Water in Ancient Mediterranean Households (Global Perspectives on Ancient Mediterranean Archaeology) [1 ed.] 9781032213972, 9781032214009, 9781003268222, 1032213973

Citation preview

Water in Ancient Mediterranean Households

This book provides the first detailed study of the water supply of households in antiquity. Chapters explore settings from Classical Greece to the Late Roman Empire across a wide variety of environments, from dry deserts and moderate Mediterranean zones to wet and temperate climates further north. The different case studies presented in each chapter are united by three intimately interconnected aspects. The first, rainwater harvesting in cisterns, provides detailed techno-hydraulic investigations of the household water supply systems. The second aspect, households and water at the margins, stresses how domestic water supply systems were successfully adapted to unusually harsh environmental conditions. The third, other waters for houses, focuses on other types of water supply systems (rivers, water-bearers, stepped pools, wells) and their life biographies. As shown by the different chapters, a careful study of a household’s water supply is a rich source of evidence for understanding everyday decisions, anxieties, and changes in life. They also build towards a greater understanding of the social inequalities that are at play in the ancient Mediterranean and beyond, providing a wealth of new research to greatly augment our understanding of water as a resource in the ancient Mediterranean. Providing a new and important perspective on a central part of everyday life in the ancient world, this book is aimed at archaeologists and historians of the ancient Mediterranean, notably the Greek and Roman worlds, especially those with an interest in ancient households and water culture. Rick Bonnie is a University Lecturer in Museology at the University of Helsinki, Finland, where he researches and teaches on museum and heritage ethics, object biographies, decolonisation and provenance issues, museum collection histories, and sensory archaeology. He is the author of Being Jewish in Galilee, 100-200 CE: An Archaeological Study (2019). Among other works, Rick recently led a project that studied the impact of past climatic changes on the rise and fall of Jewish ritual purification baths in Hasmonean-Roman Judaea through hydrological modelling and contextual archaeological analysis. Patrik Klingborg is an Associate Senior Lecturer at the Department of Archaeology and Ancient History, Uppsala University, Sweden, primarily studying water usage in the ancient Greek world. Within the framework of this, he has focused on non-monumental water sources such as cisterns and wells, as well as how water was used within ancient Greek religion. He is also part of the board of the Frontinus-Gesellschaft and participates in fieldwork by the Swedish Institute at Athens.

Global Perspectives on Ancient Mediterranean Archaeology Series editor: Jeremy Armstrong and Gijs Tol

The series’ remit embraces a broad span of time from the Mediterranean Bronze Age through the Byzantine period (c. 3200 BCE – 1453 CE). Although nominally focused on the Mediterranean Sea and areas which immediately border it, the series also welcomes studies on areas slightly further afield which are linked to the Sea by cultural, social, economic, religious, or political connections, and where the Mediterranean zone is directly relevant. A guiding principle of the series is the inclusive appreciation of all available material from a particular area, time, and culture, even if the primary emphasis is on the archaeological aspects. Finally, as suggested by its name, the series is particularly interested in publishing works which adopt a broad comparative and cross-cultural approach, as well as those which bring together concepts and themes more common in the study of archaeology from elsewhere in the world (most notably the Pacific and Australasia) with those from the Old World. Birds in Roman Life and Myth Ashleigh Green

For more information about this series, please visit: https://www.routledge.com/Global-Perspectiveson-Ancient-Mediterranean-Archaeology/book-series/GPAMA

Water in Ancient Mediterranean Households

Edited by Rick Bonnie and Patrik Klingborg

First published 2024 by Routledge 4 Park Square, Milton Park, Abingdon, Oxon OX14 4RN and by Routledge 605 Third Avenue, New York, NY 10158 Routledge is an imprint of the Taylor & Francis Group, an informa business © 2024 selection and editorial matter, Rick Bonnie and Patrik Klingborg; individual chapters, the contributors The right of Rick Bonnie and Patrik Klingborg to be identified as the authors of the editorial material, and of the authors for their individual chapters, has been asserted in accordance with sections 77 and 78 of the Copyright, Designs and Patents Act 1988. All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library ISBN: 978-1-032-21397-2 (hbk) ISBN: 978-1-032-21400-9 (pbk) ISBN: 978-1-003-26822-2 (ebk) DOI: 10.4324/9781003268222 Typeset in Times New Roman by codeMantra

Contents

List of Illustrations List of Contributors 1

Water in Ancient Mediterranean Households

vii xiii 1

RICK BONNIE AND PATRIK KLINGBORG

2

Household Water, Environment, and Economy in Ancient Piraeus

12

JANE MILLAR TULLY

3

Social Stratification and Water Sharing on Late-Hellenistic Delos

33

PATRIK KLINGBORG

4

Surveying Notion’s Residential Water Supply: Cistern Use during Hellenistic–Roman Times

53

ANGELA COMMITO

5

Breaking Out from Imagined Household Uniformity: Diverse Rainwater Harvesting Solutions in Republican-Imperial Cosa

74

ANN GLENNIE

6

Rainwater Collection Strategies in Pompeian Houses

96

GEMMA JANSEN

7

Posthumanism, Social Justice, and Pollution in the Waters of Roman Volubilis

118

MARK A. LOCICERO

8

Reusing Stepped Pools in Roman Palestinian Households RICK BONNIE

133

vi Contents   9 The Significance of Household Cisterns at Roman Dura-Europos

152

J. A. BAIRD

10 Water as Social Inequality in Late Roman Britain

172

JAMES GERRARD

Index

197

Illustrations

Figures 1.1 Map with all sites discussed in this book. Map by Patrik Klingborg. Base map: Stamen Design, under CC BY 3.0. Data by OpenStreetMap, under ODbL 2.1 Map of modern Piraeus with features of the ancient water supply. Used with permission of G. Peppas, copyright Ephorate of Antiquities of Piraeus and Islands, Hellenic Ministry of Culture and Sports 2.2 Underground water supply structures excavated for the Dimotiko Theatro Station on a reconstructed urban grid, with features mentioned in the text labeled. Cross-section after Koutis & Bendermacher-Gerousis 2015, slide 5. Copyright Ephorate of Antiquities of Piraeus and Islands, Hellenic Ministry of Culture and Sports 2.3 Interior of Well 28. Copyright Ephorate of Antiquities of Piraeus and Islands, Hellenic Ministry of Culture and Sports 2.4 Flask-shaped Cistern 1, with Well 1β’ below. Copyright Ephorate of Antiquities of Piraeus and Islands, Hellenic Ministry of Culture and Sports 2.5 Cisterns (1–5, numbered from west to east) and well (1) in Pavlos Bakoyiannis Square. Copyright Ephorate of Antiquities of Piraeus and Islands, Hellenic Ministry of Culture and Sports 2.6 Cisterns 6 (shaft, left) and 7 (flask-shaped, right). Copyright Ephorate of Antiquities of Piraeus and Islands, Hellenic Ministry of Culture and Sports 2.7 From left to right, Well 25, Tunnel 15, Well 13, Well 2, Tunnel 13, and Cistern 3. Copyright Ephorate of Antiquities of Piraeus and Islands, Hellenic Ministry of Culture and Sports 3.1 Delos with neighbouring islands and their location in Greece. Map by Author 3.2 Number of water sources per house and shop. Map by Author

5

13

17 18 20 22 23 24 34 39

viii Illustrations 3.3 Number of water sources based on size, from the largest to the smallest house. Map by Author 3.4 Types of water sources in the houses. Map by Author 3.5 Water sources within 50 meters of Shop RdT 33. Map by Author 3.6 Water sources within 50 meters of Shop RdT 27. Map by Author 3.7 Water sources within 50 meters of House VI C. Map by Author 4.1 Map of the region around Notion. Notion Archaeological Survey 4.2 City plan of Notion, showing the locations of documented cisterns. Notion Archaeological Survey 4.3 Compiled cross-sections of nine cisterns produced from laser scans. Notion Archaeological Survey 4.4 Internal view of cistern 19, showing pick marks and excavation in concentric rings. Notion Archaeological Survey 4.5 Placement and bedding of terracotta pipe inside cistern 19. Notion Archaeological Survey 4.6 Masonry opening of cistern 19. Notion Archaeological Survey 4.7 Roofing system of cistern 9 from below, with one of two terracotta pipes visible. Notion Archaeological Survey 4.8 View inside cistern 9, showing pick marks on the interbedded schist and marble bedrock surfaces and hydraulic mortar on the upper portion of shaft. Notion Archaeological Survey 4.9 View of cistern 14, looking North. Notion Archaeological Survey 4.10 Interior view of cistern 7. Notion Archaeological Survey 4.11 Roofing system of cistern 7 from below. Notion Archaeological Survey 4.12 View of partial pithos reused to articulate the opening of cistern 1. Notion Archaeological Survey 4.13 View inside cistern 12, showing hydraulic mortar on uneven surfaces of schist bedrock. Notion Archaeological Survey 4.14 Conglomerate junction block for a terracotta pipeline. Notion Archaeological Survey 4.15 Marble block, possible part of a pressurized stone pipeline. Notion Archaeological Survey 5.1 State plan of Cosa with Atrium Buildings and East and West Block Houses. Cosa Excavations 2018, adapted 5.2 Reconstructed plan of Republican Atrium Building I and North Corner Plot. Brown et al. 1993, Figs. 20 and 55, combined and adapted 5.3 Reconstructed plan of Republican and Augustan Phases of Atrium Building V/House of Diana. Fentress 2003, Figs. 5 and 15, adapted 5.4 State plan of the East and West Block houses. Bruno & Scott 1993, Fig. 3, adapted 5.5 Reconstructed plan of Republican Phase 1 West and East Block houses. Bruno & Scott 1993, Figs. 5 and 10, adapted

41 42 47 48 49 54 56 57 58 59 59 60 61 63 64 65 66 67 70 70 76 78 80 83 84

Illustrations  ix 5.6 Reconstructed plan of Republican Phase 2 West and East Block houses. Bruno & Scott 1993, Figs. 10 and 19, adapted 5.7 Reconstructed plan of Republican Phase 3 East and West Block houses. Bruno & Scott 1993, Figs. 22 and 32, adapted 5.8 Reconstructed plan of Imperial East Block house/House of the Birds. Bruno & Scott 1993, Fig. 43, adapted 5.9 Faceted plot showing the amount of water able to be collected each month (bars) and the surplus, as restricted by max cistern volume (lines), for 5, 10, 15, and 20 person households, respectively, over eight cycles of rainfall collected in the cisterns from: Atrium Building I (max. 60.69 m3); Lot 4 and 5 house (max. 24.04 m3); Lot 5 garden plot house (max. 6.82 m3); Southwest House (max. 36 m3); and House of the Skeleton (max. vol. 58.8 m3) 6.1 Schematic presentation of the rainwater collection system in the atrium. Drawing by author 6.2 Impluvium with a small drain to cistern in the middle and a large drain to street at the right, Casa del Menandro I 10, 4.14.15. Photo by author. Printed with permission of the Ministero della Cultura, Parco archeologico di Pompei 6.3 Lions as terracotta waterspouts with a dog (or lion?) on an acanthus leaf as corner spout, Casa del fabbro I 10, 7. Photo by author. Printed with permission of the Ministero della Cultura, Parco archeologico di Pompei 6.4 Cistern turned into cellar, house VIII 6, 9. Photo by author. Printed with permission of the Ministero della Cultura, Parco archeologico di Pompei 6.5 Stone slab with hole over cistern mouth and marble puteal, house VIII 6, 1. Photo by author. Printed with permission of the Ministero della Cultura, Parco archeologico di Pompei 6.6 Marble cistern head, puteal, with signs of wear, house VII 15, 2. Photo by author. Printed with permission of the Ministero della Cultura, Parco archeologico di Pompei 6.7 Holes for nails at street drain to fix a small grating. Photo by author. Printed with permission of the Ministero della Cultura, Parco archeologico di Pompei 6.8 Portico around (peristyle) garden with rain gutter at ground level and cistern mouth in portico, house I 14, 2. Photo by author. Printed with permission of the Ministero della Cultura, Parco archeologico di Pompei 6.9 Gutter in garden with access to cistern: a settling basin and in front of the opening to the cistern a perforated lead sheet attached with iron nails, house I 9, 11.12. Photo by author. Printed with permission of the Ministero della Cultura, Parco archeologico di Pompei

85 86 88

90 99

100

100 102 102 103 104

104

105

x Illustrations 6.10 Lightwell with terracotta downpipe, collection and distribution basin and cistern mouth, house I 7, 18. Photo by author. Printed with permission of the Ministero della Cultura, Parco archeologico di Pompei 6.11 Plan of Pompeii with insulae of investigation indicated, composition B. de Fraiture 6.12 Small compluvium and small impluvium underneath, house I 9, 16.17. Photo by author. Printed with permission of the Ministero della Cultura, Parco archeologico di Pompei 6.13 Terracotta puteal, raised washbasin and amphora in the portico of Casa degli amanti (I 10, 10.11); three bronze vessels, one of which was the standard vessel for drawing water from cisterns were found here, too, but are not present on this photograph. After Jashemski 1993, 50 6.14 Collection and distribution basin in corner of courtyard of Casa del Menandro (I 10, 4.14.15): (a) overview with cistern head and basin in corner. Photo by author. Printed with permission of the Ministero della Cultura, Parco archeologico di Pompei; (b) detail of basin with drain to street at bottom and drain to cistern at top. After Ling 1997, 120, plate 66 7.1 Photograph of aqueduct in Queretaro (Mexico), designed to resemble ancient Roman aqueducts in Spain 7.2 Location of Volubilis. After Bigi 2018, 167, Fig. 2, reproduced with permission 7.3 Plan of Volubilis. Drawing by author 7.4 North-East Quarter of Volubilis, with aqueduct supply lines indicated by the thick solid line. Drawing by author 7.5 Location of olive oil presses (marked by the numbers) in Volubilis, with the vast majority in the southern part of the city. After Bigi 2018, 168, Fig. 3, reproduced with permission 8.1 Sepphoris SP4, an example of stepped pool structure. Photo by author 8.2 Map of the region with sites mentioned in the text. Map by author 8.3 Jerusalem, stepped pool ‘miqweh 65’: (a) looking northwest, showing the ashlar collapse inside the pool; (b) Concentration of pottery bowls trapped under ashlar collapse, looking north. Reich 2000, Photos 2.77–78, courtesy of the Israel Exploration Society 8.4 Jericho, stepped pool area ‘A(A)209-A(A)243’: (a) overview of area, looking north; (b) plates exposed in the connecting pool A(A)243, looking east. Netzer 2001, Ills 55 and 61, courtesy of the Israel Exploration Society

106 108 110

112

113 119 120 125 126 128 134 136

138

139

Illustrations  xi 8.5 Gamla, stepped pool ‘1293’, looking east, with broken storage jars at the bottom. Photo Danny Syon/Gamla Excavations 8.6 Sepphoris, stepped pool SP1 / L85.3228, looking north, with the later installation mortared on its floor. Gordon 2018a, Photo 4.25, courtesy of Eric and Carol Meyers/Duke Sepphoris Excavation Project 9.1 Plan of the site indicating excavated area and main monuments, with approximate position of river marked to the east. Site plan adapted by author from Detweiler, YUAG, Dura-Europos Archive 9.2 Aerial photograph of Dura-Europos in 1932 from south showing position relative to the Euphrates, before building of modern dams upstream. YUAG, Dura-Europos Archive DS-2 9.3 Idealised Durene house, drawn in the 1930s by architect Henry Pearson, which brings together features found in a number of houses at Dura. Note square ‘latrine’ in the centre of the courtyard. YUAG, Dura-Europos Archive 9.4 Cistern opening (left) and partially overgrown cistern cover (right) in courtyard of house D5-E, excavated in 1930s. Photo by author, 2006 9.5 Plan of city block D5 indicating cisterns and relationships of house units of final period; by then, only four house units were present, as houses D5-A and D5-E were internally connected, as were houses D5-F and D5-F1, and houses D5-C/D5-D. Plan by author, incorporating field and archival data (after Pearson 1932 and Pillet 1931). The large collapsed cistern was that of room F1 10.1 Drapers’ Gardens: late second to third century showing timber buildings to the east of the road supplied with piped water. [1215] is the pipe illustrated in Fig. 10.2. The building to the west of the road was served by a well that existed within a private, fenced yard not shown on this plan. Courtesy of Pre-Construct Archaeology Ltd 10.2 A bored wooden pipe with lead fitting from Drapers’ Gardens. It is shown as [1215] on Figure 10.1. Courtesy of Pre-Construct Archaeology Ltd 10.3 Drapers’ Gardens in the late fourth and early fifth centuries. The open space east of the road is now occupied by two wells. The most southerly contained an important hoard of bronze, pewter, and iron vessels. Courtesy of Pre-Construct Archaeology Ltd. 10.4 The westernmost well at Drapers’ Gardens (see Fig. 10.1). Note the white clay used to backfill between the edge of the construction pit and the timber planking. Courtesy of Pre-Construct Archaeology Ltd. 10.5 The diurnal cycle of water use. After White 1977, Fig. 6.6

140

143 153 154

157 161

165

174 175

176

177 178

xii Illustrations 10.6 Plan showing the relationship between the unusually wide and deep well at Rudston and the villa’s bath building. After Stead 1980, Fig. 3 10.7 Simplified sketch section of the well at Rudston showing the various deposits mentioned in the text. After Stead 1980, Fig. 16 10.8 The late Prof. Philip Rahtz, wearing a British army ‘tin helmet’, excavating the Roman well at Pagans Hill, Somerset in the early 1950s. This image underlines not only the danger inherent in excavating wells, but also the proximity that the modern archaeologist’s experience has to that of the ancient well digger. University of York and the Estate of Prof. Philip Rahtz: CC BY-NC-SA 10.9 A sketch section of the well at Dalton Parlours. After Wrathmell & Nicholson 1990, Fig. 158

182 183

184 189

Tables 3.1 Number of sources based on size, from the largest to the smallest house 5.1 Detailed information available from publication about the Atrium Buildings’ rainwater harvesting features 5.2 Detailed information available from publication about the East and West Block houses’ rainwater harvesting features 6.1 Insulae investigated and their courtyards 9.1 Cisterns in houses at Dura-Europos. Houses in bold were internally connected with adjacent houses in the final period. List does not include cisterns in buildings which were not in use in the final period (e.g., those which might have belonged to houses which were sealed below later structures). For a full list of houses at the site, including internal connections between houses and bibliography for each, see appendix in Baird 2014

40 93 94 108

159

Contributors

J. A. Baird is Professor of Archaeology in the School of Historical Studies, Birkbeck College, University of London, UK. Rick Bonnie is a University Lecturer in Museology at the University of Helsinki, Finland, where he researches and teaches on museum and heritage ethics, object biographies, decolonisation and provenance issues, museum collection histories, and sensory archaeology. Angela Commito is a Senior Lecturer in Classical Archaeology at Union College, New York, USA. She studies urban biography, environmental archaeology, and social collapse and resilience. James Gerrard is Professor of Roman Archaeology at Newcastle University, UK, where he researches and teaches the archaeology of Roman Britain and the Western Empire. Ann Glennie is a Visiting Assistant Professor at the College of the Holy Cross, Worcester, Massachusetts, USA. Gemma Jansen is an archaeologist and independent researcher. She is specialised in Roman water systems. She lives in Maastricht, The Netherlands. Patrik Klingborg is an Associate Senior Lecturer at the Department of Archaeology and Ancient History, Uppsala University, Sweden, primarily studying water usage in the ancient Greek world. Mark Locicero is an Honorary Research Member at the Department of Ancient Mediterranean and Near Eastern Studies, University of British Columbia, Vancouver, Canada. His research uses water as a lens to investigate the relationship between natural resources and cultural values. Jane Millar Tully is a classical archaeologist and recipient of the Loeb Classical Library Foundation Postdoctoral Fellowship for 2023–2025. She is currently based at the University of Tennessee, USA.

1

Water in Ancient Mediterranean Households Rick Bonnie and Patrik Klingborg

Introduction Small-scale water supply installations such as cisterns and wells are some of the most consistently found archaeological features in household contexts across the Mediterranean and beyond. Yet, despite their vast numbers, they remain one of the least understood and investigated features of any household. It almost seems as if their functioning and meaning is too ordinary and obvious to be worthy of a thorough research interest. Instead, studies of the ancient water supply have focused primarily on how towns, neighbourhoods, and households in the Greek, Hellenistic, and Roman worlds received drinking water through such monumental structures as aqueducts and fountains.1 This volume aims to contribute to addressing this situation by focusing solely on small-scale water supply installations like cisterns and wells within households across the ancient Mediterranean and beyond. Modern scholarship’s concern with the more monumental water structures has deep roots in history. As societies in Western Europe emerged from the Middle Ages, they increasingly turned towards antiquity as an idealized past, worthy of praise and imitation. Among the many aspects that were admired, monumental water supply structures, notably aqueducts, came to be seen as markers of civilization.2 This development was strengthened during the eighteenth and nineteenth centuries—the period of Western modernity—when European nation-states and Imperialist regimes developed intricate links with ‘their’ pasts. In their development of cities and states and a consolidation of power by middle and upper classes,

1 See, e.g., Glaser 2000; Hodge 2002; Longfellow 2011; Richard 2012; Grewe 2019. Note that in The History of the Decline and Fall of the Roman Empire, Gibbon (1776–1789, 2:11) failed to mention any type of water supply infrastructure other than monumental aqueducts. Similarly, as recently as twenty-two years ago cisterns and wells received almost no attention in Wikander’s otherwise excellent handbook of ancient water technology (2000). 2 It should be noted that this renewed sense of admiration did not mean that the workings of these monumental water supply structures were being rediscovered. Coates-Stephens 2003 has pointed out how the knowledge of the hydraulic workings of Rome’s water supply continued through the Middle Ages. For a brief history of Roman fountains during the Renaissance and early modernity, see Richard 2012, 2–4. DOI: 10.4324/9781003268222-1

2  Rick Bonnie and Patrik Klingborg control of the sensorial regimes was of utmost importance.3 Thus, in a very real sense a proper water supply and drainage system became a prerequisite and marker of an ‘advanced’ society.4 Importantly, this split between the monumental and the unspectacular in modern scholarship is also reflected in a divide in antiquity in terms of equality: the haves and have nots, those who benefitted from the construction of monumental fountains and had the means to get access to piped aqueduct water versus those that had to dig wells and cisterns only to later engage in the heavy labour of drawing water. Much has been written about the former, little about the latter. Using the divides between the monumental and the non-impressive, the haves and the have nots as starting points, this book aims to fill this important knowledge gap by providing case studies of the water supply of households in antiquity from across the Mediterranean and beyond based on cisterns, wells, and other installations. As such, our hope is that this book provides new and important perspectives on the most common water supply system of households, and through this on everyday life in the ancient world. The Study of Small-Scale Water Supply Installations To explore the use of small-scale water supply installations, it is first necessary to clarify what particular features we understand to fall under this term and what these features denote. In the past, some scholars have used particular terminology for a wide range of features with different forms and function, which can lead to confusion.5 In this volume, contributions primarily focus on cisterns, wells, reservoirs, and stepped pools. We define a cistern as a statically situated waterproof container constructed above or below ground to store water. A cistern held water that had been received from an external source—usually rainwater—and was not intended to receive a constant inflow or facilitate constant outflow (see, e.g., Figures 2.4, 4.3). Stepped pools, a type of small-scale water installation particularly found in houses across the Southern Levant,6 share the characteristics of a cistern but most frequently have a flight of steps across the full width of the pool area (see Figure 8.1). Those installations with a constant in- and outflow are referred to as reservoirs, and they tend to be connected to long-distance conduits including aqueducts. A well, however, is an artificial water source—usually a shaft—dug into the ground until it reaches the water table, from which it is fed (see, e.g., Figures 2.7, 10.7). This means that water is constantly ‘produced’, while the water level in the well may rise or recede over time.

3 Stallybrass & White 1986, 125–148; Urry 2000, 94–103; Hamilakis 2013, 16–24. 4 See further Locicero, this volume. 5 E.g., Broneer 1954. See also the discussion in Klingborg 2023. For a discussion on cisterns and wells, see Wilson 2008, 285–290. On cisterns, see Klingborg 2017. See also, on wells, Hodge 2000. 6 See Bonnie, this volume.

Water in Ancient Mediterranean Households  3 In practice, the presence of small-scale water installations like cisterns and wells, as defined here, in ancient Mediterranean societies has been known, and often acknowledged, by researchers since the inception of modern scholarship.7 From the late nineteenth century onwards, these installations were primarily, if not exclusively, excavated because it was realized that they often contained rich, relatively well preserved, and frequently sealed material deposits. At first impressive finds were sought, such as sculpture or metal objects. Later there was an increased focus on obtaining sealed deposits of pottery in order to establish more precisely dateable typologies. This emphasis on the use of cistern and well deposits for ceramic typologies started during the early twentieth century, but is still ongoing in more recent decades, as regional ceramic wares are being recognized and their typologies refined.8 The sealed nature of many ancient cisterns and wells also makes them useful for sampling soils for palaeobotanical environmental reconstructions.9 Only in the 1970s, did cisterns and wells gradually start to be taken seriously for their own merits—that is, as a central part of the water supply for many communities living around the Mediterranean in antiquity. In particular, John Camp showed the potential of studying these installations by investigating the water supply of ancient Athens. Doing so he managed to use the large number of excavated cisterns and wells at the Athenian Agora to reveal information about changing groundwater levels, and, as a consequence of this, he argued, climatic changes in the past.10 Following this, in the early 1990s Werner Brinker attempted to describe cisterns in the Eastern Mediterranean, from their first appearance to the beginning of the Islamic era. Through this he managed to put his, primarily Greek, material into perspective, showing that rainwater harvesting was a pan-Mediterranean phenomenon practised over millennia.11 It also became apparent that cisterns and wells now had to be taken into consideration, as Dora Crouch and Renate Tölle-Kastenbein included them in their overarching works on ancient water supply systems.12 From the 2000s onwards, there has been an increased interest in cisterns and wells in studies of ancient Mediterranean societies. While this shift is a complex process, two factors, in particular, stand out. The first is that during recent decades, the fields of archaeology and history of ancient Mediterranean societies have become increasingly interested in non-elite dwellings, practices, and everyday life.13 This makes it necessary to take cisterns and wells into account as they were intimately connected with the daily lives of almost all the inhabitants of the ancient world.14 The second is an ever-growing awareness of the importance of   7 E.g., Gell 1810, 130, noting the cisterns in ancient Hermione. Later published in Klingborg 2021.   8 See, e.g., Miller 1974; Ritterspach 1974; Albarella et al. 1993; Zolotarev 2006.   9 See, e.g., Neumann et al. 2014; Mylona 2019; Penttinen & Mylona 2019. 10 E.g., Camp 1977. See also Crouch 1975 for an early study on Roman Palmyra, and Crouch 1984 for a study on Hellenistic Morgantina. 11 Brinker 1990. 12 Tölle-Kastenbein 1990; Crouch 1993. See also Garbrecht 2001 for a holistic study of the water system at Pergamon. 13 See, e.g., Allison 1999; 2001; Nevett 2010. 14 E.g., Jansen 2002; Galor 2004; 2007; Kamash 2010; 2012; Klingborg 2017; Karvonis 2023.

4  Rick Bonnie and Patrik Klingborg ­sustainability and the impact of climatic changes on socio-cultural developments. As a result of this, some scholars have emphasized the perceived benefits of smallscale local water supply systems in contrast to massive interventions turning lakes into deserts and rivers into dry gullies.15 Furthermore, recent studies on cisterns and wells have focused on aspects such as water availability, analysis of the installations at specific sites and in larger areas, as well as their usage at religious sites.16 Water in Ancient Mediterranean Households As indicated by the volume’s title, the contributions in this volume cover a vast time span, from Classical Greece to the Late Roman Empire, and a wide variety of environments, from dry deserts and moderate Mediterranean zones to wet and temperate climates further north (Figure 1.1). Most contributions discuss household water installations in a settlement (Chapters 2, 4, 6, 7, 9), others zoom in on a few buildings or neighbourhoods (Chapters 3, 5). Two chapters focus on several sites in a wider region (Chapters 8, 10). Yet, notwithstanding their diversity in cultural, temporal, and environmental contexts, together the contributions in this volume highlight three intimately interconnected aspects in particular: (1) the interaction between different water sources, (2) small-scale water installations as supplying water on the margins, and (3) how the control and use of water sources created and reinforced social hierarchies and inequalities. Through a focus on these aspects, the contributions show that a careful study of a household’s water supply is a rich source of evidence for understanding everyday decisions, anxieties, and changes in life. The chapters in this volume are organized diachronically as far as possible. The first contribution is Jane Millar Tully’s study (Chapter 2) of the late Classical and Hellenistic wells and cisterns in the otherwise dry Piraeus, Athens’ harbour city. Through a focus on urban gardening, Millar Tully highlights the interaction of various water sources in the city as well as how these allowed activities which the natural setting otherwise would not, literally creating gardens in an otherwise hard and rocky place. Furthermore, urban irrigation practices, while poorly understood, may have had a significant impact on a household’s water consumption previously not taken into consideration, expanding our understanding of water usage in ancient cities.17 In Chapter 3, Patrik Klingborg focuses on another aspect of urban water management, namely how a society dealt with near-constant water stress by exploring the unequal hydraulic relationships created in the Late Hellenistic period on Delos, a small and rocky island in the middle of the Aegean Sea. Long known as a dry and desolate place, a small minority of the inhabitants were in control of the vast 15 Antoniou et al. 2014, 692; Angelakis et al. 2016; Tzanakakis et al. 2020. 16 On water availability, see Klingborg & Finné 2018; Aarnio 2021. For analysis of the material at specific sites, see Braemer et al. 2015; Albrecht 2016; Klingborg 2019; Seifried 2019; Ore et al. 2020. For larger areas, see Mantellini 2015; Klingborg & Finné 2018. For usage at religious sites, see, e.g., Kamash 2008; Ehrenheim et al. 2019; Fuchs 2023; Kimmey 2023. 17 See, e.g., Connelly & Wilson 2002; Klingborg & Finné 2018.

Water in Ancient Mediterranean Households  5

Figure 1.1 Map with all sites discussed in this book. Map by Patrik Klingborg. Base map: Stamen Design, under CC BY 3.0. Data by OpenStreetMap, under ODbL.

majority of water sources, all of them vulnerable to seasonal and annual fluctuations. As a consequence of this, those with wells and cisterns must have had significant leverage on individuals who did not have access to water. This is followed by Angela Commito’s study (Chapter 4) of the site of Notion in Western Turkey during the last three centuries BC and the first century AD based on the results of recent fieldwork, adding to the available evidence. Similar to Piraeus and Delos, Notion is characterized by a lack of natural water sources, making the addition of artificial systems crucial for urban development. As in some other cities, this was solved by a combination of cisterns, many of which are published here for the first time, and a recently discovered and still undated terracotta conduit leading water under pressure. This suggests that there was both private and public interest in securing water for the site. The theme of water scarcity is continued in Ann Glennie’s chapter (Chapter 5) on rainwater harvesting at Cosa in Southern Tuscany, Italy, in the late Republic and early Empire. Just as at Notion, the lack of natural water sources at Cosa prompts questions regarding how the local population managed their need for fresh water. However, in contrast to Notion, Cosa was never provided with water from distant sources. Instead, water supply was limited to cisterns found in private and public contexts. This resulted in a vulnerability to droughts and the need for careful water management by the inhabitants, perhaps reminiscent of that discussed for Delos.

6  Rick Bonnie and Patrik Klingborg In the following contribution (Chapter 6), Gemma Jansen explores water management and the use of cisterns at Pompeii from a different perspective, as there was no lack of water in the city. Rather, she focuses on where cisterns were located in various households, and how the use of these was affected by the influx of new water resources through an aqueduct. Thus, while she focuses on the cisterns, the use of aqueduct water, or the possible lack of such a water supply, during the last days of the town looms large in the background. How did an unreliable large-scale water supply system affect the usage and abandonment of cisterns? Moving to North Africa, Mark Locicero takes another approach in Chapter 7 through what he coins as quotidian waters in the Roman city of Volubilis, Northern Morocco, primarily during the third century AD. Volubilis, like Pompeii, had an aqueduct, but here it only supplied the upper city with the more affluent houses. In contrast, the lower city, where workshops have been identified and production took place, did not receive any running water. What is more, this part of the city shows no evidence of any wells or cisterns, despite the need for abundant water there to maintain production. Thus, it appears that the inhabitants of the lower city had to carry water up a steep hill from the nearest river running by Volubilis to manufacture the wealth that allowed the elite to avoid this onerous task. Here, as also shown for other sites in this volume, secure and continuous access to water was very much a question of wealth and status. A household’s interaction with their water sources can also be viewed from how they could be multifunctional. To explore this, Rick Bonnie (Chapter 8) investigates so-called stepped pools in Roman Palestinian households from the perspective of social and cultural biographies.18 It has been suggested that these stepped pools were used primarily for Jewish ritual purification practices by the occupants of these houses. Using an object biography approach, Bonnie argues that such a singular meaning and functioning for these household water features is too short-sighted and even problematic. Examples from households in Jerusalem, Jericho, and Sepphoris suggest that stepped pools there frequently only had a relatively short original ritual lifespan. After this, they may have been given a different household function, usually related to food storage. As such, he allows us to break away from the perspective of water supply resources as never-changing, singlepurpose installations. In another example from the East, Jennifer Baird (Chapter 9) turns to the material from Dura-Europos in the Syrian desert. This city differs from those discussed in the previous chapters by having good access to water through the mighty Euphrates River. Notwithstanding this, many of the excavated houses show evidence for cisterns, presumably in order to secure water in case of emergencies such as sieges. Furthermore, Baird suggests that household independence was an important factor. In the final contribution to this volume, a socially created form of water stress is explored. In his study of wells in towns and rural villas of Roman Britain, James

18 On object biography, see, e.g., Kopytoff 1986; Gosden & Marshall 1999; van Haasteren & Groot 2013; Joy 2015.

Water in Ancient Mediterranean Households  7 Gerrard (Chapter 10) portrays a situation where access to water was a central concern despite living in a temperate climate with regular and abundant rainfall. Furthermore, his study shows how this water stress could have significant social consequences. Here the local elite desired, and presumably needed, to be part of Roman culture. Consequently, they required access to large volumes of water for bathing, and in Gerrard’s example acquiring water was achieved at great expense in terms of hard and dangerous labour, certainly performed by slaves or otherwise dependant individuals. Future Directions With regard to small-scale water supply installations in households across the ancient Mediterranean and beyond and building on the contributions to this volume, there are numerous further aspects that can be studied. For example, there is ample room for archaeologists and historians to turn towards the effect of cisterns and wells on individuals and their lives and the interaction between these and the environment around them, as well as the spatiality of the water supply. With the latter, we mean that it is frequently observed how neighbouring households may have had unequal access to water sources. Several of the contributions in this volume touch on this matter and interpret this unequal access as an indicator for social inequality and mobility. However, on a Mediterranean scale we still know little about the accessibility of private water sources in and beyond the home, even if this matter is of high importance for understanding everyday social interactions. This is also affected by how we today tend to perceive ‘private’ and ‘public’. Research has shown that in antiquity people’s conception of private and public space and features were markedly different.19 These differences need to be considered carefully in future studies of both previously considered private and public water sources. Additionally, there is a clear need for cooperative and interdisciplinary work discussing all components of the water supply in tandem among scholars within diverse fields such as archaeology, history, urban studies, geology, environmental hydrology, and climatology. As some of the contributions in this volume show, a household’s water supply in the ancient Mediterranean, as elsewhere, is highly variable and ultimately reliant on both the environment and neighbourhood planning and construction. Another underexplored aspect, raised by Gerrard, is the effect of Mediterranean household habits in a Northern European context. He also shows how it is possible to approach cisterns and wells from a more inclusive and sensitive perspective, and how they affected individuals and their lives. No longer are the wells he discusses simply providing a commodity, but they affect the humans around them in different and varied ways; for some they are a boon, for others a curse. This discussion could be expanded across the Mediterranean. How was it to live with cisterns and wells in

19 See, e.g., Wallace-Hadrill 1994; Russell 2015; Tuori & Nissin 2015.

8  Rick Bonnie and Patrik Klingborg ancient Athens or Rome? How did experiences there differ from those in the cities of North Africa? These types of studies would provide fresh and important insights into the differences of everyday life across cultures, environments, and regions. The different contributions to this volume also show the value of adopting a broader perspective in investigating the spatiality of cisterns and wells within urban areas. Where, for whom, and when is water available? While in many ways being the opposite of exploring how individuals were affected by these installations, a spatial approach can be used to highlight similar issues. This is the case, for example, in Klingborg’s chapter, where the island of Delos shows a stark difference between having access to or being deprived of water sources. In general, as any spatial distribution of wells and cisterns makes clear, access to water in both rural and urban areas was far from equal. This has been noted across regions and periods by various contributors to this volume. However, past scholarship has seemingly almost neglected access to a household’s basic drinking and sanitation needs as a matter for understanding social inequality. This is understandable considering modern Western accessibility to such water, but this aspect needs more consideration for the ancients. More emphasis is also needed on the functioning of cisterns, wells, and other small-scale water supply installations beyond the mundane in ancient Mediterranean houses—there can be little doubt that water was also used for ritual needs. There are few exceptions to this general neglect of the ritual functioning of private water sources. The clearest examples are the stepped pools found in houses across the Southern Levant, which are routinely interpreted as having a primarily ritual function for Jewish purification. In those cases, the need is rather to explore more their mundane functioning, as Bonnie discusses in his chapter. Finally, there is also a great deal of potential in cooperative efforts exploring the water supply and usage in an area from a holistic perspective. This would allow us to reconstruct the whole matrix of water usage, from the acquisition of water sources and who benefited from access to them to how excess water was disposed of. In a similar vein, interdisciplinary efforts would be able to elucidate further aspects, such as the effect of water usage on the groundwater, and consequently, for example, on vegetation and erosion. Micromorphology and analysis of silt remains could also reveal important information about the actual usage of cisterns and wells in the past. There is, thus, little doubt that while a number of recent publications, including this one, have attempted to shed further light on the ancient water supply system, and in particular, the use of non-monumental resources, there is still much that can be done. References Aarnio, N. 2021. Hydrological analysis of stepped pools in Roman Palestine, MSc thesis, Aalto University. Albarella, U., V. Ceglia & P. Roberts 1993. ‘S. Giacomo Degli Schiavoni (Molise): An early fifth century AD deposit of pottery and animal bones from Central Adriatic Italy’, Papers of the British School at Rome 61, 157–230.

Water in Ancient Mediterranean Households  9 Albrecht, N. 2016. Römerzeitliche Brunnen und Brunnenfunde im rechtsrheinischen Obergermanien und in Rätien. Studia Archaeologica Palatina 1, Leipzig. Allison, P.M., ed. 1999. The archaeology of household activities, London. Allison, P.M. 2001. ‘Using the material and written sources: Turn of the millennium approaches to Roman domestic space’, AJA 105, 181–208. Angelakis, A.N., K.S. Voudouris & I. Mariolakos 2016. ‘Groundwater utilization through the centuries focusing οn the Hellenic civilizations’, Hydrogeology Journal 24, 1311–1324. Antoniou, G., N. Kathijotes, D.S. Spyridakis & A.N. Angelakis 2014. ‘Historical development of technologies for water resources management and rainwater harvesting in the Hellenic civilizations’, International Journal of Water Resources Development 30(4), 680–693. Braemer, F., D. Gazagne & G. Davtian 2015. ‘Small-scale water systems in the fertile crescent. The role of cisterns-based water systems in an arid zone between rain-fed agriculture and stockbreeding during Roman times’, Water History 7(4), 455–471. Brinker, W. 1990. Wasserspeicherung in Zisternen. Ein Beitrag Zur Frage Der Wasserversorgung Früher Städte. Leichtweiss-Institut für Wasserbau der Technischen Universität Braunschweig, Mitteilungen 109, Braunschweig. Broneer, O. 1954. Corinth I:4. The South Stoa and its Roman successors, Princeton, NJ. Camp, J. 1977. The water supply of ancient Athens from 3000 to 86 B.C., Ph.D. thesis, Princeton University. Coates-Stephens, R. 2003. ‘The water-supply of early Medieval Rome’, in Technology, ideology, water: From Frontinus to the Renaissance and beyond. Papers from a conference at the Institutum Romanum Finlandiae, May 19-20, 2000, eds. C. Bruun & A. Saastamoinen. Acta Instituti Romani Finlandiae 31, Rome, 81–113. Connelly, J.B. & A. Wilson 2002. ‘Hellenistic and Byzantine cisterns on Geronisos Island. With a mortar analysis by C. Doherty’, Report of the Department of Antiquities Cyprus, 269–292. Crouch, D.P. 1975. ‘The water system of Palmyra’, Studia Palmyreńskie 7, 151–186. Crouch, D.P. 1984. ‘The Hellenistic water system of Morgantina, Sicily: Contributions to the history of urbanization’, AJA 88(3), 353–365. Crouch, D.P. 1993. Water management in ancient Greek cities, Oxford. Ehrenheim, H., P. Klingborg & A. Frejman 2019. ‘Water at ancient Greek sanctuaries: Medium of divine presence or commodity for mortal visitors?’, Journal of Archaeology and Ancient History 26, 1–31. Fuchs, J. 2023. ‘The water supply of the Heraion of Samos’, in Going against the flow. Wells, cisterns and water in ancient Greece, ed. P. Klingborg. ActaAth-4°, 23, Athens, 135–159. Galor, K. 2004. ‘Qumran’s plastered pools: A new perspective’, in Khirbet Qumrân et ’Aïn Feshka, Vol. 2: Études d’antropologie, de physique et de chimie, eds. J.-B. Humbert & J. Gunneweg, Göttingen, 291–320. Galor, K. 2007. ‘The stepped water installations of the Sepphoris Acropolis’, in The archaeology of difference. Gender, ethnicity, class and the ‘other’ in antiquity. Studies in honor of Eric M. Meyers, eds. D.R. Edwards & C.T. McCollough, Boston, 201–213. Garbrecht, G. 2001. Altertümer von Pergamon I:4. Stadt Und Landschaft. Die Wasserversorgung von Pergamon, Berlin. Gell, W. 1810. The itinerary of Greece: With a commentary on Pausanias and Strabo and an account of the monuments of antiquity at present existing in that country; compiled in the years 1801–06, London. Gibbon, E. 1776–1789. The history of the decline and fall of the Roman Empire, 6 Vols, London.

10  Rick Bonnie and Patrik Klingborg Glaser, F. 2000. ‘Fountains and nymphaea’, in Handbook of ancient water technology, ed. Ö. Wikander. Technology and change in history 2, Leiden, 413–451. Gosden, C. & Y. Marshall 1999. ‘The cultural biography of objects’, World Archaeology 31, 169–178. Grewe, K. 2019. Aquädukte: Wasser Für Roms Städte, 3rd ed., Daun. van Haasteren, M. & M. Groot 2013. ‘The biography of wells: A functional and ritual life history’, Journal of Archaeology in the Low Countries 4(2), 25–51. Hamilakis, Y. 2013. Archaeology and the Senses: Human Experience, Memory, and Affect, Cambridge. Hodge, A.T. 2000. ‘Wells’, in Handbook of ancient water technology, ed. Ö. Wikander. Technology and change in history 2, Leiden, 29–33. Hodge, A.T. 2002. Roman aqueducts & water supply, London. Jansen, G.C.M. 2002. Water in de Romeinse Stad: Pompeji – Herculaneum – Ostia, Leuven. Joy, J. 2015. ‘“Things in process”: Biographies of British Iron Age pits’, in Biography of objects: Aspekte eines kulturhistorischen Konzepts, ed. D. Böschung, P.-A. Kreuz & T. Kienlin, Paderborn, 125–141. Kamash, Z. 2008. ‘What lies beneath? Perceptions of the ontological paradox of water’, World Archaeology 40(2), 224–237. Kamash, Z. 2010. Archaeologies of water in the Roman Near East. Gorgias Dissertations in Near Eastern Studies 54, Piscataway, NJ. Kamash, Z. 2012. ‘Irrigation technology, society and environment in the Roman Near East’, Journal of Arid Environments 86, 65–74. Karvonis, P. 2023. ‘The water supply in the houses of Delos’, in Going against the flow. Wells, cisterns and water in ancient Greece, ed. P. Klingborg, ActaAth-4°, 23, Athens, 77–90. Kimmey, S. 2023. ‘The Nemean Wells. Water management and sanctuary deposition’, in Going against the flow. Wells, cisterns and water in ancient Greece, ed. P. Klingborg. ActaAth-4°, 23, Athens, 113–134. Klingborg, P. 2017. Greek cisterns: Water and risk in ancient Greece, 600–50 BC, Ph.D. thesis, Uppsala University. Klingborg, P. 2019. ‘Fill and chronology in ancient Greek cisterns’, Schriftenreihe der Frontinus-Gesellschaft 31, 43–63. Klingborg, P. 2021. ‘The cisterns of the Bisti Promontory at Hermione’, Opsucula 14, 135–155. Klingborg, P. 2023. ‘Wells and cisterns in Greek literature’, in Going against the flow. Wells, cisterns and water in ancient Greece, ed. P. Klingborg. ActaAth-4°, 23, Athens, 161–178. Klingborg, P. & M. Finné 2018. ‘Modelling the freshwater supply of cisterns in ancient Greece’, Water History 10(2), 113–131. Kopytoff, I. 1986. ‘The cultural biography of things: Commoditization as process’, in The social life of things: Commodities in cultural perspective, ed. A. Appadurai, Cambridge, 64–91. Longfellow, B. 2011. Roman imperialism and civic patronage: Form, meaning and ideology in monumental fountain complexes, Cambridge. Mantellini, S. 2015. ‘The implications of water storage for human settlement in Mediterranean waterless islands: The example of Pantelleria’, Environmental Archaeology 20(4), 406–424. Miller, S.G. 1974. ‘Menon’s cistern’, Hesperia 43(2), 194–245. Mylona, D. 2019. ‘Animals in the sanctuary. Mammal and fish bones from areas D and C at the sanctuary of Poseidon at Kalaureia’, Opuscula 12, 173–221.

Water in Ancient Mediterranean Households  11 Neumann, F.H., J.K. Zangenberg, Y. Shivtiel & S. Münger 2014. ‘Galilee blooming: First palynological and archaeological data from an early Byzantine cistern at Horvat Kur’, Environmental Archaeology 19(1), 39–54. Nevett, L. 2010. Domestic space in classical antiquity, Cambridge. Ore, G., H.J. Bruins & I.A. Meir 2020. ‘Ancient cisterns in the Negev highlands: Types and spatial correlation with Bronze and Iron Age sites’, Journal of Archaeological Science: Reports 30, 102227. Penttinen, A. & D. Mylona 2019. ‘Physical environment and daily life in the Sanctuary of Poseidon at Kalaureia, Poros. The bioarchaeological remains’, Opuscula 12, 159–172. Richard, J. 2012. Water for the city, fountains for the people. Monumental fountains in the Roman East. Studies in Eastern Mediterranean Archaeology 9, Turnhout. Ritterspach, A.D. 1974. ‘The Meiron cistern pottery’, Bulletin of the American Schools of Oriental Research 215, 19–29. Russell, A. 2015. The politics of public space in Republican Rome, Cambridge. Seifried, R.M. 2019. ‘Seascapes and fresh water management in rural Greece: The case of the Mani Peninsula, 1261–1821 CE’, Levant 51(2), 131–150. Stallybrass, P. & A. White 1986. The politics and poetics of transgression, Ithaca. Tölle-Kastenbein, R. 1990. Antike Wasserkultur, München. Tuori, K. & L. Nissin, eds. 2015. Public and private in the Roman house and society. Journal of Roman Archaeology Supplements 102, Portsmouth, RI. Tzanakakis, V.A., A.N. Angelakis, N.V. Paranychianakis, Y.G. Dialynas & G. Tchobanoglous 2020. ‘Challenges and opportunities for sustainable management of water resources in the Island of Crete, Greece’, Water 12(6), 1538. Urry, J. 2000. Sociology beyond societies: Mobilities for the twenty-first century, London. Wallace-Hadrill, A. 1994. Houses and society in Pompeii and Herculaneum, Princeton. Wikander, Ö., ed. 2000. Handbook of ancient water technology. Technology and change in history 2, Leiden. Wilson, A. 2008. ‘Hydraulic engineering and water supply’, in The Oxford handbook of engineering and technology in the Classical world, ed. J.P. Oleson, Oxford, 285–318. Zolotarev, M.I. 2006. ‘A Hellenistic ceramic deposit from the north-eastern sector of Chersonesos’, in Chronologies of the Black Sea Area in the period c. 400–100 BC, ed. V.F. Stolba & L. Hannestad, Aarhus, 193–216.

2

Household Water, Environment, and Economy in Ancient Piraeus Jane Millar Tully

Introduction The history of Piraeus is usually told in terms of its relationship with the sea, especially the three natural harbours—Kantharos, Zea, and Munichia—through which Athens exercised its naval power. But just as important to the city were its sources of fresh water: winter rainfall, a shallow aquifer, and the infrastructure built to capture, store, and distribute water for human use. Prior to the construction of the Hadrianic aqueduct in the second century AD, and in contrast to the massive statesponsored development of its maritime resources, initiatives to store and distribute fresh water in Piraeus prior to the Roman period seem to have been an entirely non-monumental affair. The water supply structures of Piraeus are noteworthy for the number of preserved examples and the documentary sources that contextualize them in wider socioeconomic and environmental relationships. This study explores the household wells and cisterns of Late Classical and Hellenistic Piraeus, adopting a diachronic perspective to ask how they were used and functioned over time. With several hundred built features of the ancient water supply discovered over decades of rescue excavations (Figure 2.1), it is necessary to be selective. The abundance of evidence allows for an examination at a micro-scale, focusing on a few recently excavated examples to explore how residents prioritized water storage and distribution. The history of Piraeus has so far been written in primarily political and military terms, with studies of its economy facing outward to focus on bulk imports and long-distance trade.1 It remains to be satisfactorily understood how environmental conditions and water supply infrastructure limited and enabled local economic activity on the peninsula. After an introduction to its urban development, this chapter focuses on areas of the city with potential to shed light on broader aspects of the local economy and environment thanks to archaeological and epigraphic evidence, including recent investigations around the Municipal Theatre and Munichia Hill.

1 On the history of Piraeus, see Panagos 1968; Garland 1987; Steinhauer 2012; and for a general introduction, Steinhauer 2021. Most accounts focus on the Classical period, though recent efforts are filling in earlier (Fragkopoulou 2015) and later developments (Grigoropoulos 2016). DOI: 10.4324/9781003268222-2

Household Water, Environment, and Economy in Ancient Piraeus  13

Figure 2.1 Map of modern Piraeus with features of the ancient water supply. Used with permission of G. Peppas, copyright Ephorate of Antiquities of Piraeus and Islands, Hellenic Ministry of Culture and Sports.

Environmental and Historical Background Connectivity with mainland Attica is a defining feature of the natural and cultural history of Piraeus. Strabo traces the etymology of Piraeus to its once having been an island lying ‘over against’ the mainland (πέραν).2 Geomorphological analysis suggests that the peninsula was indeed isolated by a shallow marine bay during the Late to Final Neolithic (c. 4850–3450 BC) and a lagoon in the Early to Middle Bronze Age (c. 2850–1550 BC), changes driven by a combination of isostatic sea level change and local progradation of the Cephissus river.3 Much of the ancient city spread over a low isthmus (17 m above sea level) of alluvial sediment between the mainland and the Akte peninsula (58 m above sea level), rising to its highest point at Munichia Hill (86 m above sea level). To the north and east, marshy coastal plains presented challenges to further expansion. When the Long Walls famously linked Piraeus and Athens in the 460s BC, building efforts included massive fills to level the swampy ground.4 Marshland appears in historic maps until the construction of modern drainage works at the end of the nineteenth century.5 Though the area of Piraeus had been inhabited since prehistory, dense urban settlement only took root with its development as a naval base in the fifth century BC, when it grew from a small deme into one of the busiest commercial hubs in the Mediterranean. Rather than growing organically like Athens, Piraeus was neatly divided up by a Hippodamian grid laid in the mid-fifth century, c. 460–430 BC. At

2 3 4 5

Strab. 1.3.18. Goiran et al. 2011, 531–534; Chiotis 2019, 337. Plut. Cim. 13.7; Xen. Hell. 2.4.34. Travlos 1988, 340; Conwell 2008, 5–7; Wassenhoven 2018, 50–51.

14  Jane Millar Tully the end of the fifth century, Piraeus suffered the destruction of its fortifications and several ruptures from the city of Athens. The resulting depopulation led one author to liken early fourth-century Piraeus to a hollow nutshell without a kern.6 By the third quarter of the fourth century, however, the city had rebounded into a profitable commercial centre as politicians like Eubolos and Lycurgus incentivized metic settlement and encouraged private investment in local markets. Supporting a diverse population of 20,000–25,000 inhabitants, the city spanned the Akte peninsula and rose onto Munichia Hill, with harbour facilities reaching their maximum extent c. 330 BC.7 Some contraction of settlement followed, and the city concentrated around Zea harbour in the mid-Hellenistic period, eventually becoming confined to the isthmus.8 In 87/86 BC, Piraeus was put to siege and sacked by the Roman general Sulla.9 Strabo described a much-reduced settlement in the first century AD, clustered around the harbours and Sanctuary of Zeus Soter,10 which has yet to be located archaeologically. The first century BC provides an endpoint for this chapter, since most of the water supply structures uncovered in recent excavations are capped with fills associated with the destruction in 87/86 and subsequent rebuilding efforts. Nevertheless, Piraeus continued to be a locally and regionally important commercial centre, and portions of an aqueduct and other structures associated with the Roman and Late Antique water supply have been uncovered throughout the city.11 Construction of the water supply infrastructure also took place in a dynamic interplay with the natural conditions of precipitation and geology. Attica is drier than the rest of Greece, part of the arid region that encircles the Saronic Gulf and Cyclades and receives little rain between May and September. The unique geology of Piraeus makes the peninsula more challenging, in hydrological terms, than the city of Athens. Current average annual rainfall is lower, 360 mm in Piraeus to Athens’ 400 mm. While Athens is underlain by porous limestone that rainfall penetrates to replenish local aquifers, much of Piraeus is covered in less permeable marl. But what marls lack in springs, they make up for as an excellent medium for cistern construction. Locally, Piraeus marl was a main ingredient in the waterproof lining used to coat cisterns.12 Based on excavated examples, its cisterns have proven less prone to collapse than their Athenian counterparts built into limestone.13 On Munichia Hill around 40 m above sea level, the impermeable marl gives way to thickly bedded Aktites limestone, which creates conditions favourable for springs, though none are mentioned in ancient sources.14 Nor do they mention the   6 Philiskos, PCG fr. 356.   7 Settlement fluctuations mapped by Grigoropoulos 2016, 244, Fig. 3, based on dated archaeological sites.   8 Lewis 1990, 250–251; Steinhauer 2021, 231–233.   9 Appian, Mith. 30–41; Plut. Sull. 14.6–7. 10 Strab. 9.1.15. 11 E.g., Axioti 2009 on the excavated houses at Rondiris Square. 12 Chiotis 2019, 340–343. 13 Papazachos 1990; Stoumbos 2015. 14 Chiotis 2019, 334–335. Two natural springs noted by von Eickstedt (1991, 121–122) were apparently unknown in antiquity. Several ancient quarries can still be visited in parks near the summit of the hill.

Household Water, Environment, and Economy in Ancient Piraeus  15 single stream called Tzirloneri, known for curative properties in the nineteenth century, which used to run from the Akte peninsula to a mouth southwest of Zea harbour.15 By the time Vitruvius was writing in the first century BC, aqueducts brought water from Athens to Piraeus, though apparently not of potable quality.16 At a macro-scale, then, the environment of Piraeus is poor in water sources. But on the meso- and micro-scale, its inhabitants created and maintained relatively well-watered areas that could be utilized for a wide array of activities. By offering households a measure of independence from Attica’s unpredictable precipitation, water supply systems reduced their vulnerability to drought and flood, provided those systems were well maintained.17 History of Archaeological Investigations On the surface, the settlement at ancient Piraeus was rigorously planned and executed, the only surviving urban grid securely attributed to Hippodamus of Miletus.18 Recovering the Classical grid has been a research priority from the beginning of archaeological investigations in Piraeus, and early findings informed the city’s modern planning in the nineteenth century.19 Ever since, rescue excavations, especially during the rapid commercialization of the harbour in the last sixty years, have confirmed and added to the city’s Classical plan and later expansions.20 North– south streets included drains and channels to remove wastewater,21 but accessing and storing clean water for household use appears to have been the prerogative of individual households. The result is that below its carefully laid urban grid, the peninsula’s bedrock is riddled with shafts, tunnels, and cisterns that appear to follow their own logic. While in most cases surface architecture does not survive, the density of water supply features suggests that each house would have had its own water source, probably accessed through a cistern or well mouth in the courtyard, an impression reinforced when excavated features are mapped on the reconstructed urban grid.22 The grid petered out on the slopes of Munichia Hill, but a variety of wells, cisterns, channels, and aqueducts attest to dense occupation, puncturing the rocky hill with such frequency that a nineteenth-century visitor remarked that they made it ‘dangerous to walk about the ruins’.23 Though early visitors and investigators had noted great numbers of water-storage features in Piraeus, their systematic examination has begun only recently. Klaus-Valtin von Eickstedt offered the first synthesis of water supply features (279 in total) discovered through 1981, though only about 15 Garland 1987, 145–146. 16 Vitr. 8.3.6. 17 Horden & Purcell 2000, 241, 244–247. 18 Arist. Pol. 2.1267b. 19 Curtius & Kaupert 1881, 34; Dörpfeld 1884, 279–287; Judeich 1931, Plan III. On the Hippodamian plan, see Hoepfner & Schwandner 1994, 22–50; Shipley 2005, 368–373; Gill 2006. 20 Steinhauer 2012, 96–97; see also Papadopoulou 2015. 21 Steinhauer 2021, 235–236. 22 Steinhauer 2012, 104–105; Avgerinou et al. 2016, Figure 4.4. 23 Dodwell 1819, 426.

16  Jane Millar Tully 20% could be dated with any confidence, as several had been put to modern use as cesspits or septic tanks (Sickergruben).24 Rescue excavations since then bring the total to some 388 cisterns and 345 wells and shafts. Water supply features are found in nearly every excavated plot, but very rarely excavated to their full depth for reasons of safety and time constraints.25 While von Eickstedt’s summary offered a useful overview of the distribution and morphology of the non-monumental water supply across Piraeus, it is only with recent excavations that many more cisterns and wells have been dated, and several wells excavated to their full depth, bringing into focus dynamic strategies for water supply across the centuries. Excavations between 2012 and 2015 for the Dimotiko Theatro (Municipal Theatre) Station of the Athens Metro were the largest in Piraeus history, covering 4000 m2 to a depth of up to 30 m below surface. The work focused on Agios Konstantinos Square, overseen by Panagiotis Koutis, and Pavlos Bakoyiannis Square, overseen by Giorgos Peppas (Figure 2.2). Regrettably no surface architecture or remains were preserved, having been cleared for construction of the Municipal Theatre and associated squares in the late nineteenth and early twentieth centuries. But subsurface features—mostly related to water supply—were excavated and recorded in collaboration with geologists and architectural engineers, and georeferenced maps integrated with previous excavations and the reconstructed urban grid. Ephorate archaeologists systematically recorded and studied 132 structures related to the water system including 63 wells and shafts, 37 cisterns, and 32 linear features (tunnels, pipes, and the Roman aqueduct).26 Analysis of finds is ongoing, but the addition of so many dated structures offers significant insight into the ancient water supply of Piraeus and its relationship with urban development. Diachronic Changes in Household Water Supply: New Insights from the Metro Excavations Inhabitants of Piraeus seem initially to have followed the model of Athens and dug wells.27 Of the 42 wells investigated during the Metro excavations, with an average diameter just under one meter, seven appear to have been abandoned before completion (c. 1.3 m deep) and were filled in during the late fifth or early fourth century BC. Most were dug to an average depth of fifteen meters below surface, to reach an aquifer with an absolute altitude of three meters above sea level. Well 28, 24 von Eickstedt 1991, 121–133 (discussion), 194–237 (catalogue); several revisited in Klingborg 2017, 207–212. More recent rescue excavations in Piraeus are published annually in the Αρχαιολογικόν Δελτιόν (ΑΔ), summarized with reference to the Metro excavations in Chrysoulaki et al. 2017, 424. 25 Avgerinou et al. 2016, Figure 4.4; Peppas, forthcoming. 26 On excavations for Athens Metro S.A. ‘Extension of Line 3: Haidari-Piraeus,’ including a 7.6 km tunnel with six stations and ten ventilation shafts, see Avgerinou et al. 2016, 48–50; Chrysoulaki et al. 2017; Peppas, forthcoming; and the conference proceedings from the National Technical University of Athens, May 15, 2015, available at efadyat.wordpress.com. 27 The bibliography on Athenian water supply is extensive, starting with Camp 1977; most recently summarized by Stroszeck 2021. On wells see Chiotis & Chioti 2012, 407–424; Koutsoyiannis & Mamassis 2017, 31–42; Stroszeck 2017, 43–63; on cisterns see Klingborg 2017, 57–66; Stroszeck 2023.

Household Water, Environment, and Economy in Ancient Piraeus  17

Figure 2.2 Underground water supply structures excavated for the Dimotiko Theatro Station on a reconstructed urban grid, with features mentioned in the text labeled. Crosssection after Koutis & Bendermacher-Gerousis 2015, slide 5. Copyright Ephorate of Antiquities of Piraeus and Islands, Hellenic Ministry of Culture and Sports.

rectangular in cross-section, plunges eighteen meters deep (3.5 m below sea level), with a stepped bottom that may have been aimed at increasing its capacity (Figure 2.3). This is only surpassed by circular Well 33 (5 m below sea level). Excavators believe that these deep wells were dug when the water table was lowered by

18  Jane Millar Tully

Figure 2.3 Interior of Well 28. Copyright Ephorate of Antiquities of Piraeus and Islands, Hellenic Ministry of Culture and Sports.

drought and/or overuse.28 All are roughly carved near the base, probably reflecting the difficulty of working in standing water, and extra cuttings in the walls of several wells at the level of the water table most likely represent efforts to increase groundwater infiltration.29 A general shift from wells to cisterns in Piraeus has been attributed to the fourth century BC, contemporaneous with the appearance of cisterns in the Athenian Agora, but the development of water storage in Piraeus may have followed a different trajectory than that of Athens. Thucydides mentions that the Piraeus had wells (φρέατα) at the outbreak of the Peloponnesian War in 430 BC, but no springs or fountains (κρῆναι).30 Suddenly [the plague] fell upon the city of Athens, and attacked first the population in Piraeus, so that it was said by them that the Peloponnesians had put poison in the wells: for there were as yet no public fountains there.31

28 Koutis & Bendermacher-Gerousis 2015, 3. 29 Benisi 2015; Koutis & Bendermacher-Gerousis 2015, 2. 30 On κρῆναι see generally Glaser 1983; on terminology see Tölle-Kastenbein 1985. The slightly ambiguous term φρέαρ (pl. φρέατα) refers to wells in the Classical period, but from about 200 BC is also applied to cisterns in literary sources (Klingborg 2023). 31 Thuc. 2.48.2. Translation by author.

Household Water, Environment, and Economy in Ancient Piraeus  19 Thucydides’ statement that in 430 BC there were no fountains yet might imply that some form of public water supply had been constructed before his death (c. 400 BC), as Garland suggests,32 but this remains to be confirmed archaeologically. Of course, an accurate description of the Piraeus water supply was hardly the historian’s main objective, and Thucydides goes on to say that the plague’s origins are speculative.33 Still, suspected contamination of existing water sources may have spurred the construction of new and alternative infrastructure: a shift from wells— more vulnerable to mass contamination, since it was believed that all active water sources were connected underground, as indeed local aquifers are—to cisterns built and maintained in individual households.34 In the excavated wells at Agios Konstantinos Square, some stratified deposits have been dated to the end of the fifth and the beginning of the fourth centuries BC, so if there was an interruption in the use of wells, it appears to have been short-lived. In one case, the excavation of a flask-shaped cistern (1) uncovered a well (1β’) sealed beneath it (Figure 2.4). The well contained fifth-century ceramics, leading excavators to interpret this combined feature as the result of a change in water supply during the Peloponnesian War, whereby the construction of a cistern sealed off the well and groundwater supply and opened a space for the storage of rainwater instead.35 The replacement of Well 1β’ with Cistern 1 is a remarkable case, but for the most part cisterns did not entirely replace wells. Based on preliminary analysis of well deposits excavated for the Dimotiko Theatro Station, many were still in use in the Hellenistic period, during which several new wells were also opened. Cisterns, then, were adopted alongside wells to secure an adequate water supply. The change was likely spurred by increased demographic pressure on available groundwater as the Attic population moved within the walls during the Peloponnesian War,36 and by the city’s continued growth thereafter. New cisterns could have been part of the so-called Kononian building phase of 394–391 BC, ensuring the city was self-sufficient in water while its fortifications were reconstructed and expanded. Wells alone could not support the population, especially relying on a coastal aquifer vulnerable to intrusion by seawater when depleted.37 Cisterns offered individual households an appealing alternative, a reliable source of freshwater not affected by the actions of a growing population, and one to which the bedrock of Piraeus was naturally suited.38 Individual preference certainly also played a role. While cistern water would have been clean, clear, and odourless when all components of

32 Garland 1987, 145. 33 Thuc. 2.48.3. 34 Klingborg 2017, 129–131 develops this theory with reference to risk management; see also TölleKastenbein 1990, 11; cf. Camp 1977, 148. 35 Koutis & Bendermacher-Gerousis 2015, 6. 36 Thuc. 2.17.3. 37 For saltwater intrusion, see also Klingborg, this volume. 38 Antoniou et al. 2014, 680; Chrysoulaki et al. 2017, 426–427.

20  Jane Millar Tully

Figure 2.4 Flask-shaped Cistern 1, with Well 1β’ below. Copyright Ephorate of Antiquities of Piraeus and Islands, Hellenic Ministry of Culture and Sports.

the collection and storage system were well-maintained, ancient sources record a variety of preferences regarding water sources and their respective merits.39 There is some epigraphic and literary evidence that attests to institutional frameworks for the sustainable management of water.40 Even after aqueducts began to supplement or replace wells and cisterns in Athens, it was recommended to keep those private water sources clean and operable in case war or siege cut off or contaminated the external supply.41 However, most contemporary sources concern public projects. There were elected officials to oversee the Athenian water infrastructure (κρουνών ἐπιμελητής) and decrees on its regulation.42 Writing much later, Plutarch (AD 45–120) attributes regulations on the digging and use of wells to Solon.43 The law quoted stipulates that should someone digging a well fail to hit groundwater after ten fathoms (18.3 m), that person was entitled to take six choae (20 l) twice a day from a neighbour’s well. Regrettably, there is much less on cisterns, and we do not know what consequences there may have been for broaching a neighbour’s well or cistern with a new construction (for which see archaeological examples below). Most Piraeus cisterns fall into three categories: small and conical (max. 1.6 m diam., shallow), shaft-shaped (max. 1.6 m diam., 6–8 m deep), and flask-shaped

39 Garland 1987, 215; Brinker 1990, 68; Crouch 1993, 166; Klingborg 2017, 83–86. 40 Koutsoyiannis et al. 2008; Krasilnikoff & Angelakis 2019. 41 Aen. Tact. 40.8. On cisterns and sieges, see Brinker 1990, 6; Tölle-Kastenbein 1990, 185. 42 For officials, see Arist. Ath. Pol. 43.1, for regulations, e.g., IG I3 49, c. 430 BC. Krasilnikoff & Angelakis 2019, 254–256. 43 Plut. Sol. 23.5.

Household Water, Environment, and Economy in Ancient Piraeus  21 (3.2–6.14 m diam., connected to the surface by a shaft 2–3 m deep). All were underground, saving valuable space in the small plots allotted by the urban grid (230–240 m2 including courtyards). The small and conical cisterns are earliest, carved into bedrock in the late fifth century or early fourth BC, based on associated finds. They were built near the road and unlined, so they may have served another purpose, such as waste disposal. If they were used as cisterns, this design was short-lived, and hydrological modelling suggests that such small reservoirs would have significant overflows in winter and failed to meet water demand in summer.44 Flask-shaped cisterns became the most common form of water storage in Piraeus, with narrow necks that expand underground to a flat-bottomed conical body, circular in plan, with a 10–30 m3 capacity. Their larger volume would have reduced, if not eliminated, the risk of failure and overflow.45 The shape was statically stable, with the internal pressure of the stored water absorbed by the surrounding bedrock, and hydraulic lining kept them watertight. The underground holding tank is, of course, just one part of the ancient cistern system, which collected rainwater runoff from elevated surfaces, usually the roof, through a system of eaves, channels, and pipes. Very few such features survive, though gutters, eavestroughs, and downspouts in stone, lead, or terracotta have been noted at other Greek sites.46 At Piraeus, there are several preserved well heads ( puteals) and small settling basins adjacent to the mouth of the cistern ( prolakkia), which helped prevent contaminants from entering the system. But without more preserved surface features, it is impossible to know exactly how the Piraeus cisterns operated in context. Most cisterns appear in pairs linked by a short tunnel, usually one flask-shaped cistern and one shaft with footholds.47 Excavators interpret the large flask-shaped cisterns as collection and settling tanks, and the shaft-shaped cisterns for drawing clean water, a system described by Vitruvius and Pliny the Elder.48 While flask-shaped cisterns—also described as bell-shaped, bottle-shaped, and carafeshaped—appear in several parts of the Greek world, joining them with shafts appears to have been primarily an Attic phenomenon.49 Some have separation walls in the connecting tunnels, which might have allowed cleaning or repair on one side while the other continued to provide water, or ensured a minimum water level in part of the system.50 While in some cases cistern systems were clearly constructed as a single unit, others evolved with the properties they supplied. The Metro excavations uncovered one in Pavlos Bakoyiannis Square with an unprecedented seven features—five cisterns and two wells—connected by four tunnels (Figure 2.5). Based on the slopes of cistern floors and connecting tunnels, excavators interpreted

44 Mamasis et al. 2015; on early cisterns, cf. Koutsoyiannis et al. 2008. 45 Cf. Klingborg & Finné 2018 on examples from Olynthos and Dystos. 46 Biernacka-Lubanska 1977, 29; Klingborg 2017, 34–35, 38–42. 47 Steinhauer 2012, 103–104. 48 Vitr. 8.6.12; Plin. HN. 36.52. 49 Daux 1962, 657; Klingborg 2017, 24–26, nos. 113–115 and 122. 50 Klingborg 2017, 50, n. 312.

22  Jane Millar Tully

Figure 2.5 Cisterns (1–5, numbered from west to east) and well (1) in Pavlos Bakoyiannis Square. Copyright Ephorate of Antiquities of Piraeus and Islands, Hellenic Ministry of Culture and Sports.

this as the combination of three smaller systems, probably to serve a large home or public building that merged several housing plots.51 With increasingly dense urban settlement, residents adopted several creative solutions to increase cistern capacity. The Metro excavations found some form of expansion interventions in nearly all cases. The gouging of well walls around the level of the aquifer has already been mentioned. For flask-shaped cisterns, whose main chambers could not be easily made larger, tunnels were dug into bedrock or neighbouring cisterns.52 One dead-end tunnel reached 16.3 m in length, but most stretched only about three or four meters, proportional to the diameter of the tank (Cistern 1 in Figure 2.4 is a good example).53 Around the Dimotiko Theatro Station, several excavated cisterns appear to have been joined together in the Late Classical or Early Hellenistic period, mirroring building activity on the surface as houses expanded to fill multiple lots.54 In other cases, the construction of tunnels or new cisterns incorporated existing infrastructure. This was the case for Cisterns 6 and 7 in Agios Konstantinos Square, both of which have bottom depressions associated with cleaning and sediment collection (Figure 2.6). Excavators believe that shaft-shaped Cistern 6 operated independently, 51 Peppas, forthcoming; see also Klingborg, this volume. 52 This is not an uncommon technical solution, occurring in about 20% of the flask-shaped cisterns studied by Klingborg (2017, 26). 53 Daux 1962, 657; von Eickstedt 1991, 216–217. 54 Steinhauer 2012, 100–102.

Household Water, Environment, and Economy in Ancient Piraeus  23

Figure 2.6 Cisterns 6 (shaft, left) and 7 (flask-shaped, right). Copyright Ephorate of Antiquities of Piraeus and Islands, Hellenic Ministry of Culture and Sports.

based on the bottom depression and footholds for maintenance, then was incorporated into the larger flask-shaped Cistern 7. The addition of new cisterns to a system was probably not simply to increase water storage volume—blind tunnels could do that, without the added risk of contamination from an additional opening to the surface. Rather, new openings provided additional collection points, perhaps to adapt the system to a new roof layout as properties expanded or contracted. They could also provide additional places from which to draw water, if the old opening became inaccessible or inconvenient. Not only aggregation, but also later division, is visible in the water system. In the Late Hellenistic or Early Roman period, lots were subdivided into smaller properties and parts of the water system went out of use.55 Tanks which were joining were subdivided by barrier walls. In Figure 2.5, separation walls were found in the tunnels between Cisterns 2 and 3 and Cisterns 4 and 5. In Piraeus, where the Hippodamian plan has been reconstructed from excavations throughout the city, developments in the water supply features serve as proxies for housing patterns in areas where surface architecture does not survive. The diversity of cistern shapes in Piraeus might reflect the varied origins of its inhabitants. The majority are flask-shaped as in Athens, but the remainder includes relatively more shafts, pear-shaped, and even roofed examples.56 When systems were combined and expanded, builders use dividing walls to negotiate between rainwater storage in impermeable cisterns and groundwater accumulation

55 Koutis & Bendermacher-Gerousis 2015, 7. For the single example from Athens, see Klingborg 2017, 198–199, 313. 56 Percentage of cistern types for several sites summarised in Klingborg 2017, 115, Fig. 48.

24  Jane Millar Tully

Figure 2.7 From left to right, Well 25, Tunnel 15, Well 13, Well 2, Tunnel 13, and Cistern 3. Copyright Ephorate of Antiquities of Piraeus and Islands, Hellenic Ministry of Culture and Sports.

in permeable wells. During the construction of Cistern 3, builders appear to have breached Well 3β (visible in plan view, Figure 2.2), implying the location of the earlier well may not have been known by the builders, perhaps because it had gone out of use. Rather than sealing the well and incorporating it into the cistern, they built a wall in the cistern to separate the features, 3.27 m high and lined with hydraulic mortar. An opening left at the top allowed the well to serve as overflow in cases of intense winter rains which could overwhelm the system, as much of a threat as drought to the maintenance of a clean water supply. The already large Cistern 3 is connected to Shafts 13 and 25 by two lengths of tunnel, taller between Shaft 13 and Cistern 3 and lower between Shafts 13 and 25. At some point, a wall was constructed to separate Cistern 3 from the rest of the system, after which the shafts and tunnels presumably continued to function as water collectors themselves (Figure 2.7). Another possible motivation for the changes in water systems in the fourth century, one especially resonant with modern anxieties, is drought. This was proposed based on the construction of public water infrastructure in Athens and literary and epigraphic evidence for water and grain shortages, clustered around c. 360 and 330 BC.57 Recent syntheses of high-resolution paleoclimate data have yet to confirm or soundly refute the theory.58 Cistern construction is not a direct proxy for climate, since the extent to which household water supply responded to local conditions is unknown, so cisterns and wells may or may not be good indices of water stress.

57 E.g. Dem. 50.61, see below. Panessa 1981; Camp 1982; cf. Sallares 1991, 392–396. 58 Finné et al. 2014; 2019; Norström et al. 2018; Weiberg et al. 2019. Most Attic records come from the moister eastern peninsula, see Kouli et al. 2009; Kouli 2012; Triantaphyllou et al. 2010.

Household Water, Environment, and Economy in Ancient Piraeus  25 Dendroarchaeological study of waterlogged door panels recovered from Well 65 attests to similar aridity and interannual variability as modern conditions, but without fitting into an existing chronology its utility is limited.59 Moreover, modern rainfall data show that periods of below-average rainfall occurring two or three times a century are statistically to be expected in the Eastern Mediterranean,60 and unlikely to have caused a complete overhaul of the existing water supply system. The complex arrangements of wells and cisterns throughout Piraeus suggest that the water supply responded readily to demographic pressure and the preference and means of the occupants. There may be technological explanations as well; in the mid-fourth-century Aristotle mentions a new waterproofing method for cisterns, part of a discussion of water storage as a strategy for weathering sieges, the art of which Macedonian armies were refining at the time.61 In Piraeus, constructions and technologies of different periods coexisted and operated simultaneously, with fifth-century BC wells operating alongside the second-century AD aqueduct, and Classical cisterns supplying Roman houses. Features of the Piraeus water supply demonstrate how new technologies were integrated into earlier constructions as systems transformed to meet the changing needs of the city. The following discussion introduces some economic factors in household water supply that go beyond demographics and risk management. Acknowledging that non-monumental water supply is not synonymous with domestic use, if the Piraeus water supply exceeded minimum household needs, what economic opportunities did surplus create and at what scale? Considering the water supply on Munichia Hill and its interaction with other locally important features of the unique urban landscape of Piraeus provides some final points for discussion. Water Supply in the Local Economy Historical records suggest that the demesmen, metics, visitors, and slaves who lived and worked in Piraeus pursued a variety of trades. Quarries were a major economic asset, and rather than being limited to outside the walls, small-scale quarries were integrated into the urban fabric and probably used for individual plots. The Aktites limestone outcrops on Munichia were highly exploited, with remains of ancient quarrying still visible around the summit. Throughout the city, the presence of cisterns and wells built into small-scale urban quarries testifies to their secondary use as housing plots.62 When Strabo describes Munichia as a hollow hill, he is more likely referring to quarries and their reuse than to features of the water supply: Munichia is a hill forming a peninsula, much of it full of hollows and undermined both by nature and deliberately, to receive dwellings.63 59 Barboutis et al. 2019; cf. Mariolopoulos 1962 on carbonized wood from the Acropolis (Cupressus sp.). 60 Garnsey 1988, 154–158; Sallares 1991, 392–393; Manning 2018. 61 Arist. Pol. 1330b, 4–7. Sallares 1991, 392, n. 5. 62 Langdon 2000, 248–250; see also Carusi 2019. 63 Strab. 9.1.15. Translation by author.

26  Jane Millar Tully In addition to quarrying, residents pursued commerce, industry, and, though often overlooked in a city that handled so much imported grain, agricultural production.64 In a speech by Demosthenes dated 361 BC, the speaker mentions a water shortage amid a litany of difficulties, and even allowing for some dramatic license, the implications are significant: Not only did the land produce no crop, but that year, as you all know, even water from the wells ran out, so that not a vegetable grew in the garden.65 This suggests that even in a year when field crops failed, people might still expect to irrigate their vegetable gardens and would normally reserve enough for the purpose. This point is significant for two reasons: irrigation is rarely considered in usage estimates for Greek cisterns,66 and traditional understanding of Classical and Hellenistic Greek land use limits gardens to natural springs and water-courses, assuming that water from urban wells and cisterns was a precious commodity limited to human consumption.67 However, epigraphic evidence from the late fourthcentury (338–326 BC) attests to local garden cultivation, and Piraeus’ sanctuaries profited from leasing both land and water.68 A lease record for property dedicated to an unknown deity records nine small gardens (κηπεῖα), though only the last five are preserved,69 with renters from Piraeus and Ceramicus. Positioned along the road to Munichia, the small plots could have been intensively exploited for vegetable production for the Athenian market.70 Unless gardeners had access to springs or streams not mentioned in ancient sources—unlikely, given local geology—they must have used their wells or cisterns for irrigation. The lease for one plot mentions a mulberry tree (συκάμινος), another the rental of marshland (ἕλος).71 The marsh surrounding Piraeus is usually portrayed as limiting economic opportunities rather than encouraging them, but in this case appears to have been made profitable, perhaps for pasture or reed collection. One of the public estates rented out by the deme

64 Garland (1987, 74) interpreted attestations to the celebration of the plowing festival Plerosia (IG II2 1177) and references to pasture (IG II2 2498) as evidence of arable land within the boundaries of Piraeus. 65 Dem. 50.61. Translation by author. 66 Cf. Klingborg & Finné 2018, 117–118. Plato suggests dipping up wastewater from gutters running from public fountains for watering plants (Leg. 6.76Ib–c; Criti. 117). 67 Carroll-Spillecke 1989, 85. 68 See below, IG II2 1361. 69 SEG 33.168b col. II/III. 70 Walbank 1983, 185; Hilditch 2015, 224–225. On the possibility of cultivation within the Long Walls, see Conwell 2008, 55–57. On roads, which followed ancient courses through the nineteenth century, Wassenhoven 2018, 55–56. 71 SEG 33.168b col. II/III. Mulberry tree, see line 10. This would have been the black mulberry (Morus nigra) brought from southwest Asia by the fifth century BC (Theophr. Hist. pl. 1.12.1, 5.6.1–3), and not the white mulberry (Morus alba) later imported from China, widely cultivated for silkworms in the medieval Peloponnese (Morea). Marshland, see line 19.

Household Water, Environment, and Economy in Ancient Piraeus  27 in the late fourth century was called Schoinous, probably in reference to a local abundance of reeds (σχοῖνος), and was used for pasture (ἐννόμια).72 The inscription listing gardens on the road to Munichia dates between the Battle of Chaeronea in 338 and the Athenian defeats of 322 BC, a period that saw a brief but significant growth in population that might have impacted local cultivation as well as the water supply. Well-documented grain shortages during this time could have encouraged Piraeus residents, new and old, to bring more land into cultivation. The same period was defined by religious regeneration in Athens, with a particular emphasis on foreign cults. Lycurgus sponsored infrastructure and encouraged metic settlement in Piraeus,73 especially by Egyptian and Cypriot merchants, and the Thracian worshippers (ὀργέωνες) of Bendis were given special dispensation to lease sanctuary land.74 Sanctuaries usually had access to dependable water sources for cultic purposes, and a fourth-century decree (c. 330–323 BC) records the Piraeus Bendideion, not yet located archaeologically, selling water.75 The lessee is given rights to the water for personal use, but any proceeds from its sale were to be set aside for the upkeep of sanctuary buildings.76 While quarries and sanctuaries are well studied, local landscapes of cultivation have received little attention in Piraeus, where ancient authors describe the diversity of goods flowing in from everywhere.77 At first glance, intensive cultivation seems unlikely on the densely inhabited and inhospitable peninsula, though it is attested for nearby Phaleron.78 A house and garden near the coast is recorded as renting for only 205 drachmas in fourth-century Attica, putting an enterprise like market gardening within the reach of many.79 Actual water demand for vegetable and flower crops varies widely depending on local and seasonal rates of evapotranspiration, but irrigating a small household plot of 3 m2 should require roughly 0.075 m3 per week (75 l), about half that amount in cooler months, a tiny fraction of average cistern capacity.80 Of course, a market plot might be many times that size—irrigating a garden the size of a Piraeus house plot might require 3–6 m3 per week.81 Looking beyond Attica, however, there is ample archaeological evidence for vineyards, orchards, and commercial gardens in parts of Pompeii beyond the

72 IG II2 2498. The inscription has been dated to the archonship of Archippos, 321/0 or 318/7 BC; see Conwell 1993. 73 Cf. Xen. Poroi 2.6. 74 Garland 1987, 44–45. 75 IG II2 1361, ll. 8–13. 76 Papazarkadas 2011, 196. Leasing of water rights is also attested in the fifth century in the precinct of Neleus and Basile, but there stipulated for the maintenance of an olive nursery (IG I3 84, l. 34). 77 Hermippus fr. 63 = Ath. 1.27e–28a. 78 Garland 1987, 69. 79 Carroll 2003, 29–30. One drachma is generally equivalent to the daily wage for an unskilled labourer, and by the late fourth century BC only those with personal wealth of over 2000 drachmas were among the propertied classes. 80 Klingborg & Finné 2018. 81 Water requirements calculated for the Sonoma Valley, a Mediterranean climate zone, by the University of California Cooperative Extension.

28  Jane Millar Tully reach of the aqueduct, irrigated with rainwater no doubt at a great cost to labour,82 and water infrastructure has been well studied as a proxy for agricultural development in the Roman suburbium.83 Given the economic landscape of Piraeus and the Athenian Long Walls, there is good reason to include irrigation in the uses of wells and cisterns. Looking back to the Metro excavations for the Dimotiko Theatro Station, while this chapter has focused on morphology of cisterns and wells, the ongoing analysis of their contents promises to shed further light on the relationship between water supply and local production. The anaerobic conditions of waterlogged sediments in 28 wells have preserved organic material including leaves and branches that may indicate the presence of trees or small courtyard gardens, as well as numerous seeds and fruits.84 We still have much to learn from the household water supply of Piraeus, not only about water management but also the broader environmental and economic dimensions of the port of Athens, promising future insights into the everyday life of its diverse inhabitants. Conclusion Because of the number of preserved examples, many of which have been stratigraphically excavated in recent years, the cisterns and wells of Piraeus offer significant evidence for the operation of non-monumental water supply in the ancient Mediterranean world. Below the planned streets and uniform housing blocks, subsurface water supply features reveal an array of preferences and strategies adopted by the inhabitants of Classical and Hellenistic Piraeus. As properties expanded and contracted, residents had to reckon with the decisions of past occupants as well as current neighbours in securing their daily water supply. At the edges of the city, the construction of water supply structures interacted with quarrying and cultivation, placing water supply in a broader framework of economic activity and social change. Acknowledgements This contribution owes much to the previous work and ongoing analysis by archaeologists and conservators involved in the Metro excavations under the Ephorate of Antiquities of Piraeus and Islands (Stella Chrysoulaki, Director). Thanks especially to Aimilio Bendermacher-Gerousis, Panagiotis Koutis, and Giorgos Peppas, for their willing collaboration and the contribution of figures and notes on excavations. Any errors are my own.

82 Carroll 2003, 35. 83 On the probable relation of cisterns to gardens and orchards for suburban production, see Marzano 2013, 87. 84 Chrysoulaki & Peppas 2015. Archaeobotanical analysis is in progress for inclusion in the final publication of the Metro excavations.

Household Water, Environment, and Economy in Ancient Piraeus  29 References Antoniou, G., N. Kathijotes, D.S. Spyridakis, & A.N. Angelakis 2014. ‘Historical development of technologies for water resources management and rainwater harvesting in the Hellenic civilizations’, International Journal of Water Resources Development 30(4), 680–693. Avgerinou, P., E.D. Chiotis, S. Chrysoulaki, P. Defteraios, T. Evangelou, N.M. Gigourtakis, G. Kakes, Y. Kourtzellis, P. Koutis, N. Mamassis, M. Pappa, G. Peppas & A. I. Strataridaki 2016. ‘Updated appraisal of ancient underground aqueducts in Greece’, in Underground Aqueducts Handbook, eds. A.N. Angelakis, E. Chiotis, S. Eslamian & H. Weingartner, Boca Raton, 43–62. Axioti, C. 2009. ‘Ανασκαφή οικοπέδου στην πλατεία Τερψιθέας στον Πειραιά᾽, in Από τα Μεσόγεια στον Αργοσαρωνικό: Βʹ Εφορεία Προϊστορικών και Κλασικών Αρχαιοτήτων: Το έργο μιας δεκαετίας, 1994–2003 (Πρακτικά συνεδρίου, Αθήνα, 18–20 Δεκεμβρίου 2003, eds. V. Vassilopoulou & S. Katsarou-Tziveleki, Markopoulo Mesogaias, 489–495. Barboutis, I., P. Koutis, V. Kamperidou & S. Chryssoulaki 2019. ‘Knowledge and quality of woodworking in ancient Greece revealed from a simple finding’, Pro Ligno 15, 144–151. Benisi, M. 2015. ‘Περιγραφή των γεωλογικών και υδρογεωλογικών συνθηκών της στενής και ευρύτερης περιοχής ανάπτυξης των αρχαίων συστημάτων ύδρευσης του Πειραιά, βάσει των ερευνών για τις επεκτάσεις του Μετρό’, presented at Η έρευνα των αρχαίων συστημάτων ύδρευσης του Πειραιά στο πλαίσιο των έργων του ΜΕΤΡΟ. Μια πρώτη θεώρηση, 15 Μαΐου 2015, Εθνικό Μετσόβιο Πολυτεχνείο. Available at: https://efadyat. wordpress.com. Biernacka-Lubanska, M. 1977. ‘A preliminary classification of Greek rainwater intakes’, Archeologia 28, 26–36. Brinker, W. 1990. Wasserspeicherung in Zisternen. Ein Beitrag zur Frage der Wasserversorgung früher Städte. Leichtweiss-Institut für Wasserbau der Technischen Universität Braunschweig Mitteilungen 109, Braunschweig. Camp, J. 1977. The water supply of Ancient Athens from 3000 to 86 BC, Ph.D. thesis, Princeton University. Camp, J. 1982. ‘Drought and famine in the 4th Century B.C.’, in Studies in Athenian architecture, sculpture and topography. Presented to Homer A. Thompson. Hesperia Supplements 20, Athens, 9–17. Carroll-Spillecke, M. 1989. Kēpos: der antike griechische Garten. Wohnen in der klassischen Polis 3, Munich. Carroll, M. 2003. Earthly paradises: Ancient gardens in history and archaeology, London. Carusi, C. 2019. ‘The quarries of Attica revisited’, in From document to history: Epigraphic insights into the Greco-Roman world, eds. C.F. Noreña & N. Papazarkadas, Boston, 56–69. Chiotis, E.D. 2019. ‘Integrating geology into archaeology: The water supply of Piraeus in antiquity’, Journal of Greek Archaeology 4, 337–377. Chiotis, E.D. & L. Chioti 2012. ‘Water supply of Athens in antiquity’, in Evolution of water supply through the millennia, eds. A.N. Angelakis, L.W. Mays, D. Koutsoyiannis & N. Mamassis, London, 407–442. Chrysoulaki, S., T. Evangelou, P. Koutis & G. Peppas 2017. ‘Bringing to light ancient water supply structures: The METRO rescue excavations in Piraeus’, in Cura aquarum in Greece: Proceedings of the 16th International Conference on the history of water management and hydraulic engineering in the Mediterranean region, Athens, Greece, 29–30 March 2015, ed. K. Wellbrock, Siegburg, 417–442.

30  Jane Millar Tully Chrysoulaki, S. & G. Peppas 2015. ‘Ευρήματα της ανασκαφής᾽, presented at Η έρευνα των αρχαίων συστημάτων ύδρευσης του Πειραιά στο πλαίσιο των έργων του ΜΕΤΡΟ. Μια πρώτη θεώρηση, 15 Μαΐου 2015, Εθνικό Μετσόβιο Πολυτεχνείο. Available at: https:// efadyat.wordpress.com. Conwell, D.H. 1993. ‘Topography and toponyms between Athens and Piraeus’, Journal of Ancient Topography 3, 49–62. Conwell, D.H. 2008. Connecting a city to the sea: The history of the Athenian long walls, Leiden. Crouch, D.P. 1993. Water management in ancient Greek cities, New York. Curtius, E. & J.A. Kaupert, eds. 1881. Karten von Attika, Berlin. Daux, G. 1962. ‘Chronique des fouilles et découvertes archéologiques en Gréce en 1961’, BCH 86(2), 629–632. Dodwell, E. 1819. A classical and topographical tour through Greece: During the years 1801, 1805, and 1806, London. Dörpfeld, W. 1884. ‘Ein antikes Bauwerk im Piraeus’, Mitteilungen des Deutschen Archäologischen Instituts 9, 279–287. von Eickstedt, K.-V. 1991. Beiträge zur topographie des antiken Piräus. Βιβλιοθήκη της εν Αθήναις Αρχαιολογικής Εταιρείας 188, Athens. Finné, M., M. Bar-Matthews, K. Holmgren, H. S. Sundqvist, I. Liakopoulos & Q. Zhang 2014. ‘Speleothem evidence for late Holocene climate variability and floods in southern Greece’, Quaternary Research 81(2), 213–227. Finné, M., J. Woodbridge, I. Labuhn & C.N. Roberts 2019. ‘Holocene hydro-climatic variability in the Mediterranean: A synthetic multi-proxy reconstruction’, The Holocene 29(5), 847–863. Fragkopoulou, F. 2015. ‘Piraeus: Beyond ‘known unknowns’, in Aegis: Essays in Mediterranean archaeology: Presented to Matti Egon by the scholars of the Greek Archaeological Committee UK, eds. D. Evely & Z. Theodoropoulou-Polychroniadis, Oxford, 131–136. Garland, R. 1987. The Piraeus: From the fifth to the first century BC, London. Garnsey, P. 1988. Famine and food supply in the Graeco-Roman world: Responses to risk and crisis, Cambridge. Gill, D.W.J. 2006. ‘Hippodamus and Piraeus’, Historia 55(1), 1–15. Glaser, F. 1983. Antike Brunnenbauten (κρηναι) in Griechenland, Vienna. Goiran, J.-P., K.P. Pavlopoulos, E. Fouache, M. Triantaphyllou & R. Etienne 2011. ‘Piraeus, the ancient island of Athens: Evidence from Holocene sediments and historical archives’, Geology 39(6), 531–534. Grigoropoulos, D. 2016. ‘The Piraeus from 86 BC to late antiquity: Continuity and change in the landscape, function, and economy of the port of Roman Athens’, BSA 111, 239–268. Hilditch, M.H. 2015. Kepos: Garden spaces in ancient Greece: Imagination and reality, Ph.D. thesis, University of Leicester. Hoepfner, W. & E.-L. Schwandner 1994. Haus und Stadt im klassischen Griechenland. Wohnen in der klassischen Polis 1, Munich. Horden, P. & N. Purcell 2000. The corrupting sea: A study of Mediterranean history, Oxford. Judeich, W. 1931. Topographie von Athen. Handbuch der Altertumswissenschaft 2, Munich. Klingborg, P. 2017. Greek cisterns: Water and risk in ancient Greece, 600–50 BC, Ph.D. thesis, Uppsala University. Klingborg, P. 2023. ‘Wells and cisterns in Greek literature’, in Going against the flow. Wells, cisterns and water in ancient Greece, ed. P. Klingborg. ActaAth-8°, 23, Athens, 161–178.

Household Water, Environment, and Economy in Ancient Piraeus  31 Klingborg, P. & M. Finné. 2018. ‘Modelling the freshwater supply of cisterns in ancient Greece’, Water History 10, 113–31. Kouli, K. 2012. ‘Vegetation development and human activities in Attiki (SE Greece) during the last 5,000 years’, Vegetation History and Archaeobotany 21(4), 267–278. Kouli, K., M. Triantaphyllou, K. Pavlopoulos, T. Tsourou, P. Karkanas & M.D. Dermitzakis 2009. ‘Palynological investigation of Holocene palaeoenvironmental changes in the coastal plain of Marathon (Attica, Greece)’, Geobios 42(1), 43–51. Koutis, P. & A. Bendermacher-Gerousis 2015. ‘Τα συστήματα ύδρευσης’ presented at Η έρευνα των αρχαίων συστημάτων ύδρευσης του Πειραιά στο πλαίσιο των έργων του ΜΕΤΡΟ. Μια πρώτη θεώρηση, 15 Μαΐου 2015, Εθνικό Μετσόβιο Πολυτεχνείο. Available at: https://efadyat.wordpress.com. Koutsoyiannis, D., N. Zarkadoulas, A.N. Angelakis & G. Tchobanoglous 2008. ‘Urban water management in ancient Greece: Legacies and lessons’, Journal of Water Resources Planning and Management 134, 45–54. Koutsoyiannis, D. & Ν. Mamassis 2017. ‘The water supply of Athens through the centuries’, in Cura aquarum in Greece: Proceedings of the 16th International Conference on the history of water management and hydraulic engineering in the Mediterranean region, Athens, Greece, 29–30 March 2015, ed. K. Wellbrock, Siegburg, 31–42. Krasilnikoff, J. & A.N. Angelakis 2019. ‘Water management and its judicial contexts in ancient Greece: A review from the earliest times to the Roman period’, Water Policy 21(2), 245–258. Langdon, M. 2000. ‘The quarries of Peiraieus’, ΑΔ 55, Α’, 235–250. Lewis, D. 1990. ‘Public property in the city’, in The Greek city from Homer to Alexander, eds. O. Murry & S. Price, Oxford, 245–264. Manning, S.W. 2018. ‘Some perspectives on the frequency of significant, historically forcing drought and subsistence crises in Anatolia and region’, in Water and power in past societies, ed. E. Holt, Albany, 279–295. Mariolopoulos, E.G. 1962. ‘Fluctuation of rainfall in Attica during the years of the erection of the Parthenon’, Geofisica Pura e Applicate 51, 243–262. Marzano, A. 2013. ‘Agricultural production in the hinterland of Rome’, in The Roman agricultural economy: Organization, investment, and production, eds. A. Bowman & A. Wilson, Oxford, 86–106. Norström, E., C. Katrantsiotis, M. Finné, J. Risberg, R. H. Smittenberg & S. Bjursäter 2018. ‘Biomarker Hydrogen Isotope Composition (δD) as Proxy for Holocene Hydroclimatic Change and Seismic Activity in SW Peloponnese, Greece’, Journal of Quaternary Science 33(5), 563–574. Panagos, C. 1968. Le Pirée: Étude économique et historique depuis les temps anciens jusqu’à la fin de l’Empire Romain, Athens. Panessa, G. 1981. ‘Oscillazioni e stabilità del clima nella Grecia Antica: Introduzione ad una riconstruzione paleoclimatologica’, Annali Della Scuola Normale Superiore Di Pisa 11(1), 123–158. Papadopoulou, C. 2015. ‘New discoveries in the Piraeus’, Archaeological Reports 61, 56–64. Papazachos, B.C. 1990. ‘Seismicity of the Aegean and surrounding area’, Tectonophysics 178(2), 287–308. Papazarkadas, N. 2011. Sacred and public land in ancient Athens, Oxford.

32  Jane Millar Tully Peppas, G. Forthcoming. ‘Η διαχρονική εξέλιξη των κατασκευών ύδρευσης στον αρχαίο Πειραιά: υδροληψία, διαχείριση και κατανάλωση του νερού’, presented at Επιστημονική Διημερίδα Αρχαιότητες σε Τροχιά, Μουσείο Αρχαίας Αγοράς Θεσσαλονίκης, 11–12 Απριλίου 2019. Sallares, R. 1991. The ecology of the ancient Greek world, Ithaca. Shipley, G. 2005. ‘Little boxes on the hillside: Greek town planning, Hippodamos, and polis ideology’, in The imaginary polis, January 7–10, 2004, ed. M.G. Hansen, Copenhagen, 335–403. Steinhauer, G. 2012. ‘Ancient Piraeus: The city of Themistokles and Hippodamus’, in Piraeus: Centre of shipping and culture, eds. G. Steinhauer, M.G. Malikouti & V. Tsokopoulos, Piraeus, 1–123. Steinhauer, G. 2021. ‘Piraeus: Harbors, navy, and shipping’, in The Cambridge companion to ancient Athens, eds. J. Neils & D.K. Rogers, Cambridge, 231–243. Stroszeck, J. 2017. ‘Wells in Athens: The contribution of the Kerameikos wells’, in Cura aquarum in Greece: Proceedings of the 16th International Conference on the history of water management and hydraulic engineering in the Mediterranean region, Athens, Greece, 29–30 March 2015, ed. K. Wellbrock, Siegburg, 43–88. Stroszeck, J. 2021. ‘Water and water management,’ in The Cambridge companion to ancient Athens, eds. J. Neils & D.K. Rogers, Cambridge, 110–123. Stroszeck, J. 2023. ‘The cisterns of the Athenian Kerameikos. Distribution and recent documentation,’ in Going against the flow. Wells, cisterns and water in ancient Greece, ed. P. Klingborg. ActaAth-8°, 23, Stockholm, 91–112. Stoumbos, G. 2015. ‘Τεχνικογεωλογικά χαρακτηριστικά της ‘Μάργας του Πειραιά’ – Συγκριτική ποιοτική ανάλυση της συμπεριφοράς της στις αρχαίες και σύγχρονες υπόγειες κατασκευές της περιοχής του σταθμού ‘Δημοτικό Θέατρο’,’ presented at Η έρευνα των αρχαίων συστημάτων ύδρεσης του Πειραιά στο πλαίσιο των έργων του ΜΕΤΡΟ: Μια πρώτη θεώρηση, 15 Μαΐου 2015, Εθνικό Μετσόβιο Πολυτεχνείο. Available at: https:// efadyat.wordpress.com. Tölle-Kastenbein, R. 1985. ‘Der Begriff Krene’, Archäologischer Anzeiger 100, 451–470. Tölle-Kastenbein, R. 1990. Antike Wasserkultur, München. Travlos, J.N. 1988. Bildlexikon zur Topographie des antiken Attika, Tübingen. Triantaphyllou, M.V., K. Kouli, T. Tsourou, O. Koukousioura, K. Pavlopoulos & M.D. Dermitzakis 2010. ‘Paleoenvironmental changes since 3000 BC in the coastal marsh of Vravron (Attica, SE Greece)’, Quaternary International 216(1), 14–22. Walbank, M.B. 1983. ‘Leases of sacred properties in Attica, part I’, Hesperia 52(1), 100–135. Wassenhoven, M.-E. 2018. ‘The diachronic development of the natural and man-made landscape between Athens and Thene Sea’, in Ήρως κτίστης: μνήμη Χαράλαμπου Μπούρα, eds. Μ. Κορρές, Σ. Μαμαλούκος, Κ. Ζάμπας & Φ. Μαλλούχου-Tufano, Athens, vol. 1, 49–60. Weiberg, E., A. Bevan, K. Kouli, M. Katsianis, J. Woodbridge, A. Bonnier, M. Engel, M. Finné, R. Fyfe, Y. Maniatis, A. Palmisano, S. Panajiotidis, C. N. Roberts & S. Shennan 2019. ‘Long-term trends of land use and demography in Greece: A comparative study’, The Holocene 29(5), 742–760.

3

Social Stratification and Water Sharing on Late-Hellenistic Delos Patrik Klingborg

Introduction Access to water is essential for human societies. Due to this, water sources and use points form important nodes in urban and rural landscapes, shaping, and, in turn, being shaped by, human behaviour. Yet, studies of the water supply in ancient Greece, and especially water usage, have focused almost exclusively on monumental structures, such as public fountains and aqueducts. In practice, the vast majority of the population relied on less impressive wells and cisterns. Current knowledge about water collection and usage, and how this affected society, is therefore severely skewed. This chapter provides a spatial analysis of water access in an ancient Greek urban landscape using primarily cisterns and wells through a case study of insulae II–IV and VI in Le Quartier du Théâtre on the island of Delos. The aim is to investigate who had access to water, how and from which source(s)? What does this reveal about the hydraulic relationships in the area? And how did the situation create and maintain social inequalities? To answer these questions, GIS analyses were performed based on a map of ancient Delos provided by the École française d’Athènes. Throughout the study, it was a methodological assumption that water sources would be shared in Greek urban contexts. The structures were plotted as ‘houses’ and ‘shops’, together forming ‘complexes’, using the conventional room numbering. It should be noted though that, despite these terms being widely used, the de facto function of most domestic units on Delos is unknown.1 A database of water sources in individual houses was compiled following Trümper’s Wohnen in Delos in combination with material in Exploration Archéologique de Délos 8, Karvonis’ L’eau dans la ville hellénistique de Délos and Klingborg’s Greek cisterns. Water and risk in ancient Greece 600–50 BC.2 Délos 8 and Karvonis provided further information on sources in shops. The 1 ‘Shops’ and ‘houses’ are based on how the material has been treated in earlier studies. ‘Shop’ (boutique in Délos 8; Laden in Trümper 1998) refers to a small, isolated unit (usually a single room) connected to a larger house, but it did not necessitate any commercial activity. Often a single house had several shops. ‘House’ (maison in Délos 8; Haus in Trümper 1998) refers to larger habitation units with several rooms and usually a courtyard. These are given a capital letter in Délos 8. ‘Complex’ refers to whole structures forming a unit, including houses and shops. 2 Trümper 1998; Délos 8; Karvonis 2000; Klingborg 2017. DOI: 10.4324/9781003268222-3

34  Patrik Klingborg

Figure. 3.1 Delos with neighbouring islands and their location in Greece. Map by Author.

individual water sources were denoted by Insula, House, and Room, e.g., ‘Cistern III B-a’, with a final number added when several sources of the same type were located in the same room. Additionally, the island’s three major public water sources were included: the Minoe fountain, the Inopos reservoir, and the Theatre cistern.3 Delos The island of Delos is centrally located in the Cyclades, 2.5 km from Mykonos and 500 m from Rheneia (Figure 3.1). With an area of 360 ha (5 × 1.3 km), it is a relatively small island.4 The highest peak, Mount Cynthus, reaches about 110 m above sea level.5 Geologically, Delos is composed mainly of granite.6 The climate is Mediterranean semi-arid, with a modern average precipitation of 350 mm per year, concentrated during the fall and spring; about 85% of the precipitation takes place between October and March. There is also considerable annual variation, ranging from under 150 mm to almost 700 mm per year.7 It has been argued that this pattern closely resembles precipitation patterns during antiquity, although 3 Citerne du Théâtre, see Délos 42a-b; the Minoe fountain, see Délos 4, 103–119; Glaser 1983, 15–16; 2000, 419; the Inopos reservoirs, see Siard & Braunstein 2006; Finker & Moretti 2007. 4 Moretti & Fincker 2011, 159. 5 Desruelles 2001, 554 (112 m); Desruelle 2007, 163 (114 m); Brunet et al. 2011, 697 (112 m); Moretti & Fincker 2011, 159 (113 m); Desruelles & Fouache 2015, 205 (114 m). 6 Desruelles et al. 2009, 21. 7 Desruelles 2001, 553; Desruelles & Fouache 2015, 204.

Social Stratification and Water Sharing on Late-Hellenistic Delos  35 recent climatological studies in Greece suggest fluctuations in temperature and precipitation over time.8 In general, the geology of the island is favourable to the infiltration of rainfall. This is usually absorbed by fissures in the rocky surfaces, in superficial soil deposits at hillslopes and depressions, or collected in ephemeral streams. Of these, the rock fissures are the most effective in replenishing the aquifers due to almost instant absorption.9 Rain falling on the soil deposits is partially or completely lost to evaporation. Instead, the aquifers, especially in depressions, are refilled through temporary streams, making them the hydrologically most important areas on Delos. The most significant stream is the Inopos. This minor ephemeral water course was already, by Strabo, described as small.10 With a watershed of only 30.5 ha and a length of 200 m, its pre-Hellenistic output has been calculated as no more than 0.15 m3/s at its mouth.11 Despite its small size, the Inopos was portrayed in antiquity as a continuation of the Nile.12 Together, these phenomena allow for relatively high-water levels, and even flooding, during rainy periods. Despite this, the small and loosely connected water tables do not allow for perennial springs. Furthermore, the aquifers are too small to allow storage from one year to another. The lack of precipitation in one year is therefore not compensated by the previous year being wetter.13 Periods of groundwater depletion still occur semi-regularly today without extensive usage, suggesting considerably more serious problems in antiquity.14 Historically, the island was intermittingly occupied from prehistoric times until the first steps towards a city were taken during the eighth century BC. The island first gained importance as a religious site during the Archaic period (700–480 BC), and thereafter in the fifth century BC as the centre of the Delian League. While the city itself remained small until the mid-Hellenistic period, various rulers had already constructed impressive monuments in and around the sanctuaries before this. The economic importance of the island exploded after 167 BC when the Romans declared it tax free and its population grew to over 10,000 people.15 The prosperity soon ended, however, and the town withered away after a sack by Mithridates in 88 BC and a large pirate raid in 69 BC, in combination with the emergence of new trade centres, such as Puteoli in Italy. Consequently, most of the domestic architecture dates to between the mid-second and early first century BC, largely preserving the urban fabric of the Late Hellenistic period. The island was finally abandoned in the sixth century AD.16

  8 Desruelles et al. 2003, 255; Finné & Labuhn 2023.   9 Desruelles & Fouache 2015, 207–208. 10 Strabo 10.5.2. 11 Desruelles & Fouache 2015, 208. 12 Strabo 6.2.4. 13 Desruelles 2015. 14 Desruelles et al. 2003, 256. 15 Desruelles & Fouache 2015, 203. 16 Desruelles & Fouache 2015, 205.

36  Patrik Klingborg The Water Supply In practice, the water supply on Delos was composed of four types of water sources: wells, cisterns, infiltration wells, and fountains. In total 147 hydraulic installations have been identified on the island. Out of these, 79 were in the Quartier du Théâtre.17 Karvonis attributed this large figure to the public supply being insufficient in the Hellenistic period, prompting the extensive construction of private wells and cisterns.18 As in other Greek cities, water sources could be fully publicly available, public with restricted use, or privately available. The potential use of a water source differed depending on the type. Wells are artificial water sources—usually shafts—dug into the ground until they reach the water table, from which they are fed. This means that water is constantly ‘produced’, while the water level may rise or recede over time. The water tends to be drawn from one or, in rare cases, several delimited horizontal openings. Wells may be interconnected with other wells or cisterns, for instance, for receiving overflow.19 Cisterns, however, are statically situated waterproof containers constructed above or below ground to store water. Cisterns hold water that is received from an external source—usually rainwater—and are not intended to receive a constant inflow or to facilitate constant outflow. Installations with a constant inflow and outflow may instead be referred to as reservoirs. On Delos also a hybrid between wells and cisterns existed, sometimes called ‘infiltration wells’.20 These collected rainwater as cisterns but are not waterproof. Therefore, they could draw water from an aquifer during times of lesser rainfall, as well as replenish the aquifer after heavy rains.21 In this context, ‘fountain’ refers to a water source called κρήνη in ancient Greek.22 While a krene often looked like what is called a fountain today the two are not identical. Instead, a krene is an artificially modified water source with augmentations such as a basin, roof, or staircase, although little monumentality was needed. Whether the water originated from a spring, stream, or well was not a concern. Publicly available sources, usually fountains in public areas, could be used by the population for everyday needs.23 This contrasts with restricted public sources. For instance, to access water sources in sanctuaries it would be necessary to perform ritual purification(s), often wait a specified number of days, and adhere to local purity regulations.24 These regulations would make it difficult to use water 17 Desruelles 2001, 563. For more on the water on Delos, see also Brunet 2001; 2008. 18 Karvonis 2000, 37. 19 Not to be confused with qanats, see Hodge 2000; Wilson 2008, 290–293. For the definition of a qanat and the increased use of the term, see Chiotis 2018. 20 Desruelles et al. 2003, 257; Desruelles & Fouache 2015; Karvonis 2023, 79, 84–85. 21 Desruelles & Fouache 2015, 209. 22 Wycherley 1937; Tölle-Kastenbein 1985. 23 E.g., the fountain on the North Hill in Olynthos (Olynthus 12, 98–103). 24 Ehrenheim, Klingborg & Frejman 2019. For an example of regulations concerning purifications, see CGRN 212 (= IvP II 255) concerning the cult of Athena Nikephoros at Pergamon; for specific

Social Stratification and Water Sharing on Late-Hellenistic Delos  37 sources in a sanctuary for everyday needs, even if technically allowed. Finally, private sources were found in domestic settings or workshops, placing them under the control of the property’s owner or tenant. Public Water Sources on Delos Three major public water sources existed on Delos:25 the Minoe fountain, the Inopos reservoir, and the Theatre cistern.26 Additionally, as yet unidentified, krenai are known from epigraphic evidence.27 Constructed in the sixth century BC, the Minoe is the island’s oldest public source. It is a common shaft fountain utilising groundwater, located at the island’s religious centre, between the Stoa of Antigonos (c. 250 BC), and the late-second-century BC Agora of the Italians.28 The structure measures 10.7 × 11.6 m, with a 4 × 3.75 m shaft collecting water. Access was permitted by a staircase on the southern side, towards the sanctuaries. Notably the Minoe predates by a considerable margin both most of the monumental architecture in the area as well as the second-century BC population influx. It is therefore likely that its function changed over time as its environment was transformed. Michèle Brunet, observing the fountain’s location near the sanctuaries, argued that it was primarily intended for worshipers.29 However, the krene was not found within the sacred enclosure of any of the sanctuaries.30 In addition to this, a fifthcentury BC inscription regulating its use prohibits washing, diving, and swimming. While fragmentary, these regulations resemble those for fountains in secular rather than cultic contexts. For example, a regulation preserved from the sanctuary of Asclepius at Kos instead prohibits the throwing of cakes and other things into the water of a krene.31 Furthermore, sixth-century BC fountains like the Minoe have primarily been connected to tyrants providing water for the population, not to religious sites. Considering the lack of cisterns on Delos before the fourth century BC, it seems likely that the Minoe was intended to supply the community.32 How applicable this function would be during the second century BC is difficult to appraise. The Inopos was located at 25 m above sea level on the slopes of Mount Cynthus. Functioning like an infiltration well, it was principally formed by a dam dated to c.

local purity regulations, see CGRN 90 (= IG XII.1 677) which forbids the use of sandals, or anything made of pig in the sanctuary of Alektrona at Ialysos on Rhodos. 25 Karvonis 2000, 13. See also the krene by the Prytaneion (ibid., 9). 26 Brunet 2011, 698–701. The traditional names are kept although the Inopos reservoir and Theatre cistern are infiltration wells following the definitions given here. 27 Karvonis 2000, 13. Vallois (1944, 190–192) recorded a fountain close to the Theatre, but this is no longer visible according to Karvonis (2000, 16). 28 Délos 4, 103–119; Glaser 1983, 15–16, no. 8; Brunet 2011, 698–699. On the type, see Glaser 1983, nos. 1–17. 29 Brunet 2011, 698–699. 30 The second-century BC relief dedicated to the nymphs of the Minoe does not ascertain that the fountain was sacred or within a sacred sphere. 31 IG XII.4 285. 32 As argued by Karvonis 2000, 19.

38  Patrik Klingborg 400 BC blocking its homonymous stream. Additionally, the local water table contributed to the supply. In total, the structure was 40 m long and 8–10 m wide with a volume of 1700 m3. Epigraphically attested regulations, similar to those from the Minoe, mentions an Inipophylax, a guard of the Inopos, suggest that the source needed oversight.33 Hence, the Inopos reservoir was likely primarily intended as a water source, not to protect areas below it from flooding.34 The Theatre cistern was constructed during the early third century BC. Measuring 25 m long, it had a capacity of 1150 m3. A cover, resting on eight arches, protected the water while access was facilitated through puteals. Since the Theatre cistern was not equipped with waterproof lining, the structure functioned as an infiltration well, collecting rain from the nearby theatre’s cave and the water table. Whether intentional or not, as a result, the water collected during heavy rains replenished the ground water in the surrounding Quartier du Théâtre over time. Private Water Sources The private water supply was dominated by smaller wells and cisterns. In the Quartier du Théâtre these were unevenly distributed (Figure 3.2).35 Only 13% of the shops had a water source, all of them wells,36 while 78% of the houses were equipped with at least one water source.37 Of the houses, 21 (c. 50%) had only one water source. Thirteen houses had two sources, while two houses had three. Only one house had four water sources, House II B.38 As a point of comparison, the most extreme example of water hoarding on Delos was the Maison du Lac, situated just outside the Quartier du Théâtre, with two cisterns and three wells.39 In general, having multiple sources suggests that the water supply was viewed as important. About 30% of the houses in the investigated insulae had both a cistern and a well. How these installations functioned in tandem is unknown. In House III K, the well’s opening exhibits considerably more wear than that of the cistern, suggesting uneven use. However, although both the well and cistern openings were built into walls, making refurbishment unlikely, the complex history of this house makes it uncertain whether both installations were built at the same time or

33 IG XI,2 144; Brunet 2011, 699–700; Siard 2006. 34 Desruelles 2001, 563. 35 The identification of the water sources is not always certain as the material is based on sometimes contradictory publications. 36 Two excavated shops on Delos had cisterns. Shop Θ in the Stadium quarter, presumably the remnant of an older house, while the one in l´îlot XIII in the Theatre quarter has been linked to commercial activity (Karvonis 2000, 76). 37 37 out of 48 houses. Eight houses lack a water source. This is slightly below the average for Delos, where 21% of the houses lacked a water source: see Trümper 1998, 28. Of the 38 shops in the investigated area, only five had a water source. Shops 10, 27b, 45 on the Rue du Théâtre, Shops 4 on Rue 5. See also Shops 28, 32, 40a, 40b on the Rue du Théâtre outside of the investigated area. 38 15 houses (31%): Houses II A, II B, II C, III D, III E, III I, III K, III Q, III R, III V, IV B, VI J, VI I, VI L, VI N. 39 Trümper 1998, 29, n. 150.

Social Stratification and Water Sharing on Late-Hellenistic Delos  39

Figure 3.2 Number of water sources per house and shop. Map by Author.

whether they originally even belonged to the same property.40 Similarly, in House VI I, Maison du Dionysos, the well opening is more worn than that of the cistern, but the relatively mobile puteals used here could have been replaced at some point. In House II A, which had both a well and a cistern, only the cistern was equipped with a drawing mechanism, presumably because it was used more frequently. Overall, there are no areas in the insulae which had fewer water sources than others, except perhaps the eastern half of Insula II. Of the eastern houses without a water source there, House II F is one of the largest on Delos.41 Its current condition seems to be a result of a never completed ennobling reconstruction without the addition of a cistern.42 House II E, however, had access to a well in Shop 3 on the Rue Supérieure du Théâtre. House II D did not have a water supply. It is uncertain if this indicates any difference between the two sides of the insula. It is also notable that the best-equipped house, II B, is the closest to the neighbourhood’s public source, the Theatre cistern. House size, however, did matter, with larger houses usually having more sources than smaller ones (Table 3.1). In the first quartile, 19 water sources were spread over ten houses, including seven with more than a single source. In the second quartile, only 12 sources were found in total. Just two houses had more than one source,

40 Trümper 1998, 276. 41 No. 17 of 91 in Trümper 1998. 42 Trümper 1998, 265. Possibly a well was planned.

40  Patrik Klingborg Table 3.1  Number of sources based on size, from the largest to the smallest house No.

House

Sources

No.

 1 House II B  2 House VI I  3 House III I  4 House III K  5 House III N  6 House II C  7 House II F  8 House II A  9 House II E 10 House VI N 11 House VI H 12 House III T 1st quartile

4 2 2 2 1 2 0 2 0 2 1 1

13 House IV B 14 House III J 15 House IV A 16 House III S 17 House VI O House III A 18 19 House III V 20 House III O House II D 21 22 House III L 23 House III X 24 House VI J 2nd quartile

3 1 1 1 1 1 1 1 0 0 0 2

No.

Sources

No.

Sources

1 1 2 2 2 1 1 2 2 0 0 1

37 House III R 38 House VI B 39 House III P 40 House III G 41 House III B 42 House III E 43 House VI D 44 House VI C 45 House III Y 46 House III C 47 House VI K 48 House IV C 4th quartile

House

25 House VI M 26 House III M 27 House III Q 28 House III U 29 House VI L 30 House VI A 31 House VI E 32 House III D 33 House VI F 34 House VI G 35 House III F 36 House III H 3rd quartile

House

House

Sources

3 1 1 0 1 2 1 0 0 0 1 1

seven had one, and three had none. The third quartile was slightly better equipped with 15 water sources, including five houses having two installations. With only eleven installations, the bottom quartile has the lowest number of sources. The existence of multiple water sources in larger houses has been explained in the past by properties being combined, assimilating pre-existing water sources.43 There are, however, exceptions, such as Houses II A and III U which show no sign of being formed by several earlier buildings. Moreover, some small houses had multiple water sources, for example, House III R, with three water sources, and III D, III E, and VI F with two water sources each. Yet, the number of water sources per square meter is considerably higher in the smaller houses when compared to the larger ones (Figure 3.3). By this measure, for example, the smaller Houses III R and III E had more than three times as many sources than the large House II A. If house size is correlated to the number 43 Délos 8, 39–40; Karvonis 2000, 86; 2003, 87.

Social Stratification and Water Sharing on Late-Hellenistic Delos  41

Figure 3.3 Number of water sources based on size, from the largest to the smallest house. Map by Author.

of inhabitants, including dependants, then these smaller houses had considerably better water access than most of the larger and more prestigious ones. The larger number of sources in relation to house size also suggests that the owners or occupants were more concerned with, and willing to invest into, a sufficient supply, relatively speaking. Looking at the preferred type of water source, about two-thirds of the houses were equipped with wells, half of them with cisterns, and one-third had both types (Figure 3.4).44 Four houses had two cisterns, and two of these did not have access to a well.45 Two houses had two wells, but also had cisterns. Besides wells being somewhat more popular than cisterns, there is no pattern in terms of preferences. For example, in terms of size, 85% of the cisterns are in houses larger than 200 m2, while no house smaller than 100 m2 had a cistern. Similarly, 80% of the wells are in houses larger than 200 m2 and only two (7%) in houses smaller than 100 m2.

44 Houses with a water source in Insula II (3/6 houses: 50%), Insula III (18/24 houses: 75%), Insula IV (3/3 houses: 100%), and Insula VI (13/15 houses: 87%). Houses with a well in Insula II (3/6 houses: 50%), Insula III (13/24 houses: 54%), Insula IV (2/3 houses: 66%), and Insula VI (10/15 houses: 66%). Houses with a cistern in Insula II (3/6 houses: 50%), Insula III (10/24 houses: 42%), Insula IV (2/3 houses: 66%), and Insula VI (9/15 houses: 60%). 45 Houses IIB, IIIR, IIIU, and VIF. For other houses on Delos with two cisterns, see e.g., Maison des Dauphins, the Hôtellerie, and Maison du Lac.

42  Patrik Klingborg

Figure 3.4 Types of water sources in the houses. Map by Author.

Choice of Water Source The large number of houses with a private water source shows that this was a prioritised investment. This is unsurprising since these households would hardly function without reliable water access, especially if the occupants aspired to higher social status. Houses without water must have secured access through external sources. Yet, the uneven distribution of wells and cisterns highlights that the water supply was the result of processes where individuals took decisions based on several variables. Four factors, in particular, must have been important: the volume provided in relation to seasonality, the cost of construction and maintenance, the space required, and personal preferences. Available Volume and Seasonality Wells and cisterns ensure water access in distinctly different ways; the former produce water continuously,46 while the latter store a finite volume collected at irregular intervals. The exact volume of water that wells on Delos made available is unknown as this depends on the static head (the volume when no more water can flow in), the flow rate (how quickly water replenishes), and the drawing capacity. Too many wells in an area, especially if used intensively, can diminish the available groundwater. With 36 wells within an area of 1.25 ha and an average distance 46 Tölle-Kastenbein 1985, 425–453.

Social Stratification and Water Sharing on Late-Hellenistic Delos  43 between them of only 13.5 m, this probably occurred in the Quartier du Théâtre.47 The dry climate and generally poor water table further suggest that the wells were not particularly productive. However, saltwater intrusion, often viewed as a potential issue on Delos was probably only a minor problem according to Desruelles et al. Their measurements show a salinity of no more than 0.2% in wells in the lower areas and only about 0.1% in the higher ones, despite a two-meter sea-level rise since Hellenistic times.48 In terms of volume, a well could probably provide at most a few litres per minute, although not continuously. A previously suggested rate of 7 l per minute allows for 10 m3 per day.49 Considering that wells were unlikely used around the clock, we should envision considerably lower volumes. The area’s total aquifer volume, about 7000 m3 in an average year, was another limiting factor allowing for no more than 500 l/day per well on average. This figure drops to almost nil during dry years.50 The issue of water availability for the wells is also stressed by modern measurements, as at least two wells in the Quartier du Théâtre were found to be dry in June 2001, despite negatable stress on the aquifer.51 The volume made available by a cistern is primarily limited by the precipitation and rain harvesting area (usually a building’s roof), while the size of the cistern minus the water used over time sets a limit for the availability at a specific point. On Delos this means a potential water collection of around 100 m3 per year in most cases.52 Using a common consumption figure of 20 l/day per individual, this would suffice for 14 persons.53 Both cisterns and wells are affected by seasonal changes. This is important as the peak water need during summer coincided with its lowest availability, as almost all precipitation falls between October and March. In wells the water level tends to sink during dry periods and, at times, dries out completely. Cisterns are replenished primarily during the winter and the water availability during the summer depends on the occupants’ consumption pattern throughout the year.54 The different reactions to seasonal changes mean that wells presumably produced more water over the year but were less reliable throughout the cycle. Houses with access to both cisterns and wells would benefit from the combination. Possibly wells were used to manually fill cisterns during winter as those with a documented 47 The shortest distance between two wells is 7.4 m (Wells VI B-d and VI D-b) and the longest distance is 34 m (Wells RsdT 3 and III C-e). While Strabo 3.5.7–8 suggests that the effect of too many wells in an area was known, it is unlikely that the hydro-technical mechanisms behind it were understood. 48 Desruelles 2001, 566–567; Desruelles et al. 2003, 261. See also Desruelles et al. 2012. There is no fixed safe level for drinking water salinity, but more than 0.1% is often perceived as poor quality today. 49 Uhl et al. 2009, 41. This can be compared to a continuous bucket chain, known from antiquity but impractical in domestic contexts, which provides 15–70 l per minute (Wagner & Lanoix 1959, 122; Oleson 2000, 251–263). 50 Desruelles et al. 2003, 260. 51 These wells may originally have been deeper, thus providing water during less optimal circumstances. See Desruelles 2001, 564, fig. 9; Desruelles et al. 2003, 258, fig. 5, table 1. 52 Cistern II A-d: 73 m3; Cistern III N-f: 92 m3; Cistern VI I-c: 115 m3. See Klingborg 2017, 81. 53 Klingborg & Finné 2018. 54 Klingborg & Finné 2018.

44  Patrik Klingborg volume were greatly over-dimensioned in relation to available rain harvesting surfaces.55 For example, Cistern VI I-c had a volume of almost 190 m3 and a rain harvesting surface with the potential of providing 115 m3 per year based on modern average precipitation.56 Since cisterns usually provide twice as much water annually compared to their volume when accounting for consumption, the cistern could have received about 380 m3 per year. As smaller cisterns were safer in terms of structural integrity and less expensive to construct, it is likely that Cistern VI I-c was intended to be partly filled manually from the nearby well, although this would contribute to depleting the relatively small aquifer more quickly. The volume of the cisterns, in relation to the water harvesting area also suggests that the cisterns were primarily intended for use during the summer, since they did not rely on water being drawn during winter to fulfil their potential. This runs counter to cisterns in most other Greek cities, where they would be most efficient during winter and had to be used during that time for overflow not to happen.57 Cost Cost also likely influenced the choice of water supply, as a well or cistern represented a considerable investment, although de facto sums are difficult to give.58 In general, the price would depend on the man-hours and materials needed. While we lack written evidence,59 it has been argued based on Byzantine practices that there were professional cistern builders in antiquity.60 Excavating the cavity was a work intensive process, as shown by Werner Brinker who, based on studies of the Eupalinos tunnel on Samos suggested 50 man-hours per m3 in hard lime-stone.61 This would result in over 5000 man-hours for many cisterns on Delos, about oneand-a-half years work.62 Subsequently, masons and material would be needed for the cistern walls. Since most cisterns on Delos are rectangular, they also needed a cover, which includes either arches or beams of wood or stone. Finally, a lining was necessary to make the cistern waterproof. This was probably also applied by professionals. Studies of the cistern lining at Kerameikos in Athens suggest that a technique similar to ta-de-laq in modern Morocco was used, where a mixture of olive oil and mud created a final surface when polished with a hard material such as

55 Desruelles 2001, 564 states that the cisterns were well dimensioned to the rainwater harvesting area. Instead, Desruelles et al. 2003, 260 and Desruelles & Fouache 2015, 209–210 argue that some cisterns were over-dimensioned, others under-dimensioned. 56 Trümper 1998, 301; Klingborg 2017, no. 332. 57 Klingborg & Finné 2018. 58 Karvonis (2000, 87) was surprised that more houses did not have a well considering their low cost, but he probably underestimated the investment. For the cost of constructing a cistern, see Klingborg 2017, 72–75. 59 Except for Vitr. De arch. 8.6.12–15 and Plin. HN 36.52. 60 Jantzen & Megow 1977, 175. 61 Brinker 1990, 24–25. 62 Based on ten working hours per day.

Social Stratification and Water Sharing on Late-Hellenistic Delos  45 marble.63 On Delos an inscription testifies to Aristokritos plastering a frear, a term for well or cistern used during Hellenistic times.64 Constructing a well would be considerably less labour intensive. Depending on local conditions, it would either be cut into the bedrock or dug as a large pit supported by scaffolding. Thereafter the walls could be constructed.65 In contrast to the situation in many other areas of Greece, some wells on Delos were lined, adding to the cost. It is difficult to estimate the monetary investment of constructing a well or cistern. The cost of 15–30 m3 large cisterns has been calculated to about a year’s salary for a skilled worker based on the required man-hours.66 The larger and architecturally more elaborate cisterns on Delos would be significantly more expensive. Wells were correspondingly cheaper due to their smaller size and less costly architecture. As a point of comparison, a Hellenistic inscription from Halicarnassus records donations for the construction of a frear, presumably a well in this case, with preserved sums reaching a minimum of 94 drachmas, perhaps a quarter of an annual salary.67 The total cost was almost certainly considerably higher as the inscription also includes what may be workmen or workdays. Thus, regardless of the actual sum for the construction of a cistern or well on Delos, we should consider it substantial and well beyond the means of most individuals. Space The space available was also important, both during and after construction. Building a cistern required a considerable area to be dug out and often this would have been dangerous if done near other structures due to the risk of collapse. Most cisterns were therefore probably constructed before the houses were built.68 Digging a well would be limited by similar factors, especially in those cases when it was not cut into the bedrock all the way from the mouth. Once complete, cisterns and wells used similar openings, often puteals, resulting in comparable appearances and space usages on the surface.69 Despite this, cisterns continued to have a large footprint as walls were rarely constructed on top of them. Consequently, a cistern usually continued to require a large open area after its completion. This would make it impractical to have a cistern in many shops. For example, the cistern in House III N occupies almost 30 m², an area considerably

63 Stroszeck 2014; 2017. 64 IG XI.2 154, l. 33. See also Klingborg 2017, 69; 2023. 65 The way wells were constructed was largely a local phenomenon, but see Stroszeck 2017, 48–51 for comparanda from Kerameikos in Athens. 66 Klingborg 2017, 72–75. Estimating the income during antiquity is, of course, complicated. 67 EM 199. See Isager 2002. For the identification as a well, see Klingborg 2017, 66–71; 2023, 161, n. 2. 68 See, e.g., Trümper 1998, 265 concerning House II F. 69 Although in Athens well puteals were slightly larger than those for cisterns. See Lang 1949. For examples of where puteals could be placed, see House II A for the stylobate, House VI I for the impluvium, House III K for walls, and House II B for regular rooms.

46  Patrik Klingborg larger than most shops along the Rue du Théâtre.70 In fact, even small cisterns, such as the one in House III A, which measured just under 10 m², would occupy more than half the space in most shops. But for these shops, control of the roof was presumably an equally important limitation since cisterns rely on rainwater to be filled. Consequently, there would rarely be any point in constructing one for domestic units as small as shops.71 Personal Preferences While volume and seasonality, cost, and space were important factors, the much less tangible personal preferences probably also played a role. This is reflected in the difference between Houses III A and III B. Both houses have been interpreted as canonical Normalhäuser that once formed a freestanding block to which the insula was later attached. But despite once being practically identical, at some point (possibly already in the original phase) the occupants of House III A constructed a well, while those of House III B opted for a cistern.72 This illustrates that different water-supply solutions were sometimes chosen in similar circumstances. Access to Water The unequal distribution of water created a situation where a large proportion of the population did not have access to private water sources. Most importantly, the cost, lack of space, and, in the case of cisterns, unavailability of rainwater harvesting surfaces would have impeded the occupants of shops and possibly some houses to acquire a private water source. So how did these individuals access water? Two options existed: either they relied exclusively on public sources, or they could use private sources belonging to other individuals. On Delos the exclusive use of public sources is unlikely. The Inopos reservoir and the Theatre cistern together had a volume of 2850 m3, or about 285 l/year per person during Late Hellenistic times, far too little for any one individual to survive. Moreover, this estimate might in fact be rather optimistic. Even in modern times, without any water being drawn, the Theatre cistern dries up on a regular basis. Presumably the situation was more severe in Hellenistic times when the aquifer was under considerable stress. The Inopos reservoir seems to have been more reliable, but lacking a cover it was likely polluted in various ways, including through algae growth. Finally, the Minoe fountain was certainly too small to support the whole island. Consequently, water must have been shared between complexes.73 How such sharing arrangements functioned in practice, however, is unknown. Except for specific scenarios in Solon’s archaic Athenian law, there is no direct 70 For more on House III N, see Trümper 2006. 71 Klingborg & Finné 2018. 72 Délos 8, 38; Trümper 1998, 266–267. The cistern’s oblique orientation suggests that it pre-dated the current building, but Trümper believes that these were the original structures. 73 As suggested by Hellmann 1992, 241; Trümper 1998, 28; Karvonis 2000, 75, 88.

Social Stratification and Water Sharing on Late-Hellenistic Delos  47

Figure 3.5 Water sources within 50 meters of Shop RdT 33. Map by Author.

evidence for organised private water sharing in the Greek world.74 It is clear, though, that tightly knitted hydraulic relationships must have been integral to the urban fabric of Delos. These relationships were probably characterised by two contradictory principles: great accessibility in terms of the actual number of water sources, and severe limitations due to unequal power structures. The large number of water sources in the area ensured that a well or cistern was always present within a short distance. For example, Shop 33 on the Rue du Théâtre had 14 water sources in 10 domestic units within a 50 m walking distance (Figure 3.5).75 Of these, six were cisterns and eight were wells, offering a relatively 74 This stands in contrast to the well-developed Roman legal system for water access. There is some scant evidence for the sale of water, e.g., from a sanctuary of the Nymphs in Attica (LSG 178, see Bousquet 1967, 92–95; Koerner 1974, 173). See also the text in I.Stratonikeia 1508, dated to 81 BC or soon thereafter, which van Bremen 2011 convincingly argues may have regulated access to a public water source, including sums for some individuals having access to the water ‘day and night’. While much later and from a much different context, in Praise of Antioch, Oration XI.243–248, Libanius (AD 314–393) mentions that the inhabitants of Antioch could go to their neighbours to get water rather than to the public fountains. However, this was ascribed to an abundance of water, and consequently the presence of fountains in many private houses. See also the contributions by Locicero and Baird, this volume. 75 One of these sources was located within 20 m (Well III K-a), another within 30 m (Cistern VI L-b), seven within 40 m (Cistern III N-f, Cistern VI L-c, Well VI L-c, Well VI L-f, Cistern VI M-c, Well RdT 10, and Well RdT 27b), and five within 50 m (Cistern III H-b, Cistern III Q-d, Well III Q-d, Well VI A-c, and Well RdT 45).

48  Patrik Klingborg

Figure 3.6 Water sources within 50 meters of Shop RdT 27. Map by Author.

even distribution in terms of type, allowing them to complement each other depending on the time of the year. This situation is mirrored by other domestic units along the street, such as Shop RdT 27 with nine water sources within a 50 m range (Figure 3.6).76 Other domestic units had access to fewer, and less diverse, water sources within the same distance. From House VI C, which lacked a water source of its own, only five wells could be reached within 50 m, offering a less extensive and versatile water supply than in the case of Shops 33 and 27 (Figure 3.7). Expanding the distance to 100 m from House VI C, however, increases the number of water sources dramatically to 30—17 wells and 13 cisterns. Despite the numerous choices, there can be little doubt that hydraulic relationships were also defined by power structures.77 Invariably those with secure access to water would be in a stronger position compared to those that did not, particularly during dry periods when choices became fewer. Overall, the need for water must have created a situation where large numbers of less well-off individuals were dependant on elites to such a degree that they could presumably be coerced into doing their bidding to some extent. Perhaps, considering the Italic influences on Delos, this took the form of a patron-client relationship. If correct, then the hydraulic

76 Well VI I-c, Cistern VI-I-c, Well RdT 10, Well RdT 27b, Cistern III H-b, Well VI A-c, Well VI K-a, Cistern VI L-b, and Well VI L-f. 77 See also Baird, this volume.

Social Stratification and Water Sharing on Late-Hellenistic Delos  49

Figure 3.7 Water sources within 50 meters of House VI C. Map by Author.

relationships would contribute to maintaining unequal circumstances on Delos, further strengthening the elite at the expense of the general population. Controlling a water source therefore did not only secure your own supply but also allowed the projection of power in society. Conclusions The many wells and cisterns in the Quartier du Théâtre on Delos, in combination with the limited public water supply, suggest that its inhabitants relied primarily on domestic water sources. However, while houses often had a private source, shops did not. Though some inhabitants without a water source may have relied partly on public sources, it is likely that often-private sources in other complexes were used; thus, accessing water was presumably a question of social arrangements. The number of private water sources within short distance, furthermore, indicates that social arrangements were more important than physical vicinity. In terms of choice of water source, the evidence suggests that wells were preferred in some cases, cisterns in others. While cisterns were more expensive and required more space to construct, these factors cannot have been decisive; some of the largest houses did not have a cistern, while some of the smaller ones had more than one. Multiple cisterns are also more common than multiple wells. In the past, houses with more than two water sources have been interpreted as the result of two complexes being combined but, in a number of cases, this seems unlikely.

50  Patrik Klingborg Rather the choice of number and type of water source often seems to have reflected individual preference. In the end, the uneven distribution of water sources on Delos, in particular, between houses and shops, strongly suggests the presence of hydrological power structures allowing some individuals to exert influence over others. This must have had important effects on everyday life, for example, in terms of agency, potential activities, living standard, and health. A further effect would have been a disconnect between those individuals living on the margins and those who could engage in various forms of conspicuous consumption—the haves and the have nots, as Gerrard phrases it in his chapter. Finally, the situation must have been the most critical during dry years when little water was accessible even for those with private water sources. While we know little about the living conditions for the less fortunate during such periods, it is difficult to imagine how they were not, quite literally, at the mercy of their wealthier neighbours. Acknowledgements I would like to thank all the participants in the workshop for their excellent feedback on a draft version of this chapter, Irene Vikatou for her support and suggestions, Pamela Browne for her comments and Rick Bonnie for his excellent editorial work. References Bousquet, J. 1967. ‘Deux inscriptions attiques’, BCH 91, 90–95. van Bremen, R. 2011. ‘Day and night at Stratonikeia’, in Labraunda and Karia. Proceedings of the international symposium commemorating sixty years of Swedish archaeological work in Labraunda. The Royal Swedish Academy of Letters, History and Antiquities Stockholm, November 20–21, 2008, eds. L. Karlsson & S. Carlsson. Boreas. Uppsala Studies in Ancient Mediterranean and Near Eastern Civilizations 32, Uppsala, 307–329. Brinker, W. 1990. Wasserspeicherung in Zisternen. Ein Beitrag zur Frage der Wasserversorgung früher Städte. Leichtweiss-Institut für Wasserbau der Technischen Universität Braunschweig, Mitteilungen 109, Braunschweig. Brunet, M. 2001. ‘5. L’eau à Délos. Un milieu naturel et son aménagement durant l’Antiquité’, in Délos, BCH 125, 620–627. Brunet, M. 2008. ‘La gestion de l’eau milieu urbain et rural à Délos dans l’Antiquité’, in Vers une gestion intégrée de l’eau dans l’empire romain. Actes du Colloque International Université Laval, octobre 2006, ed. E. Hermon. Atlante tematico di topografia antica, Suppl. 16, Rome, 25–32. Brunet, M. 2011. ‘L’eau dans la Délos antique: Programmes athéniens d’ingénierie hydraulique sur l’île sacrée d’Apollon’, Comptes rendus des séances de l’Académie des Inscriptions et Belles-Lettres 155(2), 695–704. Brunet, M., S. Desruelles, C. Cosandey, E. Fouache, K. Pavlopoulos & H. Siard 2003. ‘L’eau à Délos’, BCH 127, 516–525. Chiotis, E.D. 2018. ‘The Hadrianic aqueduct of Athens and the underlying tradition of hydraulic engineering’, in Great Waterworks in Roman Greece. Aqueducts and Monumental Fountain Structures. Function in Context, eds. G.A. Aristodemou & T.P. Tassios. Archaeopress Roman Archaeology 35, Oxford, 70–97.

Social Stratification and Water Sharing on Late-Hellenistic Delos  51 Délos 4 = Cayeux, L. 1911, Exploration archéologique de Délos IV. Description physique de l’île de Délos, Paris. Délos 8 = Chamonard, J. 1924. Exploration archéologique de Délos VIII. Le Quartier du théâtre. Construction et technique, Paris. Délos 42a = Fraisse, P. & J.-C. Moretti 2007. Exploration archéologique de Délos XLII. Le théâtre, Vol. 1. Texte, Paris. Délos 42b = Fraisse, P. & J.-C. Moretti 2007. Exploration archéologique de Délos XLII. Le théâtre, Vol. 2. Planches, Paris. Desruelles, S. 2001. ‘Les ressources hydriques et les aménagements antiques dans le contexte cristallin de l’île de Délos’, Topoi 11, 551–578. Desruelles, S. 2007. ‘La gestion des ressources en eau dans la ville antique de Délos (Cyclades, Grèce)’, Bulletin de l’Association de géographes français 84, 161–172. Desruelles, S., L. Chevalier, C. Cosandey, P. Karvonis & J.-C. Moretti 2003. ‘The mastery of drainage and water management at Delos’s “Theater Quarter” (Cyclades, Greece)’, in The Mediterranean world environment and history, ed. É. Fouache, Paris, 253–262. Desruelles, S. & É. Fouache 2015. ‘Water in the ancient city of Delos (Cyclades, Greece). Resources and hydraulic devices’, in La géoarchéologie française au xxie siècle, ed. N. Carcaud & G. Arnaud-Fassetta, Paris, 203–212. Desruelles, S., É. Fouache, A. Ciner, R. Dalongeville, K. Pavlopoulos, E. Kosun, Y. Coquinot & J.-L. Potdevin 2009. ‘Beachrocks and sea level changes since Middle Holocene: Comparison between the insular group of Mykonos–Delos–Rhenia (Cyclades, Greece) and the southern coast of Turkey’, Global and Planetary Change 66, 19–33. Desruelles, S., É. Fouache, R. Dalongeville, K. Pavlopoulos, J.-P. Peulvast, Y. Coquinot, J.-L. Potdevin, C. Hasenohr, M. Brunet, R. Mathieu & É. Nicot 2012. ‘Sea-level changes and shoreline reconstruction in the ancient city of Delos (Cyclades, Greece)’, Geodinamica Acta. The European Journal of Geodynamics 20, 231–239. Ehrenheim, H., P. Klingborg & A. Frejman 2019. ‘Water at ancient Greek sanctuaries: Medium of divine presence or commodity for mortal visitors’, Journal of Archaeology and Ancient History 26, 1–31. Finker, M. & J.-C. Moretti 2007. ‘Le barrage du réservoir de l’Inopos à Délos’, BCH 131, 187–228. Finné, M. & I. Labuhn 2023. ‘Hydro-climate in the Aegean from 700 BC to AD 300: Links between climate and freshwater availability’, in Going against the flow. Wells, cisterns and water in ancient Greece, ed. P. Klingborg. ActaAth-4°, 23, Athens, 31–54. Glaser, F. 1983. Antike Brunnenbauten (κρηναι) in Griechenland. Österreichische Akademie der Wissenschaften. Philosophisch-historische Klasse. Denkschriften, 161, Vienna. Hellmann, M.-C. 1992. Recherches sur le vocabulaire de l’architecture grecque, d’apres les inscriptions de Délos. BÉFAR, 278, Athens. Hodge, A.T. 2000. ‘Qanats’, in Handbook of ancient water technology, ed. Ö. Wikander. Technology and change in history 2, Leiden, 35–38. Isager, S. 2002. ‘Halikarnassos and the well of Aphrodite on EM 199, text and provenance’, in Ancient history matters. Studies presented to Jens Erik Skydsgaard on his seventieth birthday, eds. K. Ascani, V. Gabrielsen, K. Kvist & A. Holm Rasmussen. AnalRom Suppl. 30, Rome, 153–158. Jantzen, U. & W.-R. Megow 1977. ‘Eine Zisterne im Stadtgebiet von Samos’, AM 92, 171–195. Karvonis, P. 2000. L’eau dans la ville Hellénistique de Délos, MA thesis, Université de Paris X. Karvonis, P. 2023. ‘The water supply in the houses of Delos’, in Going against the flow. Wells, cisterns and water in ancient Greece, ed. P. Klingborg. ActaAth-4°, 23, Athens, 77–90.

52  Patrik Klingborg Klingborg, P. 2017. Greek cisterns. Water and risk in ancient Greece, 600–50 BC, Ph.D. thesis, Uppsala University. Klingborg, P. 2023. ‘Wells and cisterns in Greek literature’, in Going against the flow. Wells, cisterns and water in ancient Greece, ed. P. Klingborg. ActaAth-4°, 23, Athens, 161–178. Klingborg, P. & M Finné 2018. ‘Modeling the freshwater supply of cisterns in ancient Greece’, Water History 10, 113–131. Koerner, R. 1974. ‘Zu Recht und Verwaltung der griechischen Wasserversorgung nach den Inschriften’, ArchPF 22–23, 155–202. Lang, M. 1949. ‘ΙΣΘΜΙΑ ΦΡΕΑΤΩΝ. Terracotta well-heads from the Athenian Agora’, Hesperia 18, 114–127. Moretti, J.-C. & M. Fincker 2011. ‘Les réseaux d’eau courante à Délos’, in Les réseaux d’eau courante dans l’antiquité. Réparations, modifications, réutilisations, abandon, récupération. Actes du colloque international de Nancy (20-21 novembre 2009), eds. C. Abadie-Reynal, S. Provos & P. Vipard, Rennes, 159–172. Oleson, J.P. 2000. ‘Water-lifting’, in Handbook of ancient water technology, ed. Ö. Wikander. Technology and change in history 2, Leiden, 217–302. Olynthus 12 = Robinson, D.M. 1946. Excavations at Olynthus XII. Domestic and public architecture. The John Hopkins University studies in archaeology 36, Baltimore. Siard, H. 2006. ‘Un règlement trouvé dans le Réservoir de l’Inopos à Délos’, BCH 130, 329–348. Siard, H. & D. Braunstein 2006. ‘Travaux de restauration au réservoir de l’Inopos’, BCH 130, 748–751. Stroszeck, J. 2014. ‘Water management in Classical Athens: cisterns of the classical bathhouse on the Kerameikos road in front of the Dipylon’, in ΙWA regional symposium on water, wastewater and environment: traditions and culture. Patras, Greece, March 22– 25, 2014. E-Proceedings, eds. I.K. Kalavrouziotis & A.N. Angelakis, Patras, 499–507. Stroszeck, J. 2017. ‘Wells in Athens: The contribution of the Kerameikos wells’, in Cura Aquarum in Greece. Proceedings of the 16th international conference on the history of water management and hydraulic engineering in the Mediterranean region, Athens, Greece 28–30 March 2015, ed. K. Wellbrock, Siegburg, 43–88. Tölle-Kastenbein, R. 1985. ‘Der Begriff Krene’, AA 1985, 451–470. Trümper, M. 1998. Wohnen in Delos. Eine baugeschichtliche Untersuchung zum Wandel der Wohnkultur in hellenistischer Zeit. Internationale Archäologie 46, Rahden. Trümper, M. 2006. ‘Baden im späthellenistischen Delos, I: Die öffentliche Badeanlage im Quartier du Théâtre’, BCH 130, 143–229. Uhl, V.W., J.A. Baron, W.W. Davis, D.B. Warner & C.C. Seremet 2009. Groundwater development basic concepts for expanding CRS water programs, Baltimore. Vallois, R. 1944. L’architecture hellénique et hellénistique à Délos jusqu’à l’éviction des Déliens (166 av. J.-C); 1ère partie: Les monuments. Bibliothèque des Écoles françaises d’Athènes et de Rome 157, Paris. Wagner, E.G. & J.N. Lanoix 1959. Water supply for rural areas and small communities. World Health Organization Monograph Series 42, Geneva. Wilson, I.A. 2008. ‘Hydraulic engineering and water supply’, in The Oxford handbook of engineering and technology in the classical world, ed. J.P. Oleson, Oxford, 285–318. Wycherley, R.E. 1937. ‘ΠΗΓΗ and ΚΡΗΝΗ’, CR 51, 2–3.

4

Surveying Notion’s Residential Water Supply Cistern Use during Hellenistic–Roman Times Angela Commito

Introduction Notion is an ancient harbour town on the Aegean coast of Turkey, located 50 km south of Izmir (Figures 4.1 and 4.2). The town is strikingly situated on a pair of promontories that rise as much as 85 m above sea level. Below the promontories, bays to the east and west provide natural harbours; the Hales River, modern Avcı Çayı, hugs the site and empties into the western bay. The town stretches across an area of 35 ha, enclosed by 3.5 km-long fortification walls. From these walls, the city of Ephesus is clearly visible across the sea, only 15 km to the southeast. At one point both Notion and Ephesus were thriving urban centres with vibrant harbours. However, the trajectory of the life of the community at Notion diverged markedly from that of its larger and better-known neighbour. Unlike Ephesus, the fortified harbour town identified as Notion was occupied intensively for only a few centuries. Why did Ephesus survive as a city while Notion did not? How does studying short-lived cities such as Notion help us understand patterns of ancient urbanism that must have been much more common than that enjoyed by such long-lived cities as Ephesus? What do we gain by considering the biography of a ‘failed’ city? These are among the questions explored by the Notion Archaeological Survey, a research project carried out between 2014 and 2019 by the University of Michigan and Brown University.1 The project produced rich ceramic evidence for the chronology of the settlement and its involvement in regional and long-distance exchange, clarified essential aspects of the city plan and urban development, provided detailed study and documentation of major buildings and public works, and shed light on the local geology and modification of the landscape through terracing and stone quarrying. Our most important conclusion has been to demonstrate that Notion was intensively occupied for only a few centuries. Although Notion is known from a fragment of Hecataeus to have existed as a community at least by c. 500 BC,2 there is 1 The project is directed by Christopher Ratté, with Felipe Rojas serving as assistant director and the author as senior archaeologist. Project website: http://sites.lsa.umich.edu/notionsurvey/. 2 Hecateaus FrGrHist 1A, 1, F fr. 233 (apud Steph. Byz.). DOI: 10.4324/9781003268222-4

54  Angela Commito

Figure 4.1 Map of the region around Notion. Notion Archaeological Survey.

little archaeological evidence for occupation of the site before the third century BC. The visible remains must therefore belong to a radical expansion or relocation of the original settlement, possibly located closer to the harbour and the Hales River and now buried beneath river-borne silt. Even more surprising is the almost complete lack of evidence for occupation after the first century AD, a period associated with unprecedented growth in the region. The lack of evidence is especially noteworthy given that the imperial period was prosperous for the nearby oracular sanctuary of Apollo at Claros, and that in late antiquity Notion (then called Colophon) was an episcopal see, represented by bishops at the early Christian councils of Ephesus and Chalcedon in the fifth century AD.3 The surprisingly limited ‘heyday’ of Notion, lasting about ten generations from the third century BC to the first century AD, is indicated by varied archaeological evidence documented by the survey.4 Almost all of the identifiable ceramics 3 See Price & Gaddis 2007, I 298, 336; II 232. 4 The preliminary results of the survey, with references, are published in Ratté et al. 2017; 2018; 2020a; 2020b.

Surveying Notion’s Residential Water Supply  55 ­retrieved through surface collection date to between the third century BC and the first century AD.5 The best parallels for the city plan and fortifications are found in late Classical and Hellenistic cities in Western Turkey. Major features of the Bouleuterion are typical of the Classical and Hellenistic periods. Details of the architectural ornament of the Temple of Athena indicate a date not later than the Augustan period. Notable, too, is the lack of mortared rubble or brick construction and the absence of typically Roman buildings such as baths. It is not only the lack of baths, but the water supply system as a whole, that offers clues about the surprising biography of this ancient city. The ‘re-founding’ of Notion on promontories overlooking the Aegean Sea had important strategic advantages, including defence. But despite its real or perceived benefits, this new location presented a major challenge: a complete lack of fresh water. No springs have been found on the site, and the nearest source of fresh water seems to have been the Hales River, located 85 m below the highpoint of the city. How did the builders of the new town address this problem? Was this deficiency partly to blame for the large-scale abandonment of the city, which seems to have taken place only a few centuries after its re-founding? The Water Supply of Notion The results of the survey indicate that rainwater collected in cisterns formed an essential part of the water supply on which the residents of ancient Notion depended. A total of 13 certain cisterns and ten likely or possible cisterns have been identified on the site (Figure 4.2). ‘Certain’ cisterns have clearly recognizable cavities for water collection or are otherwise identifiable by the presence of hydraulic mortar and/ or a circular masonry or rock-cut shaft. Cisterns considered ‘likely’ are features with cistern-like characteristics, such as a circular arrangement of stone blocks but no visible hydraulic mortar, while those recorded as ‘possible’ are visible as large depressions in the ground. An additional three cisterns were observed outside the survey zone, immediately north of the modern highway that runs through the valley north of the city’s fortification walls. Finally, five water-related stone blocks were documented within the site (not shown on the map). The documented cisterns are located throughout the site, in both non-domestic and domestic spaces: around the Temple of Athena and the so-called Heroon, in and around the Agora, near the Bouleuterion and Theatre, and on the southern slopes of the site in what appear to have been residential areas. They must have gathered rainwater from the roofs of domestic and non-domestic buildings alike. Location does not necessarily indicate who used a cistern and for what purposes: as at any site, the presence of a cistern need not relate to the use of the building above it. Although most of the identified 5 What is remarkable is the almost total absence of recognizable wares of the later Roman period, such as African Red Slip ware, or of late Roman amphora types. These results are consistent with the pottery found in earlier excavations, both the material published by A. Demangel and R. Laumonier, and the material recovered by excavations directed by E. Atalay and M. Büyükkolancı: Demangel & Laumonier 1923; 1925; Atalay 1986; 1987; Büyükkolancı 1996.

Figure 4.2 City plan of Notion, showing the locations of documented cisterns. Notion Archaeological Survey.

56  Angela Commito

Surveying Notion’s Residential Water Supply  57

14

18

7

2

9

12

6

19

1 0

5

10 m

Figure 4.3 Compiled cross-sections of nine cisterns produced from laser scans. Notion Archaeological Survey.

cisterns are located on the central ridge of the site, there must have been many more in what we interpret to be residential areas to the north and south, in sloping terrain now covered by eroded soil and debris. Laser scans were made of twelve of the cisterns that could be entered, producing detailed three-dimensional digital models (Figure 4.3).6 The cisterns were cut into bedrock, in some places schist, in others marble, in still others, banded schist and marble. The cisterns are presumably flask-shaped, although none of the original bottoms are visible and might be rounded rather than flat. In general, they expand from a diameter of 1–2 m at the top to 3–4 m at the preserved bottom and extend down to a preserved depth of 5–7 m. Most are lined with hydraulic mortar, especially the schist surfaces. In several cisterns one or two terracotta pipes for water inflow (and possibly for outflow, too, though none has been specifically identified as such) remain in situ, laid above the bedrock and below the large stone slabs used to articulate the cistern opening. In several cases, the upper parts of the shafts are built out of rubble masonry supporting the stone slabs. One cistern was roofed using a system of stone arches. The following discussion illustrates these overall patterns in construction and highlights specific issues of interest.

6 The laser scans were made by Christian Kurtze of the Austrian Archaeological Institute. The author is grateful for his enthusiasm and expertise.

58  Angela Commito Cistern Construction and Use The basic design and construction of the cisterns at Notion are well illustrated by cistern 19, located west of the Theatre (Figures 4.3, no. 19 and 4.4–4.6). The cistern cavity is excavated entirely from the marble bedrock in a series of concentric rings expanding in diameter towards the bottom. Pick marks are well preserved on the surface of the bedrock, which may have originally been lined with hydraulic mortar, as most of the other cisterns are (especially the schist surfaces, but also some of the marble surfaces, as, for example, in cistern 18, located in a broad saddle between the Agora and the Theatre). At the time it was documented, the cistern contained some water, which, along with debris, made it impossible to measure its original depth. The present depth measured from the top of the preserved masonry opening is approximately 7 m, at which point the cistern has a maximum diameter of 3.2 m. Two terracotta pipes are located above the bedrock, which has been cut to create a kind of bedding for the pipes. This aspect of construction is much more obvious in some of the other cisterns, where the bedding is deep enough to form a narrow trench, confirming that it is intentional. Small, irregular stones ranging 10–30 cm in length are placed around the pipes and across the top of the bedrock. These create a relatively level surface on which two large, rectangular stone slabs oriented north-south are positioned at the west and east edges of the opening. The slabs are at least 1.1 m long and 20–25 cm in height; the width could not be measured. An additional, smaller set of slabs are laid east-west between the larger ones. These marble (or schist) slabs serve as transition blocks between the circular rock-cut shaft below and the rectilinear masonry superstructure on top.

Figure 4.4 Internal view of cistern 19, showing pick marks and excavation in concentric rings. Notion Archaeological Survey

Surveying Notion’s Residential Water Supply  59

Figure 4.5 Placement and bedding of terracotta pipe inside cistern 19. Notion Archaeological Survey.

Figure 4.6 Masonry opening of cistern 19. Notion Archaeological Survey.

60  Angela Commito The course laid directly on top of the slabs slightly overhangs the smaller slabs but is set back from the edge of the larger slabs by about 5 cm and is composed of four large blocks of the local conglomerate, each of which is 50 cm high and 66–70 cm long; their width could not be measured. Two courses of marble blocks, levelled with chinking stones, are preserved on top of the conglomerate blocks along the west and south sides only. The maximum height of these two courses is 47 cm. A final course is preserved on top at the southwest corner of the structure at the current ground level. The original floor level cannot be determined, and it is unclear how the floor related to the cistern opening and whether there was a puteal or other protective covering. The edges of the large north-south slabs that articulate the cistern opening are worn with what appear to be rope marks. Since these could be from modern use, they do not necessarily indicate how water was collected from the cistern originally or throughout its lifespan. This same basic construction composed of a bedrock cavity, into which terracotta pipes are placed, overlaid with levelling masonry topped by rectangular stone slabs that themselves support some kind of superstructure can be clearly seen in several other cisterns. In most of these, however, nothing is preserved above the level of the large stone slabs that provide a transition from circular shaft to rectilinear opening. These are clearly visible from the inside of cistern 9, located on a hillslope overlooking the southwest corner of the Agora, which has two terracotta pipes placed at the top of the bedrock below the levelling stones on which the large slabs are placed (Figures 4.3, no. 9, 4.7 and 4.8). The cistern is excavated from

Figure 4.7 Roofing system of cistern 9 from below, with one of two terracotta pipes visible. Notion Archaeological Survey.

Surveying Notion’s Residential Water Supply  61

Figure 4.8 View inside cistern 9, showing pick marks on the interbedded schist and marble bedrock surfaces and hydraulic mortar on the upper portion of shaft. Notion Archaeological Survey.

62  Angela Commito schist interbedded with marble, and hydraulic mortar is preserved around the upper portion of the shaft. A slightly different roofing system was also observed. Unlike the cisterns described above, which are still accessible from the original opening, cistern 21, located just south of the Theatre, appears to have been illegally excavated or otherwise entered from the side. As a result, the roofing system is still intact and visible from below. Two large schist slabs, about 1.2 m long, 60 cm wide, and 20–25 cm thick, completely cover the bedrock cavity. Semicircles are cut from their abutting edges to make a circular opening, approximately 40 cm in diameter. This opening is topped by a smaller, flat schist slab, visible through the circular opening. Because the roofing system could be observed only from below, it was not possible to obtain additional dimensions or discern details of construction. The overall scheme appears similar to that documented for two cisterns at Pergamon, in which capstones were prepared with semicircle cuttings to create a circular opening sealed with a round stone lid.7 At Notion, cistern 11, located in the northeast corner of the Agora, has a square opening, 55 cm by 60 cm, formed by three large cover slabs. Could the other cisterns, described above, have also used this roofing system? It seems unlikely, since the space between the large slabs that articulate the opening in these cisterns is usually 50 cm or so wide, too narrow to fit an additional set of slabs with cuttings that would allow for an opening on the order of 40 cm in diameter. A third roofing system was used for one of the documented cisterns at Notion, cistern 14, located on the sloping terrain between the Agora and a lower saddle to the east (Figures 4.3, no. 14 and 4.9). One standing arch of large conglomerate masonry voussoirs, oriented west-east, spans the northern portion of the cistern cavity. Cuttings for a second arch spanning the southern portion of the cavity are clearly visible on the west surface. These two arches may have supported masonry walls across which stone beams or some other type of flooring, laid perpendicularly, covered the cavity below.8 Any possible surviving superstructure has been disturbed by extensive illegal excavations. Adaptation to unusual circumstances, reuse, and remodelling are all evident in the cisterns of Notion. One rather whimsical example is cistern 7, located northeast of the Heroon, in the terraces that step down from the central spine of the site to the northern fortification walls (Figures 4.3, no. 7, 4.10 and 4.11). The bottom of the cistern cavity is excavated from schist bedrock, on top of which is a masonry wall, 1.7 m high, itself roofed with large slabs in an arrangement like that described above. However, the masonry wall abuts on a large boulder or protrusion of marble bedrock, which the cistern builders left in place, and which therefore forms the upper eastern wall of the cistern. The masonry wall is well preserved along the south, 7 Cisterns 85 and 86 in Garbrecht 2001, 21. 8 Stone or wooden beams supported by multiple arches are documented in many of the cisterns at Delos, usefully compiled in the catalogue in Klingborg 2017, 247–259, with references.

Surveying Notion’s Residential Water Supply  63

Figure 4.9 View of cistern 14, looking North. Notion Archaeological Survey.

64  Angela Commito

Figure 4.10 Interior view of cistern 7. Notion Archaeological Survey.

Surveying Notion’s Residential Water Supply  65

Figure 4.11 Roofing system of cistern 7 from below. Notion Archaeological Survey.

where it is encased in the hillside in what must have been a terracing structure. To the north, the masonry has been punched through, allowing access to the cistern cavity and preserving much of the roofing system. Two large stone beams, oriented east-west across the cavity, support at least three cover slabs. It is not clear how the cistern was accessed, since no opening is visible, and all of these slabs seem too large to have been conveniently lifted as water was needed. Possibly an opening was carved into one or more cover slabs that have not survived, on the disturbed northern edge of the cistern. Another unusual cistern is cistern 1, located just north of the Temple of Athena (Figures 4.3, no. 1 and 4.12). Though otherwise standard—flask-shaped cavity carved from bedrock, topped by terracotta pipe, levelling course, and large stone slabs—placed on these stones is the top portion of a pithos, which has been roughly cut across the belly. The wide seam between the slabs and the uneven bottom edge of the pithos is filled with small rocks, ceramic sherds, and soil. Though possibly part of the original construction, the pithos seems rather to be a later modification or addition, a perfectly circular and ready-made access system for a cistern that remains open today to a depth of over 7 m, measuring down from the pithos rim. Additional evidence for remodelling and maintenance comes from cistern 18, in the low saddle of the site, in which at least two layers of hydraulic mortar are preserved, and from cistern 6, located just north of the Heroon, the walls of which include an 80 by 80 cm patch of mortared rubble infill. Furthermore, the undulating mortared surfaces inside cistern 12, located just off the northeast corner of the

66  Angela Commito

Figure 4.12 View of partial pithos reused to articulate the opening of cistern 1. Notion Archaeological Survey.

Bouleuterion, may be the result of re-mortaring after the collapse of portions of the schist bedrock that formed the cavity (Figure 4.13). Collapse was evidently a problem at some point in the lives of many of these cisterns and the people who maintained and depended on them. There is a striking, though not surprising, correlation between the state of preservation of a cistern cavity, on the one hand, and the type of bedrock from which it was excavated, on the other. Cisterns excavated all or mostly from marble are very well preserved, while those excavated from all or mostly schist exhibit minor to very severe collapse. The difference in preservation is especially well illustrated by the three-dimensional models created through laser scanning, which is a remarkable tool for the visualization of subterranean and other features that are otherwise difficult to document (Figure 4.3). Notion’s Water Supply in Context Flask-shaped cisterns are found throughout the Mediterranean from the fifth century BC through the later Roman era,9 but in mainland Greece and the Aegean, the period of highest frequency of construction is constrained to the third and   9 Connelly & Wilson 2002, 278–280, list comparanda at Athens (fifth century BC), Olynthus (fourth century BC), Kourion (third century BC), Pergamon (probably third and second centuries BC), Morgantina (second and first centuries BC), and Punic Carthage (before 146 BC). Comparanda for vaulted cisterns topped by large, flat cover slabs exist across the eastern Mediterranean, from the

Surveying Notion’s Residential Water Supply  67

Figure 4.13 View inside cistern 12, showing hydraulic mortar on uneven surfaces of schist bedrock. Notion Archaeological Survey.

second centuries BC.10 The construction of the cisterns at Notion seems therefore to correspond to the moment when the city was re-founded and enjoyed its greatest occupation. At other Hellenistic cities such as Pergamon and Morgantina, most excavated houses had at least one cistern, located in the peristyle or courtyard.11 If the same pattern held true at Notion, there originally could have been hundreds of cisterns, one for each household. Even a low estimate of the urban population at Notion (3,500–4,000 people) corresponds to 700–800 households. However, some households may have shared a cistern or relied on water from communal sources, such as fountains or cisterns associated with public areas or buildings (e.g., in the Agora and next to the Bouleuterion). The identifiable remains of houses at Notion range from 217 to 435 m2 in area, including unroofed courtyards.12 The most clearly visible structure on the site with domestic features, in particular, an interior peristyle courtyard, is located between the Heroon and the Agora and contains cistern 8, which is completely filled in (Figure 4.2). Though separate from what appear to be the main residential areas of Hellenistic period (e.g., many cisterns on Delos) to the early Byzantine period (e.g., Syria, including Dehes on the limestone massif), including at sites across western Asia Minor. 10 Klingborg 2017, 65, with extensive catalogue. 11 Crouch 1984, 355; Tsakirgis 1984, 334–341; Wulf 1999, 20f; Garbrecht 2001, 16–40. 12 The houses at Notion were studied by Christopher Ratté and will be discussed in the final publication of the Notion Archaeological Survey.

68  Angela Commito the city, this house-like structure does have features in common with what seem to be other peristyle houses on the site, and, at 324 m2 in area, is a useful example falling somewhere between the two main conjectured house sizes. With its open courtyard of almost 39 m2, the structure would offer about 285 m2 of roof surface. It is possible that the entire roof surface was used as a catchment for the cistern. However, if the roof over the northern side of the house were slanted away from the courtyard, draining instead into a street or sewer, as has been reconstructed for houses at Olynthus, then the catchment area would be reduced to around 177 m2.13 Assuming that average annual rainfall in antiquity was close to modern precipitation records of 700 mm per year,14 the two possible roof catchments would yield water volumes of almost 200 and 124 m3, respectively. Using a 30% rate of loss to account for evaporation and for absorption by ceramic rooftiles,15 the volume of water available for collection in a cistern would be reduced to about 140 and 87 m3 annually. Although the cistern located under this structure is filled in, those that remain open have an estimated capacity of 30–50 m3, meaning that even the smaller possible roof catchment could have supplied enough water to fill the cistern twice during the year. A single filling of 30–50 m3 of water per cistern could supply 8–14 persons per year, assuming one person requires 10 l of water a day, or 4–7 persons per year, using a higher consumption rate of 20 l per day.16 Though water availability was not constant, dependent as it was on precipitation patterns with considerable seasonal, annual, and other temporal variations, the results of modelling by Patrik Klingborg and Martin Finné underscore that such variability was itself predictable, and that, if users modified their consumption in expectation and response to these variations, cisterns could provide a sustainable and reliable water supply.17 Did the residents of Notion have access to additional water sources? Though running water was available near Notion in the form of the Hales River, it is located at an elevation too low to have supplied the city, and in antiquity the sea likely penetrated farther inland than it does today, making the river an inconvenient and inappropriate source of water for the new town. In 2022, we were able to investigate what local inhabitants had long reported to be the remains of a terracotta pipeline on the hill northwest of the city. We documented substantial remains of 13 Klingborg 2017, 78, who cites and follows Cahill 2002, Fig. 23, in whose reconstruction the northern section of the house roof is angled down to a drainage alley. 14 Average annual rainfall in Izmir for the years 1981–2010 was 688.6 mm (Izmir province). https:// www.mgm.gov.tr/veridegerlendirme/il-ve-ilceler-istatistik.aspx?k=H&m=IZMIR and for 1938– 2019 was 711.1 mm (Izmir province)​://www.mgm.gov.tr/veridegerlendirme/il-ve-ilceler-istatistik. aspx?k=A&m=IZMIR. The problem with these figures is that precipitation varies significantly within the region. However, other sources cite a figure in the upper 600s mm: see, e.g., Türkeş & Erlat 2003, 1779, Table 1, for Izmir (the city, not the province), and note the comparable figures provided therein for other Aegean coastal sites such as Bodrum. 15 A 30% rate of loss is an ‘arbitrary figure’ proposed by Connelly & Wilson 2002, 287, and accepted by Klingborg 2017, 78–79, and followed here. 16 See Klingborg 2017, 80–82, for discussion and references concerning water consumption rates. 17 Klingborg & Finné 2018.

Surveying Notion’s Residential Water Supply  69 what must have been an aqueduct supplying Notion for a distance of about 2 km along the hills to the northwest of the city. The pipeline likely ran under pressure in an inverted siphon from the point where it left the hill to the northwest (elevation 110 m), across the Hales River valley, and up to the site.18 The date of the aqueduct has not yet been determined, but multiplicity in methods of water supply is not uncommon at comparable Hellenistic towns. Many, though not all, of the houses at Olynthos had cisterns, but water was also available from at least two public fountains, supplied by long-distance terracotta pressure pipelines.19 At Morgantina, residents used water piped in from springs, brought up from wells, and collected in cisterns.20 This redundancy has been interpreted as a way to make sure the bestquality water was available for drinking. This preference seems to be visible in the archaeological evidence at some sites. At Priene, for example, a site with abundant natural springs, most houses used water from fountains or had a piped-in supply, and only houses located at the outer edge of the town plateau, out of the range of these running sources, were equipped with cisterns.21 However, although ancient sources on the whole seem to prefer springs and sources of running water to rainwater collected in cisterns, cistern water was nevertheless considered appropriate for drinking and in some cases even recommended.22 In addition to the remains of the terracotta pipeline on the hill northwest of the city, a number of stone blocks that appear to have been used for water transportation or management have been found on the site. One (found near cistern 18) likely served as a junction block for a terracotta pipeline, fragments of which are still attached to the mortar that would have originally held the pipe in place (Figure 4.14). Two others (found near cistern 13) could have belonged to a stone pressure line (Figure 4.15). In contrast to the junction block, which is made from conglomerate stone, these are fashioned from local marble and have traces of mortar but no terracotta pipe. Since the diameter of the hole, 9 cm, matches the diameter of many of the terracotta pipes found in situ in the cisterns and elsewhere as surface funds, it is also possible that these blocks, too, were integrated into a terracotta pipeline. In addition, a lead pipe situated underneath a floor was revealed by illegal excavation. Finally, blocks found in the eastern portion of the Agora, just south of the axis created by the main east-west street running across the city, must have belonged to a fountain; parapet slabs at the entrance to the Sanctuary of Athena seems also to have belonged to a fountain. Taken together, these elements of the urban water

18 Stone arches in the Hales River valley north of Notion were noted and identified as part of an aqueduct in Demangel & Laumonier 1923, 358, n. 3, referring to Frontieros 1878–80, 189–190. Structures that match these descriptions have been recently observed by local residents but not yet examined by the research team. 19 Robinson 1930, 11–14 (fountain in southeast part of South Hill); Robinson 1946, 95–114 (fountain in southeast corner of agora); Crouch 1996, 171–176 (pipelines, baths, fountains); Klingborg 2017, 240–246 (cistern catalogue); Cahill 2002. 20 Crouch 1984. 21 Crouch 1996, 167; Fahlbusch 2003. 22 Klingborg 2017, 83–86, for review and discussion of ancient sources.

70  Angela Commito

Figure 4.14 Conglomerate junction block for a terracotta pipeline. Notion Archaeological Survey.

Figure 4.15 Marble block, possible part of a pressurized stone pipeline. Notion Archaeological Survey.

Surveying Notion’s Residential Water Supply  71 infrastructure suggest that a piped-in water supply complemented the supply from rainwater collected in cisterns. Even though Notion at some point was apparently supplied by a long-distance pipeline, it never seems to have been equipped with the kind of large-scale aqueduct that many towns across Western Asia Minor built during the first and second centuries AD, often with support and expertise from the Roman state. Notion may have been largely depopulated by that time, before it had a chance to participate in the aqueduct-building boom of the Roman period. At Notion, there are no obvious remains of bath complexes, and there is very limited use of mortared rubble construction. Both of these are hallmarks of Roman-period urban development, as are large-scale aqueducts. Furthermore, as noted earlier, very little of the identifiable surface pottery examined so far post-dates the first century AD. Pulling these clues together, then, one gets the sense of a city largely frozen in time, with little evidence of activity after the first century AD. The water supply system documented so far fits into an emerging picture of an urban population that did not survive long enough at a sufficiently large size to alter its urban infrastructure to meet changing region-wide behaviours regarding water use. Water Supply and Urban Biography The results of the survey, combined with considerable historical and epigraphic evidence, offer a vivid picture of the life of the ancient urban community at Notion. References by Thucydides and Aristotle make it clear that Notion (the ‘southern place’) was the port of Colophon, situated 15 km to the north.23 In addition to emphasizing Notion’s relationship with Colophon, the ancient sources also highlight its importance as a military port. An important sea battle fought off the coast of Notion in 406 BC resulted in the downfall of the Athenian general Alcibiades,24 and in 190 BC the military port was besieged by Antiochus the Great.25 A century earlier, Colophon had been attacked and destroyed in the early third century BC and its population deported to Ephesus by Lysimachus.26 Inscriptions from Claros and other sites tell us that at least some of the Colophonians were allowed to return home within a decade or so of the destruction of their city,27 and that both Colophon and Notion continued to coexist as separate but allied communities through the first half of the third century, but by the middle of the third century at the latest, Notion had begun to acquire the name of New Colophon or Colophon-by-the-Sea.28 By the end of the third century, epigraphic references to Notion disappear, and it is evident that Notion had in effect become Colophon.

23 Thuc. 3.34; Arist. Pol. 1303b, 10. 24 Xen. Hell. 1.5.11–15; Diod. 13.71.1; Hell. Oxy. 4, 1–4; Plut. Lys. 5. 25 Livy 37.26.5, 8. 26 Paus. 1.9.7; 7.34. 27 Robert & Robert 1989, 83–85. 28 Étienne & Migeotte 1998.

72  Angela Commito The ‘re-founding’ of Notion was likely initiated by the destruction of Colophon by Lysimachus. Roland Étienne and Léopold Migeotte have proposed that a significant portion of the population of Colophon may have chosen to move from Ephesus to Notion rather than back to Colophon.29 The relocation of Notion on a naturally defensible site overlooking an important military harbour may therefore have been an attempt by this community to consolidate its scattered members in a new location where they envisioned a more secure and prosperous future. What is more surprising than the ‘re-founding’ of Notion, however, is the large-scale abandonment of the city during a time of greater prosperity in Western Asia Minor than ever before. Ephesus may have attracted significant proportions of the populations of local communities, including Notion. In the relative stability of the early Roman period, the site’s strategic advantages, especially its defensibility, may have become outweighed by its disadvantages, notably the lack of fresh-water sources. While we know from historical sources that Notion continued to exist into the early Christian era, it must have been a much smaller community, perhaps situated outside the walls in a ‘lower town’ located closer to the harbour. The construction and use of the water supply system, with its dependence on cisterns and piped-in water but no large-scale aqueduct, appears to overlap with the period of greatest occupation of the city, between the third century BC and the first century AD, indicated by architecture and surface pottery. The water supply of Notion therefore demonstrates how residents created an urban infrastructure in accordance with the site’s challenging topography and provides indirect evidence for the chronology of the city’s lifespan. In writing the biography of Notion, a city that does not fit universalizing models based on long-lived cities such as Athens or Rome, it is possible to highlight the varieties of ancient urban experience, and to showcase the importance of cisterns as a water supply strategy. Rather than applying a value system based on longevity to the study of an urban community, research at Notion has focused instead on careful consideration of the many moments when its members exchanged one vision of their future for another in response to changing circumstances. At Notion, these decisions resulted in neither success nor failure per se, but in a community repeatedly transformed by the varied desires of its members. The story of water at Notion parallels the biography of this community. References Atalay, E. 1986. ‘1985 Yılı Notion Kazıları’, in VIII. Kazı Sonuçları Toplantısı 2, Ankara, 249–264. Atalay, E. 1987. ‘1986 Notion Kazıları’, in IX. Kazı Sonuçları Toplantısı 1, Ankara, 147–169. Büyükkolancı, M. 1996. ‘1994 Yılı Notion Kazıları’, in VI. Müze Kurtarma Kazıları Semineri 1995 Kuşadası, 371–381. Cahill, N. 2002. Household and city organization at Olynthus, Princeton.

29 Étienne & Migeotte 1998, 149f.

Surveying Notion’s Residential Water Supply  73 Connelly, J.B. & A.I. Wilson 2002. ‘Hellenistic and Byzantine cisterns on Geronisos Island’, in Report of the department of antiquities, Cyprus, 278–280. Crouch, D. 1984. ‘The Hellenistic water system of Morgantina, Sicily. Contributions to the history of urbanization’, AJA 88(3), 353–365. Crouch, D. 1996. ‘Priene’s streets and water supply’, in Cura Aquarum in Campania. Proceedings of the ninth international congress on the history of water management and hydraulic engineering in the Mediterranean region, Pompeii, 1–8 October 1994, eds. N. de Haan & G.C.M. Jansen. BABesch Suppl. 4, Leiden, 137–143. Demangel, R. & A. Laumonier 1923. ‘Fouilles de Notion (1921)’, BCH 47, 353–386. Demangel, R. & A. Laumonier 1925. ‘Fouilles de Notion (1921) deuxième partie’, BCH 49, 322–346. Étienne, R. & L. Migeotte 1998. ‘Colophon et les abus des fermiers des taxes’, BCH 122, 143–157. Fahlbusch, H. 2003. ‘Wasserwirtschaftliche Anlagen des antiken Priene’, in Wasserhistorische Forschungen. Schwerpunkt Antike, ed. C. Ohlig. Schriften der Deutschen Wasserhistorischen Gesellschaft 3, Siegburg, 55–80. Frontieros, A. 1878–80. Mouseion kai vivliothēkē tēs Euangelikēs scholēs 3, Smyrna. Garbrecht, G. 2001. Altertümer von Pergamon I:4. Stadt und Landschaft. Die Wasserversorgung von Pergamon, Berlin. Klingborg, P. 2017. Greek Cisterns. Water and Risk in Ancient Greece, 600–50 BC, Ph.D. thesis, Uppsala University. Klingborg, P. & M. Finné 2018. ‘Modelling the freshwater supply of cisterns in ancient Greece’, Water History 10, 133–131. Price, R. & M. Gaddis, eds. 2007. The Acts of the Council of Chalcedon. Translated Texts for Historians 45, Liverpool. Ratté, C., A. Commito & P. Knoop 2018. ‘Notion Archaeological Survey, 2016’, in 35. Araştırma Sonuçları Toplantısı, Bursa, 291–302. Ratté, C., F. Rojas & A. Commito 2017. ‘Notion Archaeological Survey, 2014–2015’, in 34. Araştırma Sonuçları Toplantısı, Edirne, 617–638. Ratté, C., F. Rojas & A. Commito 2020a. ‘New Research at Notion’, in Zwischen Bruch und Kontinuität. Architektur in Kleinasien am Übergang vom Hellenismus zur römischen Kaiserzeit, eds. U. Lohner-Urban & U. Quatember. Byzas 25, Berlin, 345–362. Ratté, C., F. Rojas & A. Commito. 2020b. ‘Notion Archaeological Survey 2017–2018’, in 37. Araştırma Sonuçları Toplantısı, Diyarbakır, 333–353. Robert, J. & L. Robert. 1989. Claros 1. Décrets héllenistiques, Paris. Robinson, D.M. 1930. Excavations at Olynthus, part 2. Architecture and sculpture: houses and other buildings. Johns Hopkins University Studies in Archaeology 9, Baltimore. Robinson, D.M. 1946. Excavations at Olynthus, part 12. Domestic and public architecture. Johns Hopkins University Studies in Archaeology 36, Baltimore. Tsakirgis, B. 1984. The domestic architecture of Morgantina in the Hellenistic and Roman periods. Ph.D. diss., Princeton University. Türkeş, M. & E. Erlat. 2003. ‘Precipitation changes and variability in Turkey linked to the North Atlantic oscillation during the period 1930–2000’, International Journal of Climatology 23, 1771–1796. Wulf, U. 1999. Altertümer von Pergamon VX:3. Die Stadtgrabung. Die hellenistischen und römischen Wohnhäuser von Pergamon, Berlin.

5

Breaking Out from Imagined Household Uniformity Diverse Rainwater Harvesting Solutions in Republican-Imperial Cosa Ann Glennie

Introduction Cosa, a colony of Latin status, was founded in Southern Tuscany by the Romans following the confiscation of land from the recently defeated, indigenous Etruscans in 273 BC.1 A 13-hectare settlement was implanted on a coastal promontory, rising 114 m above sea level, and was augmented by a small port below. Excavations of the site began in 1948, led by Frank Brown under the auspices of the American Academy in Rome (AAR). Brown’s campaigns, which extended into the late 1980s, revealed environs including the large, polygonal fortification walls with towers, the ‘arx’, a religious precinct on the highest point of the colony, the forum, replete with civic, commercial, and religious buildings, an adjacent bathhouse, and, downslope from the forum, a couple of blocks filled with domestic structures. He concluded that the settlement had persevered from the third century BC, but never prospered. It had a renewal just after 200 BC, experienced a disruption in 70 BC, and renewed again in the Julio-Claudian period before it declined in the fourth century AD.2 The narrative developed in the site’s publications relied heavily, thanks to the author Aulus Gellius, who in the second century AD declared colonies miniatures or copies of Rome,3 on finding parallels from ancient accounts about the mid-Republican city of Rome to flesh out the archaeological remains of her colony. Thus, Cosa was promoted as a type site of middle Republican Roman settlements and even, cyclically, as a model for mid-Republican Rome itself. As academic assessment of Roman colonization advanced beyond the Gellian mode, scholars found that as much as Cosa could be a valuable source for early Roman city planning and architecture, it was not a universal colonial blueprint or a ‘mini-Rome’. This re-evaluative trend was complemented by further archaeological investigation at Cosa from 1991 to 1997 led by Elizabeth Fentress under the auspices of the AAR.4 Fentress’s work clarified unsupported, generalized claims

1 Fasti Triumphales, tablet II (Degrassi 1954, 98). 2 Brown 1951; 1980; Brown et al. 1960; 1993; McCann et al. 1987; Bruno & Scott 1993. On the 70 BC disruption, see footnote 34 below. 3 Gel. 16.13. 4 Fentress 2003. DOI: 10.4324/9781003268222-5

Breaking Out from Imagined Household Uniformity  75 by Brown about the Republican chapter of the colony; for instance, her team clarified, contra Brown’s assumption, that second-century BC atrium-style structures on the forum were in fact atrium-style houses (see below). Additionally, Fentress highlighted the staying power of this settlement, documenting another Severan period revival not emphasized by Brown, and observing that though it experienced periods of population contraction and settlement nucleation, it was not abandoned until the fourteenth century AD, rather than Brown’s supposed fourth century AD. Despite the attention rightly given to the environs and chronology of the colony, the importance of Cosa’s rainwater management systems has been largely underplayed. Though the surrounding territory was well-watered, including freshwater springs at the port,5 the hilltop on which Cosa lies is completely waterless; it has no natural source of freshwater and was never provisioned with an aqueduct.6 In the publication of his first survey and mapping of Cosa’s unexcavated remains, Brown writes: It may be said without exaggeration that in ancient Cosa practically every public building and every private dwelling possessed at least one cistern for the storage of water. On the town plan some forty-five have been plotted in broken outline. A score more are known and many lie as yet undiscovered.7 As Brown noted, in addition to the traditional domestic rainwater harvesting scheme prevalent throughout the ancient Mediterranean, Cosa has a non-domestic rainwater harvesting system in order to guarantee the colony’s water supply.8 A separation exists, however, between the limited excavation and definitive conclusions forwarded about Cosa’s domestic rainwater harvesting system. Eight larger houses, the Atrium Buildings, have been identified on the forum of Cosa, but only two have been completely excavated. Two blocks of smaller houses downhill from the forum have been explored, but only six of the houses, not all contemporary, are well understood. From these handful of excavated domestic structures, two Cosan house types have been identified. Based on these it has been inferred that the internal arrangement, including the water management mechanisms, of all houses throughout the colony followed exactly one of these two plans. Closer examination shows the diversity of house layout and water management solutions among the examples of each type. This chapter presents the state of the evidence 5 See McCann et al. 1987 on the springs and a springhouse associated with one of them. 6 Despite aqueducts which utilized springs in the hinterland of and region surrounding Cosa (Bruun 1991, 273–284; Calastri 2007), the site itself never had one. 7 Brown 1951, 85. These cisterns are, however, not the only features which appear in broken outline on the corresponding plan, creating difficulty in pinpointing them. 8 Because of Cosa’s waterless nature, a supplemental non-domestic rainwater harvesting system was implemented. In total, ten large-scale reservoirs and cisterns, built between 273 and 140 BC, have been best explored and documented across the colony. These containers could hold over 3.5 million litres of water at full capacity. They seemed to have been maintained and remained functional throughout the Republican and Imperial periods, falling out of use in Late Antiquity. For more on this topic, see Glennie 2022.

76  Ann Glennie of Cosa’s domestic rainwater harvesting systems,9 stresses the importance of collecting, publishing, and evaluating data associated with ancient water management systems, and discusses what the diversity of the rainwater harvesting systems at Cosa means for its role in the study of Roman colonization and Roman archaeology more broadly. The Atrium Buildings As briefly alluded to above, two major domestic quarters have been excavated at Cosa, one of which fronts three sides of the colony’s forum (Figure 5.1). The forum

Figure 5.1 State plan of Cosa with Atrium Buildings and East and West Block Houses. Cosa Excavations 2018, adapted. 9 There is also a robust water disposal system at Cosa, both within individual houses, utilizing cesspools—sometimes several per structure (e.g., Brown et al. 1993, 59–97, 104–106, 135–138, 238–241; Bruno & Scott 1993, 21, 38–42, 47–49, 69–71, 81–82, 172–173, 181–183; Fentress 2003, 14–19, 34–55)—, and probably colony-wide, as represented by the small portion of a sewer system which has been investigated (Bruno & Scott 1993, 103, 124–127, 154–155). The numerous wastewater disposal features indicate a clear concern for removing overflow water in both domestic and non-domestic settings. Though some water probably could not have been reused, the absence of mechanisms for reusing greywater might in turn suggest that water scarcity was not an issue; this hypothesis needs to be tested out by more rigorous investigation.

Breaking Out from Imagined Household Uniformity  77 houses were explored by Brown and his team, who, though they only excavated Atrium Building I fully, established the existence of eight structures similar to the iconic atrium-style house known from other Roman sites.10 Not believing that houses would be located along a Republican forum, however, they were attributed various civic, judicial, production, and mercantile functions;11 excavators argued that during the Julio-Claudian revival of Cosa some of these became domestic structures, easily adapted because of their atrium-style design.12 Later Fentress and her team excavated Atrium Building V, or the House of Diana,13 demonstrating that since the inception of the structure in the early second century BC it had served a domestic function and continued to do so through various renovations into the first century AD.14 Though no other Atrium Buildings were completely reexplored by Fentress’s team, it seems more probable that the rest of these structures with their atrium-style plans were likewise domestic from their inception. Because of this excavation history, Atrium Building I and Atrium Building V/House of Diana provide the most complete pictures of domestic rainwater harvesting in Cosa’s Atrium Buildings in the middle Republican and early Imperial periods.15 The Republican Atrium Buildings Atrium Building I is a large structure (22.72 × 17.32 m; Figure 5.2)16 located on the northern side of Cosa’s forum, just to the east of the forum’s entrance. During the middle Republican period, it had two tabernae (Rooms 2 and 5) on its forumfacing façade flanking its fauces (Room 4) which opened into a compluviate atrium (Room 10). Off the atrium were several other rooms, including one facing westward to a perpendicular street (Room 7). The opposite façade had three more tabernae (Rooms 13, 15, and 19) facing outward to a parallel street behind the forum.17 Roughly a generation after being begun, an eastern annex to Atrium Building I was

10 Brown et al. 1993, 57–97, 237–241. As demonstrated by the volume’s state plan, the team only fully excavated Atrium Building I (Brown et al. 1993, Plan II). That only Atrium Building I was fully excavated is noted elsewhere by Brown (1980, 33). Other than some limited soundings in Atrium Building VII, the remainder of the Atrium Buildings investigated by Brown et al. (1993, Plan II) had at best small portions of their forum-facing facades and perhaps one exterior perpendicular wall traced. Clarification about the limited excavations of these other structures is offered elsewhere by Brown (1980, 33–34). 11 The Atrium Buildings are sometimes also referred to in publication as the ‘Atria Publica’; Brown 1980, 35–36; Brown et al. 1993, 70, 93, 97, 132; cf. Fentress 2000, 19–20; Sewell 2013, 82–94. 12 Brown et al. 1993, 238–241. 13 In initial excavation, the area was dubbed by Fentress and team ‘Forum V’ (Fentress 2003, 6) and later restyled the House of Diana after a dedicatory inscription to the goddess and an under-lifesized torso of her were discovered in excavation (ibid., 45–54). 14 Fentress 2003, 14–23, 34–55. 15 Limited information is presented about water features in the Atrium Buildings which were not totally excavated: on a cistern of Atrium Building II, see Brown et al. 1993, 78, 240; on the impluvium and cistern of Atrium Building VII, see ibid., 84, 86. 16 Brown et al. 1993, 59. 17 Brown et al. 1993, 61, Fig. 20.

78  Ann Glennie

Figure 5.2 Reconstructed plan of Republican Atrium Building I and North Corner Plot. Brown et al. 1993, Figs. 20 and 55, combined and adapted.

completed on the adjacent side18 which was accessed through a corridor (Room 12). Though the North Corner Plot is characterized by the excavators as its own entity, it seems possible based on this connection that in antiquity it was considered a part, or at least a subsidiary, of Atrium Building I. Aligned with the centre of the fauces and roughly in the centre of Atrium Building I’s opus signinum floored atrium was a herringbone brick tiled impluvium.19 In addition to water entering the compluvium, an underfloor channel opening beneath 18 Brown et al. 1993, Fig. 55, Plates 73–77 designate the northern half of the plot the ‘N corner plot’ and the southern half the ‘NW Porticus Annex’. Separating them in thought and designation may come from the apparent construction of the northern half about a generation before the southern half. Yet, to me it seems more useful to see these two plots as connected, so I will refer to them collectively as the North Corner Plot. 19 Brown et al. 1993, 62–63.

Breaking Out from Imagined Household Uniformity  79 the wall of the room which faced outward to the perpendicular street (Room 7) also brought water into the impluvium,20 an otherwise exceptional arrangement. A pipe with rectangular section fed water into a cistern vaulted with thirteen unmortared voussoirs.21 In the centre of the top of the vault was a drain to direct overflow into a cesspool (Room 19).22 Water could be withdrawn from the cistern via a hexagonal drawshaft aligned with the northeast corner of the impluvium; the puteal for this drawshaft does not survive.23 The North Corner Plot was built in two phases beginning with the northern half. Within that space are two tabernae (Rooms 24/26 and 25). Room 26 has a cistern beneath it, from which water could be drawn, possibly through a Vulci tufo puteal, as represented by a fragment found on the street just outside the building.24 This cistern was originally supplied by an underfloor channel carrying rainwater falling on the pavement behind Room 24. When the southern addition to the North Corner Plot was built about a generation later, a small, basin-like catchment was added inside the new room (Room 21), likely fed by a downspout, to continue to supply the cistern.25 Atrium Building V/House of Diana (22.13 × 17.25 m; Figure 5.3) is located on the forum’s long western side.26 The Republican imprint of Atrium Building V/ House of Diana reveals a somewhat different layout from Atrium Building I but similar hydraulic accommodations. The fauces and two flanking tabernae (rooms C and D) face the forum. The central atrium (room B) is surrounded by various connected spaces which radiate off of it on three sides. At the back of the interior space is an open garden plot with two additional environs (rooms N and S). The atrium had a compluviate roof which allowed water to fall into the complementary impluvium, feeding, in turn, the ‘large cistern’ below.27 Water could be withdrawn through a drawshaft along the southwestern side of the impluvium for which no puteal has

20 Brown et al. 1993, 67–68. 21 See the contributions by Commito and Klingborg, this volume. 22 Brown et al. 1993, 64. In Cosa’s published scholarship the terminology of ‘soak away pit’ and ‘cesspool’ seem to be used interchangeably to describe subterranean bedrock cut or partially stone constructed features without opus signinum lining for the disposal of both solid and liquid waste. For ease and consistency, I refer to all these kind of waste disposal features as cesspools. 23 Brown et al. 1993, 64; Collins-Clinton 2020, 7 (though mistakenly labelled here as P 10), 208, 219–220 has tentatively identified Puteal 12 (P 12) as being a fragment of Atrium Building I’s puteal because of its hexagonal shape. 24 Brown et al. 1993, 105–106; Collins-Clinton 2020, P 1 (212–213). 25 Brown et al. 1993, 137; see also Jansen, this volume. 26 Fentress 2003, 19 n. 19, 20 27 Fentress 2003, 16. See also Rabinowitz & Fentress 2002, https://www.press.umich.edu/special/ cosa/f5_p_i.html, incl. pl. 49. No dimensions are provided for this cistern, but speculating based on analogy with other domestic cisterns at Cosa and comparison of the assumedly same stadia rods used in other aboveground photography during Fentress’s campaigns which seem to be two meters in length (each subsection being 50 cm), the cistern is just over two meters in width and just under three meters in height (with unrecorded length and unrecorded dimensions of the transverse passage). This estimate, however, may not align with the characterization of the cistern as ‘large’.

Figure 5.3 Reconstructed plan of Republican and Augustan Phases of Atrium Building V/House of Diana. Fentress 2003, Figs. 5 and 15, adapted.

80  Ann Glennie

Breaking Out from Imagined Household Uniformity  81 been identified.28 In the back garden a small bathhouse, represented by room N, a terracotta tiled floor on which some kind of standalone tub could have been positioned, and room S, an elevated cistern for the provisioning of water to the bath (and possibly the garden?), is located in the southeast corner of the house plot.29 The back garden also had a niche cut into its southwestern wall which could have contained a small (c. 0.6 m length) basin30 and, slightly later in the Republican period, above it a fountain.31 It is speculated that room S and the basin/fountain would have been fed by water from the adjacent Street 5 which was at a higher elevation.32 Since the excavation of Atrium Building V/House of Diana was conducted with more rigorous adherence to stratigraphy, the nuances in the phasing of the building are better illuminated, demonstrating at least three different Republican phases. Fentress also considers the process of the original construction of Atrium Building V/ House of Diana in relation to its subterranean hydraulic features, noting how the cistern would have been excavated first, ‘serving the dual purpose of creating necessary water storage under the atrium and of cutting sufficient [sic] large blocks of bedrock to provide foundations for the walls’.33 A similar principal probably applied to the cesspools excavated beneath Atrium Building V/House of Diana and the concept could likely be broadly extended to the simultaneous creation of water containers and of building materials from the excavation of all subterranean spaces across Cosa. The Imperial Atrium Buildings Excavation campaigns in both Atrium Building I and the annexed North Corner Plot as well as in Atrium Building V/House of Diana demonstrate major overhauls during the Imperial period. Both houses show initial revamping in the Augustan/ Julio-Claudian era, when the colony as a whole seems to have experienced a rebirth after its destruction and partial abandonment around 70 BC.34 Atrium Building I now certainly absorbed the entirety of the adjacent North Corner Plot, if it had not earlier. It appears, in the absence of comment to the contrary, however, that the water collection system must have remained consistent with the Republican version of the house. No date for the abandonment of Atrium Building I and the North Corner Plot is given, but two rooms (Rooms 16 and 22) were destroyed by an earthquake;

28 Collins-Clinton 2020, 208–209. 29 Fentress 2003, 19. 30 Fentress 2003, 18. For the fragments of a stone basin that possibly sat in the niche, see CollinsClinton 2020, 199–201 (B 6a, B 6b, and B7). 31 Rabinowitz & Fentress 2002, https://www.press.umich.edu/special/cosa/f5_p_iia2.html. 32 Rabinowitz & Fentress 2002; Fentress 2003, 18, https://www.press.umich.edu/special/cosa/f5_p_ i.html for unusual bedrock platform associated with the bathrooms that is suggested to have possibly served as a settling basin for the elevated cistern (room S). 33 Fentress 2003, 20–21; cf. Bruno & Scott 1993, 17; see also Klingborg, this volume. 34 For more on this destruction, see Brown 1980, 73–75; Brown et al. 1993, 237; Bruno & Scott 1993, 147, 156–157, 201, 205; Fentress 2003, 32–34. Slane, however, has shown that there was no disruption at Cosa’s port in 70 BC, thus questioning the extent of this supposed disruption (Will & Slane 2019, 117).

82  Ann Glennie coins provide a terminus post quem of c. AD 37,35 but Fentress suggests that this event might have been the earthquake felt in Rome in AD 51.36 That suggests the house fell into at least partial disuse by the mid-first century AD. During the Augustan period, the bathrooms (rooms N and S) in the Atrium Building V/House of Diana were rebuilt, but not dramatically altered.37 The back garden was also transformed from a cultivation to a leisure area, and now bore an open gutter to collect rainwater coming off the eaves north of the outdoor space; the outlet for water collected by it could not be discerned in excavation, however, possibly having been robbed out.38 During the Augustan remodelling, most rooms were repaved, including the fine opus signinum flooring of the atrium and fauces beneath which was placed a drain to prevent overflow of the impluvium.39 A short time after, in the Claudian period, the impluvium was modified by a low wall added around it which deepened the pool;40 the overflow drain may have fallen out of use after this alteration. Not too long later, c. AD 80, the house was abandoned.41 The East and West Block Houses The second excavated domestic quarter lies about two blocks diagonally downslope from the forum (Figure 5.1). Portions of two blocks of these smaller houses there were investigated solely by Brown’s team (Figure 5.4). The remains of the houses lie on either side of one of Cosa’s streets, designated Street M, and were divided in publication into the houses of the East Block and of the West Block.42 Based on the area excavated, the two blocks are thought originally to have been divided into ten plots meant to each hold a house; however, in practice it was more common that only the upper plot of each block was occupied by a house—fronted onto Street N for the East Block and Street M for the West Block—and that the lower plot was occupied by a back garden, making each block more like five plots.43 Past publication focuses largely on the West Block’s Lots 4 and 5 and the East Block’s Northeast and Southwest Houses (on equivalent grid to Lots 2 and 3 of the West Block). Many revisions to their ground plans over time and modern agricultural activity have created a complicated stratigraphy in this area.

35 Brown et al. 1993, 241. 36 Fentress 2003, 55–57. 37 Rabinowitz & Fentress 2002, https://www.press.umich.edu/special/cosa/f5_p_iiia2.html. 38 Rabinowitz & Fentress 2002, https://www.press.umich.edu/special/cosa/f5_p_iiia2.html. The feature is not addressed in the print publication, but is on the Augustan plan (Fentress 2003, Fig. 15). See Jansen, this volume, on water collection in gardens. 39 Rabinowitz & Fentress 2002, https://www.press.umich.edu/special/cosa/f5_p_iiia.html. 40 Fentress 2003, 38–39; Rabinowitz & Fentress 2002, https://www.press.umich.edu/special/cosa/ f5_p_iv.html. 41 Fentress 2003, 63; Rabinowitz & Fentress 2002, https://www.press.umich.edu/special/cosa/f5_p_v_ house.html. 42 Bruno & Scott 1993, 6. 43 Bruno & Scott 1993, 13–14.

Breaking Out from Imagined Household Uniformity  83

Figure 5.4 State plan of the East and West Block houses. Bruno & Scott 1993, Fig. 3, adapted.

The Republican East and West Block Houses Excavators discerned three Republican phases for the West Block and two for the East Block. The Republican structures were predominately non-compluviate atrium-style houses,44 but do have a sort of central room/enclosed atrium with other rooms built around and behind it. The water supply and disposal are individualized in placement, shape, and dimensions by plot, perhaps customized to a patron’s tastes or more practically positioned to utilize pre-existing fissures in the bedrock. For instance, the West Block’s Lots 2 and 3 share a cistern, while excavators hypothesized that Lots 4 and 5 may have been occupied by one connected structure because, though most lots have both a cistern (or cistern access) and a cesspool for waste disposal, Lot 4 does not have a cistern and Lot 5 does not have a cesspool.45 These unique differences demonstrate that though built on plots of relatively the same size, the houses did not have identical interior designs.46 In the first Republican phase (c. first quarter of second century BC; Figure 5.5), cisterns, usually located towards the front portion of the house, were filled via an underfloor channel running from the cistern through the house’s entryway where it could gather water from the front eaves. The cisterns had an ovoid or square 44 Bruno & Scott 1993, 19. 45 Bruno & Scott 1993, 21. 46 See also Klingborg, this volume.

84  Ann Glennie

Figure 5.5 Reconstructed plan of Republican Phase 1 West and East Block houses. Bruno & Scott 1993, Figs. 5 and 10, adapted.

settling basin connected to this channel47—a feature not attested in the forum’s Atrium Buildings—and probably provided since in this location water was being collected at ground level instead of from the roof and therefore would have picked up more debris. Though its puteal does not survive, a drawshaft near the northeast end of Lot 5’s cistern was surrounded by a shallow square spill basin with raised margin connected to a drain which would take overflow into a cesspool;48 the drain and corresponding cesspool are not articulated on the reconstructed Republican Phase 1 house plan, but based on the state plan would have emptied into Lot 4’s cesspool (Figure 5.4). The cistern vault of Lots 2 and 3 of the West Block had fallen in and its drawshafts were not preserved.49 In the East Block, the Northeast House originally had a drawshaft located over a meter from the southwestern wall (Figures 5.4 and 5.5).50 During the second Republican phase (175–early 1st c. BC; Figure 5.6), in the West Block, the garden plots of Lots 4 and 5 were both developed into houses.51 The Lot 5 garden plot house had a slightly trapezoidal cistern with a shallow bottom depression (50 cm in diam. and 8 cm in depth),52 provisioned by two underfloor 47 Bruno & Scott 1993, 18, 36. 48 Bruno & Scott 1993, 19. 49 Bruno & Scott 1993, 19. No information is provided about the contemporary Southwest House’s drawshaft. 50 Bruno & Scott 1993, 43. 51 Bruno & Scott 1993, 65. 52 Bruno & Scott 1993, 68. The original excavators use the terminology ‘sediment basin’, but I prefer bottom depressions (see Klingborg 2017, 48–50). Though Bruno & Scott (1993, 19) generalize that

Breaking Out from Imagined Household Uniformity  85

Figure 5.6 Reconstructed plan of Republican Phase 2 West and East Block houses. Bruno & Scott 1993, Figs. 10 and 19, adapted.

channels, one to the front of the house and one in what must have been a small outdoor space at the back of the building. Water falling into this outdoor space also fed a channel to the cistern of Lot 5 proper. The drawshaft of the Lot 5 garden plot house, located at the northwest end of the cistern, was surrounded by a trapezoidal spill basin with raised margin; the overflow drain for the spill basin was not detected in excavation, but hypothesized to also run out the front of the building possibly following the line for one edge of the entryway. The puteal does not survive. Major remodelling occurred in both blocks during the third Republican phase (early first century BC; Figure 5.7). In the West Block, Lots 4 and 5 were definitively combined (in the unlikely event they had not been before) into one house, known as both the House of the Treasure, after a hoard of 2,004 coins found beneath it, and the House of Quintus Fulvius, after its possible owner (henceforth referred to as Lots 4 and 5 house).53 The northwest garden plots of the East Block, meanwhile, were occupied mostly by one large atrium-style house with expansive back garden, reoriented perpendicularly to the original blocks to front onto Street 5, and known from the human remains found in its cistern as the House of the Skeleton.54 By this period, if not earlier, water from the roofs of the Lots 4 and 5 house was funnelled into the cistern intakes by means of the terracotta roofing scheme,

each of the East and West Block cisterns had a bottom depression, only those of the Lot 5 garden plot house and Northeast House (later part of House of the Birds) are detailed in publication. 53 Bruno & Scott 1993, 79–80. 54 Bruno & Scott 1993, 79, 99–152. There is limited discussion of a house neighbouring the House of the Skeleton to the southwest and some facets of its water system, including a kitchen and bath complex with drainage pipes (ibid., 153–158).

86  Ann Glennie

Figure 5.7 Reconstructed plan of Republican Phase 3 East and West Block houses. Bruno & Scott 1993, Figs. 22 and 32, adapted.

as evidenced by limited remains of sima and spout fragments recovered there, and possibly also by downspouts.55 In that same house, the cistern remained unchanged, but the underfloor channels providing water to the cistern were repositioned a short distance away and essentially parallel to the earlier ones;56 like in the previous period, the channel to the back of the house gathered water from a space belonging to the house on Lot 5’s garden plot. A possible puteal fragment in Vulci tufo was found southwest of the wall over the cistern.57 The configuration of the Lots 4 and 5 house’s cistern’s wellhead, possibly because of the remodelling, is located awkwardly in a doorway within the house.58 The cistern in the House of the Skeleton was probably newly built in this phase,59 as the garden plot under which it was located most likely would not have had a pre-existing cistern or even cesspool. The cistern is fed by an impluvium paved with a brick herringbone pattern with a circular intake pipe covered with a lead filter plate to the southern edge of the cistern below.60 A drawshaft at the northwest corner of the impluvium has a narrow enough diameter (50 cm) to suggest to the original excavators that the skeleton in the cistern did not end up there on accident.

55 Bruno & Scott 1993, 90–92 and Plate 47. 56 Bruno & Scott 1993, 82–83. 57 Bruno & Scott 1993, 85. See also Collins-Clinton 2020, 213–214 (P 4), also 209. 58 Bruno & Scott 1993, 86 and Fig. 22. 59 Bruno & Scott 1993, 119. 60 For more on such filters, see Jansen, this volume.

Breaking Out from Imagined Household Uniformity  87 The cistern’s puteal does not survive. Because of the absence of any compluvium tiles recovered in excavation, the excavators only cautiously characterize this room as a compluviate atrium and hypothesize that the roof might have still been under construction when the colony experienced unrest c. 70 BC.61 An overflow drain beneath the pavement of the House of the Skeleton’s atrium and entryway could carry excess water from the impluvium to the sewer of Street 5.62 The Imperial East and West Block Houses Following the apparent 70 BC destruction at Cosa, the West Block does not seem to have been reoccupied in antiquity.63 In the East Block, however, an Augustan phase atrium-style house, referred to as the House of the Birds after some of the interior plaster decoration, was implanted on top of the Southwest and Northeast Houses (Figure 5.8).64 The House of the Birds utilizes aspects of these houses which it overlies, including being oriented in the same direction (and therefore perpendicular to the earlier, now defunct House of the Skeleton to its north). Exploiting the cisterns of the Southwest and Northeast Houses, this house has two new impluvia, one over the Northeast House’s cistern (Atrium 1) and another over the Southwest House’s cistern (Atrium 2). The herringbone bricked impluvium of Atrium 1 was walled in with a short (max. pres. height 38 cm) red plastered pluteus surrounding its exterior and drawshaft at the northwest corner. The cistern has a bottom depression (50 cm in diam.) centred beneath this drawshaft; an old, filled bottom depression shows that the Republican Northeast House drawshaft was likely located further to the northeast.65 The cistern below was fed by two inlets in each corner of the northwest edge of the impluvium.66 Atrium 2’s impluvium was likewise tiled with herringbone bricks and, as it abuts a main wall of the house, walled on the other three sides by a short red-plastered pluteus (height 30 cm). It also has two intakes into the cistern below, located in the northern corner and close to the western corner. A drawshaft, contrastingly detached from the impluvium, is located more or less in the middle of the cistern vault.67 The cistern that originally belonged to the Northeast House was modified so that an overflow drain could carry excess water to the rubble fill over the house’s old cesspool.68

61 Bruno & Scott 1993, 117–123. 62 Bruno & Scott 1993, 119–120. 63 Bruno & Scott 1993, 162. 64 Bruno & Scott 1993, 176 and Colour Plate I. On dating, see ibid., 185–188. 65 Bruno & Scott 1993, 43–45. 66 Bruno & Scott 1993, 172. 67 Bruno & Scott 1993, 182. 68 Bruno & Scott 1993, 172.

88  Ann Glennie

Figure 5.8 Reconstructed plan of Imperial East Block house/House of the Birds. Bruno & Scott 1993, Fig. 43, adapted.

Rainwater Harvesting and Water Supply of Cosa’s Houses The total number of houses excavated at Cosa provides a small sample from which was extrapolated the situations of the two types of houses at Cosa, including those houses’ rainwater harvesting systems. While such practice is part and parcel of the nature of archaeology (and the inability, if not irresponsibility, of excavating every single inch of a site), more transparency about some of the extrapolation done in original publication would have been welcomed. And though gaps do exist in the record due to either different recording methodologies or a lack of interest in

Breaking Out from Imagined Household Uniformity  89 quotidian water management strategies,69 enough good data was collected to allow not only for careful and acknowledged extrapolation, but also for analysis of domestic rainwater supply at Cosa. From information about the catchment area and cistern volumes of several of Cosa’s houses, coupled with modern rainfall data and the projected usage of water by 5, 10, 15, and 20 occupants, the amount of water that was harnessed and potentially retained in surplus by Atrium Building I, Lots 4 and 5 house, Lot 5 garden plot house, Southwest House, and House of the Skeleton over a period of eight years can be modelled (Figure 5.9).70 Atrium Building I, Lots 4 and 5 house, and the House of the Skeleton all had large enough roof catchments and cisterns to ensure that even with 20 occupants the water supply almost never ran out between the annual cleanings that would have required the emptying of the cisterns. The Southwest House’s catchment and cistern sometimes fully emptied with 20 occupants, but was never empty with just 15. And Lot 5 garden plot house, with its smaller floorplan and much smaller cistern, sometimes had an insufficient water supply with 10 occupants, but could easily accommodate five. These figures can give us an idea of how many people occupied each house, but by no means should be equated with an exact population of a given house. Naturally the houses with larger ground plans (Atrium Building I, Lots 4 and 5, and the House of the Skeleton) might be expected to physically hold more people, but it is also significant that the cisterns were scaled proportionately enough to the 69 See Locicero, this volume. 70 Model based on that developed by Klingborg & Finné 2018. NOAA precipitation data from Grosseto’s weather station was available for the longest contiguous period between the years 1973–1981, so that is what has been utilized here. In keeping with Klingborg & Finné 2018, the model assumes a water usage of 15 l per day per inhabitant and a 20% loss of water marshalled from the roofs, attributed to absorption rates of terra cotta tiles. What was left over could be totalled over time to determine a cistern’s surplus until the cistern was cleaned, which seems to have happened once a year, and the surplus reset. Because August was the first month of the late summer to early fall when a surplus of water typically would be created based on the Grossetto precipitation for 1973–1981—a surplus which additionally at times was a large surplus—my model assumes that the annual cleaning of Cosa’s cisterns occurred in July, resetting the surplus at the beginning of August; by contrast Klingborg and Finné 2018 assume that the annual cleaning took place in October, so that the surplus reset in November. In actuality it could have been that the ancients played this by ear depending on the local conditions and the contents of their cisterns in the late summer to early fall. cf. Thomas & Wilson 1994, 153–154.   The dimensions of Atrium Building I are 22.72 × 17.32 m (Brown et al. 1993, 59). The dimensions for each of the ten plots in a block are provided as ‘56–58 Roman feet by 28–30 Roman feet’ (Bruno & Scott 1993, 13–14), averaged (57 × 29 Roman feet), and converted into meters using the 0.2959 m per Roman foot (=16.8663 × 8.5581 m). Lot 4 and 5 house occupies two plots (33.73 × 17.16 m), Lot 5 garden plot house occupies one plot (16.87 × 8.56 m), and Southwest House occupies about one and one-third plot (21.93 × 11.41 m). House of the Skeleton is not well enough aligned with the blocks to use this metric for calculating its size; dimensions, approximately 21.5 × 16 m, come from hand measuring Fig. 32 (Bruno & Scott 1993, 108).

Figure 5.9 Faceted plot showing the amount of water able to be collected each month (bars) and the surplus, as restricted by max cistern volume (lines), for 5, 10, 15, and 20 person households, respectively, over eight cycles of rainfall collected in the cisterns from: Atrium Building I (max. 60.69 m3); Lot 4 and 5 house (max. 24.04 m3); Lot 5 garden plot house (max. 6.82 m3); Southwest House (max. 36 m3); and House of the Skeleton (max. vol. 58.8 m3).

90  Ann Glennie

Breaking Out from Imagined Household Uniformity  91 roof catchment to provide for a potentially greater number of inhabitants. Nonetheless, such accommodation does not mean that 15–20 occupied each of these larger houses—fewer or more occupants may have lived there depending on any number of personal, social, or temporal circumstances. Lot 5 garden plot house might also have held more than or fewer than the projected 5–10 people it could accommodate. In the case that house occupancy outpaced the house’s cistern’s contents, additional water could be supplied from one of several non-domestic sources across the colony. That being said, if there were fewer people per house than the estimated ranges generated by the model it could indicate that households were using more water per capita and therefore that water was probably devoted to more uses than just drinking and maybe basic hygiene. Perhaps the best physical evidence for the range of this kind of household usage beyond drinking and hygiene comes from the Augustan House of the Birds. Near its heart, the house contains a kitchen outfitted with an opus signinum basin.71 It is easy to imagine that in addition to food preparation, this unit was utilized to boil water for drinking, cleaning, or bathing. And, in fact, directly north of the kitchen there are two connected rooms which probably held a bath suite, as attested by their utilitarian herringbone brick floors and the remains of a terracotta hipbath recovered from them.72 One of these rooms was positioned over a cesspool, which eventually caused the floor to collapse—probably while the house was still in use since the doorway to the bath suite was subsequently walled off— but it is unclear if any overflow drains might have communicated between it and the bathroom for easy water disposal.73 In the northwest corner outside of the House of the Birds proper is a poorly preserved set of spaces which have been interpreted as a work yard, largely in part because of a platform for a hearth or oven found in the southwest corner.74 What kind of work carried out here is not explored, but one could imagine something like a small-scale ceramic- or metalworking installation, either of which (and any number of other private industries too) would require water during day-to-day operations. A nearby doorway into the western side of the House of the Birds allowed entry into a small vestibule and then, to its north, a washroom plastered in opus signinum for washing up after whatever labours were carried out; the room drained via an outlet into an underground channel that fed into the cesspool under the northernmost bathroom.75 Additionally, along the northern wall of the house (and just north of the bathrooms), was another shallow, rectangular opus signinum basin which drained into 71 Evidence for kitchens also exists in Atrium Building I (Brown et al. 1993, 238–240), Lots 4 and 5 house (Bruno & Scott 1993, 21, 28–29, 81), Northeast House (ibid., 69–71), Southwest House (ibid., 69), and House of the Skeleton (ibid., 123–125). 72 Bruno & Scott 1993, 39–42 and Plate 21. Bathrooms are also evident in Atrium Building I (Brown et al. 1993, 238–240), Lot 4 and 5 house (Bruno & Scott 1993, 21, 81, 95), Southwest House (ibid., 38), and House of the Skeleton (ibid., 123–127, 154–155). 73 Bruno & Scott 1993, 39, 181. 74 Bruno & Scott 1993, 167–171. 75 Bruno & Scott 1993, 183.

92  Ann Glennie the cesspool beneath it.76 There is no immediately obvious purpose of this additional water-related feature, but it and the others in the house demonstrate a keen concern of the inhabitants for using and disposing of water within their house. It seems possible to cautiously extrapolate that some of these many domestic environs which utilized water were also present in other Cosan houses—particularly kitchens and bathrooms which are directly attested, but also work spaces and washrooms. Conclusions The edited volume of which this contribution is a part has, as one of its goals, the elucidation of water management strategies that have been overlooked or minimized.77 When it comes down to it, the features of water collection (and indeed of water disposal) are relatively well documented in the publication record of Cosa, despite being maybe the furthest away from the elements of hydraulic engineering which have traditionally drawn scholarly focus. Nevertheless, if we are to bring light to domestic rainwater harvesting, we must strive to fully publish the elements of the system or be transparent about why such details are not available. Having specific measurement for hydraulic features—not just regarding the impluvia, cisterns, and intake channels and pipes, but also their catchment areas—and positioning data, such as coordinates and elevations, will allow the application of more and finer-grained analyses as well as allowing cross-site, cross-cultural, and crosstemporal comparisons which can reveal information about domestic water usage. To that end, Tables 5.1 and 5.2 provide the available metrics about the domestic rainwater harvesting features of Cosa. In this chapter, I have highlighted the lack of uniformity in the domestic rainwater harvesting systems at Cosa. The relatively detailed, but still small, sample of data from Cosa demonstrates the diversity of solutions for gathering (and disposing of water)—both among and between the two house types. The inhabitants of Cosa were both knowledgeable about and flexible with their rainwater harvesting strategies, creating individual systems in each house as topography or the inhabitants dictated rather than mechanically reduplicating the same system in all house plots of the same size. Moreover, this diversity, evidence of the implementation of specific solutions at both the community and individual level over a long settlement history, further demonstrates why we cannot approach Cosa generally or simplistically; it is not a mini-Rome or a cookie-cutter model for mid-Republican colonies or Rome.

76 Bruno & Scott 1993, 183. They also propose that the bathroom may have originally drained into this opus signinum basin and then into the cesspool, but as noted above this situation is not attested archaeologically. 77 See Bonnie & Klingborg, this volume.

Table 5.1  Detailed information available from publication about the Atrium Buildings’ rainwater harvesting features Room

Name

Dimensions (m)

AB I/N Corner Plot

7 10

Channel Impluvium

10 10

Channel Cistern

10

Overflow drain

width: 0.22 exterior: 2.07 × 2.07; interior: 1.924 × 1.924; depth: 0.274–0.324 ? × 0.26 × 0.15 6.96 × 3.26 × 2.65–2.70; vault spring: 1.85 above floor; intrados height: 0.6

10

Hexagonal drawshaft Basins Channel Cistern Cistern

AB II AB V/Diana

AB VII

18 18 24/26 B B B, A garden S garden

Impluvium Cistern Channel Basin Cistern Marble basin(s) Drainage channel Impluvium Cistern

Vol. (m3)

197–180 BC to mid-1st c. AD (?) 197–180 BC to mid-1st c. AD (?) 60.69

1.48 × 2.07 × 0.22 2.59 × ? × 3.50–?

197–180 BC to mid-1st c. AD (?) 197–180 BC to mid-1st c. AD (?) 197–180 BC to mid-1st c. AD (?)

diag.: 0.6; centre aligned to NE corner of impluvium; 0.6 from SE side of cistern

crossbar: 11.32 × 1.33 × ?; vertical: 5.18 × 1.33 × ?; SE prolongation: ? × ? × ?

Usage Dates

197–180 BC to mid-1st c. AD (?)

16.34

197–180 BC to mid-1st c. AD (?) 197–180 BC to mid-1st c. AD (?) 197–180 BC to mid-1st c. AD (?) 197–180 BC to 4th c. AD (?) 197 BC to c. AD 80 197 BC to c. AD 80 197 BC to c. AD 80 after 197 BC to c. AD 80 197 BC to c. AD 80 Claudian Period to c. AD 80 Claudian Period to c. AD 80 197–180 BC to 4th c. AD (?) 197–180 BC to 4th c. AD (?)

Breaking Out from Imagined Household Uniformity  93

Site

Block

Site

Name

Dimensions (m)

West

Lots 4 and 5 (Treasure/Q. Ful.) Lot 5 Garden Plot House

Settling basin Cistern Spill basin Cistern Bottom depression Spill basin Settling basin Cistern Cistern Bottom depression 1 Cistern Impluvium

0.5 × 0.3 × 0.17 5.5 × 1.9 × 2.3 1.4 × 1.4 × 0.08 2.90–3.05 × 1.45 × 1.58 0.50 (diam.) × 0.08 1.9 × 1.6 × ? 0.4 × 0.4 × 0.32 6 × 1.98 × 3.03

East

SW House/ Salvii NE House House of the Skeleton House of the Birds

Impluvium 1 Bottom depression 2 Impluvium 2

c. 1 diam. 7 × 3 × 2.8 ‘perfect square’ >1,