade and barter on a small scale. 1

Only years preceding the Common Era are referred to as BCE.

corpora coniugum aut liberorum servitio tradebant. Hinc ira

2

Tac. Ann. 4 . 72: Tributum iis Drusus iusserat modicum pro

et questus, et postquam non subveniebatur, remedium ex bello.

angustia rerum, ut in usus militares coria boum penderent,

Rapti qui tributo aderant milites et patibulo adfixi: Olennius

non intenta cuiusquam cura, quae firmitudo, quae mensura, donec Olennius e primipilaribus regendis Frisiis inpositus

infensos fuga praevenit, receptus castello, cui nomen Flevum. 3

Tac. Ann. 4.73: Quod ubi L. Apronio inferioris Germaniae pro

terga urorum delegit, quorum ad formam acciperentur. Id aliis

praetore cognitum, vexilla legionum e superiore provincia

quoque nationibus arduum apud Germanos difficilius tolera-

peditumque et equitum auxiliarium delectos accivit ac simul

batur, quis ingentium beluarum feraces saltus, modica domi

utrumque exercitum Rheno devectum Frisiis intulit, soluto

armenta sunt. Ac primo boves ipsos, mox agros, postremo

iam castelli obsidio et ad sua tutanda degressis rebellibus.

1

After 47, the river Rhine was no longer just a natural barrier, but also Rome’s northern frontier on the continent. The frontier zone was consolidated by the construction of forts and watchtowers on the left bank of the Rhine between Nijmegen and the North Sea coast. Military units were now stationed here for longer periods of time. A new infrastructure was created with roads and bridges to facilitate the transport of men and goods. The Roman forts attracted all kinds of people who wanted to make a living out of the army. They settled down near the forts, where settlements developed that probably were melting pots of people from throughout the Empire. In Batavian territory, a civil settlement was founded around 10 BCE.4 According to written sources, this settlement was Oppidum Batavorum, situated in Nijmegen’s modern city centre. After its destruction during the Batavian revolt in 69/70, a new settlement was established on the left bank of the river Waal. It became the new civitas capital of the Batavians: Ulpia Noviomagus.5 More to the west and slightly later, a settlement in the western coastal area known as Municipium Aelium Cananefatium became the civitas capital of the Cananefates.6 Some civil vici developed into successful small towns, like Maastricht and Heerlen, which became known as Mosa Trajectum and Coriovallum.7 Together with Ulpia Noviomagus and Municipium Aelium Cananefatium, Mosa Trajectum and Coriovallum developed into regional centres with urban characteristics. In the course of the 1st and 2nd centuries, Roman building techniques, building material, pottery and other artefacts also appeared in smaller settlements. Farmers, who until the conquest were mostly self-sufficient, now produced crops and livestock for the Roman army. With the profits they made, they built larger farmsteads and villas that were modelled on those in the more central parts of the Empire, with baths, heated rooms, marble wall coverings and wall paintings.

1.2

Theories on Romanization

The impact of the Roman occupation on the conquered peoples has been central to a debate that started at the beginning of the 20th century and later became known as ‘the Romanization debate’. Since numerous publications on this topic have seen the light, I will confine myself to some of the more important insights and to the consequences of the debate for the subject of this study.8 Until the second half of the 1980s, theories on Romanization mainly focused on changes in the social, economic and political organization. The first systematic studies on the transformation of religion in the Roman West appeared from the second half of the 1980s onwards.9 In 1993, it was one of the topics at an international conference held in Luxembourg, entitled Integration in the Early Roman West. The role of culture and ideology. As of the late 1990s, several studies on the Gallic and Germanic provinces saw the light that have contributed importantly to our knowledge of the transformation of religion under Roman rule.10 4

Except for the northern Rhine frontier, the exact bound-

remained occupied until 260/270. See for the discussion on

aries of the Batavian territory are unknown. It probably

the settlement and its name: Bogaers 1960, 276-312; Haalebos

covered the region between the line Woerden-Gorinchem in the west and the modern Dutch-German border in the

2000, 14, 35-39; Van Enckevort/Thijssen 2005a, 104-107. 6

reworked PhD-thesis of Jasper de Bruin: De Bruin 2019.

sandy soils of the modern province of North Brabant: Nico5

lay 2007, 3-4 and fig. 1.1.

7

See for the name Mosa Trajectum 2.3.5.

The name of the settlement is uncertain. It may have been

8

For this overview I follow Jeremia Pelgrom’s introduction to

Noviomagus or Batavodurum before it became known as

the subject, which provides a good and clear outline of the

Ulpia Noviomagus. The names Municipium Batavorum and

debate: Pelgrom 2009.

Ulpia Noviomagus occur on altars and grave markers from

9

Birley 1986; Derks 1991.

the second half of the 2nd century onwards. The name Muni-

10

See, for instance, Haynes 1993, 1997; Cancik/Rüpke 1997;

cipium Batavorum Ulpia Noviomagus has not been attested

Derks 1998; Spickermann et al. 2001; Spickermann 2008;

so far. For practical reasons, I use the name Ulpia Noviomagus

Van Andringa 2002, 2007, 2011.

for the settlement at Nijmegen-West that took off in 70 and

2

See for the name of the settlement Bogaers 1960, 303-309 and 2.3.2. The most recent study on the Cananefates is the

east. The southern part may have included a part of the

Central to most studies are the cult places and ritual practices of public religion. However, how and at what pace the transformation of religion and its material culture took place outside the public context has remained largely unclear. Studies on the subject only dealt with the description and classification of artefacts dissociated from their contexts. Their invisibility in the archaeological landscape, the sparse analytic studies within Roman archaeology on private forms of worship and the absence of native written sources on this subject account for our faulty knowledge.11 According to the Romanization model proposed by Martin Millett in 1990, Rome conducted a non-intervention policy.12 This implied that local elites were maintained in power, as long as they paid taxes and conformed to Roman principles with respect to the administration of their territory. Since the display of weaponry was forbidden by the lex Julia de vi publica, weapons could no longer function as symbols of power, as traditionally had been the case.13 Consequently, to express their authority and to distinguish themselves from lower social strata, native elites chose Roman artefacts and practices as new symbols of power. According to Millett, these symbols gradually penetrated the lower classes of society. This phenomenon of voluntarily embraced aspects of Roman culture became known as aemulatio, emulation or self-Romanization. Postmodernist critics rejected the passivity of the lower classes implied in Millett’s model. They argued that a society is a diverse collective of social groups, each with its own ideologies, ideas, norms and values. Moreover, Millett’s model only focused on socio-economic and political processes, while omitting ideological aspects. The idea was advanced that the acceptance of Roman culture depended primarily on the ideologies, norms and values of native groups.14 In more recent studies it is pointed out that, especially under Augustus, the concept of humanitas was used to define Roman identity.15 The degree of humanitas determined the success of local elites.16 The more ‘Roman’ elites looked and behaved, the more successful they were within the imperial administrative organisation. These theories restored the presumed leading role of local elites and, again, marginalised the role of the lower classes in processes of change. In order to explain how the lower social classes played a decisive role in the acceptance or refusal of practices and the material culture of a new culture, post-colonial theories on Romanization looked at our colonial past. Colonised peoples did not take over the language and culture of the colonizer, but created a new language and culture. The acceptance of elements of the colonizer’s culture did not mean that colonised peoples took over the ideas behind these elements. They took over what they appreciated or could use. Likewise, conquered peoples in the Roman period could embrace Roman artefacts and practices, they could attribute new meanings to such artefacts and practices, they could also refuse elements of Roman culture and emphasize their non-Roman identity.17 Described as flexible, discrepant or fragmented identities, they allowed people to function in varying situations and in different communities.18 Immigrants and natives could be portrayed on grave stelai wearing Roman clothing, native garments, or both.19 Groups and individuals could appropriate and reinterpret cultural elements and social practices, creating new cultural identities, processes referred to as creolisation, bricolage or hybridisation. It has been argued that especially in frontier zones, a variety of cultural elements and influences were picked up and mixed, resulting in a ‘frontier style’,

11

In the past, studies on Roman private religion have focused

16

Pelgrom 2009, 164.

on Pompeii, Herculaneum and Ostia and are predominant-

17

Pelgrom 2009, 165-167.

ly descriptive. See, for instance, Boyce 1937; Orr 1972; for

18

See for the term ‘discrepant identities’ Mattingly 2006, 203245; for ‘fragmenting identities’ Hingley 2005, 91-116.

private religion in Ostia Bakker 1994. 12

19

Millett 1990a, 1990b.

13

Dig. 48.6.1; Millett 1990b, 37-38.

stelai in the Rhine-Moselle area. She has demonstrated

14

Pelgrom 2009, 161-163 and note 17.

that attributing a native or non-native identity based on

15

Humanitas in the Imperial period stood for a combination

clothing is dangerous. It also appears from her study that

of qualities such as humanity, literacy and civilisation. It

there were regional differences in the way people adopted

embodied what it meant to be Roman: Woolf 1998, 54-60.

Roman-style monuments and Roman dress: Rothe 2009.

Ursula Rothe has investigated people portrayed on grave

3

a distinct frontier material culture along the borders of the Roman Empire.20 An example is the grave stele commissioned by the Syrian Barates for his wife Regina, a Catuvellaunian freedwoman at Arbeia (modern South Shields), a vicus and Roman fort on Hadrian’s Wall. The stele dates to the second half of the 2nd century and is a typical Roman funerary monument, but the sculptural style and epitaph, written in Latin and flawless Aramic, testify to the presence of a Syrian community on Hadrian’s Wall. The seated woman in the relief wears garments that resemble the ethnic clothing of a variety of groups in the western Roman provinces. As a Syrian, Barates could have chosen a typical Palmyrene funerary bust to commemorate his wife, but instead, he commissioned a typical Roman funerary monument to communicate the couple’s integration as an immigrant and an ex-slave in the diverse and complex society in the British northern limes zone. The depicted spinning paraphernalia and basket filled with wool point at Regina’s skills in spinning and wool working. She is presented as a respectable, industrious wife, which must have been an equally important message in a community where legal constraints prevented valid Roman marriage.21 When studying a group of artefacts, a large research area or a large dataset is required, as well as dateable artefacts and find-contexts.22 These conditions do not apply to the study area and the majority of statuettes that are central to this study. Moreover, although the changing appearance of statuettes suggest a transformation of cults, the identities of the owners of these statuettes remain invisible to us. Therefore, the issue of Romanization is not put central here. However, the third chapter deals with the adoption and adjustment of Roman and indigenous iconographic elements of two statuette groups. The focus of this study is on a specific aspect of religion within the wider subject of its transformation in the Roman period: how people used statuettes in everyday life. I have chosen this approach because most find-spots of statuettes are located in or near settlements. Only a very small number come from sanctuaries. It is often stated in Dutch archaeological reports that statuettes were set up in household shrines. This practice is known from Pompeii, Herculaneum and Ostia.23 Evidence of household shrines in the Roman West is scarce and nearly absent in the study area.24 This suggests that, if statuettes were set up in houses, they were placed in very simple, wooden shrines that have not survived. Cupboards may also have been used to set up statuettes, and small tables.25 In the words of Jörg Rüpke, the home was, ‘without a doubt the most important location for individual religious practices, for the consumption of objects distinguished as instruments or reminders of religious communication because they represented gods and myths.’26 It seems, therefore, fruitful to regard practices with statuettes as an important part of everyday, lived religion. 20

Hoss 2015, 136-137.

21

Carroll 2012, 281-311.

22

See for the impossibility of establishing production or

Velsen, North Holland: Lange, 2021, 104, 256 and plate

deposition dates of statuettes this chapter.

LXXV, 331, f3367. A rare example from the western prov-

See for Pompeii, for instance: Krzyszowska 2002. For Hercu-

inces is a shrine with statuettes found in situ. It is in fact

laneum: Losansky 2015. Annemarie Kaufmann-Heinimann

a niche in the exterior wall of an early 2nd-century bath

has studied find complexes consisting of bronze statuettes,

complex at Rezé (Loire-Atlantique), France. The shrine

often buried together with other metal objects. She inter-

probably protected the people living in this quarter of the

prets the find-assemblages as the contents of household

town. Three shrines have been found in Germany: two

shrines and sanctuaries and argues that the assemblages

household shrines in Xanten and one in a villa in Neuen-

were buried because of impending danger, the so-called

ahr-Ahrweiler. Fragments of a limestone aedicula have

Angstdepots, as booty, or as metal hoards intended for

been recovered from a Claudio-Neronic layer at Augst,

recycling: Kaufmann-Heinimann 1998, 182-192. The focus of

Switzerland: Kaufmann Heinimann 1998, 96-97, 186, figs.

23

instance, Juv. 9.138; Ov. Fast. 6.310; Petron. Sat. 29.8. 24

her study is on Augusta Raurica (Augst, Switzerland), but she has also incorporated find-complexes from other sites

56-57 and note 642. 25

instance, Mols 1999, 36-39, 56, 60, note 307.

tion are Roman authors who have written about statuettes

4

We know from Pompeii and Herculaneum that cupboards may have functioned as household shrines. See, for

in Europe and northern Africa. Other sources of informaand rituals with statuettes in domestic contexts. See, for

A wooden fragment of a household shrine comes from

26

Rüpke 2018, 218.

1.3

Lived religion

The notion of lived religion has been introduced in the field by Rüpke, who in 2012 started a research programme on ancient religion drawing on this concept.27 It had its precursors in the 20th century, such as la religion vécue in France and Alltagsreligion in Germany. In 2008, Meredith McQuire’s Lived Religion. Faith and Practice in Everyday Life saw the light and put the study of lived religion on a new footing. She has stated: ‘Scholars of religion, especially sociologists, must re-examine their assumptions about individuals’ religious lives. What might we discover if, instead of looking at affiliation or organizational participation, we focused first on individuals, the experiences they consider most important, and the concrete practices that make up their personal religious experience and expression? What if we think of religion, at the individual level, as an ever-changing, multifaceted, or messy – even contradictory – amalgam of beliefs and practices that are not necessarily those religious institutions consider important?’28 Her work inspired Rüpke and other ancient historians. Instead of focusing on public or polis religion, cities and peoples, a spearhead of the project is the individual as religious agent.29 So, lived religion is what this study is about: exploring how Roman period statuettes functioned in the individual’s everyday life and what the meaning and purpose of these artefacts could have been.30 I consider the concept of lived religion not the opposite of public or polis religion or a conscious deviation from norms or rules that were imposed by Roman authorities, but an indispensable complement. In my opinion, lived religion in the Roman period includes all religious and ritual activities which extended beyond what religious organizations dictated. These ritual activities must not be seen as resistance or rebellion against Roman religious order, but as the expression of what a person considered necessary or desirable at a certain moment in his or her life, for the well-being or protection of him-/herself or others. Although statuettes of Roman deities were common religious symbols, selecting statuettes for one’s house was a private matter. Moreover, the statuettes were often combined with all kinds of objects that played a significant role in people’s lives and were set up together in rooms in the house. These combinations were very personal.31 The same may also apply to ritual activities involving the deposition of statuettes in and around settlements. These statuettes, too, were often buried together with other objects. Such combinations, however, were probably not always or not solely the result of individual choice and may have been based on traditions that remain hidden from us.32

1.4

The life path of statuettes

Objects can follow different life paths, which become salient when studying their cultural biography. The idea that things, like people, have different cultural biographies has been developed by the anthropologist Igor Kopytoff (1986). If we study a person, we can draw up a psychological, professional, economic, political or social biography by selecting or discarding certain aspects. The same applies to objects. A car has a technical biography which can be traced by studying its repair record. For its economic biography we can assemble data of the value from the beginning to the end of a car’s life cycle. Cars also offer social and cultural biographies. We can study relations between its owners, or the meanings and role of a car in different societies.33

27

See on lived religion also Rüpke 2018, 211-261.

31

Rüpke 2018, 219-221.

28

McQuire 2008, 4.

32

See for these deposits 4.2.2.

29

Rüpke 2012.

33

Kopytoff 1986, 66-68.

30

See for a discussion of the concept Naerebout 2013.

5

In general, cars follow more or less the same life path. After they have left the factory, they are distributed and sold to their first owner, who after some time sells it to the next, until the last owner brings it to the junkyard. Some cars, however, acquire special meanings and become icons, symbols for nations, groups of people or individuals. This happens to specimens of, for instance, the Citroen 2CV, Volkswagen Beetle and Aston Martin. Being collector’s items, they are kept in garages, cherished by generations in a family. A few have appeared in movies, or are on display in museums. Their biographies differ from those of most other cars and, like most collector’s items, they probably do not end up at the local junkyard. The studies by David Fontijn and Fokke Gerritsen have demonstrated the usefulness of Kopytoff’s theory for examining patterns of ritual deposits of weaponry and other objects in special places in the landscape and in settlements, from the prehistoric into the Roman period. Fontijn noted that ‘as objects may accumulate special meanings on their life path, selective deposition implies that the meanings themselves vary. Thus, there must have been different kinds of biographies’.34 With respect to the life cycle of houses, Gerritsen has stated that ‘biographies can also be written for prehistoric houses, and can provide insights into the cultural dynamics of house building, habitation and house abandonment’.35 The importance of these phases appears from ritual deposits in and around houses.36 Not all metal and terracotta statuettes began their life cycle as personifications of supernatural powers that would ensure prosperity, fertility, health or protection. There is evidence of statuettes being dolls and children’s toys, especially statuettes of animals. Some have movable body parts, or holes for a small wheel axle.37 Animal figurines were also used as cheaper replacements for animal sacrifices.38 Moreover, statuettes may also have had a purely decorative function, as an exotic object or souvenir, for instance. At the end of their life path, some were offered in a sanctuary, while others were placed as a gift in a grave.39 Like pottery, weaponry and bones, the majority ended up in the ground or in the water, either as refuse, as part of an offering, or as a result of another ritual. To the last two possibilities I will return at the end of the paragraph. First, I will briefly address the scarcity of statuettes in the study area. A dataset of nearly 700 items contrasts sharply with the millions of potsherds and other artefacts that have been recovered during excavations since the beginning of the 20th century. Undoubtedly, a substantial part of the metal specimens has been melted down for recycling. In our time, many metal objects have been retrieved from the archaeological record by detectorists.40 Yet, before the invention of the portable metal detector and a long time before organised excavations started, people already collected artefacts from the past, collections that have not been preserved or documented, or only partially. An illustrative example is the collection of Johannes Smetius (1590-1651), a minister at Nijmegen and a collector of antique objects.41 In Antiquitates Neomagenses (1678), his eponymous son states having over fifty figurines made of silver, bronze, ivory, stone and clay. He also mentions having over a hundred silver, bronze, marble and glass fragments of statuettes of male and female deities, emperors, human figures and animals.42 Unfortunately, his collection got dispersed after his death, and many items have disappeared.

34

Fontijn 2002, 26-27.

started an online database with finds from amateur archae-

35

Gerritsen 2003, 38.

ologists and detectorists. See for the website:

36

Gerritsen 2003, 63-66 and table 3.5.

https://www.portable-antiquities.nl/pan/#/public.

37

See, for instance, Willemsen 2003, figs. 79-80, 84, 86.

38

Kyll 1966, 64-66; Von Gonzenbach 1967, 8.

inscribed marble and stone objects, fragments and heads

39

See for toys in ritual deposits and graves Willemsen 2003,

of statues or busts, bronze and pottery vessels, but also

135-139.

numerous small objects like pieces of jewellery, coins,

Fortunately, more and more detectorists report their finds

gems, fibulae, amulets and oil lamps, statuettes and frag-

40

41

and the exact find-spot to provincial archaeological registration points. The Portable Antiquities of the Netherlands Project or PAN project of VU University Amsterdam has

6

Smetius’ collection comprised large objects like altars,

ments of statuettes: Nellissen 2004, 20-104, 108. 42

Nellissen 2004, 81-94.

Since the large majority of statuettes come from settlements and their surroundings, their presence raises some questions. Why would people leave statuettes behind when they abandoned their habitations? Bronze statuettes as well as complete and fragmented terracotta figurines have been recovered from pits and watery places in and around settlements. Had they been discarded because they were worn or broken? May be, but if the statuette had been an object of worship this is unlikely. In the case of bronze statuettes it is inconceivable that people would discard them as refuse, no matter what their function had been, because bronze objects retained their intrinsic value and could be recycled. When studying archaeological publications on settlement excavations I noted that, like ritual deposits of bronze objects, pottery vessels and bones, statuettes and statuette fragments have been recovered from features like pits, ditches and wells.43 It could be advanced that these features often contain potsherds, bones and other artefacts, since they are known to act as artefact traps, places where artefacts pile up through post-depositional processes. Therefore, these recurring find-spots may seem to present a pattern in itself, but from the artefacts that are caught in these traps only a random distribution pattern will emerge. The same applies to objects that were accidentally lost.44 Explanations such as ‘loss’ and ‘artefact traps’, are, therefore, not a credible solution for the presence of a single object or a striking assemblage of objects in these features, such as single bones of different animals, a single complete statuette or a statuette fragment, accompanied by a metal or ceramic object, a weapon, jewellery, coins or a combination of these objects.45 In the fourth chapter it will be examined which statuettes from settlements and the surrounding landscape may have been left behind as ritual deposits. Criteria have been formulated to recognise such deposits. Until now, an analytic study of deposits of statuettes in and around settlements has not been undertaken.

1.5

Distribution and use

Bronze and terracotta statuettes have been recovered from settlements, cemeteries and the surrounding landscape, generating catalogues focusing on provenance, distribution and iconography.46 Apart from a few studies, the function and meaning of these figurines have seldom been questioned or analysed in relation to their find-spots.47 Furthermore, studies on statuettes usually focus on either bronze or clay figurines, as if they were two separate categories of artefacts, which, in my opinion, they are not.48 It is an artificial distinction that should not be made in an analytic study on distribution, function and iconography. Moreover, it provides a distorted picture of people’s preferences for certain deities, since it appears from this study that male deities are usually depicted in bronze, while female figures are predominantly depicted in clay. Therefore, the dataset of this study comprises all Roman period bronze and terracotta statuettes and identifiable statuette fragments from established find-spots in the Netherlands.49 43

See, for instance, Merrifield 1987.

44

Fontijn 2002, 37.

45

Besides deposits of complete objects, inside and outside

der-Feytmans 1979; Menzel 1960, 1966, 1986; Manfrini-Aragno 1987. 47

the study area there is also evidence of deposits compris-

Belgium De Beenhouwer 2005, 823-841; for bronze stat-

ing fragmented or destroyed artefacts, selected bones,

uettes from Roman period graves Bolla 2013; for infant

intentionally broken pottery vessels and ‘killed’ or bent metal artefacts: Cunliffe 1993, 12-13; Pleiner 1993, 161 and

figurines Derks 2014a. 48

houwer (terracotta statuettes) and Bolla (bronze statuettes).

Nieuwhof 2015, 128-148, 176, 190-193. See, for instance, Zadoks et al.1967, 1969, 1973; Zadoks/Ger-

See, for example, the studies by Van Boekel (terracotta statuettes), Kaufmann-Heinimann (bronze statuettes), De Been-

fig. 19; Clarke 1999, 24; Groot 2008, 117-133; Groot 2009; 46

See for a functional analysis of terracotta figurines from

49

Although not all find-spots are exact and sometimes based

hartl-Witteveen 1983; Boucher 1970, 1971, 1973, 1976, 1983;

on second-hand information, for instance in the case of

Boucher/Tassinari 1976; Boucher/Boucher 1988; Boucher/

several detector finds, statuettes from the terp region and

Oggiano-Bitar 1993; Van Boekel 1987, 1989, 1993, 1996; Fai-

specimens in museum collections.

7

The objective of this study is to gain more insight in the use of Roman period statuettes through a comparative analysis of find-spots, iconography and ritual behaviour in the study area, the modern Netherlands, during the Roman period. In its second chapter a short account of the development of each area is given, followed by a description of find-spots of statuettes and statuette fragments per area. In the third chapter a spatial and iconographic analysis is made of statuettes, identifiable statuette fragments and their find-spots in three areas in the Netherlands: the area north of the limes, the limes zone and the area south of the limes zone, in order to find answers to the following two research questions: 1 Can we observe relations between size, material and subject of statuettes and their find-spots in the three areas and if so, how can we explain these relations? 2 Were there preferences for specific subjects or iconographies in the three areas and if so, how can we explain these preferences? The fourth chapter focuses on rituals with statuettes. These rituals raise all kinds of questions, many of which cannot be answered. Therefore, I will focus on two questions which possibly can be answered from the data presented in this study: 3 How can we identify ritual deposits of statuettes? 4 Are there any indications for the ritual deposition of statuette fragments?

1.6

The dataset

Controlled excavations, construction works, stray finds by amateur archaeologists and coincidental finds in the Netherlands have yielded nearly seven hundred bronze and terracotta statuettes and statuette fragments. They have been divided into three main categories: deities, human figures and animals. The dataset consists of published and unpublished statuettes. The first group comprises specimens from catalogues, museum collections, archaeological studies and reports. The second group includes hitherto unpublished specimens from excavations, archaeological databases, stray finds and statuettes belonging to private collections.50 All items have been sorted by subject and find-spot and fed into a database. Although a substantial number of figurines have been found during controlled excavations, the exact find-spots of an equally substantial number have not been documented in detail. Nevertheless, in many cases it was possible to establish their find-spot in a broader sense. Statuettes of unknown provenance have not been included, but I have made a few exceptions in the case of an approximate provenance, by which I mean that the city or province is known, but not the exact find-spot. Only figurines have been incorporated that were not part of a utilitarian object. Therefore, moulds and figurative appliques on vessels, cutlery, furniture, and other objects have been excluded from this study.51 I have made this distinction in order to demarcate the dataset from objects with a practical function. The objection could be raised that also utilitarian objects like drinking vessels at some point could have a ritual function, but without a clear find-context this function usually escapes us. Furthermore, we do not know if an object with a decoration in the shape of a deity was bought because of this deity, or because the buyer simply needed that specific object and took the decoration for granted. 50

Catalogues of bronze statuettes: Zadoks et al.1967, 1969, 1973;

Including objects that may have had a ritual function, like

Zadoks/Gerhartl-Witteveen 1983. Catalogues of terracotta

a bronze statuette of Amor holding a hollow poppy, found

statuettes: Van Boekel 1987, 1989, 1993, 1996. I owe a great

in Limburg. The statuette probably was an incense burner:

deal to the authors of these catalogues, whose expertise and

Limburgs Museum, inv. nr. L02988.

accurate work have been indispensable for this study.

8

51

A few items constitute a ‘grey area’. The first group comprises figurines without traces of a fitting or a hole for a fitting. Still, they could have been part of a utilitarian object, because comparable figurines have fittings or holes, indicating that they were once part of a drinking vessel, jar, support, and so on. The second group consists of a few figurines which originally were attached to a large object such as a piece of furniture. They have been recovered from specific find-contexts and will be discussed in chapter 4. To conclude this section, it should be noted that the results of this study are based on small numbers. Therefore, I will not draw generalising conclusions based exclusively on numerical dominance. Besides ratio, only a comparison of the results with those from other studies and areas can provide meaningful insights into, for instance, the preferences of people for specific representations or ritual practices.

1.7

The areas north and south of the Rhine and the Dutch part of the Roman limes zone

The study area, the modern Netherlands, comprises three geographical zones: the areas north and south of the Rhine and the limes zone. It is commoner nowadays to define regions ecologically or in relation to the political geography of the period. For this study, one would expect an investigation of, for instance, all statuettes found in Germania Inferior. Nevertheless, I have confined the study area to the modern Netherlands. I have done this for two reasons. The first is the accessibility of finds. Terracotta statuettes usually have been preserved only fragmentarily, the fragments ending up in carton boxes on shelves in archaeological depots. For various reasons, it has proven to be very time-consuming to gather all data.52 The second reason is the political division of the study area in the Roman period, the northern region being part of ‘Germania Libera’ after 47 and the southern half initially belonging to Gallia Belgica and from around 84 to Germania Inferior (see 1.1). To what extent the political division of the study area in 47 had consequences for the availability, distribution and choice of statuettes, will be explored in the third chapter. Therefore, I distinguish between the areas north and south of the Rhine. Figurines have been found in settlements north and south of the Rhine, as well as in and around military settlements in the limes zone along the Rhine and North Sea coast. The third reason for the division into these three areas are the theories and ideas that have been advanced to account for different developments in urban, rural and military environments.53 Starting with the region north of the Rhine, this is a rural region where Roman period settlements with urban characteristics are absent. The Roman period graves that came to light in this area did not contain any statuettes.54 The only Roman forts in the north are the two forts at Velsen (North Holland). These forts will be discussed together with the military settlements in the limes zone. South of the Rhine, numerous Roman period settlements have come to light through excavations, ground surveys and geo-archaeology.55 These settlements could consist of one to several farmhouses and annexes. Farmsteads with a stone built main house are often referred to in literature as villas. They show a wide variety in layout and size.56 52

53

54

I would like to stress that this has nothing to do with the

early Medieval cemetery, dating between the 4th and 8th

staff of archaeological depots and museums. They have

centuries. In or before 1834, a Roman period statuette was

been very cooperative and helpful.

found here. See for Rhenen 2.3.3. The mother goddess fig-

See, for instance, Roymans 1995; Roymans/Derks 2011. See

urine is a stray find. Since a small number of the earliest

for Britain Mattingly 2006. See for military contexts, for

graves contained Roman coins and terra sigillata vessels,

instance, Haynes 1993, 1997, 1999, 2013.

the figurine may have come from one of these graves:

The only exception of a cemetery north of the Rhine that

Huiskes 2011, 19-26, 41-48.

has yielded a Roman period statuette is Rhenen (Utrecht).

55

See, for the latter, for instance, Jeneson 2013.

Excavations in 1951 at the Donderberg site produced an

56

See for a definition of a villa 2.3.1.

9

The four largest settlements possess urban characteristics: Nijmegen, Voorburg, Maastricht and Heerlen. It is beyond any doubt that there were cemeteries belonging to these towns but so far, they have only come to light at Nijmegen and Heerlen. Nijmegen and Voorburg became the civitas capitals of the Batavi and Cananefates: Ulpia Noviomagus, the settlement that was founded after the destruction of Oppidum Batavorum and Municipium Aelium Cananefatium, nowadays part of modern Voorburg. Nijmegen and Voorburg meet most of the criteria for urban communities as defined by Jürgen Kunow.57 Maastricht and Heerlen were vici that developed into successful small towns. Since only small parts of these settlements have been excavated, it is uncertain to what degree they meet Kunow’s criteria in the Roman period. We can only assume that all four settlements possessed the characteristics that are considered manifestations of Roman urbanism: the presence of temples, a forum, a porticus, stone monuments, baths, a defensive wall, and evidence of the use of the Latin language. Temples have only been found at Nijmegen and Maastricht, whereas basilicae or theatres have not been found in any of the four settlements. The only characteristics that have been attested with certainty in all four settlements are houses, workshops and public baths. Still, there is no doubt about their function as regional, economic centres located on important routes over land and water. Several definitions have been given for the Dutch part of the Roman limes. Since a precise demarcation of this zone to the south cannot be given, the following definition will be employed in this study: the military installations and immediately associated features (extramural settlements, cemeteries) on the river deposits on the left bank of the river Rhine, from the German border in the east to the North Sea in the west, the forts at Velsen (North Holland) and those along the North Sea coast between Katwijk (South Holland) and Aardenburg (Zeeland). The research period spans from around 50 BCE, the time of the first Roman military campaigns in this region, until the 5th century, when statuettes disappeared from the living context.

1.8

Definition of terms

• Statuette ‘Statuette’ in this study is used for all small-size freestanding representations of deities, human figures and animals. They are made predominantly of metal and clay, a few from other materials like stone, amber or chalk. The dataset also includes busts and one herm, which is a bust on a shaft. The height of the statuettes varies between 50 and 320 mm. • Bronze Apart from one silver and a few lead specimens, all metal figurines from the dataset have been termed ‘bronze’. Yet, in this case bronze is an umbrella term. Strictly speaking, bronze is a copper-tin alloy, with trace elements like arsenic, silver or bismuth. Messing is a copper-zinc alloy with characteristics that differ substantially from copper-tin alloys. Lead could be added to a copper-tin alloy, for instance to lower the melting point. Roman bronze often contains a percentage of lead between ten and twenty, but this percentage can be much higher. A copper-tin alloy with more than ten percent lead is called ‘leaded tin bronze’. By adding zinc, the product is easier to work after casting.58 The large majority of the metal statuettes are made from bronze alloys. Since the difference between these alloys is seldom visible to the naked eye, I use the term ‘bronze’ for all figurines made from copper-tin and copper-zinc alloys.

57

Kunow 1992, 143-142. See for the urban character of Ulpia Noviomagus also Driessen 2007, 148-164.

58

10

Riederer 1987, 116; 2000, 575-583.

• Terracotta The colour of unpainted clay figurines after firing varies from white to orange-red and brown. The colour of the clay is the result of its composition, the iron and calcium content in particular, and the conditions during firing.59 White clay figurines are often termed pipe-clay figurines because in early modern and modern times white clay was used for the manufacturing of smoking pipes. It is, therefore, a misconception that terracotta figurines are always orange-red or brown. The term simply means ‘fired clay’, which is why I in this study refer to all fired clay statuettes with the term ‘terracotta’. For this study, I have not examined the composition of the clay or metal alloys used, the type of moulds or the manufacture techniques of figurines that hitherto were unpublished. These data are not relevant to the research questions formulated in the previous section. • Find-spot The find-spot is the place where a statuette or statuette fragment has been found, either in the area north of the limes, south of the limes or in the limes zone. The description of find-spots in literature varies from indefinite indications such as ‘near settlement x’ to more precise descriptions such as ‘in the southwest corner of building y’. Examples of categories of find-spots are: in or near a military fort, settlement, building, cemetery, river, stream or other watery place. Special features in the landscape that existed in antiquity could be considered as find-spots with particular characteristics, but they are often hard to recognise today. The same applies to watery places which now have disappeared or dried up. Therefore, these find-spots do not fall into the category find-context (see hereafter). • Find-context In this study, the find-context is a find-spot that distinguishes itself from other find-spots through specific characteristics. These characteristics are recognisable in the archaeological record and, together, constitute different types of archaeological contexts. The find-contexts in this study comprise graves, pits, wells, cisterns, drains, ditches and post holes. According to this definition, a find-spot can contain a varying number of find-contexts. • Votive deposit Richard Bradley remarks at the beginning of his book A Geography of Offerings that all terms relating to hoards and deposits give problems and that the term votive deposit ‘was treated as a residual category made up of collections of objects whose composition resisted a practical interpretation’.’60 Those who study the ancient Mediterranean world, and the Graeco-Roman period in particular, point out that the meaning of the noun ‘votive’, is derived from the Latin verb vovere, which means ‘to promise solemnly or sacredly, to devote, dedicate, to consecrate something to a deity’. In a paper about the definition of ritual and religious concepts, Andreas Murgan has pointed out that the terms in modern archaeological literature are loosely used, as if they are interchangeable, without considering the complexity of original and actual meanings hiding behind notions such as bothros, favissa, stipe, votive deposit and so on.61 According to Murgan, a correct use of the word votive with all its compounds can only be made when there is a relation between offering and vow.62 In 1996, Jelle Bouma dedicated a chapter in his PhD thesis to the same problem, citing Tony Hackens who had studied the original meanings and use of these terms in different contexts.63 By combining literary and archaeological evidence, Hackens arrived at a typology of containers and depositions of votive material.64 59

See for a detailed explanation of the manufacture process

61

of terracotta statuettes and characteristics of workshops

60

Murgan 2016. I thank Andreas Murgan for providing the text of his paper.

in Cologne, Central-Gaul and the Rhine-Moselle area Van

62

Murgan 2016, 89-90. See on the ritual of the votum also 4.2.6.

Boekel 1987; De Beenhouwer 2005, 2014.

63

Bouma 1996, 43-50.

Bradley 2017, 1.

64

Hackens 1963, 84-97; Bouma 1996, 51 and note 149.

11

Bouma concludes from Hackens’ typology that no words in antiquity are known that distinguish between open and closed deposits, or between deposits made in- and outside sanctuaries. Neither are there specific words in antiquity to distinguish between deposits made in a sanctuary and deposits made in watery places, caves, marshes or other features in the landscape.65 Offerings were gifts for a deity or deities, which could be the deity or deities to which a sanctuary was dedicated. Offerings made in settlements could be gifts for a supernatural power, too, although this might not always have been the case with respect to closing or abandonment rituals. Peter van den Broeke prefers the more neutral term abandonment deposit over offerings to describe ritual deposits in pits filled with potsherds and other objects in association with houses. He argues that, although the pits were dug as part of a ritual, it is doubtful that their contents must be seen as gifts to a supernatural power.66 Considering the difficulties related to the term ‘votive’, I will avoid the terms votive deposit and ex-voto. Instead, I will name all deposits with an alleged ritual meaning ritual deposits, sometimes alternated by the terms special and specialised deposit, to avoid annoying repetition. • Ritual Before defining what is meant by a ritual deposit in this study, it is necessary to define ritual. Although many scholars have theorized about the nature of rituals, there is no all-comprising definition for what ritual exactly is. In Day of Shining Red: An Essay on Understanding Ritual (1980), the British anthropologist Gilbert Lewis argues that ritual and art share several similarities. Both lack a commonly agreed definition, and with both there are performers and performances, interpreters and beholders. The lack of a definition seems not to bother Lewis, since he claims that anthropologists often intuitively recognise a ritual, even without understanding its meaning or the symbols involved. He points out that understanding and interpreting symbols or behaviour is a skill we can learn, just as we can learn to interpret the tracks of animals in the snow.67 My objection to Lewis’ view is that anthropologists as well as archaeologists must provide arguments in support of their claim that a ritual is involved. Before we are able to recognise the track pattern of a mountain lion in the snow, we must learn how to identify the tracks of the mountain lion in order to distinguish them from those of other mammals. It has nothing to do with intuition, but everything with knowledge and definition. The American religious studies scholar Catherine Bell is the author of two influential works on ritual, Ritual Theory, Ritual Practice (1992) and Ritual: perspective and dimensions (2009). According to Bell, rituals in general share one or more of the following features: formalism, traditionalism, invariance, rule-governance, sacral symbolism and performance. Unfortunately, the archaeologist might only see a few elements of past rituals: the objects that have been used at the time the ritual was performed, sometimes accompanied by the remains of offerings or traces of fire, and the ritual’s ambience. Other elements of the ritual usually remain invisible: the persons involved in the ritual, the gestures, the accompanying words, the choreography, and so on. Rituals are based on tradition and custom and are often a repetition of activities from the past.68 In most cases, the archaeologist is only able to recognise rituals in the archaeological record if they are performed in the same way over a long period of time. This does not imply that the intentions, meanings and objects involved in the ritual have remained unchanged. Recognising rituals with ‘new’ objects is possible by comparing such ritual deposits to the characteristics of traditional ones, which is the central topic of the fourth chapter. The last feature Bell has formulated is sacral symbolism and involves the use of special objects when an appeal is made to supernatural powers. The appeal can be direct, by addressing them orally, or more indirect, by written dedications. The objects symbolise the appeal to or the communication with the supernatural, and include imagery and objects such as amulets.69 65

Bouma 1996, 51.

68

Bell 1997, 145-150.

66

Van den Broeke 2015, 89.

69

Bell 1997, 153-154.

67

Lewis 1980, 6-9.

12

• Ritual deposit Establishing definitions or criteria is necessary to create a point of departure. An example of the necessity to establish criteria for the study of societal phenomena is dancing. How do we define dance? How do we distinguish dance movements from everyday movements? The main difference is the way in which they are used. According to some definitions, dance movements are a form of non-verbal communication. Communication implies a form of interaction between the dancer(s) and the audience. The movements are meaningful to the audience, which also implies that we have to understand the societal context in which dance is being performed.70 The criteria in this section are based on patterns and characteristics that have emerged from previous studies on ritual deposition of pottery, bones and weaponry from the Bronze Age into the Medieval period.71 To make a reasonable cause for the distinction between ritual and secular deposits of statuettes and statuette fragments, I will take into account the composition of deposits, the condition of the objects involved and the observed deposition patterns in other studies. If a deposit of a statuette or a statuette fragment occurs repeatedly in recurring find-contexts, alone or together with other, selected artefacts, it could be intentionally deposited as part of a ritual. Thus the key words are ‘patterning’ and ‘repetition’. If a deposit contains a single statuette fragment, only fragments have been incorporated which depict an important and easily recognisable part of the statuette, which is why I have incorporated only heads and heads with the upper part of the torso. Patterning and repetition can be detected if a deposit meets at least two of the following criteria: one of the first two (1 or 2) and one of the last two (3 or 4). 1. A deposit includes at least one complete or substantially complete statuette, sometimes accompanied by a striking assemblage of other objects, or: 2. A deposit includes a statuette head/head-torso fragment, sometimes accompanied by a striking assemblage of other objects. Other fragments of the statuette are absent. 3. A statuette or statuette head/head-torso fragment is buried in or next to a building in a settlement: under a floor, demolition layer, under or near the entrance, in a corner, posthole, or: 4. A statuette or statuette head/head-torso fragment is deposited in a ditch, pit, well, drain or cistern associated with a building or settlement. Deposits with the abovementioned characteristics are considered ritual deposits, once repetition and patterning have been recognised. • Sanctuaries and alleged cult places Sanctuaries are the first place where one would expect to find statuettes. Yet, it appears from this study that less than six percent of all statuettes come from sanctuaries and alleged cult places. For the Roman period, the French historian and archaeologist John Scheid distinguishes two types of sacred places. The first category is often recognisable in the archaeological record: man-made sanctuaries, usually with a specific layout, ranging from simple religious precincts with an altar to impressive temple complexes with colonnades and secondary buildings. The second category consists of natural places that were considered to be residences the gods had chosen for themselves: groves, mountain tops, large caves, springs and deep pools.72 Of course, these natural features were the result of human choice as much as the first category and in some cases, they were turned into formal sanctuaries, like the sanctuary of Sulis-Minerva at Bath, Britain and the cult place of Sources de la Seine, Fontes Sequanae, north-west of Dijon, France. Since architectural remains are lacking, only the combination of a striking feature in the landscape and an accumulation of specific finds such as figurines can reveal the existence of sites belonging to the second category. 70

Naerebout 1997, 161-166.

71

See, for instance, Merrifield 1987; Fontijn 2002; Gerritsen

72

Scheid 2003, 63-64, 73-74.

2003; Therkorn 2004; Groot 2009; Nieuwhof 2015.

13

Sanctuaries with architectural remains and a Gallo-Roman ground plan come from Empel (North Brabant), Elst and Nijmegen-Maasplein (Gelderland).73 The first has yielded one bronze statuette, a stray find. The sanctuary at Elst has produced one bronze statuette fragment and no statuettes have been recovered from the temple precinct at Nijmegen-Maasplein.74 In Maastricht (Limburg) the northwestern corner of a temple has been discovered. Despite the impressive architectural remains that have come to light, the lay-out of this temple has to remain largely speculative, since the site is located in Maastricht’s city centre. The site has not yielded any statuettes, only a bronze applique of a male head.75 A fourth sanctuary with architectural remains stood at Buchten (Limburg).76 Just outside the sanctuary’s precinct, but near the enclosure wall, a pit was found with a bronze statuette of a cockerel and several other small metal objects. A sanctuary of which only traces of the stone foundations have been preserved stood on the premises of the Aardenburg castellum (Zeeland). On the terrain of this small, square building a fragment of a terracotta statuette of one of the Parcae came to light. Just outside the precinct a small, sandstone votive altar has been recovered.77 The existence of two sanctuaries at Colijnsplaat and Domburg (Zeeland) is inferred from roof tiles, architectural fragments, stone altars and statues that have been recovered from the sea and the beach. A bronze statuette and a fragment of a terracotta statuette may have come from these two sanctuaries. At Wijchen-Tienakker (Gelderland) a post-built structure in the vicinity of a villa may have been an openair sanctuary. This interpretation is based on the square lay-out of the structure and the discovery of a bronze statuette and a small tuff stone altar in a pit within the structure.78 Three sites in the study area could fall into Scheid’s second category, but this is uncertain. The first was situated in the vicinity of a military fort: The Hague-Scheveningseweg ( South Holland).79 The site has yielded hundreds of fragments of at least twenty terracotta statuettes. The identification of the site as a possible cult place is based on the number of statuettes and a few pits with a striking content. Excavations at Bergen op Zoom (North Brabant) have uncovered a site where hundreds of miniature vessels and at least one terracotta statuette were deposited in a fen.80 North of the Rhine, at Noordbarge (Drenthe), some twenty objects, mostly bronze and terracotta statuette fragments, were recovered from elevated terrain near a native Roman period settlement. At Hoogeloon-Kerkakkers (North-Brabant), a square terrain with an enclosure ditch, two rows of deep postholes and several pits has been interpreted as traces of a sanctuary. A terracotta statuette fragment was recovered from the enclosure ditch. What becomes clear from this overview is that the identification of these sites as cult places is largely or solely based on finds, which could result in circular reasoning. Nevertheless, the fen at Bergen op Zoom and the elevated terrain at Noordbarge constituted a distinct feature in the landscape in the Roman period and, therefore, provide an additional argument for the interpretation ‘cult place’ as defined by Scheid. The miniature vessels at Bergen op Zoom came to light during controlled excavations, but the Noordbarge statuettes were found in the 1860s by people digging for loam, which is why the presence of a cult place at Noordbarge must remain hypothetical. Two conclusions can be drawn: very few statuettes have been recovered from sanctuaries and the identification of several sites as open-air sanctuaries or cult places is uncertain. 73

See for a definition of Gallo-Roman temples and variations of the scheme, for instance, Derks 1998, 146-152.

74

Like villa sites, finds from these sanctuaries may have been

See for the ground plan of the temple Panhuysen 1996, 46-49 and figs. 11-12. See for the temple 2.3.5.

removed by collectors of antiquities. Alleged cult places on

76

Derks/De Fraiture 2015.

the sites The Hague-Scheveningseweg and Noordbarge have

77

See for Aardenburg 2.4.2.

yielded a striking number of statuettes, probably because

78

See for Wijchen 2.3.4.

they lack architectural remains that could be recognised by

79

See for The Hague-Scheveningseweg 2.4.2.

treasure hunters. The same applies to the hundreds of altars

80

The number of deposited statuettes is unknown, since

from the sanctuaries at Colijnsplaat and Domburg, which

14

have been preserved because they were hidden under water. 75

mainly tiny fragments have been found: see 2.3.5.

• Military or civil? The terms ‘military’ and ‘civil’ are essential in this study, since a distinction is made between military and non-military communities. It has been argued that before the Roman conquest a division between soldiers and civilians not really existed. Warriorship was a key value in Celtic-Germanic societies where farmers had to leave their farms and cattle to fight in times of war and then turned back to farming again.81 Therefore, the distinction between civilian and soldier that was introduced in the Roman period had a great impact on local communities. The Roman camps along the Rhine attracted craftsmen, traders and lots of other people who could make a living out of the army. Around these military forts and fortresses, canabae legionis and vici developed: settlements where people lived who provided services for the army. Research in the last decades has demonstrated that military forts could house not only soldiers, but also their families, servants and tradesmen that depended on them, while soldiers could live with their families in the associated canabae and vici.82 Data from recent excavations obscure the distinction even more. Evidence from the earliest Roman civil settlement in the Netherlands, Oppidum Batavorum (ca. 10 BCE- 69/70), suggests that the settlement comprised at least one military building.83 At Ulpia Noviomagus, Nijmegen, stamped building material indicates that the Roman army was involved in the construction of the town. It is also argued that the capital of the civitas Batavorum, may have housed small military units after the Tenth Legion had left the fortress on the Hunerberg.84 At Heerlen, too, archaeological evidence testifies to the presence of veterans and Romans.85 At Voorburg, too, the involvement of the Roman army in construction works is suggested by military stamps on building material. Also, militaria and inscriptions by soldiers testify to a military presence at Voorburg.86 According to Ian Haynes, ‘Even in provinces with a substantial number of forts and fortresses, there are good reasons to believe that soldiers were stationed in towns.’87 However, little is known about how many civilians lived in forts and how many soldiers lived with their families in canabae and vici. Moreover, it remains uncertain whether soldiers were stationed in the urban communities in the study area, let alone how many. This implies that a rigid distinction between civil and military in settlements cannot be made. As a consequence of the strict definition of the limes zone, military communities in this study are the castra and forts, as well as the associated canabae legionis and vici. Vici and urban settlements outside the limes zone are regarded as civil communities, even though the presence of soldiers and returning veterans in some of these settlements is suggested by the use of military building techniques and/or militaria or inscriptions. • Vicus It results from the distinction between military and civil that the term vicus is confusing. Besides the vici that developed near military camps, there were also civil vici. After all, the word vicus itself means no more than ‘dwelling place’. Some of these vici must be considered as regional centres within civitates.88 They were smaller than civitas capitals, but housed small production centres for the regional market and provided facilities like public baths and sanctuaries. Examples of civil vici are Maastricht, Heerlen (Limburg), Cuijk (North Brabant) and, possibly, Venlo (Limburg).89 The latter could be an example of a civil vicus developing from a military vicus. The presence of a small Augustan military post was suspected at Venlo because of its strategic location, but traces of the fort itself have not been found, only finds that point at a military presence between 19 BCE and 9.90 81

82

Nicolay 2007, 10-11, 237-244, with reference to further liter-

86

De Bruin 2019, 139-141.

ature.

87

Haynes 2013, 155 and note 44.

Van Driel-Murray 1995, 3-21; 2008, 82-86; Haynes 2013, 14-17,

88

Van Es 1994, 55.

157-160.

89

The presence of an early Roman fort at Cuijk is doubtful:

83

Van Enckevort/Heirbaut 2010, 85-91.

84

Nicolay 2007, 192.

85

Jeneson/Vos 2020, 159-160.

Haalebos et al. 2002, 23-24. See for the Augustan military post and vicus at Venlo: Van der Velde et al. 2009, 635-642. 90

Van der Velde et al. 2009, 640-641.

15

In a few cases the distinction between military and civil vici is less evident. At the site The Hague-Scheveningseweg, the second habitation phase of an initially civil settlement has yielded a large quantity of imported goods, including militaria and casting moulds for militaria such as armour fragments, swords, arrow and spear heads.91 Based on similarities with The Hague-Ockenburgh, De Bruin has suggested that the settlement was used by the military between 190 and 250.92 At Naaldwijk-Hoogwerf (South Holland), radical changes are visible in the layout and organisation of the settlement at the end of the 2nd century. A strong increase in imported pottery, Roman coins and inscription fragments points at intensified contacts with Romans, even though the quantity of imported pottery and militaria is smaller than in military vici.93 Moreover, many stamps on building material and a fragmentary bronze plate mention the Classis Germanica. Bronze fragments of a large inscription plate, as well as bronze fragments of statues of emperors stood at Naaldwijk, which would be less likely in a civil vicus.94 At Ouddorp (South Holland), the use of Roman building techniques suggests a military presence, but it is not clear to what extent the army stimulated the building of the settlement. Research has demonstrated that Ouddorp and its harbour had an important regional function, but the quantity of military finds from this settlement is not as striking as at Naaldwijk-Hoogwerf and The Hague-Scheveningseweg.95 No Roman military base has been found near Ouddorp, Naaldwijk-Hoogwerf and The Hague-Scheveningseweg (South Holland), which implies that they should fall outside the limes zone as defined above. However, in the case of Naaldwijk-Hoogwerf and The Hague-Scheveningseweg, the nature and quantity of finds strongly point at the presence of a military base nearby, which also changed these settlements after the first habitation phase. It, is therefore, fruitful to (re)assess the find-assemblages to gain more insight in the nature of such settlements and their inhabitants. In this study, Naaldwijk-Hoogwerf and The Hague-Scheveningseweg are regarded as military settlements, whereas Ouddorp is considered a civil vicus. • Production date and deposition date In the course of my research, I have noticed that in many publications no clear distinction is made between the production and deposition date, in which case the given date can be confusing. The deposition date is the moment or period in which a statuette somehow ended up in the ground or in the water. A precise production date of bronze statuettes cannot be determined. This is a consequence of the impossibility to establish an absolute production date through technical analysis. Moreover, there was no standardized process for the production of metal statuettes. They were cast hollow or solid, often with the help of a wax working model, with or without negative moulds.96 Since there is no evidence for workshops specialized in the production of bronze statuettes in the Roman West, we can only establish a date for their deposition, which is often based on the occupation period of the settlement where the statuette was found.97 For example, Ulpia Noviomagus was inhabited between 70 and 270. This provides a terminus post quem and a terminus ante quem for the deposition of artefacts. If a statuette is found together with dateable objects like coins, of which the latest was struck in 150, the deposition date of the statuette probably lies between 150 and 270. Unfortunately, such detailed find circumstances rarely occur. This implies that in most cases deposition dates can only be based on the settlement’s history, which does not add much value to our understanding of the use of statuettes. And lastly, it cannot be ruled out that some statuettes ended up in the ground after 270.98

91

92

Magendans/Waasdorp 1986, 297-300; Waasdorp 1999, 171;

statuettes Kaufmann-Heinimann 1998, 16-20; Veen 2014a,

Mijle Meijer 2011, 8.

137-138 and 3.13.1. See for a discussion on serial production:

De Bruin 2019, 103-105.

93

See for Naaldwijk-Hoogwerf 2.4.2.

94

De Bruin 2019, 105-108.

95

See for Ouddorp De Bruin 2012 and 2.3.2.

96

See for the production technique of bronze statues and

16

Kaufmann-Heinimann 1998, 18-20. 97

Bronze statuettes in the Roman West are often dated to the first two or three centuries.

98

Before 70 is also possible, but less likely.

In the case of terracotta figurines, we sometimes have a rough production date. A few modellers have signed their work. Their signature is sometimes accompanied by a consul name or the place where they were active, in which case we know in which period and where the figurine was made. Moreover, several figurines without a signature can be ascribed to a certain modeller.99 The modeller Servandus, for instance, was active in Cologne between around 150 and 180. Then the terracotta production in Cologne stopped, maybe because of an epidemic, or as a result of competition from other firms.100 This means that a figurine signed by Servandus found in Ulpia Noviomagus probably has a deposition date between 150 and 270.101

1.9

Structure of the text

This study comprises five chapters, this being the first. The second chapter is a catalogue, with the find-spots of statuettes described per area, province and municipality. The description includes the most relevant archaeological data of each find-spot and find-context, serving as the starting point for the spatial and iconographic analysis in the third chapter. The spatial analysis aims at exploring to what degree the presumed distinctiveness of the three areas is reflected by the distribution of statuettes in each area. Further, it will be examined whether there are relations between find-spot, size, material, subject and function. The paragraphs about specific iconographies of statuette groups focus on the impact of interpretatio and creolisation. Also, striking iconographies of single statuettes will be discussed. The fourth chapter is dedicated to the identification of ritual deposits of bronze and terracotta statuettes. The first inventory consists of alleged ritual deposits of statuettes in the study area. A second inventory comprises possible deposits of statuettes in Britain, in order to explore whether there are recurring depositional patterns in the two areas, which together constituted the northwestern frontier of the Roman Empire. The second issue that will be explored is the deliberate fragmentation of statuettes and deposition of statuette fragments from the Neolithic onward. The third subject that will be addressed concerns a special type of statuette deposits related to so-called magical practices. The fifth and final chapter summarises the conclusions from the four chapters and includes suggestions for further research. The Greek and Latin citations as well as the quoted translations in this study come from the Loeb Classical Library Online. For the cited Latin inscriptions I follow the EDCS online database, with references to the CIL, AE and RIB numbers, or, if not available, the EDCS number. See for the list of abbreviations page 213.

99

See, for instance, Van Boekel 1987, 207-213; De Beenhouwer

100

Van Boekel 1987, 210.

2014. 22-23.

101

See also De Beenhouwer 2005, 30.

17

2

Catalogue and description of find-spots

2.1

Introduction

The description of find-spots is preceded by a brief overview of the transformation of each area after the Roman conquest, beginning with the area north of the Roman limes, followed by the limes zone and the area south of it. An extensive historical overview of the Roman Netherlands can be found in several standard works.102 Although publications of single settlements, military forts and burial sites have been important for this study, they do not allow for many conclusions concerning the development of each area as a whole. The western North Sea coast, for example, north and south of the Rhine, has not been the subject of any synthesising study yet.103 The same applies to burial customs in the study area.104 The Lower German limes zone, stretching between the North Sea and modern Remagen in Germany, was on the tentative list of UNESCO’s World Heritage Sites for some time and has been added to the World Heritage list in 2021.105 The scarce settlements and sites that have been extensively excavated and published may present a distorted picture of historical reality. As a consequence, this chapter reflects the unbalanced distribution of the available data. Preceded by an introduction to each area, the find-spots are presented per province, in alphabetical order. Each section ends with a table of find-spots without archaeological data and a table of uncertain find-spots.

2.2

The area north of the Rhine

2.2.1

Imported goods: evidence of trade or diplomatic gifts?

The area north of the Rhine consists of the modern provinces of North Holland, Friesland, Groningen, Drenthe, Overijssel, Flevoland and the northern part of Gelderland.106 The Roman military campaigns by Drusus and Germanicus in the northern Netherlands, between 12 BCE and 16, have left little trace in the archaeological record. The evidence consists mainly of coins. Wim van Es has suggested that early coins found in southeast Friesland, northwest Overijssel and Drenthe may be indicative of Roman military troops using an overland route. This hypothetical route ran through the modern province of Drenthe, connecting the nav102

103

Although an update would be desirable. Byvanck 1943; Van

105

Es 1981. See for the province of Friesland Boeles 1951.

Kooistra/Polak 2013; Graafstal et al. 2018; Verhagen et al.

An exhibition on the Roman North Sea coast and its

2019. Lampas issued a special number in June 2020 on the

defences was held in 2016 at the National Museum of Antiquities, Leiden. See for the south-western coastal area

104

See for studies on the Dutch limes zone, for instance,

subject: Mols/Polak 2020. 106

The area where the province of Flevoland is located was

De Bruin 2019. Michael Erdrich’s study provides a broad

part of an inland sea during the Roman period. It has yield-

outline of the coastal area: Erdrich 2001.

ed Mesolithic and Neolithic finds as well as ship wrecks

An exception is Hiddink 2003. A more recent paper out-

from the Medieval period, but no Roman period figurines

lines changing burial rituals in the northwestern provinces

or other artefacts.

of the Empire from the late Iron Age onward, focusing on elite funerary assemblages: Pearce 2014.

19

Map 1 Find-spots of statuettes and statuette fragments in the Netherlands Base map: paleogeographic reconstruction of the Netherlands around 100 (Vos/Vries 2013).

20

igation route over Lake Flevo and the middle course of the river Ems, which discharges in the Wadden Sea east of Delfzijl (Groningen), nowadays part of the border between the Netherlands and Germany.107 Roman presence in the north and interaction with Frisii and Chauci have been attested by imported goods. Before 50, imported goods predominantly came from the coastal areas of the provinces North Holland, Friesland and Groningen. Between 16 and 50, most finds from North Holland come from Velsen, where two Roman forts and one, maybe two harbours were constructed that functioned between 15 and 47 (see below). After 50, an increasing number of Roman goods reflects intensified contact between the people north and south of the Rhine, which was at its peak between around 160 and 200.108 Towards the end of the 3rd century contacts between Romans and local populations shifted from the northern terp region to the eastern sandy soils of the provinces of Drenthe and Gelderland.109 The first explanation for the presence of imported goods north of the Rhine is that these goods were a means of payment made by Roman troops for food and livestock.110 The idea of structural trade between Romans and local tribes has been rejected by Michael Erdrich and Henk Hiddink.111 The latter has argued that valuable objects, weaponry and coins were diplomatic gifts to confirm agreements between local and Roman authorities, who after retreating behind the Rhine, attempted to create a buffer zone north of that river.112 Hiddink, however, does not exclude trade on a small scale and refers to the presence of goods for daily use, like pottery and tephrite querns.113 Daniël Gerrets, too, points out that the Romans needed livestock, horses, salt and wool. He also argues that in the debate on trade scholars do not differentiate between goods, while objects like bronze statuettes or wine strainers probably circulated in other circles than querns, which already were imported before the arrival of the Romans.114 At sites where specialized crafts were carried out on a somewhat larger scale, such as iron production at Schagen-Muggenburg, the presence of imported goods also demonstrates that goods were exchanged between its inhabitants, Romans and peoples overseas. Besides a means of payment or diplomatic gift, Roman goods, exotic objects such as statuettes and coins may have been brought by Germanic auxiliaries or veterans returning to their homeland. 115 A considerable number of coin hoards have been recovered from all five provinces north of the Rhine.116 Apart from a few exceptions, they consist of denarii. The composition and find circumstances of several of these hoards suggest that they were buried as ritual deposits.117 Some hoards, however, contradict this interpretation. The coin hoard of 1,266 denarii from Uitgeest, for instance, may have been a part of a veteran’s savings or maybe even a bonus. The praemium for retiring legionaries was 3,000 denarii, until Caracalla increased it to 5,000 in 215. Whether auxiliary veterans could receive a bonus is not known, but if they did, it probably consisted of denarii, too.118 It has also been suggested that Roman coins and other valuable objects were booty, captured by Germanic tribes from the Romans.119 Since most bronze statuettes from the northern coastal region are single finds, it raises the question why Frisii and Chauci would take especially statuettes of Mars and Mercury instead of Roman weaponry, militaria or other Roman artefacts. 120

107

Van Es 1981, 265 and note 581.

116

Van Es 1960; Hiddink 1999, 203-206.

108

Hiddink 1999, 205 and note 1288.

117

See, for instance, Hiddink 1999, 203 and note 1276.

109

Van Es 1981, 263-270.

118

Wesch-Klein 2011, 444. According to Nicolay auxiliary vet-

110

Van Es 1981, 265-266.

111

Erdrich 2001, 72-108; Hiddink 1999, 201-210.

112

Van Es already suggested that the Romans created a buffer

erans did not receive a discharge payment: Nicolay 2007, 164. 119

zone in the area on the right bank of the Rhine: Van Es 1981,

See for the importance of raiding in Germanic culture, for instance, Hiddink 1999, 190-191, and 203 for the interpreta-

36, 42, 266. See also Hiddink 1999, 203-207.

tion of buried hoards in the Rhine-Weser area.

113

120

Hiddink 1999, 206 and note 1291.

114

Gerrets 2010, 140-142, 206.

statuette. It cannot be ruled out that during the quarrying

115

See for the function of statuettes in the northern Nether-

of terpen more finds came to light which have not been

lands: Van Es 1981, 266; Knol 2014, 196.

reported and registered.

A few statuettes have been found together with another

21

Map 2 Find-spots of statuettes north of the Rhine Base map: paleogeographic reconstruction of the Netherlands around 100 (Vos/Vries 2013).

There is one other explanation for the presence of imported goods in the northern provinces that needs to be mentioned. Excavations have demonstrated that Roman artefacts have been recovered from post-Roman find-contexts, such as the terra sigillata sherds from the Tjitsema terp in Wijnaldum and the Oldenhove terp in Leeuwarden. The same applies to roof tiles, militaria, coins and sculpture such as the large fragment of a limestone statue or column found at Wijnaldum and the Hludana altar from the Besseburen terp at

22

Beetgum.121 The results of a recent study of terra sigillata sherds from Frisian terpen also suggest the use of terra sigillata sherds in a secondary context. It seems that especially in the 2nd and 3rd centuries, terra sigillata sherds were imported instead of complete tableware.122 It has been argued that the value and use of these sherds was symbolic rather than practical. Terra sigillata sherds have been found in assemblages that have been interpreted as ritual deposits, dating from the Roman period into the early Medieval period. They may even have been gifts or a means of payment.123 Outside Friesland, in North Holland and Schleswig-Holstein (Germany), terra sigillata has been found in early Medieval contexts, while other Roman artefacts were absent. Likewise, also Roman statuettes may have been collected after the Roman period. It seems, however, unlikely that this would apply to all statuettes from the northern terp region. The execution of most bronze statuettes from the northern coastal area indicates that they were manufactured outside the study area.124 Whether their presence in the north is the result of gift exchange, barter between Romans and the terp habitants, or evidence of Germanic auxiliaries bringing statuettes to the north is a problem that, for the time being, cannot be solved. For the question why statuettes were left behind in the northern terp settlements when they were abandoned in the course of the 3rd and 4th centuries, I propose the following solution. What or whom terp dwellers saw in the statuettes is unknown, but they could have exchanged them for other goods or have melted them down for recycling. Since they did neither of the two, the statuettes may have been left behind deliberately, either as part of an abandonment ritual or an offering to supernatural powers. The existence of such practices is suggested by metal objects, statuettes and coin hoards recovered from peat bogs and watery places outside the terp region in the northern Netherlands and beyond.125 Also finds from a few controlled excavations seem to confirm this idea, such as a bronze Mercury statuette in a ditch at Tzum (Friesland), a bronze putto from a ditch at Texel (North Holland) a terracotta bird under the entrance of a house at Schagen and a Victoria statuette in a post hole of a house at Colmschate.126 Since most statuettes from the terp region lack archaeological data, the idea of ritually deposited statuettes in these settlements and elsewhere in the north has to remain hypothetical. 2.2.2

Friesland and Groningen: the northern coastal area

Together, the provinces of Friesland and Groningen cover the northern coastal area. When exactly this region became inhabited is still open to discussion. Recently, an excavation at Wommels-Stapert (Friesland) has produced pottery that probably can be dated to the 8th or 7th century BCE. Westergo, the western part of the Frisian coastal area, was part of a tidal marsh landscape with low elevations, and therefore inhabited earlier and more intensively than the eastern part, may be even as early as the Bronze Age. Several suggestions have been made concerning the origins of these pioneers. They may have come from northwest Niedersachsen, from Drenthe or North Holland. The number of settlements east of this area is substantially smaller than in Westergo, and colonisation probably began during the late Iron Age.127 121

Gerrets 2010, 137-138.

statuettes, a wooden pedestal with an inscription, probably

122

Volkers/Polak 2016, 235-312.

to Victoria, a bronze griffin head (helmet decoration?), a

123

Nieuwhof 2020, 94-110.

bronze disc with a lion’s head (phalera or shield decora-

124

See on the difference between statuettes found north and

tion?), a Roman coin and a spearhead: Hübner 1876, 66-69.

south of the Rhine chapter 3.

A find-assemblage comprising a golden neck ring, six

See for instance, the bronze statuettes from Heeren-

silver rings, two small bronze sticks and a bronze Mercury

veen-De Knipe, Zutphen and Heerde (this chapter). See for

statuette comes from Beelen: Schoppa/Hucke 1938. Both

125

Germany the find-assemblages from Marren-Oldenburg

assemblages date to the 4th century or slightly later.

(Niedersachsen) and Beelen (Westfalen). The first was

126

See this chapter.

found during ploughing. It comprised two bronze Mars

127

Gerrets 2010, 51-52.

23

At the time of Caesar’s campaigns in Gaul, the region was predominantly inhabited by the Frisii, whose territory stretched between the rivers Ems and Rhine.128 The Chauci lived east of the Frisii. According to Ptolemy, their territory stretched between the rivers Ems and Elbe.129 It is assumed that the river Lauwers constituted a part of the western border of the territory of the Chauci. Nowadays, this river is part of the modern border between the provinces of Groningen and Friesland.130 It is a generally accepted view that the northern coastal area was gradually abandoned in the course of the 3rd century, until the area became largely deserted in the 4th century. The appearance of new pottery styles and cruciform brooches in the northern coastal area in the 5th century led some scholars to believe in an Anglo-Saxon invasion. This idea seemed to be confirmed by a burnt layer in settlements like Ezinge, on top of which small Anglo-Saxon dwellings were built. The results of Ernst Taayke’s PhD thesis on the pottery of this area, as well as excavations at Wijnaldum, suggest that it was not a ‘hostile take-over’, since the Anglo-Saxons arrived here when the area was almost completely deserted. Almost, indeed, since Taayke also pointed out that a few settlements in Groningen remained inhabited, such as Ezinge.131 This was underlined by the results of recent research, which demonstrated continuing habitation in the 4th century in a few settlements, among which Ezinge and Midlaren-De Bloemert.132 In the northern coastal area numerous pre-Roman and Roman artefacts have come to light from the 19th century onwards, as the artificial dwelling mounds, called in Dutch terpen or wierden, were quarried for their rich soil, which was used for the fertilisation of farmland.133 The majority of finds from the northern terp region were found during the quarrying of terpen. Before the construction of the sturdy dikes along the Dutch coasts, these manmade mounds with farmsteads protected people and cattle from high tides and floods. Pieter Boeles, who was the curator of the Frisian Museum during the first half of the 20th century, collected finds from the terpen and published an important overview of Friesland’s habitation history.134 Albert Egges van Giffen started investigating terpen and settlements from the beginning of the 20th century onwards. His systematic excavation method attracted national and international attention. At Ezinge (Groningen) Van Giffen carried out excavations between 1923 and 1934. These excavations were never fully published, until in 2011 a project was started to study and process Van Giffen’s data. This resulted in publications by Annet Nieuwhof on the finds from Ezinge and the habitation history of the terp region in the northern Netherlands from the late Roman period into the Medieval period.135 Other important projects were the excavations at Wijnaldum, the Frisia-project in the 1990s and the continuation of this project, the ‘Terpenproject Steilkantonderzoek Friesland 2009-2013’, focusing on research of levelled terpen.136 They have produced numerous data and resulted in publications that shed a new light on the development of settlements and society in the Dutch Germania Libera region, from the Iron Age into the Medieval period.137 The next section provides an overview of the find-spots of statuettes and statuette fragments from Friesland, Groningen and North Holland.

128

Van Es 1981, 21 and fig. 6.

129

Ptol. Geog. 2.11.8, in Lanting/Van der Plicht 1999/2000, 60.

Friesland 1991-1993, Frisia Project 1998-2000: Dongjum,

130

See, for instance, Therkorn et al. 2006, fig. 1, after Bloemers

Friesland, Peins, Friesland, Englum, Groningen. Terpen-

1983.

project Steilkantonderzoek Friesland 2009-2013: Achlum,

131

Taayke 1996, 193-201.

Jelsum, Firdgum and Oosterbeintum, Dronrijp-South and

132

Nieuwhof 2013, 54-84.

East. Other major excavations in Friesland took place in

133

Which is why many artefacts are found on farmland.

Leeuwarden (2005-2006), Arkun (2012), Wartena, Sakse-

134

Boeles 1951.

135

Nieuwhof 2013, 2014, 2015. A publication of the Ezinge excavations is in preparation.

24

136

I thank Johan Nicolay for this overview. Wijnaldum,

noord (2013), Sneek,Wommels-Stapert (2014). 137

See, for instance, Nieuwhof 2013.

Map 3 Find-spots of statuettes in Friesland and Groningen Base map: paleogeographic reconstruction of the Netherlands around 100 (Vos/Vries 2013). 1. Makkum 2. Pingjum 3. Kimswerd 4. Almenum 5. Wijnaldum 6. Herbaijum 7. Hitsum 8. Dongjum 9. Tzum 10. Spannum 11. Lions 12. Dronrijp 13. Menaldum 14. Beetgum 15. Marssum 16. Jelsum 17. Cornjum 18. Hallum 19. Ferwerd 20. Blija 21. Holwerd 22. Veenwoudsterwal 23. Idaard 24. Schettens 25. Nijlanderzijl 26. Harich 27. Heerenveen 28. Houwerzijl 29. Oldehove 30. Joeswerd 31. Ezinge 32. Krangeweer 33. Kantens 34. Rottum 35. Loppersum 36. Tolsum 37. Oosterlittens 38. Oosterbeintum 39. Enumerhoogte 40. Sneek 41. Wirdum

Friesland • Sneek-Stadsrondweg Northeast of the modern town of Sneek, excavations in an early Roman period settlement revealed two occupation phases, the first dating from the end of the 1st century BCE until the end of the 1st century, the second dating to the 2nd or 3rd century. Most finds come from the first phase.138

138

Niekus 2002, 17-22.

141

Van Boekel 1987, nr. 36.

139

Ufkes 2002, 47-51 and figs. 4.2-4.5. See for the meaning of

142

See for the mother goddess statuette Van Boekel 1987, nr.

the wheel the site Leiden-Roomburg-vicus. 140

88.

Bowman et al. 2009, 156-170.

25

The excavations have produced four terracotta figurines, representing a bovine or sheep (fig. 2.1), a cockerel or hen, an egg and a fragmented object that has been interpreted as a wheel.139 • Tolsum It was a long-held idea that the text on the Roman writing tablet from the Tolsum terp in Friesland mentions the purchase of a cow. In 2009, a Dutch-English research team from Oxford University demonstrated with digital photography that the transcription and translation by Carl WilhelmVollgraff, made in 1919, is not correct. According to the new interpretation it is a contract between a debtor and a creditor. The debtor declares having borrowed an amount of money from a certain Carus, a slave owned by a Julia Secunda. A tribune of the Fifth Legion and a soldier of a Batavian unit act as witnesses or interpreters. The presumed date of the contract is February 23 of the year 29. If the date is correct, it suggests that Romans remained in contact with the Frisians after the Frisian Revolt of 28.140 This seems to be confirmed by a find from the same terp, five terracotta fragments of a Fortuna statuette dating to the end of the 1st century.141 Together with a terracotta fragment of a mother goddess from Oosterlittens these are the only known imported terracotta statuettes in the terp region.142 • Tzum An excavation in a levelled terp (Greate Vlearen) in 1983 yielded five fragments of a human skull, which were found together with a bronze Mercury statuette (figs. 2.183-85), a cast bronze plate with a decoration in the form of two s-shaped lines, a large iron nail, one half of a key and potsherds. The latter belong to native pottery as well as wheel-thrown pottery resembling terra nigra. The bronze plate could be a bronze door fitting. Attached to it was a corroded iron object, the remains of a hinge. The objects came from the deepest level of a ditch.143 The statuette is the only specimen that has been recovered from a terp during an archaeological excavation. The metal objects are not the first striking finds from the terp. In 1849, the discovery of potsherds with geometrical decorations and a human skeleton were reported. During levelling of the terp in 1850, a coin hoard of 13 silver denarii came to light. The latest coin was struck under Commodus, around 180. A second skeleton came to light in 1959 and prompted a small excavation. Based on the finds it was established that the habitation period of the terp started in the 4th century BCE and continued into the early Medieval period. The ditch has been dated to the 2nd century, based on carbon from the ditch and the pottery.144 • Winsum The Bruggeburen terp, one of the highest in Friesland, was quarried for its fertile soil between 1865 and 1900. Early Roman pottery sherds from this terp attracted the attention of Boeles (see above), and led him to believe that a military post was built here.145 Boeles also suggested that the settlement functioned as an administrative centre for the Roman army and tax collectors. The majority of the Roman period finds date between 12 and 28, the year of the Frisian Revolt. Besides local and Roman pottery, many wine and oil amphorae came to light. The coins from the terp date from the last half of the 1st century BCE to the reign of Tiberius (14-37), the earliest being a denarius struck in 46 BCE.146 In the 19th century, a metal hoard of 6.5 kilo was discovered, consisting of gold and silver vessels. A coin hoard was found in the same terp. It consisted

143

I owe special thanks to Evert Kramer, former conservator of

144

the Fries Museum and excavator of the site in 1983. He pro-

Niehof 2015, 391.

vided the information on the objects and the excavation

145

Bos et al. 1997, 65.

from hitherto unpublished manuscripts. A bronze fibula

146

Bos et al. 1997, 66-67; Galestin 1999/2000, 229-233.

from the filling of the ditch does probably not belong to the assemblage: personal information by Evert Kramer.

26

See on the 1983 finds from Tzum also Elzinga 1984, 127-128;

of 22 denarii, the latest coin being struck in 253.147 Excavations carried out in 1997 produced a wide variety of finds, dating from the 7th century BCE to the 14th century.148 The pottery assemblage and the coins have been linked to an early Roman military presence at Winsum. The bronze asses with countermarks are typical of soldier’s wages. The range of pottery from Winsum-Bruggeburen is similar to what has been found in military forts in the Netherlands, Germany and Great Britain, an assemblage of pottery types that has been described as a ‘fortress assemblage’.149 Winsum-Bruggeburen is a location in the northern Netherlands where a military presence seems plausible, although no traces of a military camp have come to light. The exact function of the terp settlement itself is still not clear, but its strategic location and the finds suggest that it was already an important political and/or administrative centre in the region at the beginning of the 1st century.150 Also smaller objects suggest pre-Roman differentiation in the northern coastal area. An amateur-archaeologist found two decorated bronze heads of split pins, one in 1997 and the other in 1998, in two small terpen just north of Winsum, near Salverda Menaldum (Friesland). They probably belonged to two late La Tène belt hooks that were made exclusively in the Dutch river area. Until this discovery, they had never been found outside that region, which implies that goods were exchanged between local or regional elites from the north and south during the late Iron Age.151 The Bruggeburen terp has yielded no statuettes.

147

148

Van Es doubted the coin hoard was a closed find complex

149

Galestin 2001-2002, 463.

since the latest coin was struck hundred years later than

150

Gerrets 2010, 133-134.

the other coins: Van Es 1960, 88.

151

Nicolay 2005, 130-135.

See for the coins and the pottery of the 1997 excavations at Winsum-Bruggeburen Galestin 1999-2000 and 2001-2002.

27

Friesland: find-spots of statuettes without contextual data Find-spot

Material

Subject

Reference

Almenum

bronze

Apollo

Zadoks 1967, nr. 2

Beetgum

bronze

Mercury

Zadoks 1967, nr. 35

Blija

bronze

Mercury

Zadoks 1967, nr. 31

Cornjum

bronze

genius

Zadoks 1967, nr. 8

Cornjum

bronze

Mars

Zadoks 1967, nr. 21

Dongjum

bronze

Mars

Zadoks 1967, nr. 20

Dronrijp

bronze

Mars

Zadoks 1967, nr. 23

Ferwerd

bronze

Mercury

Zadoks 1967, nr. 30

Hallum

bronze

Mercury

Zadoks 1967, nr. 33

Harich

bronze

Mercury

Zadoks 1967, nr. 3152

Heerenveen-De Knipe

bronze

mounted warrior

Galestin 2007, 74-79153

Herbajum

bronze

Mars

Galestin 1994, 155-159

Hitsum

bronze

Juno(?)

Zadoks 1967, nr.12

Holwerd

bronze

Mercury

Zadoks 1967, nr. 38

Holwerd

bronze

horse

Zadoks 1967, nr. 51154

Idaard

bronze

Mercury

Zadoks 1967, nr. 39

Jelsum

bronze

Mercury

Zadoks 1967, nr. 29

Kimswerd

bronze

Mercury

Zadoks 1967, nr. 25155

Lions

bronze

Isis-Fortuna

Zadoks 1967, nr. 11156

Makkum

bronze

horse

Zadoks 1967, nr. 49157

Marssum

bronze

billy-goat

Zadoks 1967, nr. 44

Menaldum

bronze

Jupiter

Zadoks 1967, nr. 15158

Nijlanderzijl

bronze

Mars

Zadoks 1967, nr. 17

Oosterbeintum

terracotta

bird

Boeles 1951, 537159

Oosterlittens

terracotta

mother goddess

Van Boekel 1987, nr. 88

Pingjum

bronze

Mercury

Zadoks 1967, nr. 27

Schettens

bronze

Mercury

Zadoks 1967, nr. 32

Spannum

bronze

Mars

Zadoks 1967, nr. 19

Tzum

bronze

Hercules

Zadoks 1967, nr. 9

Tzum-Tolsum

terracotta

Fortuna

Van Boekel 1987, nr. 36

Tzum

bronze

Mars

Zadoks 1967, nr. 18

Wijnaldum

bronze

Minerva

Zadoks 1967, nr. 41

152

According to the catalogue the statuette depicts Apollo,

154

Fig. 2.2 and 3.5.5.

but in my opinion it represents Mercury. This is suggested

155

Figs. 2.4-2.5 and 3.4.1, 3.5.1.

by the figure’s posture as well as by the winged cap, which

156

Figs. 2.6-2.7 and 3.4.1.

the authors have interpreted as a type of hair dress. The

157

Figs. 2.8-2.9. Originally, the horse carried a rider: 3.5.5.

posture is comparable to that of nrs. 27-29, 35-36 in the

158

Figs. 2.10-11. See for the interpretation of the statuette as Jupiter-Sabazius 3.5.3.

same catalogue. 153

28

Fig. 2.3. See for the meaning of such figurines 3.5.5.

159

Fig. 2.12 and 3.5.8.

Friesland: uncertain find-spots of statuettes Find-spot

Material

Subject

Reference

Friesland

bronze

Apollo

Zadoks 1967, nr. 1

Friesland

bronze

Mars

Zadoks 1967, nr. 16

Friesland

bronze

horse

Zadoks 1967, nr. 50

Groningen • Ezinge In 1918 started the levelling of the wierde Ezinge, situated in northeast Groningen. Since the works yielded a large number of finds, Van Giffen carried out excavations between 1923 and 1930. A larger excavation between 1931 and 1934 brought to light several house plans. In his first publications on Ezinge, in 1926 and 1928, Van Giffen thought that the earliest occupation level dated to the beginning of the 1st century.160 Finds like a La Tène fibula and geometrically decorated pottery led him to correct this in 1931. Thirty years later, in 1964, an excavation confirmed the date of the earliest occupation level, the 5th or 4th century BCE. With radiocarbon dating a limewood post could be dated between 540-200 BCE. Recent research of other wierden in the Reitdiep area endorses the date of the earliest occupation level of these settlements in the 5th century BCE.161 Habitation at Ezinge continued into the Medieval period, but the settlement is an exception, since many terpen were abandoned in the course of the 3rd century. Nieuwhof’s research on Ezinge has produced some 350 pre-Roman and Roman period find assemblages that are possibly related to ritual practices, varying from terra sigillata sherds and single skulls to complete human and animal skeletons.162 The terp has yielded three bronze statuettes, including a bronze cockerel, a statuette of Jupiter, and a statuette of a standing male, a cavalryman (fig. 2.14).163 All three statuettes lack contextual data.

160

161

See for the results of Van Giffen’s excavations at Ezinge

162

Nieuwhof 2014, 162.

Nieuwhof 2014.

163

Knol 2014, 187-197, Galestin 1990, 146-152. See for the inter-

Nieuwhof 2014, 11-18.

pretation of the statuette 3.5.6.

29

Groningen: find-spots of statuettes without contextual data Find-spot

Material

Subject

Reference

Enumerhoogte

terracotta

bird

Taayke 2005, 162-163164

Houwerzijl

bronze

Isis

Zadoks 1967, nr. 10165

Joeswerd-Winsum

bronze

Mars

Noordelijk Depot Nuis, Groningen166

Kantens

bronze

horse

Noordelijk Depot Nuis, Groningen167

Krangeweer

bronze

Mercury

Zadoks 1967, nr. 10

Krangeweer

bronze

Matron

Zadoks 1967, nr. 43

Loppersum

bronze

Mercury

Zadoks 1967, nr. 34

Oldehove-Frytum

bronze

pedestal

Zadoks 1967, nr. 43

Rottum

bronze

mother goddess?

Zadoks 1967 ,nr. 24

Rottum

bronze

Mercury?

Zadoks 1967, nr. 40

Wirdum

bronze

Minerva

Zadoks 1967, nr. 42

Wirdum

chalk

bear

Miedema 1999/2000, 247-248168

Wirdum

terracotta

male figure

Knol 2005, 196169

2.2.3

North Holland: the northwestern coastal area

Starting in the late 1970s, several projects in North Holland were set up to investigate the development of landscape use and settlements from the Bronze Age onwards. Research focused on the region of the Oer-IJ estuary, between the cities of Alkmaar, Haarlem, Amsterdam and the North Sea coast. Between 1978 and 1982, a terrain of 2,200 hectares, the Assendelver Polders, was investigated. Initially, the focus was on farmsteads on the western sandy soils and later shifted to habitation on peat. The ten sites on peat yielded traces of dwellings from the 7th century BCE to the 1st and 2nd centuries.170 Modelling human and animal figures from clay is a tradition that was born during the Palaeolithic. An assemblage of fired clay animals from a Bronze Age ditch at Grootebroek (North Holland) seems, therefore, not surprising. The complete figurines represent a bull, a pig, and an undetermined animal. The fragments depict parts of cattle, too, such as horn fragments and a pig’s legs and snout.171 Terracotta specimens from Roman period sites in the north represent mostly animals. A bird figurine has been recovered from a late Roman find-context at Schagen (see hereafter).

164

Fig. 2.13 and 3.5.8. The terracotta bird was found in a dung

165

layer in the wierde Enumerhoogte. It has a hole on each

found in the same terp: Galestin 1995, 91-98.

side of the neck for a wooden stick, to make the head of

166

Fig. 2.16 and 3.5.1.

the bird wiggle. The head could be removed. Since the

167

Fig. 2.17 and 3.5.5.

original head was missing, a copy was made from the head

168

Figs. 2.21-24 and 3.5.9.

of a comparable specimen from Oosterbeintum. This bird

169

Fig. 2.25 and 3.5.7.

was on display at an exhibition entitled ‘Professor Van

170

Therkorn et al. 2006, 12-16.

Giffen and the secret of the wierden’ in 2005-2006 at the

171

Van Regteren-Altena et al. 1977, 241-254.

Groninger Museum.

30

Fig. 2.15 and 3.5.2. An almost identical statuette of Isis was

Map 4 Find-spots of statuettes in North Holland Base map: paleogeographic reconstruction of the Netherlands around 100 (Vos/Vries 2013). 1. Naarden-Het Gooi 2. Velsen 3. Castricum 4. Texel-Den Burg 5. Schagen

• Schagen From 1974 onwards, excavations at Schagen produced several Roman period farmsteads. The excavations at Schagen-Muggenburg were carried out in 1983-1984 and 1989-1990. The sites, known as Schagen-Muggenburg I and III, yielded plans of relatively rich farmsteads which could be dated to around 300 and the early to mid-4th century.172 Evidence of iron production and wool spinning explains the presence of imported luxury goods as well as imported coal and pottery from Britain. At the site Witte Paal, traces of Iron Age habitation indicate that the first pioneers settled here around 300 BCE. Here, too, iron production has been attested. Two coin hoards were found, both dating to the 3rd century. The first came to light at Schagen-Muggenburg, and consisted of four antoniniani, of which the latest was minted between 268 and 270 in Rome. A large coin hoard at Schagen-Witte Paal comprised at least fifty silver antoniniani, all dating to the 3rd century. A golden finger ring and silver hair pin came to light on the same site.173 At Schagen-Muggenburg I numerous shallow and deep pits were discovered, as well as linear features. Ten shallow pits contained complete skeletons, bone bundles, human inhumations and cremated remains covered by inverted pots. The assemblages suggest a ritual function of the pits and their contents, just like the varied contents of the deeper pits.174 Pits and linear features at Schagen-Muggenburg III share similarities with the pits and deposits at Schagen-Muggenburg I. Based on the lay-out of the dwellings, the linear features in combination with the pits and their contents, Linda Therkorn has argued that they were linked to the seasons, star alignments and constellations.175 From Schagen-Muggenburg III comes a terracotta bird (figs. 2.26-27).176 The head is missing, but a hole on each side of its neck indicates that it rested on a small wooden or metal stick. Like the heads of the birds from Enumerhoogte (Groningen) and Oosterbeintum (Friesland) this head, too, could wiggle and was removable.177 The figurine was found in a pit under the entrance of a building.178 The building may have been an outbuilding belonging to a house which was constructed around 270. 172

Therkorn 2004, 15.

173

Diederik 2002, 23-35, 58-60.

174

Therkorn 2004, 17-28.

175

Therkorn 2004, 45-55.

176

Diederik 2002, 99. I thank Martin Veen (Huis van Hilde,

(Archeocultura) for the photos and the additional documentation on the terracotta figurine. 177

Figs 2.26-27. See for the other two birds figs. 2.12-13 and 3.5.8.

178

See for the interpretation of the assemblage 4.2.2.

archaeological depot Noord-Holland) and Frans Diederik

31

• Texel-Den Burg Between 1971 and 1975, excavations were carried out on the northwestern edge of Den Burg, on the island of Texel, and produced numerous traces of continuous occupation from the end of the middle Bronze Age into the Medieval period. One house could be dated to the late Iron Age and seven to the Roman period. The houses were enclosed by ditches from the early Roman period onwards.179 One of the sites was inhabited with intervals in the Iron Age, in the Roman period between 100 and 325, and again from the 6th century onward. At the spot where an enclosure ditch of a Merovingian building cut through a late Roman ditch, a bronze statuette was found (fig. 2.180). It is uncertain whether it ended up in the ditch in the Roman period or in the early Medieval period.180 Despite the fact that the figurine is an attachment, it will be discussed in chapter 4, because of its assumed secondary function as a ritual deposit.181 North Holland: find-spots of statuettes without contextual data Find-spot

Material

Subject

Reference

Castricum

bronze

Victoria

Detector Magazine 136, 2014

Naarden-Het Gooi

lead

Minerva

Private collection182

2.2.4

Drenthe, Overijssel and the northern part of Gelderland: the northeast

The eastern part of the Netherlands is characterized by Pleistocene cover sand soils. Continental glaciers have reached the area at least three times, leaving behind deep depressions and fissures in the landscape filled with clay, sand and peat. During the last glacial period, the glaciers did not reach the Netherlands, but in the cold and arid climate, vegetation could hardly survive. As a consequence, wind-blown deposits formed a layer of one to several meters of sand. The boundaries of the later village territories were natural landscape features like stream valleys and valleys with peat or raised bogs.183 Which tribes had their territories in the eastern Netherlands is not entirely clear. Besides Groningen and Friesland, the Frisii may have inhabited also northern Drenthe.184 The Tubantes are mentioned for the first time in 14, when Germanicus and his troops are being attacked by Bructeri, Usipetes and Tubantes.185 They probably are the same as the Tuihanti, a name mentioned on two 3rd-century votive altars from Vercovicium, modern Housesteads, near Hadrian’s Wall. According to some scholars, the Tuihanti were a Frisian tribe.186 They probably gave their name to the region Twenthe, at the east side of Overijssel. The large scale iron production at Heeten (see below) has been connected with the manufacturing of arms for the Salii. This tribe is mentioned for the first time in 358 by the emperor Julian (355-360), who intercepted a Frankish tribe that wanted to settle in Toxandria, a region east of the river Scheldt. Julian describes them as ‘a people belonging to the Salii’. It is generally accepted that the Salii are connected to the Salland region in western Overijssel.187

179

I thank Philip Woltering for the data of the excavation.

183

Harsema 1987, 108.

Woltering 2000, 24-33. See for the most recent publication

184

Lanting/Van der Plicht 2009/2010, 59.

of the excavations Woltering 2017.

185

Lanting/Van der Plicht 2009/2010, 60-61. Tac. Ann. 1.51.

180

Additional information by Philip Woltering.

186

Which is disputed by others: Lanting/Van der Plicht

181

See for the interpretation of the statuette Veen 2017, 250252 and 4.2.2- 4.2.3.

182

32

2009/2010, 61. 187

See, for instance, Van Es 1967, 540-558; Lauwerier et al.1998-

I thank Marenne Zandstra for bringing this lead figurine to

99, 189 and note 97, Van Beek 2009, 83 and fig. 3.12 (after

my attention. Without a thorough examination it cannot

Hiddink 1999, fig 8.3, 221). The idea is rejected by Lanting

be excluded that the lead figurine is a post-Roman copy.

and Van der Plicht: Lanting/Van der Plicht 2009/2010, 69.

Map 5 Find-spots of statuettes in Drenthe, Overijssel and northern Gelderland. Base map: paleogeographic reconstruction of the Netherlands around 100 (Vos/Vries 2013). 1. Zuidlaren 2. Annen 3. Beilen 4. Dalfsen 5. Heeten 6. Heerde 7. Colmschate 8. Zutphen 9. Didam 10. Westervoort 11. Bennekom 12. Ede 13. Wehl 14. Noordbarge

Research in the northeast has mainly focused on settlement archaeology. In the early Roman period, settlements consisted predominantly of Einzelhöfe. Habitation was still diffuse and mobile, as a consequence of the Celtic field system, and only slowly increasing.188 It would take another century before the settlement system would transform into larger and stable settlements. The majority of these settlements were permanently occupied until the end of the 4th or the beginning of the 5th century. Hiddink distinguishes between large, medium-sized and small settlements.189 According to this classification, Wijster, Heeten, and probably also Ede-Bennekom (see below), were large settlements, consisting of ten to fifteen farmsteads. Settlements from the middle and late Roman period are well represented in this region. Numerous Bronze Age and Iron Age sites have been discovered and examined in Overijssel.190 Several Roman period settlements have been uncovered as well, including Colmschate, Dalfsen, Denekamp, Raalte and Heeten (see below).191 188

Hiddink 1999, 133.

189

Hiddink 1999, 123-156.

190

See for an overview of excavations in Overijssel between

1969-1999: Groenewoudt et al. 2006. 191

Van Beek 2011, 428-446. See for Colmschate also Hermsen 2003, 65-71 and Hermsen 2007.

33

Drenthe • Noordbarge At the location De Hoge Loo, Noordbarge, an assemblage of 24 terracotta statuette fragments and bronze objects, mostly statuettes, came to light in 1845 and 1846 when people were digging for loam. De Hoge Loo is an area of elevated terrain with a meadow at the south side, nowadays located between the residential areas of Noordbarge and Bargermeer (Emmen). The word loo means wood, and meer means lake, which dried out when a canal was dug here in the 19th century.192 The statuettes were one of the reasons why Van Giffen conducted a small excavation in the area in 1932, since he suspected the presence of a Roman period settlement.193 In 1934, he published the results of his first excavation, including the exact find-spot of the statuettes. East of the find-spot, he found a ditch and traces of a rampart with palisade, as well as sherds of Roman pottery which led him to conclude that a fortified, native settlement was situated here in the Roman period.194 Although it turned out that the rampart and ditch were of a much later date, excavations between 1972 and 1974 produced around thirty house plans, granaries, sunken huts and traces of enclosures. The excavators distinguish four periods, from the late Iron Age into the early Roman period. During the last period, in the early 1st century, the settlement consisted of an enclosure with three longhouses and several outbuildings.195 The number of statuettes found at De Hoge Loo as well as the nature and location of the site suggest it may have been an open-air sanctuary or sacred place.196 Some of the items from De Hoge Loo have disappeared in the course of time. In later publications two ‘new’ bronze figurines appeared: the head of a donkey or mule and a headless animal figurine, interpreted as a bull by Van Giffen, as a boar by Zadoks et al. (fig. 2.38).197 Both objects were found at the same spot as the others. The donkey or mule’s head (fig. 2.37) is the handle of an object and does not occur in the 1967 catalogue by Zadoks et al.198 Nowadays, only the bronze boar, the Mercury statuette, the bronze pedestal and the donkey head, as well as nine terracotta statuettes have survived (figs. 2.28-36).199 The terracotta specimens comprise two almost identical heads of Minerva, a statuette depicting three mother goddesses on a bench, a bust of a boy, a Venus statuette, the head of a female figure, probably Venus, fragments of a head depicting Bacchus or Apollo, an enthroned warrior, and the head of a female figure. The latter belonged to a statuette depicting a couple, since the right hand of another person rests on her right shoulder (figs. 2.28-29).200 The warrior wears an armour, leggings and boots (figs. 2.34-36).201 Besides the unusual iconography, its head is missing, which makes the identification impossible. Apart from the warrior statuette, the terracotta statuettes were manufactured in the Rhine-Moselle area, which implies a deposition date in the 2nd century or later.202 Two of them were manufactured between 150 and 200 and thus deposited around the end of the 2nd century or later.203

192

Van Giffen 1934, 91.

193

199

Van Giffen 1934, 88-93.

143-150. Three terracotta statuettes were not accessible. See

194

Van Giffen 1934, figs.1-2.

for these statuettes Van Boekel 1987, nrs. 74, 137, 181. See for

195

Hiddink 1999, 132-133 and fig. 5.30.

the fragmented head of Apollo/Bacchus Van Boekel 1987,

196

See for the interpretation of alleged cult places also chap-

See also Zadoks et al. 1967, nrs. 28, 45, 52; Galestin 2000,

nr. 160.

ter 1.

200

See also Van Boekel 1987, nr. 168.

197

Van Giffen 1934, 92, Zadoks et al. 1967, nr. 45.

201

See also Van Giffen 1934, fig. e. The figurine does not feature

198

The authors may have been under the impression that the donkey head is not a Roman period artefact. See Jansen

Van Boekel 1987, 207.

1849, 83-86, Pleyte 1882, 19-21 and figs XXI-XXIII. Donkeys

203

See for the production date of these statuettes Van Boekel

and mules were associated with Dionysus and often depicted in bronze, for instance as a fulcrum terminal.

34

in Van Boekel 1987. 202

1987, 433-437 and nrs. 74 and 181.

• Wijster The largest known and fully published Roman period settlement in Drenthe is Wijster.204 The earliest occupation phase begins in the last half of the 2nd century and ends around the middle of the 5th century. In the 4th century, the settlement consisted of ten houses according to Hiddink, whereas Van Es estimated the number of houses at fifteen to twenty, with 150 to 200 inhabitants.205 Two large buildings within the settlement have been interpreted as a Herrenhof, based on the presumed monumentality of the entrance of one of the houses, the dimensions of the houses, the absence of stables and the quantity of imported goods at this location which seems to be larger than elsewhere in the settlement.206 As to the economy of Wijster, the importance of cattle raising is suggested by the lay-out of the houseplans as well as by the quantity of cattle bones. The horse is well represented, too, but fifty percent of the horse bones come from foundation deposits.207 A terra sigillata bowl placed on a handmade pot in a posthole next to the southwest entrance of one of the houses has been interpreted as a deposit related to the abandonment of the house.208 The settlement has not yielded any statuettes. Drenthe: find-spots of statuettes without contextual data Find-spot

Material

Subject

Reference

Beilen

bronze

triple-horned bull

Zadoks 1967,nr.46209

Zuidlaren

bronze

Jupiter

Zadoks 1967, nr. 14

Overijssel • Colmschate An extensively examined and published settlement is Colmschate. Excavations in 1984 and 1985 at Colmschate-De Scheg yielded a Roman period settlement and two burial sites.210 Excavations at Colmschate-Skibaan between 2001-2004 at the west side of the same site, uncovered more house plans of the same settlement, which was occupied between the last half of the 1st and the end of the 4th century.211 Two of the nine coins found at Colmschate-Skibaan were struck in London, by the usurpers Carausius (286-293) and Alectus (293-296). Since they used Germanic auxiliaries to maintain their power, Hermsen suggests the presence of a veteran at Colmschate-Skibaan.212 The settlements at Colmschate may have developed because of their location at a crossroads of a north-south route along the IJssel valley and an east-west route.213 At Colmschate-Skibaan, a bronze statuette of Victoria was found in a posthole belonging to the southern wall of a house, east of the entrance.214 The house was dated to the second half of the 3rd century.215

204

Van Es 1967.

209

See for triple-horned bulls also 3.26.2.

205

Hiddink 1999,140; Van Es 2012, 18.

210

Verlinde 1987, 179-183; Verlinde 1991, 172-179; Verlinde/

206

An overview of the debate extends beyond the scope of

207 208

Erdrich 2006, 271-382.

this research, but Van Es, the excavator of Wijster, as well

211

Hermsen 2007, 255-256.

as Hiddink, have expressed some reticence towards this

212

Hermsen 2007, 251.

interpretation: Hiddink 1999, 140 and note 838.

213

Van Beek 2011, 383, 441 and fig. 12.12.

Van Es 1967, 404.

214

See 4.2.2.

Van Es 1967, 176-177 and plate 23, Van Es 2012, 22-24, Van

215

Hermsen 2003, 65-71; 2007, 55.

den Broeke 2015, 95.

35

• Dalfsen In 1960, traces of a late Roman period settlement were uncovered at Dalfsen. The settlement was situated on the north bank of the river Vecht, three kilometres to the east of the modern village of Dalfsen. The settlement, its size probably comparable to Roman period Colmschate, was occupied in the 3rd and 4th centuries.216 A bronze statuette of Mercury was discovered just outside the settlement, in a downward sloping layer of blown sand and river deposits in the river bank. The layer also contained animal bones, sherds of a native pottery vessel and Roman potsherds.217 It is not unlikely that the large bronze statuette was a ritual deposit, made in the late 4th century.218 • Heeten In 1994, an excavation at the south side of Heeten revealed traces of a large Roman settlement with an enclosure, covering a surface of 0.8 hectares. Inside the enclosure seven house plans, granaries, sunken huts and animal burials were uncovered. The occupational period of the settlement extended from the middle of the 2nd into the 5th century.219 The number of kilns and an exceptional quantity of iron slag round the settlement suggest iron production on a large scale.220 Hiddink notes that farming, too, was important for the local economy, since all primary buildings at Heeten are longhouses.221 After the analysis of plant and animal remains it was concluded that 4th -century Heeten had no ordinary subsistence economy. It was also noted that some of the more than twenty animal burials of cattle, horses and red deer may have had a ritual meaning.222 Two deposits of other objects were recovered from postholes of buildings. A ceramic situla was found in a posthole of the easternmost house. The situla was placed at the bottom of the posthole before the post was put into place.223 Two statuettes come from Heeten: a bronze figurine of a boar was discovered in the top soil at the location De Telgen.224 On the site Boeteler Enk, a fragment of a lead statuette, an arm, was also found in the top soil.225 Since the arm is covered by a long-sleeved garment, the statuette could represent the eastern deity Attis, an indigenous deity such as Intarabus, or a human figure.226 Besides civilians also soldiers could wear long-sleeved tunics, as can be seen in the relief on the tombstone of Titus Flavius Bassus from Cologne.227 Northern Gelderland In the middle and late Roman period, the quantity of imported goods in the eastern Netherlands varies considerably. The proportion of wheel-turned pottery to handmade pottery in settlements like Colmschate is 4 to 10%, versus 10 to 20% in the area 20 kilometres north of the Rhine, in the settlements Ede-Bennekom, Ede-Veldhuizen, Didam and Wehl.228 In the northern and western coastal area beyond the border of the Empire, as well as in the largest part of Drenthe, the proportion of wheel-turned to handmade pottery seldom exceeds 1%, whereas it varies south of the limes between 50 and 100%. It has been argued that Roman military presence along the Rhine prevented free trade between peoples north and south of the limes.229 It 216

217

Oral information by Wim van Es, who at the time super-

223

Verlinde/Erdrich 1998, 702.

vised the excavations at Dalfsen.

224

Van der Velde/Prangsma 2002, 14 and fig. 8.

Van Es 1960, 245-247, 1964, 167-173. The settlement has not

225

Van der Velde/Kenemans 2002, 19 and fig. 10.

been fully published. A preliminary publication saw the

226

See for Intarabus Faider-Feytmans1979, nr.52; for Attis,

light in 1964: Van Beek/Van Es 1964.

for example, the statuette from, presumably, Margraten

218

See 4.2.2.

(Limburg) which is nowadays part of the collection of the

219

Hiddink 1999, 142.

220

The iron was obtained from bog iron.

227

The tombstone is on display in the RGM, Cologne.

221

Hiddink 1999, 142.

228

Taayke 2003, 9.

222

Lauwerier et al. 1998-99, 178-190.

229

Hermsen 2007, 249-250.

36

Limburgs Museum, inv. nr. L 04380.

is also possible that people north of the Rhine, for unknown reasons, were less interested in Roman pottery, although the example of imported terra sigillata sherds suggests otherwise. To my knowledge, there is no written evidence of Roman authorities actively preventing trade between people south and north of the Rhine, but from time to time they may have controlled the flow of goods across the limes into Germania Libera.230 The higher percentage of imported goods in settlements closer to the Rhine, like Ede, Bennekom and Didam, could be a consequence of their proximity to settlements across the Rhine in the limes zone. • Didam-Kollenburg Excavations from 1995 onwards at Didam-Kollenburg have revealed traces of Iron Age occupation as well as a Roman period settlement. Fourteen one-aisled houses were uncovered, including an exceptional large building measuring 30-40 metres in length and 9.5 metres in width. The Roman period settlement was in use from the 3rd to the 5th century.231 Many iron objects have been recovered from the site, including a hammer, possibly belonging to a blacksmith. Fragments of crucibles, an unfinished bronze clasp and crucibles with traces of gold are indicative of bronze casting and gold working. A bronze Minerva and a bronze attachment came to light in the settlement. The attachment depicts a boar and once decorated an object, probably a helmet. The Minerva statuette is a detector find. • Ede-Bennekom A Roman period settlement at Ede-Bennekom was discovered during preparatory works for the construction of houses in a residential area. Excavations in 1970-1971 revealed traces of 29 buildings. The earliest houses resemble the houses of Wijster.232 The first enclosures appear in the second half of the 3rd century.233 Imported pottery and local ware point at a habitation period between 125-150 and 400-425.234 A bronze statuette of Fortuna was found between houses in the settlement and has been interpreted as a building sacrifice.235 • Ede-Veldhuizen The first traces of Roman period habitation at Ede-Veldhuizen were uncovered during preparatory construction works in 1968. Excavations in 1971 and 1972 produced traces of hundreds of farmsteads, outbuildings, granaries, wells, enclosures and pits in an area of 7.2 ha.236 Eventually, almost 180 house plans of farmsteads could be identified.237 The handmade and wheel-thrown pottery spans a period from the middle of the 1st into the 5th century, the majority dating to the 3rd and 4th centuries. Just over 14 percent of the pottery was imported.238 Two bronze statuettes, a Minerva and a human figure holding a patera, were found together in a shallow pit filled with black soil, above a well inside the settlement.239 The well was in use in the 3rd century. After it had fallen out of use the well partially collapsed and became filled with refuse, a layer of charcoal and sand, covered by a layer of black soil. In the 5th century, the statuettes were probably deposited here in a pool of water, since the area had become very wet.240 230

It has been argued that in Cisalpine Gaul, Roman author-

quadrant G-10, but no more details are known on the find

ities exercised control by periodically allowing or refusing

circumstances. See for the interpretation of the find: Van

access to goods, for example by conceding annual or sea-

Es 2012, 17. See also fig. 2.41 and 4.2.2.

sonal fairs: Haeussler 2013, 151.

236

Taayke et al. 2012, 21-22.

The other buildings were 14-25 metres in length and 6-8

237

Taayke et al. 2012, 60-77.

metres in width: Koster et al.2001, 3-20.

238

Taayke et al. 2012, 127.

232

Hiddink 1999, 141 and note 844.

239

Figs. 2.42-46a, 3.17 and 4.2.2. See for the three bronze stat-

233

Van Es et al. 1985, 543-545.

234

Van Es et al. 1985, 623-631.

235

Van Es et al. 1985, 611. The statuette was probably found in

231

uettes from Bennekom and Ede-Veldhuizen also Zadoks/ Peters 1976, 163-167. 240

Taayke et al. 2012, 128; Zadoks/Peters 1976, 163.

37

• Wehl-Oldershove Prior to the construction of a rest home at Oldershove excavations were carried out in 2002-2003. Traces of the workshop of a goldsmith and bronze caster were uncovered, as well as traces of habitation from the 2nd into the 6th or 7th century.241 The finds included imported Roman, Anglo-Saxon and Scandinavian pottery, crucibles and other evidence of bronze casting and gold and silver working. Also, the head of a terracotta Venus came to light(figs. 2.47-49).242 The statuette head was found in a pit west of the smith’s workshop. Besides the statuette head, the pit contained over 1.5 kilo of potsherds dating to the 3rd and 4th centuries, iron nails, iron slag, a large fragment of a Roman roof tile and a fragment of an iron window grill.243 The contents of the pit were deposited in the 3rd or 4th century.244 Northern Gelderland: find-spots of statuettes without contextual data Find-spot

Material

Subject

Reference

Heerde

bronze

Venus

Van Dijk 2015245

Westervoort

bronze

Victoria

Detector Magazine 136, 2014246

Zutphen

bronze

Fortuna

Private collection247

2.3

The area south of the Rhine: Gallia Belgica and Germania Inferior

2.3.1

Theories on post-conquest settlement development in Lower Germany

The region south of the Rhine comprises the provinces of South Holland and Zeeland in the west, North Brabant in the southwest and south, Utrecht in the centre, southern Gelderland in the east, and Limburg, the most southern province within the Netherlands, in the southeast. The variety of landscapes and soil types has led to considerable differences in the development of settlements, house building and the transformation of religion and cultural identities after the Roman conquest. Theories have been advanced to account for these differences in the northwestern provinces of the Empire. A key element of one of these theories, hereafter referred to as ‘the villa theory’, is the distinction between villa landscapes and non-villa landscapes.248 According to this theory, villa landscapes were situated in the fertile loess belt that runs from the Cologne region through Dutch Limburg and central Belgium into northern France. Villa landscapes were dominated by villa settlements, with an economy depending on arable farming. The surplus of grain produced by farmsteads in the loess belt for the Roman military troops along the Rhine enabled the inhabitants to accumulate capital. Therefore, these landscapes are also called ‘landscapes of opportunity’. Non-villa landscapes were regions with clay, peat or sandy soils, dominated by small post-built farmsteads, with an economy based on stock farming. 241

Prangsma et al. 2003, 31.

the modern village of Heerde. The find-spot is very likely to

242

Prangsma et al. 2003, 24-26, 28-29; Prangsma et al. 2004,

be secondary, since soil was brought here from a location

13-14.

east of the find-spot, de Groote Wetering, in the river basin

243

Prangsma et al. 2004, 9, 13-14, 19, figs. 11-12.

of the IJssel, a strongly meandering river: Van Dijk 2015, 7.

244

See also 4.2.4.

245

Fig. 2.50. The statuette is 310 mm in height and was found by a detectorist on farmland between the river IJssel and

38

See also 3.4.1 and 4.2.2. 246

Detector find, southeast of the city of Arnhem.

Map 6. Find-spots of statuettes south of the Rhine Base map: paleogeographic reconstruction of the Netherlands around 100 (Vos/Vries 2013).

247

Fig. 2.51. North of the village of Warnsveld, a detectorist

248

Roymans/Derks 2011.

found the bronze Fortuna in a field just north of a small tributary of the river IJssel. I thank Louis Swinkels for the information on the statuette. See also 4.2.2.

39

Another element of the villa theory is the presumed failure or success of urban communities. Urban communities in the non-villa landscapes were thought to be less successful than their counterparts in the Gallic hinterland. Nico Roymans added an ideological component to the theory: the reluctance of native elites to accept Roman culture and ideology. David Mattingly calls non-villa landscapes ‘landscapes of resistance’.249 Situated in the non-villa landscapes of Lower Germany were the civitas capitals of the Batavi and Cananefates. Successful urban communities were situated in the fertile loess belt: Cologne in modern Germany, Tongeren in modern Belgium and Maastricht and Heerlen in Dutch Limburg. However, as a consequence of new archaeological data, the distinction between villa- and non-villa landscapes has proven to be too clear-cut. Elements of the theory have been refined or still need refining. A first example is the concept of ‘villa’. Besides being significantly different from Roman period villas in the Mediterranean, the villas in the northwestern provinces demonstrate a wide variety in layout, size, and organisation. Therefore, Roymans and Derks have proposed a broad villa definition for the northwestern provinces, according to which ‘single farmsteads with a stone built main house’ is synonymous with ‘small villa’.250 Other elements of the villa theory that have been revised are the economic and social organisation and the diversity of house construction and settlement layout within villa landscapes.251 Recently, research has revealed a substantial number of non-villa settlements in the central part of the civitas Tungrorum, where cattle-breeding appears to have been prevalent.252 By using GIS applications, Karen Jeneson has investigated the loess landscape between the rivers Meuse and Rhine. Her research was aimed at reconstructing the Roman landscape by analysing the spatial order, settlement lay-out, settlement density and settlement patterns.253 Although it was expected that the landscape of the loess belt between Tongeren and Cologne was dominated by large villas, her study has revealed an unexpectedly dense population living in small post-built settlements.254 As to the non-villa landscapes, several buildings that were uncovered during excavations in Rijswijk (South Holland), the Kromme Rijn region (Utrecht), Hoogeloon (North Brabant), Wijchen-Tienakker, Druten-Klepperhei and Ewijk-De Grote Aalst (Gelderland) are villas according to the broad villa definition proposed by Roymans and Derks. They possess ‘villa characteristics’ such as stone-built structures, hypocausts, stucco, mosaic floors and mural paintings.255 It has been argued that in non-villa landscapes ‘traditional settlements with post-built houses make up the majority of rural sites (more than 95 %)’.256 This percentage might need some adjusting, since in the Kromme Rijn region alone, around one hundred settlements have been defined, eight of which might have been villa sites, while only a few settlements have been excavated.257 And lastly, the success or failure of settlements depends on a complex set of factors, the presence of fertile soil being just one of them. Water wells, navigable waterways and the vicinity of routes over land were determining conditions, too, for the continuation of habitation. Although the civitas capital of the Batavi was abandoned after 270, archaeological evidence has demonstrated that habitation continued on the left bank of the river Waal.258 It has even been postulated that Nijmegen kept its function as a regional centre.259 The Roman period settlement at Maastricht developed on the banks of the Meuse, whereas Heerlen and Tongeren were situated on important routes over land. In the 1990s, Roymans added a second ideological component to the villa theory: the pastoral mentality and martial values of the Lower Rhine population and the Batavi in particular, whose territory more or less 249

Enckevort 2012, 217-249.

See, for instance, Roymans 1995, 47-64; Mattingly 2006, 369, 385, 522-524.

256

Roymans/Derks 2011, 3.

250

Roymans/Derks 2011, 2-3.

257

Vos 2009, 46-48.

251

Roymans/Derks 2011, 33-37.

258

An important reason for not rebuilding the city at the same

252

Vanderhoeven 2015, 192-195.

location may have been flooding by the river Waal, which

253

Jeneson 2013, 9.

eventually changed its course and washed away a part of

254

Jeneson 2013, 155-156, fig. 537, 224-227.

the city.

255

Bloemers 1978, 51-52, 312-313, 332-343; Vos 2009, 44-54; Van

40

covered the eastern Rhine delta.260 He argued that the high valuation of cattle is reflected by the tradition of sheltering men and cattle under one roof, and the large number of ritual deposits of cattle and horse bones in and around farmhouses.261 The theory on differences between non-villa and villa landscapes also extends to the connection between cosmology and pastoral or arable farming, as has been postulated by Derks. According to Derks, local elites chose a tutelary deity for their civitas or pagus whose forces and abilities best fitted the needs and values of the community. He has argued that dedications to Hercules Magusanus primarily come from regions where pastoral farming was dominant, whereas dedications to Mars predominantly come from areas oriented towards arable farming.262 In chapter 3, it is advanced that Hercules/Hercules Magusanus was venerated by soldiers and veterans and not by Batavians in particular.263 Whether Mars was worshipped in the study area as bringer of fertility or as the god of war is difficult to assess. It is possible that the inhabitants of the study area not only appreciated his martial appearance, but also knew of his capacities as bringer of fertility. Be it as it may, the presumed connection with arable farming is not confirmed by the find-spots of the Mars statuettes in the study area. Nearly all specimens have been found in the northern coastal region, the eastern river area and the Kromme Rijn area, where pastoral farming was dominant. While the dichotomy between villa and non-villa landscapes is crumbling, one wonders to what extent the ideological components of the villa theory are still valid. 2.3.2

South Holland and Zeeland

South Holland Behind a belt of coastal dunes, the province of South Holland predominantly consists of wet moorland with peat and clay soils. The limes runs via the modern cities of Alphen aan den Rijn, Leiden and Valkenburg towards Katwijk on the North Sea coast. Excavations revealed a considerable number of late Iron Age and Roman period settlements on peat and clay. The territory of the Cananefates was situated in South Holland. Like many tribesmen, Cananefates served in the Roman army, but apart from that, little is known about them. According to Tacitus, ‘The Cananefates live in part of the island; in origin, speech, and courage they are equal to the Batavi, but inferior to them in number’.264 His remark has provoked a debate on the Frisian or Batavian origins of the Cananefates, which is based on the differences in pottery styles.265 Rien Polak and Laura Kooistra have advanced that, instead of arriving at the same time in the study area, as generally is believed, the Cananefates may have arrived later than the Batavi, or initially lived east of Utrecht, too, and moved later to the western coastal area.266 Jasper de Bruin has suggested that the Cananefates were a pluriform group from the western coastal area, mainly north of the Rhine. They settled in South Holland at the same time as the Roman troops, or slightly later.267

259

Van Enckevort/Thijssen 2014, 31-37; Hendriks et al. 2014,

The Batavi settled here after they had separated themselves

37-45. 260

See chapter 1, note 4.

261

Roymans 1995, 55.

262

Derks 1998, 100-115.

263 264

from the Chatti. 265

See for the Batavian origins of the Cananefates, for instance, Roymans 2004, 205-208; Lanting/Van der Plicht 2009/2010, 55.

See 3.8.5.

266

Polak/Kooistra 2013, 389-392.

Tac. Hist. 4.15.1: Ea gens partem insulae colit, origine lingua

267

De Bruin 2019, 147-156.

virtute par Batavis; numero superantur. Tacitus means the island formed by two arms of the Rhine and the North Sea.

41

Map 7 Find-spots of statuettes in South Holland and Zeeland Base map: paleogeographic reconstruction of the Netherlands around 100 (Vos/Vries 2013). 1. Domburg 2. Monster 3. The Hague-Waalsdorpervlakte 4. Katwijk 5. The Hague-Ypenburg 6. Koudekerk aan den Rijn 7. Ellewoutsdijk 8. Colijnsplaat 9. Ouddorp 10. Rockanje 11. Schipluiden 12. Schiedam 13. Wateringen 14. Rijswijk 15. The Hague-Ockenburgh 16. The Hague-Erasmusweg, Wateringse Veld and Nikkelwerf 17. The Hague-Scheveningseweg 18. Leiden 19. Zwammerdam 20. Alblasserdam 21. Aardenburg 22. Naaldwijk 23. Voorburg 24. Valkenburg 25. Alphen aan den Rijn

42

Epigraphic evidence testifying to the presence of Cananefates in South Holland comes from two milestones found at Monster and Rijswijk, and three fragments of a military diploma from Poeldijk.268 The diploma dates between 160 and 167 and belonged to a Cananefatian soldier who had served in the ala I Noricorum. In 1991, a research project on the Roman period landscape in Midden-Delfland started as a follow-up to the 1970s excavations at Rijswijk-De Bult, in the Midden-Delfland area. Midden-Delfland covers a surface of around 6.700 hectares and is situated between the modern towns of Rijswijk, Delft, Rotterdam and Vlaardingen. Starting point for the project was the supposition that the development of Rijswijk-De Bult was typical for other Roman period rural settlements: from a single isolated structure at the beginning of the 1st century to a complex settlement structure over time.269 To this end, a comparison was made between sites at Assendelft, situated north of the Rhine in North Holland, and sites in the Midden-Delfland area south of the Rhine. It became apparent that the sites had little in common.270 Only the lay-out of the ditch system of the Midden-Delfland sites, the Assendelft sites and Rijswijk bear a strong resemblance to each other. Heleen van Londen has argued that the Roman military was responsible for the layout of the ditch system.271 • Alphen aan den Rijn-De Schans Before the construction of a business park at the location De Schans, preliminary archaeological investigations were carried out in 1988-1990. A stretch of a Roman road and a system of ditches were discovered, dating to the 2nd and 3rd centuries. A large quantity of Roman period pottery sherds, imported and indigenous ware, building material and several hearths indicated that a settlement was situated here, but no traces of post-holes were found. It is argued that the settlement may have had a military character.272 A large, complete terracotta rattle in the shape of the head of a boy, possibly representing Bacchus or Amor, comes from a ditch parallel with the Roman road (fig. 2.186). Besides potsherds, the filling of the ditch comprised coins, fibulae, bronze and metal objects, glass, whetstone fragments and fragmented terracotta statuettes.273 • Naaldwijk-Hoogeland During excavations in 2007-2008 at the location Hoogeland, habitation traces have been found from the Roman period into the Medieval period. The occupation began around 75. In the second phase, between 100 and 125, the settlement comprised at least four farmsteads. Although the changing lay-out and organisation of the settlement suggest a Roman influence, native building traditions and material culture prevailed. Contrary to several other settlements in the area, the settlement at Hoogeland did not consolidate its growth and decline set in. In the last phase, between 190 and 225, the settlement consisted of one house with a porticus.274

268

The inscription on the milestone from Monster/Naaldwijk

Fe[l]ic(i) In[v]ic(to) Au/[g(usto)] pont(ifici)/[ma]x(imo)/

reads: Imp(eratori) Caes(ari)/M(arco) Aurel(io) Anto/nino

[t]rib(unicia)pot(estate) co(n)s(uli)/[I]I p(atri) p(atriae)

Aug(usto) pont(ifici)/max(imo) tr(ibunicia) pot(estate)

proco(n)s(uli)/[c(ivitas)] Cananefatu(m) ab [Hadr(iani)]

XVI/co(n)s(uli) I[I] et/Imp(eratori) [C]aes(ari) / L(ucio)

F(oro)/leug(a): CIL 17.2, 587 = AE 1965, 118.

Aurel(io) [Ve]ro Aug(usto) / tr(ibunicia) pot(estate) [II]

269

Van Londen 2000, 172-173.

co(n)s(uli) II/ a M[---] A E C/m(ilia) p(assuum) VII. The

270

Van Londen 2000, 174-176.

letters M.A.E.C. refer to Municipium Aelium Cananefati-

271

Van Londen 2000, 177-180.

um. The figure VII is uncertain. It has been interpreted as

272

Brouwer 1989, 326-327.

XII, but must have been between V and X. The inscription

273

Kleiterp 1991, 342-343. Apart from the terracotta rattle, the other terracotta statuette fragments are inaccessible. See

seems to have been recut: CIL 17.2, 588 = CIL 13, 9165

also 4.2.2.

=AE 2006.913. See also Bogaers 1960, 307, note 235. The inscription on the Rijswijk milestone reads: [Imp(eratori)

274

Goossens 2012, 521-528.

Caes(ari) C(aio)]/[M]essio Quin/[to] Traiano/Decio[P]io

43

Among the metal finds are a bronze, bare left arm of an almost half life-size statue, two fingers of a statue, an Amor applique, a fragment of a Corinthian helmet and a statuette of a semi-nude young woman (figs. 2.52-54). She holds a jar upside down in her right hand, while pressing the water out of her hair with the left. The statuette was found in the filling of a well that cut through a Roman period ditch. According to the excavators, in the last phase of the occupation period of the settlement, the inhabitants collected metal finds and building material from another, abandoned location for recycling.275 If so, this raises the question why the large bronze statuette was not recycled and ended up in the well in the early Medieval period.276 According to the excavators, the statuette depicts Venus. The woman may also represent a water nymph.277 It has been suggested that the figurine and the Amor applique could have been fixed onto a household shrine.278 This is very unlikely and parallels for this practice are not known. Numerous Amor appliques have been found in and outside the study area, once attached to furniture, vessels, supports, lamps and candelabras. Statuettes were set up in shrines, on tables or in niches, but not soldered onto household shrines. The water nymph/Venus figure is an attachment that once decorated the edge of a large object. The bronze fragment of the Corinthian helmet has been interpreted as the helmet of Mars.279 It could also have belonged to a statuette of Minerva, since this goddess is also depicted wearing a Corinthian helmet. • Naaldwijk- ’t Zand Heultje The site ’t Zand Heultje is located east of Hoogeland. Habitation started around the middle of the 1st century and ended somewhere in the 3rd century. In the vicinity of the settlement the presence of a road and Corbulo’s canal are suspected.280 Three main buildings, two outbuildings, five wells and several pits were uncovered. At the end of the 1st century there was an increase in imported goods.281 Stone built structures seem to have been absent in the settlement. Striking finds from ’t Zand Heultje are the relatively large number of Dressel 20 amphorae, a terracotta fragment of Venus and an unidentifiable fragment of a human figure or deity. The Venus fragment was found together with potsherds. It depicts a part of the lower body, the genitals and upper legs of a nude woman. It comes from a well belonging to a house that could be dated to the first half of the 2nd century.282 • Ouddorp In 1958-59 and 1982, Jan Trimpe Burger conducted excavations near Ouddorp, on the island of Goeree-Overflakkee, which were published in 2012.283 On the site Oude Oostdijk, a Roman period settlement of around 4.5 hectares was occupied between 80 and 225.284 As early as the late 1st century, the harbour settlement must have been an important centre within a trading network. The buildings in the settlement were constructed according to native and Roman building techniques.285 It has been argued that the settlement was the civitas capital of the Frisiavones.286 The layout of the settlement, a stamp mentioning the Classis Germanica and graffiti on pottery testify to a military presence in the settlement. 287

275

Goossens 2012, 95.

283

De Bruin et al. 2012.

276

Goossens 2012, 130-131, 295 and figs. 5.67,7.13-7.15. It appears

284

De Bruin et al. 2012, 123-124.

from fig. 5.67 that the attachment was found in filling 4,

285

De Bruin et al. 2012, 45-46.

under the two layers that closed off the well.

286

De Bruin et al. 2012, 125-128, 145-150. It has also been sug-

277

See also 3.8.7 and 4.2.3, 4.2.5.

gested that Cuijk may have been the civitas capital of the

278

Goossens 2012, 295.

Frisiavones: Haalebos et al. 2002, 30. See for Cuijk also this

279

Goossens 2012, 295.

280

Van der Feijst 2012, 10.

281

Van der Feijst 2012, 43.

282

Van der Feijst 2012, 57-58 and fig. 4.31.

44

chapter. 287

De Bruin et al. 2012, 105-108.

Fragments of three terracotta statuettes have come to light: the head of a water nymph, a bust of a girl, and three fragments of a base.288 The water nymph comes from a ditch in the settlement.289 • Rijswijk-De Bult A key site in South Holland is Rijswijk-De Bult, which was excavated in the 1960s and published by Tom Bloemers.290 In the 1st century the settlement consisted of one farmhouse. In the 2nd century, the settlement grew to three farmsteads and a building with a granary, fenced off by a palisade. By then, the settlement was surrounded by a rectangular ditch system. In the same period, there was a strong increase in imported goods. In the 3rd century, the settlement expanded from 1-1.5 to 1.85 hectares.291 The post-built-structure was replaced by a stone-built house between 210 and 240 and renovated at the westside between 240 and 270. One of the rooms had mural paintings and a hypocaust.292 A buried bowl (type Oelmann 104) came to light in the southeast corner of the foundation ditch of the house. It has been interpreted as a building sacrifice.293 Two fragments of a terracotta bird and two terracotta sherds of an unidentifiable, draped figure were recovered from the site.294 • Rockanje At Rockanje, a considerable number of finds in 1950 prompted a small-scale excavation in the next year, during which traces of buildings with a stone footing were uncovered.295 The finds from 1950 included terra sigillata sherds from the 2nd and 3rd centuries, Roman coins from the same period, a few bronze objects, large tuff blocks, many roof tile fragments, fragments of tubuli and hypocaust tiles, as well as the head of a terracotta statuette of a mother goddess.296 The head belonged to a type of mother goddess statuettes that depict an enthroned woman suckling two infants.297 • Schiedam-Kethel Some 15 kilometres south of Rijswijk, excavations at Schiedam-Kethel in 1961 produced traces of a 2nd-century farmstead with an outbuilding and a granary.298 Among the finds was a complete terracotta figurine of a horse.299 • Schipluiden-Harnaschpolder Schipluiden is situated in the centre of the Midden-Delfland region, not far from Rijswijk-De Bult. Archaeological research between 2001 and 2003 revealed traces of two Roman period settlements. The rectangular ditch system resembled that of other native settlements in the region. The northernmost settlement was occupied during the 2nd century, measuring at least 90 by 80 metres. 300 However, besides two large clusters of pits, no traces of buildings were found. Some pits contained special deposits like complete pottery vessels and animal skeletons.301 In the southernmost settlement traces of three buildings were identified.

288

Van Boekel 1987, nrs. 125, 155, 191, 268. The identification of

294

Van Boekel 1987, nrs. 166, 218, 228.

semi-nude females with a jar is problematic, which is why

295

Bogaers 1952, 5-8.

Van Boekel has categorised this type of representations as

296

Van Boekel 1987, nr. 89.

‘water nymph or Venus’. See also 3.8.7.

297

Bogaers assumed it was the head of Venus: Bogaers 1952, 7.

See for the interpretation 4.2.2.

298

Modderman 1973, 149-158.

290

Bloemers 1978.

299

Van Boekel 1987, nr. 242.

291

Bloemers 1978, 46-54, 114-117.

300

Goossens 2006, 7-78, 370.

292

Bloemers 1978, 51, 312-313, 332-343; De Bruin 2017, 226.

301

Goossens 2006, 97-128.

293

Bloemers 1978, 183; De Bruin 2017, 225.

289

45

Among the finds from the northernmost settlement is the head of a terracotta statuette. It probably depicts the head of Venus, but the face has been ‘mutilated’ by a deep, horizontal scratch where the mouth can be expected, a scratch over the eyes and one over the chin. The head comes from a disturbed find-spot in the southeast zone of the settlement. In 1989, a few fragments of terracotta statuettes had been found forty metres to the northwest.302 • The Hague-Erasmusweg In the 1990s, excavations along the Lozerlaan revealed a Roman period settlement and an open-air sanctuary at the junction with the Erasmusweg. The settlement was inhabited from the end of the 1st into the 3rd century. A bronze jug dating to the 2nd or 3rd century (type Eggers 128) was deposited a couple of metres south of the southern corner of the sanctuary, at a corner consisting of two ditches.303 The middle part of a terracotta statuette was found near the road to Wateringen. It depicts a mother goddess.304 • The Hague-Nikkelwerf In 1999, traces of a native settlement were uncovered during an emergency excavation. The occupation period could be dated to the 2nd century.305 The excavation has yielded the head of a terracotta bust of a woman, which has been erroneously interpreted as a mother goddess.306 Examples of nearly identical busts have been found at Nijmegen and Vechten.307 • The Hague-Wateringse Veld The site Wateringse Veld is situated at the southwest side of The Hague, bordered to the north by Erasmusweg. Excavations between 1996 and 2003 have yielded 26 house plans and 29 annexes of a Roman period settlement, 14 Roman period graves and four milestones from the 2nd and 3rd centuries.308 The milestones stood along the Roman road which ran between Forum Hadriani and the river Meuse. Roman period habitation began around the middle of the 1st century, coinciding more or less with the construction of Corbulo’s canal in 47 and stretched until the middle of the 3rd century.309 Farmers settled on the higher grounds. In the second quarter of the 2nd century, rectangular ditch systems appeared, possibly designed with the help of or ordered by Roman authorities, an assumption based on the use of the actus, a Roman unit of measurement, which equals 35.5 metres. Besides five unidentifiable terracotta statuette fragments, two fragments of two mother goddesses came to light. One fragment was recovered from a ditch, together with 977 pottery sherds, of which the majority consists of imported ware from the 2nd and 3rd centuries.310 • Voorburg-Arentsburg: Forum Hadriani or Municipium Aelium Cananefatium The first finds from the Roman period settlement have been reported in the early 16th century. Already in the 17th century it was suspected that the Roman period settlement Forum Hadriani, indicated as Foro Adriani on the Peutinger map, was located here. The first controlled excavations were carried out by Reuvens between 1827 and 1834 at the Arentsburg estate. Reuvens discovered stone foundations and a bath complex, but died before he could publish his findings. Based on excavations between 1908 and 1915, Holwerda concluded that the site was not Forum Hadriani, the civitas capital of the Cananefates, but a castellum. Half a 302

Reigersman-van Lidth de Jeude 2006, 145-146.

306

Van Zoolingen 2011, 25 and fig. 15.

303

Van Zoolingen 2011, 13-17.

307

See figs. 2.103-104a; Van Boekel 1987, nrs. 196-200.

304

Van Boekel 1987, nr. 80.

308

Two graves come from the site Juliahof.

See for an inventory of find-spots in the area Veen/Waas-

309

Siemons/Bulten 2014, 453-454.

dorp 2000, 46, find-spot nr. 43.

310

Siemons/Bulten 2014, 407-408.

305

46

century later, in 1966, Bogaers conducted an excavation at the site where Reuvens had dug his trenches and ascertained that the settlement had to be Forum Hadriani.311 Plans for new construction works in the south-eastern part prompted an archaeological survey through drillings in 2000-2001. The survey demonstrated the presence of, at least partly, undisturbed Roman period structures. Excavations in 2005 produced several ditch systems, wells, traces of a stone-built structure and a large quantity of finds which could be dated between 120 and the third quarter of the 3rd century. The excavators concluded that the orientation of ditches, buildings and parcels in the southern part of the settlement did not correspond with those situated in the northern and western part, where Reuvens and Holwerda had been excavating. Therefore, they suggested that the southern part was part of another settlement.312 The discovery of a harbour in 2007-2008 supported this idea.313 Therefore, the authors of the publication use the more neutral, modern toponym Voorburg-Arentsburg for this part of the settlement instead of Forum Hadriani. Voorburg-Arentsburg cannot be linked with certainty to Forum Hadriani, but one of the four milestones from the site Wateringse Veld dates to 151 and links the site archaeologically to Municipium Aelium Cananefatium.314 As a consequence of the scarcity of traces that can be dated before 120, little is known about the earliest phase of the settlement. Since the handmade vessels resemble specimens from native settlements in the area like Rijswijk-De Bult, it is assumed that the earliest settlers were Cananefates.315 Between 120 and 151, the settlement became the Cananefatian civitas capital. Its centre probably was located where Reuvens had found the baths. In this period, the orientation of the houses and the layout of the settlement changed, now showing a classical street pattern of rectangular insulae with strip houses, the short side parallel to the road. Public buildings, a central bath complex and a public porticus made their appearance. Gradually, post-built structures were replaced by stone-built houses. Between 160 and 170, the settlement was enclosed by a palisade and rampart, which later was replaced by a stone-built defensive wall with gates. In the third quarter of the 3rd century, growth of the town stagnated, but habitation probably continued into the 4th century.316 As can be expected of a civitas capital, Voorburg-Arentsburg was an important economic, social and administrative centre, which is emphasised by the discovery of a gully that was transformed into a harbour.317 Reuvens already noticed a remarkable assemblage of finds in one of the almost sixty wells. The assemblage comprised several metal objects, including coins, bronze handles of buckets or containers, hooks and a key.318 Other remarkable assemblages found in the top filling or at the bottom of wells are interpreted as offerings related to the beginning or end of the well’s life cycle. From the harbour, too, several remarkable deposits have been recovered, like a complete pottery vessel containing hazelnuts and other organic material. Other deposits include a complete quern, a human thigh bone with cut marks, ‘pierced pots’, coins and other metal objects.319 One well found during the 2005 excavations contained two complete wagon wheels. In a second well, a whole grind stone had been deposited. Under the stone, two bronze objects and a wooden object were placed. The wooden object has been misinterpreted as a dog or wolf, but it undoubtedly depicts a water creature, a dolphin.320 During the same excavation a small, limestone votive altar came to light, bearing an inscription: I. O. M. So far, no sanctuaries have been found in the settlement. The altar, measuring 10 x 10 x 16 centimetres, may have 311

Bink/Franzen 2009, 14-17; Besselsen 2014, 5-6. See for an

315

Bink/Franzen 2009, 432.

extensive study of the development of Forum Hadriani and

316

Bink/Franzen 2009, 410-417; Besselsen 2014, 6-7.

a reconstruction of the town Buijtendorp 2010.

317

Besselsen 2014, 7-8; Driessen 2014b, 149-172.

312

Bink/Franzen 2009, 413-416, 433-437.

318

Driessen 2014b, 176, fig. 8.10.

313

See for the harbour Driessen 2014a, 97-110.

319

Driessen 2014b, 167-173, 8.5-8.6 and fig. 8.7.

Besselsen 2014, 6-7 and note 15. I follow the authors of the

320

Bink/Franzen 2009, 87, 332-333, fig. 9.20. See also Lange

314

publications on Voorburg-Arentsburg and avoid the name

2017, 264, nr.125. Dolphins in Roman art are often depicted

Forum Hadriani. See for the inscription AE 2000, 1022 = AE

with sharp teeth and a non-naturalistic tail. See for the

2003, 1229 = AE 2006, 913.

interpretation also 4.2.2.

47

been used in a domestic context, although Jupiter Optimus Maximus is usually associated with official state religion. The total number of altars and tablets dedicated to Jupiter by cavalrists amounts to 39. Twenty-two are dedicated to Jupiter Optimus Maximus, sometimes in combination with a second deity. Nine altars were erected by members of the ala I Cananefatium, seven to Jupiter Optimus Maximus and two to Jupiter Dolichenus. This suggests that the supreme god Jupiter had a special meaning for the Cananefatian cavalrymen.321 Very few bronze statuettes have come to light at Voorburg-Arentsburg. They include the infant Bacchus riding an animal, a torso of a nude, seated man, most likely Mercury, a statuette of a greyhound and a stag.322 Terracotta statuettes comprise three specimens of Apollo, one head of Bacchus or Apollo, one lower part of a standing figure, may be Bacchus, two statuettes of Cybele, two of Diana, three to four statuettes of Mercury, one of Minerva, seven Venus statuettes, one water nymph or Venus, one Priapus statuette, nine mother goddesses, one thorn-puller, one statuette that possibly depicts a child, the head of a Risus bust, four pedestals and several other fragments of unidentifiable figures and animals. At least eight of the aforementioned terracotta statuettes have been recovered from the harbour during the excavations in 2007-2008.323 Besides the aforementioned ritual deposits from the wells at Voorburg-Arentsburg, the head of a Risus bust at the bottom of well 8 may have been a ritual deposit, too.324 • Wateringen-Juliahof During excavations in 2006 traces of an enclosed Roman period settlement came to light. The settlement knew two to three habitation phases. In the earliest phase, the farmstead was occupied in the first half of the 2nd century or slightly earlier. The latest phase dates between the middle of the 2nd and the second half of the 3rd century.325 Among the finds from the settlement is the torso of a terracotta statuette of an enthroned female figure, probably Minerva, Juno or Fortuna. The statuette was found in a pit among a concentration of pits that were almost entirely filled with imported potsherds dating after the middle of the 2nd century. The pits have been interpreted as refuse pits.326 Two fitting fragments constitute the pedestal of a second terracotta statuette, probably an animal. One of the two fragments may come from a well.327 South Holland: find-spots of statuettes without contextual data Find-spot

Material

Subject

Reference

Alblasserdam-Kerkstraat

terracotta

woman

Van Boekel 1987, nr. 138

The Hague-Waalsdorpervlakte

bronze

Isis with Horus

RMO328

Koudekerk a/d Rijn

bronze

mother goddess

Dijkstra/Grinsven 2007, 71329

Monster

bronze

bull

Zadoks 1969, nr. 64330

The Hague-Ypenburg

bronze

Hercules

RMO331

321

Bink/Franzen 2009, 251-252.

322

Zadoks et al. 1969, nrs. 9, 72, 78. The torso was found in the Virulylaan at Leidschendam-Voorburg, just south of

that was found in 2005 is identical to the complete specimen from Veldhoven: Bink/Franzen 2009, 213 nr. 1056. 324

323

48

See for well 8 Bink/Franzen 2009, 106-107; for the terracotta statuettes Bink/Franzen 2009, 213; for the interpretation 4.2.2.

Voorburg-Arentsburg. See for the resemblance to Mercury, Zadoks et al. 1967, nr. 25.

325

Eimermann 2009, 5.

Van Boekel 1987, nrs. 2-5, 7, 10 15, 17, 23, 51, 60, 71-72, 86, 90,

326

Eimermann 2009, 61.

92, 103-104, 128, 149, 153-154, 157, 161, 207, 215, 226, 230-231,

327

Meij/Reigersman-van Lidth de Jeude 2009, 107-108.

252, 256-258, 267. A hitherto unidentified terracotta frag-

328

RMO, inv. nr. h1964/4.1.

ment of a human figure or deity (nr. 153) has been re-in-

329

See fig. 2.55, 3.5.4 and 4.2.2. I thank Louis Swinkels for

terpreted by Van Boekel as Priapus. See for this fragment

bringing this statuette to my attention. An almost identical

and the eight statuettes from the harbour Driessen 2014c,

statuette comes from the Hunerberg at Nijmegen: figs.

489-500; Bink/Franzen 2009, 212-214. The Diana statuette

2.56-56a; Zadoks et al. 1969, nr. 38.

Zeeland In the Roman period, the province of Zeeland did not yet consist of the group of islands we know today, and the river Scheldt had its estuary north of Colijnsplaat. Although little is known about the people who lived here, it is assumed that the Frisiavones had their territory in this area. The boundaries of their territory are unknown. Moreover, it is not clear whether they shared their territory with the Sturii and the Marsaci, south of the river. Some argue that only the Frisiavones had their own civitas.332 Pliny the Elder mentions the Sturii in combination with the Marsaci: ‘In the Rhine itself, the most notable island is that of the Batavi and Cananefates, which is almost a hundred miles in length, and others are those of the Frisii, Chauci, Frisiavones, Sturii and Marsaci, which lie between Helinium and Flevum.’333 One of the earliest Roman imports from Zeeland dates to around 70. It is a sherd from a terra sigillata bowl, type Dragendorff 29.334 • Colijnsplaat and Domburg In addition to the Roman fort and vicus at Aardenburg, the province of Zeeland is particularly known for its two Roman period sanctuaries and the Nehalennia altars.335 Stone monuments came to light for the first time in 1647, after a series of severe storms, on the beach of Domburg, situated on the island of Walcheren. The Latin inscriptions and inscribed altars were addressed to a variety of deities, including Neptune, Hercules, Victoria and the native goddess Burorina. The majority, 28, were addressed to the native goddess Nehalennia. It was concluded that once a temple stood at Domburg. Centuries later, in 1970, just off the coast near Colijnsplaat, a fisherman caught fragments of two stone altars in his nets. Wondering how these altars ended up in the Oosterscheldt at a depth of 25 metres, a search expedition set out to fish for more finds. Besides Roman roof tiles and pottery, 104 altars and statues were brought in. Eventually, the total number of altars and statues that came to light at Domburg is 43, whereas 320 have been found near Colijnsplaat. Therefore, it is assumed that a second sanctuary stood at the location where the altars were dredged up from the sea bottom. Both temples were simultaneously in use, from the end of the 2nd until the middle of the 3rd century.336 Despite the large number of altars, Domburg has yielded only one bronze statuette, but the sea probably washed away small objects from the sanctuaries. A terracotta fragment of an unidentifiable figure was found by a diver at Colijnsplaat. The bronze statuette was found on the beach, near the site where the Nehalennia sanctuary once stood. It depicts a genius loci, wearing a mural crown and holding a patera in the right hand.337 Statues and statuettes of Nehalennia are scarce. She is predominantly depicted on stone altars.338 330

Found in 1777, just south of Monster, on the North Sea coast.

331

Found in 2002. See also fig. 2.57. See for a comment on the

337

statuette is lost since the 19th century. The image in Zadoks

statuette Halbertsma 2014, 309, note 29. 332

333

See for instance, Byvanck 1943, 200 and fig. 7; Roymans 2004,

Zadoks et al. 1969, nr. 19; Van Boekel 1987, nr.158. The bronze et al. 1969 is a plaster cast.

338

A stone statue of an enthroned Nehalennia is on display

25 and fig. 3.2. Lanting and Van der Plicht reject the idea that

in the National Museum of Antiquities, Leiden. Only a

the Sturii and Marsaci both had their own civitas: Lanting/Van

drawing exists of a second example: Stuart 2013, 87-89 and

der Plicht 2009/2010, 56-57. See also Willems 1981, 13, fig. 3.

vol. 2, 129, nr. 28H1. The Römisch-Germanisches Museum

Plin. HN 4.15: In Rheno autem ipso, prope c͞ in longitudinem,

in Cologne has a bronze statuette of a seated woman. She

nobilissima Batavorum insula et Cannenefatium, et aliae Fri-

wears a headdress and holds fruits in her lap. She has been

siorum, Chaucorum, Frisiavonum, Sturiorum, Marsaciorum

interpreted as Nehalennia, but this is justly disputed by

quae sternuntur inter Helinium ac Flevum.

Stefan Ritter who argues that the woman is a mother god-

334

Trimpe Burger 1973, 135.

dess: Ritter 1994, 355-256. See for the statuette https://www.

335

See for Aardenburg 2.4.2.

livius.org/pictures/germany/koln-ccaa/cologne-muse-

336

Stuart 1971, 30-31; Stuart 2013, 11, 41-42.

um-pieces/cologne-statuette-of-nehalennia/ (29-12-2016).

49

• Ellewoutsdijk A large Roman period settlement was uncovered at Ellewoutsdijk in 2001, consisting of nine houses on a terrain of almost five hectares. In 1999, a farmstead with several outbuildings had already been unearthed before the construction of the Westerscheldt tunnel. The settlement was inhabited between the second half of the 1st and the middle of the 2nd century.339 Although few imported goods have been found, the presence of spelt (Triticum spelta) and dill (Anethum graveolens) points at contacts with the Romans, who introduced spelt and dill in the northern regions. The excavations have not yielded any statuettes, but a terracotta head of Minerva was found on the bank of the Westerscheldt at Ellewoutsdijk in 1961.340 2.3.3

Utrecht

Map 8 Find-spots of statuettes in Utrecht. Base map: paleogeographic reconstruction of the Netherlands around 100 (Vos/Vries 2013). 1. Vleuten-De Meern 2. Utrecht-city 3. Houten 4. Odijk 5. Cothen 6. Jutphaas 7. Woerden 8. Rhenen* 9.Vechten 10. Wijk bij Duurstede *North of the Rhine, no statuettes have been recovered from Roman period graves or burial sites. The only exception is an early Medieval cemetery at Rhenen that has yielded a Roman period statuette. In 1834 or earlier, a mother goddess figurine was recovered from the western slope of the Rhenense Berg or Donderberg, just north of the Rhine between Utrecht and Rhenen. Excavations in 1951 at the site produced an early Medieval cemetery, dating between the 4th and 8th centuries. The site is the largest known early Medieval cemetery in the Netherlands. Over 1100 graves have been identified, dating between 375 and 750.341 Since a small number of the earliest graves contained Roman coins and terra sigillata vessels, the figurine may have come from one of the earliest graves. The stray find has been recovered from one of three circles of black soil under the top stratum 342

and depicts an enthroned mother goddess with a dog on her lap.343 339

Sier 2003, 181-183.

342

Huiskes 2011, 19-26, 41-48.

340

Van Boekel 1987, nr. 62.

343

Van Boekel 1987, nr. 81.

341

Huiskes 2011.

50

The limes runs through the province of Utrecht from east to west via Wijk bij Duurstede, the modern provincial capital city of Utrecht and Woerden. The landscape south of the limes in Utrecht is dominated by clay and peat. Situated in the mid-south of the province, the Kromme Rijn area is bordered to the north by the Utrechtse Heuvelrug, the city of Utrecht to the west and the river Lek to the south. According to Tacitus, the area was uninhabited at the time of the transfer of the Batavi, in the aftermath of Caesar’s conquest of Gaul.344 Archaeological evidence like La Tène bracelets and late Iron Age coins contradict his assertion.345 The area became the western part of the Batavian territory after the Batavi had split off from the Chatti, who lived east of the Rhine.346 The first large project in this area was undertaken in the late 1970s. The aim was the reconstruction of the habitation history of the Kromme Rijn area from the earliest human occupation up to the late Medieval period.347 Besides excavations, the project focused on mapping and identifying pre-Roman and Roman period sites within the Kromme Rijn area. This resulted in over a hundred sites, of which a limited number were excavated. In his PhD research, Wouter Vos has focused on the results of the project. He studied the spatial development of Batavian settlements by analysing settlement patterns, the social and cultural development of the Batavian communities and the interaction between the Roman military, returning veterans and native communities.348 One particular focus in his study were the excavations in the micro-regions Houten and Wijk bij Duurstede, of which the results are summarized hereafter. • Houten-Tiellandt Within the micro-region Houten, excavations have been conducted at several locations, including Wulven, Tiellandt and Houtense Wetering in the northwest and centre of Houten. The habitation traces at Houten-Tiellandt were too fragmentary and complex to distinguish different phases of habitation. The habitation period covered the 1st and 2nd centuries.349 In the centre of Houten, a stone-built structure came to light, measuring 26.5 by 4.75 metres. The earliest phase probably dates to the Flavian period and was rebuilt in stone. The walls of the building had a tuff stone footage, mural paintings, stucco, a hypocaust and glass windows. Its lay-out resembles the plan of the villa of Simpelveld in the province of Limburg.350 A remarkable find is a limestone altar dedicated to Hercules Magusanus, measuring 50 x 25 x 15 centimetres.351 It was probably found in a secondary context, since another fragment of the altar was discovered at a distance of dozens of metres. Yet, its original location may not have been far removed from its find-spot. The dedicator was a Volusius Vetonianus.352

344

Tac. Hist. 4.12: Batavi, donec trans Rhenum agebant, pars

352

From Germania Inferior two Volusii are known: the one

Chattorum, seditione domestica pulsi extrema Gallicae orae

mentioned on the altar from Houten and a Q. Volusius

vacua cultoribus simulque insulam iuxta sitam occupavere,

Ingenus, whose name occurs on an altar dedicated to

quam mare Oceanus a fronte, Rhenus amnis tergum ac latera

the matronae Alaferchuiae in the sanctuary of Eschwei-

circumluit.

ler-Fronhoven. The gentilicium Volusius is of Etruscan

See for the discussion Roymans 2004, 26-27. See for La Tène

origin. From Gallia Belgica we know a Q. Volusius Saturnus,

bracelets and late Iron Age Celtic coins in the Kromme Rijn

who was censor in 59 and legatus in 61. It has been suggest-

area Vos 2009, 183-185, notes 3-6 and figs. 5.1 and 5.4, 189-191.

ed that Volusius Ingenus’ family were granted citizenship

See Roymans 2004-24-25, note 89 and figs. 3.1-3.2; Polak/

through Saturnus, when he was censor or legatus, and thus

Kooistra 2013, 389-390.

got his gentilicium, too. The cognomen Vetonianus or Vetto-

347

Van Es/Verwers 1980, 7.

nianus is very rare in the western provinces. The name may

348

Vos 2009, 2.

refer to the Umbrian town of Vettona, to the Vetones from

349

Vos 2009, 134-139.

Hispania or to the gentilicium Vettonius, which is possibly

350

Vos 2009, 167-173.

of Celtic or Etruscan origin: Van Es/Hessing 1990, 87-88 and

351

Vos 2009, 140 and fig. 4.16. The altar was published by Van

notes 34-37.

345

346

Es and Hessing: Van Es/Hessing 1990, 81-90.

51

• Houten-Wulven At Wulven, excavations took place in 1993-1995. In the earliest habitation phase, between 15 BCE and 70, the settlement consisted of three main buildings and a substantial number of outbuildings. One of the three buildings was very large, 35 by 12 metres, and may have been a farmstead with a timber porticus. During the second habitation phase, in the 2nd century, the settlement comprised four main buildings. A bronze statuette of Isis-Fortuna was found by a detectorist at the end of the excavations.353 • Jutphaas The village of Jutphaas is situated six kilometres west of Houten. Before the planned construction of a residential area, small-scale rescue excavations were carried out between 1974 and 1976, revealing the existence of three settlements, A-C.354 The habitation period of settlement A started sometime in the late Iron Age. The earliest Roman finds from settlement A can be dated to the second and third quarter of the 1st century.355 The site was abandoned in the first quarter of the 2nd century.356 Settlement C has yielded a much higher percentage of Roman imports. This settlement was abandoned after the middle of the 2nd century.357 Both settlements have produced fragments of terracotta statuettes. Two fragments come from settlement A: the upper part of a Juno statuette and a fragment of an unidentifiable standing figure. Settlement C has yielded the lower part of a Mercury statuette, a fragment of a Venus statuette, a fragment of a mother goddess figurine, one of an unidentifiable enthroned figure as well as three fragments of an unidentifiable nude or semi-nude figure. The Venus fragment depicts the chest of a nude woman and came to light at the farmyard. The mother goddess figurine was found in the top soil above a fire place outside the northern entrance of the farmstead. The three unidentifiable fragments come from the top soil above a hearth. A fourth unidentifiable fragment was found in the filling of a creek.358 • Wijk bij Duurstede-De Horden Wijk bij Duurstede-De Horden has been excavated between 1977 and 1987, a surface of 14 hectares. Traces of habitation date from the middle Bronze Age into the beginning of the 3rd century. More than hundred structures have been identified.359 In the early Roman period the settlement consisted of two farmsteads with outbuildings. In the second half of the 1st century, the number of imported goods in the settlement increased.360 Around 150 ditches were dug alongside the houses, indicating that the rising ground water level became problematic. One of the buildings was constructed over a Bronze Age barrow and could have had a religious function. Inside, three concentrations of coins were found, which may have been part of what originally was a closed find complex.361 This building (H25) and a second, almost identical, building (H9) had a timber porticus, a phenomenon that has been attested at several locations in the Dutch river area and the northern part of the southern sandy soils. Vos has suggested that returning veterans, inspired by Roman military architecture, introduced the porticoes and the use of ceramic roof tiles in their farmhouses.362 In the vicinity of building H25 a bronze statuette of Mercury, a fragment of a terracotta statuette and several republican denarii were found. Near H9, a fragment of a terracotta statuette came to light.363 353

Vos 2009, 133 and fig. 4.12.

361

Van Es 1994, 58-60 and figs. 44-45.

354

Van Tent 1978, 199-203.

362

Vos 2009, 237-243. The number of roof tiles was too small to

355

Van Tent 1978, 202-203.

cover a complete roof. Moreover, the weight of a complete

356

Van Tent 1978, 205.

roof would be too much for a timber construction. Addi-

357

Van Tent 1978, 208-209.

358

See for all terracotta fragments from Jutphaas: Van Boekel

tional information by Wouter Vos. 363

Vos 2009, 89-96. It is possible that fragments of terracotta

1987, nrs. 41, 48, 70, 123, 150, 162, 163, 175.

statuettes have been overlooked at the time of the excava-

359

Vos 2009, 60-62.

tions. Additional information by Wouter Vos.

360

Vos 2009, 88.

52

• Wijk bij Duurstede-De Geer 800 metres from De Horden, on higher ground, a settlement at Wijk bij Duurstede-De Geer has been uncovered during excavations in 1969, 1972 and the late 1980s. The settlement covered a surface of 5.5 hectares. House plans have been identified from the late Iron Age into the early Medieval period. The majority of finds from this settlement can be dated to the 2nd and the 3rd centuries. Graffiti on pottery as well as stamped bricks point to contacts with the Roman military. Striking finds are fragments of a bronze wine strainer, a bronze dice, a stilus, three seal rings and a bronze statuette of Mars.364 Utrecht: find-spots of statuettes without contextual data Find-spot

Material

Subject

Reference

Cothen

bronze

Mars

DetectorMagazine nr. 134, 2015365

Odijk

bronze

Mercury

Zadoks 1969, nr. 45

Utrecht-city centre

bronze

Jupiter

Zadoks 1969, nr. 31366

Utrecht: uncertain find-spots of statuettes Find-spot

Material

Subject

Reference

Vechten or Wijk bij Duurstede

bronze

triple-horned bull

Zadoks 1969, nr. 65

Wijk bij Duurstede

bronze

Minerva

Zadoks 1969, nr. 56

Wijk bij Duurstede

bronze

greyhound

Zadoks 1969, nr. 73

2.3.4

Southern Gelderland

The Dutch eastern river area is composed of Holocene fluvial deposits and is bordered to the north and east by Pleistocene ice-pushed ridges. Natural borders of the region are the Rhine in the north and the basin of the Meuse in the south. Very soon after the arrival of the first troops in or shortly after 19 BCE, the eastern river area became the centre and eastern part of the civitas Batavorum with its capital in the eastern periphery.367 Together with the Kromme Rijn area, this part of the Netherlands has been intensively studied. A large regional study was started in 1978: the Eastern River Area project. One of the aims was to inventory ‘the accumulated and still increasing quantity of valuable data’.368 The project concentrated on the development of society during the Roman period, mainly within the boundaries of the Roman Empire, but also included a strip of land north of the Rhine, where Ede, Bennekom, Arnhem and Wageningen are situated. The results of the project included overviews and maps of Roman forts and possible forts or fortified sites, as well as already known and new settlements from the late Iron Age into the Merovingian period, a chronology and typology of sites, a site catalogue and an analysis of the finds.369 Most statuettes from the eastern river area come from Nijmegen and its direct surroundings. The number of statuettes found in the hinterland is small. One of the reasons could be that the majority of identified Roman period sites outside Nijmegen have not been excavated.370 It is also possible that people in the hinterland had less access to Roman goods, or lacked interest in statuettes. 364

Vos 2009, 108.

368

Willems 1981, 10.

365

Found in 2013 in a field near Cothen.

369

Willems 1981 and 1984.

366

Stray find from the Oude Gracht.

370

See for the site catalogue Willems 1981, 93-132.

367

Willems 1981, 11-15.

53

Map 9 Find-spots of statuettes in Southern Gelderland Base map: paleogeographic reconstruction of the Netherlands around 100 (Vos/Vries 2013). 1. Deest 2. Wijchen 3. Beuningen 4. Heteren 5. Elst 6. Millingen 7. Groesbeek 8. Rossum 9. Heerewaarden 10. Varik 11. GeldermalsenHondsgemet 12. Druten 13. Berg en Dal 14. Overasselt 15. Nijmegen 16. Oosterhout 17. Tiel-Passewaaij

• Deest The village of Deest is situated on the left bank of the river Waal, some 15 kilometres west of Nijmegen. Plans for sand extraction prompted an archaeological investigation in 2003, since a Roman period settlement was suspected here.371 Besides traces of a late Neolithic site, a Roman period settlement could be identified. It was occupied from the first quarter into the late 1st century. Traces of structures have not been found.372 Among the relatively large quantity of metal finds was a bronze figurine of a boar, a stray find.373 • Druten-Klepperhei Five kilometres west of Deest, at Druten-Klepperhei, a Roman villa was discovered during excavations between 1975 and 1978. The first results were published in 1978 and comprised the plans of the structures and the habitation phases.374 So far, the only finds that have been published are the wall paintings.375 The field drawings have been digitised and measurement errors have been adjusted. This has resulted in different interpretations of the settlement’s development.376 The settlement comprised two contemporary farmsteads in the first half of the 1st century. In the 2nd century, one farmstead probably was enclosed. It was partly stone-built. In the late 1st and first half of the 2nd century, the settlement consisted of several one-aisled

371

Van Bergen et al. 2008, 6-9.

374

Hulst 1978, 133-151.

372

Van Bergen et al. 2008, 57.

375

Peters et al. 1978, 152-197, Swinkels 1982, 183-192.

373

Tuijn 2006, 37.

376

Van Enckevort 2012, 232-244.

54

houses, of which one had a long porticus, plastered walls and an internal gutter.377 In the second half of the 2nd century a house was constructed with stone foundations. Post-built houses also belonged to the settlement in this period. The settlement was abandoned in the 3rd century.378 One terracotta fragment of a human figure, possibly depicting a boy, was found in the top filling of a well, together with terra sigillata sherds that could be dated around 150-200. The well went out of use at the end of the 3rd century.379 • Elst The protestant church or Grote Kerk at Elst was bombed in World War II. Before its restoration in 1953, excavations were conducted at the end of the 1940s. The excavations revealed the remains of two ‘Romano-Celtic’ temples, published by Jules Bogaers in 1955.380 Research in 2002-2003 confirmed Bogaers’ assumption that the temples were built on an earlier cult place, probably an open-air sanctuary. This was concluded from a 14 C-dating of the layer of animal bones which Bogaers already connected with a ritual function of the site before the middle of the 1st century, probably as early as the Augustan period. The discovery of a sword dating to the La Tène D2 period (70-30 BCE) suggests an even earlier date for the cult place.381 The exact construction date of the first Roman period temple is uncertain. Bogaers assumed it was destroyed by fire during the Batavian revolt in 69/70. The results of the 2002-2003 campaign have made evident that the second temple was built between 97 and 103.382 It remained in use until the beginning of the 3rd century and probably was destroyed by fire.383 With a ground plan of 31 by 23 metres, the temple at Elst is one of the largest known Gallo-Roman temples.384 During the excavations a four centimetres long bronze fragment was found in top soil. According to the excavators it depicts the upper part of Hercules’ club.385 Therefore, they assume that the temple was dedicated to Hercules or Hercules Magusanus.386 A fragment of a limestone altar was found on the site of a Roman period settlement at Elst-Brienenshof, some 700 metres south of the temple at Elst. Excavations in the late 1980s has produced traces of a settlement which was occupied from the early Iron Age into the Medieval period. Derks et al. suggest that the altar originally came from the temple precinct at Elst. Since construction debris from the temple has been moved to the settlement at Brienenshof in the early Medieval period, this is not unlikely. The excavators mention wall painting fragments, small altar fragments and an architrave block which fits in Bogaers’ reconstruction of the temple at Elst.387 Two horizontal serifs of the first letter and the top of the vertical bar plus a part of the highest horizontal arm of the second letter have been preserved. The remains of two letters have been interpreted by Derks et al. as an H and E.388 Assuming that the drawing in the publication is correct, there are more options. The first letter could be an M, an N, a V, an L followed by an I or vice versa, two I’s or an X. Besides an E, the second letter could also be an F. An alternative for the name Hercules would be the name of another deity. A plausible alternative would be the name of a native goddess. Several altars found in or around Batavian territory are dedicated to native goddesses, such as Hurstrga (see below).389 Inscriptions on 377

This building was a bathhouse.

387

See also Hulst/Jong 1989, 29-30.

378

Habermehl 2011, 196-197.

388

Derks 2008a, 139 and figs. 5.1-5.2.

379

Van Boekel 1987, nr. 9.

389

The first word of the dedication, Deae, has been engraved

380

Bogaers 1995; Derks/Hoff 2008, 6-7.

in the head of the altar to this goddess. From the 2nd cen-

381

Heeren et al. 2008, 30-31; Renswoude 2008, 75.

tury onward, dedications to goddesses often begin with the

382

Heeren et al. 2008, 30-34.

word Deae, followed by her name. Exceptions in and near

383

Derks 2008a, 135-138.

the study area are two inscriptions from Germany to the

384

Bogaers 1955, 172, 174-181, 240-241.

goddess Vagdavercustis (Kalkar, AE 2003, 1227 = AE 2005,

385

Renswoude 2008, 77.

1060 and Rindern, CIL 13, 8703); from the Netherlands to

386

Derks 2008a, 138-139. Also Bogaers assumed that Hercules

Viradecdis (Valkenburg, AE 1975, 635) and to Rura (Roer-

Magusanus was the principal god in Batavian territory:

mond, EDCS-67800024).

Bogaers 1955, 173 and notes 3-4.

55

altars can begin with the name of the dedicator, a formula or an abbreviation.390 The tree on one side of the altar does not provide any clues as to the identity of the deity for whom the altar was erected.391 And lastly, the altar from the settlement at Elst-Brienenshof could have been erected in another nearby sanctuary: the temple at Elst-Westeraam. This temple, measuring 13 x 14 metres, was built between 10 and 20 and probably was in use until the end of the 2nd century. No finds were recovered from the site that could shed a light on the deity or deities that were worshipped in this temple.392 • Geldermalsen From Geldermalsen comes a remarkable stray-find. It was discovered by a detectorist in the vicinity of the village. The small figurine is 55 mm in height and represents an African young man holding the ears of an elephant’s head in front of him. His upper body rises from a wreath of leaves on the head of the elephant, mounted on a lion’s leg (figs. 2.58-59).393 The elephant wears a feather headdress. The figurine must have been part of a group of three or four figures holding a bronze container.394 Although the figurine falls outside the scope of this research, it is mentioned here because of its exotic nature. It may have been brought to our region by a returning veteran, or by a trader following the Roman troops. A second statuette of an African man has been found at Heeswijk, North Brabant (see below).395 • Geldermalsen-Hondsgemet Around ten kilometres west of Tiel, excavations in 2003-2005 uncovered a Roman period settlement at Geldermalsen-Hondsgemet. The settlement covered an area of five hectares and was inhabited from the late Iron Age into the late Roman period. The habitations were built on both banks of a gully. In the early Roman period, imported goods were introduced in the settlement. From 120 onwards, farmsteads were no longer abandoned, but new farmhouses were built on the same spot. Special deposits dating to the earliest phase and the middle Roman period have been identified at the boundaries of the settlement. Other deposits were recovered from wells. In the late Roman period, between 270 and 350, the settlement was reorganised. Traces of metallurgy, including bronze casting, suggest that the inhabitants combined these activities with stock breeding and arable farming.396 A remarkable quantity of metal finds included bronze bracelets and crucibles dating to the late Iron Age (200-50 BCE). Although the settlement has been fully excavated, it has yielded only one terracotta statuette. The figurine is complete and depicts a lion. It has been recovered from a pit in a layer that could be dated after 130.397 • Groesbeek-Hüsenhoff Ten kilometres southeast of Nijmegen, at Groesbeek, excavations were carried out at the location Hüsenhoff. Besides an Iron Age burial site, a Roman period farmstead and ten outbuildings were uncovered. They were constructed during the late 2nd century. In the 3rd century, several outbuildings went out of use and the farmhouse was rebuilt. The second habitation phase of the settlement seems to have ended around the middle of the 3rd century. The excavation has yielded a small bronze Minerva statuette. It was found in the vicinity of one of the outbuildings east of the farmstead, next to a ditch and post alignment. Because of this figurine, the excavators assume that there was a household shrine inside the building.398 390

See for formulas on altars in the two Germanies and Gaul

394

Raepsaet-Charlier 1993. 391 392 393

56

See for a bronze container carried by four Amor figurines Swinkels 2011, 11.

Bauchhenß 2013, 147.

395

See for the two figurines also 3.8.16.

Van Enckevort 2005a, 37-50.

396

Van Renswoude/Van Kerckhove 2009.

Like the side tables with lion’s legs from Pompeii and Her-

397

Van Renswoude/Van Kerckhove 2009, 199, 201.

culaneum: Mols 1999, 44 and figs. 111-114.

398

Van der Feijst 2012, 107-108.

• Millingen Small excavations at Millingen between 1887 and 1889 revealed a Roman period stone-built structure enclosed by an elliptical ditch. Therefore, it was suspected that a watchtower or signal-station stood at site 441, Millingen-Eversberg. Willems, however, has suggested the stone building may have been part of a villa.399 An altar dedicated to Rufia Materna, embedded in one of the walls of the local church, was sold to the RMO at Leiden in the 19th century.400 On the altar a tree is depicted. Since the ‘goddess’ is the daughter of the dedicators, it is thought that Rufia Materna was a priestess. A bronze statuette of Harpocrates was found around 1925 on the right bank of the Waal, possibly on the site of the Waiboerhoeve, just opposite Millingen.401 • Nijmegen: Oppidum Batavorum Around 19 BCE, the first Roman troops settled on the Hunerberg. In 12-10 BCE, a settlement was founded in the centre of modern Nijmegen, probably Oppidum Batavorum or Batavodurum, mentioned by Tacitus and Ptolemy.402 At the beginning of the 20th century, the exact location of Oppidum Batavorum was still unknown. Holwerda conducted excavations on Kops Plateau and thought he had discovered Oppidum Batavorum. Mathé Daniels, the city archivist, suspected that a Roman period settlement was located in the city centre. Excavations between 1979 and 1981 at Kelfkensbos and Valkhof confirmed that Oppidum Batavorum was located on the Valkhof grounds in the modern city centre, covering a surface of around 25 hectares.403 Terra sigillata sherds found during the excavations in 2005-2006 and 2008 at the location St. Josephhof suggested that the settlement was built around 10 BCE, although two stamps suggest an earlier date, between 19 and 16 BCE.404 Traces of fire throughout the settlement testify to the destruction of the settlement during the Batavian Revolt of 69/70. It has been argued that the settlement was founded to transform the territory of the Batavi into a civitas, an administrative regional district, with Oppidum Batavorum as its capital.405 The settlement developed along both sides of a road running from east to west. The excavators distinguish three occupation phases and two building blocks. The earliest phase ends between 10-20, the second phase around 40 and the third in 69/70. Despite the absence of a rectangular street grid in the investigated part of the settlement, the excavators concluded that the layout was Mediterranean. The large quantity of graffiti on Arretine ware seems to indicate that in the early 1st century the houses at St. Josephhof were inhabited by veterans. In the second phase, houses had walls which were painted on the inside. Cellars, too, made their appearance in this phase. One building, constructed around 40, has been interpreted as a soldiers’ barracks. In the Neronian period, buildings were rebuilt with tuff stone footings, terracotta roof tiles and antefixes.406

399

Willems 1981, 100, 121 and note 131.

ments, Batavodurum being a prehistoric settlement in

400

Willems 1981, 128, site 445. The inscription reads: Deae

Nijmegen-West: Van Enckevort/Thijssen 2005a 97-102. This

dominae Rufiae/[M]aternae aram et /[l]ucum consacravit/

idea is disputed by Roymans: Roymans 2004, 148 and note

Mucronia Marcia /ubi omnibus annis sacrum/instituit XVI

393. See also Driessen 2007, 84 and note 433. It remains

k(alendas) Aug(ustas) / [e]t natali Maternae f(iliae) suae /

strange, however, that Tacitus would refer to a presumably

[---] I (?) N(onas) Octob(res) et Parental(ibus) /[I]X K(al-

Roman settlement with the word oppidum, whereas Caesar

endas) Martias Rufis Simil[i] /patri et [S]imi [f]il(io) [et] / 401 402

uses the term for fortified, native settlements.

Maternae [fil(iae)]: CIL 13, 8706.

403

Van Enckevort/Heirbaut 2010, 45-46.

Zadoks et al. 1969, nr. 21.

404

Van Enckevort 2012, 133.

Tacitus mentions both names: Hist. 5. 19-20. Ptolemy only

405

Van Enckevort/Thijssen 2005a, 99-100, Roymans 2004, 208-

mentions Batavodurum: Geog. 2.9.5. Harry van Enckevort and Jan Thijssen have suggested that Oppidum Batavorum

209. 406

Van Enckevort 2010a, 55-98.

and Batavodurum might have been two different settle-

57

During an excavation in the Gerard Noodtstraat in 1993, a large bronze kettle came to light. At the location Hertogstraat a comparable kettle was found, next to a set of millstones. It contained seven bronze vessels and an iron spear head.407 A Spanish wine amphora (Haltern 70) was buried straight-up behind house A9. It contained animal bones, including those of a young pig and may have been a ritual deposit, which is suggested by the position of the vessel and the pointed base, which had been deliberately cut off and was deposited near the neck.408 In 1923, an altar came to light during construction work at the northwest side of St. Canisiussingel. It was dedicated to a triad of mother goddesses whose name has only been partially preserved.409 A bronze statuette of Minerva was found in the spoil heap during excavations in 1985 at the Waalkade (fig. 2.80).410 The excavation in the Gerard Noodtstraat produced a bronze statuette of Bacchus (fig. 2.78).411 This statuette and the bronze Hercules from Velsen (see below) are among the earliest examples of bronze statuettes in the study area. In the course of the same excavation, a bronze pedestal came to light (fig. 2.79). Both objects are stray finds, but statuette and pedestal may have belonged together. Also, a bronze fragment of a right foot was found, measuring three cm in length (fig. 2.81-81a).412 The foot must have been part of a circa 30 cm high statuette of a deity or emperor. An excavation in the Hertogstraat produced a bronze billy goat (fig. 2.77). The excavations at St. Josephhof have yielded two bronze statuettes, one depicting a roe (fig. 2.82), the other a standing male on a rectangular plinth (fig. 2.83). The 63 mm high figure depicts a desultor, an acrobat-rider.413 • Nijmegen: Ulpia Noviomagus414 The first systematic excavations in the western part of Nijmegen were carried out in 1834 at the location Winseling-Fort Krayenhoff. The uncovered structures were interpreted as a public bath complex. Based on altars and dedicatory inscriptions from the site, Brunsting believed the structures were part of a temple complex. In the early 1920s, during construction works further south, the city archivist Daniels saw the foundations of a large complex, measuring 45 by 96 metres. He was able to reconstruct the foundations of two Gallo-Roman temples.415 Archaeological research remained focused on the military camps at Nijmegen, until in the 1970s excavations were carried out north of the Weurtseweg in the quarter Waterkwartier, followed by excavations in the 1980s at the corner of Weurtseweg and Bronsgeeststraat. Here the eastern enclosure ditch of the settlement was discovered, as well as a part of its city wall.416 In 1992 in the Waterkwartier quarter, a part of a large public bath complex was unearthed.417 Archaeological research continued in the 1990s, but the results have only partially been processed and published. Therefore, the settlement’s development outlined hereafter must be regarded as provisional. In 70 or 71, the Legio X Gemina arrived from Hispania and built a new legionary fortress on the Hunerberg.418 For thirty years, this legion would be the driving force behind the economy of a new settlement that 407

Van Enckevort 2010a, 97-98.

408

According to the excavators the vessel was used as a rub-

409

attachments, which usually are sandaled or bare feet: Mols 1994, 293-296.

bish container: Van Enckevort 2010a, 82. Arguments that

413

See for the interpretation of the figurine 3.10.

speak against the use of the large whine amphora as a

414

See for the name of the settlement chapter 1, note 5.

rubbish container are its value and the narrow neck.

415

Van Enckevort/Thijssen 1996, 63-64.

Daniels/Brunsting 1955 nr. 44. The inscription reads: [-----]

416

Van Enckevort/Thijssen 1996, 65-66.

ibus […..] e. Daniels’ suggestion was matribus sive, since

417

Van Enckevort/Thijssen 1996, 81-82.

there is an inscription from Bonn addressed to the mother

418

Haalebos/Willems 2005b, 73-86. According to traditional

goddesses sive matronis: Daniels/Brunsting 1955, 37.

beliefs, this legion had the task of protecting and maintain-

410

Verrijt 1985, 47-49. The statuette is missing.

ing the civitas Batavorum, but not everybody is convinced

411

See for the statuettes from Oppidum Batavorum figs. 2.59-

that a whole legion was needed for this: Erdrich 2005,

2.64.

47-49, Driessen 2007, 93.

412

58

The foot does not resemble the feet that are furniture

developed on the left bank of the Waal River: Noviomagus. It is thought that the emperor Trajan granted the settlement municipium status around 98. In inscriptions, the town is referred to as Municipium Batavorum or Ulpia Noviomagus. A few years later, between 102 and 104, Trajan transferred the Tenth Legion to Aquincum (Budapest) for his military campaigns against the Dacians, which must have been a disaster for the local economy.419 In the late 2nd century, during the reign of Marcus Aurelius (161-180), a defensive wall was constructed around the city, presumably against the increasing number of Germanic raids. Excavations have revealed traces of fire in large parts of Ulpia Noviomagus, but also in other settlements in the region, like Empel and Tongeren.420 The Tenth Legion must have been involved in the construction of public buildings like the baths and temple complex, since many stamps on building material bear the name of this legion. Apart from the location of these buildings little is known about the lay-out of Noviomagus. It covered a surface of 30-40 hectares, about half the size of the Colonia Ulpia Traiana at Xanten, Germany.421 The southern and eastern boundaries of the settlement have been identified. The western boundary has not been securely attested, but may have run east of the Rivierstraat. The Waal has flooded the northern part of the settlement, but to what degree is still unknown.422 So far, houses have been found at the south and southwest side of the settlement. The planning of the settlement is Roman, with a cardo and a decumanus maximus, but insulae seem to be absent at several locations. The unearthed house plans are predominantly strip-houses.423 At the beginning of the 2nd century, many buildings were still post-built. The two temples were built around 100, for which a terrain with pottery shops was cleared.424 After the destruction of the town, there are indications for a slow recovery during the first decades of the 3rd century. Shortly before the middle of the 3rd century, the town had been largely deserted. Recent excavations at the Voorstadslaan have yielded graves dating to the second half of the 3rd century. They indicate that habitation continued after 240, maybe in the northern or central parts of the town.425 When Rome tried to restore its power along the northern limes at the end of the 3rd century, a new fortification was built on the Valkhof hill, around 280. Coins from its defensive ditches indicate that the area continued to be inhabited.426 Outside the walls of the fort, a large cemetery developed between 300 and 700.427 Daniels mentions a limestone statue, altars and a fragment of a Jupiter column that came to light at Nijmegen-West near Fort Krayenhoff, the site which Brunsting interpreted as a temple complex.428 Three altars are dedicated to Jupiter Optimus Maximus, the matres Mopates and the matres Aufaniae. A dedicatory relief depicts Abundantia. The fragment of the Jupiter column, 45 cm in height, was found during the excavations by Reuvens in 1834. A 21 cm high limestone statuette depicts an enthroned woman with an object in her right hand and her left arm resting on a shield. Therefore, the figure has been interpreted as Minerva (fig. 2.96).429 A 9.5 cm high stone ex-voto depicts the lower body with genitals and upper legs of a man.430 Apparently, a bronze statuette of Mercury, a small bronze Cybele figurine and a terracotta statuette of Fortuna were also found at Fort Krayenhoff, but details on the find circumstances are not known.431 419

Bogaers 1967, 54-76. Whether Trajan tried to compensate

427

the city by granting market rights is doubtful. See for the

around 1,500 have been examined: Van Enckevort/Thijssen

debate on this topic: Haalebos 2000, 35-39, Van Enckevort/ Thijssen 2005a, 107.

The estimated number of graves is 5,000-10,000 of which 2014, 37-41.

428

Daniels 1927, 65-111; Daniels/Brunsting 1955, nrs. 17, 24-26,

420

Van Enckevort/Thijssen 2005a, 110.

421

40 hectares according to Kemmers/Polak: Kemmers/Polak

429

See also Daniels/Brunsting 1955, nr. 25.

2011, 252; 30 hectares according to Heirbaut: Heirbaut 2013, 12.

430

Daniels/Brunsting 1955, nr. 17 and plate IV nr.8; Zee 2005,

422

31, 38-39.

Heirbaut 2013, 11-12.

191, fig.77.

423

431

Heirbaut 2013, 39-40; Kemmers/Polak 2011, 252-253.

424

Van Enckevort/Thijssen 2005a, 110.

1987, nr. 27; for the bronze Cybele and Mercury statuettes

425

Heirbaut 2013, 162.

Zadoks et al. 1969, nrs. 16 and 44. The Cybele figurine is

426

Van Enckevort/Thijssen 2014, 37.

probably not a free standing statuette but an attachment.

Zee 2005, 192-193. See for the terracotta Fortuna Van Boekel

59

So far, no statuettes have been recovered with certainty from the precinct of the temple complex at Maasplein.432 A bronze Amor figurine and a terracotta statuette of Mars have been found outside the temple’s precinct.433 Although very few statuettes have been recovered from sanctuaries in the study area, people must have offered statuettes in the temples. A long wall fragment with niches is known from Fort Krayenhoff. A part of the enclosure wall of the temple complex at Maasplein also had niches. It has been advanced that statuettes were set up in these niches.434 Besides an altar dedicated to Fortuna, a silver finger ring was found during the excavations at Maasplein. It bears an inscription addressing the god Salus.435 These finds indicate that several deities were worshipped in Noviomagus. To which deity or deities the temples were dedicated remains unclear. A 20 cm high limestone statuette of Hercules came to light during the excavations at Maasplein (figs. 2.84-84a). The fragmentary figure may have been free standing or part of an architectural ornament and was recovered from an area situated outside the temple precinct.436 Complete, the figure must have been around 30 cm in height. It depicts the hero wearing the skin of the Nemean lion with the head of the lion as a helmet, a representation of Hercules which is known from Greek vases and statues. Earlier excavations at Maasplein have yielded eight metal statuettes. They comprise a silver statuette of a billy goat and seven bronze specimens: three statuettes of Mercury, an attribute of Mercury, a caduceus, a statuette of Venus, a Lar, and the leg of a goat or sheep (figs. 2.85-2.93). The identifiable terracotta statuettes and statuette fragments include Cybele (one), Juno (one, maybe two), Minerva (one limestone and one, maybe two terracotta figurines), Mercury (two), Mars (one), Fortuna (four to eight) and Venus (seven).437 The human figures represent two mother goddess figurines, five busts of a woman, two busts of a boy and two of a woman or a boy.438 Three figurines are probably locally made. The first depicts the lower part of a torso with the upper legs of a male figure (fig. 2.118).439 The second depicts the head of woman (fig. 2.159).440 The third will be discussed in chapter 4 (figs. 2.119-2.122). The identifiable terracotta animal figurines comprise a bull, four lions, two to three horses and a bird.441 The excavations at the site Rijn-Lekstraat have yielded a few unidentifiable fragments of terracotta statuettes, as well as a helmet that most likely belonged to a terracotta statuette of Minerva, the back side of a standing female figure, a fragment of a horse figurine and the head of a woman (ca.100 mm in height) of a 432

Since not all excavations have been published, it cannot

436

be established with certainty which finds come from the temple precinct. Trenches 1-9 from the 1992-1993 excava-

Harry van Enckevort. 437

Van Boekel 1987, nrs. 16 (Cybele), 27 and 30 (Fortuna),

tions were all located outside the temple complex: oral

113 (Venus); Van Boekel 1996, nrs. 9 (Mercury) 15 (Venus),

information by Harry van Enckevort. The upper part of a

44 (Mars). See for the identifiable, hitherto unpublished

terracotta statuette of Cybele came to light in 1921, when

terracotta fragments from Ulpia Noviomagus figs. 2.97-122, 2.160-2.161a.

Daniels discovered the temple complex. Whether the statuette was found within the precinct of the complex has not

438

Zee 2005. 192-193. The Amor figurine is a large attachment

See for the mother goddess figurines Van Boekel 1987, nrs. 73, 81bis; for the busts Van Boekel 1987, nrs. 177, 196, 198.

been documented. 433

The statuette was found in trench 4. Oral information by

439

The execution suggests that the figurine dates to the late

and comes from trench 6. The Mars statuette is the only

Medieval period or beyond. See, for instance, a few frag-

terracotta specimen in the study area and comes from

mented figurines from Amsterdam: Gawronski/Kranen-

trench 3. See for the Mars statuette also Van Boekel 1996,

donk 2018, 425-428.

nr. 44.

440

See also Van Boekel 1996, nr. 10.

434

Zee 2005, 192-193.

441

Van Boekel 1987, nrs. 208, 217, 251, 254. The fragment of

435

The inscription on the ring reads: Sal(uti) Rusticus/sutori-

lion’s manes (fig. 2.113) could be identified by comparison

bus / Noviom(agensibus)/(curia) Esser(avi) d(onum) d(ed-

with a statuette from Vechten and a fragment from Aarden-

it) d(edicavit): AE 1998, 967 = AE 2012, 981. The inscription

burg: Van Boekel 1987, nrs. 250 and supplement nr. 16.

on the altar reads: Fortun(ae) / sacrum: Van Enckevort/ Thijssen 2005a, 108.

60

large limestone statuette.442 The absence of a headdress and the hairstyle suggests that Venus or Fortuna is depicted.443 • Nijmegen-Hatert The site Hatert is located 4 kilometres southeast of Nijmegen-West. Excavations in 1979-1980 and 1986 produced traces of a settlement with a cemetery. Whereas a large part of the burial site could be examined, very few traces of the settlement have been preserved. It must have measured at least 125 by 150 metres. The earliest evidence dates to the end of the 1st century.444 From the settlement comes a terracotta pedestal with the fragmentary remains of a standing figure leaning on a column. The statuette was a stray find.445 • Nijmegen-cemeteries: Hugo de Grootstraat In the eastern part of Nijmegen several clusters of graves have been investigated. Although the exact boundaries are not known, a part of the 1st-century cemetery is roughly bordered to the north by the Barbarossastraat, to the west by the Hugo de Grootstraat, to the east by the Museum Kamstraat and to the south by the Jan van Goyenstraat. The earliest phase dates to the Augustan period. Here, the inhabitants of Oppidum Batavorum buried their dead. Until 1975, finds from this site were merely stray finds. The first excavations in the Hugo de Grootstraat in 1975 produced at least eleven 1st-century graves, but also late Roman graves, which most likely held deceased coming from the late settlement on the left bank of the Waal (see above). In the 1970s, nearly 200 cremation graves were investigated, dating to the middle of the 1st century. Excavations in the Museum Kamstraat in 1976 yielded around 150 certain pre-Flavian graves.446 During construction works in the Barbarossastraat, a terracotta Venus statuette came to light.447 • Nijmegen-cemeteries: Verpleeghuis Margriet In the late Roman period, two cemeteries developed to the west and east of the late Roman fortification on the Valkhof-Kelfkensbos, nowadays part of Nijmegen’s city centre (see above). Excavations in the eastern cemetery were conducted in 1975 and between 1980 and 1983.448 The grounds of the former Verpleeghuis Margriet were part of the eastern cemetery. Excavations in 2008 yielded two terracotta Venus statuettes. They were found in situ in graves, perhaps children’s graves (figs. 2.152-153a), dating to the end of the 3rd or the beginning of the 4th century.449 • Nijmegen-cemeteries: Ulpia Noviomagus The graves belonging to the cemetery of Ulpia Noviomagus were located at both sides of the road that ran between the town and the military camps to the east. It is the largest known Roman period burial site in the study area. The available publications focus on the southern part of the burial site and the richest graves.450 A substantial part of the site has been overbuilt and many finds have disappeared from the archaeological record, either as a result of treasure hunting or construction works in the 19th and early 20th centuries. During excavations between 1993 and 1997 it became clear that there are still many undisturbed graves, also in the area Brunsting examined in the 1930s. The site is known as ‘the cemetery under Hees’.451 Bogaers and 442

The terracotta helmet most likely belonged to a statuette of

446

Bloemers 2016, 70.

Minerva, since terracotta statuettes of Mars are very rare. If

447

Van Boekel 1987, nr. 110.

it did belong to a statuette of Mars, it would be the second

448

See for the publication of the finds from the two cemeteries Steures 2013.

specimen in the study area. 443

See for Venus statues with a comparable hairstyle Schmidt

449

217 and nrs. 271, 277, 281. 444 445

Oral information by Harry van Enckevort, Municipal Department of Archaeology, Nijmegen.

Haalebos 1990, 20-22.

450

Brunsting 1937; Koster 2013.

Van Boekel 1987, 612 and nr. 152.

451

Koster 2005, 205-208.

61

Haalebos conducted excavations in 1978 and the 1980s between Voorstadslaan and Sperwerstraat. Forty-six graves from this part of the site have been published.452 Brunsting investigated 37 of an estimated number of 12,500 graves, which he dated between 70 and the middle of the 3rd century. Based on the inhabited surface of Ulpia Noviomagus and the number of graves, Brunsting estimated the population at 5,000. Today we know that Ulpia’s cemetery is much larger than Brunsting assumed.453 A bronze hoard of 80 fragments of statues, statuettes, inscriptions and architectural ornaments came to light at Hessenberg in Nijmegen’s city centre. In the Roman period, this area was part of the cemetery of Ulpia Noviomagus. Among the fragments was a large fragment of a thyrsus, an attribute of Bacchus, and a large fragment of a Jupiter-Ammon herm (figs. 2.94-95).454 Finds from the 20th-century excavations and stray finds include four terracotta statuettes of Venus, two of Fortuna, one statuette of Luna, one of Juno and one of Cybele.455 The head of a Minerva statuette was found in the Krayenhofflaan when a grave pit was hit during digging operations. The head was found together with terra sigillata sherds dating to the middle of the 2nd century.456 The human figures include one mother goddess figurine, one bust of a woman and two statuettes depicting a boy standing between two horses. The latter two come from the same mould.457 Excavations in the Koekoekstraat in 2010 have produced a fifth statuette of Venus and a second mother goddess statuette (figs. 2.154-55a). The mother goddess figurine came from a grave dating to the 3rd century.458 The terracotta statuettes of animals comprise two cockerels, three ithyphallic dogs, a duck, three horned birds, a lion and a ram.459 Two amber figurines, a sleeping Amor and a bear, were found in situ in a grave during excavations in the 1980s.460 The grave, dating between 80 and 100, was a cremation grave of a woman in her twenties.461 • Oosterhout-Van Boetzelaerstraat Just opposite of Nijmegen, across the Waal, excavations were conducted at Oosterhout between 1996 and 2001. One of the examined sites was the location Van Boetzelaerstraat. Here, flint arrowheads, axes and scrapers testify to an early Bronze Age occupation. During the Iron Age, the site was overbuilt by farmsteads, outbuildings, wells and granaries. Certain finds suggest a continuing habitation until the middle of the 3rd century. In the 1st century, the settlement reached its maximum size of 4.5 hectares. The northern half of the linear settlement constituted a separate cluster. It was enclosed by a ditch and linked at the west side by a second ditch to the rest of the settlement. Despite the presence of Roman building material, the two-aisled farmsteads were built according to native building traditions. Numerous horse bones suggest that horse breeding was an important economic activity. A substantial number of Roman artefacts such as fibulae, hairpins, mirror fragments and perfume bottles point at an intensive contact with the Romans.462 Several ritual deposits from this ditch emphasized its symbolic nature. An interruption of the ditch at the east side was marked by a pointed post, carved from oak wood. A handmade vessel was buried next to the

452

Koster 2013.

453

Van Enckevort et al. 2000, 85-89.

454

See for the Jupiter-Ammon herm 3.8.11. See for the herm

458

Oral information by Harry van Enckevort.

and the thyrsus fragment also Veen 2014a, 2014b.

459

Van Boekel 1987, nrs. 209-210, 221-223, 229, 238-240, 249,

455

457

for a boy between two horses Van Boekel 1987, nrs. 247-248.

Venus: Van Boekel 1987, nrs. 93, 101, 108, 116. Fortuna: Van Boekel 1987, nr. 33, 1996, nr. 18. Juno: Van Boekel 1987, nr. 45.

253. 460

62

See for these figurines Koster 2013, 176-179; for bear figurines 3.5.9 and 3.7.5.

Minerva: Van Boekel 1987, nr. 61. Luna: Van Boekel 1987, nr. 456

See for the mother goddess figurine Van Boekel 1987, nr. 87;

46. Cybele: Van Boekel 1987, nr. 14.

461

Koster 2013,173, 176-179.

Van Boekel 1987, nr. 61.

462

Van den Broeke 2002a, 14-15.

post. Besides a symbolic and a practical function the ditch system probably also had a defensive function, in combination with the palisade.463 Not far from the western ditch, more ritual deposits were found, such as an iron sword, a part of a lance and iron nails at the bottom of a pit. The settlement has yielded six terracotta figurines: the upper part of a Cybele statuette, two lions, the head of an Apollo statuette, a fragment of a Fortuna statuette, the lower part of a Venus figurine and a fragment of a second specimen (figs 2.164-167a).464 The statuette of Cybele and the two lions were recovered from a ditch at the south side of the southernmost settlement (figs 2.60-60a1-3). The ditch had a north-south orientation, but its function is not clear. The Cybele statuette and the two lions were deposited in the ditch after the second quarter of the 2nd century, based on the production date of the Cybele statuette.465 Around 1990, a large bronze Hercules statuette came to light just north of the Waal, southwest of the site Oosterhout-Van Boetzelaerstraat.466 It depicts Hercules holding a drinking vessel, a type of representation of which several examples are known from the study area.467 • Overasselt-De Bullenkamp At Overasselt-Scheiwal a Roman period villa was discovered during excavations in 1934. Not far from this site, Willems describes two other find-spots of Roman artefacts: Overasselt-De Bullenkamp I and II, site numbers 353 and 356. According to Willems, De Bullenkamp I is a burial site, De Bullenkamp II a settlement.468 At site 356, the head of a terracotta Fortuna statuette was found in 1942.469 Other finds include Roman coins, bronze fibulae, a ring with an engraved gem, a bronze ring with a glass bead, bronze parts of a harness, an iron shovel and a bronze lock.470 • Tiel-Passewaaij In 1992, some 35 kilometres west of Nijmegen, rescue excavations started at the location Passewaaij. In the 1970s, two large concentrations of Roman material on farmland were mapped. Settlements and a burial site were identified at Hogeweg/Zennewijnseweg, Passewaaijse Hogeweg and Oude Tielseweg. Between 1996 and 2004 archaeological research focused on the burial site, the settlement at Oude Tielseweg and the settlement at Passewaaijse Hogeweg.471 Since only the latter has yielded statuettes, this settlement will be described in more detail. The earliest houses were constructed between 50 BCE and 50. Around 40, the layout of the houses began to change. The mainly two-aisled houses were smaller, measuring 12 to 14 metres in length. They were in use until the beginning of the 2nd century. In the 2nd century, the layout of the houses changed again. During the last habitation phase, between 270 and 450, only two houses were inhabited.472 The first imported vessels in the settlement appeared during the Augustan period. Between 10 and 40, imported pottery began to replace handmade pottery. In the period between 60/70 and 90, only 14 percent of the pottery from the settlement and ten percent from the burial site consisted of handmade pottery.473 Besides a staggering number of 1007 fibulae, a substantial quantity of militaria, parts of horse gear and eight bronze seal boxes have been recovered from the settlement.474 It has been argued that seal boxes, militaria, olive oil and fish sauce were introduced by returning veterans, whereas imported pottery was acquired through exchange.475 463

Van den Broeke 2002a, 12-14.

469

Willems 1981, 121; Van Boekel 1987, supplement nr. 2.

464

I thank Peter van den Broeke for the photos and additional

470

Willems 1981, 121-122, site nrs. 353 and 356.

information on the statuettes and the excavation.

471

See for a complete overview of the history of archaeologi-

465 466

See for the interpretation 4.2.2.

cal research at Tiel: Heeren 2009, 27-35.

The approximate coordinates are 185.30 / 431.10. See also

472

Heeren 2009,71-72.

Swinkels 2017b, 183-190.

473

Heeren 2009, 105.

467

See figs. 2.61-64, 3.8.3 and 4.2.2.

474

Heeren 2009, 120-126, 147-152.

468

Van Boekel 1987, note 552.

475

Heeren 2009, 162-163 and note 310.

63

Maaike Groot has studied ritual deposits of animal bones at Tiel-Passewaaij.476 She distinguishes five categories: complete or nearly complete skulls and skeletons, articulated limbs, combinations of the categories mentioned above, and large concentrations of disarticulated bones.477 In the settlement several special deposits have been found, like a pit with the skeleton of a crow, a horse’s skull and an iron knife, a pit with a horse’s skull and the skeleton of a dog, two horse’s skulls, two skeletons of dogs accompanied by large pottery sherds, and so on. She concludes that the practice of ritual deposits begins between 50 BCE and 40, rises to a peak between 150 and 270 and decreases after 270.478 Zennewijnen is situated southeast of Tiel-Passewaaij. In the filling of a ditch next to a late Roman house, a fragment of a lime stone altar came to light in 1930, during the mining of clay for a stone factory.479 The altar was dedicated to the goddess (I)seneucaega by a certain Ulfenus, beneficarius of the tribune of the Legio XXX Ulpia Victrix Severiana Alexandriana. Thanks to this honorary title, the dedication of the altar can be dated between 222 and 235, the reign of the emperor Severus Alexander.480 So far, this is the only known inscription addressing the goddess (I)seneucaega or Seneucaega. A salient detail is that the goddess on the altar is depicted as Diana.481 Northwest of Tiel at Kapel Avezaath, another altar came to light in 1954. It was dedicated to the Germanic goddess Hurstrga. Hurst in Germanic means ‘bush’ or ‘shrub’. The dedicator was a Batavian magistrate, a Valerius Silvester.482 Tiel-Passewaaij has yielded ten terracotta statuettes, including fragments of statuettes, and two bronze statuettes of Mars and Mercury (figs. 2.169-170). A third bronze statuette presumably depicts Diana and was discovered by a detectorist at the excavation site.483 All three bronze specimens were found in top soil. The identifiable terracotta statuettes include three heads, one of Cybele, one of Minerva and one of a mother goddess figurine, as well as the torso of a Venus statuette, the rear part of a lion and a fragment of the crest of a cockerel. The terracotta head of Minerva was recovered from a well (well 9), south of house 3, together with large pottery sherds and a metal fitting. The head of Cybele comes from a pit north of the same house. The head of a mother goddess was found in a ditch next to an outbuilding which has been interpreted as a stable.484 The Venus fragment comes from the spoil heap (figs. 2.65-68a). The heads ended up in the ground between 150 and 220 (Minerva and Cybele) and between 210 and 240 (mother goddess).485 See for the three statuette heads also chapter 4. • Tiel-Passewaaij-cemetery The burial site has produced two fragmentary stray finds, of which one depicts the crest of a cockerel.486 • Wijchen Eight kilometres southwest of Nijmegen, several excavations were carried out at the south side of Wijchen, on the north bank of a lake, Wijchens Meer. The lake was once part of a meander bend of the Meuse. During dredging in 1978, a large quantity of Roman artefacts came to light. In 1998 the municipality of Wijchen gave permission to investigate the bottom of the lake, which resulted in some 1400 finds.

476

Groot 2008, 117-158.

e /ex p(raecepto) eius /Val(erius) Silveste[r] /dec(urio)

477

Groot 2008, 120.

m(unicipii) Bat(avorum)/pos(uit) l(ibens) m(erito): AE

478

Groot 2008, 141 and tables 3.2, 3.3. See for the habitation

1958, 38 = AE 1959, 10.

phases: Heeren 2009, 49-74.

483

The statuette is inaccessible.

Heeren 2009, 27.

484

Heeren 2009, 311, table 32.

480

AE 1993, 157 = AE 1981, 657.

485

Heeren 2009, 63-70, 331, table 45. I thank Stijn Heeren for

481

See for statuettes of deities with a Roman appearance and

479

a native name 3.8.2. 482

64

Derks 2014b, 109.The inscription reads: Deae/Hurstrg(a)

the additional information. 486

Heeren 2009, 134 and 331, table 45.

A Roman period settlement at the site Martensterrein was occupied between 50 BCE and 100. Excavations at Heerenstraat and Molenveld produced traces of a late Roman period settlement and a Roman period burial site.487 Parts of a Roman period villa were uncovered at the site Tienakker in 1999-2000. Sand extraction operations had destroyed many traces and the excavators did not find the main building. They could identify two outbuildings belonging to the villa complex, the remnants of a burial mound and a structure that has been interpreted as an open-air sanctuary.488 The villa was abandoned around 280, after which a watchtower was built on the villa grounds in the early 4th century.489 The first building was situated in the southeast corner of the examined terrain. It was a rectangular building enclosing a well. West of this building, two pits were discovered.490 In the northernmost pit a decorated coarse ware jar was buried. The pit also contained an iron knife. In the other pit the remains of a wooden box with metal fittings were found. The box contained an iron knife and two plated denarii, one of Severus Alexander, struck in 222-228 and one of Maximinus Thrax, struck in 235-236. Originally, the two pits were interpreted as depositions with a ritual meaning.491 Later, Van Enckevort revised this interpretation and suggested that the two pits were cellars.492 If the pits were cellars, the two deposits may still have had a ritual meaning. In the cellars of several villas and large houses in France ritual deposits as well as household shrines have been found. The deposits consisted of different objects, but every deposit contained an iron or lead object.493 The majority of the over 1400 finds from the lake Wijchens Meer are potsherds dating between 70 and 170. The metal finds include some sixty coins, predominantly from the late Roman period, a spear head, a fragment of a stilus, a golden Stützarmfibel, a bronze finger ring and two bronze dishes.494 Although a part of the finds may have been refuse from the settlement(s) on the north bank of the lake, objects like the bronze dishes, the golden fibula and the coins most likely ended up in the water as part of a ritual. A square structure of 9 by 9 metres has been interpreted as an open-air sanctuary.495 Sherds from the postholes date the structure to the 3rd century. Inside, not far from the southwest corner, a 27 cm high tuff stone altar and a bronze Minerva statuette were found in a pit (fig. 2.69).496 A bronze Mercury statuette came to light at the north side of the lake.497 Another statuette comes from Wijchen-Alverna, where a bronze Victoria was discovered during the layout of asparagus beds.498

487

Seinen et al. 2013, 250-251.

493

Bailliot 2015, 104.

488

Heirbaut/Van Enckevort 2011, 25-48; Van Enckevort 2012,

494

Seinen et al. 2013, 256-258.

217-225.

495

Heirbaut 2011, 45-47. It is argued that the structure had no

489

Van Enckevort 2012, 225-228.

roof because of the shallow postholes. This does, however,

490

According to Van Enckevort’s reconstruction, the pits were

not exclude the possibility that it had a thatched roof. See

located inside the main building, close to what must have been the eastern wall. 491

for the interpretation also chapter 1. 496

Heirbaut/Van Enckevort 2011, 36; Van Enckevort 2012, 219-

46-48 and fig. 5.1; Van Enckevort 2012, 217-224 and fig. 114.

220. 492

Heirbaut/Van Enckevort 2011, 25-36; Heirbaut 2011, 25, See also 4.2.2.

The arguments for this new interpretation are the orien-

497

Zadoks/Gerhartl-Witteveen 1983, nr.206.

tation of the pits and their content: Van Enckevort 2012,

498

Zadoks et al.1969, nr. 62.

222-223.

65

Southern Gelderland: find-spots of statuettes without contextual data Find-spot

Material

Subject

Reference

Beuningen

bronze

Molossus (dog)

Swinkels 1993, 314-315499

Beuningen

bronze

Cerberus

Kleijen 2006, Swinkels 2015500

Nijmegen

bronze

Fortuna

RMO501

Nijmegen

bronze

Mercury

Zadoks 1969, nr. 42

Nijmegen

bronze

Mercury

Zadoks 1969, nr. 51

Nijmegen

terracotta

caricature, man

Van Boekel 1987, nr. 147

Nijmegen

bronze

Venus

RMO502

Nijmegen

terracotta

Risus bust

Van Boekel 1987, nr. 182

Nijmegen

bronze

Mercury

Zadoks 1969, nr. 47

Tiel-Badhuisplein

bronze

Mars

Zadoks 1969, nr. 33503

Varik

terracotta

pedestal

Van Boekel 1987, nr. 266

Southern Gelderland: uncertain find-spots of statuettes Find-spot

Material

Subject

Reference

Near Nijmegen

bronze

Cybele

Zadoks 1969, nr. 17

Near Nijmegen

bronze

Mercury

Zadoks 1969, nr. 43

Near Nijmegen

bronze

Mercury

Zadoks 1969, nr. 48

Rossum

terracotta

Venus, head

Van Boekel 1987, nr. 106504

499

The bronze figurine is a stray find, found near the village.

503

Swinkels has suggested that the dog is a Molossus, a breed

part of an assemblage found during the construction of

that originated from the northwest of ancient Greece and

a residential area in the 1920s. Besides the statuette, the

Albania, but nowadays no longer exists. Molossus dogs

deposit comprised a terra sigillata bowl, a bronze disc and

were mainly used to protect habitations and cattle. Some

23 fibulae. The composition suggests a deposition with

early Roman period furniture was decorated with Molos-

a ritual meaning rather than a precautionary cache. The

sus heads, like a table from Herculaneum and klinai with

2nd-century date of the bowl indicates that the objects

bronze or silver Molossus heads. See for the table from Herculaneum: Mols 1999, 46 and figs. 106-110. 500

Fig. 2.182. See for the interpretation 4.2.2. The statuette is

were deposited in or after the 2nd century. 504

Many Roman period finds have been collected at Rossum

Figs. 2.70-71. From the west side of the village. The bronze

on the bank of the Waal, including Roman coins and mil-

figurine depicts the three headed dog Cerberus. At the

itary stamps on bricks. Roymans identifies Alem/Rossum

surface traces of gold are still visible, indicating that the

with Grinnes and nearby Kessel with Vada, the two set-

figurine was gilded.

tlements that were attacked by Julius Classicus and Julius

501

RMO, inv. nr. e1947/3.18.

Civilis in 70: Roymans 2004, 144-146.

502

RMO, inv. nr. e1931/2.62.

66

Southern-Gelderland: statuettes found in riverside areas Find-spot

Material

Subject

Reference

Linge river

bronze

inscribed pedestal, Vagdavercustis

Zadoks 1969, nr. 81505

Rhine

bronze

Mercury

Zadoks 1969, nr. 52506

Waal

bronze

Mars

Zadoks 1969, nr. 34

Waal

bronze

Hercules

Swinkels 2017b507

Waal

bronze

Silenus

RMO508

Waal

bronze

dolphin

Zadoks 1969, nr. 67

Waal

bronze

pedestal

RMO509

Waal

terracotta

woman, head

Van Boekel 1987, nr. 197510

Meuse

terracotta

mother goddess, head

Van Boekel 1987, nr. 78511

2.3.5

North Brabant and Limburg

During the last glacial period, the Weichselien, the sandy and loamy soils of the southern Netherlands were covered with a layer of sand brought by the prevailing northwestern winds from the North Sea, which 12,000 years ago was not a sea, but a tundra. The sand ridges are still visible in the landscape, which today is characterised by subtle height differences and brook valleys running westwards towards the rivers. As a consequence of the rising sea level after the last glacial period, peat was formed in depressions in the landscape. One of the largest peat areas is situated on the border between North Brabant and Limburg, known as De Peel, nowadays partially a nature reserve. During glacial periods, frozen rocks were crushed and pulverized by the ice masses and the northwestern wind spread the fine particles from the dry sea and northern river bottoms over land. These fine particles, known as loess, were part of the famous fertile loess belt, of which a part runs through Dutch Limburg. According to Caesar, before his campaigns, the territory of the Eburones covered a large part of what are now the southern Netherlands, northern Belgium and the Rhineland north of Bonn. The Rhine constituted the northern and eastern border of their territory.512 After the disappearance or destruction of the Eburones, 505

Found in 1842 between Hemmen and Indoornik (Heteren,

Hupperetz 2014, 108. However, high-ranking officers were

nowadays belonging to the municipality Overbetuwe).

not the only people who could offer large bronze statu-

The dedicator of the Latin inscription is a decurio of the

ettes. See for different views on the hypothetical location

ala Vocontiorum, who addresses the native war goddess

506

of Carvium Byvanck 1943, 392-394; Van Es 1981, 103-104.

Vagdavercustis. The statuette itself has not been found. See

507

Figs. 2.61-64.

also 3.8.2.

508

Fig. 2.72.

According to the 1969 catalogue by Zadoks et al., the

509

Fig. 2.168.

statuette was found around 1925 ‘in one of the rivers near

510

Found between the Waal bridge and the railway bridge at Nijmegen.

Nijmegen’ The statuette is 23 cm in height and features in a museum catalogue of the Allard Pierson Museum at

511

together with pottery that could be dated to the 2nd or 3rd

Amsterdam. It appears from the catalogue that the find-

century.

spot of the statuette is De Bijland near Lobith, some twenty kilometres east of Nijmegen, on the right bank of the

Found in a dredging pit on the north bank of the Meuse,

512

Caes. BGall.5.24.4: Unam legionem, quam proxime trans

Rhine. The authors argue that the statuette was an offering

Padum conscripserat, et cohortes v in Eburones, quorum pars

made by a high ranking military officer who was based

maxima est inter Mosam ac Rhenum. Caes. BGall. 6.5.4: Erant

at Carvium, a Roman fort at the bifurcation of the rivers

Menapii propinqui Eburonum finibus, perpetuis paludibus

Rhine and Waal, close to the dam Drusus constructed at

silvisque muniti. See also Roymans 2004, 23, note 84.

the beginning of the 1st century: Van Oppen De Ruiter/

67

Map 10 Find-spots of statuettes in North Brabant and Limburg Base map: paleogeographic reconstruction of the Netherlands around 100 (Vos/Vries 2013). 1. Buchten 2. Kessel 3. Helden 4. Venlo 5. Escharen 6. Velp 7. Heeswijk 8. Empel 9. Kerkrade 10. Montfort 11. Maasbracht 12. Baexem 13. Melick 14. Baarlo 15. Hoogeloon 16. Veldhoven 17. Son en Breugel 18. Berghem 19. Oss 20. Bergen op Zoom 21. Maastricht 22. Margraten 23. Voerendaal 24. Heerlen 25. St. Oedenrode 26. Cuijk 27. Aalburg

new groups came to live in the region.513 One of the groups that settled in the area that now covers North Brabant probably were the Frisiavones. Their neighbours must have been the Texuandri, who, according to Pliny the Elder, lived ‘at the beginning of the Scheldt and had several names’.514 This probably means that 513

Roymans 2004, 23-26 and figs 3.1-3.2.

514

Plin. HN 4.17.106: a Scaldi incolunt extera Texuandri pluribus nominibus.

68

the group consisted of several branches. Besides the passage in Pliny, the Texuandri are known through two inscriptions. The first occurs on the Tropaeum Traiani, the war memorial at Adamklissi (modern Romania), the other on an altar from Carrawburgh (Northumberland) on Hadrian’s Wall.515 It is assumed that the territory of the Texuandri was bordered by the rivers Meuse, Scheldt and Demer, but the exact boundaries are not known. The relocation of the Ubii in the early Roman period from the region east of the Rhine to the west could imply that they lived in the north-eastern part of North Brabant and southern Limburg as well. For the Iron Age and Roman period, research in the southern Netherlands has primarily focused on the development of settlements. From the 1920s onwards, large scale excavations were carried out. Important projects were the excavations at Oss-Ussen and the Kempen project.516 Initially, the latter project focused on the area southwest of Eindhoven, the excavation of the villa at Hoogeloon in particular. Later, more sites were examined, including Empel (North Brabant) and Weert (Limburg).517 A fully published villa is the villa at Kerkrade-Holzkuil (Limburg).518 The excavations of the villa at Hoogeloon-Kerkakkers and Maasbracht have recently been published.519 Several other villas and monumental houses have been identified, some of which have been partially or fully excavated.520 North Brabant • Bergen op Zoom Excavations between 2000 and 2007 in the city centre of Bergen op Zoom uncovered a remarkable find-complex in the filling of a shallow depression. In antiquity, the depression was a fen.521 The finds could not be linked to a settlement or burial site and comprised a large quantity of sherds of miniature amphorae, medium-sized and large amphorae, fourteen small fragments of terracotta figurines, 76 coins, three glass fragments, a few amber beads and bone fragments. The amphorae found during the excavation probably were only a part of the total number of deposited vessels. At least 306 miniature amphorae and 29 larger amphorae could be identified.522 Offerings of miniature vessels are a widespread tradition and have been found in antiquity throughout Europe and the Mediterranean. Just like figurines of animals, mini-vessels are often seen as a replacement for ‘the real thing’. The fen and the finds suggest that the site was a cult place in the Roman period. The majority of the vessels were locally made. Based on the imported pottery and coins, the deposition of objects started before or around the middle of the 2nd century.523 There are reasons to believe that the amphorae were deliberately broken and subsequently removed and buried.524 Whether the terracotta figurines were deliberately broken as well, is not clear. The fragments were too small to allow for any identification of depicted figures or animals, with the exception of one fragment. It depicts the lower part of a standing male figure with four cylindrical objects, wooden barrels, piled up at its left side, and other objects placed on top of the barrels. It is Sucellus, a Celtic deity associated with fertility and forests, but also with the manufacture of wine and beer barrels. 525 He is often depicted with barrels, a hammer and/or a cornucopia.

515 516

Roymans/Derks 2015, 19. See for Oss-Ussen Van den Broeke/Van der Sanden 1987,

the site catalogue Habermehl 2011, 205-214. 521

Wesselingh 2000 and this chapter. For the Kempenproject: 517

my attention.

Slofstra et al. 1982, 1985.

522

Niemeijer 2009, 11-17, 23.

Roymans/Theuws 1993, 7-8. See for Weert Hiddink 2003,

523

Niemeijer 2009, 25 and note 35. See for the miniature ves-

2010. 518

I thank Ryan Niemeijer for bringing the find-complex to

sels also Roymans/Derks 2015, 240-241 and fig. 8.

Tichelman 2005.

524

Niemeijer 2009, 29, 41-44.

519

Hoogeloon: Roymans et al. 2015. Maasbracht: Vos et al. 2017.

525

Niemeijer 2009, 23, note 28. See for a comparable speci-

520

See for a list of villas and large houses in the loess region

men: De Beenhouwer 2005, 237, nr. 338.

69

• Berghem-Lallenberg During an excavation in 1955, traces of habitation and burials were found. The finds from the site could be dated between 180/190 and 270 and included the upper part of a mother goddess figurine with an inscription at the back: ‘Fidelis f(e)cit’. The mother goddess figurine was a stray find, but may have been a funerary gift belonging to the cremation grave that was found at the site.526 • Cuijk Situated on the left bank of the Meuse River, where the road between Nijmegen and Blerick crossed the road to Xanten, the strategic location of Cuijk must have attracted the attention of the Romans at an early stage. However, despite two or three ditches from the 1st century, convincing evidence of an early Roman fort at Cuijk is still lacking.527 In the 2nd century, Cuijk was a large settlement, covering around ten hectares, without any military architecture. Traces of two roads were uncovered, under the Grotestraat and northeast of the St. Martinus church. Two rectangular buildings have been interpreted as Gallo-Roman temples. They were constructed in the second half of the 2nd century. At the Grotestraat the foundations of a 3rd-century building, most likely a bathhouse, were uncovered. The strategic location of the settlement, its dimensions, the alleged temple complexes and bathhouse suggest that Cuijk was an important regional centre in the hinterland of Batavian territory. It has also been suggested that Cuijk was the civitas capital of the Frisiavones.528 At the beginning of the 4th century, during the reign of Diocletian (283-305) or Constantine (306-337), forts were constructed at several locations in Germania, including Nijmegen and Cuijk.529 Despite the central function of the settlement and its presumed temples, so far only one bronze Mercury statuette (fig. 2.73) and one terracotta fragment have come to light. The terracotta fragment belonged to a figure that probably represented Fortuna or Minerva.530 Both were found along the Grotestraat. The burial site has not yielded any statuettes. • Empel The discovery in 1986 of a substantial number of Celtic and Roman coins, fibulae and other metal objects, as well as tuff stone fragments, prompted excavations at Empel between 1989 and 1991.531 The excavations revealed traces of a pre-Roman cult place and a Gallo-Roman temple. The earliest phase of the cult place, most likely an open-air sanctuary, could be dated between around 100 and 15 BCE.532 In the Flavian period (69-96) a monumental stone-built Gallo-Roman temple replaced the open-air sanctuary. Although the temple was at least partly destroyed by fire in the late 2nd century, the place maintained its ritual function until around 235.533 Among the offerings at the sanctuary, many metal objects date to the 1st centuries BCE and CE. After the 1st century, the number of metal objects strongly decreased. The excavators assume that dedicators began to offer typical Roman objects like votive altars and sculpture, which have disappeared in later times, because they were melted down or burnt in lime kilns.534 The offered metal objects include over 140 examples of militaria and pieces of horse gear.535 Besides fragments of a few unidentifiable bronze statues such as a half life-size and maybe a life-size bronze statue, a 42 mm high bronze right booted foot came to light, as well as an applique of the goddess

526

Van Boekel 1987, nr. 84, Bogaers 1970b, 64-67.

530

Van Enckevort/Thijssen 2002, 69.

527

Haalebos et al. 2002, 21-24.

531

Roymans/Derks 1992, 423.

Haalebos et al. 2002, 29-30. See for Ouddorp as an alterna-

532

Roymans/Derks 1994, 14-19.

tive location of the civitas capital of the Frisiavones this

533

Roymans/Derks 1994, 19-25.

chapter.

534

Roymans/Derks 1994, 30.

Haalebos et al. 2002, 24-42.

535

Van Driel-Murray 1994, 92, Nicolay 2007, 120-124.

528

529

70

Luna and two small bronze heads, one of a bull and one of a male figure.536 Examples of comparable heads have been recovered from Iron Age contexts. They were once attached to objects like knives. Whether the heads have been deliberately separated from the objects to serve as offerings is unknown, but there might be a link between the native ‘cult of the head’ and the loose heads found in and around settlements, as will be argued below.537 The foot was part of a relief, and if it depicted a standing figure, it must have been 25-30 cm in height. The boot is an extraordinary closed specimen, bordered with the skin of a feline and a feline’s paw resting on the ankle. The heel is decorated with an engraved acanthus leaf, a lion’s head and claws. Besides emperors, deities like Mars, Diana and Bacchus are sometimes depicted wearing boots.538 Another example of a leg with a boot decorated with lion skin was part of a relief and has been found in a cellar of a building at Maastricht (see below). This figure has been interpreted as Diana.539 It has been advanced that the sanctuary and the offerings were dedicated to Hercules Magusanus. This is based on two finds. The first is a thin, silver or tin plated bronze plaque measuring 6.5 by 7 centimetres with an inscription addressed to ‘Hercules Magusenus’. The dedicator was a Julius Genialis, veteran of the Legio X Gemina.540 Originally, the plaque was attached to the base of a statue, large statuette or bust. The second find is a bronze statuette of Hercules, a stray find from the site.541 Like the specimens from Oosterhout and Velsen (see below) the figurine depicts Hercules holding a drinking cup, a Hercules bibax.542 It has been argued that in this sanctuary, returning veterans offered their gear and/or weaponry to Hercules Magusanus, not only as a token of their gratitude for his protection during their service, but also as a sort of rite de passage, marking their return into civil life.543 However, the offering of weapons is also known from other sanctuaries, like the sanctuary at Kalkar (Nordrhein-Westfalen), just across the German border, which was dedicated to the Germanic goddess of war Vagdavercustis.544 A bronze pedestal with an inscription addressing the same goddess comes from the Linge river (Gelderland).545 Therefore, it cannot be ruled out that the militaria and weapons were offered to another deity or more deities than solely Hercules Magusanus.The booted foot and the ‘loose’ heads of Luna, the male figure and bull suggest that also other deities were worshipped at Empel.546 The site has not yielded any other statuettes. • Esch-Hoogkeiteren Ten kilometres south of Den Bosch (’s-Hertogenbosch) Roman pottery was found during ploughing. Excavations between 1952 and 1961 at Kollenberg and Hoogkeiteren, produced a pre-Roman urn field and two Roman period cemeteries. The burial site at Hoogkeiteren yielded four tumuli and four rich graves. The tumuli were constructed between 175 and 250.547 In one of the graves, several amber objects were found, including a spindle, a Bacchus figurine and a satyr, as well as two pendants depicting stylized gladiator helmets.548 The grave could be dated to the first or second quarter of the 3rd century.549 Besides the amber spindle, jewellery, beads and slippers suggest that the deceased was a woman.550

536

Swinkels 1994, 86-90.

542

See for the iconography of Hercules 3.8.3.

Swinkels already suggested this possibility in the 1990s:

543

Van Driel-Murray 1994, 106.

Swinkels 1994, 89-90.

544

Bödecker 2010, 16-19.

538

Swinkels 1994, 86-89.

545

See also this chapter, 3.8.2 and 4.2.2.

539

Panhuysen 1984, 50 and fig. 14.

546

Hercules is usually depicted barefoot.

The name Magusanus is more common. The inscription

547

Van den Hurk 1986, 381-385, 9-12.

reads: Herculi Magusen(o) Julius Gen/ialis veter(anus)

548

Van den Hurk 1986, 112-115, 133, plate VI and 391, plates 4-5.

leg(ionis) X G(eminae) P(iae) F(idelis) v(otum) s(olvit)

549

Van den Hurk 1986, 122-123.

l(aetus) l(ibens) m(erito): AE 1990, 740 = AE 1994, 1281 = AE

550

Van den Hurk 1986, 18-19. See for the qualities of amber

537

540

2000, 1011. Roymans/Derks 1994, 26. 541

3.7.5.

Swinkels 1994, 84.

71

• Hoogeloon-Kerkakkers Hoogeloon is situated 15 kilometres south west of Eindhoven (North Brabant). Kerkakkers is a stretch of farmland located on a cover sand plateau. The top layer of the plateau consists of relatively fertile soil, which explains the presence of four Roman period settlements on the plateau.551 The discovery of a Roman period villa at Kerkakkers in 1979 prompted excavations between 1980 and 1987. The main building was constructed around 100, the baths at the end of the 2nd century. It is assumed that the complex fell into decline in the early 3rd century. After that, it remained partially occupied. Also, traces of 34 Roman period farmhouses were uncovered, as well as 14 outbuildings, nine wells and many pits and postholes. Around twenty farmsteads date to the 1st and 2nd centuries, the rest to the 3rd century. The settlement remained occupied until the last quarter of the 3rd century.552 A square terrain with an enclosure ditch, two rows of deep postholes and several pits, has been interpreted as the remains of a sanctuary. It was located 60 metres southwest of the settlement.553A fragment of a terracotta figurine was found in the enclosure ditch of the presumed sanctuary. It depicts a garment with drapery folds.554 Four fragments of a terracotta statuette depict one of the Parcae, the female personifications of destiny.555 The fragments were recovered from a drainage ditch of the baths which are dated to the late 2nd or early 3rd century. From the same site comes a terracotta pedestal, probably belonging to a Venus figurine, found in a shallow pit.556 Two fragments of an unidentifiable enthroned figure were found in the top soil during the excavations in 1985.557 • Hoogeloon-cemetery Kaboutersberg East of the plateau with the villa complex, a cemetery was excavated in 1988. Although a burial mound on the site was thought to be prehistoric, it appeared to be a Roman period tumulus, of which nothing had been preserved but the foundations of the tomb and a pit with limestone fragments of a so-called tower tomb. Two cremation burials have been preserved, which date from a later period: two pits with the remains of a child between seven and fifteen years old and a woman in her thirties or forties. Besides potsherds, the grave gifts included an oil lamp and a patera.558 The vicinity of the villa suggests that the villa owner(s) commissioned the construction of the tower tomb and the tumulus somewhere in the 1st or 2nd century.559 In 1986, not far from the site, a fragment of a mother goddess figurine was found.560 • Oss-Westerveld Excavations at Oss-Ussen between 1976 and 1986 revealed two burial sites and traces of habitation from the early Iron Age onwards, as well as three Roman period settlements.561 Excavations in 1990-1992 yielded a fourth settlement: Westerveld.562 It was the largest of the four, covering a surface of 7.5 hectares. The settlement was inhabited from the late Iron Age into the early 3rd century.563 In total, 37 house plans, outbuildings, granaries, 131 pits and wells were unearthed. A square ditch measuring 45 by 45 metres may have enclosed a sanctuary.564 The influx of imported goods began in the Augusto-Tiberian period and included imported pottery, glass vessels, jewellery and wine barrels.

551

Hiddink 2015, 87.

557

Van Boekel 1987, supplement nr. 11.

552

Hiddink 2015, 103-104, 116-118, 121.

558

Hiddink 2015, 89.

553

Hiddink sees a resemblance to the funerary monuments at

559

Roymans 2015, 138.

Someren and Oss-Ussen: Hiddink 2015, 114, note 58 and figs.

560

Van Boekel 1987, supplement nr. 5.

2 and 18.

561

Van der Sanden 1987, 59-66.

554

Van Boekel 1987, supplement nr. 13.

562

Wesselingh 2000, 171-182.

555

The terracotta statuettes are not mentioned in the publica-

563

Wesselingh 2000, 158-159.

tion about the villa by Roymans et al. 2015.

564

Van der Sanden 1987, 61. Wesselingh 2000, 71.

556

72

Van Boekel 1987, nr. 130 and supplement nr. 15.

During excavations in 1977 a pedestal of a terracotta statuette was found in the top filling of a well that was in use in the 2nd century.565 A fragment of a terracotta mask was found in a well near one of the houses.566 • Son en Breugel Excavations in 2016 at Son en Breugel yielded a Roman period settlement. One of the houses had a sunken byre at the east side. The excavators date the house between the last quarter of the 2nd and the beginning of the 3rd century. The sunken byre was 48 cm deep and contained 1200 potsherds, a lot of iron slag, fragments of Roman ceramic building material and the terracotta head of a mother goddess figurine.567 Technical analysis of the iron slag has made clear that metalworking took place in the settlement, but not in the sunken byre itself.568 The statuette head came to light in the southeast corner of the byre, in a layer with many finds (fig. 2.74).569 It was once part of a mother goddess suckling two infants and strongly resembles the head of the mother goddess from Rockanje ( South Holland).570 North Brabant: find-spots of statuettes without contextual data Find-spot

Material

Subject

Reference

Aalburg

bronze

Mercury

Detector Magazine 134, 2014

Aalburg-Genderen

terracotta

woman

-571

Aalburg-Genderen

bronze

Minerva

-572

Escharen

bronze

Hercules

-573

Heeswijk

bronze

African man

-574

St. Oedenrode

bronze

Mercury

Zadoks 1969, nr. 49

St. Oedenrode

terracotta

Fortuna

Van Boekel 1987, nr. 28; Beex 1973, 179, site 122575

Velp

bronze

Mars

-576

Limburg • Buchten A rescue excavation was carried out in 1976 at the location Sluisweg, west of the village of Buchten, where a business park was planned. Based on finds like pottery and Roman period building material, the presence of a Roman period villa was suspected. An extraordinary find from the 1976 excavation was a Roman period 565

Van Boekel 1987, nr. 164. Van Boekel dates the figurine

573

Stray find from a field near Escharen. I thank Godfried

to the 2nd century, arguing that the date is based on the

Scheijvens for bringing the statuettes from Heeswijk and

period the well was in use: Van Boekel 1987, 622. Since the

Escharen to my attention. There are reasons to doubt the

statuette was found in the top filling of a well, it probably was deposited in the well when it fell out of use.

authenticity of the Hercules statuette from Escharen. 574

Figs. 2.75-76. The bronze statuette was found in 2009 in

566

Wesselingh 2000, 132-133.

a ploughed field by a detectorist, William Offermans.

567

I thank Margriet Tump (BAAC) for bringing the terracotta

This area has been severely disturbed by works in 1934, to

figurine to my attention and for the documentation of the

straighten the meanders of the river Aa. See for the statu-

excavation. 568

Information by BAAC.

569

See for the interpretation 4.2.2 and 4.2.4.

570

This chapter.

571

I thank Ronald Louer for the information on the statuettes.

572

Detector find: Meffert/Scheijvens 2010, figs.154-156.

ette also 3.8.16. 575

The head of the Fortuna statuette was found in a field on the right bank of the Dommel, where traces were found of a Roman period settlement.

576

Detector find. Private collection.

73

bronze cockerel, inlaid with enamel. The statuette was found together with its bronze pedestal. Although it is one of the top items of the collection of the Limburgs Museum, until 2015, very little was known about the results of the excavation, the find-context of the cockerel and the object itself.577 Recently, the results of the excavation at Buchten have been re-examined and published. The most important features that were uncovered were the traces of a rectangular structure, measuring 35.7 by 25.9 metres. Since the ground plan of the structures did not resemble the layout of a villa, it was later advanced that the outer structure was the enclosure wall of a temple precinct. In the most recent publication on Buchten, the remains of the inner structure have been interpreted as an open walled court within the enclosure wall of a sanctuary. Coins and potsherds indicate that the sanctuary was in use between around 50 and 250.578 The sanctuary was possibly dedicated to the goddess Arcanua, since two finds bear an inscription addressed to the previously unknown native goddess. The first is a thin, bronze leaf shaped object, a palmetto, dedicated by a M(arcus?) I(ulius?) Am(---).579 The second is the pedestal that belonged to the bronze cockerel. The dedicator is a veteran of the Sixth Legion, Ulpius Verinus.580 Nine comparable statuettes of cockerels have been found: six in Britain, one in Cologne (Germany), one in Tongeren (Belgium) and one in Ezinge, Groningen.581 The cockerels were probably made in Britain by the same workshop, at the end of the 2nd or the beginning of the 3rd century.582 The cockerel and the pedestal were found in a pit outside the sanctuary, at the southeast side near the enclosure wall. Besides the cockerel and the pedestal, the pit contained five other metal objects: a fibula, a silver ring with a gem, a necklace and two coins.583 It has been suggested that the objects were buried to make room for new offerings in the sanctuary.584 This explanation seems not very plausible. First, because the objects were buried outside the precinct, whereas gifts to the god(s) were usually buried within the temple precinct. Second, if it was necessary to tidy up the sanctuary in order to make room for new donations, the pit would have contained more objects. The composition of the deposit gives the impression that the metal objects were not a precautionary cache but carefully selected. The fibula, ring and necklace could be worn together. Therefore, it seems more plausible that the jewellery, coins and inscribed statuette were an offering to the goddess Arcanua. Why the pit was dug outside the sanctuary remains the question. The latest of the two coins in the deposit was struck between 238 and 244 and provides a rough terminus post quem for the deposition around or, more likely, after the middle of the 3rd century. Since the sanctuary was in use until the middle of the 3rd century, the deposit may have been made after the sanctuary was abandoned. If so, it suggests that the place still maintained a special status. It would explain the location of the deposit outside the precinct, but still near the enclosure wall.585 • Heerlen: Coriovallum In the 19th century, the priest Joseph Habets noticed that Coriovallum on the Peutinger map had to be Heerlen. The construction of new residential areas around the turn of the 20th century brought to light many traces of Roman period buildings. Baths were discovered in 1940 and excavated by volunteers. When the largest part of the complex was already uncovered, the archaeologist Van Giffen was called in to uncover the 577

Derks/De Fraiture 2015b, 9-13.

581

See for the site Ezinge 2.2.2.

578

Derks 2015a, 37-46.

582

Hoss 2015, 169-171.

Hoss 2015, 118-119. The inscription reads: D(eae)/Arkanuae

583

Derks 2015a, 45-46, fig. 5.3 and appendix VI. The latest coin

579

(!) M(arcus) I(ulius?) Am( )/l(ibens) m(erito): AE 1983, 724

was struck under Gordian III (AD 238-244): Aarts 2015, 110-

= AE 1989, 558. 580

584

Derks 2015a, 46.

veteranus leg(ionis) VI v(otum) l(ibens) m(erito): AE 1983,

585

See also 4.2.2.

723.

74

111.

The inscription reads: Deae Arcanu(a)e Ulpius/Verinus

rest. He published his findings in 1948 and concluded that the baths were constructed around 40.586 Another excavation was carried out in the 1950s at the north and east side of the baths. Recently, the early excavations have been reassessed and processed.587 The authors come to the conclusion that the bath complex was constructed in the Neronian-early Flavian era. It underwent several alterations and its surface area expanded from 500-2000 m2. It remained in use until the 5th century.588 The vicus of Coriovallum was located at the crossing of two roads, one connecting Aachen and Xanten, the other running between Cologne and Boulogne-sur-Mer. The estimated surface area of the vicus must have been around 40-50 hectares, with four cemeteries, two of which were at least half a kilometre long. So far, archaeological evidence such as the bath house and strikingly large buildings suggests that Coriovallum may not have been just an ordinary roadside settlement. It is even argued that it might have been the capital of the Baetesi.589 The two identifiable bronze statuettes from Heerlen represent Mercury and Apollo. Also, a bronze left hand was found, 5.5 cm in length, of a statuette depicting an emperor or deity.590 A fragment of a bronze gilded staff, 23 cm in length, was found in the southeast corner of the eastern palaestra, in the vicinity of the last praefurnium (fig. 2.179).591 The staff probably belonged to a statue of Mercury, Bacchus or Jupiter. Mercury is depicted with a caduceus, Bacchus with a thyrsus and Jupiter with a sceptre.592 The statue probably was at least half-life-sized. The terracotta statuettes comprise two Venus statuettes, one Fortuna, one Minerva, one Juno, one Apollo and one fragment of the tail of a cockerel or hen.593 • Heerlen-cemetery Valkenburgerweg Two of the four cemeteries have yielded statuettes. From a cemetery west of Coriovallum at the north side of the Valkenburgerweg comes a terracotta statuette of a standing woman in a long, hooded mantle. She holds a bowl in the left hand and another object in the right. Although the head is missing, a parallel from Trier shows a woman with an elaborate coiffure of plaits. The statuette depicts a goddess or a mortal woman. Based on parallels from Trier, the statuette probably dates to the 3rd century.594 • Heerlen-cemetery Landgraaf From a grave at the location ‘De Landgraaf’, a cemetery east of Coriovallum, comes an amber statuette depicting a sleeping Amor. The grave can be dated to the 2nd century, since another funerary gift was a type of oil lamp that was made around 100 or later.595 • Helden-Schrames Archaeological research was carried out at Helden-Schrames between 2004 and 2008.596 The excavated house plans comprised one early Roman period house, five dating to the middle Roman and seven to the late Roman period.597 One of the houses, H9, with a date between 150 and 270, had a sunken byre at the east side.598

586

See for Van Giffen’s suggestion for the construction date of

592

the baths: Van Giffen/Glasbergen 1948, 231-233, 254-255. 587 588

In the Roman period, Bacchus’ followers are seldom depicted with a thyrsus: Veen 2014a, 131-135.

Jeneson/Vos 2020.

593

Van Boekel 1987, nrs. 1,44, 59, 114, 117 and 216.

Jeneson/Vos 2020, 191-193.

594

Van Boekel 1987, nr. 131.

589

Jeneson/Vos 2020, 198.

595

RMO, inv. nr. i1930/3.17. I thank Ruurd Halbertsma (RMO)

590

Zadoks et al.1969, nr. 3. The Apollo statuette was found between Heerlen and Locht, near the Roman road between

for the information. 596

591

I thank Juliëtte de Winter (BAAC) for the additional information on the excavation.

Heerlen and Aachen. See for the statuettes Zadoks et al.1969, nrs. 3, 41, 79. I thank

597

De Winter 2010, 91-107.

Karen Jeneson for the information and photo of the staff.

598

De Winter 2010, 99-102 and fig. 7.4.

75

House H9 yielded 71 metal objects, such as nails, keys, hooks, strips, a finger ring with gem, a piece of horse gear, a bucket handle and a chain. Besides metal slag, the sunken byre produced nails, a bucket handle in the shape of a dolphin, a balance scale (Roman or Frankish), a Frankish iron throwing axe and a bronze statuette of Mars, 136 mm in height.599 The majority of these objects could be dated to the middle Roman period. Apart from a fragment of melted bronze, other evidence of metallurgy was absent.600 The bronze Mars statuette was found at the bottom of a large pit in the southernmost corner of the sunken byre, just above virgin soil.601 In the same area the iron throwing axe came to light. Although the presence of metal slag, the balance scale and other metal objects in the sunken byre are indicative of metalworking in the settlement, the Mars statuette was found at the bottom of a pit. It seems, therefore, less plausible that it was buried with the intention to be melted down.602 With the terracotta mother goddess from Son en Breugel (see above), the Mars statuette is the second statuette in the study area that has been recovered from a sunken byre. • Kerkrade-Holzkuil Archaeological research carried out in 2001 at the north side of Kerkrade revealed traces of a Roman period villa. The villa was excavated in 2002 and published in 2005.603 The excavators distinguish five phases for the habitation history of the complex, from 70 until the end of the 3rd century. In the second phase the main building was replaced by a stone-built house. The enclosed terrain was at its largest at the end of the 2nd or the beginning of the 3rd century, covering circa 4.2 hectares. A bathhouse with a caldarium, frigidarium and apodyterium was built at the east side of the villa.604 A fragment of a terracotta statuette, the head of a smiling boy of the Risus type, was found in a pit in front of the southwest wing of the villa complex, together with sherds of a pre-Flavian bowl (Dragendorff 27) and cooking pots dating to the 2nd and 3rd centuries.605 • Kerkrade-Kaalheide During excavations on the Spekholzerheide plateau in 1936, a rectangular house plan was uncovered, measuring 8.8 by 17.8 metres. It was interpreted as one of the outbuildings belonging to a villa complex. The main building was identified in 1941 and excavated in 1950. Also an earlier, post-built structure was uncovered, constructed at the end of the 1st century. The villa complex remained inhabited until the early 3rd century. The complex was enlarged and renovated at least twice and measured 51.5 by 21.8 meters at its largest. During the excavations in 1950, a terracotta bust of a boy was found.606 The Risus type bust is a rattle, since the head holds a pebble.607 Another terracotta figurine, a bear, had come to light during the excavations in 1936. Apparently, the bear ended up in a private collection, which is why it does not feature in the catalogue of the museum Ehrenstein at Kerkrade. Nevertheless, the figurine is mentioned in an inventory list of finds from the excavations in 1936.608 • Kerkrade Maar-West In 1997, a terrain at Kerkrade Maar-West was investigated since the presence of a Roman villa was suspected. The excavation revealed four Roman period farmsteads. The expected main building of the villa complex was not found. The settlement was occupied between 170 and 240-250. 599

De Winter 2010, 277 and fig. 5.1.

600

De Winter 2010, 272-274.

601

De Winter 2010, 281, fig. 15.5.

602 603 604

Tichelman et al. 2005, 49-87.

605

Wiepking 2005, 193-194.

76

606

Brunsting 1950, 30-31; Koster et al. 2002, 48-49. The findspot and find circumstances of the bust have not been documented.

See also 4.2.2 and 4.2.4.

607

Van Boekel 1987, nr. 183.

Tichelman et al. 2005.

608

I thank Diane Smeets-Habets (municipality of Kerkrade) for bringing this figurine to my attention.

A remarkable feature was a pit at the west side of the terrain, measuring two by two metres. It had a timber lining and contained nine complete and near-complete pottery vessels as well as a terracotta figurine of a bear (figs. 3.11-12).609 The excavators interpreted the contents of the pit as a special deposit.610 One of the pottery vessels in the pit, a Niederbieber 33 beaker, could be dated to the 3rd century, indicating that the objects were buried in the 3rd century or later. • Maasbracht-Steenakker Before the construction of a new residential area, excavations produced traces of a Roman villa as well as habitation traces from the Iron Age and Medieval period.611 In the late 1st century, a large rectangular building was constructed. The main building measured 17 by 36 metres and had a porticus at the south side. At the end of the 2nd century, the complex was restructured. A new wing was added at the east side and a tower at the west side. The complex measured 30 by 50 metres.612 The main room of the building was decorated with wall paintings and several rooms were equipped with a hypocaust.613 The villa was abandoned or destroyed around 270. The head of a Risus bust was found in the cellar under the main building of the villa. The finds from the cellar could be dated to the late 2nd and early 3rd centuries.614 The bust is the third specimen of the Risus type found in a villa in Dutch Limburg. • Maastricht: Mosa Trajectum? The presence of the rivers Meuse and Jeker, the construction of a road which connected Cologne and Boulogne-sur-Mer, as well as the town Atuatuca Tungrorum, modern Tongeren, founded around 10 BCE, may have been the decisive factors for the foundation of the vicus Maastricht. The road ran via Maastricht to Tongeren and the Jeker was used for the transportation of building material to the town, situated 19 kilometres to the west.615 Where the river Jeker discharges into the Meuse, the Meuse could be crossed safely. Although it is said that the Latin name for Maastricht is Mosa Trajectum, the name appears for the first time in the early Medieval period. It is not indicated on the Peutinger map and Greek and Latin sources do not mention the name. Tacitus only mentions a pontem Mosae fluminis, but it is not clear whether this is the bridge that was found during dredging operations in the early 1960s.616 At the time, the remains of the bridge were dated to the late Roman period, but dendrochronological analysis has demonstrated that the bridge had known at least two earlier phases of which one dates to the first half of the 1st century.617 The foundation date of the vicus is uncertain. Traces of prehistoric habitation have been found at a depth of five to six metres under street level in the modern city centre, which makes it difficult to examine a possible continuation into the Roman period. Traces of a settlement that could be dated to the 2nd and 3rd centuries have been found on the left bank of the Meuse, north of the Onze Lieve Vrouwe-church, in the area that is bordered to the west by the Havenstraat and to the east by the Meuse. Construction works in 1840 revealed a part of a Roman bath complex in the city centre, between Stokstraat and Havenstraat. Excavations in the 1950s and 1960s uncovered not only remains of the baths which were already discovered in the 19th century, but also new structures and buildings belonging to the baths. The baths were probably constructed in the first part of the 2nd century and fell in disuse at the end of the 3rd or in the course of the 4th century.618

609

2017a, 83-127.

Dijkstra 1997, 9-17; Wiepking 1997, 18-22 and fig.10. See for the meaning of bear figurines also 3.5.9.

614

Van Boekel 1987, nr. 184.

610

See for the deposit also 4.2.2.

615

Nouwen 2012, 65.

611

Vos et al. 2017.

616

Tac. Hist. 4.66.

612

Willems 1988, 28-33.

617

Panhuysen 1996, 22-25.

613

See for an interpretation of the figural scenes Swinkels

618

Panhuysen 1996, 32-43.

77

North and south of the bath complex, more structures have been identified. During excavations in 1983 at the location Planckstraat-Hotel Derlon, the northern wall and a part of the western wall of a sanctuary or temenos were unearthed northwest of the Onze Lieve Vrouwe-church. The temenos wall was built in the first half of the 2nd century. Shortly after the middle of the 2nd century, the building was restructured. A monumental Jupiter column was erected in the middle of the courtyard. At the beginning of the 3rd century, the sanctuary fell into disrepair.619 Along the Roman road to Tongeren, traces of wooden buildings were uncovered during excavations in 1969. One of the buildings had a cellar, measuring 3.31 by 3.1 metres, constructed with local limestone. In the cellar three fragments of bronze statues came to light: the left leg of a striding figure, which must have been around 70 cm high, and two large fragments of garments, one belonging to a life-sized statue, the other to an almost life-sized statue.620 Besides the remnants of the Jupiter column, limestone remains of monumental funerary monuments, columns, grave stones, altars, sculptured architectural elements and fragments of statues came to light.621 The majority were recovered from the Meuse, where the limestone objects had been used for the construction of a dam.622 Sculptural fragments of identifiable deities include a column with Fortuna and Diana, a column with Virtus, Mars, Hercules and Roma or Honos, a fragment of a column with Bacchus, a column fragment with a portrait of, probably, Hercules, fragments of a Jupiter statue, a Jupiter column with Jupiter, Fortuna, Apollo and Hercules, a column with Mars, Venus and Hercules, a fragment of a column with Juno, a torso of, presumably, Mercury, and an altar dedicated to the three Parcae, the female personifications of destiny. Two sculptures depict animals: a bear and a lion with a prey.623 Very few bronze and terracotta statuettes have been found at Maastricht. They include a bronze statuette of Mercury, from the grounds of the Roman baths, a bronze head of a deity or male human figure from the precinct of the sanctuary, and a bronze statuette of Attis from an unknown find-spot at Maastricht.624 A terracotta fragment of a Venus statuette came to light in the late Roman fortification.625 Stray finds are a fragment of a terracotta horse and the head of a boy, discovered during excavations in 1981 at the Onze Lieve Vrouweplein southwest of the Roman period settlement.626 • Margraten-Backerbosch In 1879 and 1880, three monumental buildings were excavated by Joseph Habets at Backerbosch, near the village of Cadier en Keer. He interpreted the largest as the main building, which had a 82.7 metres long façade. It had baths and was heated by a hypocaust. Through geophysical survey and archaeological monitoring, a fourth structure and more rooms in the main building could be identified. Also another, possibly older building was discovered.627 The first building of the complex was constructed at the end of the 1st or the beginning of the 2nd century and probably burnt down in the 3rd century. 619

Panhuysen 1996, 46-49.

acorn and the fragment of a small bronze altar with an

620

Panhuysen 1996, 50, fig. 14 and note 161.

acorn. The oak tree was associated with Zeus and Jupiter.

621

Panhuysen 1996, 253-395.

In the sanctuary at Dodona, Greece, Zeus was venerated in

622

Panhuysen 1996, 119-120.

the oracular oak tree. The Jupiter column at Hausen-an-

623

Panhuysen 1996, catalogue nrs. 54, 55, 57-60, 62a-d, 64-66,

der Zaber is decorated with oak leaves and acorns: Green

75,79.

1992c, 164. Despite the Jupiter column and the objects

See for the Mercury statuette Zadoks et al. 1969, nr. 39.

with an acorn, Panhuysen’s interpretation of the head is

The Attis statuette is unpublished: Limburgs Museum

questionable, since a beardless young man would be a very

624

Venlo, inv. nr. L04380. The bronze head has a long neck. Inspection revealed that it is an attachment. Panhuysen

Van Boekel 1987, nr. 143.

suggested that the male figure represents Jupiter, possibly

626

Van Boekel 1987, nrs. 189 and 243.

because of the Jupiter column in the temple. Other objects

627

Groot 2006, 292 and fig. 6.

from the temple are a bronze weight in the shape of an

78

unusual iconography for the supreme god. 625

Besides building material and wall painting fragments, the finds included terra sigillata vessels, a silver plated bronze drinking cup with separate tripod, hair pins, several styluses (stili), as well as the lower part of a terracotta statuette of an enthroned mother goddess. The statuette was found in a pit filled with building debris in front of the villa. 628 • Voerendaal-Ten Hove The presence of a Roman period villa was already established during excavations at the end of the 19th century. A villa rustica was constructed around 50, a stone-built house. The enclosed villa settlement also comprised several post-built structures. At the end of the 1st century, a new villa complex appeared that remained into use until the end of the 3rd century. Besides the monumental main house the complex comprised several buildings, including a granary, a bathhouse and a forge.629 The excavations yielded a large fragment of a terracotta statuette, Venus standing in an aedicula, found near building H.630 Another fragment depicts a part of the upper body of a semi-nude figure, presumably male. The fragment came from a drain inside the villa compound that could be dated to the second half of the 2nd century.631 Limburg: find-spots of statuettes without contextual data Find-spot

Material

Subject

Reference

Baarlo

terracotta

Epona

Van Boekel 1987, nr. 25

Baarlo

terracotta

Epona

Van Boekel 1987, nr. 26632

Kessel-De Donk

bronze

Bacchus

Limburgs Museum633

Margraten

bronze

Attis

Limburgs Museum

Melick

terracotta

Fortuna

Van Boekel 1987, nr.31634

Montfort

terracotta

cockerel

Van Boekel 1987, nr. 213

Montfort

bronze

panther

Hupperetz 1991, 172635

Venlo

bronze

Hercules

Zadoks 1969, nr. 23

Limburg: uncertain find-spots of statuettes Find-spot

Material

Subject

Reference

Mook?

bronze

cockerel

Zadoks 1969, nr. 66

Voerendaal

bronze

Amor

Limburgs Museum636

628

Van Boekel 1987, nr. 68.

629

Willems/Kooistra 1988. See also Habermehl 2011, 214, site

with pottery and a burial were discovered in the 1970s. It is

146.

unknown whether the statuette was a grave gift belonging

630

Van Boekel 1987, nr. 115.

to the burial. Besides a Roman period cemetery, traces of a

631

Van Boekel 1987, nr. 169.

settlement have been found here as well. See for instance,

632

The two statuettes were found on the left bank of the Meuse, presumably in the vicinity of the church, where

633

634

The statuette was found at Zwarte Berg. Here, two pits

Verhoeven et al. 2010, 40 and fig. 13. 635

The panther was discovered at Montfort-Montforterbroek,

also other Roman period finds have been reported.

together with a Hallstatt spearhead and potsherds that

Inv. nr. L24291. Kessel-De Donk is situated on the left bank

have not been preserved. Not far from these objects, a coin

of the Meuse, some five kilometres south of Helden. A field

hoard was found in a pot. The panther was an applique

survey at Kessel has produced traces of the Roman period,

that once decorated a wooden object, like a chariot or

but the finds have not been published: De Winter 2010, 19 and note 7.

piece of furniture: Peters 1984, 17. See also 4.2.2 and 4.2.3. 636

Inv. nr. L29019.

79

2.4

The Dutch part of the Roman limes zone

2.4.1

The forts and associated settlements on the left bank of the Rhine

The Roman occupation in the study area started with a period of expansion during the reign of Augustus, which was characterised by the construction of large military forts on the left bank of the Rhine: Nijmegen, Xanten and Neuss. Soon, more forts appeared in the study area, at Velsen and on the left bank of the Nederrijn, Kromme Rijn and Oude Rijn. The Batavian Revolt was followed by a period of stabilisation, until raids by Germanic groups marked a new period of unrest in the third quarter of the 2nd century. New forts were built, in the study area between Katwijk and Aardenburg, to defend the North Sea coast against attacks from sea. Between the last quarter of the 3rd and the end of the 4th century, new strongholds were built along the Rhine, at Maastricht, Cuijk and Nijmegen. The distinctiveness of military culture and identity has been discussed in several publications.637 Apart from formal practices, festivals and obligatory rituals related to the imperial cult, evidence from the northern frontier zones in Lower Germany and Britain demonstrates that soldiers also worshipped local and foreign deities and performed rituals that were not related to formal religion, such as the deposition of weaponry and horse gear, for which abundant evidence has been found in Lower Germany as well as in northern Britain.638 The limes zone has yielded the highest number of statuettes and statuette fragments. The forts, canabae legionis and military vici are presented in alphabetical order, beginning with the northern limes on the left bank of the Rhine, followed by the forts on the North Sea coast. • Alphen aan den Rijn: castellum and vicus The discovery in Alphen’s city centre of a Roman period stone-built bath complex in 1959 remained practically unpublished and without any follow-up, until excavations in 1985 were carried out which produced the remains of a timber building with a gutter and drain. As a consequence of the high water table, the wood and other organic remains in the gutter and drain were well-preserved. The results of the excavations made clear that castellum Albaniana had been found, located just north of the Castellumstraat.639 Archaeological research in 2001 and 2002 yielded the following conclusions: The fort was built in 41, slightly earlier than presumed before. It went up in flames during the Batavian Revolt and was rebuilt in 70. The new fort was slightly larger and enclosed by three ditches. In 160 the fort was rebuilt again, probably after a flood. Like most castella along the limes, the fort was abandoned around 270-275.640 Digging operations for a sewer at Julianastraat in 1998 revealed a habitation layer that could be dated to the middle of the 1st century. The site has been interpreted as a vicus located near the castellum. It was destroyed by fire during the Batavian Revolt and rebuilt shortly after that. In the second half of the 2nd century, a large stone building appeared.641 Despite the enormous quantity of finds from the excavations between 1959 and 2002 only one bronze bust and one terracotta statuette came to light. A small bronze bust of Sol was discovered by a detectorist.642 A terracotta Venus statuette was found in 1981 outside the fort, in a layer dating between 75 and 100.643

637

638

80

Cancik /Rüpke.1997, Von Hesberg 1999, Haynes 1997, 1999,

Rhine-delta Nicolay 2007.

2013. See on Roman frontier communities and culture the

639

Haalebos et al. 2000, 22-28.

contributions in Breeze et al. 2015.

640

Polak et al. 2004, Zander 2013, 12.

See for Nijmegen, for instance, the ritual deposit of cavalry

641

Haalebos et al. 2000, 35-39.

helmets at Kops Plateau: Van Enckevort/Willems 1994,

642

Bakker/Bron 2013, 119, nr. 7.115.

125-137. See for deposits of weaponry and horse gear in the

643

Van Boekel 1987, nr. 102.

A human skull was recovered from one of the enclosure ditches belonging to phase 2 (c. 70-160) of the castellum and probably ended up there around 160.644 • Katwijk aan Zee-Brittenburg Between the 15th and 17th centuries, remains of a square structure became visible after heavy storms had blown away a part of the dunes. The structure is depicted in a 16th-century engraving by Abraham Ortelius. According to Holwerda, it was a Roman fort. The remains later disappeared permanently as a result of a shifting coastline. Whether the engraving really depicted a Roman fortification is uncertain. In 1981, during digging operations for the construction of sluices at Katwijk aan Zee, a geologist found Roman roof tiles and potsherds dating to the second half of the 2nd and 3rd centuries. The discovery prompted a small-scale excavation in 1982, during which a well, palisades and refuse pits were found. They probably belonged to the vicus of the Roman fort. The settlement was occupied between 160 and 240. Research in 2014 and dredging operations have not yielded any evidence of a fort. 645 Two bronze statuettes have been found at Katwijk. The first depicts Minerva, the other Apollo. The statuettes have disappeared and are only known through drawings made in the 16th century.646 • Leiden-Roomburg: castellum As early as the 16th century, finds like coins, bronze lions and a bronze statuette of Minerva at Roomburg suggested Roman period habitation in the polder. Although Holwerda suspected castellum Matilo was situated in Leiden’s city centre, new Roman finds at Roomburg prompted excavations in 1927, which brought to light defensive ditches of a fort. Excavations in 2009 aimed at finding the eastern and southern walls and ditches of the 3rd-century stone-built fort.647 Pre-Flavian pottery indicated the existence of an earlier, small fort. Around 50 it was replaced by a new fort. A stone-built fort appeared at the end of the 2nd century. A complete terracotta statuette of a water nymph was found in 1927, during Holwerda’s excavations. It comes from the filling of a ditch, probably a defensive ditch. 648 Holwerda dated the filling to the 2nd century. The production date of the statuette lies in the last quarter of the 2nd century, which provides a terminus post quem for the deposition of the statuette in the ditch.649 • Leiden-Roomburg: vicus In 1962, an excavation directly north of the castellum brought to light traces of a settlement. Excavations north and west of fort Matilo at Roomburg in the 1990’s focused on the development of the vicus and Corbulo’s canal. The canal and the vicus covered a surface of 20 hectares. Around 100, the settlement expanded westwards.650 In the northernmost ditch, which may have been part of the earliest boundary of the settlement, a complete amphora was buried, dating to the second half of the 1st century. A helmet mask from the canal possibly was a ritual deposit, too, an offering made by a soldier who had completed his 25 years of service.651 Also a silver finger ring and a thin bronze sheet came to light, both with a dedicatory inscription.652

644

645

Oral information by Rien Polak. See for the ditch (2a) Polak

649

Van Boekel 1987, nr. 124. See for the production date nr. 125.

et al. 2004, 44-46 and map B. See also figs. 4.1-2.

650

Hazenberg 2000, 41-48.

Bloemers/De Weerd 1982, 245-251. See also Vici.org/

651

Van Enckevort/Hazenberg 2000, 53.

vici/58?lang=en (1-2-2018).

652

Hazenberg 2000, 53 and figs. 33-34. The inscription on the

646

Zadoks et al. 1969, nrs. 4 and 58.

ring reads AVE AMMA. The inscription on the sheet reads

647

Brandenburgh/De Bruin 2016.

FESTI DEIS * ICON C[.

648

See for the interpretation 4.2.2.

81

The excavations in 1995-1997 produced the head of a mother goddess figurine.653 A terracotta wheel, 68 mm in diameter, with a hole, 12 mm in diameter, was found in a garden at the west side of the castellum. The interpretation of the object is uncertain.654 The wheel may have been part of a horse on wheels, or a chariot. It could have belonged to a statuette of a deity, possibly Jupiter-Taranis, with a wheel as attribute. Wheels could also be talismans. Spoked wheels could symbolize the radiate sun. At Rullstorf in Niedersachsen, Germany, a Bronze Age pit contained miniature vessels and small wheels. In the Iron Age, model wheels were buried with the dead. In Roman Britain and Gaul wheels were offered to sun or sky gods.655 • Nijmegen: the Augustan base at Hunerberg and the Flavian-Trajanic canabae legionis The earliest archaeological evidence of a Roman military camp in the Netherlands is the large, 42 hectare military base at Hunerberg. The identification of buildings from the earliest phase of the fortification is difficult, because they were overbuilt in the Flavian period by the canabae at the west side and the Flavian-Trajanic legionary fortress of the Tenth Legion in the east. In 1918, Holwerda had discovered the western gate of the Augustan base which was completely uncovered by Haalebos between 1992 and 1994. The eastern gate was discovered during Brunsting’s excavation in 1960.656 South of the via principalis between these two gates, traces of a large building came to light, measuring c. 75 by 75 metres. The building has been interpreted as the military headquarters. South of this building traces of another structure were uncovered. With a length of almost 60 metres, it was the second largest building within the fortress. It comprised a large square structure with atrium and peristylium, and presumably was the praetorium, the quarters of the army commander. The majority of the other features have been interpreted as soldiers’ barracks and officers’ quarters. 657 The composition of the garrison is unknown, due to the absence of epigraphic evidence. Recently, the conclusions by Bogaers and Haalebos from the 1980s and 1990s excavations on the site known as ‘the former grounds of the Canisius College’ have been refined.658 The construction date of the Augustan fortification, between 16 and 12 BCE, must be advanced by a few years to 19-16 BCE.659 The fortification could house up to 15,000 men, and if the adjusted construction date is correct, the military base at Hunerberg was built slightly earlier than those of Neuss, Dangstetten and Oberaden, which supposedly were built between 16 and 11 BCE. The Canisius College excavations have produced hundreds of thousands of finds.660 In 2014, the first volume was published, providing an analysis of the pottery, stamps, building material and animal bones.661 It has been suggested that the fortress was not completely abandoned after 16/12 BCE, and that a small part still was in use in the late Augustan period.662 Due to the brief occupation period of the Augustan base, the majority of finds are attributed to the Flavian-Trajanic habitation phase of the canabae legionis, although a few statuettes may have an earlier deposition date. In most cases, the production date of terracotta statuettes makes a pre-Flavian date improbable.663 Just outside the eastern corner of the Augustan fortification, an assemblage was found which Brunsting and Steures interpreted as funeral gifts belonging to a soldiers’ grave.664 According to Van Enckevort and Wil653

Find nr. 3-2-96/145. A drawing of the statuette head fea-

659

Kemmers 2006, 44-54.

tures in the report from 2000: Hazenberg 2000, 72, fig. 12.

660

See for the excavations Haalebos et al. 1995.

Van Boekel 1987, nrs. 124, 279.

661

Polak et al. 2014.

655

Green 1992c, 225-226.

662

Polak 2014, 19 and note 13, Niemeijer 2013, 71-72, 2014, 18-19.

656

Haalebos/Willems 2005a, 56-58.

663

Most terracotta statuettes in the study area were made in

657

Haalebos/Willems 2005a, 59-61.

the Rhine-Moselle area, where production centres became

658

See for earlier publications on the Hunerberg fortress, for

active from the end of the 1st century onwards: Van Boekel

654

instance: Bogaers/Haalebos 1988, Haalebos et al. 1995, 1998, Haalebos/Willems 2005b, 73-86.

82

1987, 204-210. 664

Brunsting/Steures 1992, 110.

lems the assemblage was a ritual deposit.665 The fortification has yielded no statuettes that can be attributed to the earliest phase. Bronze statuettes found in the area of the western canabae include statuettes of Fortuna, Hercules and Mercury.666 A fragment of a caduceus was found in a feature that was thought to be a sewer, but recently has been re-interpreted as a water channel.667 The fragment, 2.5 cm in length, must have belonged to a large Mercury statuette. Another bronze attribute, a 2.3 cm long torch or thunderbolt, was part of an Amor or Jupiter statuette and comes from the same feature. The find-context could be dated to the late Flavian period or slightly later.668 A bronze figurine of a lion is a stray find.669 A semi-finished bronze statuette with casting rods depicts Luna (fig. 2.123). The statuette was recovered from the same pit as a bronze figurine of a billy goat (fig. 2.125), which comes from a deeper level. They were found together with large potsherds. The terracotta statuettes include a pedestal belonging to a Venus statuette or a bust, a fragment of a Minerva statuette, two fragments of a female torso, a single pedestal, a torso of a Fortuna statuette, the head of a young woman, a rectangular plinth with two feet, a fragmentary statuette of a dwarf or caricature, the head of a cow, the head of a dog, which, based on complete parallels is ithyphallic, and several unidentifiable fragments (figs. 2.126-133). Terracotta finds from earlier excavations include statuettes of a dove and two hens.670 • Nijmegen: the fort on Kops Plateau Around 10 BCE, east of the Augustan military base, a new, much smaller military fortification was built on Kops Plateau. The excavators distinguish three to four building phases, during which the camp covered 3.5 - 4.5 hectares.671 Two roads ran across the camp, one from southwest to northeast and the other from southeast to northwest. At the crossing of these two roads, a large complex was located of which only a small part has been excavated. The complex has been interpreted as the principia. A second large complex, located at the north side of the camp, probably was the praetorium.672 The large number of officers’ dwellings and the limited space for barracks or tents is striking. The fortification has been interpreted as the quarters for the supreme commander of the army and his staff, at least during the two earliest phases. In the latest phase, a large part of the praetorium fell out of use and buildings were modified. Numerous pieces of horse gear and some iron cavalry helmets, as well as a graffito with the name of a soldier of a turma, led the excavators to think that the camp became the base for a cavalry unit.673 According to Tacitus, the Batavi had their own cavalry regiment, which was stationed ‘at home’, thus in Batavian territory.674 Although direct archaeological evidence is lacking, the excavators believe that the camp on Kops Plateau has housed the ala Batavorum. First, because it is the only known cavalry fort in this region. Second, because the camp hasn’t been burnt down during the Batavian revolt, when the ala Batavorum chose the side of the insurgents.675

665 666

Van Enckevort/Willems 1994, 131.

670

Van Boekel 1987, nrs. 225 and 235-236.

See for the Mercury and Hercules statuettes figs. 2.124 and

671

Willems/Van Enckevort 2005, 64.

2.162. See also Zadoks et al. 1969, nrs.18, 22, Haarhuis 1991,

672

See for a reconstruction of the complex Peterse 2005, 163198.

22-29. 667

668

Fig. 2.134. See for the re-interpretation Kessener/Janssen

673

Willems/Van Enckevort 2005, 64-66.

2017, 46-62.

674

Tac. Hist. 4.12: Erat et domi delectus eques, praecipuo nandi

Haarhuis 1991, 82-83. See for the sewer: Polak 2014, 26 and

studio, arma equosque retinens integris turmis Rhenum per-

fig. 1.14. 669

Haarhuis 1991, 84-85.

rumpere. 675

Willems/Van Enckevort 2005, 68.

83

Traces of annexes have been found as well. South of the camp, a building has been interpreted as stables. Numerous fragments of horse gear and graves with horses in this area suggest that, here too, cavalry units were stationed.676 Just before or during the Batavian Revolt in 69/70, the plateau was abandoned.677 Several assemblages on Kops Plateau could be ritual deposits. The earliest came from a pit south of the fort, east of the annex with stables. The pit contained a modified helmet from the Tiberian period, broken but more or less complete Gallo-Belgic butt beakers, a cork-urn, a thin-walled cup and a crucible. Four cavalry helmets came to light in the central part of the fort. The helmets were found together with pottery dating to the Claudian period. One helmet was buried with two terra nigra vessels and two helmets with a jar each. Three cavalry helmets with face masks date to the Claudian or Neronian period. Two were found together in a pit, the third was buried together with pottery and a chicken. The vessel had been deliberately destroyed before it was placed in the pit and may have contained food or drink.678 Other finds that until now have not been analysed could be ritual deposits, too, but need further investigation.679 The metal statuettes from Kops Plateau are mainly surface finds and comprise one bronze fragment of a caduceus, 40 mm in length, a lead Venus statuette, a lead figurine of a duck, a bronze statuette of a young Hercules strangling a (missing) snake, a small bronze wheel, 32 mm in diameter, two bronze pedestals, the rear part of a bronze swine, a bronze ram or billy goat, and a bronze cockerel (figs. 2.134-143, 2.148-150).680 Originally, the cockerel may have been attached to a pedestal of a Mercury statuette. A bronze lion was found near the military barracks.681 It resembles the lion found in the western canabae.682 The terracotta statuettes include three fragments of a crouching Venus, a rectangular plinth with a standing figure and several other unidentifiable fragments (figs. 2.144-147). According to the find label, the fragments of the Venus statuette come from a layer dating between 10 BCE and 10. If this date is correct, the crouching Venus is the earliest known terracotta statuette from the Roman period in the study area.683 • Nijmegen: the Flavian-Trajanic fortress on Hunerberg After the suppression of the Batavian Revolt, a standard legionary fortress was built on Hunerberg, in the northeast corner of the Augustan fortification. The excavators distinguish four to five phases during which the dimensions varied between fifteen and thirty hectares. The first three phases date to the pre-Flavian period, the other(s) to the Flavian-Trajanic period. Of the earliest phases only ditches have been found. Best known is the last stone-built phase. This fortress covered a surface of 16 hectares. The buildings found inside

676

Willems/Van Enckevort 2005, 72.

Iphicles.JPG (29-3-2017). Examples of bronze statuettes of

677

Willems/Van Enckevort 2005, 70.

the infant Hercules killing the snakes are in possession of

678

Van Enckevort/Willems 1994, 126-134.

the Cabinet des médailles et antiques de la Bibiothèque

679

Van Enckevort/Willems 1994, 131.

nationale de Paris, inv. nrs. 589-590. The museum has

680

See for the interpretation of wheels this chapter. Two other

an online database: http://medaillesetantiques.bnf.fr/ws/

metal wheels with an axle, find nr. 396/199, are modern

catalogue/app/report/index.html.

and do not belong to a Roman period cult wagon as has

681

Zadoks et al. 1969, nr. 77.

been advanced by Van Zalinge 1997, 21. In my opinion, the

682

Haarhuis 1991, 84. Although there are no visible traces of tin or lead on the other specimen, both lions could be

bronze statuette represents the infant Hercules. Consider-

attachments.

ing the popularity of Hercules and the parallels in sculpture and wall paintings of the young Hercules strangling

The fragments were recovered from a pit with finds dating

the snake, it is not very plausible that Iphicles, Hercules’

to the earliest phase of the camp: Arretine ware made

younger brother is depicted. Representations of Hercules

between 30 BCE and 25, oil lamps, pottery and a coin dat-

with his twin brother are known from Greek vases, but

ing to the Augustan period. The statuette has been errone-

usually depict a terrified Iphicles, while Hercules kills

ously identified as a thorn-puller: Van Zalinge- 1997, 23.

the snake: http://mythagora.com/bios/images/snakes.jpg https://commons.wikimedia.org/wiki/File:Heracles_and_

84

683

the fortress are the principia, measuring 93 by 65 metres, military barracks that could house six cohorts, horrea, workshops with evidence of metal working, sewers, latrines, water pipes and a well.684 Based on coins struck during the reigns of Hadrian, Antoninus Pius and Commodus, it is assumed that military units stayed at Hunerberg until the middle of the 2nd century or even longer.685 The description ‘find-spot Hunerberg’ applies to several bronze statuettes, which means that the deposition date lies between the arrival of the first troops at the end of the 1st century BCE and the second half of the 2nd century.686 This applies to a statuette of Bacchus, a small Amor figurine and four specimens depicting Hercules, Mars, Minerva and an enthroned mother goddess.687 A statuette of a horse was found in the southeast corner of the fortress, near the soldiers’ barracks.688 A caduceus comes from a pit with charcoal and terra sigillata (rouletted Argonne ware) dating to the 4th century.689 Although a late Roman date for the pit seems improbable, the available information is uncertain and does not allow for any conclusions on the deposition date or the composition of the assemblage.690 A terracotta statuette of Minerva, presumably from the site Hunerberg, is a stray find (fig. 2.158). • Nijmegen: the eastern and southern canabae legionis North of the road near the eastern gate of the Flavian-Trajanic fortress, traces of two post-built houses and six outbuildings have been uncovered. 2,500 glass fragments testified to the production of glass. Large quantities of bovine skulls were indicative of meat processing. During excavations in 1979 between the Berg en Dalseweg and the Ubbergseveldweg, traces were found of a strikingly large structure, measuring 166 by 137 metres. In previous studies, the structure has been interpreted as a forum where cattle and other livestock were traded.691 Recently, Christian Gugl and Jürgen Trumm have proposed an alternative interpretation for these large structures, based on excavation data from Windisch/Vindonissa, Switzerland, and Carnuntum, Austria. The structures at these sites share similarities such as the absence of traces of buildings and architectural ornamentation on the inside, the lay-out, dimensions and location in the vicinity of legionary fortresses. Gugl and Trumm suggest that the structures may have been open training grounds, possibly with covered porticoes, for soldiers from the fortresses.692 In the eastern canabae, northeast of the porta principalis dextra of the fortress, in a cistern or well that could be dated between 75 and 125, the fragmented head of a terracotta bust of a woman was found, together with a terracotta statuette of Fortuna.693 Other finds from the eastern canabae include a bronze pedestal (fig.2.149) and a bronze fragment of a hand holding a dolphin, which probably belonged to a statuette of Neptune, and a terracotta statuette of Jupiter.694 • Nijmegen: the industrial site Holdeurn During excavations between 1938 and 1942 a large complex for the production of pottery, bricks and tiles was unearthed on the site Holdeurn, four kilometres southeast of modern Nijmegen. Here, pottery, bricks and 684

Haalebos/Willems 2005b, 73-81.

689

Zadoks et al. 1969, nr. 80.

685

Haalebos/Willems 2005b, 85-86.

690

I thank Rien Polak for the information and documentation

686

Until a century ago, the indication ‘find-spot Hunerberg’ applied to a larger area. This implies that earlier finds may

691

Van Enckevort 2005b, 93-97, Driessen 2007, 128-135.

have come from the canabae legionis outside the military

692

Gugl/Trumm 2015. I thank Rien Polak for bringing this

on the 1960’s excavations by Brunsting.

publication to my attention.

camp and even beyond. Objects from the former Gemeentemuseum with the indication ‘find-spot Hunerberg’ come

693

See also 4.2.2; Van Boekel 1987, nrs. 37 and 200.

from the early Roman cemetery. Information by Louis

694

The pedestal was found east of the eastern gate of the

Swinkels. See for the early Roman cemetery (Hugo de

legionary fortress. See for the Jupiter statuette: Van Boekel

Grootstraat-Museum Kamstraat) this chapter.

1996, nr. 40. The bronze hand holding a dolphin was

687

Zadoks et al. 1969, nrs. 1, 7, 24, 36, 38, 57.

found at Lent, in soil that came from the eastern canabae,

688

Zadoks et al. 1969, nr. 74.

Ubbergseveldweg-Klokkenberg: Bogaers 1970a, 97-101.

85

tiles were produced, mainly for the military market. The complex was constructed and initially manned by the Legio X Gemina, which was based at the legionary fortress on Hunerberg between 70 and 104. The presence of other military units that have been at work here has been attested as well. The industrial complex must have been active for at least one and a half century. In total, six brick kilns and at least four pottery kilns have been found.695 Geophysical research with ground penetrating radar has revealed that the complex was probably much larger than assumed so far.696 Although no sanctuary has been found on the Holdeurn site, several limestone altars and one sandstone specimen came to light. One altar dedicated to Vesta by the head of the potters was placed between the kilns.697 The majority of the altars are dedicated to Jupiter Optimus Maximus and the genius loci.698 A statuette of Venus came to light during the excavations, presumably near a large building.699 A mould depicting the head of Jupiter was found near the kilns.700 • Nijmegen: cemeteries associated with military settlements Cemeteries were located to the east and south of the pre-Flavian fort on Kops Plateau and the Flavian-Trajanic legionary fortress on Hunerberg. The cemetery of the latter is likely to have included graves of the inhabitants of the canabae legionis. • Canisius College This part of the 1st-century cemetery is situated outside the southwest corner of the Augustan fortress, bordered to the north by Pater Rubenspad, to the west by Museum Kamstraat and to the south by Pater Leydekkersstraat. This area was a court of St. Canisius College. Father Ferdinand Leydekkers, a priest and geography teacher at Canisius College, uncovered several graves and assembled no less than 550 finds in 1906 and 1907.701 The excavations were published in 1932 by Willem Vermeulen, who concluded that the burial site dated to the Tiberio-Neronian period.702 During Leydekker’s excavations, a rattle in the shape of a terracotta horned bird was found in a cremation grave of a child (figs. 2.156-57).703 • Kops Plateau In 1993, a slightly damaged but complete terracotta bust of a smiling boy of the Risus type came to light east of the fort (fig. 2.151). Presumably, the statuette came from a grave.704 • Valkenburg: castellum The first large-scale excavations at Valkenburg were carried out between 1941 and 1953. Data were collected during eleven campaigns, under the supervision of Van Giffen.705 Although excavations in the 1960s and 1980s resulted in seven fairly complete plans, the interpretation of several buildings remains uncertain.706 The construction of the first fort in 40/42 has been linked to Claudius’ conquest of Britannia in 43. Castellum 1a was burnt down in 47 by Chaucian pirates and replaced in the same year by a new fort, which was destroyed during the Batavian Revolt in 69/70. Castellum 4 was constructed shortly after that. Around 100, Valkenburg 5 was built, this time with stone-built principia. Around 180 a complete stone-built fort appeared 695

Polak 2005, 243-245.

702

Vermeulen 1932.

696

Van Diepen/Polak 2009, 29-32.

703

Vermeulen 1932, 183-184, grave 66 and plate XV, Van Boekel

697

See for the purpose of the altar and dedications by craftsmen 4.2.4.

1987, nr. 241. See for the figurine also 3.8.20. 704

The pit in which it was found contained cremation remains:

698

Daniels/Brunsting 1955, nrs. 40-41, 47-48, 48a, 49.

699

Van Boekel 1987, supplement nr. 7.

705

Hingh/Vos 2005, 29-61.

700

Van Boekel 1987, nr. 194.

706

See for the plans of the phases of the fort Glasbergen 1972,

701

Vermeulen 1932, 1-3 and plates IIIA and B.

86

oral information by Rien Polak.

133-144, Hingh/Vos 2005, 107.

which remained occupied until 240.707 During excavations in 1980 three horrea were uncovered, leading the excavators to think that the fort had known a seventh phase, between 250 and 400, based on dendrochronological analysis of timbers of one of the horrea. 708 Recent re-analysis of the dendrochronological data has demonstrated that a late Roman date for the timbers is incorrect.709 Three bronze umbones were discovered in a small pit in a building in castellum 4. The umbones were placed next to each other, directly under a burnt layer. Based on the nature of the objects, their location and arrangement, the excavator has suggested that the assemblage has been intentionally buried inside the building.710 An extraordinary find came to light in 1964. It was the torso of a statue, almost one metre in length.711 Although the head, lower legs and right arm are missing, it is clear that the muscled male represents Hercules, leaning with his left arm on his club.712 A bronze statuette of Isis-Fortuna was found on top of a demolition layer belonging to phase 6 of the fort. Contrary to the find circumstances of the three umbones, Glasbergen did not describe the find circumstances of the statuette.713 Two terracotta statuettes possibly come from the castellum as well. One statuette represents Amor or the male personification of autumn.714 The other depicts a sheep or goat.715 • Valkenburg-Marktveld Excavations between 1985 and 1988 at the sites Marktveld and De Woerd have yielded evidence of buildings, enclosures, ditches, roads, gullies of the Rhine, harbour structures and a cemetery. Excavations in 1994 and 1995-1996 uncovered a part of the Roman road and six buildings belonging to the vicus. The site Marktveld is located south of the fort. The earliest finds from the site are a few bronze asses and dupondii in a purse. The earliest coin was struck round 10 BCE, the latest just before 23.716. In the 1st century, a small fort, a watch tower and horrea were built on three parcels west of the road. Soon, the easternmost parcel was used to bury the dead. In 99/100 a Roman road was constructed on the left bank, leading to the fort. In the 2nd century, farmsteads were built on the western parcels. A building measuring 7.2 by 4.9 metres was located west of the cemetery and may have been a sanctuary. It was enclosed by a horse shoe-shaped ditch with the opening to the east, where the cemetery was located. The finds in the ditch could be dated to the 2nd century. The Marktveld site fell out of use around the middle of the 3rd century.717 Besides two unidentifiable fragments, two fragments of terracotta masks and a Venus statuette of uncertain provenance, the Marktveld site has yielded ten terracotta statuettes. Six identifiable specimens and two fragments of unidentifiable goddesses have been recovered from the gully. The identifiable statuettes represent Epona (one), Fortuna (two), a mother goddess (two) and a cockerel. The settlement has yielded one fragment of an Epona statuette. A stray find from the cemetery is a mother goddess figurine, found in 1941 or before.718 • Vechten The area southwest of the village of Vechten is one of the archaeological hotspots in the study area, where excavations were carried out from 1829 onward. Here, one of the most important military camps, a large vicus and cemeteries were located. Only five percent of the fort and vicus have been excavated. The widening of

707

Glasbergen/Groenman-van Waateringe 1974, 6.

714

Van Boekel 1987, nr. 146, Van Boekel 1993, nr. 1.

708

Hingh/Vos 2005, 99-112.

715

Hingh/Vos 2005, 150. The find circumstances of the figu-

709

Oral information by Rien Polak.

710 711

rine are unknown.

Glasbergen 1972, 114-116.

716

Hingh/Vos 2005, 83, 148-149. See also Bult/Vons 1990, 70-71.

Hingh/Vos 2005, 58-59, 151-152.

717

Hallewas/Van Dierendonck 1993, 14-15, 21-27, 32-43.

712

See for the iconography of Hercules 3.8.3.

718

Van Boekel 1993, nrs. 1-13. See also Van Boekel 1987, nrs. 76

713

See also 4.2.2.

and supplement nrs. 12 and 14.

87

the A12 motorway constituted an important threat for the archaeological record and prompted a series of rescue excavations between 1970 and 1995.719 Two centuries of excavations have produced a large quantity of data, mostly still unpublished. In 1991, Polak and Wynia published a map which provided an overview of the excavations so far.720 A publication on the excavations in the 1940s saw the light in 2012.721 The Roman forts at Vechten were constructed on the bank of an old Rhine tributary within the Kromme Rijn basin. The exact construction date of the earliest fort is disputed and the proposed dates vary from 12-9 BCE to 4-5 and 14-16. The earliest phase was succeeded by three others. The only fort of which the dimensions are known is the stone-built version belonging to the third phase. With a surface of 2.6 hectares, Vechten is the largest known castellum in the study area.722 The name of the settlement, Fectio, refers to the river Vecht, which branched off from the Rhine downstream from the fort. Fectio is mentioned in an inscription on an altar dedicated to the native goddess Viradecdis.723 The altar came to light in 1869, not far from the fort.724 East of the fort, a large vicus was located, covering a surface of ten hectares. The largest part was situated south of the A12 motorway, but the settlement may even have extended further northwards. The first habitation phase began before the middle of the 1st century. Shortly after that, the Rhine flooded the area. Only after raising the ground level, the settlement was habitable again at the end of the 1st century. So far, only small parts of the vicus have been excavated. Moreover, the large quantity and complexity of features and habitation traces that came to light during the excavations in 1994-1995 are difficult to interpret and do not allow for conclusions on the layout and phasing of the settlement.725 A remarkable find from Vechten is a bronze saucepan with a graffito on the handle. The saucepan came to light when the excavators wanted to lift the last staves of the wine barrels that were used for the lining of a well.726 Like the objects recovered from wells at Voorburg-Arentsburg and on other sites, the bronze saucepan may have been a ritual deposit. Vechten has yielded ten bronze statuettes and 76 terracotta specimens, of which some 25 were found in 1977-1978 by collectors in a secondary context, the dug-out soil from a former bed of the Rhine. The soil was used for the construction works at the A12 motorway in 1979-1980. The layer that contained the Roman artefacts could be dated to the 2nd century.727 The bronze specimens represent Hercules (one), Harpocrates (one), Jupiter (one) Mars (one), Mercury (two) and Minerva (one).728 The only bronze statuette of a human figure depicts an ithyphallic, African man (figs. 2.171-72).729 Bronze statuettes of animals comprise a bull and a horse.730 Among the terracotta figurines Venus is predominant with twelve specimens, followed by Fortuna and Minerva.731 Terracotta statuettes that depict a human figure comprise busts of boys, a woman and a girl, three mother goddess figurines, an unidentified female figure and two embracing couples.732 The terracotta animals include a bull, a cockerel, a dog, a goat, a lion and a wild boar.733

719

Hessing et al. 1999, 9-11.

728

Zadoks et al. 1969, nrs. 25, 32, 35, 40, 46, 55.

720

Polak/Wynia 1991.

729

The penis was cast separately and is now missing. See for

721

Zandstra/Polak 2012.

722 723

Zandstra/Polak et al. 2012, 17-19.

Zadoks et al. 1969, nrs. 65, 75.

The inscription reads: Deae/[Vir]adecd(is) /[civ]es Tungri/

731

See for all statuettes of deities from Vechten Van Boekel

[et] nautae/[qu]i Fectione/[c]onsistunt/v(otum) s(olver-

1987, nrs. 18-19, 29-30, 32, 34-35, 38, 42-43, 52-55, 57-58, 63,

unt) l(ibentes) m(erito): CIL 13, 8815.

66, 67, 94-95, 97, 99-100, 105, 109, 111-112, 119, 122, 126 and

724

Zandstra/Polak et al. 2012, 17.

725

Vos 1999, 35-47.

726

Vos 1999, 52-59.

727

Van Boekel 1987, 245, 357, 567 and nrs. 6, 8, 18, 30, 32, 34, 43, 91, 97, 105, 111-112, 126-127, 132, 140-142, 151, 172-173, 245, 259, supplement nrs. 6 and 10.

88

the interpretation 3.8.15. 730

supplement nrs. 3, 8. 732

See for the statuette of the unidentified female figure fig 2.173 and 3.20.

733

Van Boekel 1987, nrs. 206, 212, 220, 232, 245, 250, 259, 260261.

• Vleuten-De Meern/De Hoge Woerd The first excavations were conducted in the 1830s. In 1980, finds of an excavation in 1973 were published in a short, preliminary rapport.734 As a result of a coring survey and geophysical research between 1991 and 1993, the contours of a Roman period castellum could be identified. The castellum was built in the Claudian period and remained occupied into the 3rd century. North-east of the fort, a cemetery is suspected. South-east of the castellum traces of a vicus have been uncovered.735 A bronze Mercury statuette came to light around 1944, near the Castellumlaan.736 A second stray find from the site is a bronze billy goat carrying a saddle bag.737 • Woerden Already in the 16th century, Roman artefacts were found at Woerden. In 1974, amateur archaeologists carried out a small scale archaeological excavation in Woerden’s city centre. Traces of a Roman military settlement prompted a series of excavations from 1975 onwards. In 1999, evidence of a defensive tower was discovered and by 2006, the contours of the fort, Lauri on the Peutinger map, were known. In 2007 it could be established that Laurium was 120 metres in length.738 Although the construction date of the earliest phase is uncertain, the first fort was probably built between 41 and 50 and could be related to the plans of the emperors Caligula or Claudius to invade Britannia. A second fort was situated more to the north, possibly constructed around 43 or 47. It was rebuilt after the Batavian Revolt. The fourth phase of the fort dates to the late 2nd century. Outside the fort, parts of a vicus were discovered, which came into being after 70.739 Excavations in the 1970s not only produced wooden quay wall structures, but also a ship that carried grain. It had sunk with its cargo still on board, presumably at the beginning of the 3rd century.740 New excavations yielded more ships and a dugout canoe.741 The site has yielded a 75 cm high limestone altar, dedicated to Sol Helagabalus and Minerva.742 The dedicator was a centurio, Lucius Terentius Bassus of the cohors III Breucorum. Besides a patrix with the head of Silenus, a second patrix was found, depicting Jupiter-Ammon.743 Jupiter-Ammon was particularly popular among Roman soldiers because of his apotropaic powers.744 Patrices were used to make attachments, either for ceramics or bronze vessels and other objects. The length of the Jupiter-Ammon patrix is almost nine centimetres, indicating that the head probably was a decorative element at the end of a wooden beam. Bogaers and Haalebos also mention a bronze jar, dating to the 2nd or 3rd century. It was found between the two latest quays. Unfortunately, the excavators do not provide any details on the find circumstances.745 Although the patrices suggest that bronze casters were active in Woerden, no bronze statuettes have been found. Excavations north and south of the castellum yielded two identifiable terracotta statuettes, one depicting Fortuna, the other Mercury.746 Four terracotta specimens came to light in the vicus: a head of a woman, possibly Cybele, and three unidentifiable fragments.747 • Zwammerdam Excavations between 1968 and 1977 produced an almost complete plan of a castellum, small parts of a vicus and traces of quays along an old river bed of the Rhine, in which six boats were found. The fort knew three occupation phases. The earliest military post was occupied between 47 and 69/70, when the fort went up in 734

Kalee/Isings 1980.

741

Vos et al. 2010, 150-160.

735

Kalee/Isings 1980, 5-25, Blom/Graafstal 2001, 5, 24.

742

Vos et al. 2010, 129. See for the inscription AE 1994, 1285.

736

Zadoks et al. 1969, nr. 50.

743

Bogaers/Haalebos 1979b, 99-100, Vos et al. 2010, 98-99.

737

Isings 1984, 99-102. See also 3.8.17; for a ram with a saddle

744

See for Jupiter Ammon also 3.19.

bag fig. 3.12.

745

Bogaers/Haalebos 1979a, 83.

Vos et al. 2010, 15-21.

746

Van Boekel 1987, nrs. 39, 49.

739

Blom/Vos 2008, 409-420.

747

Blom/Vos 2008, 171.

740

Bogaers/Haalebos 1979, 76-81, Vos et al. 2010, 151-152.

738

89

flames during the Batavian Revolt. Traces from this phase mainly consisted of foundation ditches of timber buildings. A second timber fort appeared, measuring 134.4 by 76.4 metres. Around 180, this fort was replaced by a slightly larger stone-built structure which remained in use until 270 or 275. Inside the fort, traces of large principia have been found, East of the fort foundations of a large building have come to light, possibly a bath complex.748 Terracotta statuettes from Zwammerdam include a complete specimen of Isis, which was found outside the fort, and a torso fragment of a Venus figurine, found along the quay at barge 2.749 A near- complete figurine of a boy was found near barge 2 along the same quay. The figurine is a rattle, because it holds two pebbles inside.750 It has been argued that the figurine wears a Gallic cloak, but in my opinion it depicts a swaddled infant (fig. 2.178). This assumption is based on the resemblance to so-called stone and terracotta infant figurines.751 The head of a mother goddess figurine was found along the quay beneath barge 6.752 Find-spots of statuettes without contextual data Find-spot

Material

Subject

Reference

Maastricht-Houtmaas

terracotta

Venus

Van Boekel 1987, nr. 143753

Nijmegen-Hunerberg

terracotta

embracing couple

Van Boekel 1987, nr. 11

Vechten

terracotta

embracing couple

Van Boekel 1987, nr. 40

Vechten

terracotta

embracing couple

Van Boekel 1987, nr.121

Subject

Reference

Uncertain find-spots of statuettes Find-spot

Material

Vechten or Wijk bij Duurstede

bronze

triple-horned bull

Zadoks 1969, nr. 65

Vleuten-De Meern

bronze

Jupiter

Zadoks 1969, nr. 30

2.4.2

The forts and associated settlements on the North Sea c o a s t 754

• North Holland: Velsen I and II, the northernmost forts Terra sigillata sherds from a World War II tank ditch in 1945, as well as numerous finds that came to light between 1951 and 1957 prompted excavations from 1964 onwards.755 The site could be dated to the reigns of Caligula (37-41) and Claudius (41-54) and was interpreted as a military base with a harbour.756 Until 1972, the 748

De Weerd/Haalebos 1973, 386-397, Haalebos 1977, 284-292,

Meuse. It covered a surface of 1.53 hectares and had ten

Haalebos et al. 2000, 15-19.

towers, of which four have been excavated. The fortifica-

749

Van Boekel 1987, nrs. 40, 121.

tion was enclosed by a ditch, which was restored at the end

750

Van Boekel 1987, nr. 188.

of the 4th or the beginning of the 5th century. The fort was

751

See for examples from France and Italy Derks 2014a, 56, fig

in use until the end of the Carolingian period. Excavations

4A, Cazanove 2013, for a swaddled infant on a terracotta

in 1982 have yielded the terracotta fragment which depicts

sarcophagus from Bavay, France: http://www.livius.org/

the right breast of a nude or semi-nude woman. See for the

pictures/france/bavay-bagacum/bavay-tomb-of-a-baby/ (1-5-2017). 752

753

Van Boekel 1987, nr. 77. See 4.2.2 for the interpretation of

The forts on the North Sea coast are not described in alphabetical order but from north to south.

nrs 77 and 188.

755

Bosman 1997, 276-284.

In the first half of the 4th century, a fortification was built

756

Morel/De Weerd 1980, 475; Bosman 1997, 8-9.

at the entrance of the Roman bridge on the left bank of the

90

excavation Panhuysen 1984, 52-64, 1996, 51-58. 754

site was known as Roman Velsen, but this changed when a couple of years later another concentration of finds indicated that an earlier fort was located here as well. The find-spot was located around 600 metres to the southeast and its finds could be dated to the Augusto-Tiberian period. Then it was decided to name the later fort Velsen II and the earlier, Augusto-Tiberian fort, Velsen I.757 Both forts were built on the left bank of the Oer-IJ. Besides the strategic location along the river, the presence of Frisian tribes in the area may have been a second reason why the Romans had chosen this location for the forts. The two forts have never been in use simultaneously.758 A large part of Velsen I has been excavated, allowing to distinguish three major building phases. During its first phase, a triangular shaped fort with a rampart and ditch was built, as well as a harbour. In the following phases an annex and more defensive ditches were added and its surface was doubled to 2 hectares. The fort was occupied between 15, the year of Germanicus’ first campaign in Germania, and 28, the Frisian Revolt. The restoration of the earliest well with wood dating to 37 indicates that the Romans returned to the area. They did not restore Velsen I, but built a new fort, approximately 600 metres to the west. The construction of Velsen II, as well as the appearance of forts on the left bank of the Rhine, was probably related to the invasion of Britannia in 43. When the emperor Claudius ordered general Corbulo to withdraw his troops behind the Rhine, the fort was deliberately made uninhabitable by burning down vital parts of the complex and by poisoning wells with cadavers.759 No sanctuaries have been found at the forts in Velsen, but several finds in wells may have had a special meaning. Next to the skeleton of a man, a complete pottery vessel was found in one of the wells. An umbo was found in another. Other wells contained bones of human hands and human skeletons.760 2.7 kilometres south of Velsen I, at Velserbroek, a cult place has been identified. On a sand ridge a wide variety of weapons, jewellery, coins, iron, silver objects and pottery were deposited from the middle Iron Age onwards, which continued after the Romans had left Velsen. Although a part of the offerings consisted of Roman import goods, no statuettes have come to light at the cult place.761 A bronze statuette of Hercules, type Hercules bibax, was a stray find from Velsen I (fig. 2.181).762 The occupation period of the fort suggests that the statuette came to Velsen between 15 and 28. • South Holland: Naaldwijk-Hoogwerf The location Hoogwerf is situated on the north bank of the former Meuse estuary, known in antiquity as the Helinium.763 Here, a more than life-size bronze hand came to light in 1933.764 The find prompted the first excavations in the 1930s, conducted by Jan Hendrik Holwerda. He unearthed a settlement measuring 100 by 130 metres and interpreted it as a native settlement, because of the absence of imported goods. The settlement could be dated to the late 1st and 2nd centuries. New excavations west and east of Holwerda’s excavations were carried out in 1977, 2003-2004 and 2006. Two to four Roman period farmsteads and other structures were uncovered, dating from the early Roman period into the 4th century. Between 175 and 190, radical changes were visible in the layout and organisation of the settlement, which was characterised by a complex system of ditches. A strong increase in imported pottery, metal objects, Roman coins and inscription fragments pointed at intensified contacts with the Romans. This was substantiated by the appearance of stone and ceramic building material and different building techniques. No less than 28 stamps of the Classis Germanica have come to light, most likely the Classis Augusta Germanica Pia Fidelis, the fleet which had its headquarters just south of Cologne, on the left bank of the Rhine.765 And lastly, one,

757

Bosman 1997, 8.

762

See for the iconography of Hercules 3.8.3.

758

Bosman 1997, 275, 284.

763

Plin. HN 4.101.

759

Bosman 2012, 361-364.

764

Van der Feijst, 2007,15.

760

Dijk 1987, Bosman 1997, 48, 318.

765

Derks 2008b, 149-157, De Bruin 2019, 105-108.

761

Bosman 1997,290-291, 318, 2012, 365-367.

91

possibly two fragments of a large inscription plate also mention the Classis Germanica. It seems, therefore, plausible that this fleet had a base at Naaldwijk. The idea that Naaldwijk-Hoogwerf was a military vicus seems justified, even though the quantity of imported pottery and militaria is less than in other military vici.766 A small handmade pot was found next to the southern wall of a house. The vessel, dating to the late Iron Age or early 1st century, was buried somewhere in the 1st century, when the house was built or abandoned, which is why the excavators have interpreted it as a ritual deposit.767 A comparison with a house plan (NZH1001) at Naaldwijk- ’t Zand Heultje suggests that the pot was buried near or under the entrance.768 Besides the bronze hand and the 28 classis inscriptions on building material, three more inscription fragments came to light, as well as several fragments of large bronze statues and appliques. Another bronze hand was found at Voorburg-Arentsburg.769 The project Römische Großbronzen am UNESCO-Welterbe Limes has demonstrated that the number of large bronze statues in military settlements along the limes must have been overwhelming.770 The statues mostly depicted emperors, on horseback or standing, addressing the troops. The latter can be deduced from the adlocutio gesture of several right hands, including those from Naaldwijk and Voorburg-Arentsburg. The emperor made the gesture at the beginning of his speech. The metal finds from Naaldwijk-Hoogwerf also include a bronze bust of Serapis.771 • South Holland: The Hague-Ockenburgh A settlement came to light during excavations by Holwerda between 1930 and 1938 in Ockenburgh, at the southwest side of The Hague, at a short distance from the Nord Sea coast. Based on the scarcity of Roman fine ware, metal and glass Holwerda concluded that the settlement was a ‘poor Batavian hamlet’. The discovery of the settlement at The Hague-Scheveningseweg (see hereafter) in the late 1980s rekindled interest in the settlement at Ockenburgh, not in the last place because of the quantity of militaria that had come to light during Holwerda’s excavations. New excavations in the 1990s enabled the excavators to conclude that the Batavian hamlet was in fact a small castellum, occupied between 150 and 180.772 Horse burials and metal horse gear fittings suggest that the fort probably housed a small cavalry unit.773 In the same period, a vicus came into existence at Ockenburgh, which was longer in use than the small castellum. It even expanded in the 3rd century, which is why it has been advanced that it housed small military units in this period. Habitation came to an end in the early 4th century.774 The question remains what the function was of small forts like Ockenburgh, since it was situated 21 kilometres south of the limes and around two kilometres inland from the North Sea coast. According to Bogaers, the Meuse estuary was protected by cavalry units, housed in small forts.775 Waasdorp and others have suggested that military sites along the coast such as Ockenburgh castellum and vicus were part of a coastal defence system, presumably to protect the coast against raids by pirates.776 Since archaeological evidence for such raids is lacking, De Bruin has advanced that these coastal military installations protected the overseas transport routes between the continent and Britannia.777 The excavations at Ockenburgh have yielded four terracotta fragments of statuettes: the head of a boy or Bacchus, the head of a mother goddess figurine, a boar and a fragment of a horse.778

766

See for the interpretation of Naaldwijk-Hoogwerf as a mil-

774

De Bruin 2019, 98-103.

itary community De Bruin 2019, 105-108.

775

Waasdorp 2012, 132-135 and notes 172-173.

Van der Feijst 2007, 26.

776

Waasdorp 2012, 142-146.

768

Van der Feijst 2007, 27 fig. 3.5 and 2012, 40, fig. 4.8.

777

De Bruin 2019, 110.

769

RMO, inv. nrs. h1933/11.1, AR 1022a.

778

The horse: Van Boekel 1987, nr. 244. The boar: Van Boekel

770

Müller 2015.

1987, nr. 225 and Waasdorp/Zee 1988, 23, nr. 7.4. Head of

771

Van der Feijst 2007, 73.

a boy/Bacchus: Waasdorp/Zee 1988, 41 nr. 7.2. Head with

772

Waasdorp 2012, 11-24.

fragment of torso of a mother goddess figurine: Van Boekel

773

Waasdorp 2012, 45.

1987, nr. 85.

767

92

• South Holland: The Hague-Scheveningseweg At the north side of the city of The Hague, along the road to Scheveningen on the North Sea coast, excavations between 1984 and 1987 produced traces of a settlement with two habitation phases. In its earliest phase it was a native settlement, occupied during the first half of the 2nd century. Finds from the settlement consisted mainly of handmade, native pottery and hardly any imported goods. The second habitation phase, from around 180 until 240, yielded a large quantity of imported goods, including militaria such as armour fragments, arrow and spear heads, daggers and swords. Therefore, the site has been interpreted as a military vicus, although no traces of a fortification have been found at Scheveningseweg.779 The settlement was abandoned around 240.780 Besides pits filled with refuse, several pits contained remarkable objects, such as a pit with a tin dish, two pieces of horse gear, two drinking cups (painted ware) and a Neolithic or early Bronze Age axe, made from jadeite.781 Another pit contained the skull of a young man.782 At the same site, over hundred fragments of terracotta statuettes came to light, the largest known assemblage of statuettes in the Netherlands so far.783 Van Boekel has managed to identify at least twenty statuettes. Publications about Scheveningseweg do not mention any specifics concerning the find circumstances of the statuettes, the location of the pits with ritual deposits, the number of pits and their precise content. According to the excavators ‘here and there among the refuse pits, there were pits which did not contain refuse, but complete objects.’784 This implies that there were more than two pits with a remarkable content, but only the content of the pit with the jadeite axe has been described in detail. Based on the presence of the statuettes and the pits, this part of the site has been interpreted as a cult place.785 Two coins from the site date to the second half of the 3rd century, which is why the excavators assumed that a new settlement was established here in this period. The deposits may have been made by the newcomers.786 The question why the statuettes were so fragmented must remain unanswered. Nine statuettes depict an identifiable deity, eight a human figure, three an animal. Among the deities, four to six represent Venus, two specimens depict Fortuna, one Cybele and one has been interpreted as Minerva with lowered visor (figs. 2.174-77).787 Among the human figures are two busts of a boy, one bust of a boy or girl, four specimens of a mother goddess and one of a dwarf.788 The three animal figurines depict two cockerels and a horse.789 • Zeeland: Aardenburg-castellum Since the 1950s, excavations in the centre of Aardenburg have yielded traces of a wall with towers and a gate, a large building and several others, including a small sanctuary. Outside the wall, traces of a stone building and other structures were found.790 Right from the beginning, the evidence caused a debate on the nature of the alleged fortified settlement. Was it a castellum, a fortified civil settlement or both? The excavator of 779

780

interpretation of the site also chapter 1.

Magendans/Waasdorp 1986, 297-300; Waasdorp 1999, 171; Mijle Meijer 2011, 8.

786

Waasdorp 1999, 158.

Waasdorp 1999, 171.Waasdorp does not exclude an earlier

787

Van Boekel 1989, 17-38 and note 12. See for an alternative

phase, since the site has not been fully excavated: Magendans/Waasdorp 1986, 297; Waasdorp 2012, 142.

interpretation of this figure 3.8.13. 788

See for Venus: Van Boekel 1989, 24, 26 nrs. 26-28, 29a-b, 31;

781

Waasdorp 1999, 158.

for Fortuna: ibid., 27, nrs. 33-34; for Minerva ibid. 28, nr. 35;

782

Although it has been suggested that the skull may have

for Cybele ibid. 30, nr. 37; for mother goddesses ibid. 17, nrs.

ended up there as a result of post-depositional processes:

14, 19, nrs. 16, 21; for busts of boys ibid. 32, nrs. 39-41; for a

Van Ginkel/Waasdorp 1992, 41. See for the interpretation of skulls in pits and ditches 4.2.2. 783

bust of a girl: ibid. nr. 34 (no image). 789

Magendans et al. 1988, 215; Van Boekel 1989, 7.

784

Magendans/Waasdorp 1986, 299.

785

Van Boekel 1989, 7; Waasdorp 1999, 158-159, 170. See for the

For cockerels Van Boekel 1989, 36, nrs. 47a-b; for a horse ibid. 37, nr. 48.

790

Besuijen/Van Dierendonck 2013, 9.

93

Aardenburg in the 1960s and 1970s, Jan Trimpe Burger, always believed in a military function of the settlement. In 2010 a re-evaluation and analysis of the excavations and its finds were carried out.791 According to recent research, the castellum was built to defend and control a sea branch with access to the North Sea. The earliest phase dates between 170 and 190, followed by a second phase between 185 and 222, during which stone-built principia were constructed. In 222 the principia were renovated and enlarged. A small, square post-built structure appeared that has been interpreted as a sacellum because of a terracotta statuette fragment from the terrain and an incomplete small sandstone altar, found just outside the precinct.792 The statuette depicts one of the Parcae.793 Around 245, the defensive ditch had silted up and several buildings, including the sacellum, were destroyed by fire. In the first quarter of the 3rd century, the fort gradually lost its military function and civilians lived inside the fort. Around 260 the fort was rebuilt and regained its military function for the defence of the southern North Sea coast. The fort remained in use until 280-285, although several finds indicate occupation into the 4th century.794 Bronze statuettes from the castellum include a statuette of Mercury, a statuette of Isis-Fortuna and two statuettes of Bacchus, one depicting the god as an adolescent, the other Bacchus as a boy riding a (missing) panther.795 The terracotta statuettes include the already mentioned specimen of one of the Parcae and two fragmentary statuettes of Venus.796 The bronze Isis-Fortuna was found 13 meters south-west of the sacellum.797 • Zeeland: Aardenburg-vicus Traces west and southeast of the fortification may have been part of the vicus. The western area had a more residential character while the south-eastern zone, with the bathhouse, had a more ‘industrial’ character. Here, evidence was found of metal working, although the iron slag and large quantities of shells may also point at activities such as the production of salt, fish sauce and glass.798 Among the twelve wells, four wells (numbers 9-12) in the eastern part of the vicus contained objects that possibly were special deposits, although they have not been described as such.799 Well 9 has yielded a bronze dish, a bronze wine dipper and strainer, two bronze saucepans, two handles of saucepans, several fragments of bronze saucepans and a fibula. From well 10 came a complete bowl (Dragendorff 33), from well 11 a complete terra sigillata mortarium (Dragendorff 45) and from well 12 several complete objects.800 Besides two moulds in the shape of a terracotta dove, two terracotta statuettes have been recovered from the vicus: the head of an unidentifiable woman and a fragment of a Minerva statuette.801 Uncertain find-spots of statuettes Find-spot

Material

Subject

Reference

Baexem-Graethemerweg

terracotta

goat/sheep

Van Boekel 1987, nr. 233802

Nijmegen-Canisius grounds

terracotta

Venus

Vermeulen 1932, 237, nr. 54803

Valkenburg-Marktveld

terracotta

mother goddess

Van Boekel 1987, nr. 76

Veldhoven-Heers

terracotta

Diana

Van Boekel 1987, nr 20804

Veldhoven-Heers

terracotta

Diana

Van Boekel 1987, nr. 21

Veldhoven-Heers

terracotta

Diana

Van Boekel 1987, nr. 22

791

Van Dierendonck/Vos 2013.

798

Van Dierendonck et al. 2013, 318-319.

792

Besuijen et al. 2013, 106-111 and figs. 6.12, 6.15.

799

See for the interpretation of objects in wells 4.2.2.

793

Van Boekel 1987, nr. 129.

800

Besuijen et al. 2013, 195-199. See for the metal objects

794

Besuijen/Van Dierendonck 2013, 31-32.

795 796

797

94

Besuijen 2008, 72.

Zadoks et al. 1969, nrs. 6, 8, 29, 53.

801

Van Boekel 1987, nrs. 56 and 136.

Van Dierendonck/Vos 2013, fig. 8.23. See for the figurines

802

In 1964, a stray find came to light at the site. it depicts a

Van Boekel 1987, nrs. 107bis, 120 and 129.

rectangular terracotta plinth with inscription. The inscrip-

Besuijen et al. 2013, 111.

tion reads ACCEPTV[S] F[EC] CCAA. A year later, another

2.5

Conclusions

The total number of statuettes from the study area amounts to 528 identifiable statuettes and statuette fragments: 196 metal, 325 terracotta, four amber, two limestone statuettes and one made of chalk.805 Together with the unidentifiable fragments, the total amounts to 691 (chart 2.1).806 89 statuettes came to light north of the Rhine, 83 identifiable and six unidentifiable specimens (chart 2.2).807 The identifiable statuettes comprise 67 bronze, 15 terracotta specimens and one chalk figurine. Since trade or barter between Romans and northern peoples only took place on a small scale, imported terracotta statuettes are almost absent in the north. Presumably locally made clay statuettes are known from a few Roman period settlements. The majority of statuettes come from the terp region in the northern coastal area, which has yielded predominantly bronze statuettes. They were found during the quarrying of terpen and were most likely brought to the north by returning Germanic auxiliaries, for instance as souvenirs or a means of payment, or by Romans as diplomatic gifts. They were left behind when terp dwellers gradually abandoned the northern coastal area in the course of the 3rd and 4th centuries. In this chapter it is advanced that these statuettes may have been ritual deposits. The only deposit with a bronze statuette that came to light during a controlled excavation has been recovered from the bottom of a ditch in a levelled terp at Tzum (Friesland).808 South of the Rhine, the total number amounts to 602 specimens, of which 446 have been identified. They comprise 131 bronze statuettes, 309 terracotta, two limestone and four amber specimens (chart 2.3).809 The uneven distribution of statuettes is striking. The rural settlement at Tiel-Passewaaij, for instance, has yielded 13 statuettes, whereas hardly any statuettes have been recovered from settlements in the hinterland of the limes zone, in spite of a demonstrable Roman influence. Examples are Cuijk (North Brabant), Rijswijk, Ouddorp (South Holland) and villas in the province of Limburg. As to military settlements, numerous statuettes have been found at Vechten, but the fort at Alphen aan den Rijn has yielded only two specimens, despite an enormous quantity of other finds. Apart from treasure hunting and recycling of bronze, other explanations for the relative scarcity of statuettes in the study area can be advanced. Most settlements and sites have only been partially excavated, which implies that excavators may have ‘missed’ the statuettes. Another reason could be that people were not interested in statuettes. What archaeologists usually find when excavating a settlement is what people left behind, the result of adverse selection or Negativauslese: household refuse, immovable objects, graves and objects that were too heavy for transportation. This then would suggest that the statuettes we find in and

803

fragment of the same statuette came to light. It depicts the

fitted a head that was discovered a couple of years later. A

legs of a standing quadruplet with a thick pelt. Based on

fragment depicting the back of a woman with a quiver was

these and other finds, like the remains of a Roman period

found during excavations by W.C. Braat in 1929, but the find

bustum grave with a coin, a glass bead, pottery sherds, a

circumstances were not recorded. It could not be estab-

coin and a sherd of a terra sigillata bowl, Bogaers suspected

lished whether the fragment belongs to the second figurine

the presence of a cemetery at Baexem, just east of the road

or to a third statuette of Diana, but they do come from

to Grathem: Bogaers 1958, 182; Bogaers 1964, 62-63.

the same mould: Van Boekel 1987, nrs. 21-22. A statuette of

See also Van Boekel 1987, 197, note 439 and nr. 107. She

Diana from Voorburg also comes from the same mould. See

suggested that the fragment may have come from the 804

for this fragment Bink/Franzen 2009, 213 and note 427.

Hunerberg cemetery.

805

See table 1.

In 1871, a substantially complete terracotta statuette of

806

In reality, these numbers are, of course, higher, as a result

Diana was found at Veldhoven-Heers, at the location

of detector finds and new excavations.

Koningshof. This site contained traces of habitation and

807

See table 2.

a cemetery. See Van Boekel 1987, 287-291 and nr. 20. On

808

See also 4.2.2.

the same site two fragments of another Diana statuette

809

See table 1.

came to light. A chest fragment was found in 1871 and

95

around settlements were left behind as refuse, but I have argued in chapter 1 that this is unlikely, because the majority of figurines were objects of worship. The majority, 85 percent, have been recovered from dry and watery places in and around settlements, of which excavators have interpreted two percent as ritual deposits. Seven percent of the statuettes come from graves and less than six percent from sanctuaries.810 The majority in the study area come from the cemeteries at Nijmegen. Despite the thousands of investigated graves at Nijmegen, only a few figurines have been found in situ, of which four in late Roman period graves. All statuettes recovered from cemeteries elsewhere in the study area are stray finds. It appears from the description of settlements and find-spots in this chapter that a substantial number of statuettes have been recovered from pits, ditches, drains, wells and other watery places. Details of the find circumstances of statuettes and statuette fragments are often missing in early archaeological publications. Also remarkable is the fact that most bronze statuettes from settlement excavations are stray finds found in top soil.

810

See for the scarcity of statuettes in graves in Belgium De Beenhouwer 2005, 823. The distribution of terracotta statuettes inside and outside the 2nd-century city walls of Roman Tongeren reveals a different pattern: 38% dwelling area, 38% cult place, 18% burial ground, 6% undefined De Beenhouwer: 2014, fig. 43.

96

In all charts, the numbers include complete and incomplete statuettes

Chart 2.1

Chart 2.2

97

Chart 2.3

Chart 2.4

98

3

Spatial and iconographic analysis

3.1

Introduction

The aim of this chapter is to answer the first two research questions of this study that have been formulated in the first chapter: 1. Can we observe relations between size, material and subject of statuettes and their find-spots in the three areas and if so, how can we explain these relations? 2. Were there preferences for specific subjects or iconographies in the three areas and if so, how can we explain these preferences? To answer the first research question, it will be examined if there is a relation between the size of statuettes and the area where they were found, and secondly, whether there is a relation between material and statuette representations in the three areas. The second topic concerns the iconography of statuettes and specific iconographic elements. It has been argued in the first chapter that creolisation is an important component in the process of shaping new cultures and identities. It will be examined to what extent this applies to the shaping of iconographies of Roman period statuettes. By combining written sources and archaeological evidence, we will first explore in how far images of deities were a new phenomenon to the inhabitants of western Europe when the Romans introduced statuettes of deities in the northwestern provinces. This is necessary for a better understanding of how deviating, non-Roman iconographic elements could appear in depictions of Roman deities and other figures.

3.2

Written sources and the archaeological evidence

Based on written sources, it is often assumed that Germanic and Gallic cults were predominantly an-iconic before the Roman conquest. The Greek historian Diodorus Siculus (c. 90-30 BCE), reporting on the Celtic attack on Delphi in 279 BCE, states: ‘Brennus, king of the Gauls, on entering a temple found no dedications of gold or silver, and when he came only upon images of stone and wood, he laughed at them, to think that men, believing that gods have human form, should set up their images in wood and stone’.811 Following Diodorus Siculus or other sources, the historian Tacitus (c. 55-c. 120) tells us in his Germania: ‘Apart from this they (the Germans) deem it incompatible with the majesty of the heavenly host to confine the gods within walls, or to mould them into any likeness of the human face: they consecrate groves and coppices, and they give the divine names to that mysterious something which is visible only to the eyes of faith.’812 811

Diod. Sic. 22.9: Ὅτι Βρέννος ὁ τῶν Γαλατῶν βασιλεὺς εἰς ναὸν

812

Tac. Germ. 9: ceterum nec cohibere parietibus deos neque in

ἐλθὼν ἀργυροῦν μὲν ἢ χρυσοῦν οὐδὲν εὗρεν ἀνάθημα, ἀγάλματα

ullam humani oris speciem adsimulare ex magnitudine cae-

δὲ μόνον λίθινα καὶ ξύλινα καταλαβὼν κατεγέλασεν ὅτι θεοὺς

lestium arbitrantur: lucos ac nemora consecrant deorumque

ἀνθρωπομόρφους εἶναι δοκοῦντες ἵστασαν αὐτοὺς ξυλίνους τε καὶ

nominibus appellant secretum illud, quod sola reverentia

λιθίνους.

vident.

99

On the other hand, a well-known contemporary of Diodorus, Julius Caesar (100-44 BCE), states in his De Bello Gallico: ‘Among the gods, they (the Gauls) most worship Mercury. There are numerous images of him.’813 Caesar’s plurima simulacra clearly refer to cult statues of Mercury.814 Not long after the Roman conquest, large images of deities in the form of cult statues appeared in sanctuaries, depicting Roman and native deities. According to Pliny the Elder, the Arverni, a Gallic tribe, paid the Greek sculptor Zenodorus no less than 40 million sestertii for a colossal bronze cult statue of Mercury.815 This must have taken place in the first half of the 1st century, because afterwards, Nero commissioned the same Zenodorus to make the colossal statue of Sol for him.816 Native peoples also worshipped their own cult statues. According to Tacitus, different tribes of the Suebi, who lived on the North-Frisian Islands in the region Schleswig-Holstein and in Jutland, worshipped Nerthus, a goddess who could be equated with ‘Mother Earth’. He describes a ritual that may have been rooted in prehistory, but his account suggests that it was still being performed in the 1st century. When Nerthus appeared in her sanctuary, she was covered with a robe and led by a priest through the land, sitting in a cart. When the cart with the goddess had returned in the sanctuary, slaves would wash the cart, the robe and the goddess in the lake, after which the slaves were drowned in the same lake.817 The goddess probably ‘appeared’ at the end of winter to ensure fertility and a good harvest in the summer. To understand Tacitus’ description of the ritual, we must realise that the goddess was a cult statue. It brings up the question whether this statue was anthropomorphic or, for instance, a simple piece of wood. Examples of the latter are known from the Mediterranean world, like the Samian xoanon of Hera, which before 600 BCE was no more than a wooden plank.818 The information by Latin and Greek authors may be confusing, but there is enough archaeological evidence of stone and metal sculpture throughout Europe from the Bronze Age onwards.819 Figurines were also made of perishable materials like wood, bone or chalk which have rarely survived in the archaeological record.820 Recovered from watery places, they are often interpreted as ritual deposits.821 In the study area, anthropomorphic wooden objects are known from the Iron Age into the early Medieval period. An example from the Roman period is made of hornbeam and was recovered from a pit at Heemskerk-Beverwijk (North Holland). This object, too, has been interpreted as a ritual deposit.822 It is unknown whether such objects represented human figures or supernatural powers. 813

814

815

Caes. BGall. 6.17: Deum maxime Mercurium colunt. Huius

inis multa cum veneratione prosequitur, laeti tunc dies, festa

sunt plurima simulacra.

loca, quaecumque adventu hospitioque dignatur. non bella

Whereas simulacrum is often used to describe a cult statue,

ineunt, non arma sumunt; clausum omne ferrum; pax et quies

imago repeatedly refers to statues and portraits of living

tunc tantum nota, tunc tantum amata, donec idem sacerdos

people or ancestors. See, for instance, Cic. Verr. 2.5.72, Caes.

satiatam conver-satione mortalium deam templo reddat. mox

BCiv. 2.5, Lucr. 5.75 and 5.308, Verg. Aen. 2.172. Tacitus uses

vehiculum et vestis et, si credere velis, numen ipsum secreto

it in this sense on eight occasions in the Annales (1.73, 3.63,

lacu abluitur. servi ministrant, quos statim idem lacus haurit

12.22, 14.12 and 14.32, 15.29, 15.44-45): Pearcy 1973, 83-122.

arcanus hinc terror sanctaque ignorantia, quid sit illud, quod

Plin. HN 34.45: verum omnem amplitudinem statuarum eius

tantum perituri vident.

generis vicit aetate nostra Zenodorus Mercurio facto in civitate

818

Clem. Al. Protr. 4.40.

Galliae Arvernis per annos decem, hs |CCCC| manipretii.

819

See for examples of Bronze Age sculpture, for instance, Bri-

816

Woolf 1998, 215.

817

See also 4.2.2. Tac. Germ. 40: Reudigni deinde et Aviones et Anglii et Varini et Eudoses et Suarines et Nuitones fluminibus

ard 1991; for Iron Age sculpture, for instance, Duceppe-Lamarre 2002. 820

See, for instance, Aldhouse-Green 2012, 69 and figs. 3.8, 4.24.3, 4.9-4.10.

aut silvis muniuntur. nec quicquam notabile in singulis, nisi quod in commune Nerthum, id est Terram matrem, colunt

821

Aldhouse-Green 2012, 90-92 and figs. 4.2-4.3.

eamque intervenire rebus hominum, invehi populis arbitran-

822

Lange 2017, 143, 148 and note 374. An example from the

tur. est in insula Oceani castum nemus, dicatumque in eo

early Medieval period comes from Heiloo (North Holland)

vehiculum, veste contectum; attingere uni sacerdoti conces-

and is made of oak wood: Lange 2017, 293 nr.165.

sum. is adesse penetrali deam intellegit vectamque bubus fem-

100

Although we are completely in the dark about who or what pre-Roman figures represented, there is a small figurine from an Iron Age site that may depict a deity. Excavations in the 1990s on the Glauberg plateau in Hesse, Germany, yielded structures and two burial mounds belonging to an important Celtic oppidum dating to the 6th or 5th century BCE.823 Several statues were recovered from the site, the most complete being a 186 cm high limestone statue of a male figure (fig. 3.1). It was found just outside one of the burial mounds and depicts an armoured standing warrior. Although known as Der Keltenfürst, it is not clear who is depicted here: the deceased, a deity or an ancestor? A bronze figurine, an attachment on the rim of a bronze jar, comes from the same site and represents an armoured male figure sitting in crossed-legged position. In this case, there are more indications that a deity is depicted.824 Several Gallo-Roman representations of seated, cross-legged male figures are known from Gaul (fig. 3.2). Although the name is only attested once by an inscription on the so-called Pillar of the Boatmen in Paris, such figures are thought to depict the antler-god Cernunnos.825

3.3

Deities, human figures and animals

The first category in the dataset comprises bronze and terracotta deities of the Roman pantheon. These deities have been so often described and depicted that the iconographies and identification are seldom problematic. In this respect, the evidence from the Netherlands does not deviate from elsewhere in the provinces.826 Statuettes of Mercury, Venus, Fortuna and Minerva dominate the spectrum. Statuettes related to eastern and Egyptian cults like that of Cybele and Isis are present as well. After the appearance of production centres of terracotta figurines in Central France and the Rhine-Moselle area, statuettes gradually began to change.827 Most changes were subtle, like Roman- looking deities with ‘non-Roman’ attributes. Also new combinations were forged, which are recognisable through translations into a syncretized iconography, such as Isis-Fortuna and Jupiter-Ammon. North and south of the Rhine, statuettes have been found with deviating iconographies. In some cases, their find-spots suggest that elements of these iconographies developed in the Roman period, while others could be rooted in pre-Roman times.828 The second category within the dataset is composed of human figures. Their function and meaning are not always easy to grasp. First, the term human figures needs to be explained. Within this category we can distinguish three groups. The first includes figures who are no deities, but no mortal beings either. They represent powers or abstract notions which in the course of time became anthropomorphic, like lares and genii. Some figures represent beings associated with a deity, such as the followers of Bacchus: satyrs, maenads and Silenus. Others embody river sources, streams and springs or they represent personifications of seasons, rivers or regions. 823

Hesse is Hessen in German.

example of a bronze statuette in the Netherlands comes

824

Aldhouse-Green interprets the figure as a subjugated war-

from the earliest military fort at Velsen, North Holland. See

rior: Aldhouse-Green 2012, 23, 30-32 and figs. 1.9-2.1.

for Velsen 2.4.2. So far, no specialized production centres

825

Deyts 1992, 13-22. See for the Pillar of the Boatmen 3.8.2.

for bronze figurines have been identified. We only know

826

See, for instance, Menzel 1960, 1966, 1986; Faider-Feytmans

from a few sites at, for instance, Nijmegen (the Nether-

1979; Fleischer 1967; Kaufmann-Heinimann 1977, 1994,

lands), Blicquy and Libergies Bons-Villers (Belgium) that

1998; Ritter 1994, 1999.

there were workshops where, besides other metal objects,

Van Boekel 1987, 204-213. According to Van Boekel, the ear-

statuettes were cast. See for a semi-fnished bronze statu-

liest examples of terracotta figurines can be dated around

ette from the canabae at Nijmegen 2.4.1 and fig. 2.123; for

827

75: Van Boekel 1987, 242-243. However, fragments of a terracotta Venus statuette from the Kops Plateau come from a layer that has been dated by the excavators between 10

workshops in Belgium Faider-Feytmans 1979, 18-20. 828

See for the meaning of triple-horned bulls and horned birds 3.8.19 and 3.8.20.

BCE and 10. See chapter 2 and figs. 2.144-145a. The earliest

101

A well-studied group within the category ‘human figures’ are enthroned women with fruits, a dog or a child in their lap. Although they are often called mother goddesses, matres or matronae, it has been argued that they are not ‘real’ goddesses, but ancestral mothers, since their proper name refers to a certain group or region, like the matronae Aufaniae and the matronae Alaverhviae in the Rhineland and the matres Hannanefatae, the matres Treverae and matres Remae outside Ubian territory.829 Inscriptions in which the term mater is used alongside matronae suggest that there was no sharp distinction between the two.830 They are often depicted as a single, enthroned woman, and occasionally as a triad. Triads of mother goddesses are well-represented in the Rhineland, particularly in the area between Neuss, Bonn and Aachen. Since single enthroned women are sometimes depicted suckling one or two infants, the term dea nutrix seems appropriate for this type of representation, although there might be other, hidden meanings involved. The native goddess Nehalennia shares several iconographic elements with these enthroned mother goddesses and, as we will see, terracotta statuettes of Fortuna and Minerva have copied their enthroned pose. To this issue I will return in paragraph 3.8.9. The second group within the category ‘human figures’ are terracotta busts of women and children, young boys in particular. They have been recovered from domestic, funerary and other ritual contexts in the northwestern provinces. The faces of the terracotta busts show little variation, which makes it unlikely that they were portraits. It seems more plausible that they were symbolic representations or ‘types’ such as the young boy or girl, the adolescent or the nubile young woman. The best represented are busts of smiling, usually bald-headed boys, sometimes wearing a crescent moon shaped amulet, a lunula. Because of their smiling faces they are known as Risus busts. Female busts depict young women with a high coiffure, wearing earrings and a necklace, and, in some cases, a lunula. The busts have been connected with burial rites and rites of passage, but many aspects of their iconography remain unclear.831 The last group within the category ‘human figures’ comprises a wide variety of imagery such as warriors, riding acrobats, worshippers, thorn pullers, embracing couples, ithyphallic or drinking African men, caricatures and dwarfs.832 Although they seem to represent non-divine figures, it remains unclear whether the function of these statuettes was religious, symbolic, apotropaic or just decorative. A few specimens will be discussed in this chapter. The third category is composed of animal statuettes. They represent domestic and exotic animals as well as animals with deviant characteristics, such as the ‘horned birds’ from Nijmegen, triple-horned bulls and ithyphallic dogs.833 Deities are often depicted or associated with domestic or exotic animals, like Mercury with a billy goat or a cockerel, Diana with a roe, stag or hunting dog, Jupiter with an eagle, Juno with a peacock and Bacchus with a panther or lion. Many suggestions have been made concerning the use and meaning of animal figurines. Figurines of domestic animals have been interpreted as symbolic offerings, children’s toys, or food for the dead. However, these interpretations do not explain the specific characteristics of some animal figurines which they do not possess in real life. The spatial analysis begins with the area north of the Rhine, followed by the area south of the Rhine and the limes zone. I will not give a description of every statuette, because the majority have been described meticulously in catalogues.

829

830

Which is why this type of representation falls into the cat-

831

See for thorn pullers, for instance, Meinecke 2016, 129-160.

egory ‘human figures’. Derks has argued that the origin of

832

Meinecke 2016, 129-160.

the cult was an anthropomorphic ancestor cult: Derks 1998,

833

See for the horned birds from Nijmegen: Van Boekel 1987,

124-127 and notes 220-222. See also Raepsaet-Charlier 2015,

750-755 and nrs. 238-241; for ithyphallic dogs Van Boekel

204-217. She has suggested that the origin of the cult may

1987, nrs. 219-223; for three-horned bulls Zadoks et al. 1967,

have been aniconic.

nr. 46 and Zadoks et al. 1969, nr. 65.

Derks 1998, 119-120.

102

3.4

The area north of the Rhine

The 89 statuettes and statuette fragments found north of the Rhine include 70 metal (79%), 18 terracotta statuettes (20 %) and one chalk specimen (1%). 83 specimens could be identified (charts 3.1- 3.3).834 The coastal region comprises two clusters of find-spots: the northwestern terp region, which was part of Frisian territory, and the area east of the river Lauwers, which probably belonged to the western border region of Chaucian territory. The majority of statuettes have been found in the northwestern part of the terp region. Here, Mars and Mercury outnumber all deities. Other statuettes represent Apollo, Jupiter and Hercules. Imported terracotta statuettes from production centres in Gaul or the Rhine-Moselle area are represented by two fragments, one of Fortuna, the other of a mother goddess figurine. Animal figurines from fired clay have been found at Schagen (North Holland), Sneek, Oosterbeintum (Friesland), and Enumerhoogte (Groningen). In the area east of the river Lauwers, imported terracotta statuettes are absent. The number of statuettes in the north-east is considerably smaller than in the northwestern terp region. Moreover, several possess deviating iconographies, like the two bronze figurines from Rottum, an Isis-Venus figurine from Houwerzijl, a small, fired clay figurine from Wirdum and a statuette of a dismounted cavalryman from Ezinge. They will be discussed below. Compared to the terp region, Drenthe, Overijssel and northern Gelderland have yielded more terracotta and bronze statuettes of female deities: Minerva and Venus, each represented by four, Victoria and Fortuna by two specimens each. Among the 15 animal statuettes from the region north of the Rhine a few specimens stand out because of their striking iconography. The first is a chalk figurine from Wirdum, which represents either a hunting brown bear or, as its greyish white colour might suggest, a polar bear (figs. 2.19-24). The iconography of the horse from Kantens (Groningen) differs distinctly from the three horse figurines from the coastal area, of which at least one and probably two, originally carried a rider (figs. 2.8-9, 2.17). Two triple-horned bulls have been found in the study area. One was found near Beilen (Drenthe), the other south of the Rhine at Vechten (Utrecht).

Chart 3.1

834

See also tables 1-2.

103

Chart 3.2

Chart 3.3

3.4.1

Spatial analysis: a relation between size and function

The bronze statuettes from the northern Netherlands present some distinct particularities. Starting with the size of the statuettes, the majority are distinctly larger than most of the statuettes found south of the Rhine. First it must be clear what is meant by ‘small’ and ‘large’. When we look at the dataset, there is a group of small statuettes, measuring predominantly between around 50 and 90 mm in height, and a group of large specimens, measuring well over 100 mm. A comparison between the Mercury and Mars statuettes north and south of the Rhine reveals a distinct difference (charts 3.4 – 3.7). North of the Rhine, the smallest of the 17 Mercury statuettes is 82 mm in height, the largest 166 mm .835 The average height is 120 mm. South of the Rhine, the bronze Mercury statuettes measure between 43 and 230 mm.836 Their average size is 80 mm.

835

836

The total number of Mercury statuettes found north of the

head and lower legs, but had an estimated height of 120

Rhine amounts to 18, but the original height of the speci-

mm. The posture and size of the statuette is comparable to

men from Idaard is unknown.

the Mercury from Kimswerd: Zadoks et al. 1967, nr. 25.

A seated statuette from Voorburg-Arentsburg misses the

104

Leaving aside the Mars statuette from Joeswerd that consists of a torso and a raised right arm (fig. 2.16), the height of the remaining eight Mars statuettes from the terp region varies between 88 and 230 mm, with an average height of 136 mm. South of the Rhine, the height of the eight Mars statuettes varies between 54 and 223 mm. The two largest specimens measure between 136 and 223 mm. The first was found buried in a sunken byre, the second comes from the river Waal. Their height is quite exceptional, compared to the other specimens south of the Rhine, which are considerably smaller.837 The average height of the eight statuettes is 114 mm, but without the two large specimens the average height is 92 mm. Most bronze statuettes found north of the Rhine are noticeably larger than the majority of specimens that came to light south of the Rhine, which is why I suspect a relation between their size and function. If they were diplomatic gifts, it seems plausible that the givers chose large specimens to make a good impression. As a means of payment, large bronze statuettes probably made the exchange of valuable goods such as cattle easier than small ones, either because of their weight in bronze or because of their quality as exotic objects.838 Some specimens, especially large specimens found in watery places or in the context of settlements, may have been ritual deposits. Still north of the Rhine, but outside the terp region, the large bronze statuettes include a Mercury from Tzum (154 mm), a Mercury from Noordbarge (115 mm), a Mercury from Dalfsen (164 mm), a Venus from Heerde (310 mm), a Minerva and a worshipper from Ede-Veldhuizen (129 mm and 115 mm respectively) and a Fortuna from Bennekom (128 mm). The statuettes from Ede-Veldhuizen and Bennekom have been interpreted as ritual deposits by the excavators.839 The find-spots and size of the three Mercury statuettes from Tzum, Noordbarge and Dalfsen as well as that of the Venus statuette from Heerde suggest that they may have been ritual deposits, too.840 The same could apply to several large bronze statuettes found south of the Rhine. Beside the two large Mars statuettes mentioned above, a statuette of Silenus (123 mm), Mercury (230 mm) and Hercules (144 mm) come from the rivers Waal and Rhine. A Luna statuette (130 mm) was recovered from a pit in the canabae at Nijmegen and a statuette of Isis-Fortuna comes from the fort at Valkenburg (135 mm).841 Another particularity is the distinct predominance of Mars and Mercury in the northern coastal region. As pointed out in the previous chapter, the statuettes in the terp region could have been brought to the north by Germanic auxiliaries, returning veterans, Romans and traders, as a souvenir, diplomatic gift or a means of payment. If Romans needed food, cattle or horses from the terp settlements, they probably exchanged goods instead of paying with coins. Be it as it may, none of these explanations accounts for the predominance of Mercury and Mars statuettes in the terp region. What could be the reason for the prevalence of these two deities? A first possibility is that statuettes of Mars and Mercury ended up in the north because it was a matter of availability at the time. If the statuettes were diplomatic gifts or a means of payment, Roman authorities or traders may have chosen these two deities because they assumed the inhabitants of the north had a predilection for them. If Frisii and Chauci indeed preferred Mercury and Mars above all other representations, they may have picked one of these two deities when they were offered a statuette. The same applies to returning veterans who wanted to bring a bronze statuette to their homeland. It appears from dedicatory inscriptions that native deities were predominantly associated with Mars and Mercury.842 Although we do not grasp the mechanisms behind the associations, it is possible that people 837

See for the Mars statuette from Helden 2.3.5.; for the Mars

for the Mercury from Noordbarge fig. 2. 39, 2.2.4 and 3.5.1;

from the river Waal 2.3.4 and Zadoks et al. 1969, nr. 34.

for the Mercury from Dalfsen 2.2.4 and 3.5.1; for the Venus

838

See 2.2.1.

statuette fig. 2. 50 and 2.2.4. See for the overview of possible

839

See for the two settlements 2.2.4. The bronze Victoria from Colmschate has also been interpreted by the excavator as

ritual deposits 4.2.2 and map 11. 841

Isis-Fortuna 2.4.1 and 4.2.2.

mm in height: Hermsen 2007, 178. 840

See for the Mercury from Tzum figs. 2.183-85, 2.2.2, 3.5.1;

See for possible deposits from rivers 4.2.2. See for the statuette of Luna fig. 2. 123, 2.4.1 and 4.2.2; for the statuette of

a ritual deposit, but this is a small statuette, measuring 67 842

Derks 1998, 95 and appendices 3.2-3.3.

105

from the northern coastal area shared the appreciation for these two deities with the people in other parts of Germania and in Gaul. Some Frisii and Chauci probably knew the deities from their service in the Roman army and may have worshipped Mars as the bringer of agricultural fertility or venerated the deity for his martial appearance and capacities. The popularity of Mercury could be a consequence of the broad range of his qualities (see below). In any case, the inhabitants of the northern coastal region must have appreciated bronze statuettes of Roman deities, because if they had only been interested in the bronze itself, they would have exchanged them for other goods or melted them down to make new objects from it. A last peculiarity concerns the execution of 34 of the 41 the statuettes of deities and human figures from the northern terp region. Seven specimens stand out because of their small size and/or peculiar iconography: the seated goddess (?) and Mercury from Rottum, the cavalryman from Ezinge, the two Isis-Venus figurines from Houwerzijl and the Jupiter-Sabazius from Menaldum (see below). The remaining specimens seem well-executed, at first sight. The figures stand in classic poses and the treatment of the body is naturalistic. Several specimens have silvered eyeballs or eyes inlaid with silver or glass.843 A Minerva statuette has a gilded helmet and aegis.844 However, closer inspection reveals several imperfections. In the case of seven male deities, the muscles of the body and buttocks are clearly indicated, while other details, the facial features in particular, are sketchily rendered, by simple incisions.845 Fourteen figures have an awkward body posture, disharmonious proportions or indistinctly, poorly modelled hands and/or feet.846 A large Isis-Fortuna statuette with silvered eyeballs has a deformed right hand, due to a casting flaw.847 Since many statuettes found south of the Rhine present all kinds of imperfections, too, such imperfections may not seem surprising. True, but those statuettes are usually small specimens, produced as an offshoot of metalworking. Among the twenty large statuettes found south of the Rhine, by which I mean specimens over 100 mm in height, only three statuettes present imperfections like clumsily modelled limbs or casting flaws.848 The relatively high number of poorly executed statuettes from the northern Netherlands, 34, may be coincidental, but if not, what could be the reason? In 1980, Annie Zadoks-Josephus Jitta suggested that Romans might have ‘dumped’ statuettes in the region north of the Rhine.849 She advanced the possibility in a short article about a matrona statuette, from Krangeweer (Loppersum, Groningen). It was found together with a statuette of Mercury.850 Based on the portrait-like features, she argued that the statuette represented the personal Juno of a mortal woman, none other than Domitia, the wife of the emperor Domitian (81-96).851 The Juno was the female equivalent of the Roman Genius, a divine power, protector and creator of life. A Juno is usually portrayed as a matron, a married woman bringing an offering. Because the Roman Senate considered him a tyrant, Domitian’s memory was condemned to oblivion after his death, a verdict that may have applied to his wife Domitia as well. Since the statuette had become ‘unsaleable’ within the Roman Empire, Zadoks suggested it was dumped in Barbaricum.852 Her suggestion brings up the question whether large statuettes with imperfections like the ones I described above, were deliberately ‘dumped’ in Germania Libera. Did the Romans assume that the barbarians would not notice the flaws? This seems difficult to reconcile with the idea that statuettes may have been diplomatic gifts. Moreover, Germanic men who had served in the Roman army must have been familiar with bronze statues of gods and emperors. Be it as it may, the inhabitants of the terp kept the statuettes for a 843

Zadoks et al. 1967, nrs. 3, 8, 11, 14, 17, 25. The bronze Mercury

848

from Tzum, too, had eyes inlaid with silver or glass: figs.

more large statuettes, which are not taken into account

2.183-185. 844

Zadoks et al. 1969, nrs. 23, 34 and 49. The dataset comprises here because of their fragmentary condition.

Zadoks et al. 1967, nr. 41.

849

Zadoks 1980-1981.

845

Zadoks et al. 1967, nrs. 12, 21, 23, 29, 31, 33, 34.

850

Zadoks et al. 1967, nrs. 34. and 43.

846

Zadoks et al. 1967, nrs. 2, 13, 17-20, 23, 25, 30, 33, 35, 36. See

851

Zadoks 1980-1981, 207 and figs. 1-2.

also figs. 2.4-2.5 and 2.183-85.

852

Zadoks 1980-1981, 210.

847

Zadoks et al. 1967, nr. 11. See also figs. 2.6-2.7.

106

reason. May be they appreciated the statuettes despite the imperfections, for instance because the fact that these statuettes represented powerful deities was more important than their imperfect technical execution.

Chart 3.4 *Hitherto unpublished statuettes are indicated with inventory number

Chart 3.5

107

Chart 3.6

Chart 3.7

108

3.5

Iconographic analysis

3.5.1

The iconography of the Mars and Mercury statuettes

In the study area, Mars is predominantly represented as a beardless youth. The Mars statuettes are, with one exception, all of the same type: a standing, nude male, wearing a crested, Corinthian helmet, holding a lance in the raised right hand and another object in the left, or with the left arm resting on a shield. Most of these attributes are usually missing. The only exception to this iconography is a Mars statuette from Friesland. Besides a double crested helmet, the god wears armour and boots and holds a (missing) lance in his raised right hand and a (missing) object in his left.853 He has a beard and drooping moustache, an iconography that could be derived from famous sculptures such as the cult statue of Mars Ultor from the Forum of Augustus in Rome.854 Other candidates are well-known sculptures from the Hadrianic period. A famous example is the statue group of Hadrian and his wife Sabina portrayed as Mars and Venus.855 Another salient detail is the cape that hangs down from the statuette’s left shoulder over the elbow and forearm. It is a paludamentum, worn by military commanders and emperors portrayed as the supreme commander. The only Mars statuette that slightly resembles the statuette from Friesland comes from Greater London, but depicts the god without the cape.856 The Mercury statuettes present a more varying iconography. The god is depicted either seated or standing, with or without a winged cap, winged boots or cape, usually holding a herald’s staff. This staff, a caduceus, is entwined by two serpents and sometimes surmounted by a pair of wings. In the past, typologies of Mercury statuettes have been drawn up, mainly based on the chlamys.857 The ubiquity of Mercury in the Roman Empire could be a consequence of his multiple qualities. He was the god of financial gain and prosperity, eloquence and divination, guide of souls in the underworld, messenger of the gods, intermediary between mortals and the divine, protector of travellers and thieves. He is equated with the Greek god Hermes, but there are some differences, to which I will return. Presumably via the Etruscan god Turms, the cult of Mercury was introduced in the early Roman Republic, together with the cult of the triad Liber, Libera and Ceres, deities associated with fertility and cultivation. The earliest known temple for Mercury was inaugurated in 495 BCE and stood on a slope of the Aventine hill, overlooking the Circus Maximus. The construction of the temple was ordered by the Sibylline Oracle after a crop failure in 496 BCE. The inaugurator of the temple also established a collegium mercatorum and made arrangements for the distribution of grain among the people. Not far from the temple, goods from the fertile Campanian farmlands entered the city through the Porta Capena, where the Aqua Mercurii was situated, a spring where on festival days and the memorial day of the inauguration of the temple, incoming goods were cleansed in honour of Mercury. The god’s association with trade can also be deduced from his name, since the Latin noun merx means commodity or merchandise.858 The importance of Mercury as protector of traders and trade must be seen in the light of an agricultural and pastoral society where the importance of trade was strongly increasing, eventually leading to a monetary economy in the 3rd century BCE.859 In the 2nd century, the cult of Mercury reached a peak under the reign of Marcus Aurelius. Coins with Mercury and the words felix, pacifer or religio Augusti in the legend indicate that the god was also associated with victory,

853

Zadoks et al. 1967, nr. 16.

854

See for Mars as a nude, beardless youth Simon 1984, 520-

857

See, for instance, Bauchhenß 1992, 537-554; KaufmannHeinimann 1998, 46-50; for the limitations of such typolo-

521; Bauchhenß 1984, 572; for the bearded Mars Simon 1984,

gies Veen 2014a, 137-138.

515-517.

858

Simon 1992, 500-501; Andreussi 1996, 245-247.

855

Louvre Museum, Paris, inv. nr. Mr 316/Ma 1009.

859

Simon 1992, 534.

856

Durham 2012, nr. 23.

109

peace and piety. The latter is emphasized by coins depicting Mercury with a patera or sacrificial animals like billy goats. Especially from the second half of the 2nd century onward, Mercury was a popular theme on sarcophagi.860 Despite the varying iconographies in the study area, nearly all Mercury statuettes of which the arms have been preserved, depict the god holding a purse in his outstretched hand. The purse or marsupium is the attribute that underlines Mercury’s first and probably most important quality in the Roman period: ensuring financial gain and prosperity. Mercury holding a money bag must have been introduced when bartering goods no longer played a part in commerce. It is, therefore, understandable that his Greek counterpart Hermes is never depicted with a purse on vases and other media. Exceptions are a few Roman copies of Greek statues, which suggest that the purse was a later addition. In the Greek world, especially in the 5th and 4th centuries BCE, Hermes is usually depicted with a kerykeion, a herald’s staff, and a petasos, a flat headdress with a broad, floppy brim, worn by travellers. Occasionally, he wears winged sandals, but seldom a winged cap.861 Most bronze Mercury statuettes from the study area originally also carried a herald’s staff, a caduceus, referring to his quality as messenger and mediator between men and gods.862 In this respect, the bronze Mercury from Tzum is exceptional. The 154 mm high, solid cast statuette depicts the god with a purse in the right hand and a (missing) caduceus in the left (figs 2.183-85). A small cape covers the upper back and part of the buttocks, but leaves the shoulders bare. The statuette stands out for several reasons. First, because it depicts a bearded Mercury, whereas the Roman god is usually represented as a beardless, young man. Second, instead of wearing a winged cap or winged sandals, wings grow from his head and ankles. Wings growing out of Mercury’s head or ankles are known from Roman copies of Greek statues, but I do not know of any examples that combine both. How can we explain the beard and this rare combination? The execution of the statuette reveals a mixture of craftsmanship, care and what seems a lack of experience in modelling human figures. Striking are the detailed rendering of the head with curly hair, wings, the (originally) inlaid eyes and the well-proportioned upper body. They contrast with the awkward body posture, the disproportionally thin, short limbs, the stiff rendering of the cape and sketchily rendered feet. Bronze statuettes were often made as an offshoot of metal working. The result depended on the modelling skills of the bronze caster and the availability of models. Although initially based on famous statues which became standard types, most bronze statuettes are unica.863 They were assembled from copied parts of different models, or they were a mix of copied parts and freely modelled elements.864 In my opinion, the statuette from Tzum is exactly that. It was assembled from two to three models. Originally, the head belonged to a statuette or applique in the shape of a bust of a bearded god or emperor. The size and eccentric position of the head suggest that it was later affixed to the body or copied body of another statuette. The hybrid origin could also explain the two pairs of wings if the legs with wings were derived from a third specimen. The cape and wings on the head were freely modelled, cast separately and added afterwards. Another striking representation is the 83 mm high figurine from Rottum, found together with a small bronze female figure (see below). It depicts a bearded, standing man with a short, flat body, long arms and large hands, thin legs and tiny feet (figs. 2.19-20). His fringe beard is flat and stylised. The muscles are not indicated and his back and buttocks are flat. The figurine has been identified as Mercury, based on the object in the right hand which, according to Zadoks et al. is a purse.865 In the raised left hand the man holds an oval, unidentifiable object with a scalloped border. It has been suggested that the form of the raised left arm, in 860

Simon 1992, 537.

861

Siebert 1990, 288-289.

862

cury depict a caduceus. The other fragments depict the head or base of the figurine.

Usually, the bronze caduceus of Mercury statuettes is miss-

863

Veen 2014a, 137-138.

ing, because it is cast and added separately. The presence

864

As a consequence of the production method, it is usually impossible to attribute a date to bronze statuettes.

of the missing attribute can be deduced from the position of the hand and fingers. Five terracotta statuettes of Mer-

110

865

Zadoks et al. 1967, nr. 40.

combination with the oval object, depicts a misunderstood cornucopia.866 In my opinion, the object could represent a patera, since it resembles the object a bronze worshipper from Ede-Veldhuizen holds in the left hand (figs. 2.44-46). Here, the scalloped border is indicated by incisions. Since all other characteristics deviate from Mercury’s known iconographies, we have to take a broader look and consider other possibilities. If the object in the right hand is not a purse, but, for instance, a jar, the statuette could depict a worshipper, priest or servant, or a figure whose identity and meaning remain hidden from us. 3.5.2

The Isis-Venus figurines from Houwerzijl

Two almost identical bronze figurines were found in a terp south-west of Houwerzijl (Groningen), twenty kilometres south-west of Rottum. One was found in 1960 during ploughing and was published in 1967. 867 Apparently, the second was found in the same terp, but the find circumstances are unknown. This statuette is part of a private collection.868 The statuettes depict a nude woman with a plume on her head, standing on a round, flat plinth (fig. 2.15). There are a few differences: the statuette from the Zadoks catalogue misses the right arm and left hand and the facial features have not been indicated. The other specimen misses the plume and both hands. The object on the head represents either a crown with ostrich feathers, or a modius, a headdress in the shape of a grain measure. Both are attributes of the Egyptian goddess Isis. Statuettes with these attributes are known from the study area.869 The crown of the Isis from Houwerzijl is simplified rather than stylised. Since Venus and nymphs are the only divine beings of the Roman pantheon who are depicted nude or semi-nude, Isis is associated here with Aphrodite/Venus. 870 In 1995, Marjan Galestin published an article on the two figurines. Based on the round flat plinth and the simplified crown, she advanced that the statuettes are lids of a vessel or 19th-century ornaments belonging to a clock or another object.871 Round flat plinths carrying statuettes do not occur in the study area. All flat plinths are rectangular or have an indeterminate shape. Moreover, these figurines and their plinth were cast in one piece, contrary to most bronze statuettes, which stand on separately cast or wooden pedestals. Therefore, it is possible, as Galestin has suggested, that the two figurines were lids of vessels and constituted a set. It must be said, though, that if the 70 mm high figurines were handles of a lid, they are disproportionately long in comparison to the small round plinth, which has a diameter of 30 mm. Galestin’s second argument for her interpretation, the stylised or simplified rendering of the crown, is not convincing. Many figurines from Roman contexts in and outside the study area present stylised or simplified characteristics. As to a 19th-century date for the figurines, there are two arguments that speak against it. First, the pitted surface, corrosion and patina. The corrosion continues on the break lines and surface, indicating that the hand and arm were already missing before the figurine ended up in the ground. If the figurines were made in the 19th century, they had been in the ground for only half a century when they were discovered, in which case the surface probably would have looked less worn. Second, the treatment of the body, the cursorily rendered details and unevenly worked surface of the figurines have little in common with the smooth, elegant ornaments of 19th-century clocks and other objects. The association of Isis with Aphrodite/Venus is unique in the study area and predominantly known from Egypt, Greece and the Near East. Whether the two Isis-Venus figurines were made there cannot be established without technical analysis.

866

Zadoks et al. 1967, 90.

867

Zadoks et al. 1967, nr. 10.

868

The statuette is inaccessible.

869

Zadoks et al. 1969, nrs. 26, 29. See for the Isis-Fortuna wear-

ing a modius Zadoks et al. 1967, nr. 11. 870

See on statues of Isis-Aphrodite/Venus, for instance, Heyob 1975, 49-50.

871

Galestin 1995, 91-98.

111

3.5.3 Jupiter-Sabazius? From Menaldum, Friesland, comes a bronze statuette of a man wearing a mantle and round cap, holding a thunderbolt in his right hand (figs. 2.10-11). The statuette has been interpreted as Jupiter-Sabazius: an identification based on the assumption that the cone shaped hat is a Phrygian cap.872 The god is of Phrygian or Thracian origin and already in early Greek sources identified with Dionysus. Written sources link Sabazius to mystery cults and ‘foreign’ gods like Attis, Corybas (the son of Cybele) and Mithras. In the Roman period, the god is predominantly associated with Jupiter.873 Sabazius in anthropomorphic guise is seldom depicted in sculpture and other media. This deity has no distinct iconography other than that of an older, bearded man with thick, curly hair. His cult is best known through bronze hands decorated with symbols, in western Europe attested by specimens from Belgium.874 The statuette from Menaldum seems to be a unique representation, since comparable examples in sculpture are unknown in the northwestern provinces. This brings up the question whether the Jupiter from Menaldum really depicts Jupiter associated with Sabazius, all the more since his cap does not resemble a Phrygian cap. The latter has a pointed crown curling down and is worn by, for instance, Jupiter Dolichenus.875 Vulcanus is often depicted wearing a pilos, a cone shaped, brimless hat, but we can exclude the possibility that the smith-god is depicted here, because of the thunderbolt. Early images of Jupiter wearing a round headdress and a mantle date to the Etruscan period. The name Tinia associated with Jupiter is attested by an Etruscan inscription on the Piacenza liver. Written sources from the Roman period refer to Tinia/Jupiter throwing the thunderbolt.876 Because it is not very likely that an Etruscan statuette ended up in Friesland, I propose another solution to the problem. In the Roman period, Jupiter is often depicted wearing a wreath, in sculpture, coins and gems.877 A large bronze statuette (540 mm in height) of Jupiter was allegedly found at Bree (Belgian Limburg), but the find circumstances are suspect.878 The god has thick curly hair, which on the top of his head is held together by what Faider-Feytmans describes as un cordon torsadé, intertwined rope.879 There are several statues and statuettes of Jupiter with a comparable hairstyle (fig. 3.3). It is possible that the maker of the statuette from Menaldum has seen statues or coins depicting Jupiter with this kind of hairstyle or wreath, and misinterpreted it as a cap or hat. Another possibility is that he combined characteristics of different deities, in this case Jupiter and Vulcanus. 3.5.4 A bronze mother goddess north of the Rhine? Whereas terracotta statuettes of mother goddesses occur frequently in the northwestern provinces, bronze specimens are very scarce and attested by only three examples in the study area, one found north and two found south of the Rhine. The first figurine came to light at Rottum (Groningen), together with a bronze statuette of a male figure in a wierde (figs. 2.18-20).880 The statuette measures 55 mm. South of the Rhine, two bronze examples have been recovered from the Hunerberg at Nijmegen (Gelderland) and Koudekerk 872

Zadoks et al. 1967, nr. 15.

873

Gicheva 1997, 1069-1071.

875

Bauchhenß 1997, 472-478.

874

Faider-Feytmans 1979, nrs. 59-60. Nr. 59 was found near

876

Camporeale 1997, 400-421.

Rumst (Antwerp). The find-spot of nr. 60 is unknown. The

877

Camporeale 1997, 432-435 and nrs. 109, 114, 134-149.

bronze hand found near Tournai (Doornik, Henegouwen)

878

The person who sold the statuette possibly lied about the

2004, 2-3; for Limburg Verhoeven/Ector 1984, 138-149.

find-spot, since very few traces of Roman period occupa-

is, at least partly, a forgery: Faider-Feytmans 1979, 212. The

tion have been attested at Bree: Faider-Feytmans 1979, 49.

National Museum of Antiquities at Leiden has a specimen of unknown provenance. In the study area, miniature tools

879

Faider-Feytmans 1979, 49-50 and nr. 1.

and symbols from graves in Nijmegen and Limburg have

880

See also Zadoks et al. 1967, nr. 24.

been linked to the Sabazius-cult. See for Nijmegen Steures

112

aan den Rijn ( South Holland) (figs. 2.55-56a). The latter two examples are almost identical and measure c. 44 mm.881 All three statuettes depict an enthroned woman. They have the same posture, with the head inclined slightly forwards. However, there are distinct differences. The mother goddesses from Nijmegen and Koudekerk wear a high headdress with a convex, vertically ribbed border. The headdress covers the hair and ears. Both wear a traditional, long Roman mantle, a stola, with a palla, a garment comparable to a broad scarf which covers the shoulders and arms. This is the dress of a Roman matron. They hold a basket filled with fruits on their knees. Outside the study area, four bronze figurines of seated women with fruits or flowers in their lap do not wear a headdress, but probably still represent mother goddesses. Three come from Reims, Bavay and Pont-de-Beauvoisin (northern France). Two specimens also hold a patera. The fourth specimen comes from Trier and strongly resembles the figurines from Nijmegen and Koudekerk and nearly has the same height, 40 mm.882 Terracotta statuettes of mother goddesses in the study area are usually depicted as single, enthroned women with fruits or a dog in their lap, or women suckling one or two infants. Rare variants depict a standing woman supporting a child with one arm and holding a jar in the other, or a standing woman holding a basket with an infant in front of her body. They often wear large, round bonnets or other headdresses, a diadem, or they have the hair dressed high above the forehead.883 The enthroned woman from Rottum is depicted without any object, infant or dog. Unlike other mother goddess figurines, she does not wear a headdress, diadem or the modest Roman ‘outfit for matrons’, but a long garment and a mantle that is wrapped around her body, leaving the right shoulder and both arms bare. The ends of the mantle are gathered into a knot on the left shoulder. Both hands are resting on the arms of the chair. The most striking characteristic is her long hair, which she wears in an elaborate hairstyle with three braids. Two braids frame her forehead and hang down over the shoulders. The third braid falls down from the back of her head. A comparable braided hairstyle is known from an Iron Age bog body known as Elling woman, who was found in the Bjeldskovdal bog, ten kilometres west of Silkeborg, Denmark, where she lived between 350 and 100 BCE.884 Considering the resemblance to the two bronze mother goddesses, it is unlikely that the female figurine dates to the Iron Age. The hairstyle and clothing suggest that a native woman is depicted. She may represent a local version of a mother goddess or a native goddess. The male figure that was found together with this female figurine has been interpreted as Mercury. As had been suggested in 3.5.1, this is uncertain, although it is an attractive idea that the two figurines were buried together because they were seen as a couple, like Mercury and the Celtic goddess Rosmerta.885 3.5.5

Mounted warriors

A bronze statuette of a mounted warrior was found by a detectorist in 2004 in a field near De Knipe, Heerenveen, Friesland (fig. 2.3).The statuette measures 70 mm in height and depicts a mounted warrior holding a (missing) lance in his right hand and, presumably, a (missing) shield in the left.886 Mounted warriors exist 881

See for the statuette from Nijmegen, Zadoks et al. 1969, nr. 38.

882

De Ridder 1913, nr. 1078; Faider-Feytmans 1948, 392-393 and fig. 3; Babelon/Blanchet, 1895, nr. 93; Menzel 1966, 40 and

884

Silkeborg Museum Denmark. See also

885

Figs. 2.19-20 and 2.2.2. The two deities are mentioned

http://www.tollundman.dk/Ellingkvinden.asp (3-12-2016).

Taf. 40, nr. 85. It is not clear whether the figurine from Trier

together in inscriptions (CIL XIII, 4311, 4705, 4683) from

wears a hoodie that reaches halfway down her back, or

north-east France, and are depicted as a couple in reliefs from Autun, Strasbourg, Bath and Gloucester.

that she has long hair: Veen 2019, 155-161. See for a bronze mother goddess from Cologne, 2.3.2. 883

886

Galestin 2006, 74-79.

Van Boekel 1987, 99-123. See also 3.8.8 for the iconography of mother goddesses.

113

in a wide variety of representations of which several depict warriors on horseback. They have been predominantly found in Britain and France, but comparable examples have also been recovered from Iron Age find-spots in northern Italy.887 Among the wide variety in representation of mounted warriors, the stylised Romano-Celtic type is well-represented.888 The statuette from Heerenveen bears little resemblance to these stylised bronze warrior-horsemen statuettes. Yet, it clearly represents an attacking warrior on a horse. So far, this is the first complete specimen found in the study area.889 Several single horse statuettes and ‘loose’ horsemen from Britain originally also depicted mounted warriors, but horse and rider became separated because they had been cast separately. This may also apply to two statuettes from the terp region. Two of the four bronze horses from the northern coastal area most likely carried a rider.890 The other two do not bear such indications. Starting with the last two, one specimen represents an ancestor of the magnificent Friesian horse.891 The second was found at Kantens (Groningen), west of the river Ems. It depicts a horse with big, bulging eyes, a thin tail and two legs which constitute a broad support surface (fig. 2.17).892 The first of the other two specimens came to light at Makkum, Friesland, and represents a rearing horse (figs. 2.8-9). It measures 144 mm in height and 166 mm in length. The statuette was cast in two pieces and has two depressions on the back, indicating that the horse carried a rider that covered the casting joint. A third depression, in the middle of the belly, suggests a vanquished enemy underneath the horse.893 A soldier riding down an enemy is particularly known from early Imperial tombstones, like those of Titus Flavius Bassus and Caius Romanius Capito, both of the ala Noricorum, and the tombstone of Longinus Sdapeze of the ala prima Thracum.894 The second figurine is much smaller and came to light at Holwerd, Friesland. It measures 49 mm in height and 53 mm in length and represents a stepping horse (fig. 2.2).895 Among the three horse figurines that have been found in Friesland, this representation stands out. The two other specimens depict horses true to nature, but the figurine from Holwerd is strongly stylised. It is depicted with three legs on the ground, the right foreleg raised. It has a thin body, a flat neck and a short head with a topknot between the ears. The figurine is a stray find, without any contextual data, but has been incorporated in the 1967 catalogue by Zadoks et al., which implies that the authors date the figurine to the Roman period. However, according to the online collection database of the National Museum of Antiquities at Leiden the horse dates to the Medieval period and has been interpreted as a children’s toy. This is based on finds like a terracotta horse from a wierde in Groningen, dating between the 12th and 15th century, and a 12th-century terracotta stick horse, found during excavations in the Pfalz Werla, in Lower Saxony, Germany.896 Since the horse figurine from Holwerd has no hole in its body for a stick, it is not a stick horse. Moreover, late Iron Age or early Roman period figurines from England and the Danube region bear a stronger resemblance to the horse from Holwerd than the Medieval terracotta horses from Groningen and Werla. Unfortunately, the majority do not come from controlled excavations. Some have been found by detectorists and are for sale on the internet (fig. 3.4).897 The owners/auctioneers date the figurines to the 1st century BCE or early Roman period, but these dates may not be reliable. These horses and the horse from Holwerd share several similarities, such as their thin bodies, a raised tail and a topknot between the ears. They are depicted stepping, which is suggested by one raised foreleg and three legs standing on the ground. Usually they carry a warrior, often cast separately. It is, therefore, not unlikely 887

888

889

Zampieri/Lavarone 2000, 62-66 nrs. 33-53 and tav. X-XI.

892

Groninger Museum, inv. nr. 1991-X5.

These figurines are dated between the 5th and 3rd century

893

Zadoks et al. 1967, 118.

BCE.

894

The first was found in Cologne, the second in Mainz, the

See, for instance, Toynbee 1964, 119 and plate XXXI, c-d, 895

Zadoks et al. 1967, nr. 51.

See for terracotta specimens from Belgium De Beenhouwer

896

Willemsen 1998, 71-72 and fig. 81, 114-115 and fig. 86.

2005, nrs. 524, 527, 835.

897

http://www.ancientresource.com/lots/celtic-german-

890

Zadoks et al. 1967, nrs. 49 and 51.

891

Zadoks et al. 1967, nr. 50.

114

third in Colchester.

Pitts 1979, nrs. 136-141, Green 1992c, 123.

ic-frankish/celtic-horsemen-warriors.html (16-12-2016).

that the bronze horse figurine from Holwerd once carried a rider, too. Based on the strong resemblance to the figurines from Britain and the Danube region a date for the Holwerd figurine in the late Iron Age or early Roman period seems more plausible than a late Medieval date. Another question that merits attention is who these mounted warriors represent. It has been argued that they are provincial versions of Mars on a horse.898 Although the Roman god Mars is never depicted on horseback in the heart of the Empire, a bronze statuette found at Martlesham (Suffolk) originally depicted a figure riding down an enemy. According to the accompanying inscription, the statuette was dedicated to Mars Corotiacus by a woman called Simplicia.899 There is ample evidence of other deities on horseback. A distinct group of comparable representations, a deity who is riding down enemies or monsters, is depicted on Jupiter columns, all but one from Gaul: Quémigny-sur-Seine (Côte-d’Or), Luxeuil (Haute-Saône), Meaux (Seine-etMarne), Neschers (Puy-de-Dôme) and Obernberg (Bayern).900 At Sainte-Sabine in Burgundy, clay figurines of horsemen were offered at a shrine dedicated to Apollo Belenus, Apollo associated with the Celtic Sun-god Belenus.901 In Britain, bronze horseman figurines were found in a Roman-period temple complex at Brigstock (Northants).902 It is unknown to which deity the temple was dedicated, but apparently there was a horseman cult. The temple was built in the 3rd century and remained in use until the late 4th century.903 A type of enamelled brooch, known as ‘the horse and rider’ type, dates to the 3rd century. It depicts a figure on horseback. Such brooches have been mostly recovered from rural sites and small towns in East Anglia and the south-west (Dorset, Wiltshire, Somerset), often as part of a ritual deposit.904 Further evidence from Britain and France demonstrates that a Celtic horseman warrior cult existed already in prehistory. Mounted male and female warriors appear on pre-Roman Celtic coins and as representations in sanctuaries.905 In a sanctuary at Mouriès (Bouches-du-Rhône), horsemen were incised in stone. From a sanctuary at Entremont (Haute-Savoie), where a shrine stood with severed heads, comes a relief with a horseman carrying a severed head.906 The horse from Holwerd was probably made in Britain, where this type of figurines may have been linked to a horseman warrior cult. Whether the (last) owners of the three statuettes under discussion were adherents to such a cult is doubtful, since evidence of a horseman warrior cult in the study area is lacking. It seems more plausible that they were soldiers or returning veterans who liked the statuettes because they reminded them of their time in the Roman army. As to the statuette from Heerenveen, we have to bear in mind that the find-spot is situated outside the terp region, in an area of peat land, where no traces of Roman period settlements have been found. The statuette may have ended up there when farmland was fertilised with soil from the terp region, or, like many other objects, it was buried in the peat bog as a ritual deposit. 3.5.6

The warrior from Ezinge

One of the three statuettes from Ezinge represents a soldier in the Roman army (fig. 2.14). The statuette was discovered with a metal detector in 1989, in a section of the wierde which had been quarried for soil. The bronze statuette measures 95 mm in height and represents a standing bearded man. His posture is not martial.907 He reaches sideward with a slightly outstretched right arm. The right hand has broken off. The 898

899

Toynbee 1964, 119, 153-154, Henig 1995, 51, Galestin 2006,

901

Green 1992c, 123.

74-79. See also Durham 2012, 3.12, nrs. 162, 821.

902

Green 1992c, 122-123.

Deo Marti Corotiaco Simplicia pro se (votum) p(osuit)

903

http://www.british-history.ac.uk/rchme/northants/vol1/

l(ibens) m(erito): RIB 213. The maker of the bronze stat-

pp20-24 (2-1-2017).

uette, a certain Glaucus, has signed his work, which is

904

Mattingly 2006, 486-487.

exceptional. The horse and rider have broken off, except

905

See for a coin with a mounted female warrior for instance, Aldhouse-Green 2004, 27, fig. 1.12.

for one hoof. Only the base has been preserved, with the 900

enemy lying on his back.

906

Green 1992c, 122.

Green 1992c, 128, 192-193.

907

Galestin 1990, 146-152.

115

lower left arm is severely corroded and hangs down. The man wears a scale armour (lorica squamata) with overlapping scales from iron or bronze which were sewn onto a leather or fabric shirt, and a belt around his waist. Under his armour he wears a tunic with short sleeves. He has short, thin legs and his feet are not visible under the layer of corrosion. Galestin dates the statuette to the 4th century, based on the scale armour. Style is only a valid dating instrument if comparable specimens are known from dateable contexts, which is not the case. The date she proposes is based on Robinson’s description of a late version of the scale armour with ‘half-length sleeves and a long skirt terminating just above the knee’.908 For this type Robinson refers to 4th-century reliefs on the arches of Galerius in Thessaloniki, Greece (fig. 3.5), and Constantine in Rome. Despite the resemblance, the scale armour of our soldier from Ezinge reaches to the upper thigh and is clearly shorter than the scale armour of the soldier in the relief on Galerius’ arch. Moreover, Robinson’s description of the scale armour on a 1st-century grave stone at Colchester seems also applicable to the scale armour of our soldier from Ezinge: ‘(…) one can see a simple short-sleeved, hip-length garment that probably remained in fashion until the 3rd century with little or no change (…).’909 Robinson refers here to the tombstone of the duplicarius Longinus Sdapeze from the ala prima Thracum. The scale armour of our soldier is a bit longer than hip-length, but also resembles the one depicted in the 1st-century tombstone. Apparently, the simple, short-sleeved scale armour with slightly varying lengths was in use for over three centuries, which makes it an unreliable dating instrument. There is no doubt that the statuette represents a soldier. Whether a cavalry or infantry man is depicted cannot be deduced from his scale armour, which was a piece of equipment used by all kinds of army units.910 Nevertheless, his posture seems to hint at a specific type of representation. Galestin mentions a bronze statuette of a cavalryman who holds a horse by the reins.911 The statuette was found in the Saalburg castellum and possibly dates to the 2nd century.912 Several reliefs depict dismounted soldiers holding their horses by the reins: on the columns of Trajan and Marcus Aurelius, on sarcophagi and tombstones, the already mentioned arch of Galerius, and a grave marker from Nijmegen.913 The men on Galerius’ arch are listening to an address from the emperor.914 The non-martial posture of the soldier from Ezinge is comparable to the Saalburg statuette and the men in these reliefs, suggesting that, originally, the soldier from Ezinge may also have held a horse by the reins. The representations date from the late 1st into the 4th century and do not offer any clues as to the production or deposition date of the statuette from Ezinge. The three statuettes from Ezinge, the soldier, the Jupiter statuette and the bronze cockerel, come from the same site and are stray finds. We are in the dark about the provenance and production or deposition date of the first two, but the bronze cockerel provides a clue to its provenance and a rough terminus post quem for its deposition date. Comparable specimens have been found at Buchten, Limburg (one), Tongeren, Belgium (one), Cologne, Germany (one) and Britain (six). The distribution pattern and the manufacturing method suggest that they were made in the same workshop, which most likely stood in Britain. The production date of the bronze cockerels lies between the late 2nd and early 3rd century.915 This implies that the cockerel ended up at Ezinge around the beginning of the 3rd century or later. It is possible that one of the inhabitants of the settlement purchased the figurine in Britain.

908

Robinson 1975, 161 and fig. 123.

912

Robinson 1975, 93 and fig. 122.

909

Robinson 1975, 159.

913

Daniels/Brunsting 1955, plate 5, nr. 11.

910

Nicolay 2007, 20.

914

Robinson 1975, 93.

Galestin 1990, 150. A terracotta statuette of a cavalryman

915

Hoss et al. 2015, 159-171.

911

standing beside his horse is known from Belgium: De Beenhouwer 2005, nr. 523. See also Von Gonzenbach 1995, 212, 215.

116

3.5.7

A male figurine from Wirdum

A small terracotta statuette came to light during the quarrying of a wierde at Wirdum, Groningen, at a depth of five metres (fig. 2.25). At the same level, Roman period pottery was found, as well as a chalk figurine of a bear, which will be discussed below. The terracotta figurine is 47 mm in height and depicts a standing male figure. The right foot is missing, the left sticks out from under the tree-trunk shaped body. It has an excessively large head, short hair, two tiny stumps for arms and no legs. The eyes are two circular holes, made by piercing a round stick in the clay. The nose and mouth are clearly indicated, the latter by a broad horizontal incision. It is assumed that the figurine dates to the Roman period, based on the pottery that was found at the same level.916 Since it is a fired clay figurine, it was meant to last for a longer period. A figurine that bears a faint resemblance to the small tree-trunk man came to light in North Brabant. It is 125 mm in height and has been carved from oak wood (fig. 3.6). The wooden object was discovered during the construction of the Volkerak locks near Willemstad in 1968 and was found between the roots of an oak tree, in a peat layer. The wood could be dated to 5300 BCE.917 The figurine is probably much older than the specimen from Wirdum, but both have a large head, a tree-trunk shaped body, no legs and two little bulges where the arms can be expected. Considering the simplified physical details, it is possible that such figurines were children’s toys, simply made for fun, but another function must be taken into consideration as well. First, both figurines bare some resemblance to a tree. The veneration of the tree hails back to old belief systems. In many cults, the tree symbolised the circle of life, since a tree is ‘dead’ during the winter and comes to live in the spring. Trees were also seen as the link between the upper and lower world.918 According to Pliny the Elder trees were once the temples of deities.919 Second, the emphasis on the head is particularly known from sculpture in stone and bronze of figures with over-large heads, or consisting of heads alone. Such heads have been found in pre-Roman and Roman period contexts in central and northwestern Europe. The heads were thought to be the most important part of a living creature and often represent deities or divine beings.920 Therefore, the figurine from Wirdum could also represent a supernatural power. 3.5.8

Domestic animals

Terracotta animal figurines have predominantly been found in the terp region. They mostly represent domestic animals, farm animals in particular. The other statuettes comprise a bronze eagle from the terp region, a bronze triple-horned bull from Beilen (Drenthe), and two bronze boars from Noordbarge (Drenthe) and Heeten (Overijssel). The most remarkable representation is the bear from Wirdum (see below). Although the period extends beyond the scope of this study, it is worthwhile mentioning a Bronze Age assemblage of clay animals found at Grootebroek (North Holland) in the filling of a Bronze Age ditch, together with pottery fragments, pieces of flint, stone and bone fragments. The assemblage comprised twenty items, including, two complete animals, a fragmented specimen and fifteen small fragments. They depict cattle, pig and one unidentifiable animal.921 The figurines were modelled from clay, tempered with grog, fine sand and stone grit. Only the latter was not locally available and must have come from Pleistocene soils.922 The figurines had been fired in oxidizing conditions. Since the break surfaces of the fragments were also oxi916

Knol et al. 2005, 196. Although a pre-Roman or early Medi-

919

Plin. HN 12.2.

eval date cannot be ruled out: additional information by

920

Green 1992c, 114-115.

Egge Knol (Groninger Museum).

921

Van Regteren-Altena et al. 1977, 241-242.

917

Verhart 1993, 24-26.

922

Van Regteren-Altena et al. 1977, 245.

918

Green 1992c, 212-213.

117

dized at the core, the figurines must have been broken before or during firing. The function of the figurines and the reason for breakage are not clear. They may have fallen into the fire by accident, but the firing of these figurines could also have had a special meaning. If the breakage was intentional, the figurines may be the remnants of a ritual. This is supported by the find-context of the broken animals, a ditch. Another example of a late Bronze Age deposit of clay figurines comes from Rullstorf, Niedersachsen, Germany. The figurines were recovered from a pit, together with miniature vessels and small wheels. One of the figurines depicts a pig.923 The vessels and wheels in the pit suggests that the deposit had a ritual meaning.924 The four terracotta figurines from Sneek represent a bovine or sheep, a chicken, an egg and, presumably, a fragment of a small wheel. If it is a wheel, it may have been part of a toy, such as a horse on wheels or a chariot, but it is not a perfectly round wheel. It could also have had a symbolic or apotropaic function.925 The egg, too, could have had a special meaning, symbolising the beginning of new life, but it may also be a fake egg, which was, and still is used, to stimulate hens to lay eggs.926 The figurines ended up in the ground in the 1st century. Whether they were also made in the Roman period is uncertain. According to the excavators, the chicken/hen figurine must have been made in the Roman period, which is based on the assumption that the Romans introduced the chicken to our region. Others argue that the chicken was introduced in the late Iron Age.927 Whether the function of the figurines is profane or symbolic cannot be determined. They could have been toys, like the terracotta birds that will be discussed hereafter. It has also been advanced that figurines of bovines had a symbolic or ritual function related to the importance of dung for fertilisation.928 According to the excavators, the figurines from Sneek were locally made. This could also apply to three clay bird figurines from the northern coastal area. Two come from the northern terp region, Oosterbeintum and Enumerhoogte, the third was found during excavations at Schagen-Muggenburg (figs. 2.12-13, 2.26-27). 929 Although the three figurines are not identical, they share one remarkable feature: all three have a circular hole on both sides of the body under the neck. The holes were meant for inserting a metal pin or wooden peg. Apparently the head of the bird rested on the peg/pin to make it wiggle. The heads were removable, which explains why the heads from Schagen and Enumerhoogte are missing. Only the head of the bird from Oosterbeintum has been preserved.930 Boeles interpreted the bird as a toy, a rattle.931 It is not clear why he thought so. Maybe at the time, the bird still held pebbles or fired clay balls in its hollow body. It is also possible that he based his assumption on statuettes that contained pebbles, such as the terracotta duck and the so-called horned birds from Roman period burial sites at Nijmegen.932 The three birds from the northern provinces were probably children’s toys. They seem to have been a regional phenomenon, since birds with movable heads are not known from other parts in the study area. Yet, animal figurines with movable body parts were not unique, as is demonstrated by two Roman period wooden mice from Egypt. As the animal was pulled forward, the tail and lower jaw moved up and down.933 A much earlier example from Egypt comes from a tomb in Thebes, dating to the New Kingdom (1550-1070 BCE): a wooden toy cat with movable jaw (fig. 3.7).934 From the Roman period, dolls with movable limbs are known from Roman catacombs, Ostia and other places in Italy, Spain and Switzerland.935

923

Gebers 1995, 59, Ufkes 2002, 52.

928

IJzereef 1981, 22-23.

924

As far as I know, the pit and the objects have not been

929

See for the three birds also 2.2.2 and 2.2.3.

published.

930

The head of the bird from Enumerhoogte is a modern copy

925

See for the meaning of wheels 2.4.1: Leiden-Roomburg-the

of the head of the bird from Oosterbeintum.

vicus.

931

926

Niekus/Huisman 2002, 50-51.

932

927

Niekus/Huisman 2002, 48-50. According to Groot, chickens

118

Boeles 1951, 537 and plate XXXVI, nr. 5. Van Boekel 1987, 738, nrs. 229, 238-241. They were found in or before 1950.

may have been introduced in the late Iron Age, but little

933

Willemsen 2003, 122-123 and fig. 78.

is known about the introduction and distribution: Groot

934

British Museum, inv. nr. EA15671.

2008, 197.

935

Willemsen 2003, 82-90 and figs. 57-59.

3.5.9

The bear from Wirdum

Besides the tree-trunk figurine, the site Wirdum has yielded a figurine of a bear (figs. 2.21-23). It measures 130 mm in length and 79 mm in height. The lower legs are missing. The round eyes are indicated by incisions and a round piercing in the centre. The function of two tiny round piercings in its belly is uncertain (fig. 2.24). They may have held metal pins to attach the figurine to a base, but the diameter of the holes is less than a millimetre. In 1968, Van Giffen had the figurine examined. An X-ray-interference test revealed that the figurine was carved from chalk. Chalk is absent in the study area, but occurs in Denmark, northern Germany and Greenland. The cliffs of Dover (Britain) and the coast of Normandy (France) are also composed of chalk. The bear is the only chalk figurine from the study area, but Iron Age and Roman period chalk figurines are known from Britain.936 Why the maker chose chalk for his creation is an intriguing question. It is possible that he wanted to depict a polar bear (Ursus maritimus) or that he liked the material because it is soft and relatively easy to work. Considering the absence of polar bears in Europe, it seems more plausible that a European brown bear (Ursus arctos arctos) is depicted. After all, the only obvious exterior difference between a polar bear and a brown bear is the colour of its fur.937 The European brown bear was present in the study area until the 11th century.938 In the Roman period, the presence of brown bears in Germania Inferior is attested by a statue of a man with a bear, from Xanten. The statue base bears an inscription to the god Silvanus. The dedicator is an ursarius, a bear catcher, in the Legio XXX Ulpia Victrix Severiana Alexandriana.939 Bears are were associated with the goddess Artio, a fertility goddess whose cult is attested by a bronze statuette group with an inscription from Muri, Switzerland. It depicts the seated goddess with a large bear and fruit in her lap (fig. 3.10).940 Two terracotta bear figurines have been found in the southern Netherlands (fig. 3.11). A third bear is an amber figurine from a cremation grave in Nijmegen.941 If the bear from Wirdum depicts a polar bear, its presence and function are difficult to explain. The posture of the figurine demonstrates that the maker was skilled and familiar with bears, because it depicts a hunting bear, ready to grab its prey from the water, or from a breathing hole in the ice. Carved from stone or ivory from whale and walrus tusks, figurines of polar bears have been predominantly found within the Arctic Circle. The bear from Wirdum comes from the same site as the small tree trunk figurine and may have ended up there in the Roman period, but even this is uncertain.942 The question whether the figurine depicts a brown bear or a polar bear remains unsolved, just like its function. The bear could be linked to a cult, but the playful figurine may also have been a children’s toy or a decorative object, as is suggested by the two piercings in its belly.

936

See, for instance, Aldhouse-Green 2012, 17, 69 and figs. 3.8,

ed to swimming. http://www.nsrl.ttu.edu/about/Outreach/

4.8-4.10. 937

Exhibits/Bear%20Exhibit.pdf (16-12-2016).

Polar bears and brown bears are genetically closely related.

938

Buiks 1997, 72.

It is thought that polar bears descend from a group of

939

Bogaers 1984, 2-4, CIL 13, 8639.

brown bears that became isolated by glaciers near Siberia

940

Deyts 1992, 48; Green 1992, 139 and fig. 10; Green 1992c, 188189.

between 100,000-150,000 years ago and underwent a series of rapid evolutionary changes in order to survive in the

941

became sharper and they developed larger feet with short-

See for the terracotta bear figurines 2.3.5, for the amber bear 2.3.4 and Koster 2013, 179.

Arctic Circle. The colour of their fur changed, their teeth 942

The history of Dutch whaling in the Arctic Ocean spans

er, curved, non-retractable claws for a better grip when

a period of centuries, beginning in the late 16th century.

running on ice: There are also smaller changes like the

A crew member of one of the whaling ships may have

shape of the skull, which became flatter and better adapt-

brought the figure to the northern coastal area.

119

3.6

Conclusions on the area north of the Rhine

The distribution map presents two distinct clusters of find-spots: the northwestern terp region in Friesland, the territory of the Frisii, and the north-eastern terp region in Groningen, which probably belonged to the territory of the Chauci. The other find-spots are situated in North Holland, Drenthe, the eastern part of Overijssel and northern Gelderland on the right bank of the Rhine. The bronze statuettes from the northern terp region may have been brought to the north by Germanic auxiliaries or veterans, as a religious object, souvenir or curiosum, or by Romans as a means of payment or diplomatic gift. These presumed functions as well as the absence of structured trade between Romans and peoples in the north could explain the absence of terracotta figurines, as well as the larger size and better execution of the bronze statuettes. The average size of statuettes is distinctly larger than that of statuettes found south of the Rhine. This is probably related to their function. They were larger because they were diplomatic gifts, souvenirs or a means of payment when they were brought to the north. However, although well-executed at first sight, the bronze statuettes from the northern terp region present imperfections and casting flaws. It is suggested that Romans picked these specimens on purpose, assuming that the recipients would not notice the flaws. Another explanation could be that they appreciated the statuettes anyway and took the imperfections for granted. Fewer statuettes come from the eastern part of the terp region. Here, more figurines have come to light with an iconography that deviates from the Roman repertoire, such as the horse from Kantens, the terracotta tree-trunk figure and the bear from Wirdum, the so-called mother goddess and Mercury from Rottum. Elements of the iconography of the horse and the figurines from Wirdum and Rottum seem to hark back to local or pre-Roman traditions, but this is difficult to substantiate, due to the lack of reference material. The analysis has also revealed that north of the Rhine, representations of Mercury and Mars outnumber all other deities, particularly in the terp region. With 18 specimens, Mercury is the best represented deity, which probably can be explained by his qualities as a bringer of prosperity and financial gain. Mars is represented by nine bronze specimens. The occurrence of Mars in rural regions could be linked with his original capacities as a bringer of agricultural fertility, as has been advanced.943 They may also have appreciated the martial qualities and appearance of Mars. If the statuettes were diplomatic gifts for tribal leaders or a means of payment, Romans and traders must have assumed or known that the Frisii and Chauci preferred these two deities over others. Hoards of denarii in the northern Netherlands as well as inscriptions outside the study area attest to the service of men from northern tribes in the Roman army. A group of statuettes, too, suggests a connection between the inhabitants of the terp region and life in the military: the mounted warrior from Heerenveen, the Roman period soldier and the bronze cockerel from Ezinge, the rearing horse from Makkum that originally depicted a mounted warrior riding down an enemy, and the small horse from Holwerd, which probably once carried a warrior too. The presence of animal figurines in rural regions is not surprising and related to people’s dependence on domestic animals for food, hides and wool. Yet, the function of these statuettes is not always clear and may have changed in the course of the life path of figurines. This appears from the Bronze Age assemblages at Grootebroek (North Holland) and Rullstorf (Germany). Initially, the animal figurines from Grootebroek may have been children’s toys, but they were deliberately broken before they ended up in a fire, possibly as part of a ritual. The figurines from Rullstorf were most likely remnants of a Bronze Age ritual, since they were buried together with miniature vessels and wheels.

943

Simon 1984, 505-506. See for the capacity of Mars as a purveyor of agricultural fertility Roymans 1995, 55-58; Derks 1998, 95-105. See also 2.3.1.

120

The three terracotta birds with movable heads from Schagen, Oosterbeintum and Enumerhoogte most likely were children’s toys. The bird from Schagen ended up in a pit as part of a ritual deposit under the entrance of a building.

3.7

The area south of the Rhine: spatial analysis

3.7.1

The rural settlements

The 132 statuettes, complete and incomplete, that have been recovered from rural settlements and their surroundings include 80 terracotta (61%), and 52 bronze (39 %) specimens, of which 104 could be identified. Fifty-seven figurines depict a deity, 24 a human figure and 23 represent animals (charts 3.8-3.10). 944 Mercury prevails with ten specimens, followed by Mars (six) and Hercules (five). Among the female figures, mother goddesses and Venus are best represented by nine and six specimens respectively. The large number of statuettes in the eastern river area is a consequence of the proximity of military settlements and of the population density in the region. At least twenty specimens come from rural settlements around Nijmegen. Striking is the relatively higher proportion of bronze specimens in comparison to military and urban settlements and their cemeteries (see below). The villas themselves have yielded only a few statuettes, predominantly terracotta specimens. It is often argued that terracotta statuettes were used by poor people who could not afford bronze statuettes. Yet, the finds in villas in the study area demonstrate that they occur in wealthier contexts, too. In Britain, 13 terracotta statuettes were found in a rich grave in Colchester.945 Striking are the three Risus busts from the villas at Kerkrade-Holzkuil, Kerkrade-Kaalheide and Maasbracht-Steenakker. Unfortunately, two of the three busts lack archaeological data. Two terracotta bear-figurines (figs. 3.11-12) come from Kerkrade, one from a presumed villa at Kerkrade-Maar-West and one from Kerkrade-Kaalheide.946 Also remarkable are the three terracotta heads of female deities from Tiel-Passewaaij (Gelderland) and the terracotta head of a mother goddess from Son en Breughel (North Brabant), which were all found without any other belonging fragments. The terracotta head from Son en Breughel as well as a bronze Mars statuette from Helden (Limburg) were recovered from a sunken byre. Three statuettes come from sanctuaries or alleged cult places. At Bergen-op-Zoom (Zeeland) the fragmented terracotta statuette of Sucellus was recovered from a site which in antiquity was a fen. Here, people offered miniature vessels, figurines, coins and so on. The bronze Minerva from the villa at Wijchen-Tienakker was buried together with a small altar in a pit within a structure that has been interpreted as an open air sanctuary. The bronze cockerel from Buchten (Limburg) was also buried, together with coins and jewellery, in a pit just outside a sanctuary dedicated to the native goddess Arcanua. In the category ‘find-spots from unknown archaeological contexts’ several statuettes have been found on or in river banks: the terracotta head of Minerva on the bank of the Westerscheldt at Ellewoutsdijk (Zeeland), the head of a terracotta mother goddess on the bank of the Meuse at Heerewaarden-Lith (North Brabant), a bronze Mercury from the bank of the Dommel at St-Oedenrode (North Brabant), the bronze mother goddess from the left bank of the Rhine at Koudekerk aan den Rijn and a bronze Mercury from the left bank of the Meuse at Kessel-De Donk (Limburg).

944

See also table 3.

gifts it has been assumed that the grave is a child’s grave,

945

The figurines were deposited together with a decorated

but this is still open for debate: Eckardt 1999, 57-90, espe-

funerary couch, imported pottery, glass vessels, an oil lamp and a bronze patera. Because of a feeding bottle among the

cially 79 and note 119. 946

See also 2.3.5.

121

With the exception of mother goddesses, statuettes of native deities and altars dedicated to either double-named deities or native ones exclusively come from rural areas. Epona is represented by two terracotta statuettes, Sucellus by one terracotta fragment and Vagdavercustis by an inscribed bronze pedestal. The name Arcanua occurs in two inscriptions just outside the sanctuary at Buchten (Limburg).947 A dedication to Haeva and Hercules Magusanus comes from Overbetuwe (Gelderland).948 Dedications to Exomna, Hurstrga, (I)Seneucaega, Sandraudiga and Rura are attested in the study area by one specimen each.

Chart 3.8

Chart 3.9

947

See for Buchten also 2.3.5.

appendix 3.1. See for all inscriptions to Hercules Magusa-

948

Roymans 2009, 233 and. fig. 10. Derks mentions five

nus also Bauchhenß 2008 and 3.8.5.

more dedications to Hercules Magusanus: Derks 1998, 249,

122

Chart 3.10

3.7.2

The urban settlements

The total number of statuettes, complete and fragmented, from the four urban settlements amounts to 165, including 140 terracotta figurines, (85%), 22 metal specimens (13%), and three made of limestone (2%). The 106 identifiable statuettes comprise 55 deities, 30 human figures and 21 animals.949 Ulpia Noviomagus has yielded 91 statuettes and Voorburg 59, together 150 statuettes. Only small parts of Roman Maastricht and Heerlen have been excavated, which explains why those two settlements together have yielded only 15 specimens.950 Among the male deities, Mercury is best represented by 11 statuettes and one attribute, a caduceus. Since towns were centres of trade, the presence of Mercury in his role as god of commerce and financial gain is not surprising. Nevertheless, Mercury is also well represented in rural and military settlements, where he also may have been worshipped for his other qualities, as a god of good fortune, abundance and fertility. These aspects are indicated by attributes related to good fortune, such as the Mercury statuettes with phalli from Pompeii and Tongeren (see below), or a Mercury bust with tintinnabula from Orange, France.951 Apollo and Bacchus are represented by five and two specimens respectively. At Voorburg, three terracotta Apollo statuettes came to light, while the god is absent in Ulpia Noviomagus. 952 From Heerlen come one bronze and one terracotta specimen. Apollo is the god of light, prophecy and oracles, healing, plague and disease, music, song and poetry. Bacchus, too, had many capacities, but was probably best known as the god of wining and dining. Their connection with symposia could explain the presence of both deities in urban settlements, where elites came together to feast, to converse and listen to music and recitations of poetry. The two Apollo statuettes from Heerlen could also point at a veneration of Apollo as the god of healing. Heerlen is situated some twenty kilometres from Aachen, where a bath complex with a sanctuary for the Celtic god Grannus was located.953 Dedicatory inscriptions to Apol949

See charts 3.11-3.13 and table 4.

950

The statuettes from the associated cemeteries not included.

951

Rolland 1965, 53, nr. 65.

952

Only specimens have been taken into account of which

the identification is certain. The actual number of Apollo figurines could be higher. 953

Derks 1998, 100 and note 100.

123

lo-Grannus outside the study area come from Erp (Nordrhein-Westfalen, Germany) and from Grand (Vosges, France).954 Terracotta Venus and mother goddess figurines prevail in urban settlements with 17 and 13 specimens respectively. Nine mother goddesses have been recovered from Voorburg. Despite the proximity of the territory of the Ubii, only two mother goddesses and two to three altars dedicated to mother goddesses have come to light at Ulpia Noviomagus.955 One of the altars is dedicated to the matres Aufaniae, the other to the matres Mopates. The dedication of the third altar is uncertain.956 Fortuna and Minerva are represented by four specimens each. Four Fortuna statuettes came to light at Ulpia Noviomagus, whereas the goddess is absent at Voorburg-Arentsburg. Among the human figures, two representations stand out. The first is the bronze Lar, which is the only specimen from an established context in the study area.957 Statuettes of Lares are rare in the northwestern Roman provinces. This is not surprising, considering the Italic origin of the cult.958 The second is a hand-modelled clay figurine. It shares some similarities with clay figurines from France, Britain and Germany. Presumably, these figurines were involved in rituals that constitute one of the topics of chapter 4. Terracotta busts are better represented at Ulpia Noviomagus than at Voorburg-Arentsburg, where only one bust has come to light. It has been recovered from the bottom of a well: the head of a Risus bust.959 The nine specimens from Ulpia include seven busts of a woman, two of a boy and two unidentifiable specimens. No statuettes have been recovered with certainty from the precinct of the temple complex at Nijmegen-Maasplein. Apart from one limestone statuette, no other statuettes of Hercules or altars dedicated to Hercules or Hercules Magusanus have come to light in the vicinity of the temples or elsewhere in the civitas capital of the Batavi. The bronze head of a male figure from the precinct of the temple at Maastricht (Limburg), was found together with a bronze weight in the shape of an acorn and a fragment of a bronze altar with an acorn on top.960 Another small bronze head was recovered from the Empel sanctuary in North Brabant. Whether such heads depict a human figure or a deity is unclear. The iconography of both heads is unknown to us, but it has been suggested that they could testify to the veneration of the head.961 During construction work, a bronze hoard consisting of 80 fragments was discovered at Hessenberg, not far from the Roman road that connected the civitas capital and the military camps to the east.962 The location Hessenberg was once part of the burial site of Ulpia Noviomagus.963 Samples taken from the fragments contained large amounts of cuprite (Cu2O), indicating oxidizing conditions. This means that the bronze fragments, before they were buried, must have been exposed to an oxygen-rich milieu for a longer period, like a smouldering fire. Many pieces were covered by a dark crust. They are deformed, half-melted or melted beyond recognition. The ceramic chips stuck to many fragments are also indicative of a fire. Therefore, it has been advanced that looters collected the fragments from (grave?) monuments, houses and public buildings

954

A bronze pedestal with an inscription to Apollo Grannus

Augustus, the cult spread and became also popular among

was found in the Rhine near Nijmegen: CIL 13, 8712; Spick-

city dwellers: Kaufmann-Heinimann 1998, 182-183 and

ermann 2008, 231. See for the inscriptions outside the study

955

appendix 1, table 1.

area: Erp, CIL 12,7975; Grand, AE 1937, 55 = AE 2013, 1102 = AE

959

See 2.3.2.

2013, 1105 and AE 1982, 714 = AE 2006, 856 = AE 2013, 1102.

960

See for the site 2.3.5; for the acorn fig. 2.189.

The figurines from graves not taken into account. See for

961

See for the head from the Empel sanctuary Swinkels 1994,

the other figurines Daniels/Brunsting 1955, nrs. 38-39, 44. 956

Daniels/Brunsting 1955, 35 and plate VIII nrs. 38-39.

957

A second specimen comes from an unknown context:

89-90. 962

Veen 2010; for the publication of the hoard Veen 2014a,

Zadoks et al. 1983, nr. 205. 958

Originally, Lares were guardian spirits of the fields, hearths

See for the publication of the inscription fragments Eck/ 129-147; Veen 2014b, 227-260.

963

Although the fragments have been recovered from Ulpia’s

and crossroads, compita. Their festival, the Compitalia,

cemetery, their exact provenance is unknown, which is

was celebrated after work in the field had finished. Under

why the hoard is mentioned here.

124

after a fire had destroyed a large part of Ulpia Noviomagus in the second half of the 2nd century. For some reason they had to bury the fragments and never came back to retrieve their booty. The identifiable items consist of sculpture fragments, such as a piece of a life-size arm or leg and the upper part of a thyrsus, heavy ornamental pieces and inscription fragments, including four of a monumental bronze tablet. One of the sculpture fragments is a large part of a Jupiter-Ammon bust and will be discussed below.964 The ritual deposits in the harbour and wells at Voorburg-Arentsburg suggest that practices of ritual deposition were not confined to rural settlements. In Roman London, finds from the Walbrook Stream have been interpreted as ritually deposited objects, too.965

Chart 3.11

Chart 3.12

964

See 3.8.11.

965

Mattingly 2006, 315-316: Fittock 2015.

125

Chart 3.13

3.7.3

T h e l i m e s z o n e

The total number of statuettes and statuette fragments from the limes zone amounts to 236, including 182 terracotta (77%) and 54 metal (23%) specimens, of which 177 could be identified.966 One hundred one specimens represent a deity. 44 depict human figures and 32 an animal (charts 3.13-3.16). Among the male deities, Mercury is best represented by 11 statuettes and three caducei. Hercules is attested by five specimens. Venus outnumbers all deities with 28 statuettes, followed by Fortuna (15), mother goddesses (14) and Minerva (13). Striking in the frontier zone is the scarcity of statuettes depicting Mars (two) and Jupiter (three), considering the relation between these deities and victory in battle. Nevertheless, the number of dedicatory inscriptions in the limes one is substantial, especially from the industrial site at Holdeurn.967 As the personification of victory, Victoria is frequently depicted on coins, but statuettes of this goddess are rare. They are absent in the Dutch limes zone, which also applies to dedicatory inscriptions to her.968 Although military troops stimulated the spread of cults in the conquered provinces, statuettes of non-Roman deities such as Harpocrates, Attis and Cybele are scarce in the limes zone and also occur in rural and urban settlements. Terracotta animal figurines are well represented and may have been cheap substitutes for, among other things, offerings of livestock, as has been suggested by Nikolaus Kyll and Victorine von Gonzenbach.969 Bronze animal figurines such as billy goats, rams and cockerels are associated with Mercury. The presence of horse figurines could be related to the presence of cavalry units that were based in the frontier zone. So far, the only sanctuary with architectural remains in the limes zone is the sanctuary in the military settlement at Aardenburg, where in its vicinity a bronze statuette of Isis-Fortuna and a terracotta fragment of a statuette of one of the Parcae came to light.970 A cult place without architectural remains in the limes zone is the presumed cult place at The Hague-Scheveningseweg, where over twenty terracotta statuettes have been recovered from a terrain with several pits and statuette fragments.971 966

See also table 6.

967

See for this site also 2.3.4.

968

Three specimens are known in the study area: Castricum

969

Kyll 1966, 64-66; Von Gonzenbach 1967, 8.

(North Holland), Colmschate (Overijssel) and Wijchen-Al-

970

See 2.4.2.

verna (southern Gelderland). Dedications to Victoria are

971

The soil from the excavation was sieved to recover also tiny

Dalheim ( CIL 13, 4056 = AE 1992, 1249) and Osterburken (CIL 13, 6577 = AE 1896, 15).

known from Castleford (RIB 628 = AE 1892, 98), Lanchester

fragments: Van Boekel 1989, note 3. See for this site also 1.8

(CIL 07, 443 = RIB 1086), Greetland (CIL 07, 200 = RIB 627),

and 2.3.2.

126

Besides the dangers soldiers had to face in battle, they also had to deal with an often unknown, hostile environment and climate. These circumstances may explain the presence of objects with an apotropaic meaning in military settlements, like phallic amulets and a small but specific category of statuettes. They represent dwarfs and ithyphallic dogs and will be discussed below. Ritual deposits occur in and around early 1st-century military camps as well as in 3rd-century forts, implying that it was a practice that continued without interruption, performed by soldiers from different parts of the Empire.

Chart 3.14

Chart 3.15

127

Chart 3.16

3.7.4

The cemeteries

The total number of statuettes from cemeteries amounts to 47, including 43 terracotta (91%) and four amber (9%) specimens (charts 3.17-3.19).972 Venus and mother goddesses are best represented by eight and five specimens. The 16 animal figurines comprise 15 terracotta specimens and one amber figurine. The majority of the statuettes come from cemeteries at Nijmegen. The remainder come from Tiel (one), Veldhoven-Heers (three), Esch (one), Hoogeloon (one) and Heerlen (four). In the past, research on cemeteries at Nijmegen has been fragmentary and focused on the grave goods.973 Research on mortuary rituals along the lines of Roymans, Hiddink and Pearce could shed a light on burial rituals in Nijmegen’s Roman period cemeteries and beyond.974 To this day, little is known about their development, spatial organisation, social differentiation and so on.975 The six graves with statuettes found in situ at Nijmegen do not provide any clues for the use of statuettes through time, but it might not be coincidental that four of the six graves date to the late 3rd or 4th century.976 All other terracotta statuettes from cemeteries at Nijmegen are stray finds. The dates attributed to them by Van Boekel mostly refer to the production date of the statuettes.977 Considering the over 1,500 investigated Roman period graves at Nijmegen and the small number of statuettes from cemeteries and graves, it can be concluded that it was not customary in this part of the study area to deposit statuettes in graves in the Roman period.978 Remarkably, the statuettes from cemeteries comprise exclusively terracotta statuettes and four amber specimens. Bronze statuettes are absent. Here, too, there is a preference for Venus and mother goddesses, but also animals are well represented and constitute over a third (40%) of the total number of statuettes.

972 973

See also table 5. See, for instance, Vermeulen 1932; Brunsting 1937; Stuart

northwestern provinces: Pearce 2014, 99-111. 975

1977; Koster 2013; Steures 2013; Van Enckevort/Heirbaut 974

See for a concise overview of the development of the cemetery at Nijmegen-West Koster 2013, 208-211.

2013. An exception is Bloemers 2016.

976

See for the Nijmegen-cemeteries 2.3.4.

Roymans on mortuary ritual in La Tène cemeteries in

977

Van Boekel 1987 nrs. 14, 45-46, 61, 81, 87, 93, 101, 108, 110, 116,

northern Gaul: Roymans 1990, 217-259; Hiddink on rituals in late Iron Age and Roman period cemeteries in the Meuse-Demert-Scheldt region: Hiddink 2003; Pearce on the sequence of ritual in cremation burials in the Roman

128

197, 209-210, 221-223, 229, 238-240, 247, 249, 253. 978

See for the number of investigated graves Van Enckevort/ Thijssen 2014, 37-41.

With the exception of Bacchus and Amor, male deities are absent. Amor is represented by two amber specimens, Bacchus by one. Besides Amor’s usual iconography, standing while stringing his bow, statues and small figurines also depict him sleeping. As a small sculpture in funerary contexts, the iconography probably symbolises death as a temporary phase or death as a metaphor for eternal sleep. 979 The presence of Bacchus can be explained by his link to the cycle of life. Mercury as the guide of souls to the underworld, is absent in cemeteries in the study area.980 Margherita Bolla has demonstrated that a small number of graves outside the study area contained Mercury figurines.981 In the study area, more than a third of the figurines from cemeteries depict animals. Besides domestic animals, striking representations are four horned birds and three ithyphallic dogs. Their presence in graves could be related to their function as protectors of the deceased during their journey into the underworld. Three of the four amber figurines from cemeteries in the study area come from the graves of wealthy women.982

Chart 3.17

Chart 3.18 979

See for this type of statue Hemingway/Stone 2017, 46-63.

982

980

See for Bacchus in the study area Veen 2014a, 2014b.

ed to amber 3.7.5. See for ithyphallic dogs from graves also

981

Bolla 2013, see also below.

De Beenhouwer 2005, 825.

See for the amber figurines 2.3.4; for the qualities attribut-

129

Chart 3.19

3.7.5 Cemeteries: a relation between find-context and material The cemeteries in the study area have yielded predominantly terracotta statuettes of female figures (86%). Metal figurines are absent: the remainder consist of amber specimens. Bolla has investigated graves throughout the Roman Empire and found no more than eighty graves which contained a bronze or lead figurine. This is a remarkably small number, considering the huge number of burials from the Roman period.983 The majority have been recovered from cemeteries at the fringes of the Empire. Examples closest to the study area come from Xanten, Krefeld-Gellep, Cologne (Germany) and Tongeren-Koninksem (Belgium). Among the eighty statuettes Venus and Mercury prevail. Also striking is the number of bronze animal figurines in late Roman graves. They predominantly represent dogs and cockerels, but also unusual or exotic representations occur. Examples are a crane, an owl and a hippo carrying a cobra on his back from graves in Cologne and Bingen (Germany), and a mouse from a grave in York (Britain). 984 Fontijn has noted that bronze objects seldom occur in middle and late Bronze Age burials in northwestern Europe. However, this does not explain the absence of metal statuettes in Roman period graves, since metal vessels, jewellery and other metal objects were not exceptional as funerary gifts.985 Bolla has argued that the scarcity of bronze statuettes in graves results from a combination of the nature of metal itself and the presence of a dead body. Placing a bronze statuette of a deity in a grave together with a dead body would ‘infect’ the deity.986 The properties attributed to metal, and to lead in particular, may also be the reason why lead figurines and metal curse tablets in the Greco-Roman world were placed in graves to manipulate or harm living individuals.987 One of the richest graves from Ulpia Noviomagus’ burial site contained 13 amber objects, including two figurative representations. The first depicts a sleeping Amor, the second a bear.988 The grave could be dated between 80 and 100 and contained the remains of a woman in her twenties.989 A comparable amber figurine 983

Bolla 2013, 1.

985

Fontijn 2002, 81-84, 211-212 and tables 10.1-10.2.

984

Cologne: from the burial site near the St. Severin church, a

986

Bolla 2013, 16-17.

ram on a pedestal, a dog on a pedestal, found together with

987

See for metal curse tablets and figurines, for instance, Edmonds 2019, 55-63.

a coin from a grave dating to 188-189. The crane comes from a destroyed grave. Krefeld-Gellep: a 155 mm high Venus

988

A comparable amber bear figurine was found in a 3rd-cen-

statuette on a pedestal and an owl on a pedestal, set up

tury grave at Elsdorf-Esch, just north of Cologne: Koster

beside a wooden coffin. Xanten: a now lost figure wearing

2013, 179.

a winged helmet (Flügelhelm), probably Mercury, from a grave with a sarcophagus containing the remains of an adult man: Bolla 2013,10-11, 23-24.

130

of a sleeping Amor has been recovered from a 2nd-century burial at Heerlen-Landgraaf.990 A fourth amber sculpture from a burial in a tumulus at Hoogkeiteren, Esch (North Brabant), depicts a drunken Bacchus supported by a young satyr. The burial was dated to the first or second quarter of the 3rd century. Besides the amber Bacchus, other amber objects included a spindle and two pendants depicting stylized gladiator helmets.991 Funerary gifts like the amber spindle, jewellery, beads and slippers suggest that the deceased was a woman.992 The occurrence of amber objects in graves of women suggest a relation between death, women and amber. Ovid tells us that the two sisters of Phaëton, after his death, turned into poplar trees. Their tears ‘hardened into amber by the sun drop down from the new-made trees. The clear river receives them and bears them onward, one day to be worn by the brides of Rome’.993 Pliny the Elder mentions amber not only as a highly valued luxury, but also describes the medicinal and apotropaic qualities of amber, which should be worn by babies as an amulet.994 Because of these properties, the amber figurines may also have been regarded as suitable to accompany the deceased to the underworld. And since amber was expensive, it is not surprising that amber sculptures and objects come from graves of wealthy women, who may also had worn amber jewellery during their life time.

3.8

The area south of the Rhine: iconographic analysis

3.8.1

A relation between material and subject

South of the Rhine, terracotta statuettes prevail. It appears from the analysis that female figures such as Venus, mother goddesses and Fortuna are predominantly made of clay. Terracotta statuettes of Minerva slightly outnumber bronze specimens. All four Victoria statuettes from the study area are made of bronze.995 Two examples from Belgium demonstrate that the goddess is sporadically depicted in terracotta.996 Nearly all statuettes of Jupiter, Mars and Hercules are made of bronze.997 Outside the study area, a few terracotta specimens of Hercules and Mars are known.998 Apollo, Bacchus and Mercury are predominantly represented in bronze and occasionally in clay. All three statuettes of genii in the study area are made of bronze, but an example from Belgium demonstrates that clay specimens also occur.999 Since statuettes were commodities, we have to look at economic mechanisms such as market demand and supply, which most likely functioned in the same way in antiquity as in our time. How production centres and workshops organised the manufacture of bronze and terracotta statuettes is unknown. Neither do we know if modellers exchanged moulds or how often they worked on commission, but it appears from written sources and archaeological evidence that craftsmen in antiquity made what customers wanted and also reacted to external factors such as unrest, disease, economic stress, the dedication of a new sanctuary or the approach of a festival. A significant passage in the New Testament tells us that when the apostle Paul and a group of his disciples stayed at Ephesus (modern Turkey), a certain Demetrius organised an uproar. He was a silversmith 989

Koster 2013, 173-179.

994

Plin. HN 37.12.49-51.

990

See 2.3.5.

995

See table 1.

Van den Hurk 1986, 112-115, 122-123, 133, plate VI and 391,

996

De Beenhouwer 2005, nrs. 312, 825.

plates 4-5.

997

See also Kyll 1966, 42-43.

992

Van den Hurk 1986, 18-19.

998

See for Mars also De Beenhouwer 2005, nrs. 337, 395-396;

993

Ov. Met. 2.364-366: inde fluunt lacrimae, stillataque sole

991

rigescunt de ramis electra novis, quae lucidus amnis excipit et

for a terracotta Hercules from Colchester Eckardt 1999, 65. 999

De Beenhouwer 2005, nr. 340.

nuribus mittit gestanda Latini.

131

and the craftsmen of his workshop sold silver miniature shrines, copies of the famous Artemis temple. The miniatures of the temple, which then was one of the seven wonders of the ancient world, brought in a lot of business. Demetrius and his men, fearing that the preaching Christians would jeopardise their profits, incited a riot.1000 Archaeological evidence has been presented by Frederick Naerebout in a publication about religious choice. He discusses a goldsmith’s hoard, known as the Galjub hoard. It came to light just north of Cairo, in 1912 or 1913 and included a set of tools and 106 bronze statuettes used as goldsmith’s models for jewellery and medallion-shaped relief plaques. The models comprised Greek and Egyptian motifs, bodies, heads and dress fragments, implying that the goldsmith offered his clientele the possibility to choose the deity they preferred. Moreover, the modular nature of the hoard shows that the goldsmith could easily turn a Greek deity into an Egyptian one and vice versa, by changing head, torso or legs. 1001 In the same article, Naerebout also compares oil lamps from Syria-Palestine, where a producer of oil lamps offered his clientele Jewish and Christian motifs, as well as ‘neutral’ ones. Examples of sarcophagi, too, demonstrate that workshops offered their clients the possibility to express their individual wishes.1002 These cases support the argument that bronze statuettes, too, were made according to the individual wishes of customers. It is most likely one of the reasons why most bronze statuettes are unica.1003 Von Gonzenbach has argued that terracotta statuettes initially were produced for the military market, which, according to her, explains why Roman and ‘male-oriented motifs’ dominated the 1st century. She points out that the change of motifs and the appearance of new consumer groups were a phenomenon typical of the 2nd century.1004 Whether this could apply to the study area cannot be established, since the dataset contains very few statuettes that can be dated with certainty to the 1st century. We can only observe that the repertoire of bronze and terracotta statuettes in Gaul and Germania begins to diverge from the end of the 1st century onward. Imagery of single mother goddesses and triads of mother goddesses becomes a new and influential theme in terracotta and stone sculpture.1005 Bronze mother goddesses, inside and outside the study area, are rare.1006 1000

Acts 19, 23-28.

the kiln. Wax working models were freely modelled, they

1001

Naerebout 2010, 55-61 and note 22.

were assembled from copied parts of different models,

1002

Zanker/Ewald 2004, 51-52.

or they were a combination of copied parts and freely

1003

Another reason is the lack of a standardized production

modelled elements: Kaufmann-Heinimann 1998, 16-20;

method for small bronze sculptures. They were cast by the

Mattusch 1996a, 21-23; Mattusch 1996b, 8-9. See for prob-

direct or indirect lost-wax casting method. Larger sculp-

lems with the dating of statuettes 1.8. See for an example

tures were usually cast hollow, which was less expensive.

of a composite bronze figurine the Mercury statuette from

Small figurines were cast solid or hollow. The direct casting

Tzum: 2.2.2. See for a discussion on serial production: Kauf-

method involved a wax model onto which thin layers of clay were applied. Placed in the kiln, the wax burnt out

mann-Heinimann 1998, 18-20. 1004

Von Gonzenbach 1995, 400-428. According to others, the

and molten metal was poured into the clay mould. In the

cheaper terracotta statuettes of the central Gallic potteries

case of indirect casting, clay or plaster moulds were taken

were a form of spiritual resistance to the Roman order:

in separate sections (head, body, limbs) from a full-size

Drinkwater/Vertet 1992, 27, Eckardt 1999, 61 and notes

model, the original. The moulds were lined with a layer

22-23. Drinkwater and Vertet are not the only scholars who

of bee wax, either by applying wax slabs onto the moulds

believe that examples of provincial sculpture testify to

or by pouring or brushing wax into the moulds. When the

spiritual resistance, as has been demonstrated by contribu-

moulds were removed the wax sections were put togeth-

tions to a colloquium held in 2001, entitled Romanisation und Resistenz: Noelke et al. 2003.

er. This wax working model allowed the artisan to make changes before casting, by adding elements and by cutting

1005

It will be advanced in paragraph 3.8.9 that terracotta figu-

into the surface of the wax model. The wax model was

rines of Fortuna and Minerva copied the enthroned pose of

divided into sections again and thin layers of clay were

mother goddesses, whereas bronze specimens maintained

applied onto the surface of each section and placed inside

their standing pose. It will also be argued in the same para-

132

The most obvious explanation for the use of clay for the production of figurines is that this material was easily accessible and cheaper than the raw materials needed for bronze, which is why it was used to satisfy the increasing demand for statuettes. Terracotta statuettes were made with moulds that were used over and over again. The moulds were often derived from a model or archetype and could be adapted. Strong resemblances between statuettes of Fortuna and Minerva, for instance, show that they were derived from the same archetype.1007 This, however, does not solve the question why there was an increasing demand for specific female deities and mother goddesses in the northwestern provinces. In other words: who were the buyers of these statuettes? It appears from the analysis that the demand for terracotta figurines of Venus, mother goddesses, Fortuna, Minerva, Cybele and Juno must have been stronger than the demand for male figures such as Jupiter and Apollo.1008 The question which consumer group(s) were behind the demand for statuettes of female figures is a tricky one. If statuettes were initially produced for the military market, this could imply that the buyers were mostly male, but this is doubtful. We know that women could live inside forts and that merchants, craftsmen and soldiers could live with their families in canabae and vici. Moreover, the ratio men to women in these settlements is unknown.1009 It has also been argued that the users of figurines of females and busts of children were women, asking the gods for fertility and the health of their children.1010 If busts of children were used in rites of passage, which is uncertain, there is no reason to assume that men were less involved in such rituals than women. And lastly, there does not have to be a link between the gender of a deity and the gender of the worshippers, which is demonstrated by the dedicators of the inscribed Nehalennia altars, which were all male. For the time being, only epigraphy can help us to gain insight into the gender of the users of statuettes, altars and other religious objects. 3.8.2 Interpretatio and creolisation In an often cited fragment from De Bello Gallico, Caesar equates the Gallic gods with Roman ones: ‘Among the gods, they most worship Mercury. There are numerous images of him; they declare him the inventor of all arts, the guide for every road and journey, and they deem him to have the greatest influence for all money-making and traffic. After him they set Apollo, Mars, Jupiter, and Minerva. Of these deities they have almost the same idea as all other nations: Apollo drives away diseases, Minerva supplies the first principles of arts and crafts, Jupiter holds the empire of heaven, Mars controls wars.’ 1011 Interpretatio, replacing the names of the other’s native deities by the names of one’s own gods, is not a typical Roman phenomenon. It was already practiced centuries before Caesar, for instance by the Greek historian graph that not only Roman goddesses were worshipped as

structure of military communities: Phang 2001, 231-243,

mother goddesses but also the native goddess Nehalennia.

Greene 2012, 107-108. I thank Emily Hemelrijk for bringing

1006

See 3.5.4 and table 1.

1007

De Beenhouwer 2005; De Beenhouwer 2014, 50-60.

1010

Van Boekel 1987, 238.

1008

See table 1.

1011

Caes. BGall. 6.17: Deum maxime Mercurium colunt. Huius

1009

See 1.8. Soldiers in the Roman army could not marry, but

sunt plurima simulacra: hunc omnium inventorem artium

soldier’s epitaphs, papyri and military diplomas attest

ferunt, hunc viarum atque itinerum ducem, hunc ad quaestus

to their union with free women, freedwomen and slave

pecuniae mercaturasque habere vim maximam arbitrantur.

women. The Vindolanda tablets offer us a glimpse of the

Post hunc Apollinem et Martem et Iovem et Minervam. De his

role of elite women along the northern limes. Wives of

eandem fere, quam reliquae gentes, habent opinionem: Apolli-

army commanders, for instance, played a role in the cel-

nem morbos depellere, Minervam operum atque artificiorum

ebration of female festivals like the matronalia in forts.

initia tradere, Iovem imperium caelestium tenere, Martem

This seems to stress the domesticity of women, but it also

bella regere.

the publications by Greene and Phang to my attention.

testifies to the integration of officers’ wives in the social

133

Herodotus (484-425 BCE) who described the Scythian pantheon in his Historiae.1012 In the Roman period, the visibility of interpretatio is clearest in the form of epigraphic name-pairing, associating the name of a classical deity with a native one. Examples in the study area of dedications to gods with a native name paired with a Roman name are Hercules Magusanus, Mars Halamardus, Mercurius Avernus, Mercurius Friausius and Apollo Grannus.1013 It appears from the dates of such dedications that this process started in the early 1st century and continued into the 3rd century. Based on his analysis of inscriptions to deities with single and double names from Roman Gaul, Derks assumed at the time that the names of these deities in the inscriptions were indications of change and that the names of local deities gradually would be replaced by Roman counterparts.1014 In the same publication, he comes to the conclusion that ‘The association of names (...) enabled the North Gallic communities to preserve something of the original identity of their precious god and, at the same time, to emphasize their loyalty to the Roman cause.’1015 Others went a step further, by arguing that name pairing was as an act of cultural imperialism, a colonial act by a ruling power.1016 Also attempts have been made to explain name-pairing from a linguistic point of view.1017 Conquered and colonized peoples created new cultures by taking over elements of the invader’s or colonizer’s culture, especially those elements they wanted or needed. This does not imply that they also took over the ideas, meanings and ideologies behind those elements. The aim of this section is to examine interpretatio and creolisation in representations of deities in sculpture. Since only epigraphy can tell us with certainty who is depicted, I will discuss imagery of deities accompanied by inscriptions: Roman deities depicted with native deities, deities with Roman and non-Roman characteristics and Roman-looking deities with the name of a native deity on the pedestal. The first monument that will be discussed is the Pillar of the Boatmen from the time of Tiberius, which has been analysed by William van Andringa.1018 The monument was found in detached pieces in the Notre Dame, Paris. The inscription tells us that it was erected by a collegium of boatmen in Paris and dedicated to Jupiter and the emperor Tiberius.1019 Alongside the Roman gods Jupiter, Vulcan, Mercury, Fortuna, Mars and Venus, also Gallic gods are portrayed on the Pillar, together with their names: Enus, Cernunnos, Smertrios(?) and Tarvos Trigaranus, the bull with three cranes on his back. The Gallic deities have retained their non-Roman iconographies. A statue of Jupiter, the Roman supreme god, once stood at the top.1020 Also historic scenes are depicted on the blocks: bearded and beardless men, probably the older and younger boatmen themselves, bringing a sacrifice to a cult statue or the emperor. Although it is unknown whether the Pillar originally stood in a temple or in the headquarters of the college, the inscription tells us that it is a public dedication, financed with college funds.1021 Van Andringa rejects the idea that representations of Roman gods accompanied by native deities are examples of the persistence of indigenous religion. He argues that their association represents evidence of a new religious order.1022 The ancestral Gallic gods on the Pillar were relegated to second place, as Van Andringa has stated, to establish the new order, the power of Rome.1023 By dedicating the monument to Jupiter and the emperor, the Parisian boatmen testified in public to their adherence to the Roman emperor and the Roman supreme god. The portraits of Roman and Gallic deities were accompanied by their names because it was 1012

Hdt. Hist. 4.59.

1013

Derks 1998, 249, 250, 252, 254.

1014

Derks 1998, 91.

1018

Van Andringa 2011.

1015

Derks 1998, 242.

1019

Van Andringa 2011, 122, CIL 13, 3026.

1016

See, for instance, Webster 1995.

1020

Van Andringa 2011, 124-125.

See, for instance, the research programme F.E.R.C.A.N.

1021

Van Andringa 2011, 122-124, 126 and figs. 4.16-17.

(Fontes epigraphici religionum celticarum antiquarum),

1022

Van Andringa 2011, 124-130. See also Raepsaet-Charlier 2015,

1017

received strong criticism by Raepsaet-Charlier: 2015, 184191.

187 and note 64.

launched in 1998 by the Austrian Academy of Sciences. Attempts to ‘translate’ Celtic theonyms and epithets have

134

1023

Van Andringa 2011, 128.

important that onlookers understood the message of the monument and the new hierarchy, with Jupiter and the emperor at the summit. 1024 The Pillar could be regarded as evidence of perfect integration, but its erection may also have been motivated by pragmatism, to safeguard the boatmen’s professional interests. The examples that will be discussed hereafter are private dedications: statuettes and reliefs of deities with Roman iconographies, while the inscription indicates that a native god is depicted. Here, the use of Roman iconographies serves another purpose. The dedications demonstrate how people took over iconographies of Roman deities for their ‘own’, native deities. In this respect it is important to realise that, if workshops responded to the demands of their clientele, this implies that these were the iconographies people wanted. A bronze statuette from Muri (Switzerland) depicts a standing woman wearing a chiton and mantle which is fastened on her left shoulder. The iconography strongly resembles that of the Roman goddess Fortuna. The two hands are missing, but they possibly held a patera and a cornucopia, like Fortuna. Without the inscription on the pedestal we would not have known that the statuette represents the goddess Naria. The same applies to a statuette from Champoulet (France). The standing woman bears a striking resemblance to Fortuna, but according to the inscribed pedestal, Rosmerta is depicted. The attributes in her hands are missing, but since she is depicted in Gaul as the companion of Mercury, she may have held a marsupium.1025 A statuette of a male deity with ‘mixed’ Roman and native characteristics from Noville-lez-Bastogne in Belgium depicts a young male with long curly hair, standing in contrapposto, wearing a long-sleeved tunic and a nebris, in this case not the skin of a fawn but a wolf skin.1026 He holds a (now missing) lance in his raised right hand. The posture of the figure is comparable to many standing figures in Roman sculpture, but the lance in the uplifted right hand is typical for Mars. The long curly hair and the nebris are characteristics of Bacchus, while the nebris is draped like a chlamys, a garment often worn by Mercury. According to the inscription, the statuette depicts the Gallic god Enarabus, also known as Intarabus.1027 Another example of a deity with mixed characteristics is a bronze statuette from Burgundy, France. If statuettes of Roman-looking deities with the inscribed name of a native deity are difficult to understand, without an inscription the interpretation is even more problematic. Occasionally, the find-spot of the sculpture might give away which deity is depicted. This statuette is such an example. It depicts a woman standing in a boat with a prow in the shape of a duck’s head. The appearance of the woman is Roman. She has a diadem in her hair and wears a long garment and a palla. The statuette was found at the precinct of a sanctuary dedicated to the goddess Sequana, at Fontes Sequanae, Burgundy. Therefore, it has been argued that she represents Sequana, the divine personification of the river Seine.1028 If the statuette had been a stray find, we would have wondered which Roman goddess is depicted. An altar depicting a native deity with a Roman iconography from the study area comes from Tiel-Zennewijnen.1029 The altar is dedicated to the goddess (I)Seneucaega, whose name is attested once in the study area. We do not know anything about this goddess, but in the relief she is depicted as Diana with a deer by her side. This suggests that the capacities of (I)Seneucaega came closest to those of the Roman hunter-goddess. A statuette pedestal found in the Linge River (southern Gelderland) bears a Latin inscription of a dedicator serving in the British army who addresses the Germanic war goddess Vagdavercustis.1030 The statuette has not been found. A sanctuary dedicated to Vagdavercustis was situated some forty kilometres east of 1024

Van Andringa 2011, 125-128.

1026

Faider-Feytmans 1979, 69 and nr. 52.

1025

Kaufmann-Heinimann 2012, 4-5. The inscription on the

1027

The inscription reads: Deo/Intarabo/Ategniomarus/

first statuette reads: Deae/Nariae /reg(io) Arure(nsis) /

Respecti/ v(otum) s(olvit) l(ibens) m(erito). AE 1965, 28a1.

cur(ante) Feroc(e)/ l(iberto): CIL 13, 5161 = AE 2010, 993.

1028

See, for instance, Deyts 1992, 74; Green 1992c, 188-189.

The inscription on the second statuette reads: Aug(usto)

1029

See 2.3.4.

sac{c}r(um) / d(e)ae/ Rosmer/t(a)e/Dubno /caratiaci/Mar-

1030

Zadoks et al. 1969, nr. 81 and 2.3.4. The inscription reads:

ross(us) Marulli/ filius/ v(otum) s(olvit) l(ibens) m(erito)/

Deae /Vagdavercusti /Sim[p]licius /Super /dec(urio)/alae

d(e) s(uo) d(edit). AE 1980, 643.

Vocontior(um) /exerci[t]uus /Britannici: CIL 13, 8805.

135

Nijmegen on the Kalkarberg near Kalkar (Germany). At the site, an altar dedicated to the goddess was found in 1830.1031 The sanctuary was in use from the middle Augustan period into the 4th century. At the temple precinct, pottery, bones, weapons, helmets, militaria and 110 gilded bronze fragments of several statues came to light. Just like the weapons found at the Empel sanctuary, the weapons and other militaria were probably dedicated by soldiers or veterans to the deity of the sanctuary.1032 Whether one of the statues depicted Vagdavercustis could not be established. It would be quite exceptional if the statuette belonging to the pedestal from the Linge River represented the Germanic war goddess, since images of this goddess are unknown. It is possible that the soldier who dedicated the statuette to Vagdavercustis may have chosen a representation of Minerva, whose martial appearance must have come closest to that of the native war goddess. These examples suggest that Roman iconographies were chosen to ‘give a face’ to native deities who until then may have been an-iconic. Four of the five examples are dedications to goddesses. Apparently, the iconographies of important Roman deities were regarded as fitting for them. The dedicators of the discussed statuettes did not combine the name of a Roman and a native deity on the pedestals. For Lower Germany, Derks has observed that name-pairing does not occur in inscriptions to goddesses.1033 In Britannia, however, the association of Sulis and Minerva occurs frequently in inscriptions from her sanctuary. Of the cited examples, the only statuette that depicts a male deity is Intarabus. In most dedicatory inscriptions he is invoked alone.1034 In one inscription, from Trier, Intarabus is associated with Mars.1035 Despite the ‘mixed’ characteristics of Bacchus and Mercury, he is not associated with these deities in the epigraphic record. 3.8.3

The iconography of the Hercules statuettes

With seven statuettes found in and near the limes zone, Hercules is the best represented male deity south of the Rhine, after Mercury. Among the varying iconographies of the Hercules statuettes, different types can be distinguished. The first type depicts a strongly stylised or simplified figure which is only recognisable as Hercules through the lion skin hanging over his left arm. In the raised right arm of which the upper arm is often missing, he held a club. To this type belongs the figurine from Hunerberg, Nijmegen.1036 Emma Durham’s online database of bronze statuettes from Britain contains several comparable examples.1037 The second statuette also comes from Nijmegen (fig. 2.135). It is an example of a relatively small group of representations that depict Hercules’ first heroic deed. We see the infant Hercules strangling the (now missing) snake(s) Hera has sent to kill the child.1038 The young and beardless Hercules from Ypenburg, The Hague (fig. 2.57), belongs to a type derived from Roman copies of a Greek statue, made by Lysippus.1039 The most famous copy of the statue probably is the Hercules Farnese.1040 Lysippus’ statue depicts an exhausted Hercules after having accomplished his eleventh labour, stealing the apples of the Hesperides. The beholder’s attention is drawn to the back of the statue, where the hero hides the apples in his right hand. Since eating from the apples ensured immortality, Lisippus 1031

Busch/Schäfer 2014, 12. Not the goddess, but the worship-

1038

Boardman et al. 1988, 830-831 and nrs. 1628, 1633-1634, 1638.

pers are depicted in the relief.

1039

RMO, inv. nr. h2013/7.1. It was found in 2002 and acquired

1032

Bödecker 2010, 16-19; Bödecker/Sarge 2014, 92-94.

by the RMO in 2013: Halbertsma 2014, 309 and notes 29-30.

1033

Derks 1998, 93 and table 3.2.

See also the gilded Hercules statue found at the Forum

1034

See, for instance, an inscription from Dalheim AE 1992,

Boarium in Rome, now in the Musei Capitolini, Rome. The

1248; from Niersbach CIL 13, 4128; from Echternach AE

statuette from Ypenburg does not hold the apples of the

1965, 199; from Trier AE 1915, 72 = AE 1916, 26. 1035

CIL 13, 3653 = AE 1896, 100.

1036

Zadoks et al. 1969, nr. 24.

1037

Durham 2012, nrs. 75-77, 87.

136

Hesperides in his hand. 1040

A marble copy of the Hercules Farnese is in the Museo Archeologico Nazionale at Naples.

may have hinted at Hercules’ future deification.1041 The myth of the Hesperides is the eleventh labour, the first of the two additional labours Hercules had to perform because Eurystheus had discounted two of the ten previous labours.1042 The limestone statue of Hercules from Valkenburg demonstrates that the maker must have been acquainted with Roman copies of Lysippus’ statue, but the hand Hercules hides behind his back is empty.1043 The Hercules from Ypenburg does not hide his hand behind his back. The largest group constitutes a type known as Hercules bibax, Hercules holding a drinking cup in his hand and often wears a wreath. The posture of this type is also derived from Roman copies of Lysippus’ statue, but now the hero is depicted with one or two extra attributes. Hercules holds a drinking cup in his hand and often wears a wreath. With seven statuettes, including a Hercules mingens, a urinating Hercules, this type prevails in the study area.1044 Three statuettes from the study area depict Hercules wearing a wreath: a bronze specimen from Venlo, the urinating Hercules from Nijmegen-Hugo de Grootstraat and the statuette from Oosterhout (figs 2.61-64). Statues and statuettes of these types have been found throughout the Empire.1045 A wreath was worn during symposia and banquets. Therefore, Hercules’ wreath probably refers to the ritual of the symposium and thus to Hercules’ deified status. The statuette from Venlo wears a wreath of ivy-leaves, which was an important attribute of Dionysus and his cult. Myrtle wreaths were worn during banquets and awarded to athletes. Myrtle was sacred to Aphrodite and Dionysus gave a myrtle plant to Hades, as a compensation for the liberation of his mother Semele from the underworld.1046 Swinkels has argued that the wreath of the Oosterhout statuette probably depicts poplar leaves, since written sources tell us that the poplar was Hercules’ favourite tree.1047 In the case of the statuette from the Hugo de Grootstraat at Nijmegen, the leaves of the wreath are unidentifiable. Since Hercules in Greek literature is staged as drunk on several occasions, the meaning of the drinking cup often point at his love for wine.1048 Several statues depict the hero in a clearly drunken state, staggering and urinating.1049 Yet, the drinking cup in Hercules’ hand also has another meaning. In a group of Pompeian wall paintings, Hercules holds a drinking cup, while he is surrounded by gods, which, just like the wreath, points at his deification. In these paintings the hero shows no signs of drunkenness. The same applies to two representations of Hercules with a drinking cup in northwestern Europe. The first belongs to the Jupiter column at Mainz, Germany, the second to a stone altar in front of the temple of Sulis-Minerva at Bath, England.1050 Besides the wreath and the drinking cup, a third element that connects Hercules with the symposium is his association with Dionysus. An explanation for this has been given by Louis Foucher. After their deification, Hercules and Dionysus feature together in myths, either as friends or as opponents. They were both sons of Zeus and a mortal woman and both were deified as a reward for their deeds. A popular theme in the Graeco-Roman world is the drinking contest with Dionysus, a scene that is depicted on Greek vases and in several mosaics.1051 Doubtless, the association with dining and feasting was also appreciated by people in the study area. An explanation for the popularity of Hercules with a drinking cup in military and non-military settlements may 1041

Swinkels 2017b, 185 and note 6.

1046

Foucher 2000, 202-203.

1042

Graves 1984, 145-152.

1047

Verg. Ecl. 7.21, Phaed. Fab. 3.17, Plin. HN 12.3: Swinkels 2017b,

1043

See for the statue from Valkenburg Hingh/Vos 2005, 58-59.

1044

Figs. 2.61-64, 2.162, 2.181 and Zadoks et al. 1969, nrs. 22-23,

187 and note 13. 1048

25. The statuette of the seated Hercules from Vechten may

1045

For instance, in Euripides’ tragedy Alcestis: Halbertsma 2014, 306-307 and notes 16-17.

have held a drinking cup in his hand too, but both arms are

1049

Boardman et al. 1988, 770-772 and nrs. 879, 882, 898, 904.

missing.

1050

Swinkels 1994, 84-85.

Boardman et al. 1988, nrs. 779-780, 782-783. Hercules with

1051

Foucher 2000, 201-206.

a (sometimes missing) drinking cup or kantharos: nrs. 794, 823, 844, 881-883. Hercules urinating: nrs. 897-898, 904.

137

be that soldiers and veterans could identify themselves with the hero, who after his death had become one of the gods as a reward for his heroic deeds and now could rest and feast. The last statuette is the limestone Hercules from Ulpia Noviomagus (figs. 2.84-84a). Despite its damaged condition, it clearly depicts the hero standing, wearing the lion skin as a harness and the head as a helmet.1052 The lion skin refers to the first of Hercules’ twelve labours, the killing of the Nemean lion. It is an important attribute, which he wears as a cape or carries folded over his arm. The limestone sculpture is the only representation of the demigod that has been found in the civitas capital of the Batavi. 3.8.4

The iconography of Hercules Magusanus

Six dedications in the study area associate Hercules with the native deity Magusanus.1053 One of them is the missing sculpture from the Empel sanctuary. The silver-plated metal sheet with inscription from Empel was once attached to the pedestal of a sculpture and measures circa 100 by 90 mm.1054 The dedicator is a veteran of the Tenth Legion Gemina.1055 Outside the study area, but still in Germania Inferior, a statue from Xanten and an altar from Bonn bear an inscription to Hercules Magusanus. In view of the aforementioned iconographies of Hercules, it brings up the question which type was chosen for the missing sculpture from Empel. The iconography of the statue from Xanten and the figure on the altar from Bonn is a frequently occurring type and depicts Hercules as a bearded, muscled, nude man. In both cases the dedicator was a member of the Roman military. The statue from Xanten dates to the first half of the 3rd century and depicts the hero holding the apples of the Hesperides in his outstretched left hand, just like the gilded Hercules statue from the Forum Boarium in Rome.1056 The altar from Bonn is a roofed aedicula and dates to the second half of the 2nd century.1057 Hercules stands in the aedicula, with his club to his right. In his left hand he holds Cerberus, the monstrous, three-headed dog of the underworld on a leash.1058 This representation depicts the twelfth labour of Hercules, the capture of Cerberus from the underworld.1059 Reliefs that depict this myth are known from Alzey (Germany), Augst (Switzerland) and Velletri (Italy).1060 A slightly different representation, but still depicting the Hesperides myth, is found on coins dating to 261, issued by Postumus (260-269).1061 The text on the coins reads ‘Herculi Magusano’. Here, Hercules is depicted leaning on his club which stands on his left, while he hides his right arm behind his back. Not every Germanic or Gallic individual may have understood why Hercules holds his hand behind his back, but the statues from Valkenburg and Xanten as well as Postumus’ coins depict Hercules after having accomplished the eleventh labour. There are of course other options for the iconography of the missing sculpture from Empel, but Hercules Magusanus as the resting hero is not an unlikely one. 1052

A rare and well-known bust of Commodus depicts the

no)/L(ucius) Vib(ius) Castus/ im(m)un(is) le(gionis) XXX.

emperor wearing the lion skin as a short cape, with the head of the lion as a helmet: Palazzo dei Conservatori, 1053

1054

deo Herculi Mag(usano) / M(arcus) Naevius Minervi / nus

Derks 1998, 98, 249. A yet unpublished temple complex at

optio princi/pis leg(ionis) I M(inerviae) P(iae) F(idelis)

Herwen-Hemeling (Gelderland) has yielded at least two

[[S(everianae) A(lexandrianae)]]/ imp(eratore) d(omino)

more inscriptions mentioning Hercules Magusanus.

n(ostro) [[Alexandro]] / Aug(usto) II et Marcel / lo co(n)

Roymans/Derks 1994, 25-26 and plate 5. See for the site

s(ulibus) v(otum) s(olvit) l(ibens) m(erito). AE 1971, 282. 1058

The inscription reads: Herculi/ Magusen(o)/Iulius Gen/

Horn 1970, 234-235; Derks 1998, 113-114. A second inscribed altar from Bonn is dedicated to Hercules or, perhaps, to

ialis veter(anus) leg(ionis) X G(eminae) P(iae) F(idel-

1056

The inscription reads: In h(onorem) d(omus) d(ivinae) /

Rome.

Empel also 2.3.5. 1055

CIL 13, 8610. 1057

Hercules Magusanus, but the inscription is badly damaged.

is)/v(otum) s(olvit l(aetus) l(ibens) m(erito). AE 1990, 740

1059

Graves 1984, 152-158.

= AE 1994, 1281 = AE 2000, 1011.

1060

Horn 1970, 242-243 and figs. 4-7.

The inscription on the statue reads: Herculi Mag(usa-

1061

RIC 139.

138

3.8.5

T h e v e n e r a t i o n o f H e r c u l e s M a g u s a n u s 1062

Roymans and Derks have suggested that the sanctuaries at Empel and Elst were dedicated to Hercules/Hercules Magusanus. In chapter two it has been argued that the evidence for this assumption is thin.1063 In 1977, amateur-archaeologists reported large tuff stone blocks and wall remains at Kessel. Among the remains that had come to light during dredging operations at Kessel were blocks of a foliated frieze, a fragment of a column and a fragment of a Corinthian pilaster capital.1064 Roymans infers from these architectural elements the existence of a Gallo-Roman temple at Kessel. He assumes that this temple, too, was dedicated to Hercules Magusanus.1065 Nevertheless, the facts are less unambiguous than they have been presented. In this paragraph I will discuss what we can deduce with certainty from the inscriptions to Hercules Magusanus. Apart from two controversial inscriptions from Utrecht (see hereafter), the EDCS database yields 17 inscriptions: 13 from Germania Inferior and Gallia Belgica, two from Dacia, one from Britannia, and one from Rome.1066 Six dedications come from the study area, 11 from find-spots outside the study area, from sites with a Roman fort or near a Roman fort.1067 Eight dedications were made by soldiers or veterans of which the names do not provide any clues about their origins. The same applies to the appellation of the auxiliary units in which they served, the ala II Pannoniorum (AE 1977, 704), the ala Tungrorum (CIL 07, 1090) and the numerus Brittonum (CIL 13, 8492). The first inscription dates between 151 and 250, the second between 138 and 161 and the third between 201 and 300. Although units were levied among ethnic groups in the early Imperial period, recent research has demonstrated that from the pre-Claudian period onward increasingly more non-ethnic recruits served in a unit with an appellation other than their own origin.1068 The assumption that the summus magistratus from Ruimel was a Batavian is based on his father’s name, Vihirmas, and the dedicator’s office, magistrate of the Batavian civitas.1069 Vihirmas is a Germanic or Celtic name, but we cannot ascertain that it is a Batavian name. It also raises the question whether a Batavian identity was a prerequisite for becoming the highest magistrate of the civitas Batavorum. It is unlikely that the Romans would appoint a Batavian to this office after the Batavian Revolt, which is why Bogaers dated the altar before 69.1070 If the altar was erected at a later date, the magistrate most likely was not a Batavian, or

1062

See for the cult of Hercules/Hercules Magusanus within

god of the Ubians, based on the inscriptions from Bonn:

the auxilia also Haynes 2013, 232-235.

Bauchhenß 2008, 94-96. Detector Magazine published a

1063

See for Elst 2.3.4; for Empel 2.3.5.

bronze ring, found by a detectorist. Therefore, the find-

1064

Roymans 2004, 134-137 and figs.71.7-7.18, 7.20-7.22.

spot is not mentioned. Apparently, the inscription reads:

1065

Roymans 2004, 134, 144.

HER MAG: Detector Magazine 175, 25 (March 2021). The

1066

The reading of an inscription from Pannonia Superior,

dedicants of the inscriptions to Hercules Magusanus from

modern Neudorf an der Donau (CIL 03, 13391) is uncertain and, therefore, has not been taken into account. The M

Herwen-Hemeling are soldiers, too. 1067

Two come from sites in the hinterland of Nijmegen,

after Herculi has been interpreted as Magusano in the

Waardenburg and Elten. The remaining four come from

EDCS and as Marcus by others. The 17 remaining inscrip-

Empel (North Brabant), Houten (Utrecht), Ruimel (North

tions include: Empel (AE 1990, 740 = AE 1994, 1281 = AE

Brabant) and Westkapelle (Zeeland): Derks 1998, 249.

2000, 101), Waardenburg (AE 1994, 1282), Houten (AE 1994,

Roman forts outside the study area are Bonn, Xanten, Mumrills, Gherla, Cologne-Deutz.

1284), Ruimel (CIL 13, 8771), Elten (CIL 13, 8705), Domburg (CIL 13, 8777), Bonn (CIL 13, 8010), Bonn ( AE 1971, 282),

1068

Zandstra 2019, 84-86, 88-90.

Cologne-Deutz (CIL 13, 8492), Cologne (CIL 13, 10027, 212c),

1069

Bogaers 1960, 268; Derks 1998, 84-85.

Xanten (CIL 13, 08610), Xanten (AE 1977, 570), Tongeren

1070

Bogaers 1960, 270. See for a date to the first half of the 1st

(CIL 13, 10027, 212a), Mumrills (CIL 07, 1090, RIB 2140),

century Roymans 2004, 64 and note 203.

Rome (CIL 06, 31162), Ciumafaia (AE 1977, 702), Gherla (AE 1977, 704). Bauchhenß discusses the same inscriptions, contra Horn, who argued that Hercules Magusanus was a

139

at least not for some time after 69. Before becoming the highest magistrate of the civitas Batavorum, Flavus must have had a career in the military, where he learned Latin and latinised or changed his name into Flavus, which means light-haired. His ‘Romanised’ identity was undoubtedly one of the reasons why the Romans appointed him to the office of magistrate. All this, however, does not prove that Flavus was a Batavian. Only one of the 17 votive inscriptions explicitly mentions the origins of the dedicators: the altar from Rome, erected by Batavian and Thracian cavalrymen. Derks has argued that dedicants seldom refer to their origo in dedicatory inscriptions, but that collective dedications are an exception to this ‘rule’.1071 This then paves the way for his theory that the five inscriptions to Hercules Magusanus in the study area were made by Batavians and that Hercules Magusanus was their tutelary god.1072 What speaks against Derks’ argumentation is that the region where the inscriptions were found was scattered with Roman forts and associated extramural settlements, where soldiers and veterans from different parts of the Empire were housed for a shorter or longer period. The territory of the Batavians was also close to the territories of other tribes, such as the Cananefates, Tungri, Ubii and Menapii. Moreover, so far, Hercules Magusanus is absent in inscriptions from the civitas capital of the Batavians, Ulpia Noviomagus. The Hercules cult already existed in the Roman army before troops settled along the Rhine, because the Hercules statuette from Velsen (North Holland) entered the study area between 15 and 28.1073 When Germanic and/or Celtic soldiers became acquainted with his cult in the army, they bestowed the epithet Magusanus upon the hero. The origin and meaning of the epithet Magusanus are uncertain.1074 If Magusanus was a Germanic or Celtic deity and not just an epithet such as Victor, he probably had capacities comparable to those of the Roman Hercules. Hercules Magusanus retained the iconography of Hercules, which appears from the statues and altar discussed above. Which Germanic and/or Celtic tribes may have worshipped Magusanus is unknown. The inscriptions from the study area were found in and near the limes zone, which was a melting pot of ethnic identities. The majority of the inscriptions testify to the military identity of the dedicants. Therefore, the conclusion seems justified that Hercules Magusanus was first and foremost venerated by soldiers and veterans whose origins in most cases remain hidden from us. Whether the sanctuaries at Empel, Elst and the alleged temple at Kessel were dedicated to Hercules Magusanus is debatable and based on one inscription and one statuette (Empel) and on one small bronze fragment (Elst). 1075 Even if there once stood a temple at Kessel, there are no archaeological indications for a dedication to Hercules or Hercules Magusanus. As Bogaers already noted in 1960, there is no evidence for the assumption that Hercules Magusanus was the tutelary god of the Batavi.1076 3.8.6

Two controversial dedications

Two dedications to Hercules Magusanus from the study area do not occur in Derks’ overview.1077 They came to light during excavations in 1929 by Van Giffen at Domplein in Utrecht’s city centre, where once stood a Roman fort, Traiectum. Two large and three smaller fragments of a sandstone slab were found, covering a Medieval tomb. The fragments turned out to be inscribed.1078 The two largest fragments were published by Wilhelm Vollgraff in 1931.1079

1071

Derks 2004, 54 and 68, table C.

1072

Roymans/Derks 1994, 32-33 and table 1; Derks 1998, 98, 112.

1073

See for the site Velsen 2.4.2.

1075

See for the site Elst 2.3.4, for Empel 2.3.5.

1074

See, for instance, Toorians, who has advanced that the

1076

Bogaers 1960, 271.

Batavians were a Celtic-speaking group that shifted to

1077

See Derks 1998, 249.

Germanic after they settled on the left bank of the Rhine

1078

Hoekstra/Wynia 1997, 125-168, especially note 25.

around the middle of the 1st century BCE. He argues that

1079

Vollgraff 1931.

140

the epithet Magusanus is a Germanified and Romanised Celtic name: Toorians 2003-13-28.

The reading of the inscription is problematic because of the ligatures and numerous curved lines which probably were engraved at a later date in the surface of the stones. Vollgraff believed to see dedications to Germanic and Celtic gods, among them two to Hercules Magusanus. His interpretation was immediately rejected by Byvanck, who stated that the inscription was ‘a mystification’.1080 Despite the controversy, the two dedications to Hercules Magusanus were published in the Année Epigraphique.1081 Bogaers took up the study of the inscription and gave lectures about the subject, but never dedicated an article to it. It appears from his archives that he asked the opinion of several colleagues from the Netherlands and abroad, stating that he was intrigued by the inscription, but that he did not believe in Vollgraff’s interpretation.1082 In 2017, the author of an article about the inscription attempted to shed a light on its date.1083 The preliminary results of scans with infrared light suggest that the Latin inscription dates to the early Medieval period. Its content is still unclear. It might be a dedication related to the foundation of the St. Salvator church at Domplein, but more research is needed to substantiate this presumption. 3.8.7

Venus and water nymphs

North of the Rhine four Venus statuettes have been found. They include two terracotta and two bronze specimens. The bronze specimens measure 168 and 310 mm in length. The first was discovered in a terp in West-Friesland, the other in the river basin of the IJssel at Heerde (Overijssel). The two terracotta specimens come from Noordbarge (Drenthe) and Wehl (northern Gelderland).1084 The 62 Venus statuettes found south of the Rhine comprise 59 terracotta specimens, one made of lead and two bronze specimens. The goddess outnumbers all deities and human figures.1085 The eight Venus statuettes from cemeteries come from two burial sites at Nijmegen. It results from the analysis that Venus is seldom depicted in bronze. The three metal specimens, two come from Nijmegen.1086 The lead Venus statuette has been recovered from Kops Plateau. From the same site comes the earliest known terracotta figurine in the study area, a crouching Venus.1087 The statuette ended up on Kops Plateau before 15.1088 The Venus statuettes comprise several types, such as the Venus pudica and Venus Cnidia, Venus Anadyomene, Venus with Amor, Venus standing in an aedicula and a crouching Venus. Many types hail back to Roman copies of famous Greek statues. Venus Anadyomene was a painting made by Apelles which is lost, but has been described by Pliny the Elder.1089 A well-founded explanation for the overwhelming popularity of Venus cannot be given here. Besides being the embodiment of love, beauty and sexuality she may have been worshipped as bringer of fertility, healing, as a goddess associated with water and as a mother goddess. The latter will be discussed in 3.8.9. A problematic category consists of a rare group of semi-nude women holding a jar or leaning on a jar from which water streams. In general, statues and statuettes of nude or semi-nude women with a jar are rare.1090 They have been interpreted as water nymphs or as Venus holding a jar. Van Boekel has classified them

1080

Byvanck 1932/1933, 193-198.

1087

1081

AE 1977 (1981), 142, 539-540.

thorn-puller, as has been suggested by Van Zalinge 1997,

1082

See for a linguistic comment on Vollgraff’s reading of the

23. The left hand on her left thigh may be Amor’s hand, or

Because of the hand, the statuette cannot depict a

inscription Toorians 1994.

her own (figs. 2.144-145a). The hand is rather large, which

1083

Moed 2017.

makes the first option, the hand of the child Amor, less

1084

See 2.2.4.

1085

See also table 1.

1088

Van Zalinge 1997, 25.

1086

The statuette from Heerde was found North of the Rhine:

1089

Plin. HN 35.91.

2.2.4 and fig. 2.50.

1090

Schmidt 1997, 197-203 and nrs. 11, 70, 96.

likely.

141

as a separate group: ‘Water nymph or Venus’.1091 The terracotta statuettes from the study area lean on a jar supported by a column. Besides water nymphs, young nude or semi-nude women may also depict personifications of rivers, lakes and springs.1092 Venus, too, is often associated with water. The goddess’ connection with water could be related to the way she was born. According to Hesiod (8th-7th century BCE), she was created from or born of the sea foam, the type of representation known as Venus Anadyomene.1093 In Gaul and Britain, numerous Venus statuettes have been recovered from streams and springs. It has been argued that the figurines represent an unknown native goddess in classical guise, who was equated with Venus because of the similarity of their functions. Therefore, she has been referred to as ‘pseudo-Venus’.1094 Also other connections of Venus with water are known. One of Venus’ many epithets is Cloacina, Venus the Purifier. Apparently, her shrine stood in the Forum Romanum, adjacent to the Cloaca Maxima.1095 Whoever the statuettes may depict, we can only observe that the distinction between Venus and water nymphs became blurred. In the study area this is illustrated by a bronze statuette from Naaldwijk-Hoogeland (figs. 2.52-54). It depicts a semi-nude young woman holding a jar in one hand while she presses the water out of her hair with the other. She usually does this with both hands, sometimes with one hand.1096 Because of this gesture, the bronze statuette has been interpreted as Venus.1097 The statuette is an applique that once decorated a large object, like a bronze basin for water. The connection with water is also suggested by the find-contexts of figurines that have been interpreted as Venus or water nymphs. The applique comes from the filling of a well. A terracotta statuette head of a water nymphs was found in a ditch at Ouddorp and a complete specimen in a ditch at Leiden-Roomburg. It will be advanced in chapter 4 that they may have been ritual deposits.1098 3.8.8

The iconography of mother goddesses

The first terracotta figurines of mother goddesses probably reached the study area in the late 1st century.1099 The 41 statuettes of mother goddesses found south of the Rhine include 39 terracotta and two bronze specimens, the latter two being almost identical.1100 Thirty-one come from military and non-military settlements (76%). Five specimens (12%) have been recovered from cemeteries and five (12%) from sanctuaries or alleged cult places. Günther Schauerte has pointed out that of 1000 terracotta mother goddess statuettes in Germania Inferior, Gallia Belgica, the northern region of Germania Superior, Raetia and Noricum, 70% have been recovered from established find-spots. 33.5% come from sanctuaries, 15.7% from military and non-military settlements, 14.9% from cemeteries and 5.9% from pottery workshops.1101 Although the numbers in the study area are too small for a valid comparison, the distribution pattern differs. Schauerte has observed that many terracotta statuettes of mother goddesses and Venus were offered at sanctuaries and sacred places in Gallia Belgica,. In Bretagne, France, hundreds of terracotta figurines of mother goddesses of the dea nutrix type and Venus were offered at places with dolmen, megaliths. In Ubian territory they are almost absent in

1091

Van Boekel 1987, 568-571 and nrs. 124-128.

1092

Which is why Schmidt describes nr. 11 and its copies as

1098

See for Naaldwijk-Hoogeland and Ouddorp 2.3.2; for Leiden-Roomburg 2.4.1; for ritual deposits 4.2.2. See also

‘vermutlich Brunnenfiguren’: Schmidt 1997, 197.

Van Boekel 1987, 571.

1093

Hes. Theog. 190-200.

1099

1094

Van Boekel 1987, 353-354.

Jenkins 1958, 60-63.

1100

1095

See for bronze mother-goddesses 3.5.4. See for the bronze

Plaut. Curc.470-475; Hopkins 2012, 96.

specimens from Nijmegen and Koudekerk figs. 2.55-56a.

1096

Schmidt 1997, 206-208 and nrs. 132, 138-139, 141a-b, 145-146,

The figurine from Koudekerk was buried on the left bank

151, 165-166. 1097

Goossens 2012, 295 and fig. 7.13.

142

of the Rhine: see also 2.3.2 and 4.2.2. 1101

Schauerte 1987, 90-91.

sanctuaries. Schauerte has argued that this may be a consequence of different Gallic and Germanic offering practices.1102 The find-spots of several statuettes of Venus, Fortuna, Victoria and mother goddesses in the study area and Britain suggest that they were deliberately deposited in significant places.1103 The statuettes of mother goddesses in the study area comprise three triads and 34 single women, which corresponds with the predominance of single mother goddesses in Gaul and Germany. It was once thought that only triads of women represented mother goddesses or matronae, but this view has changed.1104 The importance of the number three was already a common feature in Indo-European society and also plays an important role in Celtic, Greek and Roman mythology. The Moirai/Parcae and the Erinyes/Dirae where triads, just like the Gorgones and the Sirens. The Irish-Celtic pantheon had several triads of gods, like the three craftsman gods or the three female personifications of Ireland. The repetition of number probably had the function of intensification.1105 The incongruence in number and iconography of some representations shows that one figure could stand for three and vice versa. An altar relief from Bonn depicts one matrona, whereas the accompanying inscription is addressed to the matronae Aufaniae. On an altar from Colijnsplaat three seated woman are depicted, while the inscription is addressed to the goddess Nehalennia in singular form.1106 Triads of mother goddesses sit on a bench. Sometimes, the woman in the middle does not wear a bonnet and has her hair hanging down. Therefore, it is thought that she is younger than the other two women and that the three women represent three generations.1107 Figurines of single mother goddesses from the study area predominantly depict women sitting on a chair with a high backrest, holding a basket with fruits or a dog in their lap. Several wear an Ubian dress or bonnet. Others wear diadems, either crescent-shaped or triangular.1108 A few suckle one or two infants, which is why this type is known as the dea nutrix type. Statuettes of standing women with an infant are rare in the study area. They are represented by two specimens, holding a child in one arm and a jug in the other.1109 The jug suggests a link with a water nymph or personification of a well, river or stream.1110 A figurine from Valkenburg-Marktveld possesses a rare iconographic element. It depicts an enthroned mother goddess in Ubian dress. At her right stands a small figure, wearing a hooded cloak. He holds an object in his right hand that looks like a bow. The band across his chest could be the strap of a baldric or quiver.1111 Sometimes mother goddesses are depicted with children, holding one child in her arm, while the other stands next to her.1112 Yet, this mother goddess does not hold a child in her arm, and the small figure does not stand on the ground but on a base. The man wears a cloak, which could imply that he is not Roman, but not necessarily so.1113 Lehner has suggested that the man represents the dedicator.1114 If Lehner is right, his short stature could refer to his humble status as worshipper. The depiction of worshippers in sculpture is a well-known phenomenon in the Roman Empire. In the Roman West, examples of monuments where besides deities processions with worshippers are depicted are a pillar erected by vicani in Jouars-Pontchartrain (Île-de-France) and the Pillar of the Boatmen in Paris,

1102

Schauerte 1987, 91 and note 143.

1107

Schauerte 1987, 68; Van Boekel 1987, 104, fig. 86 and nr. 74.

1103

See 4.2.2.

1108

See for a detailed description of the iconography of mother

1104

Schauerte 1987, 58-61.

goddesses Horn 1987 36-37, 43-47; Schauerte 1987, 56-81;

1105

Green 1992c, 214.

Van Boekel 1987, 433-457. See for altars dedicated to matro-

1106

Horn 1987, 44-45 and notes 55-56; Stuart 2013, 31; for the

nae from Cologne Thomas 2014.

inscription AE 1997, 1159. The dedication is in singular form

1109

Van Boekel 1987, nrs. 90-91 and 453, fig. 102.

and the three women do no possess the usual iconography

1110

See also Schauerte 1987, 81-82, and 3.8.7.

of Nehalennia. Therefore, Stuart suggests that the person

1111

Van Boekel 1987, nr. 76.

who engraved the inscription only knew the formula Deae

1112

See, for instance, Von Gonzenbach 1986, Taf. 5, nr. 1, Taf. 7 nr. 7.

Nehalen(n)iae. Since the altar relief depicts three mother goddesses he thinks the altar was brought to Zeeland from

1113

See Rothe 2009.

Cologne.

1114

Van Boekel 1987, 470.

143

discussed in this chapter.1115 In the Rhineland, worshippers can be seen on altars dedicated to matronae and other deities.1116 An altar dedicated to the matronae Aufaniae, dated to the second half of the 2nd century, depicts not only the three matronae in the upper register, but in the lower register also the man and woman who have erected the altar: Quintus Caldinius Celsus and his wife, in Ubian dress.1117 A variation of this practice may be the deposition of two bronze statuettes at Ede-Veldhuizen. The first statuette depicts a male worshipper, the second the goddess Minerva. The worshipper has bare feet and wears a long tunic, which is wrapped around his body, leaving one shoulder bare. He holds a patera in his left hand and probably held another object in his missing right hand.1118 Both statuettes were found in a layer above a well that belonged to a settlement that was occupied between the 2nd and 5th centuries. The well was in use in the 3rd century.1119 The layer above the well contained no other finds. The statuettes were probably deposited in the layer in the 5th century, which by then had turned into a pool, since the area had become very wet.1120 Although it is uncertain whether the two were deposited at the same time, the similarities between the two statuettes and the absence of other finds suggests that they were deposited together. An explanation for this phenomenon could be that the person(s) who offered the statuettes identified himself/ herself /themselves with the worshipper. By offering the two statuettes, a special bond was forged between the worshipper(s) and the deity. It reinforced the power of the offering, which now had become eternal. 3.8.9

Venerated as mother goddesses

It has been argued that Fortuna, Minerva, Juno and Cybele were venerated as mother goddesses.1121 The examples that will be discussed in this paragraph support this idea. We have seen that mother goddesses and matronae are usually depicted seated while holding a child, fruits or a dog in their lap. Minerva does not share the attributes of Fortuna or mother goddesses, but Van Boekel already observed that in the Rhine-Moselle area the goddess is more often depicted enthroned than standing.1122 Large statues of Fortuna and Minerva in the centre of the Empire usually depict the goddesses standing, holding attributes in their hands. 1123 The standing pose also applies to nearly all bronze statuettes of Fortuna and Minerva from the study area.1124 The enthroned pose is typical for prehistoric mother goddesses such as Cybele who may have been depicted as early as the beginning of the 6th millennium BCE (figs. 3.13-3.14). Enthroned Roman period mother goddesses, now decently dressed, still echo the iconography of the nude, seated goddess from Anatolia. As to terracotta figurines of Cybele, Van Boekel distinguishes three types: Cybele enthroned, Cybele riding a lion and Cybele standing.1125 Seven of the eight terracotta statuettes from the study area belong to the first type, Cybele enthroned.1126 The only bronze specimen from the study area depicts a standing Cybele.1127 Several representations of Fortuna, Juno and Minerva in terracotta and stone depict the goddesses seated. Famous for his statuettes of mother goddesses was the modeller Pistillus, who had a workshop in Autun,

1115

See 3.8.2.

1116

See for terracotta statuettes of worshippers in the Roman

1123

Huiusce Diei from Largo Argentina, Rome, and depictions

provinces, for instance, Von Gonzenbach 1986, Taf. 55, 1, (Allmendingen, Switzerland), 125, 1-2 (Cologne). 1117 1118

Exceptions are the large cult statue of the seated Fortuna of the Capitoline trias, with a seated Juno and Minerva.

1124

Carroll 2013, 223-224.

An exception is a bronze seated Fortuna statuette from Nijmegen: Zadoks et al. 1973, nr.6.

A comparable statuette was found in Hainaut, Belgium:

1125

Van Boekel 1987, 276.

Faider-Feytmans 1979, 91 and plate 57, nr. 95.

1126

Van Boekel 1987, nrs 14-19 and the statuette from Ooster-

1119

Taayke et al. 2012, 128; Zadoks/Peters 1976, 163.

1120

See for this site also 2.2.4.

1121

Kyll 1966, 42-43; Schauerte 1987, 88.

1122

Van Boekel 1987, 408.

144

hout (fig. 2.60). 1127

Zadoks et al. 1969 nr. 17.

France.1128 He made a statuette group of three seated Fortuna’s on a bench, each holding a cornucopia and patera.1129 This representation alludes to the three mother goddesses on a bench.1130 Two fragmentarily preserved statuettes of Fortuna from the study area depict the goddess with a child standing beside her.1131 Since the children are not winged, they do not represent Eros or another divine being. These children may also refer to Fortuna’s capacity as protector of children and the family. In this respect, Nehalennia cannot remain unmentioned. Altars dedicated to the native goddess often depict her seated in an aedicula, sometimes holding fruits in her lap with a dog and a basket filled with fruits beside her.1132 Like Fortuna, Nehalennia is occasionally depicted with a rudder or a cornucopia.1133 An altar from Domburg dedicated to ‘the Neihalenninae’ depict three women seated on a bench.1134 An iconographic element that Nehalennia does not share with mother goddesses and the Roman goddesses Fortuna, Juno and Minerva is the short shoulder cape, a garment considered typical of the coastal area of Zeeland.1135 The shared characteristics and symbols of prosperity, fertility and domesticity suggest that the native goddess, besides her capacity as protector of travellers, possessed the qualities of a mother goddess. It appears from the combination of attributes that the iconography of Nehalennia was ‘modelled’ on representations of mother goddesses and combined with Roman elements such as the aedicula and cornucopia. The sanctuaries at Domburg and Colijnsplaat (Zeeland) were probably simultaneously in use from the end of the 2nd until the middle of the 3rd century.1136 Although the cult of Nehalennia may have existed before the Roman conquest, the iconography and date of the sanctuaries suggest that the cult flourished between the late 2nd and middle of the 3rd century. The exchange of iconographic elements demonstrates how they blended together to create new representations. Apart from the three bronze mother goddesses that have been discussed in this chapter, the process of creolisation in the study area mainly concerns sculpture in terracotta and stone.1137 Bronze statuettes of Roman goddesses hardly underwent any changes. Venus does not share the enthroned pose of mother goddesses and she is not depicted with the aforementioned attributes, but her nudity is reminiscent of the mother goddess archetype, which may have been sufficient for native peoples to see in her a mother goddess and/or bringer of fertility. 3.8.10 A horse-riding acrobat During excavations at the St. Josephhof in the city centre of Nijmegen in 2008, a remarkable bronze statuette came to light (fig. 2.83).1138 It was found under the collapsed roof of a house which was burnt down during the Batavian Revolt in 69/70. The figurine is cast solid and measures 63 mm in height. It depicts a man standing on a flat plinth with a hole in the middle. Apart from a loin cloth and brimless cap he is nude. He holds a stick or club in his raised right hand. His ribs are indicated, which implies that he is skinny. Underneath the plinth, there is an eye at each side for leather cords, of which one fragment has been preserved. The statuette is not unique. Comparable specimens have been found at Tongeren (Belgium), Xanten, Bonn (Germany), Canterbury, Wroxeter, Colchester (Britain), Clermont-l’Hérault and Vaison-la-Romaine (France).1139 They vary 1128

Excavations in 2010 yielded one of his kilns with moulds

1134

Stuart 2013, 31 and nr. 23 (H16). CIL 13, 8798 = AE 2014, 888.

bearing his signature: https://humanhist.com/association/

1135

Hondius-Crone 1955, 102; Carroll 2013, 227-228 and fig. 4.

atelier-et-fours-de-pistillus/ (30-5-2018).

1136

See 2.3.2.

1129

Van Boekel 1987, 310.

1137

See 2.3.2 for the disputed interpretation of a bronze ‘Neha-

1130

See for the iconograpy of mother goddesses 3.8.8.

1131

Van Boekel 1987, 315 and nrs. 35-36.

1132

See, for instance, the catalogue of Stuart 2013: 5-30, 33, 36-43, 45-53, 56, 58-61.

1133

Stuart 2013, 30.

lennia’ from Cologne. 1138

This paragraph is the abridged version of an earlier publication: Veen 2010.

1139

See for a complete description of these figurines: Hanel/ Cerdán 2012, 339-352.

145

in height from 52 to 72 mm and can be dated to the second and third quarter of the 1st century, based on the find-spots.1140 In the past, the statuettes have been interpreted as Hercules, Mars, or one of the Dioscuri.1141 More recently, the figurines have been identified as desultores or cursores, horse-riding acrobats who performed their feats of daring in horse and chariot racing during circus games.1142 Roman period images of desultores occur since the 1st century BCE on oil lamps, coins, reliefs and sarcophagi. From Nijmegen come two oil lamps with desultores. Here we see how they are usually depicted: riding a galloping horse, holding a stick or whip in one hand, often leading a second horse by the reins. They usually only wear a loin cloth and a cone-shaped, brimless cap, a pileus. In ancient Greece, the pilos was a felt travelling hat. In the Roman period the pileus was given to slaves during the ceremony of their liberation. Therefore, it is not unlikely that desultores were liberated slaves. The proceedings of horse races during circus games have been described by the archbishop and scholar Isidore of Seville (560-636).1143 There were two varieties of the race. The desultor jumped off his horse during the race and ran the last part towards the finish, or he jumped off his galloping horse and jumped on another to finish the race. The second variety probably came via Greece to Etruria, where horse and chariot races were held with jockeys and men jumping on and off chariots during the race.1144 They are depicted in wall paintings in Etruscan tombs, such as the Tomba del Colle Casuccini in Chiusi and the Tomba delle Olimpiadi in Tarquinia, the first dating to the beginning and the second to the end of the 6th century BCE.1145 Etruscan images also testify to the military origins of the sports, like a wall painting in the Tomba delle Bighe and terracotta plaques from Cerveteri, which depict a mounted cavalryman leading a second horse by the reins. The rider could replace his horse during battle, when it was tired or wounded. Evidence from written sources like Caesar and Florus demonstrates that warriors jumping on and off horses during battle was a practice also known to Gallic and Germanic warriors.1146 Florus tells us how the Teutonic king Teutobod went to battle with four to six horses and jumped off one horse onto another when it got tired.1147 The military origins of the sport and the presumed non-Roman descent of desultores could be a reason for their popularity among soldiers based in Germania, Gallia and Britannia. Horse races and other games were probably organised to entertain soldiers in between fights and other tasks. The early date of the statuettes excludes any link with the monumental stone-built circuses and amphitheatres that were constructed in the 2nd century at Xanten, Trier, Arles and Colchester, where a circus was discovered in 2005.1148 Nevertheless, there must have been simple predecessors. After all, for horse or chariot racing only a flat piece of terrain and wooden tribunes or seats are needed, which often leave no distinct traces in the archaeological record.

1140

1141

This concerns the deposition date. Whether the figurines

adsuefecerunt, ad quos se celeriter, cum usus est, recipiunt;

were made in the same period cannot be established.

neque eorum moribus turpius quidquam aut inertius habetur,

Reinach 1894, 131, nr. 136; Rolland 1965, 69; Henig et al. 1987, 360-362.

1142

1143 1144

quam ephippiis uti. 1147

Flor. Epit. 1.38.10: Certe rex ipse Teutobodus quaternos

Hanel/Cerdán 2012, 488. This also explains why the man is

senosque equos transilire solitus, vix unum, cum fugeret,

thin: Just like a modern jockey, a desultor had to be small

ascendit, proximoque in saltu conprehensus insigne spec-

and light.

taculum triumphi fuit. Quippe vir proceritatis eximiae super

Isid. Orig. 18. 39.1. Although of later date, a frieze depicting an apobates

tropaea sua eminebat. 1148

At Nijmegen, too, traces of an amphitheatre have been

race is known from the Parthenon. See for this frieze, for

found, which have not been published yet. It was probably

instance, Neils/Schultz 2012.

constructed somewhere between 70 and 100, a period

1145

Thuillier 1989, 33-39 and figs. 1-2.

that coincides with the arrival and departure of the Tenth

1146

Caes. BGall. 4.2.3: Equestribus proeliis saepe ex equis desiliunt

Legion.

ac pedibus proeliantur, equosque eodem remanere vestigio

146

Now the iconography of the figurines is clear, the question remains how people used them and to what end. The statuettes from Nijmegen, Bonn and Canterbury have bronze eyelets under the plinth, for leather straps. As mentioned earlier, the figurine from Nijmegen still has a fragment of leather strap in one of the eyelets. The straps were probably used to attach an object, like tintinnabula, bells, as has been suggested by Norbert Hanel.1149 Tintinnabula had a ceremonial and/or apotropaic function and were attached to objects and sculpture.1150 The preserved fragment of a wooden pin in the hole of the plinth under the statuette from Bonn makes clear what the function was of the round hole in the plinth. It implies that the statuettes could be held in the hand or placed on or in an object. If the figurines had a ritual function, they may have been set up in or on a shrine, either by fans of the riders, asking the gods to let their favourite win, or by the desultores themselves, hoping to survive and win the race. 3.8.11

Jupiter-Ammon in Ulpia Noviomagus

Among the fragments of the bronze hoard from Ulpia Noviomagus a large, heavy piece represents the head of a male. The fragment, 170 mm in height, weighs 1482 grams (fig. 2.95).1151 It is approximately one eighth part of a life-sized head. The horn, ear and curls were cast solid. The solid cast parts and the unusual thickness of the walls (6-11 mm), explain the weight of the head. The surface is partly covered by a greyish brown crust. The brown spot under the ear is corrosion from an iron pin stuck into the fragment, possibly one of the chaplets that were stuck into the clay core before casting.1152 The fragment represents a part of the right side of a male head, with an animal’s ear, a ram’s horn, thick, curly hair, a beard and a diadem or ribbon around his forehead. The curly hair, beard, ram’s horn and diadem make it easy to identify the head as Jupiter-Ammon, because it is the usual iconographic representation of this god.1153 First, a brief explanation is required about the way the Egyptian god Amon, Hammon or Amun, assimilated with the Roman god Jupiter. Greek colonists founded the colony of Cyrene in North-Africa in the 7th century BCE, incorporated the Egyptian god Amon into their religion and transformed him into Zeus-Ammon. The cult penetrated into the Peloponnesus and round 460 BCE, the poet Pindar introduced the god into the Greek city of Thebes. In 331 BCE, Alexander the Great undertook a pilgrimage to the Ammon oracle in Siwa, in the Libyan Desert. Afterwards he issued coins depicting himself with horns. The god remained popular with Alexander’s successors for a while, but in Greece the cult lost its appeal. The earliest known image of Jupiter-Ammon in the Roman Empire comes from the Forum of Augustus in Rome.1154 The marble clipeus-decorations in the porticoes of the Forum referred to Alexander the Great, who after the battle of the Granicus River in 334 BCE hung the captured shields of his Persian enemies on the walls of the Parthenon in Athens.1155 In Rome, Marius placed a shield of the defeated Cimbri in the Forum Romanum in 101 BCE.1156 The barbarian with torc on the clipeus-decorations in Augustus’ Forum

1149 1150

that was published in 2014: Veen 2014a.

Hanel 1995, 83. Dierichs 2010. An example of a bronze calyx Mercury bust

1152

with tintinnabula comes from Orange, France: Rolland 1965, 53, nr. 65. Hans Jucker mentions three other Mercury

tusch 1996a, 10-16; Mattusch 1996b, 20-26; 3.13.1. 1153

Sometimes the god wears no diadem and has human ears.

calyx busts that originally had tintinnabula. Since the

Most representations of Ammon as a beardless young man

tintinnabula were missing, it was believed that the chains

refer to Alexander the Great and his successors: Leclant/ Clerc 1981, 688-689.

were for lamps or weights. The three busts come from Gaul or the Rhineland: Jucker 1961, 161-162. A calyx bust of Jupi-

1154

The temple was dedicated in 2 BC, but the forum itself was already in use in 5 BC: Spannagel 1999, 20.

ter Ammon from Vieu-en-Valromey (Ain) probably also 1151

See for a detailed explanation of the casting process: Mat-

had tintinnabula attached to it: Veen 2014b, 137 and note 57.

1155

Goldbeck 2015, 29 and note 118.

See also 2.3.4. This paragraph is a summary of an article

1156

Cic. de Orat. 2.266.

147

probably referred to the conquest of Gaul, the head of Jupiter-Ammon to the conquest of Egypt.1157 Soon, the clipeus-decorations were imitated in fora in the Roman provinces. Here, Augustus’ visual language was perceived differently and the clipeus-decorations became a symbol for the imperial cult.1158 Later, also this connotation lost its meaning and clipei became a purely decorative motif. The depiction of the god on militaria like phalerae and pteryges can be explained by the original apotropaic quality of the god, which is attested by hymns on the walls of Egyptian sanctuaries, stating that the god Amon ‘gives back life to those who live in fear, brings joy to wounded hearts, protects from evil and diseases and rescues those who are lost at sea’.1159 The apotropaic quality of Jupiter-Ammon in the Roman period appears from his representation in combination with Medusa, whose apotropaic qualities were renowned throughout antiquity.1160 On funerary altars and urns the head of Jupiter-Ammon is depicted on the corners and protruding elements, which are the most fragile parts.1161 Apart from protecting the object itself, the god also protected the dead on their journey in the afterlife. The army stimulated the spread of foreign cults.1162 A likely candidate for the spread of the Jupiter-Ammon cult was the Legio III Cyrenaica, since it was stationed for 130 years in Egypt to keep the peace after Augustus had conquered Egypt in 30 BCE. In 106, the legion was redeployed from Egypt to Bostra (Syria). Inscriptions demonstrate that the legion ‘imported’ their gods Ammon-Serapis or Zeus/Jupiter-Ammon. Coins with the head of the god were issued between the middle of the 2nd and the middle of the 3rd century, indicating that the god had become the tutelary god of the metropolis.1163 The presence of a vexillatio of this legion in Lower Germany between 200 and 270 is attested by an inscription from Iversheim (Nordrhein-Westfalen, Germany). The dedicator was the signifer of the army unit.1164 The dates of two representations of Jupiter-Ammon from Germania Inferior and Gallia Belgica suggest an earlier contact with the cult.1165 The fragment weighs 1.5 kilo. If it was once part of a head in three-dimensional form, it must have weighed at least twelve kilos.1166 The weight and the fact that only the head of Jupiter-Ammon was depicted in the Roman period indicates that the fragment must have been part of a herm, bust or relief.1167 There are several examples of stone herms of Jupiter Ammon, some of which are double herms, combined with another deity.1168 The popularity of Jupiter-Ammon as a herm can be explained by the original function of herms as an apotropaic medium, bearing the head of an apotropaic god.1169

1157

Zanker 1968, 13-14 and note 58. See for an overview of more recent and alternative interpretations Goldbeck 2015,

pteryges: Leclant/Clerc 1981, 677, nr. 85 and 85b-c. 1161

1158

See, for instance, Leclant/Clerc 1981, 674-675 and nrs. 55, 57, 60-61, 66b.

28-33. First in Central Italy, followed by Spain and Portugal,

1162

See, for instance, Stoll 2011, 468-473; Haynes 2013, 227-236.

France, Switzerland, the cities on the Adriatic coast and

1163

Stoll 2011, 470-471. I thank Ton Derks for bringing this paper to my attention.

North-Africa: Goldbeck 2015, 60-143; Verzàr 1977, 38-41; Simón 1990, 143-162; Cassari 1998, 391-407; Cassari 2004.

1164

Drinkwater 1971, 325-326. The inscription reads:----] Q(uin-

1159

Tran Tam Tinh 1971, 24. See also Leclant/Clerc 1981, 686.

tus) Coel[i]us? [Procu]leianus s[ig(nifer)]/ v(otum) s(olvit)

1160

Leclant/Clerc 1981, 669. See for Jupiter-Ammon and Medu-

l(ibens) m(erito) / leg(ionis) III Cyr(enaicae)[---]. AE 1968,

sa on public buildings Verzàr 1977, 35, 38; Simón 1990, 143162; Budischovsky 1973, 202-216. On funerary monuments:

392. 1165

An applique of a Jupiter-Ammon herm (LVR-Landesmuse-

Leclant/Clerc 1981, 674, 686-687 and nrs. 55-56; Grimm

um Bonn. Inv. nr. 30827) from Xanten or its environment

1969, 68-69 and 232-233, Taf. 50 and 52. Herms: a stone

dates to the 1st century. A funerary monument decorated

Jupiter-Ammon herm from Kassel, Germany, has the head

with the head of Jupiter Ammon from Neumagen (RLM

of Medusa engraved in the right shoulder: Leclant/Clerc

Trier. Inv. Nr. 11643) dates to 130. Grimm 1969, 126-127, 232233 and figs. 50, 54.3.

1981, 672, nr. 29. Statuettes: a bronze statuette of Minerva from Vienne. The goddess has a gorgoneion on her chest

1166

The fragment is rather flat, but may be deformed by heat.

and a Jupiter-Ammon ascending from her back: Budischo-

1167

Although the god is depicted frontally in small reliefs such

vsky 1973, 213. See for phalerae: Matz 1932, 12-13, nr. 5; for

148

as phalerae.

Examples of stone specimens from the region come from Lechenich (Erftstadt) and Cologne.1170 The first was a bust found in 1846, in the fields nearby the village of Lechenich.1171 The head from Cologne was found in 1953, in the city centre. 1172 It was recovered from a well in the praetorium and was not a head in three-dimensional form but part of a decorative element.1173 Alongside their protective, apotropaic role by roadsides, at crossings and on boundaries of land, herms were used as garden ornaments and architectural elements, set up in niches or in rows in galleries.1174 The two stone heads from Cologne and Lechenich have more or less the same size as the bronze head, being respectively 42 and 45 cm in height. Their resemblance to the head from Nijmegen suggests that the latter was cast in the Rhineland rather than in Gaul.1175 The numerous militaria with Jupiter-Ammon heads as well as the Jupiter-Ammon head from the praetorium at Cologne highlight the link between the god and the army. Like the head from Cologne, the bronze Jupiter-Ammon may once have decorated the garden, or a house or tomb of a military officer in Ulpia Noviomagus. 3.8.12 A mysterious lady from Vechten A terracotta fragment from Vechten, 65 mm in height, depicts the upper part of a young woman who touches her chin with the thumb and index finger of her right hand (fig. 2.173).1176 She wears a garment, a diadem and earrings. Apart from this fragment only one other example of this statuette is known, from Cologne.1177 A comparable, but not identical gesture is known from Harpocrates, the son of Isis, which refers to the hieroglyphic symbol for ‘child’.1178 The Roman authors Varro and Ovid misinterpreted the gesture, assuming that the god made the gesture because he was the god of silence and secrecy.1179 Yet, a connection between the woman from Vechten and the Isis-cult is not very likely, since she has no Egyptianizing characteristics or attributes. If the young woman depicts a goddess, she cannot be Venus, because she wears a tunic. 1168

Leclant/Clerc 1981, nrs.16-33. Double herms of Jupiter-Am-

1174

Neudecker 1988, 105. Among the numerous sculptures for

mon and Bacchus: 680, nrs. 134-135; herms of Jupiter-Am-

his villas, Cicero ordered bronze herms with Pentelic mar-

mon and a goddess: 680, nrs. 139-140; herms of Jupiter-Am-

ble shafts for his villa in Tusculum: Neudecker 1988, 12-13.

mon and Heracles: 682, nr. 150; herms of Jupiter-Ammon

In a Dionysian garden in the Casa degli Amorini dorati at

and a deity with bull’s horns: 682, nr. 156.

Pompeii (VI 16,7) stood a Jupiter-Ammon herm and several

1169

See, for instance, Neudecker 2011.

Bacchus herms: Zanker 1995, 175-176 and figs. 93-94. See for

1170

Grimm dates the herm from Lechenich to the end of the 1st or the beginning of the 2nd century and the herm from

70 stone herms from the Trier region: Wrede 1972. 1175

lier 1994, figs.1-2; Leclant/Clerc 1981, nr. 33.

Cologne to the second half of the 2nd century. Arguments to support these dates are not given: Grimm 1969, 68-69,

1176

1177

1171

1178

finger. See, for instance, the statue of Harpocrates from the

the find circumstances or a possible date: Lersch 1846, 116-

Villa Adriana at Tivoli, now in the Musei Capitolini, inv. nr.

Lehner 1905, 10 and Taf. XXXIV, 6.

1173

Harpocrates’ gesture, however, is made with the index

Laurenz Lersch discusses the find, but does not mention 121 and Taf. IV, 1-2. Hans Lehner does not give a date either:

1172

Binsfeld 1964, 29 Taf. 1. I thank Jan de Beenhouwer for the information.

Cologne: Ristow 1970, 82, nr. 28 and Taf.24, 2-2; Leclant/ Clerc 1981, 672.

I thank Marenne Zandstra for bringing this figurine to my attention.

168-169 and 174-175, nrs. 60 and 71, Taf. 48-49, 2-3. Ristow, Leclant and Clerc cite Grimm for the date of the herm from

See for representations of Jupiter-Ammon in Gaul: Cheval-

Scu 646. 1179

Var. L. 5.57: Hi dei idem qui Aegypti Serapis et Isis, etsi

Otto Doppelfeld was the director of the 1953 excavations.

Harpocrates digito significat, ut taceam. Idem principes in

He does not mention the herm in his preliminary report:

Latio Saturnus et Ops. Ov. Met. 9.690-692: cum qua latrator

Doppelfeld 1956, 83-96.

Anubis sanctaque Bubastis, variusque coloribus Apis, quique

See for the head also Grimm 1969, 168.

premit vocem digitoque silentia suadet.

149

Therefore, I propose an alternative interpretation. In the Greek world, another gesture, but not identical, is known from representations of the Muses. On an Attic hydria dating to 450-425 BCE, three muses are depicted. One plays the lyre, the other the harp, while the Muse in the middle listens while she touches her lips with the index finger of her left hand.1180 On a lekythos dating between 420 and 410 BCE, we see Erato listening to a cithara player, with her hand under her chin.1181 On a sarcophagus dating to the middle of the 2nd century, the nine Muses are depicted. Polymnia, the fifth figure seen from left to right, leans on a rock, with her right hand under her chin (fig. 3.8).1182 Except for the gesture of the finger the figurine does not provide any other clues. If the figurine from Vechten depicts a Muse, it may have been part of a group. 3.8.13 Minerva or a gladiator? From the site The Hague-Scheveningseweg comes the upper part of a statuette that has been interpreted as Minerva (figs. 2.174-77).1183 The fragment is 74 mm in height and depicts a figure wearing a helmet with lowered visor. Whereas Van Boekel argues that Minerva with lowered visor is depicted and not a gladiator, I suppose the latter is the case. This idea is based on resemblances to bronze and terracotta statuettes of Roman period gladiators, while representations of Minerva with lowered visor are unknown.1184 Van Boekel’s interpretation is based on two arguments. She sees a string of hair on each side falling over the shoulders at the front. Second, she supposes the statuette is enthroned, based on a second fragment that has not been published. To start with the latter, if the two fragments belong to the same statuette she has a point, since gladiators are usually depicted in action and not seated. Nevertheless, an enthroned Minerva with lowered visor seems just as contradictory as an enthroned gladiator. As to her first argument, for the braids or strings of hair more explanations could be advanced.1185 Gladiators were usually of ‘barbarian’ descent, and may have had long hair, worn in braids. Second, the depicted gladiator could be a woman, since there were also female gladiators.1186 3.8.14 The thorn-puller The Spinario or thorn-puller is a well-known statue by an unknown artist and of unknown date, of which several bronze and marble copies exist from the Roman period and beyond.1187 Two of the best known specimens are the bronze statue from the Capitoline Museum in Rome and the marble statue from the Castellani collection, now in the British Museum. The statue depicts a young, nude boy sitting on a rock, trying to withdraw a thorn from the sole of his left foot. Since it seems to be a representation without any religious connotation, it raises several questions, such as why Central-Gaulish workshops started to make small terracotta copies of the thorn-puller in the 1st century and how this motif changed from a ‘genre-motif’ with a purely decorative character into a representation with a special meaning. This is at least suggested by the find-context of several statuettes of thorn-pullers. A terracotta specimen was found in a pit in a sanctuary at Aalst (Belgium).1188

1180

Bonamici 1992, 661 and 385, nr. 14.

1181

Bonamici 1992, 667 and 386, nr. 81.

1182

Musée du Louvre, inv. nr. MR 880.

1183

Van Boekel 1989, 29-30. See for the site The Hague-Scheve-

1184

2016, 49, figs 6-7; Van den Hengel 2016, 84, figs. 1-2. 1185

The fragment is slightly abraded, which makes it difficult to see. It cannot be ruled out that the two strings are not locks of hair but, for instance, leather straps.

ningseweg 2.4.2.

1186

Drouen 2016, 79-81.

See for statuettes of gladiators Durham 2012, nrs. 170-171; De

1187

See for thorn-pullers Meinecke 2016.

Beenhouwer 2005, 247, 329; Koster 2016, 52, fig. 1; Hunink

1188

Van Boekel 1987, 160.

150

Although examples from Harelbeke and Asse (Belgium) come from disturbed or unknown find-spots, it is assumed that there was a sanctuary in the vicinity.1189 Thorn-puller figurines could also be funerary gifts and have been recovered from a burial site in Salzburg (Austria).1190 The only specimen from the study area comes from a non-documented find-spot at Voorburg-Arentsburg, where no burial site has been discovered so far.1191 If the thorn-puller had a special meaning for people in the northwestern provinces, there might be a link with the presumed function of busts of boys, in which case they were offered as part of a rite of passage. It is also possible that the statuettes originally had a purely decorative function, just like the original statue. It became meaningful in another way, for instance because it was owned and cherished by a loved one, and, therefore, accompanied the deceased on his or her journey to the underworld. 3.8.15 The power of being different Besides the thorn-puller, more representations in the study area present characteristics that we cannot link directly to a cult or a religious function: an ithyphallic black man, dwarfs and ithyphallic dogs.1192 They appear in different contexts in the study area. The first, an ithyphallic black man, comes from Vechten. It is a small, ca. 50 mm high bronze figurine depicting a nude black man, apparently a pygmy or dwarf, sitting on a stool (figs. 2.171-172). He had a large, erect penis, which is now missing.1193 In order to understand the development and meaning of such representations, we have to look at mosaics and wall paintings in the centre of the Empire. John Clarke distinguishes three functions of pygmies and other African men in the Roman world: they were an amusing part of a decorative scheme, they could represent the colonial Other, or they were thought to avert demons.1194 Representations of conquered peoples in the Roman world deviated in many aspects from the Roman ideal. Already in the 6th century BCE, artists depicted pygmies and African people in comic situations. On the foot of the François Vase, the famous volute krater from Etruria, pygmies battle with cranes. Whereas African men were still portrayed as normal human beings in the Palestrina mosaic, which dates to c. 100 BCE, by the end of the 1st century BCE they were often depicted as clowns. In the Augustan period, artists placed pygmies in Egyptian landscapes.1195 It was not the ethnic pygmy Roman artists portrayed, but the pygmaios, the pygmy-dwarf from Greek and Hellenistic art, to express their feelings towards barbarians.1196 In the past, scholars have tried to explain antique representations of physical abnormality by attributing them specific functions. According to some, they were medical teaching aids, while others thought that they were used as amulets to protect the owner from the depicted deformities. During the last decades, a growing corpus of studies on deformity in antiquity has seen the light, offering new views on the perception and interpretation of dwarfism and physical abnormality in ancient societies. It is, for instance, argued that children with unusual physical characteristics were seen as marked out by the gods. If they managed to survive, their deformity was seen as a sign that they were either particularly blessed or gifted with special powers and marked out for an extraordinary fate.1197 Because of their otherness, pygmies, people with dwarfism and physical deformations such as hunchbacks were attributed apotropaic powers.1198 This could explain why Roman emperors, and ‘bad’ emperors like Commodus and Elagabalus in particular, favoured the company of deformed people, usually slaves or entertainers, who could become court jesters and even imperial confidants.1199 1189

De Beenhouwer 2005, 136, 150-151, 160, 173, 832.

1195

Clarke 2005, 156-157. See also Versluys 2002.

1190

Van Boekel 1987, 160.

1196

Clarke 2005, 161.

1191

Van Boekel 1987, 158-160 and nr. 92.

1197

Milburn 2007, 1-2.

1192

See for ithyphallic dogs 3.8.18.

1198

Clarke 2005, 157-158.

1193

Oral information by Ruurd Halbertsma.

1199

Trentin 2009, 140.

1194

Clarke 2005, 155.

151

The philologist Detlev Fehling has demonstrated that ethology provides an adequate explanation for the importance of erected phalli. In the animal world, meerkats, baboons and other monkeys sit on the edges of their territory with an erect penis, to impress and scare off intruding males from other groups.1200 Apparently, the image of an erected phallus inspires feelings of protection and safety. In the Greek world, the protective and apotropaic function of the phallus can be seen in herms and statues of Priapus. At Pompeii, in the Casa del Menandro, an African man with a huge phallus is depicted in a mosaic at the entrance of the caldarium. In Campania, Rome and Ostia phalluses are depicted at entrances of baths and other buildings, on walls, bridges, corners and all places of potential danger.1201 The verb ‘to fascinate’ is derived from the Latin word fascinum, the phallic amulet that was used to distract or repel demons and the evil eye.1202 The numerous phallus amulets we find in all parts of the Empire confirm their intensive and wide-spread use. Besides these amulets, multiple phalli could be attached to tintinnabula, or to statuettes of human, animal and divine bodies, creating images which we nowadays may find strange or confusing. 1203 These images must have comforted and reassured the Roman period viewer whose world was full of danger, disease and insecurity. Finds from the northwestern provinces demonstrate that representations of dwarfs and figures with physical abnormalities were not limited to the centre of the Empire. The highest concentration of dwarf statuettes comes from the Trier area. Twenty-one specimens came to light at the Lenus Mars sanctuary at Trier. Several specimens were found in Trier-Altbachtal, Möhn, Dhronecken and Gusenburg.1204 A few others have been found in funerary contexts in Cologne, Jagsthausen and Trier.1205 A striking example of a bronze statuette with multiple phalli comes from Tongeren (Belgium).1206 It depicts a seated Mercury with a phallus on his head. His nose is a phallus, he holds a phallus in each hand and the last phallus was where a (now missing) phallus could be expected. The hole in the phallus on his head indicates that the 88 mm high statuette was a pendant for an amulet. The bronze figurine from Vechten may also have been an amulet. Also from Belgium comes a 250 mm high bronze statuette of a male figure with a distorted body, a large penis and furry animal legs. His curly hair is bound together by a shawl, folded in a way a woman would bind her hair together. Based on the pointed ears and animal legs the figure has been interpreted as Pan. The statuette was found at the compound of a villa at Willemeau (Hainaut), together with several other bronze objects.1207 The contradicting, exaggerated female and male characteristics of the depicted figure as well as the deformed body were probably not only meant to be humorous and may have had an apotropaic function. Altogether, five statuettes of dwarfs or caricatures have been found in the study area.1208 Two specimens come from Vechten: the bronze figurine of the ithyphallic African man and a well-preserved terracotta statuette of a dwarf holding a scroll. He wears a cloak and the head contains a pebble, for rattling. The scroll has been interpreted as an enchiridion, a handbook for cures and medicines.1209 It points at his literacy and knowledge of medicine and underlines his otherness, since these skills were unusual for people with a deviating physical appearance. The apotropaic and/or ritual function of the figurine can be derived from its function as a rattle. Fragmentarily preserved specimens come from The Hague-Scheveningseweg and Nijmegen. All five have been found near or in military settlements.1210 Another category of apotropaic artefacts that seem to be typical for military settlements and cemeteries are the numerous phallic amulets that have been found at Nijmegen and elsewhere.1211 1200

Fehling 1974, 8-9.

1206

Faider-Feytmans 1979, nr. 51.

1201

Johns 1982, 63-64; Clarke 1995, 131-134; Clarke 2005, 158-159.

1207

Faider-Feytmans 1979, nr.58.

1202

Clarke 2005, 159.

1208

Figs. 2.130-131, 2.171-172 and Van Boekel 1987, nrs. 24, 147,

1203

See, for instance, Johns 1982, figs. 10-14. See 3.8.10 for bronze

1204

1205

possibly also nr. 204, 1989, nr. 45.

Mercury busts with tintinnabula.

1209

Van Boekel 1987, 297-298 and nr. 24.

See for statuettes of grotesque figures in northern Italy,

1210

See also table 1.

Britain, Switzerland and France, Von Gonzenbach 1986,

1211

Zadoks et al. 1973, nrs. 77, 79-83, 85-86, 89, 96, 102. See for

Taf. 27, 1-2, 28, 1-2, 29, 1-2, 89, 2, 92, 2-4.

phallic amulets from forts in Britain: Green 1978, nrs. 140-

Van Boekel 1987, 298.

146.

152

3.8.16 ‘Exotic’ representations Not all representations of black men were meant to be humorous or apotropaic. Some examples depict Africans with normal physical characteristics. Two bronze statuettes of African men from the study area were found by detectorists in North Brabant and Gelderland.1212 The first came to light at Heeswijk (North Brabant). It is a realistically rendered African man, not a caricature, standing with a drinking cup in his left hand (figs. 2.75-76).1213 He is young, with a well-proportioned body and nude, except for a loincloth. The right arm and lower left leg are missing. The expression on his face could indicate that he is drunk, but this is uncertain. Like the figurine that will be discussed hereafter, the statuette probably was once part of a support for a bowl or drinking cup and may have been seen a curiosum, brought home by a soldier or veteran for his relatives, who probably had never seen an African man before. The second statuette was found in the vicinity of Geldermalsen (Gelderland) and depicts an African man wearing a tunic (figs. 2.58-59).1214 Both arms are outstretched forward. His hands rest on the ears of a large elephant’s head. Instead of human legs, the man has one lion’s leg. The statuette was part of a group of three or four, together constituting the support for a container, a bowl or drinking cup. It is a type of representation that may refer to the triumphal processions of Dionysus or emperors. A coin struck under Tiberius in 36 or 37 depicts the deified emperor Augustus in a quadriga pulled by four elephants wearing a feather headdress.1215 The realistic rendering of the elephant and the African man suggests that the object was made by someone who had seen African people and elephants, for instance in triumphal processions. Just like the other statuette, the object was probably brought by a soldier, returning veteran or merchant.1216 Representations of African men were used for supports, anthropomorphic containers, lamps and so-called silent servants or Stummer Diener, statues of slaves holding a tray. 3.8.17 Animals with a mission A stray find from De Meern (Utrecht) represents a bronze billy goat carrying a saddle with a bag on each side.1217 The bulging bags have a closing knob and are clearly full. The figurine stands on a small rectangular plinth. The billy goat is an animal associated with Mercury. Several statuette groups depict Mercury together with billy goats or rams.1218 The god in his function as patron god of financial gain is often depicted holding a marsupium, a purse or pouch, in his outstretched hand. Therefore, the bags were probably filled with money. The iconography of the figurine from De Meern is rare, but not unique. Two small bronze rams from Nijmegen (fig. 3.12) and Tongeren, Belgium, also carry a saddle with two bags. A comparable figurine comes from an unknown find-spot at Rouen, France.1219 As Clasina Isings has suggested, the billy goat may have belonged to a group of statuettes with Mercury. If so, the filled saddle bag on the back of the animal must be seen as a reinforcement of Mercury’s function as the god of commercial success and abundance. Yet, it is also possible that the specimens from Rouen, Tongeren, Nijmegen and De Meern were statuettes in their own right, an iconographic ‘short-cut’. By portraying an animal that was associated with the god, carrying bags filled with money, it was not necessary to depict Mercury himself. The mission of the animal was clear to the beholder. 1212

Several bronze statuettes of African men have been found

Derks 1998, 227-230, note 60 and fig. 5.3.

in Britain: see Durham’s online database, nrs. 837-838, 886,

1217

Isings 1984, 99-102. See for the site also 2.3.3.

1144.

1218

See, for instance, a group from Augst, Switzerland, Bavay,

1213

See also 2.3.5.

1214

See also 2.3.4.

1215

Swinkels 2011, 10-11.

1216

See for the wooden writing tablets and bronze seal boxes

France and Schwarzenacker, Germany: Isings 1984, 102. 1219

Isings 1984, 101 and notes 5-6; Zadoks et al. 1973, nr. 48; Faider-Feytmans 1979, nr. 113.

153

3.8.18 Ithyphallic dogs In the northwestern provinces figurines of seated, ithyphallic dogs mostly come from cemeteries. A few specimens were recovered from settlements and sanctuaries. It appears from a comparison between examples from the study area, Belgium, France and Switzerland that their iconography presents hardly any variation.1220 In the study area, three were recovered from the cemetery of Ulpia Noviomagus and two from military settlements at Nijmegen and Vechten.1221 The fifth specimen was found near Nijmegen, together with pottery and a statuette of a terracotta cockerel. Details of the find circumstances are not known, but the assemblage may have been grave goods or an offering.1222 Dogs have always been a companion of man. Besides their significance as hunting and guard dogs, they symbolised loyalty and were thought to possess healing capacities, based on the belief that their saliva contained curative properties. Not only Roman and Celtic deities connected with hunting are often accompanied by dogs, but also female deities like Nehalennia and mother goddesses. Here, the dogs are depicted as domestic protectors. Dogs were also connected with death, probably because of their scavenger habits. In Roman religion, the three-headed dog Cerberus was the guardian of the underworld.1223 The ithyphallic element reinforces the role of the dog as guardian with its capacity to avert demons or the evil eye. Considering the connection between dogs and death and their role as guardian, it is understandable that ithyphallic dogs have been mainly found in and around cemeteries and military settlements in and outside the study area. 3.8.19 Triple-horned bulls Two bronze triple-horned bulls came to light in the study area, one in Vechten or Wijk bij Duurstede (Utrecht), the other in Beilen (Drenthe).1224 Symbolising power and fertility, the veneration of bulls can be traced back to the Greek Bronze Age. Around forty triple-horned bulls are known from the Roman West. The majority come from the eastern part of Gaul, north of Lyon. A few examples have been found in Britain and Central Europe. Most triple-horned bulls are made of bronze, a few of stone and clay. Since they appear in different find-contexts, theories have been advanced about the deity they are associated with and the meaning of the triplication of the horns. The phenomenon of triplism has already been briefly mentioned in 3.8.8. In Iron Age and Roman period societies, triplication is attested by deposits of three objects, like the three umbones in the castellum at Valkenburg (South Holland).1225 The meaning of repetition is not only emphasis or intensification, but also the magic symbolism of the number three. Examples besides the already mentioned triads of deities and mythological figures are triple-faced or triple-headed beings.1226 Evidence from various places suggests a link between the triple-horned bull and a solar god. At Beire-leChâtel (Côte d’Or, France), stone triple-horned bulls from the sanctuary of the Celtic goddess Ianuaria are associated with a radiate divinity who has been interpreted as a Celtic version of Apollo.1227 On a bronze sceptre fitting from Willingham Fen (Cambridgeshire, England) we see the head of a triple-horned bull together 1220

See, for instance, Von Gonzenbach 1986/1995, 229-231, Taf.

1223

Green 1992c, 82-84.

41, 1, 42, 1, 44, 1-2, 4-5, 55,3, 110-112; De Beenhouwer 2005, 153,

1224

Zadoks et al. 1967, nr. 46; Zadoks et al. 1969, nr. 65; Bogaers 1962, 579-581.

1218, 1249, 1269, 1273; Liéger 1997, 64-65 and planche 94, 1-3. 1221

From the cemetery of Ulpia Noviomagus: Van Boekel 1987

1225

See for this site 2.4.1.

nrs. 221-223; from Vechten: Van Boekel 1987 nr. 220. See for

1226

Green 1992c, 214-215.

the specimen from the Canisius excavations at Nijmegen

1227

Green 1992a, 180-182.

figs. 2.128-128a. 1222

Van Boekel 1987 nr. 219; for the cockerel Van Boekel 1987 nr. 211.

154

with a Celtic solar deity holding a lance in the left raised hand while he is trampling on a fallen enemy with his left foot.1228 Triple-horned bulls are also known from graves. A terracotta specimen was found in the aforementioned grave at Colchester, which, among other gifts, contained thirteen terracotta figurines.1229 The other specimen comes from a grave in Cutry (Meurthe-et-Moselle, France).1230 Both graves date to the 1st century.1231 The two triple-horned bulls from the study area are stray finds and do not provide any clues as to their function. The specimen from Beilen (Drenthe) was found in a rural region north of the Rhine, just north of the fully excavated settlement at Wijster, where no bronze or terracotta statuettes came to light. The findspot of the triple-horned bull from Utrecht is uncertain. It may have been found at Vechten or at Wijk bij Duurstede. At Vechten a Roman military fort was situated, whereas at Wijk bij Duurstede only traces of rural settlements have been found. 3.8.20 Horned birds All four horned birds from the study area come from cemeteries at Nijmegen: three from the burial site of Ulpia Noviomagus and one specimen from a cremation grave in the early Roman cemetery.1232 The latter grave dates to the Claudian period (41-54).1233 Examination of the skeletal remains revealed that the deceased was a child was between 7 and 14 years old.1234 The other funerary gifts included terra nigra pottery, two small colour-coated beakers an oil lamp and a small bow brooch (fig. 2.156). The statuettes depict a composite, horned animal with the round body of a duck-like bird (fig. 2.157). The figurines deviate in several aspects from other terracotta specimens found south of the Rhine. They were not made from the same pure, white clay as most statuettes in the study area, they are modelled free-hand and they do not stand on plinths.1235 The significance of horns and horned deities hark back to the Bronze Age. The imagery is not restricted to Celtic religion and is also known from the Greek mainland, Crete, Egypt, Mesopotamia and India. In the northwestern provinces, an Iron Age stone statue of a double-faced, horned figure came to light at Holzgerlingen, Germany. On the Gundestrup cauldron, a cross-legged sitting figure wears a horned helmet (fig. 3.9).1236 A bronze helmet with horns dates to the 1st century BCE and was found in the Thames near Waterloo Bridge, London. It is thought that the horns symbolized virility, fertility and aggression.1237 In the Roman period, the majority of horned images predominantly come from Britain and depict horned, ithyphallic gods with spears and shields. On a bronze fitting the Roman god Mercury is depicted with horns instead of wings. The fitting was found among the votive objects from the Uley shrines on West Hill, Gloucestershire.1238 In the study area, too, triple horned bulls and terracotta horned birds demonstrate that the powers attributed to horns persisted in the Roman period. It has been suggested that horned birds were given to the dead, to place the deceased under the protection of a horned god such as Cernunnos.1239

1228

The sceptre fitting was part of a hoard, including two stat-

1233

See for this site 2.3.4.

uettes of two helmeted horsemen (the horses are missing),

1234

I thank Louis Swinkels for the information.

two bronze statuettes of mounted soldiers, a bronze owl,

1235

Van Boekel 1987, 696.

a bronze eagle a bust of a bearded man, and several other

1236

The silver cauldron was found at Gundestrup in a peat bog in Jutland, Denmark. It is assumed that it was made in

objects made of bronze, jet and amber, placed in a wooden

Romania or Bulgaria in the late Iron Age: Green 1992c, 109.

box: Rostovtseff/Mattingly 1923, 91-109. 1229

See 3.8.

1237

1230

Green 1992c, 120.

Liéger 1997, 39, 65-66 and plate 55.

1238

Green 1992c, 43, 51-53; Woodward/Leach 1993 and fig. 85.2.

1231

Eckardt 1999, 60, Billoret 1976, 352-353.

1239

Van Boekel 1987, 696 and notes 341, 345-346.

1232

Van Boekel 1987, nrs. 238-241.

155

Three of the four figurines are rattles. Animal-shaped rattles are known from the Hellenistic and Roman period. Pig-rattles have been found throughout the Mediterranean. In some cases they could have been toys, but they were also thought to possess apotropaic powers. The initial function of the horned bird-rattle in the above mentioned child’s grave may have been a toy. Despite the fragility of terracotta, several damaged terracotta rattles from non-funerary contexts imply that they most likely were discarded, broken toys.1240 Yet, three of the four graves are not necessarily child’s graves, and dog rattles are known from adult burials in Kempten, Germany and Courroux, France.1241 Considering the presumed apotropaic qualities of rattles and horns, the horned bird-rattles, either in a primary or a secondary function, were given to ward off evil and to protect the deceased on their journey into the underworld.

3.9

Conclusions on the area south of the Rhine

The analysis has revealed relations between size, subject, find-spot and material. A relation between findspot and size emerges from a comparison between the sizes of bronze Mars and Mercury statuettes north and south of the Rhine. The majority of statuettes found south of the Rhine are considerably smaller and come from military and rural settlements. They were made as an offshoot of metalworking and were purchased for personal use. A few specimens are distinctly larger and have been recovered from rivers and other watery places, which raises the question whether they were deposited in the water as offerings. A relation between subject and material is in fact a relation between material and gender. Female figures such as Venus and mother goddesses were predominantly depicted in clay. Most bronze statuettes represent male deities and clearly outnumber bronze representations of Venus and mother goddesses. Mother goddesses became a new and popular subject of terracotta workshops from the end of the 1st century onward. Their iconography apparently had an impact on the iconography of Roman goddesses such as Fortuna and Minerva, but also on the native goddess Nehalennia. Terracotta figurines of the Roman goddesses copied the enthroned pose of mother goddesses while their bronze counterparts maintained their standing pose. It has been advanced that the Roman goddesses were attributed the qualities of mother goddesses. The iconography of Nehalennia is composed of iconographic elements of both Roman goddesses and mother goddesses. Also other examples of Roman-looking deities with the inscribed name of a native deity have been discussed. An explanation for this phenomenon could be that these native deities were an-iconic before the Roman conquest. People borrowed the iconography of a Roman god or goddess whose capacities were considered to be functionally equivalent. It has been concluded that the discussed statuette groups represent examples of creolisation. The find-spots of several statuettes suggest a relation with specific representations. A relation between the military and Hercules is indicated by the find-spots of Hercules statuettes and inscriptions to Hercules Magusanus. Representations with presumably protective and apotropaic qualities such as dwarfs, ithyphallic dogs and horned birds come from cemeteries or military settlements. It results from the analysis that the number of statuettes from graves is very small. Considering the large number of graves that have been investigated, it can be concluded that people in the study area avoided placing statuettes in graves. Bronze statuettes are absent. This is not a local phenomenon and applies to the vast majority of graves throughout the Roman Empire. It has been argued that bronze statuettes of deities were thought to be more susceptible to contamination with death than bronze vessels, weaponry and jewellery. The occurrence of amber in graves of women could be explained by the apotropaic qualities attributed to amber. 1240

Willemsen 2003, 55, 59.

1242

Groot 2008, 104; Clarke 1999, 23.

1241

Van Boekel 1987, 239-240.

1243

Bradley 2017, 8-30.

156

4

Rituals with statuettes ‘Archaeologists often complain that they lack sufficient information to offer persuasive interpretations of the past. Sometimes this is true, but here is a case in which they have too much material at their disposal and not enough ideas with which to address it.’ Richard Bradley in A Geography of Offerings (2017)

4.1

Introduction

This chapter is dedicated to three rituals. The first two are closely related: ritual deposition and deliberate fragmentation. The first will be discussed in paragraph 4.2. Rituals involving the deliberate fragmentation of figurines is the subject of paragraph 4.3. Paragraph 4.4 is dedicated to so-called magical practices. Until the end of the last millennium, the ritual deposition of metal objects, bones and other artefacts was regarded with scepticism by archaeologists and seen as ‘illogical’ or ‘non-functional’.1242 In his book on ritual deposits between the first millennium BCE and the late Medieval period, Richard Bradley sketches in A Chapter of Accidents, how scholars stick to practical interpretations of metal hoards in rivers and other watery places, by explaining them as ‘accidents’: cargos of sunken vessels or inundations.1243 In 1999, the British archaeologist Simon Clarke pointed out that in antiquity ritual behaviour was perceived as perfectly logical and rational. He concluded his lecture on the re-interpretation of deposits in pits at Newstead, Scotland, with the following statement: ‘Perhaps for the benefit of some Romanists it is still necessary to labour the point that what we call ritual activity was widespread in the ancient world, invading areas of life which in our own experience appear wholly secular. Deposition on Iron Age and Roman period sites was controlled in large part by behaviour patterns that to our mind were illogical responses to superstition or religious belief. But it only really stands out as outlandishly ritualistic from the standpoint of the outsider. To members of that society it was perfectly rational behaviour, which supported a sense of social identity, was integrated with the economic system and meshed perfectly with technological understanding. We need to move beyond the mere identification of ritual. We should be able to take that as read. What is required is a completely new set of questions about Roman period society, into which consideration of the belief system is fully integrated.’1244 Ritual deposition from the prehistory into the Roman period and beyond has been central to an increasing number of publications in the Netherlands and Britain.1245 Publications on Roman period deposits in the Netherlands mainly focus on animal bones or rituals related to the life cycle of habitations.1246 It appears from studies on Bronze Age, Iron Age and Roman period deposits of metal hoards, pottery and bones that the motifs, the objects involved and the choreography of these rituals may have changed over time, but 1244

Clarke 1999, 27.

in Britain: Ross/Featcham 1976; Merrifield 1987; Cunliffe

1245

See on ritual deposits in the Netherlands: Verhart 1993;

1993; Haynes 1997; Clarke 1999; Fulford 2001; Fulford/Timby

Fontijn 2002; Gerritsen 2003; Van den Broeke 1977, 2002a, 2002b, 2004, 2015; Van Enckevort/Willems 1994; Lauwer-

2001; Fittock 2015; Crease 2015. 1246

See for animal bone deposits in the Netherlands Therkorn

ier 2002, 2004; Groot 2008; Nicolay 2007, 2014; Van Hoof

2004; Groot 2008; Nieuwhof 2015; for rituals related to the

2007; Quast 2009; Nieuwhof 2015. See for ritual deposits

life cycle of houses Van Hoof 2007.

157

recurring patterns suggest that some traditions or conventions survived.1247 Only in the course of the last two decades, archaeologists have begun to recognise ritual deposits of statuettes in the context of settlements.1248 We do not know why these statuettes were chosen, who performed these rituals, how they were carried out, for whom or for how many people. In this chapter, I will attempt to answer the third and fourth research question of this study: 3. How can we identify ritual deposits of statuettes? 4. Are there any indications for the ritual deposition of statuette fragments? It appears from the previous chapters that statuettes could have had different functions before they entered the archaeological record, but the archaeologist can distinguish only two functions with certainty, based on the find-spots of statuettes: they could end up as grave gifts or as offerings in a sanctuary. We have seen that this applies to only 13 percent of the dataset. The majority, 87 percent, come from find-spots on land or were recovered from watery places.1249 A part may have ended up in their find-spot through natural processes or human activity such as ploughing. I have argued in the first chapter that many statuettes were cultic objects from the start, which makes it unlikely that they were discarded as refuse. This is also contradicted by archaeological evidence and written sources (see below). However, to create a point of departure for the distinction between presumed ritual deposits and secular deposits, criteria have been formulated: a deposit could be a ritual deposit if it meets at least two of the following criteria: one of the first two (1 or 2) and one of the last two (3 or 4). 1. A deposit includes at least one complete or substantially complete statuette, sometimes accompanied by a striking assemblage of other objects, or: 2. A deposit includes a statuette head/head-torso fragment, sometimes accompanied by a striking assemblage of other objects. Other fragments of the statuette are absent. 3. A statuette or statuette head/head-torso fragment is buried in or next to a building in a settlement: under a floor, demolition layer, under or near the entrance, in a corner, posthole, or: 4. A statuette or statuette head/head-torso fragment is deposited in a ditch, pit, well, drain or cistern associated with a building or settlement.

4.2

Ritual deposition

4.2.1

A long history

Prior to the discussion of Roman period deposits, I will briefly dwell on the studies of Fontijn and Gerritsen. Fontijn deals with the selective deposition of weaponry and ornaments during the Bronze Age in the Meuse-Demer-Scheldt region, nowadays part of the southern Netherlands and northern Belgium. Gerritsen has studied the development of communities and the interaction with the landscape in the same area, from the late Bronze Age into the Roman period. A chapter in Gerritsen’s research is dedicated to the social and spatial organisation of the house. In this chapter he also discusses deposits of organic material and artefacts related to the life cycle of houses. 1247

See, for instance, the find-spots of Bronze Age Iron Age and

1248

early Roman period deposits in Gerritsen 2003, especially tables 3.5, 3.8, 3.10-3.14; for the find-spots of selected animal bones in ditches, wells and pits in the Roman period Groot 2008, 120-138.

158

Hermsen 2003, 65-71; Van Es 2012, 16-17; Taayke et al.2013, 128-129; Driessen 2014b, 176-178.

1249

These locations are still ‘wet’, or may have been so in antiquity.

Fontijn’s study on deposits of bronze weaponry and ornaments demonstrates that selective deposition is attested as early as the Neolithic. An important aspect of selective deposition is the choice of special places in the landscape. Wet locations were chosen for the deposition of bronze swords, daggers and axes. They include major rivers, streams, peat bogs and swamps. Dry places that were chosen were high points and plateaus, places in the periphery of burial sites and settlements, farmyards and houses. Not only the places for deposition were selected, also the objects themselves. There must have been reasons why some objects were deposited in wet or dry natural places, while others were placed in graves. Swords, for instance, were predominantly deposited in rivers. Axes, sickles and weapons seldom occurred in graves. Sickles were deposited on farmyards, axes were not.1250 Although we can only speculate on the nature of the conventions and reasons behind it, the persistence of comparable rituals with weaponry and other bronze objects in the Roman period indicate that they were important and powerful.1251 In Gerritsen’s study, the cultural biography of the house and ritual deposits related to the life cycle of houses demonstrate that these deposits were made in recurrent find-contexts, around entrances, in postholes, wall ditches and corners of houses.1252 The deposits were made when houses were constructed or abandoned. The patterns that emerge from their analyses show that depositional practices were based on well-considered choices made by people trying to make sense of their world. Although these choices, the meanings and objects involved may have changed, some depositional conventions have persisted into the Roman period. This appears from recurring locations of ritual deposits in and around Roman period settlements and the surrounding landscape and from the fact that objects were selected to accompany the deceased while others were avoided. The recurring locations of ritual deposits have been noticed by others, too, and constitute the starting point for the discussion of ritual deposits of statuettes in the study area as well as in Britain. 4.2.2

Categories of ritual deposits

In the study area, 33 possible ritual deposits of statuettes and statuette fragments have been identified of which 31 come from controlled excavations. The find circumstances of some alleged special deposits have not been described in detail. Examples are Maasbracht, Leiden-Roomburg, Ouddorp, Valkenburg-castellum and Alphen aan den Rijn-De Schans. Because of the paucity of contextual data, it remains uncertain whether these statuettes were deposited as part of a ritual. This is illustrated by the following example. The bronze statuette of Isis-Fortuna from the Valkenburg castellum once stood in the fort.1253 It was left behind when the fort was abandoned, perhaps in haste or panic. Despite the value of a large bronze statuette this cannot be ruled out. Glasbergen, supervisor of the 1962 excavations of the Valkenburg castellum, suggested that three bronze umbones were intentionally buried in the fort. The umbones were found under the burnt layer that marked the end of the fourth phase of the castellum. Glasbergen gives a detailed description of their location, orientation and find circumstances. The deposit was possibly linked to the end of the fourth phase or the construction of the new fort in the fifth phase. In contrast, his account of the find circumstances of the bronze statuette is brief and limited to the statement that the statuette was found in ‘a demolition layer belonging to phase 6 or later’.1254 Like the umbones, the statuette may have been a ritual deposit related to the end of the sixth occupation period.1255 The alleged ritual deposits in the study area have been divided into the following categories, for which I have partly followed Ralph Merrifield’s division of ritual deposits in Britain.1256 1250

Fontijn 2002, 211-217, fig. 10.1 and tables 10.1-10.2.

1253

See for this site 2.4.1.

1251

See for Roman period deposition of weaponry and horse

1254

Glasbergen 1972, 116.

gear in the study area Nicolay 2007.

1255

See for the presumed seventh phase of the fort 2.4.1.

Gerritsen 2003, 63-66 and table 3.5.

1256

Merrifield 1987, 22-57.

1252

159

1. 2. 3. 4. 5.

Deposits in sanctuaries or alleged cult places Deposits found in association with a building or buildings Deposits from ditches belonging to a house, other building or settlement Deposits from wells, drains and cisterns Deposits from banks, old and existing river beds, rivers, streams and gullies

Categories 1 and 5 will not be commented on in detail. The problems regarding the identification of cult places have been discussed in the first chapter.1257 Category 5 comprises a substantial number of statuettes which have been recovered from watery places.1258 The criteria formulated in this study cannot be applied to this category. Besides one exception, all bronze statuettes in this group are rather large.1259 They may have been buried as a precautionary cache or they were deposited in the water as part of a ritual, which has also been advanced for swords, helmets and other metal objects found in rivers and streams. Categories 2-4 comprise deposits associated with the life cycle of buildings and settlement features and will be discussed hereafter. The aim of the overview is to highlight recurrent characteristics of deposits of statuettes, patterns that have already been noticed in studies on ritual deposits of other artefacts.1260 The deposits comprise complete and substantially complete bronze and terracotta statuettes as well as statuette heads, often found together with other objects. ‘Date’ in the overviews is the deposition date, often a terminus post quem related to the estimated construction date of the building or settlement where the deposit was found. ‘Par.’ refers to the paragraphs in this study. Category 1: deposits from sanctuaries and alleged cult-places Find-spot

Context

Composition deposit, condition statuette

Selection

Date

Par.

4 bronze, 9 terracotta statuettes

Mercury, boar, donkey, base, Minerva, Venus, mother goddesses, woman, Apollo/ Bacchus, bust of boy, warrior

After 150?

2.2.4

CATEGORY 1 N. of Rhine

Sanctuary or alleged cult place

1 Noordbarge

Alleged cult place

S. of Rhine

Sanctuary/alleged cult place

2 Bergen op Zoom

Cult place, fen

Terracotta, incomplete Miniature vessels, amphorae sherds, 13 statuette fragments, coins, beads

Sucellus

After 150

2.3.5

3 Buchten

Sanctuary, pit outside sanctuary

Bronze, complete, silver ring, necklace, two coins, fibula

cockerel

After 250

2.3.5

4 The HagueScheveningseweg

Alleged cult place

ca. 20 terracotta statuettes, incomplete

Venus, Fortuna, Cybele, gladiator, busts of boys, bust of boy or girl, mother goddesses, dwarf, cockerels, horse

After 250?

2.4.2

5 Wijchen-Tienakker

Alleged cult place, pit

Bronze, complete Tuff stone altar

Minerva

After 200

2.3.4

1257

See 1.8.

1258

See, for deposits in watery places, for instance, Fontijn 2002; Nicolay 2007, 181-189; Crease 2015; Bradley 2017.

1259

An exception is the tiny bronze mother goddess from

160

Koudekerk. See for this figurine 2.3.2 and 3.5.4. 1260

Gerritsen 2002; Fontijn 2003; Nicolay 2007; Van Hoof 2007; Groot 2008; Nieuwhof 2015.

Category 2: deposits in association with a building or buildings1261 Deposits in the context of dwellings have been attested from the early Iron Age onward and became increasingly more common. For the Iron Age and Roman period, examples are known from find-spots north and south of the Rhine.1262 Leon van Hoof has investigated Roman period deposits associated with houses south of the Rhine. He distinguishes three zones of house building traditions, the villas in the loess-belt excluded. The deposits in each zone have their own characteristics. 1263 Ritual deposits related to the life cycle of a house or settlement feature such as wells and ditches comprise two categories. The first consists of deposits made before or during the construction of a building or settlement feature and are known as foundation deposits.1264 The second regards deposits that were made when the house or settlement was abandoned, so-called abandonment or closing deposits. Deposits were probably also made during the phase of habitation, but recognising the difference between these categories is difficult.1265 In some cases, foundation deposits are easier to identify. When objects are placed at the bottom of a posthole, before the post was put into place, the deposit can be linked to the beginning of the life cycle of the building. The same applies to deposits at the bottom of wall ditches, wells, in areas around entrances and in pits under floors.1266 Iron Age and Roman Iron Age foundation deposits of pottery vessels and animal bone come from Assendelft and Schagen-Muggenburg (North Holland). At Heeten (Overijssel) a foundation deposit is known from a late Roman house. A situla was placed at the bottom of a posthole before the post was put into place.1267 Examples of foundation deposits south of the Rhine are known from Midden-Delfland ( South Holland), Oss (North-Brabant), Tiel and Oosterhout (Gelderland).1268 North of the Rhine, three deposits of statuettes are associated with a building or settlement. At Colmschate (Overijssel), deposit 6 is a bronze Victoria statuette which was placed in a posthole, directly east of the entrance. Deposit 8 is a bronze Fortuna statuette, discovered near the entrance of a building at Ede-Bennekom (Northern-Gelderland). Deposit 10, from Schagen (North Holland), was discovered in a pit under the floor. The deposit comprised a terracotta bird figurine, a piece of iron, a complete handmade bowl, several blue beads and one black specimen. Deposit 11, from Wehl (Northern Gelderland), was the filling of a pit west of a gold smith’s workshop and contained besides potsherds and pieces of iron, the head of a terracotta Venus figurine.1269 Deposit 9 consisted of two bronze statuettes: Minerva and an offerant. They had been deposited in a shallow pit above a well at Ede-Veldhuizen (Northern Gelderland), during the latest phase of the settlement. A large bronze Mercury statuette, deposit 7, was found at Dalfsen (Overijssel), together with potsherds discovered in a river bank at the edge of a settlement. Deposits in sunken byres seem typical for Kempen, the region that covers the north-eastern part of Belgium and the south-eastern part of the Dutch province of North Brabant. Here, deposits have been found in the top layers of sunken byres or in postholes, where they were placed after the post had been removed. The latter are, therefore, interpreted as the result of an abandonment ritual.1270 The same could apply to deposit 18, the bronze statuette from the castellum at Valkenburg, which was found in a demolition layer that demar-

1261

Buildings include houses and outbuildings or annexes.

Nieuwhof 2015, 201-209; Van Hoof 2007, 255-270.

With ‘house’ I mean house or farmhouse. Gerritsen defines

1263

Van Hoof 2007, 255-270.

the house as follows: ‘The term house or farmhouse refers

1264

See Gerritsen 2003, 64-65, fig. 3.21 and table 3.5.

to a building used primarily for habitation by humans,

1265

Merrifield 1987, 50-51; Gerritsen 2003, 63-64, 79-86, 96-102; Van den Broeke 2002b, 2015.

and in the case of byre-houses by humans and animals

1262

together: Gerritsen 2003, 38. For the difference between a

1266

Gerritsen 2003, 65.

villa and a farmhouse I use the definition of Roymans and

1267

See for this site 2.2.4.

Derks: a farmhouse is a villa if it has a stone-built main

1268

Gerritsen 2003, 64 and table 3.5.

house: Roymans/Derks 2011, 2-3.

1269

See also 4.2.4.

Gerritsen 2003, 63-70, 79-86, 95-102 and tables 3.5, 3.7-3.8;

1270

Van Hoof 2007, 264.

161

cated the end of the sixth occupation phase of the fort and may have been the remains of an abandonment ritual. A deposit of axes in sunken byres is known from Brecht (Antwerp), where a deposit of two bronze axes and one stone axe was recovered from the sunken byre of a farmstead that could be dated to the end of the 2nd or the first half of the 3rd century. The excavators consider the axes as a deposit with a ritual meaning. They also mention ritual deposits of whetstones and grind stones in sunken byres in East-Flanders.1271 In the study area, Nieuwhof has noticed that several deposits north of the Rhine were made in byres. These deposits date to the Iron Age, but she does not rule out the possibility that others were made in the Roman period.1272 South of the Rhine, two deposits of statuettes were found in the corner of sunken byres, deposits 12 and 16, at Helden (Limburg) and Son en Breugel (North Brabant). Deposit 12 is a bronze Mars statuette. Deposit 16 is the head of a terracotta mother goddess. Therefore, I consider it likely that the terracotta statuette head, the bronze Mars and the Frankish throwing axe in sunken byres in the southern Netherlands were also ritual deposits.1273 Whether the two deposits were linked to the life cycle of the building cannot be established on the basis of the available data. The presence of metal objects and metal slag could point at a link between metalworking and the statuettes.1274 It appears from chapter 2 that pits in settlements are often interpreted as refuse pits. Nevertheless, ‘rubbish pit’ seems an unlikely interpretation in the case of a number of pits in Roman period settlements in the study area. Pits with remarkable contents have been observed in and around farmsteads from the Late Bronze Age into the Roman period. Gerritsen has presented a table with an overview of Late Bronze Age and Iron Age deposits in or next to farmhouses in Gelderland, North Brabant and Limburg. Pits with complete pots or large quantities of potsherds and other objects in Iron Age settlements have been linked to abandonment rituals.1275 Rituals related to the abandonment of the house seem to have been quite elaborate in the early and middle Iron Age, but examples from the late Iron Age and Roman period are also known.1276 Pits were dug inside or outside houses and filled with domestic artefacts. Sometimes, fire would be lighted in the pits, or the artefacts were first exposed to fire at another location and then deposited in the pits. The artefacts comprised potsherds, loam, fragments of grinding stones, cooking stones, hearth stones, spindle whorls and loom weights. Metal objects or burnt bone seldom occur in these deposits, whereas wooden objects and unburnt bone may have been part of the depositions, too, but usually remain archaeologically invisible.1277 Van den Broeke has advanced that burning domestic artefacts may not have been a ‘standard procedure’ in Iron Age abandonment rituals. He has suggested that there might have been a central location inside the main house where domestic artefacts were burnt, but not necessarily so, after which parts of the deposits were distributed over other locations in and around the farmhouse and all outbuildings, wells and pits.1278 A pit with a large quantity of potsherds and complete statuettes are deposits 19-20 from the western canabae at Nijmegen.1279 The bronze Luna statuette is a semi-finished specimen (fig. 2.123). The bronze billy goat (fig. 2.125) came from a deeper level. At the time, the excavator, Haalebos, wondered in his report about

1271

Bracke et al. 2015, 65-66 and fig. 2.

Holland): Van den Broeke 2015, 91-92. See for late Iron Age

1272

Nieuwhof 2015, 203.

and Roman period deposits that may have been related

1273

See for these sites 2.3.5.

to abandonment rituals at Ezinge (Groningen): Nieuwhof

1274

See 4.2.4.

1275

See Van den Broeke 2002b, 2015; Gerritsen 2003, especially

1277

Gerritsen 2003, 97, 99-100.

93-102.

1278

Van den Broeke 2015, 83-96.

Examples of closing rituals in late Iron Age settlements are

1279

The deeper level of the deposit with the billy goat indicates

1276

2015, 203-205.

Geleen (Limburg), Oss (North Brabant) and Oosterhout

that this deposit was made earlier than the deposit with

(Gelderland): Gerritsen 2003, table 3.14. Examples dating

the Luna statuette, which was found in the top layer of the

to the Roman and late Roman period are Maasland-Duif-

pit, together with remains of burnt limestone and charcoal.

polder (South Holland) and Schagen-Muggenburg (North

See also the next note.

162

the content of the pit.1280 Besides a large quantity of potsherds, deposit 11 from Wehl contained iron objects, a large Roman roof tile fragment and the statuette head of a Venus figurine. Deposit 13, from Kerkrade Holzkuil, is the filling of a pit in front of the southwest wing of the villa complex. The filling comprised the head of a terracotta Risus bust and sherds of a pre-Flavian terra sigillata bowl and cooking pots dating to the 2nd and 3rd centuries. Other fragments of the figurine were absent. A pit with a timber lining on the terrain of a Roman period villa at Kerkrade-Maar West, deposit 14, contained nine complete pots and a complete terracotta figurine of a bear. Find-spot

Context

Composition deposit, condition statuette

Selection

Date

Par.

CATEGORY 2 N. of Rhine

Building(s)/settlement

6 Colmschate

House, near entrance

Bronze, complete

Victoria

After 250

2.2.4

7 Dalfsen

Settlement, river bank

Bronze, complete Potsherds

Mercury

After 300

2.2.4

8 Ede-B.

House, near entrance

Bronze, complete

Fortuna

After 400

2.2.4

9 Ede-V.

Settlement, wet area

Bronze, complete Two statuettes

Minerva, offerant

After 400

2.2.4

10 Schagen

Outbuilding, under entrance

Terracotta, head missing. Metal object, beads, complete pot

bird

After 270

2.2.3

11 Wehl-Old.

Workshop, pit

Terracotta, head. 1.5 kg potsherds, iron window grill, nails, iron slag, large Roman roof tile fragment

Venus

After 300

2.2.4

S. of Rhine

Building(s)/settlement

12 Helden

Sunken byre, corner

Bronze, complete

Mars

After 150

2.3.5

13 Kerkrade- Holzkuil

Villa, pit

Terracotta, head Potsherds, 1st-3rd century AD

Risus bust

After 300

2.3.5

14 Kerkrade-Maar-West

Villa, pit

Terracotta, complete. Nine complete pots

bear

After 300

2.3.5

15 Maasbracht

Villa, cellar

Terracotta, head*

Risus bust

After 150

2.3.5

16 Son en Breugel

Sunken byre, corner

Terracotta, head

mother goddess

After 175

2.3.5

17 Tiel-Passewaaij

House, pit

Terracotta, head

Cybele

150-220

2.3.4

18 Valkenburgcastellum

Roman fort, demolition layer

Bronze, complete

Isis-Fortuna

After 240

2.4.1

19 Nijmegen

W. canabae, pit

Bronze, complete Large potsherds

Luna

70-100

2.4.1

20 Nijmegen

W. canabae, same pit, but at a deeper level

Bronze, complete

billy goat

70-100

2.4.1

*The head is not mentioned in Vos et al. 2017.

1280

In his report of 18 June 1992 Haalebos wrote: ‘De kuil,waarin

zijn ook enkele grote brokken aangetroffen.’ Although Haale-

(of beter waarop) Luna is gevonden lijkt wat merkwaardig en

bos suggests in this report that the large pottery sherds in

bevatte enkele in scherven vrijwel complete (?) potten, een’pl-

the pit could be complete pots, complete or substantially

ek met wat houtskool en mogelijk enkele stukjes crematie.

complete specimens have not been recovered from the pit:

Veel witte puntjes maken echter de indruk geen verbrand

oral information by Rien Polak.

been te zijn maar restjes van verbrande kalksteen. Hiervan

163

Category 3: deposits in ditches Ditches could be dry, wet or filled with water. Yet, the presence of water was probably not the reason why objects were deposited in ditches. Offerings were made in ditches because they were boundaries. Enclosure and defensive ditches demarcated the space around houses, settlements, forts and sanctuaries. These kinds of boundaries have always played an important role. In the early Mediterranean world, boundaries and crossroads were marked with stones or columns and later by herms, stone shafts with male genitals and a head.1281 The stones and herms were anointed or decorated with garlands, which we know from written sources such as Theophrastus (ca. 371-287 BCE) and Tibullus (?-19 BCE).1282 The veneration of boundaries is not restricted to the Mediterranean world. Ditch deposits have been attested in communities throughout Europe from the Neolithic onward.1283 Enclosures and enclosure ditches around sanctuaries and burial sites can be interpreted as an expression of territoriality, a demarcation between the territory of a god or the deceased and the rest of the world. Enclosure ditches belonging to military settlements such as Roman forts had a primarily defensive purpose. Although enclosure ditches around settlements have been interpreted as defensive ditches, too, Gerritsen has pointed out that nucleation and the appearance of enclosures in the 1st century BCE were not primarily intended for protection. Contrary to dispersed farmsteads, nucleated settlements became a symbol for local communities. Enclosures emphasized the collective identity of the settlement by separating it from the outside world.1284 Objects and fragments of objects that have been recovered from ditches constitute a more problematic group than the other categories, especially when a detailed description of the find circumstances is absent and the excavations were carried out decades ago. Moreover, ditches can cover a relatively large surface and are excellent artefact traps, because of their lower level in relation to the ground level of the environment. Nevertheless, the composition of assemblages or the undamaged condition of objects could indicate that objects were deposited as part of a ritual, which could apply to the following deposits. North of the Rhine, several deposits comprise complete or fragmented human skulls while other parts of the skeleton are missing. Also striking is the presence of bronze and iron objects such as keys, nails and knifes in the deposits.1285 Deposits 21 and 22 came to light during controlled excavations. The deposit from Texel (North Holland) is a bronze attachment in the shape of a putto. The winged boy represents the per1281

1282

1283

1284

According to Plato, herms were a Greek invention by Hip-

1285

See for ritual deposits of human skulls in the northern Neth-

parchus, the tyrant of Athens in the late 6th century BCE:

erlands Nieuwhof 2015. See for ritual deposits with one or

Pl. Hipparch. 228d. The Greek historian Thucydides tells us

more nails: Tzum, deposit 22, Wehl, deposit 11; for a deposit

that herms were found in great numbers in Athens, in the

with nails and a an iron sword the site Oosterhout-Van

porches of houses and in temples: Thuc. 6. 27. First, heads

Boetzelaerstraat (2.3.4); for a deposit from Belgium: Sber-

of the god Hermes were placed on the shafts, after the 5th

champs, 4.3.3; for a deposit with nails and keys in Britain:

century BCE also heads of other deities. Besides the apot-

Holt Down-Hampshire, deposit 6A. Several deposits in the

ropaic function of herms, the genitals symbolised fertility

study area and Belgium contained knives: see for a deposit

or territorial aggression. Hamilton 2010, 414-415. See also

of animal bones and a knife the site Tiel-Passewaaij, 2.3.4;

3.8.15 for Fehling on this subject. In the Roman period, the

for a foundation deposit containing a knife in the corner

heads on the shafts were often portrait busts, but not nec-

of a Roman period house in Midden-Delfland (South Hol-

essarily so. See, for instance, Wrede 1972, 40-80.

land) Gerritsen 2003, 66; for a deposit of a jar and a knife in

Theo. Char. 16.5: καὶ τῶν λιπαρῶν λίθων τῶν ἐν ταῖς τριόδοις

a pit at Wijchen, 2.3.4; for a knife and a whetstone (?) in a

παριὼν ἐκ τῆς ληκύθου ἔλαιον καταχεῖν καὶ ἐπὶ γόνατα πεσὼν

posthole of a late Iron Age house at Ekeren (Antwerp) Ger-

καὶ προσκυνήσας ἀπαλλάττεσθαι. Tib. 1.11-12: nam veneror, seu

ritsen 2003, 66. Several deposits in the study area, Britain

stipes habet desertus in agris seu vetus in trivio florida serta

and France contained keys. See for Voorburg-Arentsburg

lapis.

2.3.2; for Holt Down-Hampshire: 4.2.5; for a deposit in the

Merrifield 1987, 38-39; Spielmann 2008, 40; Groot 2008, 121-

cellar of a domus in Reims 4.4. See for keys in Scandinavian

128, 130, 133.

deposits Bradley 2017, 12.

Gerritsen 2003, 188-189, 197.

164

sonification of one of the Seasons and once was part of a support or a side table. Deposit 22 from Tzum (Friesland) comprised a complete bronze Mercury statuette, five fragments of a human skull, a bronze plate, a large iron nail, one half of a key and potsherds. The objects came from the deepest level of a ditch. The presence of the fragmented human skull and the composition of the deposit suggest that the assemblage was an offering and not scrap metal. South of the Rhine, at Alphen aan den Rijn-castellum, a human skull was recovered from one of the enclosure ditches belonging to phase 2 (c. 70-160) of the castellum and ended up there around 160. No other bones were found.1286 A deposit of three terracotta statuettes, Cybele and her two lions (see also figs 2.60-2.60a1-3), come from a ditch at Oosterhout (Gelderland). The figurines were found at a short distance from each other. The function of the ditch is not clear. A near life-size, complete terracotta head of a boy, possibly depicting a young Bacchus or Amor, is a rattle and was found in a ditch at Alphen aan den Rijn-De Schans (fig. 2.186). The rattle was probably too big for a toy. Its intact state implies that the head was carefully put in place and not discarded as waste. The ceremonial and ritual function of rattles as well as its undamaged condition suggest that the terracotta head was a special deposit. Other finds from the ditch, the coins, fibulae and statuette fragments, may have been ritually deposited as well, but this is uncertain. The terracotta head in a ditch at Tiel-Passewaaij (deposit 27). is the third example of a deposited statuette head in this settlement. The other two were recovered from a well and a pit, north and south of the same house. No other fragments of the statuettes were found in these settlement features. Find-spot

Context

Composition deposit, condition statuette

Selection

Date

Par.

CATEGORY 3 N. of Rhine Ditches

21 Texel

House, enclosure ditch

Bronze, applique

putto

After 200

2.2.3

22 Tzum

Terp, ditch

Bronze, complete

Mercury

?

2.2.2

Fragments of human skull, bronze fitting, iron nail, part of iron key S. of Rhine

23 Alphen a/d Rijn-De Schans

Settlement, ditch parallel to road

Terracotta, head, complete Coins, fibulae, statuette fragments, whetstone fragments

boy

After 200

2.3.2

24 Leiden-Roomburg-castellum

Defensive (?) ditch

Terracotta, complete

water nymph

After 200

2.4.1

25 Oosterhout-Van Boetzelaerstr.

Settlement, ditch

Terracotta, head plus upper part torso Three statuettes

Cybele with two lions

After 200

2.3.4

26 Ouddorp

Settlement, ditch

Terracotta, head

water nymph

150-225

2.3.2

27 Tiel-Passewaaij

Stable, ditch

Terracotta, head

mater

210-240

2.3.4

1286

Oral information by Rien Polak. See for the ditch (2a) Polak et al. 2004, 44-46 and map B. See also 2.3.2 and figs. 4.1-4.2.

165

Category 4: deposits in wells, drains and cisterns There are plenty examples of deposits related to the life cycle of wells and cisterns.1287 They mark the beginning or end of the period in which they were used. Deposits in wells of animal skulls are known from non-military and military settlements in Britain. They were easier to distinguish from dumped refuse when they were found together with relatively complete pottery vessels, which is why David Graham and Martin Millet, the excavators of wells and cisterns at Neatham, Hampshire, recognized a recurrent pattern of deposited pots and animal bones. They noticed that some deposits were made after the well had been dug, while a second deposit was made when the water in the well had become foul.1288 It has been argued that statuettes from drains could have been ‘water-oriented’ offerings.1289 A comparable pattern of these practices can be observed in wells in the study area. Several remarkable assemblages at Voorburg come from the top filling or bottom of wells and have been interpreted as offerings related to the beginning or end of the well’s life cycle. In one of the wells a whole grind stone had been deposited. Under the stone, two bronze objects and a wooden object were placed. The wooden object depicts a dolphin and probably was a decorative element belonging to a boat or ship. The harbour, too, has yielded remarkable deposits.1290 Besides complete bronze and ceramic vessels and other objects, statuettes and statuette heads have been found in wells, too. Examples come from Naaldwijk-Hoogeland, Voorburg-Arentsburg, Tiel-Passewaaij and Aardenburg, deposits 28-31. The large bronze water nymph or Venus from the filling of a well at Naaldwijk-Hoogeland is an applique that once was attached to a large object (see paragraph 4.2.3 and figs. 2.51-54).1291 The woman pours water from a jar. As stated in the third chapter, Venus, just like water nymphs, was associated with water, which is an extra argument for the idea that the attachment is a special deposit. From the bottom of a well at Voorburg comes the head of a Risus bust. Category 5: deposits from river areas, rivers, streams and gullies A substantial number of bronze and terracotta statuettes come from fossil and existing river beds, streams and gullies in or near settlements. Some terracotta statuettes, usually fragmented specimens, have been found in watery places among household waste. Although the ritual nature of these deposits is uncertain, arguments can be advanced that speak in favour of a special deposit. Starting with the incomplete terracotta figurines, it is striking that seven of the 15 are heads, five of a mother goddess and one of Juno, who also may have been venerated as a mother goddess.1292 Second, as stated before, the majority of statuettes were purchased to bring fortune or to ensure protection, which makes it unlikely that they were discarded as refuse, even when they were worn or broken. Archaeological evidence supports this. Yosef Garfinkel has examined cultic objects and figurines that were buried in or near sanctuaries in the Neolithic Near East. Based on their condition, he has argued that they were buried because they were broken or worn-out and therefore no longer suited for religious practices. Important for this study is his argument that, once an object was used in a ritual, it could no longer be turned into an object for daily use. 1287

Kaufmann-Heinimann has noted that a substantial num-

cae, from the villa at Hoogeloon-Kerkakkers and the frag-

ber of assemblages have been recovered from wells, but

ment of a terracotta statuette of a semi-nude figure, found

she interprets these Brunnenfunde as Versteckfunde and

in a drain on the compound of the villa at Voerendaal-Ten

not as ritual deposits: Kaufmann-Heinimann 1998, 189 and

Hove.

note 657. Several finds from wells, drains and cisterns are

1288

Merrifield 1987, 45-46.

left out in this overview, because they do not meet the cri-

1289

Fittock 2015, 122 and note 42.

teria formulated in this chapter. A fragment of a terracotta

1290

See for Voorburg also 2.3.2.

statuette, possibly depicting a young boy, was found in

1291

See for the site Naaldwijk-Hoogeland also 2.3.2.

the upper part of a well in the pars rustica of the villa of

1292

Other fragments of the statuettes may have been over

Druten-Klepperhei. Other finds from wells and draining ditches are four fragments of a statuette of one of the Par-

166

looked, but the majority have been analysed by experts.

Find-spot

Context

Composition deposit, condition statuette

Selection

Date

Par.

CATEGORY 4 S. of Rhine

Wells/drains/cisterns

28 Naaldwijk-Hoogeland

Top filling well

Bronze, applique

water nymph/ Venus

After 200

2.3.2

29 Nijmegen

E. canabae, well/cistern

Terracotta, incomplete. Two statuettes

Fortuna, head woman

75-125?

2.4.1

30 Tiel-Passewaaij

Top filling well

Terracotta, head. Bronze fitting, potsherds

Minerva

150-220

2.3.4

31 Voorburg-A

Bottom well

Terracotta, head

Risus bust

After 200

2.3.2

Deposits without arch. data 32 Tiel-Badhuisplein

-

Bronze, complete. 23 fibulae, terra sigillata bowl

Mars

After 100

2.3.4

33 Montfort-Montforterbroek

Swampy area near stream

Bronze, applique, Hallstatt spear head, potsherds

panther

After 235?

4.2.3

The process was irreversible.1293 Roymans has argued the same for offerings such as Gallic coins in pre-Roman Gallic societies and Roman period sanctuaries.1294 In the Roman period, too, religious objects were protected and could not be alienated.1295 When there was no room left for new offerings in a sanctuary, the donated objects were buried by priests at the sanctuary’s precinct to prevent them from being recirculated or put to some profane use. Sacred and ritual objects were also regarded as potentially dangerous.1296 In the previous chapter a ritual is described with a statue of the native goddess Nerthus.1297 Slaves would wash the statue in a lake and bring it back to the sanctuary. Thereafter, the slaves were ‘immediately swallowed by the same lake’ which I believe implies that the slaves were instantly killed by drowning them.1298 The last part of the phrase explains why they had to be killed: ‘arcanus hinc terror sanctaque ignorantia, quid sit illud, quod tantum perituri vident’: because the slaves had witnessed this ‘mysterious terror’ they had to die.1299 Since the statue was a sacred object, it is also possible that the slaves had to be killed because they had seen and touched ‘the goddess’. Therefore, they had to be removed from society.1300 As to the complete terracotta rattle from Zwammerdam (fig. 2.178), the figurine could have been a toy, but since it depicts a swaddled infant, it seems more likely that it was an offering made to ensure healing or health for a baby. The bronze statuettes in the overview are mostly accidental finds of which the exact find-spot in many cases is unknown.1301 Moreover, rivers and their tributaries were far more dynamic in antiquity, since there were no levees or dikes to regulate water levels. Rivers could change their course and streams and gullies could dry up. As a consequence, there were considerable differences in the study area between landscapes

1293

Garfinkel 1994, 159-188. See also Chapman 2000, 24.

1294

Roymans 1990, 83.

1295

Scheid 2003, 64.

1296

Nieuwhof 2015, 104-105.

1297 1298

arcanus hinc terror sanctaque ignorantia, quid sit illud, quod tantum perituri vident. 1299

Bradley has advanced that the slaves were killed to keep the location of the sanctuary secret: Bradley 2017, 17-18.

See also 3.2.

1300

Nieuwhof 2015, 104-105.

Tac. Germ. 40: servi ministrant, quos statim idem lacus haurit

1301

Which is why river finds are not indicated in map 11.

167

2000 years ago and today, especially the Lower Rhine delta. Whether all bronze statuettes were ritual deposits remains uncertain. It cannot be ruled out that some were buried in times of unrest and later were washed away by the river. Since they were objects of worship, I believe that the majority were deposited in the water as an offering, like the numerous examples of swords, helmets and other metal objects that have been recovered from rivers, river banks, old river beds and other watery places. Category 5: north of the Rhine Find-spot

Material

Subject

Par.

Heerde- near river IJssel

Bronze, complete

Venus

2.2.4

Zutphen-near tributary river IJssel

Bronze, complete

Fortuna

2.2.4

Find-spot

Material

Subject

Par.

Heerewaarden, Meuse river bank

terracotta

head, mother goddess

2.3.4

Jutphaas, river

terracotta

head, Juno

2.3.3

Koudekerk, Rhine, river bank

bronze, complete

mother goddess

2.3.2

Linge, river

bronze

pedestal, Vagdavercustis

2.3.4

Rhine, river

bronze, complete

Attis

2.3.4

Rhine, river

bronze, complete

Mercury

2.3.4

Valkenburg-Marktveld, gully

terracotta

Fortuna

2.4.1

Valkenburg-Marktveld, gully

terracotta

head, mother goddess

2.4.1

Voorburg-Arentsburg, harbour

terracotta

Venus

2.3.2

Voorburg-Arentsburg, harbour

terracotta

Venus

2.3.2

Voorburg-Arentsburg, harbour

terracotta

head, mother goddess

2.3.2

Voorburg-Arentsburg, harbour

terracotta

head, mother goddess

2.3.2

Voorburg-Arentsburg, harbour

terracotta

mother goddess

2.3.2

Voorburg-Arentsburg, harbour

terracotta

mother goddess

2.3.2

Voorburg-Arentsburg, harbour

terracotta

Mercury

2.3.2

Voorburg-Arentsburg, harbour

terracotta

Mercury

2.3.2

Waal, river

bronze, complete

Mars

2.3.4

Waal, river

bronze, complete

Silenus

2.3.4

Waal, river

bronze, complete

Hercules

2.3.4

Waal, river

bronze, complete

dolphin

2.3.4

Waal, river

bronze

pedestal

2.3.4

Waal, river

terracotta

head, woman

2.3.4

Zwammerdam, fossil bed Rhine

terracotta

Venus

2.4.1

Zwammerdam, fossil bed Rhine

terracotta

head, mother goddess

2.4.1

Zwammerdam, fossil bed Rhine

terracotta, complete

swaddled infant

2.4.1

Category 5: south of the Rhine

168

Map 11 Find-spots of alleged ritual deposits categories 1-4 north and south of the Rhine Base map: paleogeographic reconstruction of the Netherlands around 100 (Vos/Vries 2013). 1. Naaldwijk 2. Valkenburg 3. Texel-Den Burg 4. Ede 5. Bennekom 6. Tiel-Badhuisplein 7. Wijchen 8. Nijmegen-west canabae 9. Helden 10. Montforterbroek 11. Buchten 12. Colmschate 13. Dalfsen 14. Tzum 15. Ouddorp 16. The Hague-Scheveningseweg 17. Voorburg 18. Leiden 19. Alphen a/d Rijn-De Schans 20. Schagen 21. Tiel-Passewaaij 22. Oosterhout (Gld) 23. Nijmegen-east-canabae 24. Wehl 25. Son en Breugel 26. Bergen op Zoom 27. Maasbracht 28. Kerkrade 29. Noordbarge

4.2.3

A secondary use of appliques

Among the aforementioned deposits, three bronze objects in the study area are not free-standing statuettes but appliques that once were attached to an object. The first is the bronze water nymph or Venus from the fill-

169

ing of a well at Naaldwijk-Hoogeland (South Holland, deposit 28, figs. 2.52-54). The second is the bronze putto from an enclosure ditch at Den Burg, on the island of Texel (North Holland, deposit 21, fig. 2.80). The third is a bronze panther, standing on a disc (deposit 33, fig. 2.187). It was discovered in a wet area near a stream, the Vlootbeek in Central-Limburg. Apparently, the panther was buried with one or more pottery vessels. To which objects the appliques belonged is uncertain. The water nymph/Venus was originally an attachment on the edge of a large object, such as a large bronze basin. The bronze putto was part of the support of a drinking vessel or the leg of a side table.1302 The panther originally decorated a large piece of furniture or a chariot, but later became a weight: its body and the disc on which it stands, were originally hollow, but later filled with lead. It was found together with a Hallstatt spearhead and potsherds. Earlier at the same location a pot filled with 36 silver denarii had come to light, of which the latest were struck under Severus Alexander.1303 The find-contexts suggest that the appliques have been deliberately broken off and were offered in a special place: a well, an enclosure ditch and a swampy area near a stream. The two deposits from Naaldwijk and Texel were made in the late Roman or early Medieval period. If the coin hoard and panther were part of the same deposit, the panther ended up in the ground after 235. 4.2.4

Deposits related to crafts?

In Penteskouphia, Greece, just southwest of ancient Corinth, over 1000 terracotta plaques were recovered from a nearby sanctuary that was dedicated to Poseidon. The majority of the plaques date to the 6th century BCE. They show scenes with Poseidon and Amphitrite, as well as scenes with craftsmen at work, such as potters with wheels and kilns and people digging for clay, woodcutters, ships, and so on. Many plaques have piercings for suspension. The plaques were offered by craftsmen addressing the sea god to support them in their work.1304 Crafts are often depicted in reliefs on tomb stones, sarcophagi and workshop signs.1305 Imagery of craftsmen at work in the Roman West is rare and has not been attested in the study area. Nevertheless, the presence of an altar on the industrial site Holdeurn (see 2.4.1), just south of Nijmegen, is related to the fabrication of tiles and pottery on the site. The altar is dedicated to Vesta, goddess of fire and hearths. The dedicator, a Julius Victor, had erected the altar pro se, for himself. Being the magister figulorum, he was head of the potters.1306 The altar was placed on a wall between the kilns and protected by roof tiles.1307 The location implies that the altar had been erected for the success of the firing process, or to ward off fire breakouts. Despite the lack of imagery, there might be three examples of deposits related to metalworking. The head of the terracotta mother goddess statuette from Son en Breugel, North Brabant (deposit 16 and 2.3.5) was buried in the southeast corner of a sunken byre, were also a lot of iron slag was found (deposit 16). The large bronze statuette of Mars from Helden-Schrames, Limburg (deposit 12 and 2.3.5), was buried in the corner of a sunken byre, where also a Frankish throwing axe came to light. Metal slag, nails, a bucket handle and a balance scale in the byre indicate that metalworking took place in the settlement. Yet, considering the location and the number of ritual deposits of axes and other objects in sunken byres, it seems not very likely that the statuette and the axe were buried to be recycled.1308 The pit near a goldsmith’s workshop at Wehl-Oldershove, Northern Gelderland (deposit 11 and 2.2.4), contained potsherds, iron objects, iron slag and the terracotta head of a Venus statuette.

1302

Veen 2017, 250-252.

1303

Peters 1984, 17; Hupperetz 1991, 172.

1304

Stissi 2002, 76-95, 313; Van Straten 1985, 224 and fig. 7.

1306

Janssen 1846, 337-339.

1305

A famous example from Rome is the tomb of the Haterii,

1307

Janssen 1846, plate XI, fig. 2.

a family of builders from the Flavian period. The reliefs

1308

See 4.2.2.

170

show scenes of the construction of buildings such as the Colosseum. See also Zimmer 1982.

Although there are more explanations for the burial of the three statuettes at Wehl, Son en Breugel and Helden, one of them could be that they were buried to watch over metallurgy activities in the settlements. 4.2.5

A comparison with British datasets

When studying cultural tradition and custom, a comparison with other Roman period datasets is indispensable. In the case of statuettes and statuette deposits, this has proven to be highly problematic. For example, a distinct difference between the dataset of this study and other datasets of Roman period statuettes is that the latter comprise either terracotta or bronze statuettes. This distinction provides a distorted picture with regard to the occurrence of and alleged preferences for deities and other types of representation.1309 Moreover, it is problematic to compare datasets of different entities, like datasets of statuettes collected from a single site or those from a region defined by modern borders. Yet, ignoring the existence of such datasets is not an option. Even a comparison between ‘small’ and ‘large’ datasets may help to detect certain tendencies, as we will see below. Besides the studies mentioned in this section, I did not find any other research on ritual deposits of Roman period statuettes in West-European settlement contexts, which hampers the study of patterning on a larger scale. In his study on terracotta workshops and statuettes from Britain and the continent, Frank Jenkins made a catalogue of terracotta statuettes with a description of the find-spots in Britain, which he completed in 1977.1310 The terracotta statuettes from Britain in Jenkins’ catalogue mostly come from excavations in the 1960s or earlier.1311 A comparison between the terracotta statuettes from the Netherlands and Jenkins’ dataset reveals some similarities, such as a strong preference for Venus and mother goddesses. A salient detail is that all identifiable mother goddess figurines from Jenkins’ dataset are of the dea nutrix type.1312 Striking differences between the two datasets are the small number of terracotta Minerva statuettes and the notable absence of terracotta Fortuna figurines in the British dataset, whereas Durham’s database contains 13 bronze specimens.1313 Matthew Fittock has examined the terracotta statuettes from Roman London. The London assemblage comprises 168 statuettes: 109 deities, 28 animals, 4 human figures and 27 specimens that fall in the category ‘other’.1314 He has compared the assemblage with datasets from the remainder of Britain, France and Switzerland. This comparison leads him to conclude that the London assemblage ‘reveals a rather distinctive pattern of consumption within the settlement compared with the rest of Roman Britain and continental Europe.’ 1315 Fittock, too, has observed the high number of Venus statuettes, 85 specimens on a total of 109 deities. He has also noted a relatively high proportion of Venus figurines around the Walbrook Stream that could have been deposits ‘as part of a water-oriented healing ritual or votive practice’.1316 Extensive excavations have been carried out at Silchester and Hampshire between 1890 and 1909, and from 1997 onwards by Michael Fulford and Amanda Clarke.1317 The early excavations have yielded numerous deposits from pits and wells consisting of one or more complete pottery vessels, pewter and bronze vessels, metal hoards other than coins, with ‘probable ritual associations’, over fifty complete dog skulls, skulls of pig and horse, bones of cattle and a few human remains.1318 Although not recognised as such in the late 1890s, also two deposits of statuettes were found. The first, a bronze figure that may represent the infant Hercules, was found in a ‘trench’ that ran between the forum and the north south-street to its east. Besides the bronze

1309 1310 1311

See chapter 1.8.

https://intarch.ac.uk/journal/issue31/2/3.25.html.

Jenkins 1977, 280-416.

1314

Fittock 2015, 118.

The results of the earlier excavations at the temple com-

1315

Fittock 2015, 118.

plex of Springfield, Kent, have been reassessed: Jarrett

1316

Fittock 2015, 119.

2008.

1317

Fulford 2001, 201.

1312

Jenkins 1977, 281-298.

1318

Fulford 2001, 202-203. See for the ritual nature of the metal

1313

Jenkins 1977, 280. See for Fortuna in Durham’s database:

hoards Fulford 2001, 206 and note 71.

171

statuette, the deposit contained pottery fragments and an iron screw. A ‘mutilated’ terracotta Venus figurine was found in a pit in an unspecified insula in the eastern part of the town.1319 Besides these deposits, Fulford has investigated more deposits in five other major settlements in southern Britain. A deposit in a timber-lined pit at London comprised eight substantially complete pots, among which three incense pots, two figurines, one, maybe both, representing Venus. The head and feet of both statuettes are missing.1320 The deposits from the six towns span a period from the late 1st century BCE into the 5th century.1321 Two more recent studies have provided additional information on deposits: Susheela Crease’s PhD thesis on deposits in watery places in Britain and Richard Jarrett’s reassessment of the finds from the temple complex at Springhead, Kent.1322 The information from Britain, however scanty, suggests that several statuettes were special deposits. For the identification of these deposits the same criteria have been applied as for the statuettes in the study area, although the descriptions in the available literature are sometimes brief or sketchy. Alleged statuette deposits in Britain Find-spot

Context

Composition/ condition statuette

Selection

Date

Reference

CATEGORY 1 Sanctuary/alleged cult place 1A Springhead- Kent

Temple complex, under building

Terracotta, lower body with base

Venus

After 100

Jenkins 1977, 333, Jarrett 2008, 268

2A SpringheadKent

Temple complex, ditch belonging to temple

Terracotta, base with feet, bronze arm, terracotta foreleg

Venus, human figure, horse

After 100

Jenkins 1977, 333, Jarrett 2008, 265-266

3A Flitwick-Bedfordshire Cult place on bank of ancient stream or bog

Ca. 25 terracotta statuettes. 21 of Venus, including 10 heads, 1 small spoked bronze wheel, 4 gems

Venus, mother goddess, Minerva, Bacchus, human figure

?

Jenkins 1977, 316-317, Fadden 2010

Find-spot

Composition/ condition statuette

Selection

Date

Reference

Terracotta, incomplete, head and lower legs missing Terracotta, incomplete, coins, latest AD350

Venus

?

Jenkins 1977, 344

Venus

After 350

Jenkins 1977, 330

Context CATEGORY 2 Building(s)/settlement

4A CaerwentMonmouthshire 5A Hemel HampshireHertfordshire

House, bottom foundation Villa, room

6A Holt DownHampshire

Villa, ‘rubbish pit’

Terracotta, incomplete, keys, nails, potsherds, coins 1st -4th c.

Venus

4th c. or later

7A Silchester-Hampshire

Settlement, pit in insula

Terracotta, incomplete

Venus

?

Jenkins 1977, 327, Taylor/Collingwood 1927, 208 Fulford 2001, 202-203

8A Silchester-Hampshire

Settlement, trench

Bronze

Infant Hercules?

?

Fulford 2001, 202-203

9A ChesterholmVindolanda

Vicus, floor level married quarters soldiers?

Terracotta, lacks head and base

Venus

After 3rd c.

Jenkins, 1977, 354, Birley 1973, 116, nr. 11.

1319

Fulford 2001, 203 and notes 23-24.

1322

Jarrett 2008, Crease 2015.

1320

Fulford 2001, 207.

1323

Although Crease doubts the ritual nature of the two

1321

Fulford 2001, 213.

172

deposits from Southwark, she does not rule out the possi-

Find-spot

Context

Composition/ condition statuette

Selection

Date

Reference

Terracotta, incomplete, building material, refuse

Venus

Late Roman

Crease 2015, 160, 354

CATEGORY 3 Ditch 10A Southwark, London

Settlement on south bank of Thames, ditch CATEGORY 4 Wells/drains/cisterns

11A CaerleonMonmouthshire

Main drain canabae legionis

Terracotta, at least 2 statuettes, incomplete

mother goddesses

After 140

Jenkins 1977, 290

12A South Shields-Durham

Well inside fort

Terracotta, incomplete

Venus

?

Jenkins 1977, 323

13A Wickford-Essex

Roman period well

Terracotta, complete

Venus

2nd-3rd century?

Jenkins 1977, 325

14A ChesterholmVindolanda

Vicus, above water-channel mansion

Terracotta, head, sherds of miniature vessel Gillam 42

Venus

Late Roman

Jenkins 1977, 354, Birley 1970, 144, fig. 5, nr.6

CATEGORY 5 River bank, gully, 15A Chichester-Sussex

Bank of river Lavant near Roman fort

Terracotta, substantially complete

Female figure

?

Jenkins 1977, 303, Taylor 1944, 152-154

16A Southwark, London

Settlement on south bank of Thames, pit or gully

Terracotta, incomplete, building material, refuse

Venus

Late Roman

Crease 2015, 160, 3541323

4.2.6

Placating the gods: in anticipation or afterwards?

It has been argued that offerings were made after the gods had granted a wish.1324 This assumption is based on a well-known Roman ritual, the ritual of the vow, the votum. According to Derks, the ritual of the vow was also known in Gaul, which he deduces from a fragment in De Bello Gallico. In this passage, Caesar states that Gauls would bring offerings at sacred places if their gods had granted them what they had asked for. They promised their god of war the spoils of war if they were victorious and to sacrifice the captured animals.1325 In

bility: ‘Whilst these finds appear to represent the remains

finds, nothing else from Southwark suggests ritual activity.

of a possible shrine or at the very least iconic figures of

These final two sites are less overt in any practices of ritual

classical religion, the recovery of these items in a ditch

deposition; however their evidence is still noteworthy and

and pit or gully with associated middens dating to the late

adds to the wider debates regarding practices of ritual

Roman period suggests the discard of either the classical

deposition within this study zone to be explored in the

religions in favour of another faith, or the demolition and/

remainder of this chapter.’: Crease 2015, 160.

or renovation of a domestic shrine or small temple that

1324

Derks 1991, Derks 1998, 217-231.

was subsequently demolished. However, aside from these

1325

Derks 1995, 123; Caes. BGall. 6.16-17.

173

my opinion, this fragment cannot serve as evidence for such practices in pre-Roman Gaul, because Caesar’s account could have been based on his own frame of reference, the rituals he was familiar with.1326 To understand how the votum worked in the centre of the Empire, the most important source of information are the acta of the fratres Arvales, a college of priests in Rome. Scheid has analysed the acta of this priesthood, which consist of over 200 inscriptions engraved on marble tables, found during excavations in a sacred grove, five miles outside Rome.1327 One of the duties of the fratres Arvales was to bring public sacrifices to ensure the welfare of the emperor. It was a highly formalized ritual that involved stating the names of the deities that were addressed, the favours that were asked for and the offerings that would be made afterwards. This part of this ritual was the nuncupatio, which would be engraved on a marble table and put on display. The offerings after the wish was granted always consisted of animal sacrifices, sometimes accompanied by gold, silver or a statue. If the vow was not fulfilled, the priests could decide not to make an offering, which underlines the contractual nature of the vow.1328 This, however, was a ritual carried out in Rome, by Roman priests and concerned the public votum. To what extent the formal character and rules of this ritual were maintained in the northwestern provinces is difficult to determine. Moreover, despite the numerous inscribed altars and other votive objects with formulas that testify to the practice of the votum, there is no reason to assume that all offerings were made afterwards. In a paper on offerings of infant figurines at river sources and their sanctuaries, such as Fontes Sequanae (Côte-d’Or, France), the French archaeologist Simone Deyts stated that these offerings could be made in advance. Early offerings at the source of the Seine were made between 40 BCE and 30 and comprised figurines and sculpted body parts of wood and stone.1329 Deyts pointed out that images of sick children were placed in the sanctuaries by parents as a ‘plea for help’, offerings in anticipation of divine help.1330 Derks disputes this, arguing that the offerings were made in the Roman period and that a plea for help does not correspond with the practice of the votum.1331 He supports his argumentation with inscriptions containing the words contra votum on graves of children who had died, despite the offerings their parents had promised the gods/God.1332 The majority of these inscriptions come from northern Italy: Rome, Milan and Aquileia, followed by Lazio, Campania, Hungary and the Balkans. One inscription comes from Trier, a second from Worms.1333 Most inscriptions date to the late Roman and early Medieval period.1334 Around fifty percent of the inscriptions mention a deceased child and a substantial number were made by Christians.1335 The question that comes to mind is whether contra votum inscriptions on graves of deceased children can be linked to offerings of infant figurines as part of the Roman votum ritual as described above, particularly with respect to Christians. Whereas 89 infant figurines have been found in Gaul and Germany, only one contra votum inscription comes from Gallia Belgica and one from Germania Superior.1336 Moreover, whereas infant figurines were offered 1326

this publication to my attention.

Another example is Tacitus’ statement that some of the Suebi venerated the Egyptian goddess Isis. Apparently, the

1330

Deyts 2004, 237, in Derks 2014a, 59 and note 49.

author had witnessed a sacrificial ritual with a boat, or

1331

Derks 2014a, 59.

heard about it: pars Sueborum et Isidi sacrifat; unde causa

1332

Derks 2014a, 60.

et origo peregrino sacro parum compero nisi quod signum

1333

AE 2007, 997; CIL 13, 6250 = AE 1994, 1301.

ipsum in modum liburnae figuratum: Some of the Suevi also

1334

The EDCS database search term contra votum has yielded

sacrifice to Isis. Of the occasion and origin of this foreign

238 inscriptions with the text contra votum. The earliest

rite I have discovered nothing, but that the image, which is

inscription dates to the 1st century, the latest to the middle of the 6th century.

fashioned like a light galley, indicates an imported worship: Tac.Germ. 9.2.

1335

1327

Scheid 1990.

1336

1328

Derks 1995, 111-127.

ny Derks 2014a, 52. See for the figurine of a swaddled infant

1329

Derks 2014a, 58 and note 40. I thank Ton Derks for bringing

from Zwammerdam 2.4.1 and 4.2.2.

174

Ehmig 2013, 321, 325. See for a table with infant figurines from Gaul and Germa-

from the late 1st century BCE onward, most contra votum inscriptions date to the late Roman and early Medieval period, including the inscription from Trier, which dates to the 5th century.1337 Another issue is the reason why people would mention a votum in a funerary inscription. According to Derks, the contra votum epitaphs testify to the anger and grief of the parents because they blamed God/the gods for the death of their child.1338 I would argue the opposite: the parents mention the votum in the epitaphs to testify to their piety, to show the world that they respected and accepted the will of God/the gods and that they were not angry with God/the gods. And lastly, there is also archaeological evidence of offerings that were made in advance. Foundation deposits were made to persuade supernatural powers to guard over the construction of the house, or to protect its future residents. An example is the ceramic situla at Heeten at the bottom of a posthole of a Roman period house, which was an offering made before the construction of the house was completed.1339 Besides the protection of loved ones and the curation of sick children there were probably many more situations which would incite people to invoke the help of a deity in advance, hoping that their gifts could persuade divine powers to act on their behalf. According to Scheid ‘Offerings are wrongly lumped together (…). Many of these objects were given as tokens of thanks, or to conciliate a god or to pay homage rather than in the fulfilment of a vow.’1340

4.3

Deposition of statuette fragments and deliberate fragmentation

4.3.1

Broken statuettes in the Neolithic and Bronze Age

Among the alleged deposits from the study area and Britain are several statuette heads. Before dealing with the question whether deliberate fragmentation was an existing practice in the northwestern provinces, the history and practice of deliberate fragmentation will be discussed. Deliberate fragmentation of statuettes and the deposition of fragmented figurines is part of a set of rituals that can be traced back to the Neolithic and Bronze Age. The Balkan archaeologist John Chapman published a study in 2000 on the ritual burial of fragmentary pottery vessels and figurines in the Neolithic and Copper Age Balkans. Here, selective deposition and deliberate fragmentation were part of daily life, aimed at connecting the living with the ancestors.1341 Among his explanations for broken, incomplete objects in the archaeological record two could be relevant to this study: 1. Objects are ritually ‘killed’ before deposition. 2. Objects are broken deliberately to be dispersed throughout the settlement and beyond as part of rituals related to enchainment. The first explanation refers to the practice of deliberately ‘killing’ artefacts, which was done by breaking them prior to their deposition. The reasons for this include the impossibility of reuse, fear of pollution, and, in funerary contexts, the wish to avoid association with the deceased. This practice is attested in Bronze Age Greece, where drinking cups were smashed in front of tombs, but also in Bronze Age barrows in Britain, where pottery sherds were deposited with the burial.1342 The swords of a late Bronze Age weapon hoard from

1337

Derks 2014a, 58, 60 and fig. 7.

1340

Scheid 2003, 100.

1338

Derks 2014a, 60-61.

1341

Chapman 2000, 23-79.

1339

See for this site 2.2.4.

1342

Chapman 2000, 23-25.

175

Pulle (Antwerp, Belgium) were intentionally broken and some had been burnt before deposition, possibly because they were ‘polluted’.1343 In the Roman period, the ritual of killing objects is repeatedly attested in Britain by pits or shafts with bent (‘killed’) swords.1344 In the study area at Alphen aan den Rijn, a comparison between the distribution of bronze and iron objects over the river bank and the terrain of the fort showed that nearly fifty percent came from the river bank. Among the objects were bent weapons, some of which bearing the name of the owner. The bent weapons have been interpreted as ritual deposits.1345 An example of deliberately broken and burnt statuettes before deposition in and near tombs in the Ardennes region in Belgium has been demonstrated by Jan de Beenhouwer.1346 After the Roman period, the deliberate deformation or destruction of objects continued, which is demonstrated by a gold hoard with deliberately coiled neck-rings from the Frankish period at Beilen (Drenthe).1347 Chapman’s second explanation not only provides a reason for the intentional breakage of objects but also for their dispersal. Deliberate fragmentation of objects involves the use of fragments for social practices of enchainment, creating relationships between the living and the ancestors or supernatural powers. Hereafter, I will briefly discuss examples of rituals with fragmented figurines in the Neolithic and Copper Age Balkans, followed by a healing ritual with figurines in Karnak, Egypt, from around 1500 into the 3rd century BCE. Fragmented figurines are known from excavated Bulgarian tells dating to the 4th and 3rd millennia BCE: Goljamo Delchevo, Vinica and Ovcharovo. Two thirds of the over 400 figurines at all sites were fragmentary. Chapman noted that the missing parts were removed from the tell occupation area and deposited at some distance from the dwelling area.1348 The Dolnoslav tell in South Central Bulgaria was excavated between 1983 and 1991. The dataset includes 500 anthropomorphic figurines, over 200 clay, bone and marble zoomorphic figurines, as well as single body parts, such as ears, busts or arms. Over 90 percent of the figurines were broken. Eighty percent of the corpus was deposited in a fragmentary state in a central part of the site. Fifteen percent of all fragments could be refitted within the site. Since the site was almost fully excavated, it has been argued that parts of the figurines were taken off site, or that only parts of figurines already broken on other sites were introduced into the Dolnoslav tell.1349 Outside the early Neolithic settlement of Grǎdinile, a pit was excavated containing a dozen deliberately broken, figurine heads. Since none of the heads matched the parts of figurines found inside the settlement, Chapman assumed that they were the result of a ritual performed beyond the site.1350 When deliberate fragmentation in the Balkans was recognised for the first time in the 1930s, it was discovered that some clay figurines were made in two parts, the upper and lower part held together with a peg. Apparently, these manufacturing techniques facilitated deliberate fragmentation. In the middle Neolithic period, body parts such as legs and heads were made as ‘figurines’ in their own right, as is known from sites in the Peloponnese, Greece. Separately modelled heads and bodies are also known from the late Copper Age.1351 As to the deposits of fragmented figurines, Chapman has argued that the context is crucial to their interpretation. Deposits of figurine fragments in pits, especially on a previously occupied site, must be seen as an exchange between the living and the ancestors. The figurine fragments in the pits connect the living to the ancestors, while matching fragments deposited elsewhere, are enchained to the same ancestors. Another

1343

Fontijn 2002, 169-170.

1344

See for ‘killed’ swords, for instance, Fontijn 2002, 169, 172,

1347

Quast 2009, 221. See for the hoard from Beilen also Waterbolk/Glasbergen 1955; Zadoks 1955.

235-236; Pleiner 1993, 160-161 and fig. 19. See for ‘ritually

1348

Chapman 2000, 55-56 and table 3.4.

S-bent’ swords that were bent after their owner was cre-

1349

Chapman 2000, 58-60; Chapman/Gaydarska 2006.

mated: Buchwald 2005, 120, 127-128.

1350

Chapman 2000, 64.

1345

Polak et al. 2004, 203.

1351

Chapman 2000, 69-72 and fig. 3.6.

1346

See 4.3.3.

176

example are the anthropomorphic figurines from the Selevac site in modern Serbia, which have been more often found in production areas and pits near houses than inside houses.1352 The fragmentation of figurines allows for categorisation with regard to gender and a preference for body parts or complete figurines. Many Neolithic figurines depict females with exaggerated breasts, hips, buttocks or genitals. In that respect, the Hamangia figurines from Durankuluk in north-east Bulgaria are no exception. They have become famous, not in the last place for two stone figurines, ‘the Thinker’ and his consort, which have become ‘the face’ of the Hamangia culture (ca. 5200-4500 BCE). They are not typical for the majority of the figurines, which represent naked, faceless female figurines with emphasized breasts and buttocks, a pregnant stomach and pubic triangle. 58 of the figurines came from graves, 42 from settlements. Whereas none of the figurines from settlements were complete, a substantial number of graves contained complete statuettes, a third of the group.1353 The second element, the choice of body parts, is demonstrated by Chapman in a table showing the distribution of body parts in relation to their find-context, settlements and graves. It reveals a clear preference for torso/leg fragments and complete figurines in graves. In settlements, complete figurines constitute the smallest category, whereas there is a preference for torso and leg fragments, followed by heads.1354 These examples from Chapman’s study illustrate how selective deposition was not limited to complete objects, but also involved selective deposition of fragments in specific places. The next paragraph is dedicated to a study on Egyptian female figurines. They do not represent supernatural beings, but human females that were used in healing rituals, after which they were deliberately broken and removed from the sanctuary because they were considered dangerous. 4.3.2

A healing ritual with female figurines in Egypt

Rituals with intentionally broken figurines of females have been studied by Elisabeth Waraksa. She has investigated the ritual use of 42 female clay figurines from the Mut temple precinct at Karnak, Egypt. The complex is located on the east bank of the Nile and contains six temples, including the Mut temple. The goddess Mut was the wife of Amun-Ra, whose temple is situated 325 metres north of the Mut precinct. The temple of Mut is first mentioned in the tomb of a high official who lived from around 1525 into the 15th century BCE. How long the temple exactly was in use is not known, but stelae dating to the reigns of the emperors Augustus and Tiberius mention construction work at the temple precinct. Excavations between 2001 and 2004 by Johns Hopkins University in the Mut temple and the surrounding area covered circa 1725 square meters. The figurines under discussion come from the surrounding area of the temple and have been deposited there from around 1525 into the 3rd century BCE.1355 Statuettes of nude females have been found throughout Egypt, dating from the 6th millennium BCE into the Roman period. Clay figurines of nude females have been recovered from houses and temples in the Nile Valley, burial sites in the western oases and mining sites in the eastern desert, the Sinai Peninsula and Nubian forts.1356 It was suggested in the 1930s that figurines of nude females were used in religious practices to promote fertility.1357 Waraksa has argued that the function of these figurines involved more than just ensuring fertility. She based her arguments on textual and archaeological evidence which provided more insight in their function. Typical for the figurines was their fragmentary state, representing either upper parts or lower bodies of nude females, exaggerated hips and pubic triangles. On many figurines traces of paint, red in particular, were visible. Furthermore, the figurines showed clear signs of deliberate breakage. Two to three figurines depicted a woman lying on bed, with a baby beside her.1358 1352

Chapman 2000, 72-73.

1356

Waraksa 2009, 12-13 and note 57.

1353

Chapman 2000, 75-78 and table 3.8.

1357

Waraksa 2009, 13-14 and note 68.

1354

Chapman 2000, 78 and table 3.9.

1358

Waraksa 2009, 20-38.

1355

Waraksa 2009, 1-12.

177

To sum up Waraksa’s conclusions, the figurines were used in rituals of healing and had two functions. They represented generic female, healthy persons who had to protect the invoked goddess against the disease or affliction she was being asked to fight. The goddess could be Isis, as mentioned in two spells, Mut or others. Only at the moment that the goddess was invoked in the ritual, the figurine briefly became the goddess through the magical words of the practitioner. When the figurine had absorbed the disease or repelled the affliction, it became a dangerous object and had to be broken before it could be removed. This explains why all figurines were found in refuse zones with large amounts of discarded pottery at a distance from the main sanctuary.1359 4.3.3

The Roman period: deliberate fragmentation in Belgium

An example of a Roman period ritual involving the deliberate breakage of statuettes comes from modern Belgium and seems restricted to tombs. Three find-spots of deliberately broken statuettes are situated in the Ardennes region in south-eastern Belgium. At Sberchamps, on top of a La Tène burial mound, an oval pit was dug in the 2nd century. In the filling of the pit iron nails, coins, a terra sigillata fragment, charcoal and many sherds of terracotta statuettes were found, but no cremation remains. Among the nine statuettes that could be identified, two depict Fortuna, one standing, one enthroned, and the third is a mother goddess with a dog in her lap. The fragments of the mother goddess and the standing Fortuna statuettes could be refitted into complete specimens. The statuettes had been smashed to pieces and showed traces of firing and heavy blows. The chronology of the ritual could be reconstructed as follows: first, the statuettes were damaged by violent blows, after which they were smashed to pieces and thrown into a fire, together with the other objects. Finally, the remains were buried in the pit on top of the burial mound. Three pits near graves at Longler-Respet (Neufchâteau), situated four kilometres from Sberchamps, contained 300 small fragments of at least 21 statuettes. Many sherds had scratches made with a dentated tool and some showed traces of firing. Finally, at Ortho (La Roche-en-Ardenne), several statuettes were found that had been smashed to pieces.1360 Here, too, some of the sherds showed traces of heavy blows.1361 4.3.4

Deposition of statuette fragments in the Netherlands

After having explored rituals with fragmented figurines in other parts of Europe and Egypt from the Neolithic into the Roman period, it is now time to discuss whether rituals with statuette fragments occurred in the study area. Most terracotta statuettes that have come to light during excavations in the Netherlands are incomplete. Complete specimens usually come from graves or they have been carefully deposited, like the terracotta water nymph from a defensive ditch at Leiden-Roomburg and the terracotta rattle from Alpen a/d Rijn-De Schans. The fragility of terracotta does not explain why archaeologists find single fragments while other fragments of the same statuette are absent. A first explanation could be that the other fragments became dispersed, by human activity, bioturbation or other natural processes. It is also possible that statuette fragments were not recognised during the analysis of finds and ended up between the potsherds. The third explanation is that there were no other fragments.

1359

Waraksa 2009, 102-104, 167-169.

1360

The context of the finds is unknown.

178

1361

De Beenhouwer 1990, 9-21.

As to the first explanation for single statuette fragments, it is possible that one fragment has been separated from others by water or sediment transport. The option that other statuette fragments have been overlooked is unlikely, since the finds from deposits 1-31 have been analysed by experts. If a single statuette fragment depicts the head of a deity, it could be a ritual deposit. According to the definition formulated in this study, we are dealing with a ritual deposit if the statuette head meets one of the two other criteria: the find-spot and/or presence of other, remarkable objects. Single statuette heads come from settlement features and buildings in Tiel-Passewaaij (deposits 17, 27, 30), Son en Breugel (deposit 16), Ouddorp (deposit 26) and Voorburg (deposit 31). Statuette heads that are part of a striking assemblage are the statuette heads from the pits at Wehl and Kerkrade-Holzkuil (deposits 11 and 13). the three statuettes from Oosterhout (deposit 25) and deposit 3A from Flitwick-Bedfordshire, which contained 25 terracotta statuettes, including ten heads of Venus. The broken figurines in the three deposits from Belgium (4.3.3) were part of a ritual which also involved mutilation and deliberate breakage. If we accept that statuette heads were deposited as part of a ritual, it raises the question whether statuettes were deliberately broken prior to their deposition. In most cases it is impossible to recognise such practices. Only the following observations can be made: if a statuette head was the desired body part, obtaining it through breakage would be more difficult if the mould was fashioned from a single piece.1362 Breaking the figurine at a wider point, the torso, would be safer, to make sure that the head remained undamaged.1363 ‘Loose’ bronze heads are rare, but not absent in the dataset. The three bronze heads from sanctuaries were once attached to an object: the small head of a male figure from the Empel sanctuary, the head of a male figure from the temple at Maastricht and the donkey head in the Noordbarge assemblage (fig. 2.37).1364 The deposits comprise two terracotta heads of Minerva, one of Venus, one head of a woman and a fragmentary head of Apollo or Bacchus (figs. 2.28-32). The deposit(s) also include(s) two statuettes with missing heads: the bronze boar and the enthroned terracotta warrior statuette (figs. 2.34-36 and 2.38). The number of heads in the Noordbarge deposit(s) is striking, but the find circumstances of the statuettes do not allow for any conclusions.1365 Is there a link between the deposition of statuette heads and busts in the study area and the veneration of the head? In central Italy, between the 6th and 2th centuries BCE, clay heads were produced for religious purposes.1366 In Germany and France, the emphasis on heads is attested by Iron Age stone sculpture and images such as the Pfalzfeld pillar, the Heidelberg head and the sanctuaries of Roquepertuse and Entremont (France). Examples of stone heads from the Roman period are known from Eype, Carvoran, Caerwent and Wroxeter (Britain) and Leichlingen (Germany). A bronze statuette from Bouray (Seine-et-Oise, France) depicts a god with an enormous head and a tiny body.1367 Diodorus Siculus (ca. 90 - 30 BCE) and Strabo (ca. 64 BCE - 23) mention the importance of the head to the northern peoples. Since their accounts on this subject are quite similar, I only cite Diodorus here: ‘When their enemies fall they cut off their heads and fasten them about the necks of their horses; and turning over to their attendants the arms of their opponents, all covered with blood, they carry them off as booty, singing a paean over them and striking up a song of victory, and these first-fruits of battle they fasten by nails upon their houses, just as men do, in certain kinds of hunting, with the heads of wild beasts they have mastered. The heads of their most distinguished enemies they embalm in cedar-oil and

1362

See, for instance, the moulds from Toulon-sur-Allier: Van

1365

Boekel 1987, 217-218, 222, 224, figs. 10-11, 15, 17. 1363

1364

The bronze head from Maastricht is on display in Hotel Derlon, Maastricht. See for the find circumstances of the

The Cybele statuette from Oosterhout, for example, is bro-

donkey head 2.2.4.

ken at a wide point. See fig. 2.60.

1366

Rüpke 2018, 84-86 and notes 6-8.

See for the bronze head from the Empel sanctuary and the

1367

Green 1992c, 114-115.

bronze head from the temple at Maastricht 2.3.5.

179

carefully preserve in a chest, and these they exhibit to strangers, gravely maintaining that in exchange for this head some one of their ancestors, or their father, or the man himself, refused the offer of a great sum of money.’ 1368 Ritual deposits of animal and human skulls in the study area also testify to the important role of the head in Iron Age and Roman contexts, even though we do not grasp their meaning. In the northern terp region, in the Englum terp, Nieuwhof has identified an Iron Age deposit consisting of eight human skulls.1369 Other Iron Age deposits from the terp comprise a dog skull in the hearth of a house, a dog skull and a fragment of a human skull in a creek.1370 Still north of the Rhine, a substantial number of deposits of animal bone with ritual intent have been identified at a late Roman settlement at Schagen-Muggenburg (North-Holland).1371 South of the Rhine, Groot has identified skull deposits in the Roman period settlement at Tiel-Passewaaij: two horse skulls in a large boundary ditch, a cattle skull in a house ditch, a dog skull in a ditch, the skull of a mare in a well, and two sheep skulls in another.1372 A horse skull was found in a ditch at Geldermalsen-Hondsgemet (Gelderland). On the same site, a horse skull was buried together with two lower legs.1373 The Roman period settlement at The Hague-Scheveningseweg has yielded the skull of a young man, which was found in a pit.1374 Another human skull comes from one of the enclosure ditches belonging to the castellum at Alphen aan den Rijn (figs. 4.1-2).1375 In Britain, over fifty dog skulls were discovered during the excavations between 1890-1909 at Silchester (Hampshire), several coming from pits and wells.1376 The skull of a child aged twelve to fourteen was found in a well in insula VI of the town. The skull of a cat was found at the bottom of a well, the skulls of a horse, sheep and oxen at the bottom of another, the skull of a pig in a pit in a house.1377 Deposits of statuette heads could be related to the deposition of human skulls. Even if there is a connection, many questions still remain unanswered. Were deposits of skulls and deposits of statuette heads part of the same ritual? Were such rituals a continuation of the Iron Age rituals with skulls that have been attested in the Netherlands and Britain? What was the meaning of skulls and statuette heads in such deposits? Were statuette heads a substitute for skulls? Were skulls and statuette heads a pars pro toto? To cite Richard Bradley again: we have too much material at our disposal and not enough ideas with which to address it. Another issue that remains unsolved is whether only statuette heads were deposited as part of a ritual. Considering the fact that settlement features such as ditches and wells can act as artefact traps, I have attempted to minimize the coincidence factor by narrowing down the definition in this study to statuette heads or heads plus a part of the torso. Other fragments have been excluded, since they depict varying parts of a statuette. Yet, I believe that some of these fragments may have been deposited as part of a ritual because they depict specific body parts like breasts, female genitals or the lower body. Examples are the lower body of a Venus statuette from a well at Naaldwijk-’t Zand Heultje and a Venus fragment from Jutphaas.1378

1368

Diod. Sic. 5.29: τῶν δὲ πεσόντων πολεμίων τὰς κεφαλὰς

1371

Therkorn 2004, 27-55.

ἀφαιροῦντες περιάπτουσι τοῖς αὐχέσι τῶν ἵππων· τὰ δὲ σκῦλα τοῖς

1372

Groot 2008, 121-125.

θεράπουσι παραδόντες ᾑμαγμένα λαφυραγωγοῦσιν, ἐπιπαιανίζοντες

1373

Groot 2009, 5.

καὶ ᾄδοντες ὕμνον ἐπινίκιον, καὶ τὰ ἀκροθίνια ταῦτα ταῖς οἰκίαις

1374

Van Ginkel/Waasdorp 1992, 41. See for this site 2.4.2.

προσηλοῦσιν ὥσπερ οἱ ἐν κυνηγίοις τισὶ κεχειρωμένοι τὰ θηρία. τῶν

1375

See for this site 2.4.1.

δ᾿ ἐπιφανεστάτων πολεμίων κεδρώσαντες τὰς κεφαλὰς ἐπιμελῶς

1376

Fulford 1999, 202.

τηροῦσιν ἐν λάρνακι, καὶ τοῖς ξένοις ἐπιδεικνύουσι σεμνυνόμενοι

1377

Fulford 1999, 203, 205.

διότι τῆσδε τῆς κεφαλῆς τῶν προγόνων τις ἢ πατὴρ ἢ καὶ αὐτὸς

1378

See for Naaldwijk-’t Zand Heultje 2.3.2; for the Venus-frag-

πολλὰ χρήματα διδόμενα οὐκ ἔλαβε. See also Strab. Geo. 7.4.5.

ment Van Boekel 1987, nr. 123. See for a British example of

1369

Nieuwhof 2015, 127-133.

deposited fragments depicting body parts Fadden 2010.

1370

Nieuwhof 2015, 205, 217, 317.

180

4.3.5

Deposition of bronze caducei

Besides the deposition of statuette heads, also distinctive attributes of deities may have been deposited, possibly as a pars pro toto. An example is the herald’s staff, the caduceus, carried by Mercury. In Colchester, Uley and Gloucestershire (Britain) and Martigny (Switzerland), single caducei have been found in sanctuaries dedicated to Mercury. It is assumed that not all caducei originally belonged to Mercury statuettes, which is why it is argued that they were offerings in their own right.1379 They were cast separately and had been attached to an object, for instance to function as an amulet. The latter can be deduced from scratches on the back of these caducei. In the study area, four bronze caducei have come to light, all four from Nijmegen: three from military camps and one from Ulpia Noviomagus. One of the four caducei that will be discussed hereafter may have been an amulet. Whether the other three originally belonged to a statuette of Mercury cannot be established. The first bronze caduceus (fig. 2.139) comes from the early Roman military camp on Kops Plateau, Nijmegen, and was found above one of the ditches of the first phase of the fort.1380 Scratches on the back of this caduceus suggest that the object was an amulet.1381 The second came to light in the north-east corner of the legionary fortress on the Hunerberg, in a pit with charcoal, a coin and 4th-century rouletted Argonne ware.1382 Its preserved upper part is 112 mm in length. In view of the abandonment of the legionary fort around the middle of the 2nd century, a deposition date in the 4th century for the caduceus seems improbable, but cannot be excluded.1383 The third caduceus comes from a late Flavian water channel on the grounds of Canisius College. The feature has a terminus post quem of 89.1384 The length of the preserved top of the staff is 25 mm (fig. 2.134). From the channel also another bronze object came to light. It depicts a torch and may have belonged to an Amor figurine.1385 The fourth caduceus, from Maasplein (fig. 2.91), will not be discussed here, since no data on the find circumstances are available yet. Unfortunately, the data on the find circumstances of the three caducei are scanty. Nevertheless, the presence of these bronze objects in a ditch, a water channel and a pit, while the rest of the statuette is missing, could suggest that the caducei were intentionally deposited.

4.4

Magical practices with figurines

The word ‘magical’ in the title of this paragraph is not without problems. What is the difference between the practices that have been discussed in the previous paragraphs and the rituals that we label as magical? What is the difference between a prayer and a spell, between invoking Jupiter or a chthonic deity? Where does religion stop and magic begin? Many attempts have been made to define what magic is. In the previous century, Hendrik Versnel has proposed the following definition for magical practices: ‘Just like religions, ‘magical’ practices or expressions may share some though not all family resemblances. This means that we may accept a broad, polythetic or prototypical definition of magic, based on a ‘common sense’ collection of features (…). It is a specific manifestation of religion, usually ‘negative’, manipulative and coercive, with short-term and often individual goals, whereas most dedications were associated with positive aspects of religion.’ 1386

1379

Van Zalinge 1997, 20 and note 21. See also Durham 2012, nrs.

earlier level.

1038-1040, 702, 706-710.

1383

See 2.4.1; Zadoks et al. 1969, nr. 80.

See for this site 2.4.1.

1384

See 2.4.1.

1381

Van Zalinge 1997, 20.

1385

Haarhuis 1991, 82-83.

1382

The 4th-century sherds may have sunken into a deeper,

1386

Versnel 1991, 184, 186; Adams 2006, 1-2.

1380

181

In a recent publication on magic in the Greco-Roman world, Drawing Down the Moon, Radcliffe Edmonds points out that we can only come up with etic definitions, since the enormous temporal and cultural differences between our modern cultures and those of the ancient Graeco-Roman world do not allow for an insider perspective.1387 Fortunately, ancient authors occasionally provide evidence from the inside: descriptions of what they considered as magic. And there is also the phenomenon of self-labelling, either of oneself or of one’s practices as magic.1388 In this respect, an important source of information are the Greek Magical Papyri, a corpus of papyri with magical spells, hymns and instructions for magical rituals from Graeco-Roman Egypt, dating from the 2nd century BCE to the 5th century. The figurines that constitute the subject of this paragraph possess characteristics that distinguish them from the other statuettes of the dataset. These characteristics include their anthropomorphic but indistinct shape and the tiny piercings in the head and/or rest of the body, which is why they are often called voodoo-dolls. To understand their function, I will first briefly discuss rituals with such figurines in the Mediterranean, where they have been found in association with curse tablets. Rituals with figurines and spells in the Mediterranean are known from the late 6th century BCE onwards. Here, the figurines are accompanied by inscribed curse tablets, known as defixiones. The Greek verb κατάδειν, defigere in Latin, mean to bind, tie or nail.. The texts on the lead sheets are usually binding spells.1389 The corpus of curse tablets comprises over 1700 examples, mostly written in Greek. The earliest date to the late 6th or early 5th century BCE and come from the Greek colony of Selinus in Sicily. Around the end of the 8th century the writing of curse tablets has come to an end.1390 Figurines and spells have often been recovered from graves, but they have also been found in sanctuaries. 1391 Graves symbolise the realm of the chthonic deities, the Underworld. In defixiones the help of chthonic and non-chthonic deities can be invoked. Five categories of spells have been defined, which I will briefly mention here. Defixiones judiciariae are litigating spells, aimed at harming or denaturing the opponent(s) in trials. They mostly date to the 5th and 4th centuries BCE. Many of them have been recovered from the Agora and the Kerameikos cemetery in Athens, but examples are known from many centuries and places throughout the ancient Graeco-Roman world. Defixiones amatoriae are erotic spells aimed at separation or attraction. Defixiones agonisticae are athletic and circus spells aimed at other competitors. Trade spells are aimed at rivals in trade and date to the classical and Hellenistic period, but are particularly known from the Greek Magical Papyri. The last category comprises ‘prayers for justice’, often petitions made by victims of theft who want restitution of stolen goods, or revenge.1392 Finds from the Kerameikos cemetery provide an important clue for understanding the relation between the figurines and the accompanying curse tablets. In a grave dating to around 400 BCE, a text with a list of names of nine men was found, with the addition that it also applied to ‘everyone who would act as witness or accuser’. The text was engraved in the lid of a small box, which contained a lead male figurine. It had a clearly modelled penis and the hands were tied behind its back. One of the names from the list was engraved in the right thigh of the figurine: Mnesimachos. The meaning of the find assemblage is clear: The text alone was not enough. Apparently it was necessary to model a figurine with tied hands to demonstrate the fate Mnesimachos would be facing. The lid of the box had two piercings, made with a sharp, pointy object, most likely a nail. Comparable examples of figurines with names were recovered from other graves in the vicinity.1393 1387

Edmonds 2019, 6-7.

1388

Edmonds 2019, 19-32.

temple at Mainz and the temple of Jupiter Heliopolitanus at Carnuntum: Witteyer 2003; Gassner 2010.

1389

1392

Graf 1996, 111.

1390

Ogden 1999, 5.

curse tablets dealing with theft come from Roman Britain,

1391

A well-known find-spot of defixiones in Western Europe is

from the Sulis-Minerva sanctuary at Bath and the Mercury

the sanctuary of Sulis-Minerva at Bath. Clay figurines from Roman period temples in Germany that have been interpreted as Zauberpuppen come from the Isis/Magna Mater

182

Ogden 1999, 3-5, 31; Graf 1996, 110. The largest number of

sanctuary at Uley: Adams 2006, 7-8. 1393

Graf 1996, 122-124.

Not only the containers, also figurines were pierced with nails. A well-known example from Egypt is the clay figurine of a kneeling young woman, nowadays in the Musée du Louvre, Paris (fig. 4.3). The statuette was found in a vessel, together with an inscribed curse tablet. Her hands are tied behind her back. Her eyes, mouth, skull, torso hands and genitals are pierced with nails. The text contains precise instructions for the ritual. It involved the modelling of two figurines, one male, one female. Magical words had to be written on all body parts of the female figurine. Then, the figurine had to be pierced with thirteen bronze nails for the head, ears, eyes, mouth, torso, hands, genitals and soles of the feet. Each time the following words had to be repeated: I pierce NN, I pierce her brains, to make her think of me and me alone. I pierce NN, I pierce her ears, …and so on. After that, instructions followed for the making of the curse tablet and the deposition of the figurine and tablet in the grave of a murdered person or someone who died too young, at sunset, together with a bunch of seasonal flowers.1394 In this case, it appears from the description that the needles were not meant to kill the subject of the spell, but to force her to abandon her man for the practitioner of the spell. Recent finds from Roman period villas, houses and sanctuaries in western Europe demonstrate that these so-called magical rituals were not restricted to the Mediterranean. Twelve handmade clay figurines have been found in Germany, Britain and France. Unlike the figurine from the Musée du Louvre, they all have a crude, vaguely human form, without any details. Edmonds explains the reason: the less direct and detailed the image, the more powerful the link to the original.1395 Excavations in Nijmegen-West (Ulpia Noviomagus) in 1995, yielded a thirteenth specimen. The clay figurine is 103 mm in height and depicts a standing male figure with two stumps for arms, partly preserved legs, protruding buttocks and genitals (fig. 2.119). The eyes and mouth are indicated by small indentations or piercings. The nostrils in the large nose are tiny piercings. Small piercings are also visible in the area of the ears, above the clearly indicated genitals, in the stomach, under the buttocks, arms and right foot (figs. 2.120-22). Since the figurine depicts a male, the intended victim of the ritual was a man. Although not linked to this figurine, a curse tablet came to light in 2016 in the same area.1396 The lead strip, 103 mm by 50 mm, bears a text of two lines. Some letters are written in mirror-writing, others are upside down. It is an example of so-called Schadenszauber, directed to a person whose name is mentioned in the text: a woman named Vitilia who is cursed to die.1397 Besides the curse tablet from Nijmegen another specimen came to light at Bodegraven (South Holland). Here, traces have come to light of a small military fort, which was constructed around the middle of the 1st century. The tablet has the shape of a tabula ansata and mentions 20 names of individuals, divided over three columns. Earlier, Haalebos suggested that the names refer to members of military units.1398 Although this remains uncertain, Swinkels has pointed at the diversity of the names, which include Celtic, Germanic and Greek roots. With the help of another curse tablet from Chagnon, France, he has suggested the following reconstruction of the last line: ‘may they have been wiped out in the same way as this.’1399 From Vechten (Utrecht) comes a lead strip with the text Smertuccus Amaionis f(ilius).1400 Sometimes, simple sketches, symbols or letters are engraved instead of words. An example is the lead curse tablet found on the premises of castellum Matilo at Leiden (South Holland). The letters OBA on the tablet possibly refer to the Hebrew name Jahwe or Jehova.1401 1394 1395

Graf 1996, 124-125.

1398

Haalebos/Polak 2007, 114-122.

Which is why individualized figurines and portraits were

1399

sic ext(r)iti s(i)nt quomodo h(oc) e(s)t. Swinkels 2018, 197-

not used for defixiones: Edmonds 2019, 63 and note 14. A

1396

217.

few elaborately formed figurines come from the fountain

1400

CIL 13, 8822. See also Byvanck 1935, 175 nr. 294.

of the goddess Anna Perenna in Rome: Edmonds 2019, 63.

1401

Raven 2010, 183. See for the tablet also Brandenburgh/Hess-

The figurine symbolises a man, while the intended victim

ing 2014, 54.

of the curse tablet from Nijmegen-West is a woman. 1397

Vit(i)lia PV[2-3]+ · adnuntio tibi ad istas manes: Weiß-König 2018, 161-165.

183

Recently, Magali Bailliot has published the twelve magical figurines from western Europe, the specimen from Nijmegen not included.1402 The figurines have several characteristics in common, such as the hand-modelled, vaguely defined body, head and limbs and the piercings.1403 The conclusion of Bailliot’s survey is that the figurines from western Europe often come from important places and have been recovered from large buildings, exceptional villas, sanctuaries and provincial capitals.1404 To start with the latter, in Mainz, Mogantiacum, the capital of Germania superior, three clay figurines were recovered from the sanctuary of Isis and Magna Mater. The first figurine comes from a well outside the temenos. The second figurine has disintegrated. The third depicts a victim mentioned in a rolled-up sheet found in the vicinity. The figurine has male and female genitals.1405 In the cellar of a domus or schola, in Durocortorum, the capital of Gallia Belgica (Reims, Champagne-Ardennes, France), three deposits, including two figurines, were placed against the north wall. One of the two figurines was made of lead and placed in a container made of two terra sigillata bowls. An iron key was placed against the bowl, probably referring to the act of sealing.1406 Two villas have yielded magical figurines. The first was found in the villa at Fishbourne (Sussex, England), the second comes from a villa in Piriac-surMer (Pays-de-la-Loire, France). The figurine from Nijmegen-West comes from the filling of a cellar and was found together with bricks, terra sigillata and other potsherds. Since processing of the data from this part of Ulpia Noviomagus is still pending, nothing is known about the nature and dimensions of the building the cellar belonged to. For the time being, the only thing that can be observed is that it is not the only figurine that was found in a cellar. Two comparable figurines come from the cellar of a domus or schola in Reims. In a broader context, our figurine came to light in Ulpia Noviomagus, the civitas capital of the Batavi, which corresponds with Bailliot’s conclusion that this type of figurines often come from important places. However, this could also be a consequence of the focus of excavators on important places like provincial and civitas capitals. Moreover, being places where a lot of different people lived together in a small area, such towns were a fertile breeding ground for friction, frustration and rivalry.

4.5

Conclusions

Based on literature about recurring find-spots of ritual deposits and the criteria formulated in the first chapter, five categories have been distinguished which contain statuettes and statuette heads, 33 deposits in total. The first category comprises deposits from sanctuaries and alleged cult places. The other categories are deposits made in association with houses or other buildings, deposits in ditches and deposits in wells, cisterns and drains. The last category comprises statuettes from rivers, river banks, streams and gullies. They include mostly complete bronze statuettes, one bronze pedestal, one complete terracotta rattle, seven terracotta heads and a few fragmented specimens. It is advanced in this chapter that besides statuettes, also statuette heads were deposited in settlements features like pits, wells and ditches. They may have been deposited as part of a ritual related to the life cycle of a feature, building or settlement. The earliest deposits from a dateable context are deposits 19-20, from the western canabae at Nijmegen. The two deposits come from the same Flavian pit but probably were two separate deposits. Slightly later 1402

Bailliot 2015, 93-111.

1403

Whereas West-European figurines are mostly made of

and piercing. Some of the figurines had been placed in containers: Faraone 1991, figs. 3-13.

clay, Christopher Faraone’s overview shows that all Greek

1404

Bailliot 2015, 105-106.

specimens are metal figurines, either bronze or lead. A clay

1405

Bailliot 2015, 100.

figurine comes from Puteoli, Italy. The rituals with these

1406

Bailliot 2015, 103-104.

figurines include binding or twisting of the limbs, burning

184

may be deposit 30 from a well or cistern in the eastern canabae, between 75 and 125. All other deposits from dateable contexts were made after 150, mostly in the 3rd and 4th centuries or later. A comparison has been made between the Dutch dataset and British datasets of terracotta and bronze statuettes. The British and Dutch datasets share a strong preference for terracotta female figures, Venus and mother goddess figurines in particular, in deposits as well as elsewhere. Whereas the mother goddess figurines from Jenkins’ and Fittock’s dataset are exclusively represented by the dea nutrix type, in the study area there is a distinct preference for seated mother goddesses with fruit or a dog on their lap, while mother goddesses of the dea nutrix type are represented by only four specimens. Since terracotta Fortuna statuettes frequently occur on the continent, the absence of terracotta Fortuna statuettes in the British datasets is striking, all the more because Durham’s database yields 13 bronze Fortuna statuettes. Based on these datasets and available studies on the subject, 17 possible ritual deposits of statuettes have been identified in Britain. The only site where statuette heads came to light is the cult-place at Flitwick-Bedfordshire, where 21 terracotta Venus figurines were found, of which ten Venus heads. The second part of this chapter begins with on rituals that involve deliberate breakage and deposition of statuette fragments. Among Chapman’s explanations for the presence of deliberately broken objects in the archaeological record in the Neolithic, deliberate breaking (‘killing’), dispersion and deposition of broken objects seem also have been part of rituals that were performed in later periods. A study of Egyptian Bronze Age clay figurines from the Mut precinct demonstrates that figurines depicting nude, healthy females were used in healing rituals after which they were deliberately broken and deposited at a distance from the main sanctuary. In the Belgian Ardennes region, deliberately mutilated and broken terracotta figures were subsequently burnt and deposited in the Roman period, as part of rituals that were performed in the context of graves. The last paragraph is dedicated to magical practices with figurines. Such practice were often aimed at manipulating, harming and even killing individuals. Through papyri and engraved texts in lead curse tablets we know that the figurines symbolised living individuals. In the northwestern provinces they were hand-modelled from clay, whereas specimens in the Mediterranean were often made of lead. Twelve of such figurines have been recovered from large buildings, villas and sanctuaries in Germany, France and Britain. A thirteenth clay specimen came to light in the filling of a cellar at Nijmegen-West. The clay figurines have small piercings all over their body, which is why they are often referred to as ‘voodoo-dolls’. While piercing the figurine with needles or nails, the practitioner invoked the help of supernatural powers. These powers could be deities like Mercury or Neptune, or chthonic deities and demons. The latter could be the reason why in Greece many figurines have been recovered from graves in ancient cemeteries. Contact with the underworld could explain why Roman period figurines have been recovered from cellars, like two figurines from Reims and the figurine from Nijmegen-West.

185

5

Summary and conclusions

Central to this research are Roman period bronze and terracotta figurines of deities, human figures and animals, the majority coming from established find-spots in the Netherlands. From the 1960s onward, bronze statuettes have generated numerous catalogues and studies focusing mainly on typologies and production centres. Terracotta statuettes received less attention and the first comprehensive study in the Netherlands on this subject saw the light in 1987. Bronze and terracotta statuettes have not been studied together and seldom in relation to their find-spots. Roman period statuettes have been found throughout Europe, in Asia Minor and North-Africa. In northwestern Europe, the statuettes were introduced by Roman troops at the end of the 1st century BCE. They were used in different ways. Besides offerings in sanctuaries they were set up in household shrines and niches in houses, deposited in graves as funerary gifts and deposited in dry and wet places in and around settlements as part of a ritual. The anthropologist Kopytoff has advanced that we can draw up cultural biographies of objects. Prior to their deposition in a sanctuary, grave or other place, statuettes could have had other functions, which usually remain invisible to us. They could have been a diplomatic gift, a means of payment, a toy, a curiosum or souvenir. In some cases, a previous function can be deduced from specific characteristics of a statuette. An example is the terracotta bird from Schagen, which originally was a toy, but ended up as a ritual deposit under the entrance of a building. This study is dedicated to the use of statuettes in people’s everyday life, the daily praxis of religion, a concept known as lived religion. In this study it will be explored which statuette representations people preferred in different areas and what they did with them. The daily praxis of religion must be seen as an indispensable complement in the study of religion. It focuses on the individual as a religious agent and the rituals he or she performed in the context of his/her house or settlement. In order to explore this topic, four research questions were formulated, the first two being: 1. Can we observe relations between size, material and subject of statuettes and their find-spots in the three areas and if so, how can we explain these relations? 2. Were there preferences for specific subjects or iconographies in the three areas and if so, how can we explain these preferences? In order to answer these questions, chapters 2 and 3 are dedicated to a description of the find-spots and a spatial and iconographic analysis of statuettes and subjects in the study area, which I have divided into three parts: the area north of the Rhine, the area south of the Rhine and the Dutch part of the Roman limes zone. The area north of the Rhine was after 47 no longer part of the Roman Empire, whereas the area south of the Rhine remained for at least three centuries under Roman rule. Besides numerous small rural settlements, the region south of the Rhine also comprised villa complexes. Here, four settlements developed into regional centres with urban characteristics: Nijmegen (Ulpia Noviomagus), Voorburg (Municipium Aelium Cananefatium), Maastricht and Heerlen (Coriovallum). The limes zone comprises all forts, military vici and canabae legionis on the left bank of the Rhine between Nijmegen and the North Sea, as well as the forts and extramural settlements between Katwijk and Aardenburg along the North Sea coast. The spatial and iconographic analysis has revealed relations between find-spot, size, material and subject, as well as preferences for subjects and iconographies, although the small numbers must always be regarded with caution.

187

A relation between find-spot and size results from a comparison between the areas north and south of the Rhine. The dataset presents two groups of bronze statuettes. A group ‘large’, measuring over 110 mm in height, and a group ‘small’, between 50 and 90 mm. The average size of bronze statuettes from the northern terp region appears to be distinctly larger than the average height of bronze statuettes from military and non-military settlements south of the Rhine. It is argued in this chapter that this could be related to the function of large bronze statuettes, which were brought to the north by Romans and traders as diplomatic gift or as a means of payment. Most small bronze statuettes have been recovered from find-spots in and around settlements south of the Rhine, in the limes zone in particular. Also, a relation between material and subject has been observed, a relation that is not tied to a specific area and constitutes in fact a relation between material and gender. It results from the dataset that statuettes of female figures are mostly made of clay, whereas males are usually depicted in bronze. Since the demand for figurines increased in the course of the 1st century, clay was cheap and easily accessible and, thus, the most logical choice for the production of figurines. A relation between find-context and material emerges from the find-spots of statuettes in cemeteries in the study area and beyond. The 44 identifiable figurines from cemeteries and graves in the study area, comprise 40 terracotta and four amber specimens. Bronze statuettes are absent. The avoidance of bronze statuettes in graves is attested throughout the Roman Empire. Outside the study area, a few examples, sometimes made of lead, come from the fringes of the Empire. Although bronze objects seem already to have been avoided in graves in the Bronze Age, this is not a satisfying explanation for the absence of metal statuettes in Roman period graves, because bronze vessels and other bronze objects often occur in graves. It is not unlikely that bronze statuettes in graves were avoided to prevent the deities they represent from being ‘infected’ by death. Since terracotta figurines of female figures occasionally occur in late Roman graves in the study area, it is possible that terracotta was attributed different properties than bronze, or that the idea of infection by death started to change in the late Roman period. Amber figurines in the study area have been found in rich graves of women, which may be explained by the healing and apotropaic qualities of amber and the fact that only wealthy women could afford amber objects and jewellery during their life time. A relation between find-spot and subject becomes evident north of the Rhine, where the majority of the bronze statuettes represent Mars and Mercury. South of the Rhine, there were distinct preferences for figurines of female figures, mother goddesses and Venus in particular. A comparison with British datasets has revealed the same preferences for Venus and mother goddesses, with two differences. The mother goddesses in the British dataset are all of the same dea nutrix type, whereas mother goddesses in the study area predominantly depict women sitting on a chair with a high backrest, holding a dog or basket with fruits in their lap. A second difference is the absence of terracotta Fortuna statuettes in British datasets, whereas the goddess is well-represented in bronze. The relatively high number of Hercules statuettes from find-spots in or near military settlements points at the popularity of the hero among soldiers and veterans. This also appears from dedications mentioning Hercules and Hercules Magusanus. Most dedications to Hercules Magusanus come from military contexts in Germania Inferior, Gallia Belgica, Britannia and Dacia and were made by soldiers and veterans. Only one of the 17 dedicatory inscriptions to Hercules Magusanus explicitly refer to the Batavian (and Thracian) origins of the dedicators. Inscriptions mentioning Hercules/Hercules Magusanus have not been found in or near the civitas capital of the Batavi, Ulpia Noviomagus. Based on the available evidence it is argued that Hercules/ Hercules Magusanus was venerated by soldiers and veterans in the northwestern provinces and not by Batavians in particular. Another relation between find-spot and iconography is demonstrated by a distinct category of humans and animals. Their iconographies are characterised by physical deviations. The group humans comprise dwarfs and pygmies, the latter with a large erect penis. The animals are ithyphallic dogs and horned birds. They have been recovered from military settlements and cemeteries. This relation can be explained by the apotropaic function of these figures.

188

The Roman occupation had not only far-reaching consequences for the economic, social and political organization of the conquered peoples but also for native ideologies and religion, which became the subject of the so-called Romanization debate. The academic discussion focuses on the role of social classes, the expression and creation of identities and the acceptance, transformation or refusal of elements of new material culture. By selecting elements of a culture, new cultures and identities were created and new iconographies were forged. To explore these processes in sculpture, two groups of statuettes have been examined: deities with Roman and non-Roman characteristics and Roman-looking deities with the name of a native deity or a double name on the pedestal. Mother goddesses are an example of the first group, of which at least a part represented ancestral mothers, since their proper name refers to a certain group or region, like the matronae Aufaniae in the Rhineland or the matres Treverae outside Ubian territory. It has been observed that terracotta statuettes of the Roman goddesses Fortuna and Minerva took over the enthroned pose of mother goddesses while their bronze counterparts maintained their standing pose. It seems, therefore, plausible that these Roman goddesses were also attributed the qualities of mother goddesses. The overwhelming popularity of Venus could be explained in the same way. Although she is usually depicted standing, her naked appearance comes closest to the mother goddess archetype. The second group regards sculptures of Roman-looking deities while the inscription mentions a native deity. An example in the study area is (I)Seneucaega, who possesses the iconography of the Roman hunter-goddess Diana. A second example is the native goddess Nehalennia, whose iconography developed in the 2nd century. She is usually depicted in an aedicula, with the enthroned pose and attributes of mother goddess figurines and, sometimes, with the attributes of Fortuna. Examples outside the study area are statuettes of the native goddesses Naria (Switzerland) and Rosmerta (France) who are depicted as Fortuna. An example of a native male deity is Intarabus (Belgium), who is depicted with iconographic elements of Bacchus, Mars and Mercury. Statues, an altar and coins with the name Hercules-Magusanus depict the god as the Roman Hercules. The fourth chapter focuses on the ritual deposition of statuettes and statuette fragments. It is advanced in this study that, like pottery, metal objects and bones, statuettes and statuette fragments, too, were selected for ritual deposition. Based on these presumptions, the last two research questions have been formulated: 3. How can we identify ritual deposits of statuettes? 4. Are there any indications for the ritual deposition of statuette fragments? Statuettes and statuette heads have been recovered from recurring find-spots in settlements, in or in association with a building, in pits, wells and ditches. On the edges or in the vicinity of settlements, statuettes have been recovered from wet contexts such as river banks and fossil river beds. Their recognition is only possible if deposition patterns can be detected. A problem is that find-spots of discarded or lost artefacts also present a pattern, since they often end up in low-lying areas as a result of post-depositional processes. Distinguishing between these post-depositional patterns and patterns of ritual behaviour is possible if we take into account the composition of deposits, the condition of the objects involved and the observed deposition patterns in other studies. To this end, criteria have been formulated for the identification of ritual deposits of statuettes and statuette fragments. I have attempted to minimize the coincidence factor by narrowing down the definition in this study to statuette heads or heads plus a part of the torso. This has resulted in five categories of deposits and the identification of 33 possible deposits of statuettes and statuette heads north and south of the Rhine. The first category comprises the deposits from sanctuaries and alleged cult places. It is explained in the first chapter that the identification of a few alleged cult places is uncertain, because it is mainly based on the presence of statuettes. Categories 2-4 consist of ritual deposits in association with buildings, deposits in ditches, wells, drains or cisterns. A few deposits belonging to these categories have been recognised at the time by the excavators. Some have been found in direct association with a building, in a posthole, near or under the entrance of a

189

building. These rituals are particularly known from the Iron Age and continue into the late Roman period. The deposits were made at the beginning, end, or during the habitation period. Deposits in wells, pits and ditches are often related to the life cycle of these features, of a building or a settlement. The composition of several deposits in these categories often contain recurring or remarkable objects. Examples of recurring objects are other statuettes, complete pottery vessels or large quantities of potsherds, dateable different centuries, coins and fibulae, bronze or iron items such as fittings, nails and keys. These artefacts also occur in other deposits in and outside the study area. Category 5 comprises deposits of bronze and terracotta statuettes from riverside areas, river banks, streams and gullies. The identification of these deposits is uncertain for several reasons, including the dynamics of the landscape and the lack of archaeological data. In order to detect comparable deposition patterns of figurines outside the study area, the criteria formulated in the first chapter have been applied to British datasets. This resulted in the identification of 17 ritual deposits. A salient detail is that nearly all British deposits consist of terracotta Venus statuettes. The Dutch deposits show a preference for female figures, but not for Venus in particular. Chapter 4 begins with three studies dedicated to the deliberate breakage of figurines. They include rituals in the Neolithic Balkans, the Egyptian Bronze Age and the Roman period in the Belgian Ardennes. An early Neolithic deposit of deliberately broken figurine heads at Grǎdinile, Romania, demonstrates that the ritual deposition of figurine heads goes further back than the Iron Age and points at a very ancient and widespread tradition. Eleven deposits in the study area comprise terracotta statuette heads, while other fragments of the same statuette are absent. It is advanced that the deposition of statuette heads and the ritual deposition of skulls in and outside the study area could be related. There could also be a connection with the veneration of the head, which is attested by literary sources, sculpture and sanctuaries in Italy, France and Germany. In the last section of chapter 4 a group of remarkable ritual deposits is discussed. The deposits consist of hand-modelled figurines. They are usually made of clay and are often referred to as ‘voodoo dolls’, because it appears from archaeological and epigraphic evidence that the figurines were pierced with nails or needles by individuals who wanted to manipulate or harm another, living person. The figurines symbolise the victim of the curse. A substantial number of lead and clay figurines are known from Greece and Italy, where they are often found together with a curse tablet, usually an inscribed lead strip or sheet. The figurines were used in rituals which involved invoking the help of supernatural powers such as chthonic deities or demons, while piercing the figurine. In western Europe, twelve specimens have been recovered from large buildings, villas and sanctuaries in Germany, France and Britain. A thirteenth specimen with distinct male genitals has been recovered from the cellar of a building at Nijmegen-West. An inscribed curse tablet came to light at Nijmegen-West in 2016. It is the fourth specimen from the study area, the clay figurine the first.

Discussion and suggestions for further research The division of the study area into three areas was made because I presumed that political, cultural and ideological differences in these areas could have an impact on the choice of statuette representations. The results of the spatial and iconographic analysis demonstrate that there are differences in the distribution of statuette representations between the regions north and south of the Rhine. The analysis has revealed some patterns, differences and preferences, subtle as they may be. Several questions have to remain unanswered, such as: Is the veneration of female figures such as Venus and mother goddesses typical for northwestern Europe or was it a more widespread phenomenon? Were Roman goddesses also venerated as mother goddesses in other parts of the Empire? Why do statuettes seldom occur in graves in the study area and if they do, why predominantly in late Roman graves? Why were most ritual deposits of statuettes made in the late Roman period? These are all questions that would benefit from further research, for which a comparison with European studies and other datasets is indispensable.

190

Another question that came to mind when I was studying excavation reports is the purpose of pits filled with potsherds, metal, and/or other objects. The description of the filling of a pit is usually followed by the interpretation ‘refuse pit’. The contents of several pits I encountered in publications did not struck me as a mixture of household refuse. First, because some pits contained potsherds from several centuries as well as fragmented or complete metal objects and second, because a single statuette fragment or bust was also part of the filling. In order to corroborate or falsify the idea that such pits are not refuse pits, it could be worthwhile to reassess data from published excavations in and outside the study area, in order to compare the location and contents of pits in settlements, with and without statuette fragments. With the exception of wells and, in some cases, ditches, nearly all bronze and terracotta statuettes recovered from watery places lack contextual data. These places involve river areas, fossil or existing rivers, river beds and banks, streams and gullies. I have advanced that these statuettes could have been buried near or deposited in the water, as is now generally excepted for metalwork, weaponry, helmets and other objects. In many cases, the precise location of the find-spot of the statuettes is unknown. If they were ritual deposits, establishing whether this was a special place in antiquity is often no longer possible, because of landscape dynamics and human activity in modern times. A fortunate trend is the increasing collaboration between metal-detectorists and professional archaeologists, institutions and provincial registration points. The registration and mapping of find-spots of artefacts could help identifying depositional sites in the landscape. By studying these sites, it might be possible to find out whether they possessed specific characteristics that could explain why people buried or deposited artefacts on these particular locations. It could also be fruitful to examine whether find-spots of metal artefacts from the Roman period and beyond correspond to Iron or Bronze Age depositional sites, as has also been suggested by Fontijn. If people deposited not only complete statuettes but also single statuette fragments, the number of deposits in and around settlements is much higher. In this research, for instance, I have not taken into account the fragments of Venus figurines that depict a breast, genitals or lower body. Yet, their presence in watery places could indicate that they were deposited as dedications. This has already been suggested for fragments of Venus statuettes from Roman London. Even statuette fragments in settlement features that do not depict a particular body part could have been a ritual deposit. Since the dispersion of statuette fragments and potsherds in settlement features is known from the prehistoric Balkans, abandonment rituals in the pre-Roman Iron Age and Roman period settlements could echo these practices. Like potsherds and other objects, statuette fragments, too, may have been deposited in different places around a house before it was abandoned. If we wish to catch a glimpse of the ideas and believes of people in antiquity, questioning our observations in the archaeological record could be more rewarding than ignoring what defies a straightforward interpretation.

191

Abbreviations ADC AJA APL ArchCl AAS AWN BAAC BABesch BAR BAR Int. Series BEFAR BJb BLAN BROB BSBH CCJ EDCS EPRO HOP JALC JRA JRA JRS KJb KNOB LANX LIMC NAR NOWELE OMROL RAM RAAP RGM RM RMO ROB TRAC VF ZAR

192

Archeologisch Dienstencentrum Amersfoort American Journal of Archaeology Analecta Praehistorica Leidensia Archeologia Classica Amsterdam Archaeological Studies Archeologische Werkgemeenschap Nederland Bureau voor Bouwhistorie, Archeologie, Architectuurhistorie en Cultuurhistorie Bulletin Antieke Beschaving British Archaeological Reports BAR International Series Bibliothèque des Écoles Françaises d’Athènes et de Rome Bonner Jahrbücher Bureau Leefomgevingskwaliteit/Archeologie Nijmegen Berichten van de Rijskdienst voor het Oudheidkundig Bodemonderzoek Bijdragen tot de studie van het Brabants Heem Cambridge Classical Journal Epigrafische Datenbank Clauss-Slaby Etudes préliminaires aux Religions orientales dans l’Empire romain Haagse Oudheidkundige Publicaties Journal of Archaeology in the Low Countries Journal of Roman Archaeology Supp. SeriesJRA Supplementary Series Journal of Roman Studies Kölner Jahrbuch Koninklijke Nederlandse Oudheidkundige Bond Lanx. Rivista elettronica della Scuola di Specializzazione in Archeologia, Università degli Studi di Milano Lexicon Iconographicum Mythologiae Classicae Nederlandse Archeologische Rapporten Northwestern European Language Evolution Oudheidkundige Mededelingen uit het Rijksmuseum van Oudheden te Leiden Rapportages Archeologische Monumentenzorg Regionaal Archeologisch Archiverings Project Römisch-Germanisches Museum (Cologne) Mitteilungen des Deutschen Archäologischen Instituts. Abteilung Rom Rijksmuseum van Oudheden (Leiden) Rijksdienst voor het Oudheidkundig Bodemonderzoek Theoretical Roman Archaeology Conference De Vrije Fries Zuidnederlandse Archeologische Rapporten

Bibliography Ancient literary sources Apuleius Apologia 63 Caesar De Bello Civili 2.5 De Bello Gallico 4.2.3, 5.24.4, 6.5.4, 6.16, 6.17 Cicero De Oratore 2.266 in Verrem (actio 2) 2.5.72 Clement of Alexandria Protrepticus 4.40 Digesta 48.6.1 Diodorus Siculus Bibliotheca Historica 5.29, 22.9 Florus Epitome 1.38.10 Herodotus Historiae 4.59 Hesiod Theogonia 190-200 Isidorus Hispalensis Origines 18.39.1 Juvenalis Satyrae 9.138 Lucretius De rerum natura 5.75, 5.308 Ovid Fasti 6.310 Metamorphoses 2.364-366, 9.690-692 Petronius Satyrica 29.8 Phaedrus Fabulae Aesopiae 3.17 Plato Hipparchus 228d Plautus Curculio 470-475 Pliny the Elder Naturalis Historia 4.15, 4.17.106, 12.2, 12.3, 16.2-3, 34.45, 35.91, 37.12.49-51

193

Ptolemy Geographia 2.9.5, 2.11.8 Tacitus Annales 1.51, 1.73, 3.63, 4 . 72 -74, 12.22, 14.12, 14.32, 15.29, 15.44-45 Germania 9, 29.1, 40 Historiae 4.12, 4.15, 4.56, 4.66, 5.19-20 Theophrastus Charakteres 16.5 Thucydides Historiae 6. 27 Tibullus Elegiae 1.11-12 Varro De lingua Latina 5.57 Vergil Aeneis 2.172 Eclogae 7.61

Other literature Aarts, J. 2015, Munten, in Derks/De Fraiture 2015, 106-112. Adams, G.W. 2006, The social and cultural implications of curse tablets (defixiones) in Britain and on the continent (Studia Humanoria Tartuensia 7.A.5), 1-14. ut.ee/klassik/sht/2006/adams2.pdf (1-12-2020) Aldhouse-Green, M.J. 2012, An Archaeology of Images. Iconology and cosmology in Iron Age and Roman Europe, London/New York2. Aldhouse-Green, M. 2014, Dying for the Gods. Human sacrifice in Iron Age and Roman Europe, Stroud3. Andreussi, M. 1996, Mercurius, aedes, in Steinby, E.M. 1996 (ed.), Lexicon Topographicum Urbis Romae III, Rome, 245-247. Babelon, E/J.A. Blanchet 1895, Catalogue des bronzes antiques de la Bibliothèque nationale, Paris. Bailliot, M. 2015, Roman Magic Figurines from the Western Provinces of the Roman Empire: An Archaeological Survey, Britannia 46, 93-110. Bakker, J.T. 1994, Living and Working with the Gods. Studies of evidence for private religion and its mother goddessial environment in the city of Ostia (100-500 AD), Amsterdam. Bastiaensen A.A./S. Langereis/L.G.J.M. Nellisen 1999 (translation), Johannes Smetius, Nijmegen, stad der Bataven II, Nijmegen. Bauchhenß, G.1984, Ares/Mars, Mars in den nordwestlichen Provinzen (Gallien, Germanien, Britannien), LIMC II, 559-580. Bauchhenß, G. 1992, Mercurius, LIMC VI, 537-554. Bauchhenß, G. 2008, Hercules in Gallien – facts and fiction, in Haeussler, R./A.C. King 2008 (eds), Continuity and Innovation in Religion in the Roman West, vol. 2 (JRA Supp. series 67), Porthsmouth/Rhode Island, 91-102. Bauchhenß, G. 2013, Füllhorner und andere Nebenseitemotiven, in Hofeneder, A./P. de Bernardo Stempel et al. 2013 (eds), Théonymie celtique, cultes, interpretatio , X. Workshop F.E.R.C.A.N., Paris 24.-26. Mai 2010, Wien. Bauchhenß, G. 2014, Die Weihealtäre aus Bonn, in Busch/Schäfer 2014 (eds), 155-178. Bauchhenß, G./G.Neumann 1987 (eds), Matronen und verwandte Gottheiten. Ergebnisse eines Kolloquiums veranstaltet von der Göttinger Akademie Kommission für die Altertumskunde Mittel- und Nordeuropas (BJb, Beiheft 44), Bonn/Cologne.

194

Beex, G.A.C. 1973, Finds in North Brabant, BROB 23, 159-188. Bell, C. 1997, Ritual. Perspectives and Dimensions, New York. Besselsen, E. 2014, Onderzoekskader en onderzoeksvragen, in Driessen/Besselsen 2014, 6-8. Besuijen, G.P.A. 2008, Rodanum. A Study of the Roman Settlement at Aardenburg and its Metal Finds, Leiden. Besuijen, G.P.A./R.M. van Dierendonck 2013, Inleiding, in Van Dierendonck/Vos 2013, 9-19. Besuijen, G.P.A./R.M. van Dierendonck 2013, De geschiedenis van het onderzoek, in Van Dierendonck/Vos 2013, 21-34. Besuijen, G.P.A./R.M. van Dierendonck/W.K.Vos/N.J.G. de Visser 2013, Sporen en structuren, inVan Dierendonck/Vos 2013, 87-208. Bink, M./P.F.J. Franzen 2009, Forum Hadriani. Definitief Archeologisch Onderzoek (BAAC-Rapport A-05.0125), ’s-Hertogenbosch/Deventer. Binsfeld, W. 1964, Zu den römischen Töpfereien am Rudolfplatz in Köln, KJb 7, 19-33. Birley, E. 1986, The deities of Roman Britain, in Haase, W. 1986 (ed.), Aufstieg und Niedergang der RömischenWelt II.18.1, Berlin, 3-112. Birley, R.E. 1970, Excavations at Chesterholm-Vindolanda 1967-1969, Archaeologia Aeliana 4th series 48, 97-156. Birley, R.E. 1973, Vindolanda-Chesterholm 1969-1972. Some important finds from the vicus, ArchaeologiaAeliana 5th series 1, 116-117. Billoret, R.1976, Informations archéologiques. Cutry, Gallia 34-2, 352-353. Bloemers, J.H.F. 1978, Rijswijk (Z.H.)‚‘De Bult‘, Eine Siedlung der Cananefaten I-III (Nederlandse Oudheden 8), Amersfoort. Bloemers, J.H.F. 2016, Four Approaches to the Analysis of (Pre-)Roman Nijmegen.Aspects of cultural evolution, acculturation, contextual function and continuity (Nederlandse Oudheden 19), Amersfoort. Bloemers, J.H.F./M.D. de Weerd 1982, Het Romeinse kampdorp van de Brittenburg, Leids Jaarboekje 1983, 245-251. Blom, E./E.P. Graafstal 2001, Aanvullend Archeologisch Onderzoek Hoge Woerd (ADC-rapport 73), Amersfoort. Blom, E./W.Vos 2008 (eds), Woerden-Hoochwoert. De opgravingen 2002-2004 in het Romeinse castellum Laurium, de vicus en het schip‚‘de Woerden 7‘ (ADC-rapport 910), Amersfoort. Boardman, J./O. Palagia/S. Woodford 1988, Herakles, LIMC IV, 738-831. Bödecker, S. 2010, Waffen für Vagdavercustis. Ausgrabungen im Tempelbezirk auf dem Kalkarberg, Kreis Kleve, Der Limes 4, Heft 2, 16-19. Boeles, P.C.J.A. 1951, Friesland tot de elfde eeuw. Zijn oudste beschaving en geschiedenis Leeuwarden2. Bogaers, J.E.A.T. 1952, Bewoning uit de Romeinse tijd, 2e helft van de 2e en 1e helft van de 3e eeuw na Chr., Rockanje (Prov. Zuid-Holland), BROB 3, 5-8. Bogaers, J.E.A.T. 1955, De Gallo-Romeinse tempels te Elst in de Over-Betuwe, The Hague. Bogaers, J.E.A.T. 1958, Baexem, Bulletin KNOB 11, 182. Bogaers, J.E.A.T. 1960, Civitas en stad van de Bataven en Canninefaten, BROB 10-11, 263-317. Bogaers, J.E.A.T. 1962, Un taureau à trois cornes provenant de Beilen, province de Drenthe, BROB 12, 579-581. Bogaers, J.E.A.T. 1963, Ruraemundensia, BROB 12, 57-86. Bogaers, J.E.A.T. 1964, Baexem, Bulletin KNOB 17, 62-63. Bogaers, J.E.A.T. 1967, Die Besatzungstruppen des Legionslagers von Nijmegen im 2. Jahrhundert nach Christus, in Internationales Limeskongress. Studien zu den Militärgrenzen Roms: Vorträge des 6. Internationalen Limeskongresses in Süddeutschland, Cologne, 54-76. Bogaers, J.E.A.T. 1970a, Romeins Nijmegen tussen ‘castra’ en ‘oppidum’, Numaga 17-3/4, 97-101. Bogaers, J.E.A.T. 1970b, Terug naar de Lallenberg, Brabants Heem 22-2, 57-67. Bogaers, J.E.A.T. 1984, Bedenkingen bij een beer uit Xanten, Westerheem 33, 2-4. Bogaers, J.E.A.T./J.K. Haalebos 1975, Problemen rond het Kops Plateau, OMROL 56, 127-178. Bogaers, J.E.A.T./J.K. Haalebos 1979a, De Romeinse opgravingen op het terrein van het voormalig St.-Jozefpensionaat te Woerden, Heemtijdinghen 15, 76-84.

195

Bogaers, J.E.A.T./J.K. Haalebos 1979b, Woerden, Bulletin KNOB 78, 99-100. Bogaers, J.E.A.T./J.K. Haalebos 1980, Graven in Hatert, Numaga 27-1, 1-9. Bogaers, J.E.A.T./J.K. Haalebos 1986, Aan de grens van Ulpia Noviomagus. Opgravingen in Nijmegen-west (Bronsgeeststraat, Dijkstraat, 1985), Numaga 33-1, 1-10. Bogaers, J.E.A.T./J.K. Haalebos 1988, Opgravingen op het terrein van het voormalige Canisiuscollege, 1987, Numaga 35, 25-41. Bolla, M. 2013, Bronzetti in contesti funerari di età romana, LANX 15, 1-50. Bonamici, M. 1992, Mousa, mousai (Etruria), LIMC VI, 681-685. Bos, J.M./M.J.L. Niekus/J.Scheffer/T.B. Volkers 1997, Opgravingen in Winsum-Bruggeburen. Romeinen in Friesland!, Paleo Aktueel 9, 65-69. Bosman, A.V.A.J. 1997, Het culturele vondstmateriaal van de vroeg-Romeinse versterking Velsen 1, PhD thesis, Amsterdam. Bosman, A. 2012, Romeinen en Velsen - castello cui nomen Flevum, Westerheem, special Kennemerland (6), 61, 357-369. Boucher, S. 1970, Bronzes (grecs, hellénistiques et étrusques) des musées de Lyon II, Lyon. Boucher, S. 1971, Bronzes antiques. Vienne (Inventaire des collections publiques antiques 17), Paris. Boucher, S. 1973, Bronzes romains du Musée des Beaux-Arts de Lyon IV, Lyon. Boucher, S. 1976, Recherches sur les bronzes figurés de Gaule pré-romaine et romaine, Rome. Boucher, S. 1983, Les bronzes figurés antiques, Lyon. Boucher, S. /J.P. Boucher 1988, Bronzes antiques I. Musée d’Évreux. Collections archéologiques, Évreux. Boucher, S./H. Oggiano-Bitar 1993, Le Trésor des bronzes de Bavay, Lille. Boucher, S./S. Tassinari 1976, Bronzes Antiques I, Paris. Bowman, A/R.Tomlin/K. Worp 2009, Emptio Bovis Frisica: The ‘Frisian Ox Sale’ Reconsidered, JRS 99, 156-170. Bouma, J.G. 1996, Religio votiva: the archaeology of Latial votive religion The 5th-3rd c. BC votive deposit southwest of the main temple at ‘Satricum’ Borgo Le Ferriere. The votive deposit in a diachromic and synchronic perspective, PhD thesis, Groningen. Bracke, M./G. de Mulder/W. de Clercq et al. 2015, Ritueel gedeponeerde bijlen in een Romeinse potstal aan de Ringlaan in Brecht (prov. Antwerpen), Lunula. Archaeologia protohistorica 23, 63-68. Bradley, R. 2017, An Archaeology of Offerings. Deposits of valuables in the landscapes of ancient Europe, Oxford/Philadelphia. Brandenburgh, C.R/J. de Bruin 2016 (eds), Met de voeten in het water. Archeologisch onderzoek aan de oostzijde van castellum Matilo te Leiden, Leiden. Brandenburgh C.R. J./W.A.M. Hessing 2014, Matilo-Rodenburg-Roomburg. De Roomburgerpolder: van Romeins castellum tot moderne woonwijk, Leiden. Briard, J. 1991, Mythes et symboles de l’Europe préceltique. Les religions de l’âge du bronze (2500 - 800 av. J.-C.), Paris. Brouwer, M. 1989, Alphen a/d Rijn-De Schans, in Archeologische Kroniek van Zuid-Holland over 1988. Holland 21-6, 326-327. Brunsting, H. 1937, Het grafveld onder Hees bij Nijmegen. Een bijdrage tot de kennis van Ulpia Noviomagus, Amsterdam. Brunsting. H./D.C. Steures 1994, The Lone Watchman. The find circumstances of the Augustan iron helmet from Nijmegen and the date of the Augustan legionary fortress, OMROL 72, 101-11. Buchwald, V.F. 2005, Iron and Steel in Ancient Times, Copenhagen. Budischovsky, M.C. 1973, Jupiter-Ammon et Méduse dans les forums du nord de l’Adriatique, Aquileia nostra 144, 201-220. Buijtendorp, T.M. 2010, Forum Hadriani. De vergeten stad van Hadrianus. Ontwikkeling, uiterlijk en betekenis van het ‘Nederlandse Pompeji’, Amsterdam. PhD-thesis VU University Amsterdam. Buiks, C.1997, Laatmiddeleeuws landschap en veldnamen in de baronie van Breda, Assen.

196

Bult, E.J./P.Vons 1990, Een gesloten Romeinse muntvondst op het Marktveld bij Valkenburg (ZH): Archeologische aanwijzingen voor militaire aanwezigheid in de Rijnmond vóór 30 n. Chr., in Bult, E.J./D.P. Hallewas 1990 (eds) Graven bij Valkenburg III. Het archeologisch onderzoek in 1987 en 1988, Delft, 57-71. Busch, A.W./A. Schäfer 2014 (eds), Römische Weihealtäre im Kontext, Cologne. Byvanck, A.W. 1932/1933, De inscriptionibus Traiectensibus nuper repertis, Mnemosyne 60, 193-198. Byvanck, A.W. 1935, Excerpta Romana. De bronnen der Romeinsche geschiedenis van Nederland 2, The Hague. Byvanck, A.W. 1943, Nederland in den Romeinschen Tijd 1-2, Leiden. Cancik, H./J. Rüpke 1997 (eds), Römische Reichsreligion und Provinzialreligion, Tübingen. Camporeale, G. 1997, Zeus/Tinia, LIMC VIII, 400-421. Caroll, M. 2013, Ethnische Tracht und römische Kleidung am Niederrhein, in Wieczorek, A./ R. Schulz/M. Tellenbach (eds), Die Macht der Toga. Mode im Römischen Weltreich, Regensburg, 223-228. Caroll, M. 2012, The Insignia of Women: Dress, Gender and Identity on the Roman Funerary Monument of Regina at Arbeia, Archaeological Journal 169-1, 281-311. https://www.researchgate.net/publication/285949313_’_The_Insignia_of_Women_’_Dress_Gender_and_Identity_on_the_Roman_Funerary_ Monument_of_Regina_from_Arbeia. Accessed 30-11-2019. Cassari, P. 1998, Sui clipei del Foro di Augusto, ArchCl 50, 391-407. Cassari, P. 2004, Iuppiter Ammon e Medusa nell’Adriatico nordorientale: simbologia imperiale nella decorazione architettonica forense, Rome. Cazanove, O. de 2013, Enfants au maillot en contexte cultuel en Italie et en Gaule, Les Dossiers d’Archéologie 356, 8-13. Chapman, J. C. 2000, Fragmentation in Archaeology. People, places and broken objects in the prehistory of South Eastern Europe, London/New York. Chapman, J.C. /B. Gaydarska 2011, Parts and Wholes: object categorisation and fragmentation in prehistoric context. http://archaeologydataservice.ac.uk/archives/view/partwhole_ba_2006/overview.cfm (9-11-2016). Chevallier, R. 1994, Un buste en bronze de Jupiter Ammon, Bulletin de la société nationale des antiquaires, 128-137. Clarke, J.R. 1995, Looking at Lovemaking. Constructions of sexuality in Roman art. 100 BC – AD 250, Los Angeles. Clarke, J.R. 2005, Three Uses of Pygmies and Aethiops at Pompeii: Decorating, ‘Othering’ and Warding off Demons, in Bricault, L./M.J.Versluys/P. Meyboom 2005 (eds), Nile into Tiber. Egypt in the Roman World. Proceedings of the 3rd International Conference of Isis Studies Leiden, May 11-14 2005, Leiden, 155-162. Clarke, S. 1999, In Search of a Different Roman Period: the Finds Assemblage at the Newstead Military Complex, in Fincham, G./G. Harrison et al. 1999 (eds), TRAC 99. Proceedings of the Ninth Annual Theoretical Roman Archaeology Conference Durham 1999, Oxford, 22-29. Crease, S.M.E. 2015, Rethinking Ritual Traditions: Interpreting Structured Deposition in Watery Contexts in Late  Pre-Roman Iron Age and Roman Britain, online PhD thesis, London, http://discovery.ucl. ac.uk/1466183/1/Susheela%20Crease%20Final%20thesis%20doc..pdf (16-12-2016) Cunliffe, B.W. 1993, Fertility, Propitiation and the Gods in the British Iron Age, Amsterdam. Daniels, M.P.M. 1927, Romeinsch Nijmegen. II. Ulpia Noviomagus, OMROL 8, 65-111. Daniels, M.P.M./H. Brunsting 1955, De monumenten van steen (Beschrijving van de verzamelingen in het Rijksmuseum Kam te Nijmegen IV. Overdruk uit OMROL 36), Leiden. Dasen, V. 1993, Dwarfs in Ancient Egypt and Greece, Oxford. De Beenhouwer, J. 1990, Les statuettes gallo-romaines en terre cuite trouvées dans une fosse sacrale à Sberchamps, Archéo-Situla 5, 8-22.

197

De Beenhouwer, J. 2005, De Gallo-Romeinse terracottastatuetten van Belgische vindplaatsen in het ruimer kader van de Noordwest-europese terracotta-industrie. PhD. diss. Catholic University of Leuven, http://hdl. handle.net/1979/189 (17-1-2015). De Beenhouwer, J. 2014, Atuatuca. Terracotta figurines and devotion in Roman Tongeren (Publications of the Gallo-Roman Museum 5), Tongeren. De Bruin, J./G.P.A. Besuijen/H.A.R. Siemons/R.J. van Zoolingen 2012, Goedereede Oude Oostdijk, een havenplaats uit de Romeinse tijd, Leiden. De Bruin, J. 2019, Border Communities at the Edge of the Roman Empire. Processes of change in the civitas Cananefatium (Amsterdam Archaeological Studies 28), Amsterdam. De Ridder, A.H.P. 1913, Les bronzes du Louvre. Tome premier: les figurines, Paris. Derks, T. 1991, The Perception of the Roman Pantheon by a Native Elite: the example of votive inscriptions from Lower Germany, in Roymans, N./F. Theuws 1991 (eds), Images of the Past. Studies on Ancient Societies in Northwestern Europe (Studies in Prae- and Protohistory 7), Amsterdam, 235-265. Derks, T. 1995, The Ritual of the Vow in Gallo-Roman Religion, in Metzler et al. 1995, 111-127. Derks, T. 1998, Gods, Temples and Ritual Practices. The transformation of religious ideas and values in Roman Gaul, Amsterdam. Derks, T. 2003, Twee Romeinse grafstèles uit Houten-Molenzoom en hun betekenis voor de romanisering van de grensstreek, Jaarboek Oud-Utrecht, 5-32. Derks, T. 2004, Beelden en zelfbeelden van Bataven: de epigrafische bronnen, in De Bataven, een verdwenen volk, Amsterdam/Nijmegen, 40-70. Derks, T. 2005, Romeinse religie en de heiligdommen van Elst, in Van Enckevort/Thijssen 2005b, 20-31. Derks, T. 2008a, Samenvatting en conclusies, in Derks/Van Kerckhove/Hoff 2008, 135-148. Derks, T. 2008b, Inscripties op brons, in Van der Feijst, L./J. de Bruin/E. Blom (eds), De nederzetting te Naaldwijk II. Terug naar de sporen van Holwerda (ADC-monografie 4), Amersfoort, 149-161. Derks, T. 2014a, Seeking Divine Protection Against Untimely Death: Infant votives from Roman Gaul and Germany, in Carroll, M./E.J. Graham 2014 (eds), Infant Death and Health in Roman Italy and Beyond (JRA Supp. Series 96), Portsmouth, 47-68. Derks, T. 2014b, Rome en de archeologie van heiligdommen in de Lage Landen, in Hupperetz et al. 2014, 106-110. Derks, T. 2014c, Die Weihealtäre aus den Nehalennia-Heiligtümern und verwandten ländlichen Tempelbezirken, in Busch/Schäfer 2014, 199-219. Derks, T. 2015a, De sporen, in Derks/De Fraiture 2015, 37-46. Derks, T. 2015b, Inscripties en graffiti, in Derks/De Fraiture 2015, 149-154. Derks, T. et al. 2008, Sporen en structuren, in Derks/Van Kerckhove/Hoff 2008 (eds), 29-59. Derks, T./B. de Fraiture 2015 (eds), Een Romeins heiligdom op een vroeg-middeleeuws grafveld bij Buchten (L.). Verslag van een archeologisch noodonderzoek (1976), Amersfoort. Derks, T./B. de Fraiture 2015b, Inleiding, in Derks/De Fraiture 2015 (eds), 9-13. Derks, T./.P. Hoff, Kaders, in Derks/Van Kerckhove/Hoff 2008 (eds), 3-24. Derks, T./J. van Kerckhove/P. Hoff 2008 (eds), Nieuw archeologisch onderzoek rond de Grote Kerk van Elst, gemeente Overbetuwe (2002-2003) (ZAR 31), Amsterdam. Derks, T./N. Roymans 2009 (eds), Ethnic Constructs in Antiquity. The Role of Power and Tradition, Amsterdam. Deyts, S. 1992, Images des dieux de la Gaule, Paris. Deyts, S. 2004, La femme et l’enfant au maillot en Gaule. Iconographie et épigraphie, in Dasen, V. 2004 (ed.), Naissance et petite enfance dans l’Antiquité, Freiburg/Göttingen, 227-237. De Weerd, M.D./J.K. Haalebos 1973, Schepen voor het opscheppen. Het scheepsarcheologisch onderzoek te Zwammerdam, Spiegel Historiael 8, 386-397. De Winter, J. de 2010, Archeologisch onderzoek op het plangebied Schrames te Helden. Bewoningssporen van het neolithicum tot de late middeleeuwen (BAAC-rapport A-07.0204), ’s-Hertogenbosch.

198

Diederik, F. 2002, Schervengericht. Een onderzoek naar inheems aardewerk uit de late derde en de vierde eeuw in de Kop van Noord-Holland (AWN-reeks 9), Steenwijk. Dierichs, A. 2010, ‘Tintinnabulum’, Brill’s New Pauly online. Dijk, N. 1987, De menselijke skeletresten en beenderen uit Velsen 1. Een voorlopig rapport, Amsterdam, University of Amsterdam, internal paper. Dijkstra, J., Inleiding, doelstelling en methode, in Van Enckevort, H./K. van der Graaf 1997 (eds), Aanvullend Archeologisch Onderzoek Kerkrade-Winckelen, verkaveling Maar-West (ROB Rapportages Archeologische Monumentenzorg 51), Amersfoort, 9-18. Dijkstra, M.F.P./P.F.A.van Grinsven 2007, De vroeg-middeleeuwse nederzetting in Koudekerk aan den Rijn; een bijna vergeten opgraving in de Lagewaardse Polder (Renus reeks 1), Leiden. Doppelfeld, O. 1956, Römische Groβbauten unter dem Kölner Rathaus: Vorbericht über die Rathausgrabung des Jahres 1953, Germania 34, 83-96. Driessen, M. 2007, Bouwen om te blijven. De topografie, bewoningscontinuïtiet en monumentaliteit van Romeins Nijmegen, Amersfoort. Driessen, M. 2014a, Sporen en structuren uit de Romeinse tijd, in Driessen/Besselsen 2014 (eds), 97-132. Driessen, M. 2014b, Deposities in (water)putten, in Driessen/Besselsen 2014 (eds), 149-198. Driessen, M. 2014c, Terracotta beeldjes en gezichtsmaskers, in Driessen/Besselsen 2014 (eds), 489-500. Driessen, M./E. Besselsen 2014 (eds), Voorburg-Arentsburg, een Romeinse havenstad tussen Rijn en Maas, vol. I and II (Themata 7), Amsterdam. Drinkwater, J.F. 1971, A new inscription and the legionary issues of Gallienus and Victorinus, Numismatic Chronicle 11, 325-326. Drinkwater, J./H.Vertet 1992: ‘Opportunity’ or ‘Opposition’ in Roman Gaul?, in Wood, M./F. Queiroga 1992 (eds), Current Research on the Romanization of the Western Provinces (BAR Int. Series 575), Oxford, 25-28. Drouen, L. 2016, Vrouwelijke gladiatoren, in Koster/Mols 2016 (eds), 79-82. Duceppe-Lamarre, A. 2002, Unité ou pluralité de la sculpture celtique hallstattienne et laténienne en pierre en Europe continentale du VIIe au 1er s. av. J.-C., Documents d’archéologie méridionale (en ligne 25), 285-318. URL: https://journals.openedition.org/dam/395 (8-6-2015). Durham, E. 2012, Depicting the Gods: Metal Figurines in Roman Britain. http://intarch.ac.uk/journal/issue31/ durham_index.html (16-10-2016). Eckardt, H. 1999, The Colchester ‘child’s grave’, Britannia 30, 57-90. Edmonds III, R.G. 2019, Drawing Down the Moon. Magic in the ancient Greco-Roman world, Princeton/Woodstock. Ehmig, U. 2013, Pro & contra. Erfüllte und unerfüllte Gelübde in lateinischen Inschriften, file:///C:/Users/ Christel.Veen/Downloads/2013_Pro%20&%20contra%20(1).pdf (2-2-2018). Eimermann, E. 2009 (ed.), Cananefaatse boeren op de noordelijke oeverwal van de Gantel. Een archeologische opgraving aan de Juliahof te Wateringen, gemeente Westland (ADC-rapport 822), Amersfoort. Elzinga, G. 1984, Archeologische Afdeling, VF 64, 126-135. Erdkamp, P. 2007 (ed.), A Companion to the Roman Army (Blackwell Companions to the Ancient World), Malden/Oxford/Victoria. Erdrich, M. 2001, Rom und die Barbaren (Römisch-Germanische Forschungen 58), Mainz am Rhein. Erdrich, M 2005, De reactie, in Willems et al. 2005 (eds), 47-49. Erdrich, M./H.M. Van der Velde 2007, Metalen voorwerpen uit Heeten, in Velde, H.M. van der (ed.), Germanen, Franken en Saksen in Salland. Archeologisch en landschappelijk onderzoek naar de geschiedenis van het landschap en nederzettingsresten uit de Romeinse tijd en Vroege Middeleeuwen in Centraal Salland (ADC-monografie 1), 233-244. Fadden, K.J. 2010, Report on Romano British artefacts found on the surface of Ruxox Farm in the County of Bedfordshire, Ampthill and District Archaeological and Local History Society, Ampthill. http://www.mooncarrot.org.uk/adalhs/downloads/Scatter%20Finds%20from%20Ruxox%20W.pdf (5-3-2018).

199

Faider-Feytmans, G. 1948, La mater de Bavai, Gallia 6, 1948, 392-393. Faider-Feytmans, G. 1979, Les Bronzes romaines de Belgique, Mainz. Faraone, C.A. 1991, Binding and Burying the Forces of Evil: The Defensive Use of ‘Voodoo Dolls’ in Ancient Greece, Classical Antiquity 10-2, 165-220. Fehling, D. 1974, Ethologische Überlegungen auf dem Gebiet der Altertumskunde. Phallische Demonstration, Fernsicht, Steinigung (Monographien zur klassischen Altertumswissenschaft 61), Munich. Fittock, M.G. 2015, Broken Deities: the Pipe-Clay Figurines from Roman London, Britannia 46, 111-134. Flamman, J.P./T.A. Goossens 2006, Schipluiden, ‘Harnaschpolder’. De inrichting en bewoning van het landschap in de Romeinse tijd (125 - 270 na Chr.) (ADC-rapport 625), Amersfoort. Fleischer, R. 1967, Die römischen Bronzen aus Österreich, Mainz. Fleischer, R. 1984, Aphrodite, LIMC II, 57-154. Fontijn, D.R. 2002, Sacrificial Landscapes. Cultural biographies of persons, objects and ‘natural’ places in the Bronze Age of the southern Netherlands, C. 2300-600 BC (APL 33/34), Leiden. Foucher, L. 2000, Concours de boisson entre Dionysus et Héracle, Syria Archéologie. Art et histoire 77, 201-210. Fremersdorf, F. 1957, Das Römergrab in Weiden bei Köln, Cologne. Fulford, M. 2001, Links with the Past: Pervasive ‘Ritual’ Behaviour in Roman Britain, Britannia 32, 199-218. Fulford, M./J. Timby 2001, Timing Devices, Fermentation Vessels, ‘Ritual’ Piercings? A consideration of deliberately ‘holed’ pots from Silchester and elsewhere, Britannia 32, 293-297. Galestin, M.C. 1990, Een soldaat uit Ezinge in Romeinse dienst, Groningse volksalmanak. Historisch jaarboek voor Groningen, 146-152. Galestin, M.C. 1994, A New Mars from the Netherlands, in Akten der 10. Internationalen Tagung über antike Bronzen. Freiburg 18.-22. Juli 1988 (Forschungen und Bericht zur Vor- und Frühgeschichte in Baden-Württemberg 45), Stuttgart, 155-159. Galestin, M.C. 1995, Egyptisch godinnenpaar op Groninger wierde, Paleo Aktueel 6, 91-98. Galestin, M.C. 1999/2000, Winsum-Bruggeburen, First Report on the Excavation. An Early Roman outpost among the Frisians? Part one: The Roman coins, Palaeohistoria 41/42, 225-235. Galestin, M.C. 2001/2002, Winsum-Bruggeburen, Second Report on the Excavation: The Roman pottery, Palaeohistoria 43/44, 435-467. Galestin, M.C. 2000, Ensembles of Roman Figural Bronzes from the Netherlands, KJb 33,143-150. Galestin, M.C. 2007, Romeinse goden in Friesland, Paleo Aktueel 18, 74-79. Galestin, M.C. 2009, Het Romeinse schrijfplankje van Tolsum eindelijk ontcijferd, Paleo Aktueel 20, 69-76. Garfinkel, Y. 1994, Ritual Burial of Cultic Objects: the Earliest Evidence, Cambridge Archaeological Journal 4 (2), 159-188. Garland, R. 1995, The Eye of the Beholder: Deformity and Disability in the Graeco-Roman World, Ithaca. Gassner, V. 2010, Eine Zauberpuppe aus dem Heiligtum des Iuppiter Heliopolitanus in Carnuntum, in Anodos, Studies of the Ancient World, In honour of Werner Jobst 8, 2008 (Trnava 2010), 221-230. Gawronsky, J/P.Kranendonk 2018, Spul.Catalogus archeologische vondsten Noord/Zuidlijn Amsterdam, Amsterdam. Gebers, W. 1995. Fünfzehn Jahre Grabung Rullstorf - eine Bilanz. Berichte zur Denkmalpflege in Niedersachsen 14, 56–60. Geerts, R.CA./H.A.P. Veerman 2012 (eds), Romeinse bewoning tussen IJzertijdgraven. Een archeologische opgraving te Groesbeek-Hüsenhoff (ADC-rapport 2687), Amersfoort. Gerrets, D.A. 2010, Op de grens van land en water. Dynamiek van landschap en samenleving in Frisia gedurende de Romeinse tijd en de volksverhuizingstijd, Groningen. Gerritsen, F. 2003, Local Identities. Landscape and community in the late prehistoric Meuse-Demer-Scheldt region (AS 9), Amsterdam. Gicheva, R.1997, Sabazius, LIMC VIII 1, 1068-1071.

200

Glasbergen, W. 1972, De Romeinse castella te Valkenburg Z.H. De opgravingen in de dorpsheuvel in 1962 (Cingula 1), Groningen. Glasbergen, W./W. Groenman-van Waateringe 1974, The Pre-Flavian Garrisons of Valkenburg Z.H.: Fabriculae and bipartite barracks (Cingula 2), Amsterdam. Goldbeck, V. 2015, Fora augusta. Das Augustusforum und Seine Rezeption in Westen des Imperium Romanum (Eikoniká Kunstwissenschaftliche Beiträge 5), Düsseldorf. Goldsworthy, A./I. Haynes 1999 (eds), The Roman Army as a Community (JRA Supp. Series 34), Ann Arbor. Goossens, T.A. 2012 (ed.), Van akkerland tot Hooghwerf. Onderzoek naar de bewoning in de ijzertijd, inheemsRomeinse tijd, de middeleeuwen en de nieuwe tijd op de haakwal van Naaldwijk (plangebied Hoogeland, gemeente Westland) (Archol-rapport 167), Leiden. Graf, F. 1996, Gottesnähe und Schadenzauber. Die Magie in der griechisch-römischen Antike, Munich. Graves, R. 1984, The Greek myths 1-2, Middlesex/New York. Green, M.J. 1976, The Religions of Civilian Roman Britain (BAR 24), Oxford. Green, M.J. 1977, ‘Theomorphism’ and the Role of Divine Animals in Romano-British Cult Art, in Munby, J./M. Henig 1977 (eds), Roman Life and Art in Britain (BAR 41), Oxford. Green, M.J. 1978, Small Cult Objects from the Military Areas of Roman Britain, (BAR 52), Oxford. Green, M.J. 1992a, Symbol & Image in Celtic Religious Art , London/New York2. Green, M.J. 1992b, Animals in Celtic Life and Myth, London. Green, M.J. 1992c, Dictionary of Celtic Myth and Legend, London. Greene, E.M. 2012, Sulpicia Lepidina and Elizabeth Custer: A Cross-cultural Analogy for the Social Role of Women on a Military Frontier, in Duggan, M./F. McIntosh/D.J. Rohl 2011 (eds) TRAC 2011. Proceedings of the Twenty First Annual Theoretical Roman Archaeology Conference, Oxford, 105-114. https://www.academia.edu/11712262/2012._Sulpicia_Lepidina_and_Elizabeth_Custer_A_Cross-cultural_Analogy_for_the_Social_Roles_of_Women_on_a_Military_Frontier_in_M._Duggan_F._McIntosh_and_D._Rohl_eds._Theoretical_Roman_Archaeology_Conference_2011_Oxbow._105-14 (12-102016). Grimm, G. 1969, Die Zeugnisse Ägyptischer Religion und Kunstelemente im Römischen Deutschland, Leiden. Groenewoudt, B.J./R.M. van Heeringen/G.H. Scheepstra 2006 (eds), Het zandeilandenrijk van Overijssel (Nederlandse Archeologische Rapporten 22), Amersfoort. Groot, M. 2007, De rol van dieren in een Bataafse plattelandsgemeenschap: dierlijk bot uit Passewaaij, in Roymans/Derks/Heeren 2007 (eds), 177-192. Groot, M. 2008, Animals in Ritual and Economy in a Roman Frontier Community. Excavations in Tiel-Passewaaij (AS 12), Amsterdam. Groot, M. 2009, Searching for Patterns among Special Animal Deposits in the Dutch River Area during the Roman Period, JALC, 1-18. http://dpc.uba.uva.nl/jalc/01/nr02/a03 (8-2-2014). Groot, T. de 2006, Roman Villae in the Loess Area of the Dutch Province of Limburg: an Analysis of their Number, Distribution and Preservation, BROB 46, 275-301. Gugl, C./ J. Trumm 2015, The Secret of Squares - Interpreting Large-Scale Square Structures at Roman Legionary Forts, in Vagalinski, L./N. Sharankov 2015 (eds), Limes XXII. Proceedings of the 22nd International Congress of Roman Frontier Studies, Ruse, Bulgaria, September 2012 (Bulletin of the National Institute of Archaeology XLII), Sofia, 103-110. Haalebos, J.K. 1977, Zwammerdam-Nigrum Pullum, ein Auxiliarkastell am Niedergermanischen Limes (Cingula 3), Amsterdam. Haalebos, J.K. 2000, Mosterd na de maaltijd. Een vergeten jubileum: Traianus en het jaar 98 na Chr. in Nijmegen, Numaga 47, 9-41. Haalebos, J.K. et al. 1995, Castra en Canabae. Opgravingen op het terrein van het voormalige Canisiuscollege te Nijmegen,1987-1994, Nijmegen.

201

Haalebos, J.K. et al. 1998, Centuriae onder centuriae hof. Opgravingen achter het hoofdgebouw van het voormalige Canisius College te Nijmegen,1995-1997, Nijmegen. Haalebos, J.K. et al. 2000, Alphen aan den Rijn-Albaniana, 1998-1999. Opgravingen in de Julianastraat, de Castellumstraat, op Het Eiland en onder het St. Jorisplein, Nijmegen. Haalebos, J.K. et al. 2002, De Romeinen in Cuijk, in Enckevort/Thijssen 2002 (eds), 21-47. Haalebos, J.K./M. Polak 2007, Een lijst met Romeinse namen uit Bodegraven, Westerheem-3, 114-122. Haalebos, J.K./W.J.H. Willems 2005a, De grote legerplaats op de Hunerberg, in Willems et al. 2005 (eds), 56-62. Haalebos, J.K./W.J.H. Willems 2005b, De legioensvestingen op de Hunerberg, in Willems et al. 2005 (eds), 73-86. Haarhuis, H.F.A. 1991, Romeinse figurale bronzen van het Lage Veld, unpublished thesis, Nijmegen. Habermehl, D. 2011, Settling in a Changing World, Villa Development in the Northern Provinces of the Roman Empire, Amsterdam. Hackens, T. 1963, Favisae, in Université de Louvain, ed. 1963, Études Étrusco-Italiques. Mélanges pour le 25e anniversaire de la chaire d’Étruscologie à l’Université de Louvain. Louvain-la-Neuve, 71–99. Haeussler, R. 2013, Becoming Roman? Diverging identities and experiences in ancient northwest Italy, Walnut Creek, CA. Halbertsma, R.B. 2014, Een held met vele gezichten: Heracles in de antieke kunst, Lampas 47-4, 299-310. Hallewas, D.P/ R.M. van Dierendonck 1993, The Valkenburg-Marktveld and Valkenburg-the Woerd excavations 1985-1986: a Preliminary Report, in Van Dierendonck/Hallewas/Waugh 1993 (eds), 11-46. Hamilton R. 2010, Herms, in The Oxford Encyclopedia of Ancient Greece and Rome, 414-415. Hanel, N. 1995, Vetera I. Die Funde aus den römischen Lagern auf dem Fürstenberg bei Xanten. Cologne. Hanel, N./A.M. Cerdán 2012, Kunstreiter (cursores, desultores) in der römischen Kleinplastik. Zur Identifizierung eines Statuettentyps, RM 118, 339-352. Harsema, O.H. 1976, Noordbarge, Bulletin KNOB 75, 52-55. Harsema, O.H. 1987, Change and Continuity in Rural Settlements in Drenthe from the Neolithic onwards. A reconsideration of traditional and current opinions, Palaeohistoria 29, 103-119. Haynes, I. 1993, The Romanisation of Religion in the Auxilia of the Roman Imperial Army from Augustus to Septimius Severus, Britannia 24, 141-157. Haynes, I. 1997, Religion in the Roman Army: Unifying Aspects and Regional Trends, in Rüpke J./H. Cancik 1997 (eds), Römische Reichsreligion und Provinzialreligion, Tübingen, 113-126. Haynes, I. 1999, Military Service and Cultural Identity in the Auxilia, in Goldsworthy A./I Haynes 1999 (eds), The Roman Army as a Community (JRA Supp. Series 34), Ann Arbor, 165-174. Haynes, I. 2013, Blood of the Provinces. The Roman auxilia and the making of provincial society from Augustus to the Severans, Oxford. Hazenberg, T. 2000, Leiden-Roomburg 1995-1997: archeologisch onderzoek naar het kanaal van Corbulo en de vicus van het castellum Matilo (RAM 77), Amersfoort. Heeren, S. 2009, Romanisering van rurale gemeenschappen in de Civitas Batavorum. De casus Tiel-Passewaaij (NAR 36), Amersfoort. Heeren, S. et al. 2008, Sporen en structuren, in Derks/Van Kerckhove/Hoff 2008 (eds), 29-59. Heirbaut E.N.A/ H. van Enckevort 2011 (eds), De verdwenen villa van de Tienakker. Archeologisch onderzoek naar het Romeinse verleden van Tienakker (Archeologische Berichten Wijchen, Rapport 4), Nijmegen. Heirbaut, E.N.A. 2013 (ed.), De Zuidwestelijke hoek van Ulpia Noviomagus in kaart gebracht (Archeologische Berichten Nijmegen 41 en 42), Nijmegen. Hendriks, J./A. den Braven/H. van Enckevort/J. Thijssen 2014, Een noordelijk steunpunt. Vroegmiddeleeuws Nijmegen vanuit archeologisch perspectief, in Peterse/Verhoeven et al. 2014 , 43-69. Hemingway, S./R. Stone, 2017, The New York Sleeping Eros: A Hellenistic Statue and Its Ancient Restoration, 46-63. https://doi.org/10.4000/techne.1266.

202

Henig, M./G. Webster/R. Wilkins 1987, A Bronze Dioscurus from Wroxeter and its Fellow from Canterbury, The Antiquaries Journal 67, 360-362. Henig, M. 1995, The Art of Roman Britain, London. Hermsen, I. 2003, De Victoria van Colmschate-Skibaan: een Romeins godenbeeldje als bouwoffer, in Overijssels Erfgoed, Zwolle, 65-71. Hermsen, I. 2007, Een afdaling in het verleden: archeologisch onderzoek van bewoningsresten uit de prehistorie en de Romeinse tijd op het terrein Colmschate-Skibaan (gemeente Deventer) (Rapportage Archeologie Deventer 19), Deventer, 42-268. Heyob, S.K. 1975, The Cult of Isis among Women in the Graeco-Roman World (EPRO 51), Leiden. Hiddink, H.A. 1999, Germaanse samenlevingen tussen Rijn en Weser. 1e eeuw – tot 4de eeuw na Christus, Amsterdam. Hiddink, H.A. 2003, Het grafritueel in de Late IJzertijd en Romeinse tijd in het Maas-Demer-Scheldegebied, in het bijzonder van twee grafvelden bij Weert (ZAR 11), Amsterdam. Hiddink, H.A. 2010, Opgravingen op Kampershoek noord bij Weert (ZAR 39), Amsterdam. Hiddink, H.A. 2015, The Villa Settlement of Hoogeloon-Kerkakkers, in Roymans et al. 2015 (eds), 87-124. Hingh, A. de/W. Vos 2005, Romeinen in Valkenburg (ZH). De opgravingsgeschiedenis en het archeologische onderzoek van Praetorium Agrippinae, Leiden. Hingley, R. 2005, Globalizing Roman Culture. Unity, diversity and empire, London/New York. Hobbs, R. 1997, Late Roman Precious Metal Deposits, c. AD 200-700: Changes over Time and Space, PhD thesis UCL, London. Hoekstra, T.J./S.L. Wynia 1997, ‘Dan zullen velen juichen’. Geschiedenis van twintig jaar conflicten over het Domplein te Utrecht (1929-1949), Jaarboek Oud-Utrecht, 125-168. Hondius-Crone, A. 1955, The Temple of Nehalennia at Domburg, Amsterdam. Hopkins, J. 2012 The sacred sewer: tradition and religion in the Cloaca Maxima, in Bradley, M. 2012 (ed.), Rome, Pollution and Propriety: Dirt, Disease and Hygiene in the Eternal City from Antiquity to Modernity, Cambridge/New York, 81-102. Horn, H.G. 1970, Eine Weihung für Hercules Magusanus aus Bonn, BJb 170, 233-251. Horn, H.G. 1987, Bildendenkmäler des Matronenkultes im Ubiergebiet, in Bauchhenß/Neumann 1987 (eds), 31-54. Hoss, S. 2015, Frontier Finds – Military Fashions, in Breeze, D./R.H. Jones/I.A. Oltean 2015 (eds), Understanding Roman Frontiers. Papers offered to Professor Bill Hanson on the occasion of his retirement, 135-153. Hoss, S. et al. 2015, Bronzen beeldje van een haan met voetstuk, in Derks/De Fraiture 2015 (eds), 159-171. Hübner, E. 1876, Römische Alterthümer aus den Oldenburgischen, BJb 57, 66-69. Huiskes, B. 2011, Eeuwige rust op de Donderberg. Een groot vroegmiddeleeuws grafveld bij Rhenen, Leiden. Hulst, R.S. 1978, Druten-Klepperhei, Vorbericht der Ausgrabungen einer römischen Villa, BROB 28, 133-151. Hunink, V. 2016, Helden in de arena. Martialis over gladiatoren, in Koster/Mols 2016 (eds), 43-52. Hupperetz, W. 1991, Het Midden-Limburgse Maasdal in de Romeinse tijd. Platteland tussen Atuatuca en Colonia Ulpia Traiana, unpublished thesis, Nijmegen. Hupperetz, W./O.E. Kuiper/F.G. Naerebout/M.J. Versluys 2014 (eds), Van Rome naar Romeins, Zwolle. Isings, C. 1984, Geldtransport voor Mercurius, Westerheem 33-3, 99-102. Janssen, L.J.F. 1846, Ontdekking van Romeinsche hypocaustum’s, outaren met opschriften enz., te Holledoorn, onder Groesbeek, Oudheidkundige Medeelingen 1-4, 323-346. Janssen, L.J.F. 1849, Drentsche Oudheden, Utrecht. Jansen R./L.P. Louwe Kooymans 2007 (eds), Van contract tot wetenschap. Tien jaar archeologisch onderzoek door Archol BV, 1997-2007, Leiden. Jarrett, R. 2008, Reappraising Penn and Harker: A reassessment of the finds from excavations at Roman Springhead, published between 1957 and 1984, and interpretations made about their use in past activities, master’s thesis, Durham, http://etheses.dur.ac.uk/2168/(29-11-2016).

203

Jeneson, C.F. 2013, Exploring the Roman Villa World between Tongeren and Cologne. A landscape archaeological approach, PhD thesis, Amsterdam. Jeneson, C.F./W.K.Vos 2020 (eds), Roman Bathing in Coriovallum. The thermae of Heerlen revisited (NAR 65), Amersfoort. Jenkins, F. 1977, Clay Statuettes of the Roman Western Provinces, unpublished PhD thesis, Canterbury. Johns, C. 1982, Sex or symbol. Erotic images of Greece and Rome, Austin. Jücker, H. 1961, Das Bildnis im Blätterkelch. Geschichte und Bedeutung einer römischen Porträtform, Lausanne/ Freiburg. Kalee, C.A./C. Isings 1980, Beknopt verslag van een opgraving in De Meern, Jaarboek Oud Utrecht, 5-25. Kaufmann-Heinimann, A. 1977, Die Römischen Bronzen aus der Schweiz I. Augst, Mainz. Kaufmann-Heinimann, A. 1994, Die Römischen Bronzen aus der Schweiz V. Neufunde und Nachträge, Mainz am Rhein. Kaufmann-Heinimann, A. 1998, Götter und Lararien aus Augusta Raurica. Herstellung, Zusammenhänge und sakrale Funktion figürlicher Bronzen in einer römischen Stadt, Augst. Kaufmann-Heinimann, A. 2012, Ikonographie und Stil. Zu Tracht und Ausstattung einheimischer Gottheiten in den Nordwestprovinzen, in Bronzes grecs et romains, recherches récentes - Hommage à Claude Rolley, Paris, 2-28 http://inha.revues.org/3964 (8-3-2014). Kemmers, F. 2006, Coins for a Legion. An analysis of the coin finds from the Augustan legionary fortress and Flavian canabae legionis at Nijmegen (Studien zu Fundmünzen der Antike 21), Mainz am Rhein. Kemmers F./M. Polak 2011, Ulpia Noviomagus Batavorum. De Romeinse stad aan de Waal, in De Haan, N. /S.T.A.M. Mols 2011 (eds), Cultuurgeschiedenis van de Oudheid, Zwolle, 249-257. Kessener, P./M.P.J. Janssens 2017, Onderzoek ten behoeve van de UNESCO nominatie Romeins aquaduct Gemeente Berg en Dal en gemeente Nijmegen. Een archeologisch wetenschappelijke bureaustudie (RAAP-rapport 3319), Weesp. Kleijen, F. 2016, Cerberus, de hellehond, Jaarverslag AWN afdeling Nijmegen en omstreken 39, 16-17. Kleiterp, M. 1991, Alphen-De Schans, Holland, regionaal-historisch tijdschrift 6, 342-343. Knol, E./A.C. Bardet/W. Prummel 2005 (eds), Professor Van Giffen en het geheim van de wierden, Groningen, exhibition catalogue. Knol, E. 2014, Ezinge: metaal uit een opgraving zonder detector, in Nieuwhof, A. (ed.), En dan in hun geheel. De vondsten uit de opgravingen in de wierde Ezinge, Groningen, 187-199. Koeling, J. 2002, Cuijk in de prehistorie, in Van Enckevort/Thijssen 2002 (eds), 15-19. Kolbe, H.G. 1961, Neue Inschrifte aus Bonn, BJb 167, 85-107. Kopytoff, I. 1986, The Cultural Biography of Things. Commoditization as process, in Appadurai, A., The Social Life of Things. Commodities in cultural perspective, Cambridge, 64-91. Koster, A./H. Tomas/W.J.H. Verwers 2001, Didam-Kollenburg in de laat-Romeinse tijd, Lichtevoorde. Koster, A. 2005, Het grafveld van Ulpia-Noviomagus in Nijmegen-West, in Willems et al.2005 (eds), 205-210. Koster, A. 2013, The Cemetery of Noviomagus and the Wealthy Burials of the Municipal Elite (Museum Het Valkhof XIV), Nijmegen. Koster, A. 2016, De gladiatoren, in Koster/Mols 2016 (eds), 53-66. Koster A./S.T.A.M. Mols 2016 (eds), Gladiatoren, helden van het Colosseum, Nijmegen. Koster, A./K. Peterse/L.J.F. Swinkels 2002, Romeins Nijmegen boven het maaiveld. Reconstructies van de verdwenen architectuur, Nijmegen. Krzyszowska, A. 2002, Les cultes privés à Pompéi, Wroclaw. Kunow, J. 1992, Zentralität und Urbanität in der Germania Inferior des 2. Jahrhunderts n. Chr., in Schalles, H.J. (ed.), Die römische Stadt im 2. Jahrhundert n. Chr. (Xantener Berichte 2), Cologne, 143-152. Kyll, N. 1966, Heidnische Weihe- und Votivgaben aus der Römerzeit des Trierer Landes, Trier.

204

Laken, L. 2010, Kandelabers in Oppidum Batavorum, in Enckevort/Heirbout 2010 (eds), 129-158. Lange, S. 2017, Uit het juiste hout gesneden. Houten gebruiksvoorwerpen uit archaeologische context tot 1300 n.Chr. (NAR 54), Amersfoort. Lange, S. 2021, The Wooden Artefacts from the Early Roman Fort Velsen 1 (NAR 69), Amersfoort. Langeveld, M./A. Luksen-IJtsma/P. Weterings 2010, Een goede buur? LR46 en LR49: definitief archeologisch onderzoek naar de vicus, grafvelden, infrastructuur en een inheemse nederzetting in de omgeving van het Romeinse castellum in De Meern, deelgebied ‘De Woerd’ (Gemeente Utrecht) (Basisrapportage Archeologie 19), Utrecht. Lanting, J.N./ J. van der Plicht 2009/2010, De 14C-chronologie van de Nederlandse pre- en protohistorie VI: Romeinse tijd en Merovingische periode, deel A: historische bronnen en chronologische schema’s, Palaeohistoria 51/52, 27-168. Laurence, R./S.E. Cleary/G. Sears 2011, The City in the Roman West, c. 250 BC – c. AD 250, Cambridge/New York. Lauwerier, R.C.C.M./B.J. Groenewoudt/O. Brinkkemper/F.J. Laarman 1998-99, Between Ritual and Economics: Animal and Plants in a Fourth-century Native Settlement at Heeten, the Netherlands, BROB 43, 155-198. Leclant J./G. Clerc 1981, Ammon, LIMC I, 666-689. Lehner, H. 1905, Das Provinzialmuseum in Bonn I. Die römischen Skulpturen, Bonn. Lersch, L. 1846, Juppiter Ammon. Sandstein Herme von Lechenich bei Bonn, BJb 9, 116-121. Lewis, G. 1980, Day of Shining Red: an Essay on Understanding Ritual, Cambridge. Liéger, A. 1997, La nécropole de Cutry (Meurthe-et-Moselle), Nancy. Losansky, G. 2015, Die Obergeschossareale der Stadthäuser in Herculaneum, Wiesbaden. Magendans, J.R./J.A. Waasdorp 1986, Archeologie in Den Haag in 1985, Jaarboek Die Haghe, 297-300. Magendans, J.R./P.J.W.M. Schulten/J.A. Waasdorp 1988, Archeologie in Den Haag in 1987, Jaarboek Die Haghe, 215-218. Manfrini-Aragno, I. 1987, Bacchus dans les bronzes hellénistiques et romains. Les artisans et leur repertoire, Lausanne. Mattingly, D. 2006, An Imperial Possession. Britain in the Roman Empire, London/New York. Mattusch, C.C. 1996a, Classical Bronzes. The Art and Craft of Greek and Roman Statuary, Ithaca/London. Mattusch, C.C. 1996b, The Fire of Hephaistos. Large Classical Bronzes from North American Collections, Cambridge. Matz, F. 1932, Die Lauersforter Phalerae, Berlin/Leipzig. Meffert, M./G. Scheijvens 2010 (eds), Jaarverslag Meldpunt Archeologische Bodemvondsten van de provincie Noord-Brabant 2007-2008, ’s-Hertogenbosch. Meinecke, K. 2016, Antike Dornausziehergruppen, BABesch 91, 129-160. Menzel, H. 1960, Die Römischen Bronzen aus Deutschland I, Speyer, Mainz. Menzel, H. 1966, Die Römischen Bronzen aus Deutschland II, Trier, Mainz. Menzel, H. 1986, Die Römischen Bronzen aus Deutschland III, Bonn, Mainz. Merrifield, R. 1987, The Archaeology of Ritual and Magic, London. Metzler, J./M. Millet/N. Roymans/J. Slofstra (eds) 1995, Integration in the Early Roman West. The role of culture and ideology (Dossiers d’archéologie du Musée National d’Histoire et d’Art 3), Luxembourg. Miedema, M. 1999/2000, West-Fivelingo. 600 v. Chr. -1900 n. Chr.: Archeologische kartering en beschrijving van 2500 jaar bewoning in Midden-Groningen, Palaeohistoria 41/42, 235-445. Milburn, O. 2007, Marked out for greatness? Perceptions of deformity and physical impairment in ancient China, (Monumenta Serica 55), 1-22. URL: https://www.jstor.org/stable/40727673 (20-5-2021). Millet, M., 1990a, The Romanization of Britain. An essay in archaeological interpretation, Cambridge. Millett, M. 1990b, Romanization. Historical issues and archaeological interpretation, in Blagg, T.F.C./M. Millett (eds), The Early Roman Empire in the West, Oxford, 35-41. Modderman, P.J.R. 1973, A Native Farmstead from the Roman Period near Kethel, Municipality of Schiedam, Province of South Holland, BROB 23, 149-158.

205

Moed, F. 2017, Exit Albiobola. Een nieuw epigrafusch onderzoek naar de naam Utrecht, ARCH-Magazine 1-2017, 42-45. Mols, S.T.A.M. 1994, Furniture Attachments Shaped Like Human Feet, in Akten der 10. Internationalen Tagung über antike Bronzen, Freiburg. 18.-22.Juli 1988, Stuttgart, 293-296. Mols, S.T.A.M. 1999, Wooden Furniture in Herculaneum. Form, technique and context, Amsterdam. Mols, S.T.A.M./A.M. Gerhartl-Witteveen/H. Kars et al. (eds) 1995, Acta of the 12th International Congress on Ancient Bronzes, Nijmegen. Mols, S.T.A.M/M. Polak 2020 (eds), De Romeinse Limes in Nederland, Lampas 53-2, Hilversum. Morel, J./M.D. de Weerd 1980, Early Roman Harbours in Velsen, in Hanson, W.S./L.J.F. Keppie (eds), Roman Frontier Studies 1979, Papers presented to the 12th International Congress of Roman Frontier Studies (BAR Int. Series 71), Oxford, 475-494. Murgan, A.M. 2016, Bothros, Favissa und Co. - Von rituellen Löchern und ihren Bezeichnungen, in Bolder-Boos, M/D. Maschek 2016 (eds), Orte der Forschung, Orte des Glaubens. Neue Perspective für Heiligtümern in Italien von der Archaik bis zum späten Republik. Publikation der internationalen Tagung vom 19. und 20. Juli 2013, Bonn, 85-98. Müller, S. 2015 (ed.), Gebrochener Glanz. Römische Groβbronzen am UNESCO-Welterbe Limes, Mainz. Naerebout, F.G. 1997, Attractive Performances. Ancient Greek Dance: Three Preliminary Studies, Amsterdam. Naerebout, F.G. 2010, How Do You Want Your Goddess? From the Galjub hoard to a general vision on religious choice in Hellenistic and Roman Egypt, in Bricault, L./M.J.Versluys 2010 (eds), Isis on the Nile. Egyptian Gods in Hellenistic and Roman Egypt. Proceedings of the IVth International Conference of Isis Studies, Liège, November 27-29 2008, Leiden, 55-73. Naerebout, F.G. 2013, Can these dry bones live? Some thoughts on ancient lived religion. Unpublished paper for the conference Lived Religion: Studying Religious Practice, October 24, 2013, Leiden. Neils, J/P. Schultz 2012, Erechtheus and the Apobates Race on the Parthenon Frieze (North XI–XII), AJA 116-2, 195–207. Nellissen, L. 2004, Nijmeegse Oudheden (translation): Johannes Smetius, Antiquitates Neomagenses, Tilburg. Neudecker, R. 1988, Die Skulpturenausstattung römischer Villen in Italien (Beiträge zur Erschlieβung hellenistischer und kaiserzeitlicher Skulptur und Architektur 9), Mainz am Rhein. Neudecker, R. 2011, Herms, Brill’s New Pauly online. Nicolay, J.A.W. 2005, Nieuw licht op de Late IJzertijd: twee vermoedelijke sierknoppen van gordelhaken uit het Friese terpengebied, Paleo Aktueel 19, 130-135. Nicolay, J.A.W. 2007 Armed Batavians. Use and significance of weaponry and horse gear from non-military contexts in the Rhine-delta (50 BC to 450 AD), Amsterdam. Niekus, M.J.L.Th. 2002, Resultaten, in Niekus/Huisman 2002, 15-23. Niekus, M.J.L.Th./M.A. Huisman 2002, Een huisterpje uit de Romeinse tijd in het klei-veengebied. Een archeologische opgraving in het tracé van de Stadsrondweg Oost te Sneek (Fr.) (ARC-publicaties 53), Groningen. Niemeijer, R.A.J. 2009, Bergen op Zoom - Paradeplaats: een bijzonder vondstcomplex met mini-amphoren (Auxiliaria 9), Nijmegen. Niemeijer, R.A.J. 2013, Die westlichen Wehranlagen des augusteischen Lagers auf dem Hunerberg in Nijmegen: Die Ausgrabungen der Radboud Universiteit 1987-1997, KJb 46, 35-74. Niemeijer, R.A.J. 2014, Das Tafelgeschirr aus dem grossen augusteischen Lager auf dem Hunerberg in Nijmegen: Die Ausgrabungen der Radboud Universiteit 1987-1997, KJb 47, 7-40. Nieuwhof, A. 2013, Anglo-Saxon immigration or continuity, Ezinge and the coastal area of the northern Netherland in the Migration Period, JALC 5, 53-83. Nieuwhof, A. 2014 (ed.), En dan in hun geheel. De vondsten uit de opgravingen in de wierde Ezinge, Groningen. Nieuwhof A. 2015, Eight Human Skulls in a Dung Heap and More. Ritual practice in the terp region of the northern Netherlands, 600 BC-AD 300, Groningen.

206

Nieuwhof, A. 2020, Luxury tableware? Terra sigillata in the coastal region of the northern Netherlands, in Rubel A./H-U Voß 2020 (eds), Experiencing the Frontier and the Frontier of Exerience. Barbarian perspectives and Roman strategies to deal with new threats (Archaeopress Roman Archaeology 76), Oxford, 94-110. http://doi.org/10.32028/9781789696813-8 (18-4-2021) Noelke, P./F. Neumann-Steckner/B. Schneider 2003 (eds), Romanisation und Resistenz in Plastik, Architektur und Inschriften der Provinzen des Imperium Romanum. Neue Funde und Forschungen, Mainz am Rhein. Nouwen, R. 2012, Tongeren, een Romeinse stad in het land van de Tungri, Tongeren. Ogden, D. 1999, Binding Spells: Curse Tablets and Voodoo Dolls in the Greek and Roman Worlds, in Flint, V./R. Gordon/G. Luck/D. Ogden 1999 (eds), The Athlone History of Witchcraft and Magic in Europe 2. Ancient Greece and Rome, London, 1-90. Pallottino, M. 1945, La scuola di Vulca (= Saggi di Antichità 3, Roma 1979), Rome. Panhuysen, T.A.S.M. 1984, Maastricht staat op zijn verleden (Stichting Historische Reeks Maastricht), Maastricht. Panhuysen, T.A.S.M. 1996, Romeins Maastricht en zijn beelden (Corpus Signorum Imperii Romani. Corpus van de Romeinse Beeldhouwkunst. Nederland - Germania Inferior - Maastricht), Maastricht/Assen. Pearce, J. 2014, A ‘Civilised’ Death? The Interpretation of Provincial Roman Grave Good Assemblages, in Brandt, J.R./ M. Prusac/ H. Roland 2014 (eds), Death and Changing Rituals. Function and Meaning in Ancient Funerary Practices (Studies in Funerary Archaeology 7), Oxford/Philadelphia, 223-247. Pearcy, L.T. 1973, Tacitus’ Use of Species, Imago, Effigies, and Simulacrum. https://repository.brynmawr.edu/cgi/viewcontent.cgi?referer=&httpsredir=1&article=1116&context=dissertations (9-1-2017). Pelgrom, J. 2009, Het romanisatiedebat: een inleiding, Lampas 42-3, 159-171. Peters, W.J.Th. 1984, De panter van Montfort, in De Waele, J.A. /L.J.F. Swinkels/E.M. Moormann 1984 (eds), Om de tuin geleid. Feestbundel aangeboden aan prof.dr. W.J.Th. Peters ter gelegenheid van zijn vijfenzestigste verjaardag, Nijmegen, 17. Peters, W.J.Th./L.J.F. Swinkels/E.M. Moormann 1978, Die Wandmalereien der römischen Villa von Druten und die Frage der Felderdekoration in den europäischen römischen Provinzen, BROB 28, 153-197. Peterse, K. 2005, Luxury Living in the Praetorium on the Kops Plateau in Nijmegen: Quotations of Mediterranean Principles in Roman Provincial Architecture, BABesch 80, 163-198. Peterse, H./D. Verhoeven et al. 2014 (eds), Het Valkhof, 2000 jaar geschiedenis, Nijmegen. Phang, S.E. 2001, The Marriage of Roman Soldiers (13 B.C. – A.D. 235). Law and Family in the Imperial Army, Leiden/Boston/Cologne. Pitts, L.P. 1979, Roman Bronze Figurines of the Catuvellauni and Trinovantes (BAR 60), Oxford. Pleiner, R. 1993, The Celtic Sword, Oxford. Pleyte W. 1882, Nederlandsche oudheden van de vroegste tijd tot op Karel den Groote: Drenthe, Leiden. Polak, M. 2005, Lokale nijverheid, baksteen, aardewerk, metaal en glas, in Willems et al. 2005 (eds), 243-249. Polak, M. 2014 (ed.), Opgravingen op het terrein van het voormalige Canisiuscollege in Nijmegen 1. 1987-1997 (Auxiliaria 14), Nijmegen. Polak, M./R.P.J. Kloosterman/R.A.J. Niemeijer et al. 2004, Alphen aan den Rijn-Albanaiana, 2001-2002. Opgravingen tussen de Castellumstraat, het Omloopkanaal en de Oude Rijn, Nijmegen. Polak, M./L.I. Kooistra 2013 (2015), A Sustainable Frontier? The Establishment of the Roman Frontier in the Rhine Delta Part 1: From the End of the Iron Age to the Death of Tiberius (c. 50 BC-AD 37), Jahrbuch des Römisch-Germanischen Zentralmuseums 60, 355-458. Prangsma, N.M. et al. 2003, Archeologisch onderzoek bij Zorgcentrum Oldershove te Wehl (ADC-rapport 172), Amersfoort. Prangsma, N.M. et al. 2004, Zorgcentrum Oldershove te Wehl. Archeologisch onderzoek bij Zorgcentrum Oldershove te Wehl, een vervolg ( ADC-rapport 297), Amersfoort.

207

Quast, D. 2009, Velp und verwandte Schatzfunde des frühen 5. Jahrhunderts, Acta Praehistorica et Archaeologica 41, 207-230. Raepsaet-Charlier, M.-Th. 1993, Diis deabusque sacrum. Formulaire votif et datation dans les Trois Gaules et les deux Germanies (Gallia romana 1), Paris. Raepsaet-Charlier, M.-Th. 2015, Cultes et territoire, Mères et Matrones, dieux “celtiques”: quelques aspects de la religion dans les provinces romaines de Gaule et de Germanie à la lumière de travaux récents, L’Antiquité Classique 84, 173-226. Raven, M.J. 2010, Egyptische magie. Op zoek naar het toverboek van Toth, Zutphen. Reigersman-Van Lidth de Jeude, F. 2006, Keramische objecten, in Flamman J.P./T.A. Goossens 2006 (eds), Schipluiden, ‘Harnaschpolder’. De inrichting en bewoning van het landschap in de Romeinse tijd (125 - 270 na Chr.) (ADC-rapport 625), Amersfoort, 145-146. Reinach, S. 1894, Bronzes figurés de la Gaule Romaine, Paris. Riederer, J. 1987, Archäologie und Chemie. Einblicke in die Vergangenheit. Berlin. Riederer, J. 2000, Der Beitrag zur Metallanalyse zur Bestimmung römischer Statuettenwerkstätten, KJb 33, 575-583. Ristow, G. 1970, Römische Götterdenkmäler in Köln. Neuerwerbungen und Neufeunde, KJb 11, 79-84. Ritter, S. 1994, Die antiken Bronzen im Römisch-Germanischen Museum Köln: Die Statuetten aus Köln, KJb 27, 355-356. Robinson, H.R. 1975, The Armour of Imperial Rome, New York. Rolland, H. 1965, Bronzes antiques de Haute Provence (XVIIIe supplément à Gallia), Paris. Rose, M.L. 2003, The Staff of Oedipus. Transforming Disability in Ancient Greece, Ann Arbor. Ross, A./R. Feachem 1976, Ritual rubbish? The Newstead pits, in Megaw, J.V.S. (ed.), To Illustrate the Monuments, London, 228-237. Rostovtseff, M./H. Mattingly 1923, Commodus-Hercules in Britain, JRS 13, 91-109. Rothe, U. 2009, Dress and Cultural Identity in the Rhine-Moselle Region of the Roman Empire (BAR Int. Series 2038), Oxford. Roymans, N. 1990, Tribal Societies in Northern Gaul. An anthropological perspective (Cingula 12), Amsterdam. Roymans, N. 1995, The Integration of Lower Rhine Populations in the Roman Empire, in Metzler et al. 1995, 47-63. Roymans, N. 2004, Ethnic Identity and Imperial Power. The Batavians in the early Roman Empire. (AS 10), Amsterdam. Roymans, N. 2015, Roman Grave Monument at the Kaboutersberg, in Roymans et al. 2015, 125-140. Roymans, N./T. Derks 1992, De tempel van Empel, een Bataafse cultusplaats, Spiegel Historiael 27, 423-428. Roymans, N./T. Derks 1994, Het heiligdom te Empel. Algemene beschouwingen, in Roymans/Derks 1994 (eds), 10-38. Roymans, N./T. Derks 2009 (eds), Ethnic Constructs in Antiquity. The role of power and tradition. (AS 13), Amsterdam. Roymans, N./T. Derks 2011, Introductory Essay, in Roymans/Derks 2011 (eds), 1-44. Roymans, N./T. Derks 2011 (eds), Villa Landscapes in the Roman North. Economy, culture and lifestyle (AS 17), Amsterdam. Roymans, N./T. Derks 2015, Texuandri, Tungri, Germani. Different levels of ethnic belonging, in Roymans et al. 2015 (eds), 19-32. Roymans, N./T. Derks/S. Heeren 2007 (eds), Een Bataafse gemeenschap in de wereld van het Romeinse rijk. Opgravingen te Tiel-Passewaaij, Utrecht. Roymans, N./T. Derks/H. Hiddink 2015 (eds), The Roman Villa of Hoogeloon and the Archaeology of the Periphery (AS 22), Amsterdam. Roymans, N./F. Theuws 1993 (eds), Een en al zand. Twee jaar graven naar het Brabantse verleden, ’s-Hertogenbosch.

208

Rüpke, J. 2012, Lived Ancient Religion: Questioning ‘Cults’ and ‘Polis Religion’. Presentation of a new research program,https://www.uni-erfurt.de/fileadmin/public-docs/Max-Weber-Kolleg/6-pdfs/projekte/2012-Ruepke_Lived_anc_rel.pdf (28-8-2017) Rüpke, J. 2018, Pantheon, New Jersey. Schauerte, G. 1987, Darstellungen mütterlicher Gottheiten in den römischen Nordwestprovinzen, in Bauchhenß/Neumann 1987 (eds), 55-102. Scheid, J. 1990, Romulus et ses frères. Le collège des frères Arvales, modèle du culte public dans la Rome des empéreurs (BEFAR 275), Rome. Scheid, J. 2003, An Introduction in Roman Religion, Bloomington Indiana. Schmidt, E. 1997, Venus, LIMC VIII, 192-230. Schoppa, H./K. Hucke 1936, Ein merkwürdiger Schatzfund von Beelen, Kr. Warendorf, Westfalen 21, 403–409. Schuuring, I./C. Tolsma 2004, Verslag bureauonderzoek archeologische waarden Tiel RIO-Vahstal (Zuidnederlandse Archeologische Notities 9), Amsterdam. Seinen, P./T. de Groot/H. van Enckevort 2013, Geheimen van het Wijchens Meer. Romeinse vondsten uit natte context, Westerheem 62-5, 249-260. Siebert, G. 1990, Hermes, LIMC V, 285-289. Siemons, H/E.E.B. Bulten 2014 (eds), Archeologie in het Wateringse Veld, gemeente Den Haag. Van steentijd tot nieuwe tijd. HOP 17, Den Haag. Sier, M.M. (ed.) 2003, Ellewoutsdijk in de Romeinse tijd (ADC-rapport 200), Amersfoort. Simón, F.M. 1990, Iconografia y propaganda ideológica. Júpiter Amón y Medusa en los foros imperiales, in Neronia IV, Alejandro Magno, modelo de los emperadores romanos (Collection Latomus 209), Brussels, 143-162. Simon, E. 1984, Ares/Mars, LIMC II, 505-559. Simon, E. 1992, Mercurius, LIMC VI, 500-537. Sinn, F. 2003, Die Grabkammer in Köln-Weiden. Formen Stadtrömischer Privatgräber für den lokalen Grabbrauch, in Noelke, P. (ed.) 2003, Romanisation und Resistenz in Plastik, Architektur und Inschriften der Provinzen des Imperium Romanum. Neue Funde und Forschungen, Mainz am Rhein, 307-325. Slofstra, J. 1982, Een inheems-Romeinse villa op de Kerkakkers bij Hoogeloon, in Slofstra, J./H.H. van Regteren-Altena/N. Roymans/F. Teuws 1982 (eds), Het Kempenproject. Een regionaal-archeologisch onderzoeksnprogramma (BSBH 22), Waalre, 102-112. Slofstra, J./H. H. van Regteren-Altena/N. Roymans/F. Teuws 1985 (eds), Het Kempenproject 2. Een regionaal-archeologisch onderzoeksnprogramma (BSBH 27), Waalre. Spannagel, M. 1999, Exemplaria Principis. Untersuchungen zu Entstehung und Ausstattung des Augustusforums, Heidelberg. Spielmann, K.A. 2008, Crafting the Sacred: Ritual places and paraphernalia in small-scale societies, in Wells, E.C./P. A. McAnany (eds) Dimensions of Ritual Economy (Research in Economic Anthropology 27), Bingley, 37-72. Spickermann, W./H. Cancik/J. Rüpke 2001 (eds), Religion in den Germanischen Provinzen Roms. Tübingen. Spickermann, W. 2008, Germania Inferior. Religionsgeschichte des römischen Germanien II. Tübingen. Steures, D.C. 2004, Godsdiensten in laat-Romeins Nijmegen. Vereersters van Dionysus, Sabazius, Christus en een paalidool, Westerheem 53-1, 2-3. Steures, D.C. 2013, The Late Roman Cemeteries of Nijmegen. Stray finds and excavations 1947-1983 (Description of the archaeological collections in Museum Het Valkhof at Nijmegen XV), Amersfoort. Stissi, V.V. 2002, Pottery to the People. The production, distribution and consumption of decorated pottery in the Greek world in the Archaic period (650-480 BC), unpublished PhD thesis, Amsterdam. Stoll, O. 2007, Der Gott der arabischen Legion: Zeus Ammon-Serapis und die legio III Cyrenaica in der römischen Provinz Arabia, in Blois, L. de/E. Lo Cascio (eds) The Impact of the Roman Army (200 B.C. – A.D.

209

476): Economic, Social, Political, Religious and Cultural Aspects. Proceedings of the Sixth Workshop of the International Network Impact of Empire (Roman Empire, 200 B.C. – A.D. 476), Capri, March 29 – April 2, 2005 (Impact of Empire 6), Leiden, 439-461. Stoll, O. 2011, The Religion of the Armies, in Erdkamp 2011 (ed.), 451-476. Stuart, P. et al. 1971, Deae Nehalenniae, Middelburg, exhibition catalogue. Stuart, P. 1977, Een Romeins grafveld uit de eerste eeuw te Nijmegen. Onversierde terra sigillata en gewoon aardewerk (Beschrijving van de verzamelingen in het Rijksmuseum Kam te Nijmegen VIII), Nijmegen. Stuart, P. 2013, Nehalennia van Domburg. Geschiedenis van de monumenten I-II, Utrecht. Swinkels, L.J.F. 1982, Mural Paintings from a Roman Villa at Druten, the Netherlands, in Liversidge, J. 1982 (ed.), Roman Provincial Wall Painting of the Western Empire (BAR Int. Series 149), Oxford, 183-192. Swinkels, L.J.F. 1993, Een hond met een vraagteken, Westerheem 43-6, 314-315. Swinkels, L.J.F. 1994, Een vergoddelijkte Hercules en enkele andere bronsfiguren, in Roymans/Derks 1994 (eds), 82-91. Swinkels, L.J.F. 2011, Pracht en praal op kleine schaal. Een Romeins miniatuurtje uit de Betuwe, Gelders Erfgoed 2011-1, 10-11. Swinkels, L.J.F. 2015, Een hellehond uit Beuningen, Nieuwsbrief Vrienden Museum Het Valkhof 41, februari 2015. Swinkels, L.J.F. 2017a, The Wall Painting Fragments, in Vos, W.K,/C.C. Bakels/T.A. Goossens (eds), The Roman Villa at Maasbracht. The archaeology and history of a Roman settlement on the banks of the river Meuse (province of Limburg, The Netherlands). In memory of Willem J.H. Willems (APL 46), Leiden, 83-127. Swinkels, L.J.F. 2017b, Een bronzen Hercules uit Nijmegen-Oosterhout, Jaarboek Numaga 2017, 183-190. Swinkels, L.J.F. 2018, Extriti sint. Een nieuwe lezing van het vervloekingstablet van Bodegraven, Lampas 2018, 51-3, 197-217. Taayke, E. 1996, Die einheimische Keramik der nördlichen Niederlande, 600 v.Chr. bis 300 n.Chr, BROB 42, 9-208. Taayke, E. 2003, Wir nennen sie Franken und sie lebten nördlich des Rheins, 2.-5.Jh., in Taayke, E./ J.H. Looijenga/O.H. Harsema/H.R. Reinders 2003 (eds), Essays on the Early Franks (Groningen Archaeological Studies 1), Groningen, 1-23. Taayke, E. 2005, Het noordelijke kustgebied in de IJzertijd en Romeinse tijd, in Knol et al. 2005, 153-163. Taayke, E./C. Peen/M. van der Harst-Domburg/W.K.Vos 2012, Ede vol erven. Germaanse bewoning op de rand van een wereldrijk.(500 v.Chr. tot 500 na Chr.), Leiden. Taylor, M.V./R.G. Collingwood 1927, Roman Britain in 1927, JRS 17, 184-219. Taylor, M.V. 1944, A Clay Figurine from Chichester, The Antiquaries Journal 24, 152-154. Therkorn, L.L. 2004, Landscaping the Powers of Darkness & Light: 600 BC - 350 AD settlement concerns of North Holland in wider perspective, unpublished PhD thesis, Amsterdam. Therkorn, L.L./E.A. Besselsen/J.F.S. Oversteegen 2006, Assendelver Polders Revisited. Excavations 1997, Amsterdam. Thomas, R. 2014, Denkmäler der Matronenverehrung in der CCAA (Köln), KJb 47, 91-178. Thuillier, J.P. 1989, Les desultores de l’Italie antique, Comptes rendus des séances de l’Académie des Inscriptions et Belles-Lettres 133, 33-53. Tichelman, G. 2005 (ed.), Het villacomplex Kerkrade-Holzkuil (ADC-rapport 155), Amersfoort. Tichelman, G. 2016, Romeinse tijd in Limburg. Een actuele kennisstand van de Romeinse tijd inLimburg aan de hand van archeologisch onderzoek tussen 2007 en 2013. http://www.sam-limburg.nl/document/442/ romeinse-tijd-in-limburg.html (1-1-2018) Toorians, L. 1994, De mythe van Albiobola, Mededelingen van de Stichting A.G. van Hamel voor Keltische Studies 4, 98-100. Toorians, L. 2003, Magusanus and the ‘Old Lad’: a case of Germanicised Celtic, NOWELE 42, 13-28. Toynbee, J.M.C. 1964, Art in Britain under the Romans, Oxford. Tran Tam Tinh, V. 1971, Le culte des divinités orientales à Herculaneum (EPRO 17), Leiden.

210

Trentin, L. 2009, What’s In a Hump? Re-examining the hunchback in the villa Albani-Torlonia, CCJ 55, 130-156. Trentin, L. 2011, Deformity in the Roman Imperial Court, Greece & Rome, Second Series 58-2, 195-208. Trentin, L. 2015, The Hunchback in Hellenistic and Roman Art, London/New York. Trimpe Burger, J. 1973, The Islands of Zeeland and South Holland in Roman Times, BROB 23, 135-148. Tuijn, W. 2006, Een laatste vondst uit de Uivermeertjes van Deest, in Jaarverslag AWN 2005 AWN-Nijmegen e.o., Nijmegen, 37. Ufkes, A. 2002, Keramische objecten, in Niekus/Huisman 2002, 47-65. Van Andringa, W. 2002, La religion en Gaule romaine. Piété et politique (1er-3e siècle ap. J.-C.), Paris. Van Andringa, W. 2007, Religion and Integration of Cities in the Empire in the Second Century AD, in Rüpke, J. (ed.), A Companion to Roman Religion, Oxford, 83-95. Van Andringa, W. 2011, New Combinations and New Statuses. The indigenous gods in the pantheons of the cities of Roman Gaul, in North, J.A./S. R. F. Price (eds), The Religious History of the Roman Empire: Pagans, Jews, and Christians (Oxford Readings in Classical Studies), Oxford/New York, 109-137. Van Beek, R. 2009, Reliëf in tijd en ruimte. Interdisciplinair onderzoek naar bewoning en landschap van Oost-Nederland tussen vroege prehistorie en middeleeuwen, Enschede. Van Beek, R. /W.A. van Es 1964, Nederzettingssporen uit de Laat-Romeinse keizertijd bij Dalfsen (Ov.), Westerheem 13.1, 13-28. Van Bergen, A./A. Janssen/W. Tuijn 2008, Verslag van de verzamelde gegevens en vondsten van de Romeinse vindplaats nabij de boerderij ‘De Uiversnest’, te Deest, gemeente Druten (Gld.), Nijmegen. Van Boekel, G.M.E.C. 1987, Roman Terracotta Figurines and Masks from the Netherlands, Heerhugowaard. Van Boekel, G.M.E.C. 1989, Terracottabeeldjes van de Scheveningseweg. Romeinse vondsten uit ’s Gravenhage II (VOM-reeks 1989, 3), The Hague. Van Boekel, G.M.E.C. 1993, Terracotta figurines and Masks, in Van Dierendonck/Hallewas/Waugh 1993, 82-120. Van Boekel, G.M.E.C. 1996, Romeinse terracotta’s (Museumstukken 6), Nijmegen. Van den Broeke, P.W. 1980, Een rijk gevulde kuil met nederzettingsmateriaal uit de IJzertijd, gevonden te Geleen, prov. Limburg, APL 13, 101-113. Van den Broeke, P.W. 2002a, Vindplaatsen in vogelvlucht. Beknopt overzicht van het archeologische onderzoek in de Waalsprong 1996-2001 (Archeologische Berichten Nijmegen. Rapport 1), Nijmegen. Van den Broeke, P.W. 2002b, Een vurig afscheid? Aanwijzingen voor verlatingsrituelen in ijzertijdnederzettingen, in Fokkens, H./R. Jansen (eds), 2000 jaar bewoningsdynamiek. Brons- en ijzertijdbewoning in het Maas-Demer-Scheldegebied, Leiden, 45-61. Van den Broeke, P.W. 2004, Rituelen in de Waalsprong, Ulpia Noviomagus 9, Nijmegen. Van den Broeke, P.W. 2015, Het verlatingsritueel: een poging tot reconstructie, in Ball, E.A.G./S. Arnoldussen (eds) Metaaltijden 2. Bijdragen in de studie van de metaaltijden, Leiden, 83-99. Van den Broeke, P. W. /W.A.B. van der Sanden 1987 (eds), Getekend zand. Tien jaar archeologisch onderzoek in Oss-Ussen (Bijdragen tot de studie van het Brabants heem 31), Waalre. Van den Hengel, L. 2016, Gladatorenspelen in de Romeinse tijd, in Koster/Mols 2016 (eds), 83-92. Van den Hurk, L.J.A.M. 1986, The Tumuli from the Roman Period of Esch, Province of North Brabant, PhD thesis, Nijmegen. Van der Feijst, L. 2007, De vicus te Naaldwijk, Amsterdam, http://www.ethesis.net/naaldwijk/naaldwijk.htm (5-5-2016). Van der Feijst, L.2012 (ed.), Vechten tegen het wassende water in de Romeinse tijd. Een archeologische opgraving in plangebied Hoogeland-Oost ’t Zand Heultje te Naaldwijk (ADC-rapport 3186), Amersfoort. Van der Feijst, L.2012, Metaal uit de Romeinse tijd, in Geerts, R.CA./H.A.P. Veerman 2012 (eds), Romeinse bewoning tussen IJzertijdgraven. Een archeologische opgraving te Groesbeek-Hüsenhoff (ADC rapport 2687), Amersfoort, 105-114.

211

Van der Gaauws, P.G./H. van Londen 1992, De Hoge Woerd; een boor- en weerstandsonderzoek nuur het Romeinse castellum van De Meern (RAAP-rapport 63), Amsterdam. Vanderhoeven, A. 2015, Town-Country Relations from the Perspective of Roman Tongeren, in Roymans/ Derks/Hiddink 2015 (eds), 189-206. Van der Mijle Meijer, R.A.2011, Scheveningseweg gemeente Den Haag. Archeologische begeleiding rioolvervanging (Haagse Archeologische Rapportage 1116), The Hague. Van der Sanden, W.A.B.1987, Oss-Ussen: de nederzettingen, in Broeke/Sanden 1987 (eds), 53-66. Van der Velde, H.M. 2002, Raalte, De Zegge, Heeten, Hordelman en Heeten, De Telgen, Overijssels Erfgoed, 42-44. Van der Velde, H.M./M.C. Kenemans 2002, Aanvullend onderzoek op een toekomstige industrielocatie op de Boeteler Enk, Zegge IV (ADC-rapport 137), Amersfoort. Van der Velde, H.M./S. Ostkamp/H.A.P. Veldman/S.Wyns 2009 (eds), Venlo aan de Maas: van vicus tot stad. Sporen van een Romeinse nederzetting en stadsontwikkeling uit de Middeleeuwen en Nieuwe Tijd in het plangebied Maasboulevard (ADC-monografie 7-2), Amersfoort. Van der Velde, H.M./N.M. Prangsma 2002, Aanvullend archeologisch onderzoek op de planlokatie ‘De Telgen’ te Heeten (gem. Raalte) (ADC-rapport 138), Amersfoort. Van Diepen, L. 2015, Aardewerk, in Derks/Fraiture 2015 (eds), 86-106. Van Diepen, L./M. Polak 2009, Resten van de steen- en pottenbakkerij van het Romeinse leger op Landgoed Holthurnsche Hof in Berg en Dal (gem. Groesbeek) (Auxiliaria 8), Nijmegen. Van Dierendonck, R.M./D.P. Hallewas/K.E. Waugh (eds) 1993, The Valkenburg excavations 1985-1988. Introduction and detail studies (Nederlandse Oudheden 15/Valkenburg project 1), Amersfoort. Van Dierendonck, R.M.//W.K.Vos 2013 (eds), De Romeinse agglomeratie Aardenburg. Onderzoek naar de ontwikkeling, structuur en datering van de Romeinse castella en hun omgeving, opgegraven in de periode 1955-heden (Hazenberg Archeologische Serie 3), Zwolle. Van Dierendock, R.M./ W.K.Vos/G.P.A. Besuijen 2013, Synthese: zonering, karakter en chronologie van Romeins Aardenburg, in Van Dierendonck/Vos 2013 (eds),287-344. Van Dijk, M. 2015, De Venus van Heerde. Romeins mysterie aan de IJssel, unpublished thesis, Nijmegen. Van Driel-Murray, C. 1994, Wapentuig voor Hercules, in Roymans/Derks 1994 (eds), 92-107. Van Driel-Murray, C. 1995, Gender in question, in Rush, P. 1995 (ed.), Theoretical Roman Archaeology Conference: Second Conference Proceedings 1992, Avebury, 3-21. Van Driel-Murray, C. 2008, Those who wait at home: the effect of recruitment on women in the Lower Rhine area, in Brandl, U. 2008 (ed.) Frauen und römisches Militär (BAR Int. Series 1759), Oxford, 82-91. Van Enckevort, H. 2005a, De Romeinse cultusplaats: de sporen, in Enckevort/Thijssen 2005b (eds), 37-50. Van Enckevort, H. 2005b, Het forum, in Willems et al. 2005 (eds), 93-97. Van Enckevort, H. 2011, Archeologisch onderzoek aan de Hugo de Grootstraat, (Archeologische Berichten Nijmegen. Rapport 27), Nijmegen. Van Enckevort, H. 2012, Gebundelde Sporen, Zwolle. Van Enckevort, H. 2010a, Sporen in de vroeg-Romeinse tijd, in Van Enckevort/Heirbaut 2010 (eds), 55-98. Van Enckevort, H. 2010b, Een ijzeren masker uit de kelder, in Van Enckevort/Heirbaut 2010 (eds), 216-220. Van Enckevort, H./J. Haalebos/J. Thijssen 2000, Nijmegen, legerplaats en stad in het achterland van de Romeinse limes, Abcoude. Van Enckevort, H./T. Hazenberg 1997, Romeins masker uit de klei getrokken, Spiegel Historiael 1-32, 38-39. Van Enckevort H./E.N.A. Heirbaut 2010 (eds), Opkomst en ondergang van Oppidum Batavorum, hoofdplaats van de Bataven. Opgravingen op de St. Josephhof in Nijmegen 1 (Archeologische Berichten Nijmegen. Rapport 16), Nijmegen. Van Enckevort H./E.N.A. Heirbaut 2013, Soldaten und Zivilisten - Frührömische Gräber aus Nijmegen, in Berke, S./T. Mattern 2013 (eds), Römische Gräber augusteischer und tiberischer Zeit im Westen des Imperiums. Akten der Tagung vom 11. bis 14. November 2010 in Trier (Contributions to the Study of Ancient World Cultures 63), Göttingen, 109-122.

212

Van Enckevort, H./J. Hendriks 2015, Het platteland van Romeins Limburg, in Tummers, P./A.J. Bijsterveld/A. Knotter/L. Wessels (eds), Limburg. Een geschiedenis I, Maastricht, 115-152. Van Enckevort, H./J. Thijssen 1996, Graven met beleid. Gemeentelijk archeologisch onderzoek in Nijmegen 19891995, Abcoude. Van Enckevort, H./J. Thijssen 2002 (eds), Cuijk. Een regionaal centrum in de Romeinse tijd (Archeologische Berichten Nijmegen 5), Utrecht. Van Enckevort, H./J. Thijssen 2005a, De stad, in Willems et al. 2005 (eds), 97-111. Van Enckevort, H./J. Thijssen 2005b (eds), In de schaduw van het Noorderlicht. De Gallo-Romeinse tempel van Elst-Westeraam, Abcoude. Van Enckevort, H./J. Thijssen 2005c, De vierde eeuw, in Willems et al. 2005 (eds), 129-143. Van Enckevort, H./J. Thijssen 2014, Het Valkhof en omgeving tot het einde van de Romeinse tijd, in Peterse/ Verhoeven et al. 2014 (eds), 23-41. Van Enckevort, H./W.J.H. Willems 1994, Roman Cavalry Helmets in Ritual Hoards from the Kops Plateau at Nijmegen, the Netherlands, Journal of Roman Military Equipment Studies 5, 125-137. Van Es, W.A. 1960, De Romeinse muntvondsten uit de drie noordelijke provincies. Een periodisering der relaties (Varia Bio-Archaeologica 13. Scripta Academica Groningana), Groningen. Van Es, W.A. 1964, Mercurius at Dalfsen, BABesch 39, 167-173. Van Es, W.A. 1967, Wijster. A native village beyond the imperial frontier. 150-450 AD, Groningen. Van Es, W.A. 1981, De Romeinen in Nederland2, Haarlem. Van Es, W.A. 1994, Verhaal, in Es, W.A. van/W.A.M. Hessing 1994 (eds), Romeinen, Friezen en Franken in het hart van Nederland, Giethoorn, 22-119. Van Es, W.A. 2005, Romeinse importen in de wierden van Groningen en Friesland, in Knol et al. 2005 (eds), 167-181. Van Es, W.A. 2012, Van Wijster naar Ede over de Rijn (Reuvenslezing 24), Leiden. Van Es, W.A./W.A.M. Hessing 1980, Herculi Magusano, een wijaltaar uit Houten, in Feestbundel aangeboden aan prof. dr. D.P. Blok, Hilversum, 81-90 (= ROB overdruk nr. 364). Van Es, W.A./W.A.M. Hessing 1994 (eds), Romeinen, Friezen en Franken in het hart van Nederland, Giethoorn. Van Es, W.A./W.J.H. Verwers 1980, Excavations at Dorestad 1. The harbour: Hoogstraat 1 (Nederlandse Oudheden 9), Amersfoort. Van Es, W.A./M. Miedema/S.L. Wynia 1985, Eine Siedlung der römischen Kaiserzeit in Bennekom, Provinz Gelderland, BROB 35, 533-652. Van Giffen, A.E. 1934, Oudheidkundige aanteekeningen over Drentsche vondsten, Nieuwe Drentsche Volksalmanak 52, 85-93. Van Giffen, A.E. /W. Glasbergen 1948, Thermen en castella te Heerlen-Coriovallum, L’Antiquité Classique 17 (Miscellanea Philologica Historica et archaelogia in honorem Huberti Van De Weerd), 199-236. Van Ginkel, E.J./J.A. Waasdorp 1992 (eds), De archeologie van Den Haag. Deel 2: de Romeinse tijd (VOM-reeks 1992, 4), The Hague. Van Hoof, L. 2007, Variaties op een rechthoek. Huizenbouwtradities en huisoffers in Romeins Nederland, in Jansen/Louwe Kooymans 2007 (eds), 255-270. Van Londen, H. 2000, Landscape and Water Management: Midden-Delfland, a Region South of the Limes, in Grünewald, T. (ed.) Germania Inferior. Beiträge des deutsch-niederländischen Kolloquiums im Regionalmuseum Xanten, 21.-24. September 1999 (= Reallexikon der germanischen Altertumskunde, Erganzungsband 28) Berlin, 169-184. Van Meij, L./W.F. Reigersman-van Lidth de Jeude 2009, Aardewerk, in Eimermann, E. 2009 (ed.), 97-112. Van Oppen de Ruiter, B.F. /W. Hupperetz 2014, Een mooie Mercurius, in Hupperetz et al. 2014 (eds), 108. Van Regteren Altena, J.H./P.J.A. van Mensch/G. IJzereef 1977, Bronze Age Clay Animals from Grotebroek, in Beek van, B.L./R.W. Brandt/W. Groenman-van Waateringe (eds), Ex Horreo (IPP 1951-1976), Amsterdam, 241-254.

213

Van Renswoude, J. 2008, Metaal, in Derks et al. 2008, 74-82. Van Renswoude J./J. van Kerckhove 2009 (eds), Opgravingen in Geldermalsen-Hondsgemet (ZAR 35), Zwolle. Van Straten, F. 1985, Om de hulp der goden, Hermeneus 57-4, 224. Van Tent, W.J. 1978, A Native Settlement at Jutphaas, Municipality of Nieuwegein, BROB 28, 199-239. Van Veen, M.M.A./J. A. Waasdorp 2000, Archeologische-geologische kaart van Den Haag (Haagse Oudheidkundige Publicaties 5), The Hague. Van Zalinge, A.1997, Het Kops Plateau. Religieuze voorwerpen in een Romeins legerkamp, unpublished thesis, Leiden. Van Zoolingen, R.J. 2011, Rural Cult Places in the Civitas Cananefatium, JALC 3-1/2, www.jalc.nl (18-12-2015). Veen, C.W.A. 2010, Een ruiter op een skateboard, in Van Enckevort/Heirbout 2010 (eds), 213-216. Veen, C.W.A. 2014a, Bacchus and Jupiter-Ammon, two bronze sculptures from Roman Nijmegen, The Netherlands, BABesch 89, 129-147. Veen, C.W.A 2014b, Een opmerkelijk bronsdepot, in Heirbaut, E. 2014 (ed.), Opgravingen op de Hessenberg 1. Onderzoek naar het grafveld van Ulpia Noviomagus (Archeologische Berichten Nijmegen, Rapport 49), Nijmegen, 227-260. Veen, C.W.A. 2017, Een beeldje uit een greppel, in Woltering, P.J., Bewoningsgeschiedenis van Texel, De opgravingen aan de Beatrixlaan in Den Burg (Late IJzertijd - Romeinse tijd: 100 v.Chr.-325 na Chr.) (Noord-Hollandse Archeologische Publicaties 3), Castricum, 250-252.  Veen, C.W.A. 2019, Machtige moeders, Numaga Jaarboek 2019, 155-161. Verhart, L. 1993, De prehistorie van Nederland, Amsterdam. Verhoeven, A./F. Ector 1984, Sabazius in Noord-Brabant?, Brabants Heem 36, 138-149. Verlinde, A.D. 1987, Colmschate, in Archeologische Kroniek van Overijssel over 1986 (Overdruk ROB 296), Zwolle, 169-187. Verlinde, A.D. 1991, Huizen uit de Bronstijd en vroege IJzertijd te Colmschate, in Fokkens H./N. Roymans 1991 (eds), Nederzettingen uit de bronstijd en de vroege ijzertijd in de Lage Landen (NAR 13), Amersfoort, 31-40. Verlinde, A.D./M. Erdrich 1998, Eine germanische Siedlung der späten Kaiserzeit mit umwehrter Anlage und umfangreicher Eisenindustrie in Heeten, Provinz Overijssel, Niederlande, Germania 76, 693-719. Verlinde, A.D./M. Erdrich 2006, Het Germaanse grafveld te Deventer-Colmschate, opgraving 1984, in Groenewoudt, B.J./R.M. van Heeringen /G.H. Scheepstra (eds), Het zandeilandenrijk van Overijssel, Amersfoort, 271355. Vermeulen, W. 1932, Een Romeinsch grafveld op den Hunnerberg te Nijmegen (uit den tijd van Tiberius-Nero), Amsterdam. Verrijt, M. 1985, Opgravingen tussen Waalkade en het Groene Balkon in Nijmegen, in Jaarverslag AWN 1985, Nijmegen, 47-49. Versluys, M.J. 2002, Aegyptiaca Romana. Nilotic Scenes and the Roman Views of Egypt (Religions in the Graeco-Roman World 144), Leiden. Versnel, H.S. 1976, Two Types of Roman Devotion, Mnemosyne 29-4, 365-410. Versnel, H.S. 1991, Some Reflexions on the Relationship Magic-Religion, Numen 38-2, 177-197. Verzàr, M. 1977, Aventicum II. Un temple du culte impérial, Avenches. Volkers, T.B. 2015/2016, Die Terra Sigillata aus friesischen Terpen (mit einem Beitrag von M. Polak), Palaeohistoria 57/58, 235-312, Groningen. Vollgraff, C.W. 1931, Inscriptiones Traiectenses, Mnemosyne, New Series 59-3, 249-265. Von Gonzenbach, V.C.1967, Kleinvotive des zweiten Jahrhunderts n. Chr. und Militär in Vindonissa, Jahresbericht /Gesellschaft Pro Vindonissa, 7-30. Von Gonzenbach, V.C.1986, Die römischen Terracotten in der Schweiz: Untersuchungen zu Zeitstellung, Typologie und Ursprung der mittelgallischen Tonstatuetten. Band B. Katalog und Tafeln, Bern. Von Gonzenbach, V.C. 1995, Die römischen Terracotten in der Schweiz: Untersuchungen zu Zeitstellung, Typologie und Ursprung der mittelgallischen Tonstatuetten. Band A, Tübingen/Basel.

214

Von Hesberg, H. 1999 (ed.), Das Militär als Kulturträger in römischer Zeit, Cologne. Vos, W.K. 1999, Een toevalsvondst in een waterput, in Hessing, W/M. Polak/W.K.Vos/S.L.Wynia 1999, Romeinen langs de snelweg. Bouwstenen voor Vechtens verleden, Abcoude2, 53-58. Vos, W.K. 2009, Bataafs platteland. Het Romeinse nederzettingslandschap in het Nederlandse Kromme-Rijngebied (NAR 35), Amersfoort. Vos, W.K./E. Blom/T. Hazenberg 2010, Romeinen in Woerden. Het archeologisch onderzoek naar de militaire bezetting en de scheepvaart van Laurium, Leiden. Vos, P./S de Vries 2013, 2e generatie palaeogeografische kaarten van Nederland (versie 2.0), Utrecht. www. archeologieinnederland.nl (10-2-2014). Waasdorp, J.A. 1999, Van Romeinse soldaten en Cananefaten. Gebruiksvoorwerpen van de Scheveningseweg (VOM-reeks 2), The Hague. Waasdorp, J.A. 2012, (ed.), Den Haag Ockenburgh. Een fortificatie als onderdeel van de Romeinse kustverdediging (Haagse Oudheidkundige Publicaties 13), The Hague. Waasdorp, J.A./K. Zee 1988, De vergeten verzamelingen van Ockenburgh: Romeinse vondsten uit ’s-Gravenhage (VOM-reeks 1988, 4), The Hague. Waraksa, E. 2009, Female Figurines from the Mut Precinct. Context and ritual function, Göttingen. Waterbolk, H.T./W. Glasbergen 1955, Der spätrömische Goldschatz von Beilen. Fundbericht und Beschreibung. Palaeohistoria 4, 81-101. Webster, J. 1995, Interpretatio: Roman Word Power and the Celtic Gods, Britannia 26, 153-161. Wesch-Klein, G. 2011, Recruits and Veterans, in Erdkamp, P. 2007 (ed.), 435-450. Wesselingh, D.A. 2000, Native Neighbours. Local settlement system and social structure in the Roman period at Oss, The Netherlands (APL 32), Leiden. Weiß-König, S. 2018, Een loden vloektablet uit Ulpia Noviomagus, Numaga Jaarboek 2018, 161-165. Wiepking, C. 1997, De vondsten, in Enckevort, H. van/K. van der Graaf 1997 (eds), Aanvullend Archeologisch Onderzoek Kerkrade-Winckelen, verkaveling Maar-West (RAM 51), Amersfoort, 18-22. Wiepking, C. 2005, Aardewerk, in Tichelman, G. 2005 (ed.), 177-219. Willems, W.J.H. 1981, Romans and Batavians. A regional study in the Dutch eastern river area I, BROB 31, Amersfoort, 7-217. Willems, W.J.H. 1984, Romans and Batavians. A regional study in the Dutch eastern river area II, BROB 34, 39-331. Willems, W.J.H. 1986, De Romeinse villa te Voerendaal. Opgraving 1985, Archeologie in Limburg 28, 143-150. Willems, W.J.H. 1988, Reiche Gutsherren in Maasbracht, in Hiller, H. (ed.), Villa rustica. Römische Gutshöfe im Rhein-Maas Gebiet, Freiburg, 28-33. Willems, W.J.H. 2005, De kampementen van de hulptroepen, in Willems et al. 2005 (eds), 70-73. Willems, W.J.H. 2005, Het kampdorp, in Willems et al. 2005 (eds), 89-93. Willems, W.J.H./H.van Enckevort 2005, Het Kops Plateau, in Willems et al. 2005 (eds), 63-70. Willems, W.J.H./H.van Enckevort/J.K. Haalebos/J. Thijssen 2005 (eds), Nijmegen, geschiedenis van de oudste stad van Nederland. Prehistorie en Oudheid, Wormer. Willems, W.J.H./L.I. Kooistra 1987, De Romeinse villa te Voerendaal. Opgraving 1986, Archeologie in Limburg 32, 29-38. Willems, W.J.H./L.I. Kooistra 1988, De Romeinse villa te Voerendaal. Opgraving 1987, Archeologie in Limburg 37, 137-147. Willems, W.J.H./H. van Enckevort/ P.W. van den Broeke et al. 2009, Batavian Capital at the Imperial Frontier (JRA Suppl. Series 73), Portsmouth, Rhode Island. Willemsen, A. 1998, Kinder delijt. Middeleeuws speelgoed in de Nederlanden, Nijmegen. Willemsen, A. 2003, Romeins speelgoed. Kinderen in een wereldrijk, Zutphen. Witteyer, M. 2003, Göttlicher Baugrund – Die Kultstätte für Isis und Mother goddess Magna unter der Römerpassage in Mainz, Mainz.

215

Woltering, P.J. 2000, The Archaeology of Texel. Four studies on settlement and landscape (1350 BC – AD 1500), Heerhugowaard. Woltering, P.J. 2017, Bewoningsgeschiedenis van Texel. De opgravingen aan de Beatrixlaan in Den Burg (Late IJzertijd-Romeinse tijd: 100 v. Chr-325 na Chr.) (Noord-Hollandse Archeologische Pubicaties 3), Castricum. Woodward, A./P. Leach 1993, The Uley Shrines; Excavation of a ritual complex on West Hill, Uley, Glooucestershire 1977-9, English Heritage Archaeological Report 17. Woolf, G. 1995, The Formation of Roman Provincial Cultures, in Metzler et al. 1995 (eds), 9-18. Woolf, G. 1998, Becoming Roman. The Origins of Provincial Civilization in Gaul, Cambridge. Wrede H. 1972, Die spätantike Hermengalerie von Welschbillig, Berlin. IJzereef, F.G. 1981. Bronze Age animal bones from Bovenkarspel: the excavation at Het Valkje. Amersfoort (Nederlandse oudheden 10, Project Noord-Holland 1). Zadoks, A.N. 1955, The Late Roman Gold Hoard of Beilen II. The coins, Palaeohistoria 4, 103-112. Zadoks, A.N. 1980-1981, Keizerin te Krangeweer, Groningse Volksalmanak. Historisch jaarboek voor Groningen, 207-211. Zadoks, A.N./W.J.Th. Peters/W.A. van Es 1967, Roman Bronze Statuettes from the Netherlands I. Statuettes Found North of the Limes, Groningen. Zadoks, A.N./W.J.Th. Peters/W.A. van Es 1969, Roman Bronze Statuettes from the Netherlands II. Statuettes Found South of the Limes, Groningen. Zadoks, A.N./W.J.Th. Peters/A.M. Witteveen 1973, Description of the Collections in the Rijksmuseum G.M. Kam, Nijmegen VII. The Figural Bronzes, Nijmegen. Zadoks, A.N./W.J.T. Peters 1976, Three Roman Bronze Statuettes from Ede and Bennekom (Gelderland), BROB 26, 163-167. Zadoks, A.N./A.M. Gerhartl-Witteveen 1983, The Figural Bronzes. Supplement, Nijmegen. Zampieri, G./B. Lavarone 2000, Bronzi antichi del Museo Archeologico di Padova. Museum catalogue, Rome. Zander, M. 2013, Het Romeinse castellum Albaniana, in Bakker, P./J.W.Bron 2013, Gered uit de grond. Romeinse vondsten van castellum Albaniana, Leiden, 11-12. Zandstra, M.J.M./M. Polak 2012, De Romeinse versterkingen in Vechten-Fectio. Het archeologisch onderzoek in 1946-1947, Nijmegen. Zandstra, M.J.M. 2019, Miles away from home. Material culture as a guide to the composition and deployment of the Roman army in the Lower Rhine area during the 1st century AD, Nijmegen. Zanker, P. 1968, Forum Augustum. Das Bildprogramm, Tübingen. Zanker, P. 1995, Pompeji. Stadtbild und Wohngeschmack, Mainz. Zanker, P./B.C. Ewald 2004, Mit Mythen leben. Die Bilderwelt der römischen Sarkophage, München. Zee, K. 2005, Inheemse en Romeinse religie, in Willems et al. 2005 (eds), 183-196. Zimmer, G. 1982, Römische Berufsdarstellungen (Archäologische Forschungen 12), Berlin.

216

List of maps, figures and drawings Maps: Rien Polak Source map: Vos, P. & S. de Vries 2013: 2e generatie palaeogeografische kaarten van Nederland (version 2.0). Deltares, Utrecht. www.archeologieinnederland.nl (10-2-2014). Map 1 Map 2 Map 3 Map 4 Map 5 Map 6 Map 7 Map 8 Map 9 Map 10 Map 11

Map of the Netherlands Map of the northern Netherlands Map of Friesland and Groningen Map of North Holland Map of Drenthe, Overijssel and northern Gelderland Map of the southern Netherlands Map of South Holland and Zeeland Map of Utrecht Map of southern Gelderland Map of North Brabant and Limburg Map of the Netherlands with ritual deposits

Figures not to scale; a-numbers refer to drawings.

Figures chapter 2 Fig. 2.1 Fig. 2.2 Fig. 2.3 Fig. 2.4 Fig. 2.5 Fig. 2.6 Fig. 2.7 Figs. 2.8-9 Figs. 2.10-11 Fig. 2.12 Fig. 2.13 Fig. 2.14 Fig. 2.15 Fig. 2.16

Terracotta bovine or sheep, Sneek. Inv. nr. F.2006-XI-1903. Photo: Yvonne Boonstra. Gemeente Súdwest-Fryslân. Bronze horse Holwerd. Inv.nr. 1921/10.166. Photo: National Museum of Antiquities, Leiden (RMO). Bronze warrior, Heerenveen-De Knipe. Inv. nr. 2007-II-720. Photo: Collectie Fries Museum, Leeuwarden. Bronze Mercury, Kimswerd. Inv. nr. 50-55. Photo: Fries Museum, Leeuwarden. Collectie Het Koninklijk Fries Genootschap. Detail Mercury Kimswerd. Bronze Isis-Fortuna, Lions. Inv. nr. 55B-23. Photo: Fries Museum, Leeuwarden. Collectie Het Koninklijk Fries Genootschap. Detail Isis-Fortuna. Bronze rearing horse with missing rider, Makkum. Inv. nr. 155-I. Photos: Fries Museum, Leeuwarden. Collectie Het Koninklijk Fries Genootschap. Bronze Jupiter, Menaldum. Inv. nr. 45-152. Photos: Fries Museum, Leeuwarden. Collectie Het Koninklijk Fries Genootschap. Terracotta bird, Oosterbeintum. Inv. nr. 28bis-III. Photo: Fries Museum, Leeuwarden. Collectie Het Koninklijk Fries Genootschap. Terracotta bird, Enumerhoogte. Inv. nr. 1910/01-00001. Photo: Marten de Leeuw. Collectie Groninger Museum, Groningen. Bronze dismounted cavalryman, Ezinge. Inv. nr. 1919-VII.3. Photo: Collectie Groninger Museum, Groningen. Bronze Isis, Houwerzijl. Inv. nr. 1964/VI.2. Photo: Marten de Leeuw. Collectie Groninger Museum, Groningen. Bronze Mars, Joeswerd. Inv.nr. 1990-V.8.

217

Fig. 2.17 Fig. 2.18 Figs. 2.19-20 Figs. 2.21-23 Fig. 2.24 Fig. 2.25 Fig. 2.26 Fig. 2.27 Fig. 2.28 Fig. 2.29 Fig. 2.30 Fig. 2.31 Fig. 2.32 Fig. 2.33 Figs. 2.34-36 Fig. 2.37 Fig. 2.38 Fig. 2.39 Fig. 2.40 Fig. 2.41 Figs. 2.42-43 Figs. 2.44-45 Fig. 2.46 Fig. 2.46a Figs. 2.47-49

218

Photo: Collectie Groninger Museum, Groningen. Bronze horse, Kantens. Inv. nr. 1991-X.5. Photo: Marten de Leeuw. Collectie Groninger Museum, Groningen. Bronze mother goddess, Rottum. Inv. nr. 1915/VI.2. Photo: Marten de Leeuw. Collectie Groninger Museum, Groningen. Bronze Mercurius, Rottum. Inv. nr. 1916/I.1. Photos: Marten de Leeuw. Collectie Groninger Museum, Groningen. Chalk bear, Wirdum. Inv. nr. 1922- XII 5. Photos: Stephan Mols. Noordelijk Archeologisch Depot, Nuis. Detail bear. Terracotta male figure, Wirdum. Inv.nr. 1920/05/0008. Photo: Marten de Leeuw. Collectie Groninger Museum, Groningen. Terracotta bird, Schagen. Inv. nr. PNH 4096-09. Photo: Kees Zwaan. Provinciaal Depot voor Archeologie Noord-Holland Detail terracotta bird. Terracotta head woman, part of couple, Noordbarge. Inv. nr. 1855/I.76g. Photo: Stephan Mols. Noordelijk Archeologisch Depot, Nuis. After Boekel 1987, nr. 168. Terracotta head of Apollo/Bacchus, Noordbarge. Inv. nr. 1855/I.76i. Photo: Stephan Mols. Noordelijk Archeologisch Depot, Nuis. Terracotta head Minerva, Noordbarge. Inv. nr. 1855/I.76f1. Photo: Stephan Mols. Noordelijk Archeologisch Depot, Nuis. Terracotta head Minerva, Noordbarge. Inv. nr. 1855/I.76f. Photo: Stephan Mols. Noordelijk Archeologisch Depot, Nuis. Terracotta Venus, Noordbarge. Inv. nr. 1855/I.76c. Photo: Stephan Mols. Noordelijk Archeologisch Depot, Nuis. Terracotta warrior, Noordbarge. Inv. nr. 1855-I.76h. Photos: Michiel Rooke. Noordelijk Archeologisch Depot, Nuis. Bronze donkey head, Noordbarge. Inv. nr. 1854/I.31. Drents Museum, Assen. Bronze boar, Noordbarge. Inv. nr. 1854/I.30. Drents Museum, Assen. Bronze Mercury, Noordbarge. Inv. nr. 1855/I.76. Drents Museum, Assen. Bronze pedestal, Noordbarge. Inv. nr. 1855/I.76. Photo: Jaap Beuker. Drents Museum, Assen. Bronze Fortuna, Bennekom. Inv. nr. PDB.2002.2.EDE.1970.71. Photo: Ronny Meijers. Collectie Museum Het Valkhof, Nijmegen, Provinciaal Depot voor Bodemvondsten Gelderland. Bronze Minerva, Ede-Veldhuizen. Inv. nr. PDB.2002.1.EDEV.1175. Photos: Ronny Meijers. Collectie Museum Het Valkhof, Nijmegen, Provinciaal Depot voor Bodemvondsten Gelderland. Bronze worshipper, Ede-Veldhuizen. Inv. nr. PDB.2002.1.EDEV.1176. Photos: Ronny Meijers. Collectie Museum Het Valkhof, Nijmegen, Provinciaal Depot voor Bodemvondsten Gelderland. Detail patera. Drawing: René Reijnen. Terracotta head of Venus, Wehl. Inv. nr. PDB.2005.11.231.

Fig. 2.47a Fig. 2.50 Fig. 2.51 Figs. 2.52-54 Fig. 2.55 Fig. 2.56 Fig. 2.56a Fig. 2.57 Figs. 2.58-59 Fig. 2.60

Fig. 2.60a1 Fig. 2.60a2 Fig. 2.60a3 Fig. 2.61 Figs. 2.62-64 Figs. 2.65-66 Fig. 2.66a Fig. 2.67 Fig. 2.67a Fig. 2.68 Fig. 2.68a Fig. 2.69 Figs. 2.70-71 Fig. 2.72 Fig. 2.73

Photos: Helma Molenaars. Collectie Museum Het Valkhof, Nijmegen, Provinciaal Depot voor Bodemvondsten Gelderland. Drawing: René Reijnen. Bronze Venus, Heerde. Inv. nr. 1999.1. Photo: Ronny Meijers. Collectie Museum Het Valkhof, Nijmegen. Bronze Fortuna, Zutphen. Private collection. Photo: Ronny Meijers. Bronze water nymph, Naaldwijk. Inv. nr. 27826. Photos: Provinciaal Archeologisch Depot Zuid-Holland. Bronze mother goddess, Koudekerk a/d Rijn. Photo: ROB. Private collection. Bronze mother goddess, Hunerberg Nijmegen. Inv. nr. BE.I.5. Photo: Ronny Meijers. Collectie Museum Het Valkhof, Nijmegen. Drawing: Acker Stratingh 1849, plate VI-3. Bronze Hercules, Ypenburg. Inv. nr. h2013/7.I. National Museum of Antiquities, Leiden (RMO). Bronze man with elephant, Geldermalsen. Inv. nr. 2010.29. Photos: Ronny Meijers. Collectie Museum Het Valkhof, Nijmegen. Terracotta Cybele with two lions. Inv. nr. Bo5.184.AA3.97. Lion 1 Inv. nr. Bo5.184.AA3.98. Lion 2 Inv. nr. Bo5.184.AA2.80. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Drawing: René Reijnen. Drawing: René Reijnen. Drawing: René Reijnen. Bronze Hercules, Oosterhout. Inv. nr. 2015.136. Photo: Ronny Meijers. Collectie Museum Het Valkhof, Nijmegen. Photo: Ronny Meijers. Collectie Museum Het Valkhof, Nijmegen. Terracotta head Minerva, Tiel-Passewaaij. Inv. nr. PDB.2011.1.V045.133. Photos: Helma Molenaars. Collectie Museum Het Valkhof, Nijmegen, Provinciaal Depot voor Bodemvondsten Gelderland. Drawing: René Reijnen. Terracotta head mother goddess, Tiel-Passewaaij. Inv. nr. PDB.2011.1.V166.025. Photo: Helma Molenaars. Collectie Museum Het Valkhof, Nijmegen, Provinciaal Depot voor Bodemvondsten Gelderland. Drawing: René Reijnen. Terracotta head Cybele, Tiel-Passewaaij. Inv. nr. PDB.2011.1.V128.246. Photo: Helma Molenaars. Collectie Museum Het Valkhof, Nijmegen, Provinciaal Depot voor Bodemvondsten Gelderland. Drawing: René Reijnen. Bronze Minerva, Wijchen-Tienakker. Inv. nr. Ti1.18.11. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Bronze Cerberus, Beuningen. Inv. nr. 2011.6. Photos: Ronny Meijers. Collectie Museum Het Valkhof, Nijmegen. Bronze Silenus, river Waal. Inv. nr. e1931/2.61. National Museum of Antiquities, Leiden (RMO). Bronze Mercury, Cuijk. Inv. nr. Cu1.04.127.

219

Fig. 2.74 Figs. 2.75-76

Photo: Rob Mols. Museum Ceuclum. Terracotta head mother goddess. Inv.nr.: Photo: Geert Oldenmenger. Provinciaal depot Noord-Brabant. Bronze African man with drinking cup, Heeswijk. Photos: Godfried Scheijvens. Private collection.

Oppidum Batavorum Fig. 2.77 Fig. 2.78 Fig. 2.79 Fig. 2.80 Fig. 2.81 Fig. 2.81a Fig. 2.82 Fig. 2.83

Bronze billy goat. Inv. nr. Ht3-195. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Bacchus. Inv. nr. Gn1.17.018. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Bronze pedestal. Inv. nr. Gn1.7.3.95. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Bronze Minerva, Waalkade. Private collection. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Bronze foot. Inv. nr. Jo1.12.42. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Drawing: René Reijnen. Bronze roe. Inv. nr. Jo1.35.0.25. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen Bronze desultor. Inv. nr. Jo1.15.180. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen.

Ulpia Noviomagus Fig. 2.84 Fig. 2.84a Fig. 2.85 Fig. 2.86 Figs. 2.87-88 Fig. 2.89 Fig. 2.90 Fig. 2.91 Fig. 2.92 Fig. 2.93

220

Limestone Hercules, Maasplein. Inv. nr. Mp1.4.2.189. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Drawing: René Reijnen. Silver billy goat, Rivierstraat. Inv. nr. Ri1.1.2.16. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Bronze Mercury, Maasplein. Inv. nr. Mp1.17.2.180. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Bronze Mercury, Bronsgeeststraat. Inv. nr. PDB.1993.5.ULP.1985.648. Photos: Helma Molenaars. Collectie Museum Het Valkhof, Nijmegen, Provinciaal Depot voor Bodemvondsten Gelderland. Bronze Mercury, Weurtseweg. Inv. nr. Ww1.103.1. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Bronze Venus, Maasplein. Inv. nr. Mp1.4.1.75. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Bronze caduceus Mercury, Maasplein. Inv. nr. Mp1.13.238. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Bronze Lar, Weurtseweg. Inv. nr. Ww1.4.482. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Bronze leg of goat or sheep, Rivierstraat. Inv. nr. Ri1.2.1.11. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen.

Fig. 2.94 Fig. 2.95 Fig. 2.96 Fig. 2.97 Fig. 2.97a Fig. 2.98 Fig. 2.98a Fig. 2.99 Fig. 2.99a Fig. 2.100 Fig. 2.100a Fig. 2.101 Fig. 2.101a Fig. 2.102 Fig. 2.103 Fig. 2.104 Fig. 2.104a Fig. 2.105 Fig. 2.105a Fig. 2.106 Fig. 2.106a Fig. 2.107 Fig. 2.107a Fig. 2.108 Fig. 2.108a Fig. 2.109 Fig. 2.109a Fig. 2.110 Fig. 2.110a

Bronze fragment of thyrsus, Hessenberg. Inv. nr. Hb4.2.009.Brons03. Photo: Ronny Meijers. Museum Het Valkhof, Nijmegen. Bureau Leefomgevingskwaliteit/ Archeologie, Nijmegen. Bronze fragment of herm Jupiter-Ammon, Hessenberg. Inv. nr. Hb4.2.009.Brons01. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Limestone Minerva, Nijmegen-West. Inv. nr. NVV I. National Museum of Antiquities, Leiden (RMO). Terracotta fragment of Fortuna? Maasplein. Inv. nr. Mp1.1.27.121. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Drawing: René Reijnen. Terracotta fragment of Fortuna? Weurtseweg. Inv. nr. Ww1.1.19.2.1097. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Drawing: René Reijnen. Terracotta fragment of Minerva/Fortuna? Weurtseweg. Inv. nr. Ww1.2.4.405. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Drawing: René Reijnen. Terracotta fragment of Fortuna? Maasplein. Inv. nr. Mp1.5.1.135. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Drawing: René Reijnen. Terracotta fragment of Fortuna? Two fragments, Maasplein. Inv. nrs. Mp1.4.39.326/ Mp1.6.104.597. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Drawing: René Reijnen. Terracotta fragment of Minerva/Fortuna? Weurtseweg. Inv. nr. Ww1.23.2.647. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Terracotta bust of woman, Maasplein. Inv. nr. Mp1.3.247. Photo: author. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Terracotta bust of woman, Brugkwartier. Inv. nr. WD7.00644KER. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Drawing: René Reijnen. Terracotta bust of woman, Weurtseweg. Inv. nr. Ww1. 4.3.440. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Drawing: René Reijnen. Terracotta bust of woman? Weurtseweg. Inv. nr. Ww1. 7.1.1535. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Drawing: René Reijnen. Terracotta Venus, Maasplein. Inv. nr. Mp1.6.104.579. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Drawing: René Reijnen. Terracotta Juno? Maasplein. Inv. nr. Mp1.18.23.240. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Drawing: René Reijnen. Terracotta Mercury, Weurtseweg. Inv. nr. Ww1.23.2.625. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Drawing: René Reijnen. Terracotta Venus, Maasplein. Inv. nr. Mp1.4.39.386. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Drawing: René Reijnen.

221

Fig. 2.111 Fig. 2.111a Fig. 2.112 Fig. 2.112a Fig. 2.113 Fig. 2.113a Fig. 2.114 Fig. 2.114a Fig. 2.115 Fig. 2.116 Fig. 2.117 Fig. 2.117a Fig. 2.118 Fig. 2.119 Figs. 2.120-22

Terracotta Venus, Weurtseweg. Inv. nr. Ww1.3.1.654. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Drawing: René Reijnen. Terracotta bust of boy, Maasplein. Inv. nr. Mp1.3.0.209. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Drawing: René Reijnen. Terracotta lion’s mane, Rivierstraat. Inv. nr. Ri1.6.43.233. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Drawing: René Reijnen. Terracotta horse, Brugkwartier. Inv. nr. WD7.00602. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Drawing: René Reijnen. Terracotta bust of woman/girl? Maasplein. Inv. nr. Mp1.13.35.482. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Terracotta bust of human figure, Maasplein. Inv. nr. Mp1.13.24.415. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Terracotta lion, front legs, Maasplein. Inv. nr. Mp1. 4. Find from soil heap. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Drawing: René Reijnen. Terracotta hand-modelled male figure, Maasplein. Inv. nr. Mp1.5.74.466. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Terracotta voodoo doll, Weurtseweg. Inv. nr. Ww1.9.P153.1617. Photo: Ronny Meijers. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Details. Photos: author.

Canisius excavations Fig. 2.123 Fig. 2.124 Fig. 2.125 Fig. 2.126 Fig. 2.126a Fig. 2.127 Fig. 2.127a Fig. 2.128

222

Bronze Luna. Inv. nr. PDB.2012.1.CA.1992.6203.m. Photo: Ronny Meijers. Collectie Museum Het Valkhof, Nijmegen, Provinciaal Depot voor Bodemvondsten Gelderland. Bronze Mercury. Inv. nr. PDB.2012.1.CA.1988.1321.m. Photo: Ronny Meijers. Collectie Museum Het Valkhof, Nijmegen, Provinciaal Depot voor Bodemvondsten Gelderland. Bronze billy goat. Inv. nr. PDB.2012.1.CA.1992.6204.m. Photo: Ronny Meijers. Collectie Museum Het Valkhof, Nijmegen, Provinciaal Depot voor Bodemvondsten Gelderland. Terracotta Fortuna. Inv. nr. PDB.2012.1.CA.1990.4151. Photo: Helma Molenaars. Collectie Museum Het Valkhof, Nijmegen, Provinciaal Depot voor Bodemvondsten Gelderland. Drawing: René Reijnen. Terracotta Minerva. Inv. nr. PDB.2012.1.CA.1993.6949. Photo: Helma Molenaars. Collectie Museum Het Valkhof, Nijmegen, Provinciaal Depot voor Bodemvondsten Gelderland. Drawing: René Reijnen. Terracotta ithyphallic dog. Inv. nr. PDB.2012.1.CA.1989.3057. Photo: Helma Molenaars. Collectie Museum Het Valkhof, Nijmegen, Provinciaal Depot voor Bodemvondsten Gelderland.

Fig. 2.128a Fig. 2.129 Fig. 2.129a Fig. 2.130 Fig. 2.130a Fig. 2.131 Fig. 2.132 Fig. 2.132a Fig. 2.133 Fig. 2.133a Fig. 2.134

Drawing: René Reijnen. Terracotta head of cow. Inv. nr. PDB.2012.1.CA.1996.9528. Photo: Helma Molenaars. Collectie Museum Het Valkhof, Nijmegen, Provinciaal Depot voor Bodemvondsten Gelderland. Drawing: René Reijnen. Four fragments, terracotta caricature. Inv. nr. PDB.2012.1.CA.1990.3441. Photo: Helma Molenaars. Collectie Museum Het Valkhof, Nijmegen, Provinciaal Depot voor Bodemvondsten Gelderland. Drawing: René Reijnen. Detail: head of caricature. Terracotta plinth with feet. Inv.nr. PDB.2012.1.CA.1989.3105. Collectie Museum Het Valkhof, Nijmegen, Provinciaal Depot voor Bodemvondsten Gelderland Drawing: René Reijnen. Terracotta head woman. Inv.nr. PDB.2012.1.CA.1996.9106. Collectie Museum Het Valkhof, Nijmegen, Provinciaal Depot voor Bodemvondsten Gelderland. Drawing: René Reijnen. Bronze caduceus Mercury. Inv. nr. PDB.2012.1.CA.1987.356. Collectie Museum Het Valkhof, Nijmegen, Provinciaal Depot voor Bodemvondsten Gelderland.

Kops Plateau and eastern canabae Fig. 2.135

Bronze infant Hercules. Inv. nr. PDB.1998.2.KH.361.31. Photo: Ronny Meijers. Collectie Museum Het Valkhof, Nijmegen, Provinciaal Depot voor Bodemvondsten Gelderland. Figs. 2.136-37 Lead Venus. Inv. nr. PDB.1998.2.KH.426.2. Photos: Ronny Meijers. Collectie Museum Het Valkhof, Nijmegen, Provinciaal Depot voor Bodemvondsten Gelderland. Fig. 2.138 Lead duck. Inv. nr. PDB.1998.2.KH.270.90. Photo: author. Collectie Museum Het Valkhof, Nijmegen, Provinciaal Depot voor Bodemvondsten Gelderland. Fig. 2.139 Bronze caduceus Mercury. Inv. nr. PDB.1998.2.KH.350.63. Collectie Museum Het Valkhof, Nijmegen, Provinciaal Depot voor Bodemvondsten Gelderland. Fig. 2.140 Bronze togatus Inv. nr. Ub14.2.2.66. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Fig. 2.141 Bronze cockerel. Inv. nr. PDB.1998.2.KH.426.113. Collectie Museum Het Valkhof, Nijmegen, Provinciaal Depot voor Bodemvondsten Gelderland. Fig. 2.142 Bronze ram. Inv. nr. Ub5.14.2.68. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Fig. 2.143 Bronze swine. Inv. nr. Sw9.10. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Figs. 2.144-45 Terracotta crouching Venus. Inv. nr. PDB.1998.2.KH.430.89. Photos: Helma Molenaars. Collectie Museum Het Valkhof, Nijmegen, Provinciaal Depot voor Bodemvondsten Gelderland. Fig. 2.145a Drawing: René Reijnen. Fig. 2.146 Terracotta foot on rectangular plinth. Inv. nr. PDB.1998.2.KH.431.91. Collectie Museum Het Valkhof, Nijmegen, Provinciaal Depot voor Bodemvondsten Gelderland. Fig. 2.147 Terracotta Jupiter. Inv. nr. PDB.1998.2.KH.236.30.

223

Fig. 2.148 Fig. 2.149 Fig. 2.150

Photo: author. Collectie Museum Het Valkhof, Nijmegen, Provinciaal Depot voor Bodemvondsten Gelderland. Bronze pedestal. Inv. nr. PMK.1990.2.7. Collectie Museum Het Valkhof, Nijmegen. Bronze pedestal. Inv. nr. Ub5.13.1.50. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Bronze wheel. Inv. nr. PDB.1998.2.KH.300.3. Photo: author. Collectie Museum Het Valkhof, Nijmegen, Provinciaal Depot voor Bodemvondsten Gelderland.

Cemeteries Nijmegen Kops Plateau Fig. 2.151

Terracotta bust of boy. Inv. nr. PDB.1998.2.KH.455.62. Photo: author. Collectie Museum Het Valkhof, Nijmegen, Provinciaal Depot voor Bodemvondsten Gelderland.

Verpleeghuis Margriet Fig. 2.152 Fig. 2.152a Fig. 2.153 Fig. 2.153a

Terracotta Venus. Inv.nr. Hg4.06. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Drawing: Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Terracotta Venus. Inv.nr. Hg4.11.449 Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Drawing: Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen.

Ulpia Noviomagus Fig. 2.154 Fig. 2.154a Fig. 2.155 Fig. 2.155a

Terracotta Venus, Koekoekstraat. Inv. nr. Vr8.00021. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Drawing: René Reijnen. Terracotta mother goddess statuette. Inv. nr. Vr8.0560. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Drawing: René Reijnen.

Canisius grounds Fig. 2.156 Fig. 2.157

224

Child’s grave with horned bird. Inv. nr. Graf.Vermeulen.1932.66. Photo: Ronny Meijers. Collectie Museum Het Valkhof, Nijmegen. Detail: terracotta horned bird. Inv. nr. CC.390.

Supplements Nijmegen Hunerberg Fig. 2.158

Terracotta Minerva. Inv. nr. e1931/2.56. National Museum of Antiquities, Leiden (RMO).

Ulpia Noviomagus Fig. 2.159 Fig. 2.160 Fig. 2.160a Fig. 2.161 Fig. 2.161a

Terracotta head of woman. Maasplein. Inv.nr. Mp1.6.1.174. Photo: author. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Terracotta Venus? Weurtseweg. Inv.nr. WD7.00399KER. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Drawing: René Reijnen. Terracotta horse. Maasplein. Two fragments. Inv.nrs. Mp1.3.102.574 and 597. Drawing: René Reijnen.

Hugo de Grootstraat Fig. 2.162 Fig. 2.163

Bronze Hercules. Hugo de Grootstraat. Inv. nr. BE.I.94. Collectie Museum Het Valkhof, Nijmegen. Canisiussingel Bronze claw right leg cockerel. Inv. nr. Ca3.05. Photo: Martijn Spinder. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen.

Oosterhout-Van Boetzelaerstraat Fig. 2.164 Fig. 2.164a Fig. 2.165 Fig. 2.165a Fig. 2.166 Fig. 2.166a Fig. 2.167 Fig. 2.167a Fig. 2.168

Terracotta head of Apollo. Inv. nr. Bo5.068.035. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Drawing: René Reijnen. Terracotta Venus. Inv. nr. Bo5.164.0216KER. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Drawing: René Reijnen. Terracotta Venus. Inv.nr. Bo5.164.216. Drawing: René Reijnen. Terracotta fragment of Fortuna. Inv. nr. Bo5.108.137. Photo: Rob Mols. Bureau Leefomgevingskwaliteit/Archeologie, Nijmegen. Drawing: René Reijnen. Bronze pedestal. River Waal. Inv.nr. e1906/1.2. Photo: author. National Museum of Antiquities, Leiden (RMO).

225

Miscellaneous Fig. 2.169 Fig. 2.170 Figs. 2.171-72 Fig. 2.173 Figs. 2.174-77 Fig. 2.178 Fig. 2.179 Fig. 2.180 Fig. 2.181 Fig. 2.182 Figs. 2.183-85 Fig. 2.186 Fig. 2.187

Bronze Mars, Tiel-Passewaaij. Inv. nr. PDB.2011.1.V168.005. Photo: Ronny Meijers. Collectie Museum Het Valkhof, Nijmegen, Provinciaal Depot voor Bodemvondsten Gelderland. Bronze Mercury, Tiel-Passewaaij. Inv. nr. PDB.2011.1.V131.008. Photo: author. Collectie Museum Het Valkhof, Nijmegen, Provinciaal Depot voor Bodemvondsten Gelderland. Bronze African man, Vechten. Inv. nr. f1926/12.44. Photos: National Museum of Antiquities, Leiden (RMO). Terracotta woman, Vechten. Inv. nr. 3.1.B2. Photo: Frans de Vries. Provincie Utrecht. Terracotta gladiator, The Hague-Scheveningseweg. Inv. nr. SCH84-4659. Photos: Johan van der Helm. Gemeente Den Haag, afdeling Archeologie. Terracotta swaddled infant, Zwammerdam. Inv. nr. ZW 1972.1021. Photo: Stephan Mols. Provinciaal Archeologisch Depot Zuid-Holland. Gilded bronze staff, Heerlen. Inv. nr. 02514. Collectie Thermenmuseum Heerlen. Copyright: Thermenmuseum Heerlen. Bronze putto applique, Texel-Den Burg. Inv. nr. 5354-01. Photo: Rob van Eerden. Provinciaal Depot voor Archeologie Noord-Holland. Bronze Hercules, Velsen 1. Inv. nr. g2008/6.1974-zn29. National Museum of Antiquities, Leiden (RMO). Bronze Mars.Tiel-Badhuisplein. Inv. nr. OKT 231/B6. Collection: Flipje en Streekmuseum Tiel. Bronze Mercury, Tzum-Greate Vlearen. Inv.nr. 1983-XI-224-244. Photos: Stephan Mols. Collection: Fries Museum, Leeuwarden. Large terracotta rattle, possibly the head of a young Bacchus or Amor. Alphen a/d Rijn-De Schans. Inv. nr. RMO h1990/12.10. Photo: Stephan Mols. Archeologisch depot Zuid-Holland Alphen a/d Rijn. Bronze panther, attachment. Montfort-Montforterbroek. Inv.nr. HVR 362. Photo: author. Heemkundevereniging Roerstreek, Roerstreekmuseum St.Odiliënberg.

Figures chapter 3 Fig. 3.1 Fig. 3.2 Fig. 3.3 Fig. 3.4 Fig. 3.5

226

Keltenfürst Glauberg. Photo: Ekkehard Wolf. https://commons.wikimedia.org/wiki/File:Keltenfuerst_Glauberg_2011.jpg?uselang=nl Deity, Etang-sur-Arroux. Photo: PHGCOM. https://commons.wikimedia.org/wiki/File:God_of_Etang_sur_Arroux_possible_depiction_of Cernunnos.jpg Bronze statuette of Jupiter. https://commons.m.wikimedia.org/wiki/File:Roman_bronze_ statuette_of_Jupiter,_2nd_century_AD,_Staatliche_Antikensammlungen,_Munich_ (8958000484).jpg Mounted warrior, Britain. http://www.ancientresource.com/images/celtic/celtic-horse-man11100b.jpg Detail Arch of Galerius, Thessaloniki, Greece. Photo: Mark J. Nelson. https://commons.wikimedia.org/wiki/File:20160516_292_thessaloniki.jpg

Fig. 3.6 Fig. 3.7 Fig. 3.8 Fig. 3.9 Fig. 3.10

Figs. 3.11 Fig. 3.12 Fig. 3.13 Fig. 3.14

Wooden figure, Willemstad. Inv. nr. 015526. National Museum of Antiquities, Leiden (RMO). Wooden toy cat, Thebes, Egypt, Inv. nr. 1885, 0618.11. British Museum London © Trustees of the British Museum. Muses sarcophagus. Musée du Louvre. Inv.nr. MR 880 (Ma 475). Photo: Jastrow. Musée du Louvre, Paris. https://commons.wikimedia.org/wiki/File:Muses_sarcophagus_Louvre_MR880.jpg Detail Gundestrup cauldron. Photo: Kern8. https://commons.wikimedia.org/wiki/File:Gundestrup_Cernunnos.jpg Muri statuette group. Acc.nr. 16170/16210. Photo: Sandstein. Historisches Museum, Bern. https://upload.wikimedia.org/wikipedia/commons/3/3e/HMB_-_Muri_statuette_group_-_ Artio_2.jpg Terracotta bear, Kerkrade Maar-West. Inv.nr. 01553/001. Photo: Diane Habets-Smeets. Gemeente Kerkrade. Ram carrying saddle bags. Inv.nr. BE.I.30. Collectie Museum Het Valkhof, Nijmegen. Seated mother goddess, Çatalhöyük, 6000-5500 BC. Museum of Anatolian Civilizations Ankara. Inv. nr. B19-36. https://upload.wikimedia.org/wikipedia/commons/7/7d/Ankara_Muzeum_B19-36.jpg Cybele enthroned, c. 50 AD, Getty Museum Malibu. Inv. nr. 57.AA.19. https://en.wikipedia.org/wiki/Cybele#/media/File:Cybele_Getty_Villa_57.AA.19.jpg

Figures chapter 4 Figs. 4.1-2 Fig. 4.3

Human skull. Alphen aan den Rijn. Photo: Rien Polak. Terracotta voodoo doll, Egypt. Inv.nr. E27145b. Musée du Louvre, Paris. https://commons.m.wikimedia.org/wiki/File:Voodoo_doll_Louvre_E27145b.jpg

227

Figures

229

Figures Chapter 2

Fig. 2.1

Fig. 2.3

Fig. 2.2

Fig. 2.5

Fig. 2.4

231

Fig. 2.6

232

Fig. 2.7

Fig. 2.8

Fig. 2.9

233

Fig. 2.11

Fig. 2.10

Fig. 2.12

Fig. 2.13

Fig. 2.14

234

Fig. 2.16

Fig. 2.17

Fig. 2.15

Fig. 2.18

Fig. 2.19

Fig. 2.20

235

Fig. 2.21

Fig. 2.22

236

Fig. 2.23

Fig. 2.24

237

Fig. 2.25

Fig. 2.26

Fig. 2.27

Fig. 2.28

238

Fig. 2.29

Fig. 2.30

Fig. 2.31

Fig. 2.32

Fig. 2.33

Fig. 2.34

Fig. 2.35

Fig. 2.36

239

Fig. 2.37

Fig. 2.38

240

Fig. 2.40

Fig. 2.39

Fig. 2.41

241

Fig. 2.42

242

Fig. 2.43

Fig. 2.44

Fig. 2.45

Fig. 2.46

Fig. 2.45a

243

Fig. 2.47

Fig. 2.47a

244

Fig. 2.48

Fig. 2.49

Fig. 2.50

245

Fig. 2.51

Fig. 2.52

246

Fig. 2.53

Fig. 2.54

Fig. 2.55

Fig. 2.57

Fig. 2.56

Fig. 2.58

Fig. 2.56a

Fig. 2.59

Fig. 2.60

247

Fig. 2.60a1

Fig. 2.60a2

Fig. 2.60a3

248

Fig. 2.61

249

Fig. 2.62

Fig. 2.65

Fig. 2.66a

250

Fig. 2.64

Fig. 2.63

Fig. 2.66

Fig. 2.67

Fig. 2.67a

251

Fig. 2.68

Fig. 2.68a

Fig. 2.69

Fig. 2.70

252

Fig. 2.71

Fig. 2.72

Fig. 2.73

253

Fig. 2.74

Fig. 2.75

Fig. 2.77

254

Fig. 2.76

Fig. 2.78

Fig. 2.79

Fig. 2.80

255

Fig. 2.81

Fig. 2.81a

Fig. 2.82

Fig. 2.83

256

Fig. 2.84

Fig. 2.84a

257

Fig. 2.85

Fig. 2.86

Fig. 2.87

Fig. 2.88

258

Fig. 2.89

Fig. 2.90

Fig. 2.91

Fig. 2.92

259

Fig. 2.93

Fig. 2.94

Fig. 2.95

260

Fig. 2.96

Fig. 2.97

Fig. 2.97a

261

Fig. 2.98

Fig. 2.98a

262

Fig. 2.99

Fig. 2.99a

263

Fig. 2.100

Fig. 2.100a

264

Fig. 2.101

Fig. 2.101a

265

Fig. 2.102

Fig. 2.103

266

Fig. 2.104

Fig. 2.104a

267

Fig. 2.105

Fig. 2.105a

Fig. 2.106

Fig. 2.106a

268

Fig. 2.107

Fig. 2.107a

269

Fig. 2.108

Fig. 2.108a

Fig. 2.109

270

Fig. 2.109a

Fig. 2.110

Fig. 2.110a

Fig. 2.111

Fig. 2.111a

271

Fig. 2.112

Fig. 2.112a

Fig. 2.113

Fig. 2.114

272

Fig. 2.113a

Fig. 2.114a

273

Fig. 2.116

Fig. 2.115 Fig. 2.117

Fig. 2.117a

Fig. 2.118

274

Fig. 2.119

Fig. 2.120

Fig. 2.121

275

Fig. 2.122

Fig. 2.123

276

Fig. 2.124

Fig. 2.125

277

Fig. 2.126

Fig. 2.126a

Fig. 2.127

278

Fig. 2.127a

Fig. 2.128

Fig. 2.128a

Fig. 2.129

Fig. 2.129a

279

Fig. 2.130

Fig. 2.131

Fig. 2.131a

280

Fig. 2.132

Fig. 2.132a

Fig. 2.133

Fig. 2.133a

Fig. 2.134

Fig. 2.135

281

Fig. 2.136

Fig. 2.137

Fig. 2.138

Fig. 2.139

282

Fig. 2.140

Fig. 2.142

Fig. 2.141

Fig. 2.143

283

Fig. 2.144

Fig. 2.145

284

Fig. 2.145a

285

Fig. 2.146

Fig. 2.147

Fig. 2.148

Fig. 2.149

286

Fig. 2.150

Fig. 2.151

Fig. 2.152

Fig. 2.152a

287

Fig. 2.153

Fig. 2.153a

288

Fig. 2.154

Fig. 2.154a

289

Fig. 2.155

Fig. 2.155a

290

Fig. 2.156

Fig. 2.157

291

Fig. 2.158

Fig. 2.159

292

Fig. 2.160

Fig. 2.160a

293

Fig. 2.161

Fig. 2.161a

294

Fig. 2.163

Fig. 2.162

Fig. 2.164

Fig. 2.164a

295

Fig. 2.165

Fig. 2.166

296

Fig. 2.165a

Fig. 2.166a

Fig. 2.167

Fig. 2.167a

Fig. 2.168

297

Fig. 2.170 Fig. 2.169

Fig. 2.171

Fig. 2.174

298

Fig. 2.172

Fig. 2.173

Fig. 2.175

Fig. 2.176

Fig. 2.177

Fig. 2.178

Fig. 2.180

Fig. 2.179

Fig. 2.181

Fig. 2.182

299

Fig. 2.183

Fig. 2.184

Fig. 2.186

Fig. 2.187

300

Fig. 2.185

Figures Chapter 3

Fig. 3.2 Fig. 3.1

Fig. 3.3

Fig. 3.4 Fig. 3.5

301

Fig. 3.7 Fig. 3.6

Fig. 3.8

Fig. 3.9

302

Fig. 3.10

Fig. 3.11

303

Fig. 3.12

Fig. 3.13

Fig. 3.14

304

Figures Chapter 4

Fig. 4.1

Fig. 4.2

305

Fig. 4.3

306

Tables Table 1 The three areas Subject Deities

Material

South of the Rhine

metal

terracotta limestone amber

Amor

3

2

Apollo

4

Attis

2

Bacchus

7

5

chalk

total 6

2

1

2 10

7 5

4

2

Fortuna

4

25

11

Harpocrates

2

5 3

1 1 1

5

13

1

4

1

1

2

4

Juno

1

7

Jupiter

5

1

Jupiter-Ammon

1

Luna

1

1

1

Mars

17

1

7

9

Mercurius

44

12

24

Minerva

13

17

11

Neptunus

1

4

1

2

1

Priapus

1

1

5

Victoria

4

1

2

7

3

3

6

1

Parcae

1

2

2

4

Venus

1 1

Isis-Fortuna

Vagdavercustis

4 29

1

3

Sucellus

2 15

7

Isis

1

9

3

9

1

1

8

4

Sol

5

3

1

Serapis

1

5

1

12

Total

total

2

Cybele

Hercules

total

4

Diana Epona

North of the Rhine Limes zone

1 1

2

2

18

18

14

56

7

13

31

1

1

1

2 1

1

1 1

1

1

61

34

4

1

3

1 1

1

1 28

66 4

Human figures boy between horses

2

2

bust of boy

23

13

bust of child

1

bust of girl

4

bust of human bust or head of woman caricature of man or dwarf

1

desultor

1

embracing couple

2 1

2

2

2

15

10

9

23

1

1

2

4

5

15

5

5

2

4 1 4

1 1

3

4

307

genius

3

gladiator

1

human or fantasy figure

1

3

1

1

1

lar

1

mater

3

matrona

1

African man

1

offerant

1

satyr

1

silenus

1

thorn-puller togatus

1

1 1 41

27

1 3

14

1

1 1

1

1

1

1 1

1

1

1 1

1

44 1

1

1

voodoo-doll

1

1

1 1

warrior enthroned

1

1

1

horse + missing warrior

1

1

1

warrior mounted

1

1

1

warrior standing

1

water nymph/personific. river

1

1 4

3

1 2

5

Animals bear billy goat

2

bird boar

1

1

3

1

3

1

6

3

3

2

1

2

6 4

bovine

1

bovine or sheep

1

4 2

6 6

3

6

1

1

1

1

bull

1

3

3

cockerel

5

13

10

dog

5

5

5

dolphin

1

1

1

dove or other bird duck

1

eagle

1

2

1

4

6

18

4

3

1

1

1

1 1

4 2 1

goat/sheep

5

3

2

5

hen

3

1

2

3

4

4

6

5

5

13

6

4

2

6

12

10

4

14

horned bird horse

7

ithyphallic dog lion

2

panther

1

ram

1

stag/roe

2

triple-horned bull

2

Subtotal

4 3

1 1

1

1

1

2

197

325

18

126

2

4

1

270

2 2

1

1

2

83

176

529

Unidentifiable fragments Deities or human figures Human figures Animals Subtotal Total

308

6

144

5

5

13

13

18

144

215

469

2

4

1

156

6

426

89

162 176

691

Table 2 North of the Rhine Subject Deities

Material metal

terracotta

limestone

Total amber

chalk

Amor Apollo

2

2

Attis Bacchus Cybele Diana Epona Fortuna

2

1

3

Harpocrates Hercules

1

1

Isis

2

2

Isis-Fortuna

1

1

Juno

1

1

Jupiter

3

3

Jupiter-Ammon Luna Mars Mercurius Minerva

9

9

18

18

5

2

7

Venus

2

2

4

Victoria

3

Neptunus Parcae Serapis Sol Sucellus Vagdavercustis 3

Human figures boy between horses bust of boy

1

1

1

1

bust of child bust of girl bust of human bust or head of woman caricature of man or dwarf desultor embracing couple genius

1

1

gladiator human or fantasy figure

1

1

2

3

lar mater

1

matrona

1

1

offerant

1

1

satyr or young Pan

1

1

African man

309

Silenus thorn-puller togatus voodoo-doll warrior enthroned

1

1

horse and (missing) warrior

1

1

warrior mounted

1

1

warrior standing

1

1

water nymph/personification river Animals bear billy goat

1 1

bird boar

1 1

3

3

2

2

bovine bovine or sheep

1

1

1

2

bull cockerel

1

dog dolphin dove or other bird duck eagle

1

1

3

3

goat/sheep hen horned bird horse ithyphallic dog lion panther ram stag/roe triple-horned bull Subtotal

1

1

67

15

3

3

1

83

Unidentifiable fragments Deities or human figures

6

Animals Subtotal Total

310

3

3

70

18

6 1

89

Table 3 Rural settlements south of the Rhine Subject Deities Amor

Material metal

terracotta

2

Apollo 2

Bacchus

1

Cybele

1

Diana

1

Total amber

chalk 2

1

Attis

limestone

1 2 1

2

3 1

Epona

2

2

Fortuna

4

4

Harpocrates

1

1

Hercules

5

5

Isis

1

1

Isis-Fortuna

1

Juno

1 1

1

Jupiter Jupiter-Ammon Luna Mars

6

Mercurius

9

1

10

6

Minerva

4

2

6

1

1

1

1

Neptunus Parcae Serapis Sol Sucellus Vagdavercustis

1

Venus Victoria

1 6

1

6 1

Human figures boy between horses bust of boy

7

7

1

1

2

2

bust of child bust of girl bust of human bust or head of woman caricature of man or dwarf desultor embracing couple genius

1

1

gladiator human or fantasy figure lar mater

1

8

9

matrona African man

1

1

offerant satyr

311

silenus

1

1

thorn-puller togatus voodoo-doll warrior enthroned horse + missing warrior warrior mounted warrior standing water nymph/personification river

1

1

2

2

2

1

1

Animals bear billy goat bird boar

1

1

bovine bovine or sheep bull

1

cockerel

3

dog

4

dolphin

1

dove or other bird

1 3

6 4 1

1

1

horse

1

1

ithyphallic dog

1

1

lion

4

4

51

53

104

1

26

27

1

1

duck eagle goat/sheep hen horned bird

panther ram stag/roe triple-horned bull Subtotal Unidentifiable fragments Deities or human figures Animals Subtotal Total

312

1

27

28

52

80

132

Table 4 Urban settlements south of the Rhine Subject Deities

Material metal

terracotta

1

4

Total

limestone

amber

chalk

Amor Apollo

5

Attis Bacchus

2

2

Cybele

3

3

Diana

2

2

4

4

Epona Fortuna Harpocrates Hercules

1

1

Isis Isis-Fortuna Juno

2

2

Jupiter Jupiter-Ammon

1

1

Luna Mars Mercurius

7

Minerva

1

1

5

12

3

1

4

Neptunus Parcae Priapus

1

1

16

17

4

4

bust of human

2

2

bust or head of woman

7

7

13

13

Serapis Sol Sucellus Vagdavercustis Venus

1

Victoria Human figures boy between horses bust of boy bust of child bust of girl

caricature of man or dwarf embracing couple genius gladiator human or fantasy figure lar mater

1

1

African man offerant satyr silenus

313

thorn-puller

1

1

1

1

1

1

togatus voodoo-doll warrior enthroned horse + missing warrior warrior mounted warrior standing water nymph/personification river Animals bear billy goat

2

bird

2 1

1

2

2

2

2

boar bovine bovine or sheep bull cockerel dog

1

1

dolphin dove or other bird

1

1

2

2

4

4

5

5

duck eagle goat/sheep hen horned bird horse ithyphallic dog lion panther ram stag/roe

1

1

triple-horned bull Subtotal

17

87

2

106

5

41

1

47

Unidentifiable fragments Deities or human figures Human figures

3

3

Animals

9

9

Subtotal Total

314

5

53

1

59

22

140

3

165

Table 5 Cemeteries south of the Rhine Subject Deities

Material metal

terracotta

Amor

limestone

Total amber 2

chalk 2

Apollo Attis Bacchus

1

1

Cybele

1

1

Diane

2

2

2

2

1

1

1

1

1

1

8

8

boy between horses

2

2

bust of boy

1

1

1

1

5

5

Epona Fortuna Harpocrates Hercules Isis Isis-Fortuna Juno Jupiter Jupiter-Ammon Luna Mars Mercurius Minerva Neptunus Parcae Serapis Sol Sucellus Vagdavercustis Venus Victoria Human figures

bust of child bust of girl bust of human bust or head of woman caricature of man or dwarf desultor embracing couple genius gladiator human or fantasy figure lar mater matrona African man offerant satyr

315

silenus thorn-puller togatus voodoo-doll warrior enthroned horse + missing warrior warrior mounted warrior standing water nymph/personification river Animals bear

1

1

billy goat bird

1

1

2

2

boar bovine bovine or sheep bull cockerel dog dolphin dove or other bird

1

1

duck

1

1

eagle goat/sheep hen

1

1

horned bird

4

4

horse ithyphallic dog

3

3

lion

1

1

1

1

panther ram stag/roe triple-horned bull Subtotal

40

4

44

Unidentifiable fragments Deities or human figures

2

2

1

1

Human figures Animals Subtotal Total

316

3 43

3 4

47

Table 6 Dutch limes zone Subject Deities

Material metal

terracotta

limestone

Total amber

chalk

Amor

1

1

Apollo

1

1

Attis Bacchus

3

Cybele

3 3

3

Diana Epona

2

2

14

15

Fortuna

1

Harpocrates

1

1

Hercules

5

5

Isis Isis-Fortuna

1 2

Juno Jupiter

2

1 2

2

2

1

3

Jupiter-Ammon Luna

1

Mars

2

Mercurius

8

6

14

Minerva

4

9

13

Neptunus

1

Parcae

1 2

1 1

1

Serapis

1

1

Sol

1

1

Sucellus Vagdavercustis Venus

1

27

28

bust of boy

9

9

bust of child

1

1

bust of girl

2

2

5

5

4

5

3

3

Victoria Human figures boy between horses

bust of human bust or head of woman caricature of man or dwarf

1

desultor embracing couple genius

1

gladiator

1 1

1

13

14

lar mater

1

matrona African man offerant satyr silenus

317

thorn-puller togatus

1

1

voodoo-doll warrior enthroned horse + missing warrior warrior mounted warrior standing water nymph/personification river

2

2

Animals bear billy goat

2

2

bird boar

1

bovine

2

3

1

1

1

1

5

6

1

1

bovine or sheep bull cockerel

1

dog dolphin dove or other bird duck

1

1

eagle goat/sheep

2

2

hen

2

2

3

5

2

2

2

4

horned bird horse

2

ithyphallic dog lion

2

panther ram

1

1

1

1

stag/roe triple-horned bull Subtotal

50

127

177

4

53

57

Unidentifiable fragments Deities or human figures Animals Subtotal Total

318

2

2

4

55

59

54

182

236