Unknown Crusader Castles

UNKNOWN CRUSADER CASTLES

This page intentionally left blank

Unknown Crusader Castles

Kristian Molin

Hambledon and London New York and London

Hambledon and London 102 Gloucester Avenue London NWI SHX (UK) 838 Broadway New York NY 10003-4812 (USA) First Published 2001 ISBN i 85285 261 5 Copyright © Kristian Molin 2001 The moral right of the author has been asserted All rights reserved. Without limiting the rights under copyrights reserved above, no part of this publication may be reproduced, stored in or introduced into a retrieval system, or transmitted in any form or by any means (electronic, mechanical, photocopying, recording or otherwise), without the prior written permission of both the copyright owner and the above publisher of this book. A description of this book is available from the British Library and the Library of Congress Typeset by Carnegie Publishing, Carnegie House Chatsworth Road Lancaster, LAI 4SL Printed on acid-free paper and bound in Great Britain by Cambridge University Press

Contents Illustrations Acknowledgements Introduction

vii xi xiii

I THE HOLY LAND, 1187-129!

1 Warfare

3

2 Military Architecture

13

3 Fortifications and External Security

38

4 Fortifications and Internal Security

54

5 The Offensive Functions of Fortifications

64

6 The Military Orders

79

II THE KINGDOM OF CYPRUS, 1191-1374

7 Warfare

89

8 Military Architecture

95

9 Fortifications and Internal Security

108

10 Fortifications and External Security

115

11 Fortifications, Naval Warfare and the Crusades

124

12 The Military Orders

130

III

CILICIAN ARMENIA, 1187-1375

13 Warfare

137

14 Military Architecture

145

15 Fortifications and External Security

152

16 Fortifications and Internal Security

168

17 The Military Orders

175

VI

UNKNOWN CRUSADER CASTLES

IV PRANKISH GREECE, 1204-1380

18 Warfare

191

19 Military Architecture

203

20 Fortifications, Naval Warfare and the Crusades

227

21 The Offensive Functions of Inland Fortifications

236

22 The Defensive Functions of Inland Fortifications

243

23 Fortifications and Internal Security

253

24 The Military Orders

262

V THE NON-MILITARY FUNCTIONS OF FORTIFICATIONS

25 Residences

271

26 Prisons

277

27 Taxation, Justice and Administration

280

28 Agriculture and Daily Life

283

29 Trade

290

30 The Church, Pilgrims and Crusaders

295

31 Conclusion

299

Abbreviations

308

Notes

309

Bibliography

389

Index

4os

Illustrations Maps 1 The eastern Mediterranean

xv

2

The Holy Land

5

3

Cyprus

4

Cilician Armenia

139

5

The Aegean

195

92

Figures 1 Crac des Chevaliers (from W. Muller-Wiener, Castles of the Crusaders, trans. J. Brownjohn (London, 1966), p. 61) 2

15

'Atlit (from D. Pringle, Secular Buildings in the Crusader Kingdom of Jerusalem: An Archaeological Gazetteer (Cambridge, 1997), p. 22) 17

3 Ascalon (from Pringle, Gazetteer, p. 21)

21

4

Acre (from Muller-Wiener, Castles of the Crusaders, p. 73)

25

5

Caesarea (from Pringle, Gazetteer, p. 44)

27

6

Goliath (from P. Deschamps, Chateaux des croisés en Terre Sainte, vol. 3, La Defense du comté de Tripoli et de la principauté D'Antioche (Paris, 1973), p. 311) 29

7

Chastel Rouge (from Muller-Wiener, Castles of the Crusaders, p. 52)

30

Kyrenia (from C. Enlart, Gothic Art and the Renaissance in Cyprus, trans, and ed. D. Hunt (London, 1987), p. 423)

97

St Hilarion (from Enlart, Gothic Art, plan VII)

99

8 9 10

Kantara (from Enlart, Gothic Art, p. 470)

11 Servantikar (after Robert W. Edwards, The Fortifications of Armenian Cilicia (Washington, DC, 1987), p. 215)

101 146

VIII

UNKNOWN CRUSADER

CASTLES

12 Yilan (from Muller-Wiener, Castles of the Crusaders, p. 78)

149

13 Amouda (from Edwards, Fortifications, p. 58)

177

14 Baghras (from Muller-Wiener, Castles of the Crusaders, p. 49)

184

15 Mistra (from A. Bon, La Morée franque: recherches historiques, topographiques et archéologiques sur la prindpauté d'Achaïe, 1204-1430 (Paris, 1969), plate 142)

205

16 Patras (from Bon, La Morée franque, p. 671)

209

17 Chlemoutsi (from Muller-Wiener, Castles of the Crusaders, p. 83) 213 18 Karytaina (from Bon, La Morée franque, plate 66)

216

19 Haliartos (from P. Lock, 'The Prankish Towers of Central Greece', ABSA, 81 (1986), pp. 101-23, at p. 114). Reproduced with permission of the British School at Athens. 220 20

Domestic hall at Montfort (from Pringle, Gazetteer, p. 74)

274

21 Margat toll station (from Deschamps, La defense du comté de Tripoli, p. 285)

293

Plates 1 Chastel Blanc: the keep (Jonathan Phillips) 2 Antioch: the city walls and the citadel (Denys Pringle) 3

Caesarea: the city walls built by Louis IX (Denys Pringle}

4

Sidon: the sea castle (Denys Pringle)

5

Gibelet: the citadel (Denys Pringle)

6

Tortosa (Denys Pringle)

7

Chastel Rouge: gateway through perimeter wall into the castle enclosure (Denys Pringle)

8

Chastel Rouge: the central keep and perimeter walls (Denys Pringle)

9

Markopoulo: a typical Latin tower in Greece (Coutauld Institute)

10 The Red Tower: a typical Latin tower in the Holy Land (Denys Pringle) n

Qaqun: the central keep (Denys Pringle)

12 Chastel Blanc: the chapel inside the keep (Denys Pringle)

ILLUSTRATIONS

IX

13 Crac des Chevaliers: the great hall (Denys Pringle) 14 Crac des Chevaliers: distant view showing the rich farmland of the surrounding hills (Jonathan Phillips) 15 Crac des Chevaliers: the central courtyard (Jonathan Phillips) 16 Recordane: the fortified mill (Denys Pringle) 17 Recordane: general view of the mill (Denys Pringle) 18 Margat: the east double walls of the inner castle, with round keep at the far end (Jonathan Phillips) 19 St Hilarion: general view of the castle (Courtauld Institute) 20

St Hilarion: the Byzantine defences of the lower bailey (Denys Pringle)

21 Buffavento: the lower structures of the castle (C. Enlart) 22 Kantara: view of the castle from the south east (C. Enlart) 23 Kantara: aerial view (C. Enlart) 24 Baghras: view of the castle from the east (Courtauld Institute) 25 Silifke: the southern defences (Courtauld Institute) 26

Servantikar: the east front of the castle (R. W. Edwards)

27 Servantikar: the main entrance (R. W. Edwards) 28 Karytaina: view of the castle from the east (Courtauld Institute) 29 Arcadia from the south (Courtauld Institute) 30 Androusa: the keep (Courtauld Institute) 31 Thessaloniki: the city walls (Courtauld Institute) 32 Yuan: general view from the fortress (R. W. Edwards) 33 Corycus: view of the sea castle from the land castle (R. W. Edwards)

This page intentionally left blank

Acknowledgements When I originally began to work on the crusades, I was fortunate enough to have Graham Loud as my supervisor at the University of Leeds. His help and guidance have proved invaluable over the past eight years, and I am very grateful to him. I would also like to express my sincere thanks to Bernard Hamilton for the kindness he has shown me. Many other people have, directly or indirectly, made the publication of this book possible. From my time in Canterbury, I would like to thank Michael Simmons and Richard Bales for all their generosity and support. I am deeply indebted to Jonathan Phillips for his considerable help, not least his suggestion that I submit this book to Hambledon and London, where Martin Sheppard and Tony Morris have been very supportive and above all patient. During the past year I have been lucky to work alongside Peter Edbury, Helen Nicholson and Denys Pringle at Cardiff University. They have provided me with much advice on the completion of this book. I would like to express my gratitude to them. For his help with the computing I am, as always, indebted to Jase. The bulk of the plans and illustrations were very generously provided by Jonathan Phillips, Juliette Constantinou, Peter Lock and, above all, Denys Pringle. I am also grateful to Dumbarton Oaks and Deutscher Kunstverlag for allowing me to use illustrations from R. W. Edwards' The Fortifications of Armenian Cilicia and W. Muller-Wiener's Castles of the Crusaders. Finally, I owe the greatest debt to my parents, Carol and Ella. Without their love, encouragement and patience over many years, this book would never have been completed, and it is dedicated to all of them.

This page intentionally left blank

Introduction In November 1095 Pope Urban II launched the First Crusade at Clermont in south central France. The popularity of his message turned out to be enormous, as tens of thousands of people joined the expedition to liberate Jerusalem from the Muslims in order to cleanse their own souls and, hopefully, achieve salvation after death. Having suffered enormous hardships as it travelled from western Europe through the Balkans, across Asia Minor and into Syria, the First Crusade finally captured Jerusalem in July 1099. During this gruelling journey its participants passed through several different regions whose history was to become inexorably linked with that of the future crusader states. Stretching from the Adriatic to Constantinople and dominating most eastern Mediterranean islands, including Cyprus, the Byzantine empire viewed the arrival of this and future Crusades with mixed feelings. On the one hand, the empire's inhabitants were fellow Christians and therefore grateful for any aid they might receive against the Turkish threat to the east. On the other hand, they were of course Orthodox Greeks rather than Catholics, and were consequently separated from western Europe by profound differences in religion, culture and outlook, which often caused tension or even open warfare. After they crossed the Bosphorus the armies of the First Crusade entered an even more alien world as they gradually began to move into Muslim territories. During the period covered by this book, these lands can loosely be divided into the three key regions of Asia Minor, Syria and Egypt. Asia Minor had originally been Byzantine, but by the 10905 most of it had been lost for good because of a series of Turkish advances achieved in the course of the eleventh century. During the next two hundred years the most dominant people in the region were normally the Seljuk Turks. However, after the mid fourteenth century the various Turkish peoples of Asia Minor were eventually unified under the rule of the Ottomans, who subsequently expanded their power in such a spectacular fashion that by 1520 they had created an empire stretching from North Africa to the Balkans via Egypt, Syria, Turkey and northern Greece. But at the time of the First Crusade and for a long time thereafter no such unity existed. The Muslims who inhabited the other key regions of Syria and Egypt were sporadically at war

XIV

U N K N O W N CRUSADER CASTLES

both with the Turks of Asia Minor and with each other. Unlike western Europe, where urban growth was still extremely limited, these political struggles were largely centred around the most important cities of the Muslim world. Inland these included Aleppo, Damascus and Cairo, all of which had grown extremely rich because of their trade links with Asia or Africa. Through ports such as Alexandria, Acre and Tyre they also had economic connections with Italy and the Byzantine empire, which ensured that the eastern Mediterranean was far richer and far more sophisticated than any of the north European lands inhabited by most members of the First Crusade. The region also differed from the west in its cultural diversity. The Muslim peoples were not just divided geographically or politically, they varied considerably in terms of their ethnic background and according to whether they adhered to the Sunnite or to the Shi'ite faith. The eastern Mediterranean also had a notable population of Jews, plus several nonCatholic Christian groups, many of whom lived under Muslim rule. These included various Orthodox communities, as well as Maronites, Nestorians and Syrian Jacobites. The most important such group for the purposes of this book was the Armenians. They inhabited a large part of what is now south-eastern Turkey and northern Syria, and in particular the region to the north west of Antioch known as Cilicia. They had consequently lived under Byzantine rule until the Turkish incursions of the eleventh century. Although the Greek emperors continued to enjoy some sporadic authority over them during the next hundred years, their distance from Constantinople ensured that by the late twelfth century the Armenians had effectively become independent. This move toward independence was greatly aided by the local terrain, for Cilicia itself was surrounded by mountains to the north, east and west, and by the Mediterranean to the south. These barriers helped to protect the Armenians both from the Greeks and from the Seljuk Turks. Like the Byzantine Greeks, they enjoyed a love-hate relationship with the crusading newcomers, for they were fellow Christians yet they were non-Catholics with their own language, history and cultural identity. This was the world which the members of the First Crusade entered in 1099. Rather than feeling overawed by the rich and complex society they had encountered, those crusaders who chose to remain after the fall of Jerusalem wasted no time in consolidating the military gains they had made. Former Muslim territories in western Syria and Palestine were gradually conquered and turned into four new crusader states: the kingdom of Jerusalem, the principality of Antioch, and the counties of Tripoli and Edessa. The westerners who undertook these campaigns were mostly French, but others came from Italy, Germany and Spain. Former Muslim ports such as Acre

MAP i. The eastern Mediterranean, showing approximate frontiers in the early thirteenth century

XVI

UNKNOWN CRUSADER CASTLES

now fell under Christian control and their lucrative Asiatic trade links came to be dominated by the Italian maritime cities of Venice, Genoa and Pisa. In the course of the twelfth and thirteenth centuries many crusader lands were also held by the Hospitallers, Templars and Teutonic Knights. These were Military Orders, groups of western knights who confronted the Muslims in battle yet followed an austere and celibate lifestyle under the protection of the pope. Like more traditional monastic orders, they grew wealthy on the donations of land or money made by their European patrons, and these resources enabled them to make a substantial contribution to the military activities of the crusader states. After these states had been established, their twelfth-century history was largely determined by the attitude taken towards them by the Muslim rulers of Aleppo, Damascus and Cairo. These three key cities were frequently at war, and at such times their rulers were prepared to negotiate peace treaties with the Latins, or even seek Christian aid against each other. Although its field army was relatively small, this enabled the kingdom of Jerusalem to prosper for most of the early twelfth century. However, in the north the rulers of Aleppo gradually proved to be more pugnacious toward Antioch and Edessa, and eventually in 1144 the latter city was recaptured by the Muslims. By the early 11505 the entire county of Edessa had disappeared, the failed Second Crusade (1148-49) doing nothing to prevent this. During this period Nur al-Din, the ruler of Aleppo, also managed to extend his rule to Damascus, threatening the security of Jerusalem itself. Worse was to come after Nur al-Din's death in 1174, for eventually his Syrian lands were annexed by Saladin, the ruler of Egypt. This meant that the Christians could no longer rely on the divisions amongst the Muslims to keep them safe, for Cairo, Damascus and Aleppo were now all controlled by one man. Saladin's empire eventually gave him such an overwhelming military superiority over the Latins that he was able to annihilate the crusader field army at the battle of Hattin in July 1187. Thereafter the entire kingdom of Jerusalem, including the holy city itself, was recaptured by the Muslims except for the heavily fortified port of Tyre. Although the county of Tripoli and the principality of Antioch did not suffer quite as badly, they too lost considerable amounts of land, including the corridor of coastal land which actually connected them, when Saladin launched another devastating campaign against them in 1188. Despite these disasters, the crusader states still managed to survive into the thirteenth century. The participants of the Third Crusade (1189-92), and in particular Richard Lionheart, managed to reconquer the coastal regions of the kingdom of Jerusalem between Tyre and Jaffa, including the vital port of Acre. It was also at this time that Richard conquered the island

INTRODUCTION

XVII

of Cyprus from the Greeks, thereby creating an important new crusader state which came under Latin rule. These successes were followed by the death of Saladin in 1193, which caused his empire to be divided between various members of his family, known collectively as the Ayyubids. As a result the situation reverted to that of the early twelfth century, for the cities of Aleppo, Damascus and Cairo were once again ruled by different individuals whose squabbles took the pressure off the Latins. It was not until the second half of the thirteenth century that the Muslim world was reunified once again. The remaining coastal territories still held by the Latins were finally wiped out in 1291. These losses could not be prevented by the numerous crusades sent to the region after 1200, most notably the Fifth Crusade (1217-21) and the Crusade of Louis IX of France (1248-54), both of which made failed attempts to invade the sultanate of Egypt. Meanwhile, new crusader states had been set up in Greece after the Fourth Crusade (1202-4), a joint expedition involving Venetian naval forces and contingents of French and Italian knights. Its original purpose had been to build on the achievements of the Third Crusade and attack the Muslims of Egypt, but instead it ended up capturing Constantinople from the Greeks. This diversion was caused by numerous political and economic factors, such as the request for western aid by a Greek claimant to the imperial throne and the Venetian desire to augment its share of the lucrative trade passing through the Aegean. It also reflected the growing tension between the Greeks and the Latins, as there was often mutual distrust between these two peoples because of their differences in language, culture and religion. Both this expedition and Richard I's earlier conquest of Cyprus reflected a growing willingness in the west to direct crusades against people who were not necessarily Muslim but were nevertheless considered schismatic opponents of Catholicism. After the fall of Constantinople in 1204 some parts of the former Byzantine empire were carved up into a number of new western states, whilst others remained under Greek control. The period until 1380 witnessed a protracted struggle between Latin newcomers and Greeks trying to recreate the Byzantine empire. On the mainland this struggle generally favoured the Greeks, but by the late fourteenth century both sides found themselves being confronted by an even stronger opponent, the Ottoman Turks, who gradually swallowed up all Christian territories around the Aegean regardless of whether they were Greek or Latin. Meanwhile, the Muslim armies which had already driven the Franks out of the Holy Land in 1291 also conquered Cilician Armenia in 1375. This was also a period of crisis for the kingdom of Cyprus, for in 1374 it was invaded by Genoese forces hoping to use the island as a base from which to dominate local trade and to undermine the influence

XVIII

UNKNOWN CRUSADER CASTLES

of their great rivals, the Venetians. This highly destructive episode marked the beginning of the end for Cyprus as a fully independent Prankish crusader state, for it resulted in the Genoese occupation of Famagusta until 1464, after which the entire island fell under Venetian control from 1489 onwards. Finally, in 1571, Cyprus was conquered by the Ottoman Turks, thereby destroying the last Christian domain in the eastern Mediterranean whose Prankish settlers could trace their ancestry back to the earliest crusader states created by the members of the First Crusade. This book is a study of the various military and non-military functions of fortifications occupied by Latin or Armenian Christians who settled in the Holy Land, Cyprus, Cilician Armenia and Greece between 1187 and c. 1380. The political and military upheavals of the period ensured that fortifications played a prominent role in its history. Yet the study of such structures still has a number of important gaps which this book aims to fill. There has been a tendency in the past for scholars to concentrate their efforts on castles in the Holy Land, and in particular famous and wellpreserved structures such as Crac des Chevaliers in Syria. Although this imbalance is being addressed by the work of scholars such as Denys Pringle, Ronnie Ellenblum and Adrian Boas, who in recent years have excavated, studied or recorded very many smaller crusader fortifications and domestic structures, their studies are still focused on the Holy Land itself, and especially on the kingdom of Jerusalem. A primary aim of this book is to shed more light on fortifications built or occupied by Latins and Armenians in the more obscure areas of Cyprus, Greece and Cilician Armenia. The second, and closely related aim is to make up for the lack of studies which deal with crusader fortifications from the thirteenth century onwards, even though this period, as we have seen, represented a massive expansion of crusading endeavour as Latins settled on Cyprus and in the former Byzantine territories around the Aegean. Within the Holy Land itself there has often been a tendency for historians to focus on the period between 1095 and the end of the Third Crusade in 1192. This is perhaps understandable, bearing in mind that some of the most famous military encounters of the middle ages, most notably the First Crusade and Richard Lionheart's struggles with Saladin, took place during this period. It may also reflect the nature of the sources, and in particular the fact that William of Tyre's well-known history of the Holy Land, which contains so many clear and accessible accounts of castles being built, besieged or destroyed, ended in 1184. Thus for example, R. C. Smail's famous work on crusader warfare dealt with the period from 1097 to 1193.l In recent years scholars such as C. Marshall, whose book on warfare between 1192 and 1291 represents a continuation of Smail's work, have begun to address this shortfall.2 However,

INTRODUCTION

XIX

this book hopes to extend the work of Marshall and others by taking a further look at the Holy Land in the thirteenth century, and by including the post-1291 history of Cyprus, Greece and Armenia. Hence each chapter begins with a brief section on warfare in these regions, to show how the nature of local fighting influenced the appearance and functions of local fortifications. The third aim of this book is to provide an analytical rather than a purely descriptive study of fortifications in the eastern Mediterranean. Once again, this is something which is already being done in the Holy Land. Alongside purely archaeological or architectural works such as D. Pringle's recent survey of secular structures in the kingdom of Jerusalem,3 some scholars have also researched the ways in which castles were actually used by the Latins, and how they fitted into the wider military and political history of the Latin East. The earliest and most influential book of this kind was undoubtedly Smail's Crusading Warfare, but, as we have seen, this has now been complemented by Marshall's important study of the period after 1192. Other books to appear in recent years which have viewed castles in the same light include Hugh Kennedy's Crusader Castles (Cambridge, 1994), plus numerous publications by leading archaeologists such as Pringle's The Red Tower (London, 1986) and Ellenblum's Prankish Rural Settlement in the Latin Kingdom of Jerusalem (Cambridge, 1998). Beyond the Holy Land itself, the picture is very different. Studies already exist devoted to the visible remains of fortifications in Cyprus, Greece and Cilician Armenia,4 but apart from giving brief historical outlines of the sites they refer to, these works are largely devoted to listing and describing surviving structures. What these fortifications were actually used for, rather than what they looked like, is a topic which has barely been touched upon, and it is one of the principal aims of this book to make up for this silence. At times it will become apparent that this can be a difficult task, because some areas, most notably Prankish Greece and Cilician Armenia, are lacking both in reliable historical sources and in recent subsurface archaeological excavation. No studies exist for these areas which can compare with the very detailed research undertaken in the kingdom of Jerusalem, where a wealth of information from papal letters, contemporary chroniclers or surviving charters, combined with the meticulous investigations of leading archaeologists, has sometimes made it possible to recreate the history of certain crusader castles virtually year by year. Despite the limited scope for this type of research outside the Holy Land, a combination of archaeological and historical sources can shed much new light on the functions of local fortifications. For example, during the last thirty years the archaeologists Hansgerd Hellenkemper and Robert Edwards

XX

U N K N O W N CRUSADER CASTLES

have produced comprehensive studies which list and describe virtually every single surviving stronghold in Cilician Armenia. Yet because there are very few traces of any urban fortifications left in this region, they have rejected the notion that the Armenians constructed defences of this kind, preferring instead to live in or near remote mountain castles. I hope to show that a closer scrutiny of the historical evidence actually indicates that the Armenians did dwell inside cities, whose older Byzantine ramparts they were happy to repair or rebuild. This would suggest that Edwards and Hellenkemper have relied too much on the archaeological evidence (or its absence) to draw their conclusions, without studying the written sources as well.5 While there are chapters in each of the parts of this book which deal with military architecture, it is my intention to do more than simply describe fortifications, and to look at a variety of sources in order to highlight the ways in which such structures were actually used to defend or conquer new land, to suppress hostile locals or to impose central authority over recalcitrant vassals. This is an approach which has not been used before with regard to Greece, Cyprus and Cilician Armenia, but in the Holy Land I hope to build on the research already carried out in this field by Smail, Marshall and others. An exploration of strongpoints in the eastern Mediterranean can also give us greater insight into how local Christian settlers actually lived. In the past there has sometimes been a tendency amongst scholars to focus on the purely military functions of fortifications, and the ways in which they could withstand sieges, dominate strategic valleys or form elaborate networks of intervisible castles. Whilst these topics will be discussed, it is important to bear in mind that many years or decades might pass before a fortress came under siege or found itself involved in any kind of warfare. Consequently this book is intended to redress the balance by considering the many non-military functions of strongpoints as residences, prisons, courthouses or centres of trade and agriculture. This topic is dealt with at length in part five, but throughout the book it will become apparent that it is misleading to view castles and warfare as separate from other aspects of medieval society, such as trade and farming. In part four, for example, the numerous rural towers which were constructed by Latin settlers in Greece will be considered. It is tempting to see these towers as evidence of chronic local warfare. Whilst it certainly seems to be true that such structures could provide shelter against pirates or Turkish raiders, it should be borne in mind that their construction would have been expensive and time-consuming, and therefore only possible at a time when local lords were relatively rich and their estates were untouched by external attackers. In the Holy Land this approach has already led to many important new discoveries by Ellenblum, who has shown that Prankish settlers built countless smaller fortified or

INTRODUCTION

XXI

semi-fortified structures which were simply intended for local farming and defence. Strongpoints of this kind had no importance whatsoever in terms of the large-scale warfare between Richard Lionheart and Saladin which has become so famous in the popular view of the Crusades, yet they accounted for the vast majority of fortifications built by Christian settlers after the First Crusade. The vast majority of research already carried out on crusader fortifications, as is already clear, has concentrated on the Holy Land, and in particular on the kingdom of Jerusalem (or Acre, as it is often known after 1187). This book nevertheless contains a chapter on that region, partly because it is the most famous and important area of crusading endeavour, and partly because it gives this research as a wide a scope as possible, so that comparisons can be made between the functions of fortifications in the Holy Land and those situated in other parts of the eastern Mediterranean. It is hoped that this will show how the famous castles built by the crusaders in the Holy Land were sometimes very different from, but sometimes very similar to, the many less well-known Strongpoints constructed in Greece, Cyprus and Cilician Armenia. This book deals with the period from 1187 to 1380. From what has already been said it will be clear that much work has already been done on the Holy Land, particularly during the twelfth century, yet after 1187 the Latins acquired far more crusader territories around the eastern Mediterranean than they had held before the battle of Hattin. It was also a time when Cilician Armenia gained independence from Byzantine authority, being transformed into another Christian state which had to face the growing pressure applied by powerful Muslim neighbours. Many more Christian Strongpoints were built or occupied during this period than the hundred years immediately after the First Crusade. The year 1380 seems a natural ending point for this book for a number of political, military and cultural reasons: Cilician Armenia fell in 1375, bringing its history as an independent state to a close. Cyprus was invaded by Genoa in 1374, and although the Genoese were only able to conquer a small part of the kingdom, this period signalled the end of the island's status as a prosperous crusader state, for its economy was in decline and its political life gradually came to be dominated by Italians rather than by the Prankish families who had originally settled there after 1192. In Greece, meanwhile, the tide had definitely turned in favour of the Ottoman Turks, for by the end of the fourteenth century they had already conquered most of the Byzantine empire and were beginning to penetrate the remaining Latin states in south-western Greece. Technologically speaking, 1380 also seems an appropriate date at which to stop, because by that point gunpowder was already known in

XXII

UNKNOWN CRUSADER CASTLES

Europe. Although it would be a long time before it revolutionised warfare in its entirety, this new discovery heralded the beginning of the end for the kind of medieval fortifications dealt with in this book. Hence both politically and technologically the late fourteenth century marked the end of an old era and the beginning of a new one.

PART ONE

The Holy Land 1187-1291

This page intentionally left blank

1

Warfare The history of the crusader states in the Holy Land between 1187 and 1291 can be divided into two roughly equal parts. First, the period from 1189 to the early 12405 witnessed a series of efforts by the Franks to restore the losses suffered after the battle of Hattin (1187), which had resulted in the erosion of Christian territories around Antioch and Tripoli and the loss of all land in the south apart from Tyre. The Third Crusade (1189-92), whose most prominent participant was undoubtedly Richard Lionheart, achieved the reconquest of Acre and a coastal strip as far south as Jaffa.1 Five years later German crusaders took Beirut,2 and the Embriaco lords of Gibelet reoccupied their old castle through diplomacy.3 In 1204 another peace treaty4 confirmed Prankish possession of Jaffa (lost temporarily in 1197) ,5 Lydda, Ramlah and Nazareth, and at about the same time the Templars and Hospitallers consolidated their position in the county of Tripoli by strengthening their local castles and launching punitive raids against neighbouring Muslims.6 Further gains were made during the crusade of Frederick II (1228-29), whose treaty with the Egyptians acknowledged Christian control over Jerusalem, Bethlehem, Nazareth, the territory of Toron and those parts of the lordship of Sidon previously held by the Muslims.7 Finally, during the separate but overlapping crusades of Theobald of Champagne (1239-40) and Richard of Cornwall (1240-41), complex negotiations with Egypt and Damascus led to the restoration of Belvoir, Toron (Tibnin), Tiberias, Chateauneuf (Hunin), Beaufort, Saphet and Cave de Tyron.8 The Franks also reoccupied Ascalon, whose citadel was rebuilt, and Jerusalem, which had been lost briefly in 1239-40.' After the successes of the early thirteenth century, the next fifty years represented a period of sporadic retreat. In 1244 the truce with Cairo broke down and Jerusalem was lost to the Khwarizmians, a violent tribe of nomadic horsemen who were allied to the Egyptians and supposedly numbered between 12,000 and 20,000 troops.10 Shortly afterwards this alliance defeated the Franks at the battle of La Forbie, enabling the Egyptians to retake much of southern Palestine, including Ascalon and Tiberias in 1247.11 Further Muslim conquests were temporarily halted by the first Crusade of Louis IX (1248-54), who refortified several sites in the Latin East, although his

4

U N K N O W N CRUSADER CASTLES

disastrous campaign in Egypt prevented him from actually regaining lost strongholds.12 Moreover, within ten years of Louis' departure the Egyptian sultan Baybars (1260-77) had resumed the offensive, so that by 1271 Christian territories had been reduced to a narrow coastal strip between Latakia in the north and 'Atlit (Pilgrims' Castle) in the south.13 Many of the remaining outposts in this region were gradually picked off during the 12805, until Acre itself was finally lost in 1291.14 From the Muslim point of view, these triumphs represented the culmination of a long period of change. Politically the Muslims had been hopelessly divided during most of the late twelfth and early thirteenth centuries, for the Ayyubid sultans of Egypt were almost constantly embroiled in power struggles with the Ayyubid rulers of Damascus and Aleppo, and were at times also challenged by smaller powers such as the lords of Kerak. Inevitably, disputes of this kind were most common whenever a leading political figure died, hence the Frankish territorial gains of 1197-1204,1229 and 1239-41 can be linked to the internal clashes which followed the deaths of the Egyptian sultans Saladin (1193), al-Adil (1218) and al-Kamil (1238).15 These problems began to resolve themselves in the mid thirteenth century, for in 1250 the Ayyubids of Egypt were overthrown by the Mamluks. The Mamluks were soldier slaves who had been acquired in childhood by individual Ayyubid rulers and then brought up to form an extremely loyal and highly skilled military bodyguard. This made them very effective soldiers, but tensions inevitably arose whenever an Ayyubid ruler died, for his Mamluks could then find themselves being ousted from key military posts by the Mamluks of his successor. Problems of this kind resulted in the Mamluk seizure of power in 1250, and subsequently led to a decade of political instability in Egypt as rival Mamluks, lacking a strong hereditary precedent, struggled for power. Whilst this period benefited the Christians in the short term, in 1260 it led to the accession of the Mamluk sultan Baybars, who was to prove one of their greatest opponents. Furthermore, the year 1260 witnessed the arrival of the Mongols, who, having spent the previous decade advancing westwards, captured both Damascus and Aleppo. As a result, Saladin's Ayyubid descendants in Syria were virtually all driven out of power, so that when a Mamluk army from Egypt managed to defeat the Mongols in the autumn of 1260, Baybars was able to extend his power over Aleppo and Damascus with relative ease. From 1260 onwards Egypt and Syria were therefore once again united under one ruler, and the crusader states found themselves surrounded as they had been in the time of Saladin.16 These dramatic political changes inevitably influenced the military effectiveness of the Muslims. Before 1250 the Ayyubids could theoretically raise

WARFARE

MAP 2.

The Holy Land

5

6

UNKNOWN CRUSADER CASTLES

a combined field army of around 22,000 men, about half of whom came from Egypt.17 In practice, of course, the Egyptians and the Syrians spent more time using these troops against each other than the Franks, who prospered 'because of the constant discord of the [Muslim] princes of the land, which was highly favourable to the Christians'.18 Alternatively, one Muslim power might even ally itself with the Christians against another, as was the case at the battle of La Forbie in 1244, when the ruler of Damascus was prepared to join forces with the kingdom of Jerusalem against the Egyptians and Khwarizmians.19 Similar problems had occurred even during the reign of Saladin, who was only able to keep his cumbersome and regionalised empire together because of his own forceful personality. After the battle of Hattin Saladin often found it extremely difficult to maintain his field army, which was largely composed of seasonal troops who served the sultan during the summer in exchange for land or money, but then went home for several months as soon as winter set in. Even when they were present, some of these forces turned out to be ill-disciplined, badly trained and more interested in loot than complex notions of holy war. Drawn from all over Muslim Syria, they often expressed greater loyalty toward their local rulers, such as the Zengid rivals of Saladin in the north, than they did toward the empire-building projects of Saladin himself. Regional loyalties and personal aspirations also affected the various members of Saladin's own family and entourage, who openly argued over strategies in the middle of campaigns. Between 1187 and 1192 these serious weaknesses help to explain why Saladin failed to capture all Christian lands in the wake of Hattin, proved incapable of defending Acre against the armies of the Third Crusade, and ultimately had to return large parts of the coast to Richard Lionheart. Without a well-controlled army which stayed with him throughout the year, Saladin could not complete the task of capturing powerful Latin strongholds like Tyre or wiping out crusader forces in battle, being eventually forced to reach a negotiated settlement with the leaders of the Third Crusaded Another reason why Saladin failed to defeat the Third Crusade was his inability to match the overwhelming naval power of the Latins, despite his considerable efforts at building up a sizeable Egyptian fleet.21 In the thirteenth century there were sporadic attempts to remedy this weakness, but Prankish possession of Cyprus and the Levantine coast made it difficult for Egyptian war galleys to reach, let alone blockade, the sea lanes which connected the crusader states with western Europe. A notable example of the troubles faced by the Muslims was Baybars' disastrous naval raid on Limassol in 1271, involving eleven to fourteen Egyptian galleys. Before they even reached Limassol these vessels were shipwrecked off the coast because

WARFARE

7

of the incompetence of their sailors, some of whom were taken prisoner. This was the last major naval offensive attempted by the Mamluks before 1291. Apart from illustrating just how far the Muslims had fallen behind the Latins in this particular form of warfare, it helps to explain why coastal strongpoints such as Acre held out far longer than crusader fortifications inland.22 On land, however, it was a very different story, for by the mid thirteenth century many of the weaknesses which had plagued Saladin's field army were beginning to disappear. The empire which Baybars created after 1260 was similar in size to that of Saladin, but far more centralised, as the regional Ayyubid rulers in Syria had virtually all been removed by the Mongols. Hence Baybars was able to push ahead with reforms which improved military training, equipment and administration. He also built on policies begun during the 12405 which ensured that the halqa, or cavalry elements within his army, continued to grow.23 At this time the quality and quantity of archers also increased steadily, so that by 1289, for example, Sultan Kalavun (1279-90) was reputedly able to deploy thirty such troops against each arrow slit in the city walls of Tripoli.24 Furthermore, greater centralisation enabled Baybars to use the resources of his empire more effectively than Saladin had done, causing the overall size of field armies to rise to between 12,000 and 40,000 troops. Whilst this latter figure may not be reliable, the former probably is, implying that major Mamluk campaigns launched against the Franks after 1260 rarely involved fewer than 10,000 men.25 Most late thirteenth-century offensives carried out by the Mamluks were also better organised than those of Saladin. Although internal wrangles still affected the Muslims from time to time, the majority of crusader strongpoints fell during the key reigns of Baybars (1260-77) and Kalavun (1279-90), whose leadership on campaign never seems to have been seriously questioned by their own followers. This resulted in some very carefully prepared and highly successful attacks on crusader castles by these two rulers. During his campaign against the Templar stronghold of Saphet (1266), for example, Baybars patiently supervised the siege weapons, set up field hospitals for the wounded and offered financial rewards for the bravest attackers, whilst simultaneously sending out contingents to carry out diversionary raids which prevented the Franks from relieving the castle. Even though it took several weeks of hard fighting, the siege was eventually brought to a successful conclusion without any overt criticism of Baybars' leadership.26 Two other factors also improved the fighting capacity of Muslim armies at this time. First, greater use was made of Mamluks, who were originally soldier slaves imported by individual sultans from Central Asia and the Black Sea region. These troops were trained as professional soldiers from

8

U N K N O W N C R U S A D E R CASTLES

an early age and were instilled with a deeper sense of loyally toward their own sultan than most seasonal or mercenary forces. Although Mamluks had already been used for many years, their numbers increased significantly in the course of the thirteenth century. It is difficult to say how many Mamluks were accumulated by each sultan. Kalavun is reputed to have owned between 6000 and 12,000, whilst the average number for most of his contemporaries seems to have been between 2000 and 4ooo.27 These troops formed the core of late thirteenth-century field armies. Although Mamluk sultans were still reluctant to undertake sieges in midwinter, they did make all-year-round campaigning, such as Baybars' raids on the fortress of Margat in January 1270, much easier.28 Secondly, the Muslims became increasingly adept at using mines and catapults to capture Christian strongpoints, including Acre itself, whose outer walls were either sapped or bombarded to the point of collapse during the final siege of i29i.29 Earlier, the Mamluks had captured a number of crusader castles by digging mines beneath their walls, including the Hospitaller fortress of Margat (1285), forced to surrender when the enormous inner tower guarding the southern tip of the fortress was in danger of collapsing.30 Whereas the basic art of digging siege tunnels had changed little over the centuries, more profound developments affected catapults during the crusader period. In the course of the twelfth century more and more large catapults, or trebuchets, came to be fired by heavy counterweights filled with sand or stones rather than by men pulling on ropes, and as a result stronger and more accurate weapons could be built. A recent reconstruction of a trebuchet proved capable of hurling objects (including a piano) weighing up to 500 kilogrammes over considerable distances.3! Ammunition weighing almost this much was used in some of the most decisive sieges of the thirteenth century. At Acre in 1291, for example, catapults were used which could easily launch stones weighing nearly 200 kilogrammes.32 Ammunition of this kind was deadliest if it could be aimed against exactly the same part of a castle's ramparts again and again, thereby literally shaking it to destruction. In order to achieve such accuracy, the weight of the stones fired by a trebuchet needed to be calculated in advance so that it did not vary.33 Furthermore, if it was to be capable of throwing such large stones over a long period without falling apart from the strain, the trebuchet itself had to extremely tall and sturdy. Indeed, the modern reconstruction just referred to stood to a height of sixty feet.34 Hence trebuchets needed to be handled by skilled siege engineers, and required large amounts of money and logistical planning for their upkeep. After Baybars had united both Syria and Egypt under his personal rule in 1260 these resources became available, enabling him to construct large numbers

WARFARE

9

of trebuchets which could be used against the Christians. In 1268, for example, Baybars is said to have deployed twenty-six catapults against the castle of Beaufort. These weapons also played a prominent role in Baybars' attacks on Saphet (1266) and Crac des Chevaliers (1271), and were regularly brought along on campaigns in large prefabricated sections which were normally stored in key centres such as Damascus.35 The military and political improvements which took place in the Muslim world between 1187 and 1291 form a direct contrast with contemporary events inside the crusader states. Whereas the Mamluk conquests of the late thirteenth century were the result of tangible military reforms, Prankish territorial gains made before 1241 were an illusion, achieved almost entirely through negotiation rather than success on the battlefield. In reality the Franks were often weak, and much of this weakness stemmed from the political situation within the kingdom of Jerusalem. In 1225 Isabella, the heiress to the kingdom, married the German emperor Frederick II, and as a result the throne passed to their son Conrad (1228-54) and grandson Conradin (1254-68). But normally Frederick and his Hohenstaufen heirs were either absent in Germany or too young to rule in their own right, or both. As a result, a succession of regents and lieutenants had to be appointed to govern the kingdom in their place, but these representatives often proved incapable of controlling (or were themselves controlled by) local factions within the crusader states. The death of Conradin and the extinction of the Hohenstaufen line in 1268 should have clarified matters, but instead the throne came to be disputed between the rival claimants Hugh III of Cyprus (1268-84) and Charles of Anjou, brother of Louis IX of France. It was not until Charles died in 1285 that Hugh Ill's son Henry II (1285-1324) could succeed to the kingdom in peace, and Acre finally came under the control of an undisputed monarch who actually lived in the Latin East.36 During the long period of weak or absent rulers between the 12205 and the 12805 there were numerous internal clashes in the Latin East both in the kingdom of Jerusalem and in the north.37 After Frederick II's marriage to Isabella in 1225, the emperor came to the Holy Land between 1228 and 1229. This campaign was not simply intended as a crusade against the Muslims, it also gave Frederick the opportunity to impose his authority (and that of his newly born son Conrad) over the kingdom of Jerusalem. But certain factions in the East, and in particular the powerful Ibelin family, were deeply hostile toward the idea of Hohenstaufen rule, whilst many churchmen, most notably the Templars, were opposed to Frederick because his arguments with the papacy had already led Pope Gregory IX to excommunicate him. Shortly after Frederick's arrival in the East fighting broke out between his supporters and Ibelin sympathisers both on the mainland

10

UNKNOWN CRUSADER

CASTLES

and on Cyprus. Frederick returned to Europe in 1229 but continued to send imperial troops from Lombardy under the command of his marshal, Richard Filangieri, to try to defeat the Ibelins. During the winter of 1231-32 Richard Filangieri made a failed attempt to capture the Ibelin stronghold of Beirut, and by the spring of 1233 the imperialists had also been defeated on Cyprus. But after these setbacks Frederick's supporters continued to hold Tyre, whilst the Ibelin faction dominated Acre. It was not until 1242 that the Ibelins finally managed to capture Tyre by stealth and bring a formal end to hostilities. For more than a decade the kingdom of Jerusalem was therefore not only ruled by an absent German dynasty but was also deeply divided within itself.38 While the damage caused by this type of warfare may have been exaggerated in the past, it must still have caused a lot of troops and resources to be squandered which could otherwise have been used in the struggle with the Muslims. The Latins could ill afford to waste their time on internal struggles, for the sources make it clear that they were often heavily outnumbered by their Muslim opponents; a situation which can only have deteriorated as Egyptian armies grew in size after the mid thirteenth century. Already at the battle of La Forbie (1244), where we have seen that the Khwarizmians alone contributed up to 20,000 cavalry troops, it seems that there were no more than 2000 Latin knights present; a figure which probably represented the absolute maximum for the Holy Land.39 Similarly, in 1291 the Muslim besiegers of Acre are said to have outnumbered the city's entire population of 40,000 people, and as a result the 200 knights and 500 infantry sent there by King Henry II of Cyprus and Jerusalem (1285-1324) must have seemed like a drop in the ocean.4o Attempts by the Franks to remedy this problem by appealing for help from the West were fraught with difficulties, for the arrival and departure of crusading expeditions, which were anyway only a temporary solution to the lack of soldiers, could not be predicted accurately. This was vividly illustrated by King Andrew II of Hungary's decision to leave the Fifth Crusade (1217-21) before it had even reached Egypt, where numerous subsequent disputes between fellow Christians, combined with Frederick II's failure to appear despite repeated promises to do so, eventually contributed to the collapse of the entire expedition.41 Even when crusaders did turn up, harsh weather, disease and food poisoning could take a heavy toll; problems which affected numerous expeditions, including that of Edward, the future king of England (1271-72) .42 This period also witnessed a rapid escalation in the cost of warfare generally, and in particular horses, whose 'price tripled between 1140 and 1180, and had doubled again by i22o'.43 During the 12605, when the income of the French King Louis IX stood at around 250,000

WARFARE

11

Hvres tournois per year, the total annual cost of maintaining a mere hundred French knights whom Louis had stationed in Acre came to 10,000 Hvres tournois. This was a cripplingly expensive bill which, so the evidence suggests, even Louis could not always afford to pay on his own.44 Thus the cost of equipping and transporting expeditions to the East became so great that crusading armies shrank in size, and could only be financed by the wealthiest European lords.45 These factors, combined with the fact that most crusaders lacked any experience of fighting in the Latin East, weakened the effectiveness of expeditions. In 1269, for example, an Aragonese crusade to the Holy Land enabled the Franks to launch a raid against Muslim villages near Montfort with an army containing 130 knights.46 This force was far too small, however, to recapture any Christian territories lost to the Mamluks, or indeed risk a direct confrontation with Baybars, who was said to have had such a large field army close by that one contingent alone numbered 15,000 men. Moreover, during skirmishes just outside Acre, some Aragonese leaders of the campaign, declaring that they had come to fight for Christ, were needlessly killed because they simply charged into the enemy ranks and were immediately cut down by Muslim troops ravaging the area. The crusade achieved nothing permanent, wasted Christian lives, and provoked Baybars into carrying out a damaging counter-raid against Acre.47 Temporary crusades could therefore create more problems than they solved, for they did not remedy the permanent shortfall of troops, and even if they did succeed in regaining territory this had to be defended by local Franks who subsequently found themselves even more thinly spread out than before. On top of this there were the usual security problems common to virtually all parts of medieval Europe. Incidents of localised raiding, violence and crime were common, and there was also the more remote but nevertheless real possibility of an uprising by Muslim peasants under Latin rule. Indeed, during the years or decades which separated full-scale invasions by Saladin, Baybars or Kalavun, this must have been the type of fighting which affected most Christian territories. We are therefore left with the impression of a region where Latin settlers virtually always found themselves heavily outnumbered by external aggressors, but also had to cope with the possibility of lawlessness or rebellion within their own lands. It is against this background that the various functions of crusader fortifications should be discussed, for castles and urban defences were relied on by the Christians to protect their property, maintain their hold on the land and above all make up for their lack of soldiers in the field. Unlike European crusades, such structures lasted for decades or even centuries, and remained firmly under the control of local settlers. They were therefore relied upon to solve

12

U N K N O W N C R U S A D E R CASTLES

the numerous military weaknesses which have been described in this chapter, and as such became vital to the continued existence of the entire crusading movement in the Holy Land.

2

Military Architecture During the thirteenth century the Latins occupied a vast array of fortifications in the Holy Land, most of which were either built on top of, or incorporated into, earlier Arabic, Byzantine, Roman or even older defences. Indeed, some strongpoints, such as the vast fifth-century ramparts at Antioch, barely contained any Prankish building work at all, whilst fortresses like Montfort and 'Atlit, which were built on previously unoccupied or long abandoned sites, were extremely rare. These complexities make it preferable to divide a few Latin strongpoints into general categories depending on size, location and function, rather than describing all fortifications built in the area individually. First, the Franks held a number of major castles which were constructed on a very large scale in order to withstand substantial Muslim invasion forces. In the north, the best preserved and most famous such fortifications were the Hospitaller strongholds of Margat and Crac des Chevaliers. The outer walls of Margat enclosed an extensive triangular mountain spur which dominated the coastal route between Tripoli and the principality of Antioch. Indeed, Margat covered such a large area that it contained a small town, situated to the north of the inner citadel and separated from it by a rock-cut moat. The citadel itself consisted of numerous towers and buildings ranged around an inner courtyard, many of which date from the period after 1186, when the Hospitallers bought Margat from its original secular owner. The most important of these structures was the keep, a huge round tower with walls at least 5.5 metres thick which stood at the exposed southern tip of the site and therefore guarded the likeliest point of attack.1 The layout of Margat's defences compares very closely with that of Crac des Chevaliers, another former baronial castle which was also situated on a mountain spur and was acquired by the Hospitallers in ii42.2 Thereafter the Order added several flanking towers and an enormous talus to the south and west sides of the original enclosed courtyard, and then constructed a whole new curtain wall around the entire site. As was the case at Margat, the keep was placed at the most vulnerable point along the south side of the castle, where it formed the central flanking tower of the inner bailey.3 These major alterations appear to have been undertaken after the earthquakes of

14

U N K N O W N C R U S A D E R CASTLES

1170,1200/01 and 1202/03 but to have been completed by around 1220.4 The earthquake of 1170 was particularly devastating, for according to one Muslim source, it left 'several fortresses of the Franks ... such as Hisn al-Akrad (Crac des Chevaliers), Safitha (Chastel Blanc), al-Arimah [and] Arkas ... plunged into an ocean of ruins'.5 Bearing in mind that the Hospitallers did not acquire Margat until 1186, this suggests that the Order carried out its building programme in the thirty years following Saladin's invasion of Syria. This is made all the likelier by the fact that virtually identical improvements were made by the Hospitallers at the fortress of Silifke, an Armenian stronghold which they did not acquire until 1210.6 Not far from Crac des Chevaliers and Margat, the Templars held the equally substantial strongholds of Tortosa and Chastel Blanc (Safita), which were also considerably rebuilt during the crusader period. This was done at a much earlier stage, however, for a document dating from 1152 reveals that the Templars had probably already acquired Chastel Blanc by this stage, and were in the process of constructing new defences at Tortosa.7 Unlike Crac des Chevaliers and Margat, where all new Hospitaller towers were either round or horseshoe shaped and donjons were attached to adjacent ramparts, Tortosa was provided with a massive rectangular keep which stood in isolation at the north-west corner of the site. During a further phase which appears to date from the early thirteenth century, and therefore coincided with the Hospitallers' work at Margat and Crac des Chevaliers, this keep was provided with two flanking corner towers placed at the water's edge. At this point the keep was also protected from potential land attacks by the addition of two successive ditches and curtain walls to the north, east and south and was flanked by two corner towers situated at the water's edge. Some idea of the sheer scale of these defences can be gained from those sections of the inner curtain wall which still survive to their original height of 25.5 metres.8 A similar process seems to have taken place at the hill top fortress of Chastel Blanc, for the impressive rectangular keep which stands at the summit was probably constructed in the middle years of the twelfth century, or perhaps after the earthquake of 1170, whilst some at least of the outer curtain walls and elaborate gateways defending the lower slopes were probably constructed later.9 Indeed, there may have been several phases of improvement and reparation, for we have seen that in 1200/01 and 1202/03 more earthquakes struck the region, whilst in 1218 an Aleppine invasion force attacking Chastel Blanc supposedly 'destroyed its towers'.10 Meanwhile, the Templars were responsible for the construction of another rectangular keep and related defences at the fortress of al-'Arimah, which occupied a long ridge above the coastal plain between Tripoli and Tortosa. Little

MILITARY ARCHITECTURE

15

Crac des Chevaliers. Note the double walls and complex entrance passage. The revetments and vaulted undercrofts of the inner bailey strengthened the castle against siege weapons and earthquakes.

16

UNKNOWN CRUSADER CASTLES

is known about the history of this fortress, although it had probably already been acquired by the Templars with Chastel Blanc at some point before 1152.n This group of Templar bases, along with both Margat and Crac des Chevaliers, survived numerous Muslim incursions in the course of the thirteenth century before finally succumbing to the Mamluks from 1271 onwards. In that year Crac des Chevaliers and Chastel Blanc fell,12 followed by Margat in 128513 and Tortosa in i29i.14 Al-'Arimah may have been lost in 1271, for hi 1282 a peace treaty between the Muslims and the Templars stated that the Order still possessed some estates nearby, but that the fortress itself now belonged to the sultan of Egypt.15 In the kingdom of Jerusalem, there were a number of major strongholds which resembled Tortosa, Chastel Blanc, Crac des Chevaliers and Margat. Tortosa was similar to the Templar fortress of 'Atlit (Pilgrims' Castle), in that both could rely on the sea to protect them from the west but needed very extensive defences against attackers approaching from inland across the relatively flat coastal terrain. 'Atlit, which was constructed from scratch in 1217-18 with the aid of the Teutonic Knights and members of the Fifth Crusade, was better off than Tortosa because it actually lay on a promontory jutting out into the sea, but it still had vast man-made defences to protect its landward side. These consisted of an outer ditch and rampart with three gate towers, and a much higher inner wall whose two flanking towers were so tall that their defenders could spot potential attackers eight miles away. Together these towers acted as the inner donjon of the castle, in much the same way that the inner southern towers of Crac des Chevaliers did.16 Like Tortosa, 'Atlit also proved to be so well defended that it was never taken by force, being simply evacuated by the Templars in August i2pi.17 The appearance of both Tortosa and 'Atlit was determined by their proximity to the sea, but there were also fortresses in the south situated on hill tops and which required successive defensive rings to protect them on all sides. About half a century after they completed Chastel Blanc, the Templars rebuilt Saphet, an important Galilean castle which they had originally acquired around 1168 and which was restored to them as a ruin in the peace treaty of 1240.18 The new fortress built after 1240 consisted of an outer wall, approximately 22 metres high and 825 metres long, which Tan around a second and much higher rampart dominated by a large, probably circular keep. The exact appearance of this keep is difficult to establish, for it was later replaced by a similar Mamluk structure before a nineteenthcentury earthquake flattened the entire fortress. It may have been one of the seven towers along the inner rampart mentioned in De constructione castri Saphet, an invaluable contemporary account of the castle's construction and appearance. This document indicates that the outer wall was also

MILITARY ARCHITECTURE

V

'Atlit: the fortress and the town (or bourg).

flanked by seven towers, and was provided with underground tunnels which led to several casemates guarding the castle's outer moat.19 When it was besieged by the Muslims in 1266, this massive stronghold managed to withstand six weeks of almost constant mining and bombardment, only capitulating once sultan Baybars had managed to sow discord amongst its defenders.20 Such resilience helps to explain why the Military Orders had already begun to erect similar concentric castles in the second half of the twelfth century, particularly at Belvoir, an almost symmetrical fortress composed of two successive walls forming a square within a square. Belvoir was built by the Hospitallers between 1168 and 1189 on a lofty hilltop guarding the Jordan crossings south of Lake Tiberias. In 1189 it was lost to Saladin. Although it was theoretically restored to the Franks in 1241, it seems unlikely that the Hospitallers actually reoccupied it during the brief period before the Egyptian and Khwarizmian conquests of 1244-47.^ The inland castles mentioned so far had concentric defences intended to protect them on all sides. There were also a number of important strongholds located in more isolated surroundings with less regular fortifications. In

18

U N K N O W N CRUSADER CASTLES

1255 Alexander IV, acting on behalf of the monks who lived there, granted the Hospitallers Mount Tabor, a strategic hill top which lay within sight of Belvoir and to some extent must therefore have been intended to compensate the Order for the loss of this latter castle.22 In 1211 the Egyptian sultan al-Adil had ringed the summit of Mount Tabor with a massive curtain wall, 1750 metres in length and flanked by ten towers.23 Seven years later, following a failed Latin attack on the castle,24 the Muslims decided to demolish this rampart and abandon the site rather than risk a more successful Prankish assault in the future.25 One might therefore expect that when the Hospitallers arrived more than three decades later they would have attempted to repair and regarrison al-Adil's wall, but there is no archaeological evidence to suggest that plans to build new defences were ever put into action. Instead it seems that the Hospitallers merely occupied (and perhaps fortified) the small monastery in the south-east corner of al-Adil's castle, and simply relied on the height of Mount Tabor's summit and the steepness of its slopes to keep attackers out. To some extent this may have been done because of inadequate troops and resources, for in 1263 Baybars seems to have captured the site with such ease that more extensive man made defences would clearly have been desirable.26 The Muslims found it far harder to capture heavily fortified mountain strongholds such as the castle of Beaufort, situated in the mountainous interior south east of Sidon. Beaufort could rely on sheer cliffs hundreds of metres high to protect it from the east, whilst steep gorges deepened by the Franks defended it from the north and north west. Consequently, the outcrop occupied by this castle formed an isolated stronghold which only needed flanking towers and multiple ramparts along its more exposed southern face. This fortress, lost to Saladin in 1190, was regained by the Franks in 1240 and garrisoned by them until Baybars conquered it for good in 1268. During this period the Franks strengthened the castle's defences by building a separate citadel which was situated on the southern plateau opposite the castle's main entrance.27 Another impressive example of a mountain castle built by the Franks was Montfort. This fortress lay at the heart of an extensive lordship held by the Teutonic Knights a few miles north east of Acre and acted as the Order's headquarters until it was captured by Baybars in 1271. Montfort was built on a precipitous spur formally acquired by the German knights in 1228 and therefore dates entirely from the thirteenth century.28 Its defences were arranged around a large 'D' shaped keep which stood at the eastern end of the fortified spur and was isolated by rock-cut ditches. Access to this keep was also blocked by a wall along the north and west slopes of the castle, although the apparent absence of any medieval remains on the south

MILITARY ARCHITECTURE

19

side suggest that this slope was considered steep enough not to require additional man-made defences.29 This 'semi-concentric' arrangement would not have been feasible at the more gently-sloping sites of Saphet, Crac des Chevaliers, Margat and Belvoir.30 To the north of Montfort, the castle of Akkar (Gibelcar), situated in the county of Tripoli, also occupied a mountain spur which only needed extensive defences where it could be approached from the neighbouring mountains. Its keep, a relatively simple quadrangular structure, also resembled that of Montfort in that it was placed at the neck of the ridge connecting the castle with surrounding hills. This ridge had also been severed by a rock-cut moat, whereas the rest of the site was so inaccessible that it was apparently only protected by a single curtain wall following the edge of the spur.31 Perhaps the most extreme use of nature to protect a stronghold occurred not at larger fortresses such as Akkar or Montfort but at Cave de Tyron, a cave fortress carved into the side of a mountain opposite the town of Sidon. Even though it only housed a tiny Prankish garrison, it must have been virtually impregnable, for it lay hundreds of metres above sea level and could only be approached along a path barely one metre wide.32 Both Akkar and Cave de Tyron had been occupied by the Franks at various times since the twelfth century, and the former was not lost until 1271, whilst the latter seems to have been abandoned by 1260.33 Apart from castles, the Franks also built or occupied a number of citadels intended to defend existing urban settlements, and these could often be as powerful as individual fortresses. Even though it has already been discussed in the context of major Syrian castles, Tortosa was strictly speaking a citadel rather than an independent fortress, for a much older walled town had already existed next to it before the Templars built their new keep from 1152 onwards. By contrast, 'Atlit was built from scratch by the Templars in 1217-18, effectively becoming a citadel once a walled settlement sprang up just beyond its outer defences. Thus the distinction between castles and citadels is not always such an easy or useful one to make, but if we define most citadels as strongholds which were attached to important walled settlements already in existence before the time of the First Crusade, it is clear that the Latins possessed a number of such structures both inland and near the sea. Along the coast, the citadels of Arsuf, Beirut, Jaffa and Ascalon all resembled Tortosa in that they were placed as close to the shore as possible, in order to protect them from the west, but needed fairly elaborate defences on their landward side. This was clearly the case at Beirut, whose citadel had been left intact by Saladin in 1187, was recaptured and repaired by German crusaders in 1197, and then strengthened even further by the new Ibelin lord of the city in the early thirteenth century.34 When the German pilgrim Willbrand of

2O

UNKNOWN CRUSADER CASTLES

Oldenburg passed this castle in 1211, he described it as a large, imposing structure, built on a rocky knoll and defended by a deep ditch, several towers and two successive curtain walls.35 Investigations carried out early last century appear to confirm that it did indeed stand next to the shore on the site of a later Ottoman fort, whilst present rebuilding work in Beirut has uncovered some of the castle's masonry which is similar to that at Belvoir.36 To the south of Beirut, Arsuf, Jaffa and Ascalon were all partially or totally destroyed by Saladin's forces at the time of the Third Crusade.37 But between 1191 and 1192 Richard Lionheart reoccupied these sites along with Caesarea, and he may have carried out some repairs both there and at Arsuf. He also built more extensive new fortifications at Ascalon and Jaffa,38 the latter successfully withstanding a Muslim attack shortly afterwards.39 The sheer speed with which Richard erected these defences suggests that they generally amounted to a limited reconstruction of citadels rather than city walls, although at Ascalon the inner curtain wall around the town was hastily rebuilt, even if many of the powerful outworks which had protected the site in the twelfth century must surely have been left in ruins.40 Even these efforts proved short-lived, for in 1192 Ascalon was flattened as part of Richard's peace treaty with Saladin, and five years later Jaffa was sacked and demolished by the Muslims.41 Caesarea and Arsuf, meanwhile, may only have been defended by citadels during the next couple of decades, for during the Fifth Crusade (1218) the city walls of the former were rebuilt, whilst in 1211 the inhabitants of the latter still lived in constant fear of bandits, implying that the urban defences of these settlements were either inadequate or possibly even incomplete.42 Subsequent efforts to fortify Caesarea, Ascalon, Jaffa and Arsuf were just as patchy. The fortifications of Ascalon lay in ruins from 1192 until the crusade of 1240-41, when reconstruction of the citadel was begun by Theobald of Champagne and completed by Richard of Cornwall.43 Thereafter the citadel was entrusted to the Hospitallers, but Ascalon was lost again to the Egyptians a mere six years later.44 Further north the ramparts erected at Caesarea in 1218 were destroyed by the ruler of Damascus within a few months of their completion.45 Both this settlement and Jaffa, which had apparently remained unfortified since 1197, were given new citadels during the Emperor Frederick II's rebuilding programme in the late i22os.46 A quarter of a century later Louis IX considered it necessary to provide Caesarea and Jaffa with massive new urban fortifications, suggesting that for the previous twenty-five years these places had primarily relied on Frederick's citadels to protect them. Even if their city walls had been repaired since Frederick's time, they were now falling into decay again.47 A similar situation may have prevailed at Arsuf, for although the citadel was strengthened by

MILITARY ARCHITECTURE

21

Ascalon: the city walls and citadel.

John of Ibelin in 1240, and later described as 'very strong' by Joinville,48 the urban defences were considerably improved after the Hospitallers acquired the site in i26i.49 This again implies that whilst Arsuf must have had walls around it for much of the thirteenth century, these were not always in a particularly good state, and that the Ibelins, unable to foot the maintenance bill any longer, therefore sold the lordship to a Military Order. By this point the military situation had deteriorated to such an extent that not even the Hospitallers could hang on to Arsuf, which fell to Baybars in 1265 along with Caesarea. Three years later, Jaffa was also lost for good.50

22

UNKNOWN

C R U S A D E R CASTLES

The citadels of Jaffa, Arsuf and Ascalon no longer survive, but at the latter site a combination of historical and archaeological evidence suggests that the stronghold constructed during the Crusades of Theobald of Champagne and Richard of Cornwall amounted to an impressive structure with towers, ditches and double walls built out of ashlar blocks and recycled Roman columns. It was probably situated on a mound in the north-west corner of the city, where it is still possible to make out two approximately rectangular lines of defence, plus the remains of a masonry glacis.51 At Arsuf the citadel stood on an almost identical hill to that of Ascalon, being separated from the surrounding town by a deep moat and connected to the port via a long staircase.52 Although it has long since vanished, the citadel of Jaffa should be placed in the same category,53 as indeed should that of Tiberias, a pre-iiSj Prankish settlement which, like Ascalon, only came under effective crusader rule in the thirteenth century between the peace treaty of 1241 and the Egyptian conquests of 1247. During that brief spell its Prankish lord, Odo of Montbéliard, rebuilt the town's citadel, which was defended by a massive wall and moat on one side but could rely on Lake Tiberias (rather than the Mediterranean) to shield it on the other.5* Whilst the coastal citadels were situated at the water's edge, others made even greater use of the sea as a means of protection. That of Caesarea, whose thirteenth-century history has already been touched upon, stood on a narrow promontory whose neck was defended by powerful ramparts and a water-filled moat.55 This castle represented a scaled down version of 'Atlit, as did the baronial fortress of Nephin, which was situated just south of Tripoli on a small headland defended by 'twelve good towers' and separated from the coastal plain by two rock-hewn ditches.56 For most of the twelfth century Nephin was held by the same baronial family as that which controlled Maraclea, another coastal settlement located between Tortosa and Margat. This site does not seem to have been particularly well fortified, for in 1188 it was evacuated just before the arrival of Saladin, and in 1271 Baybars probably occupied it soon after the fall of Crac des Chevaliers.57 However, the inhabitants may have used an isolated rock, situated about fifty metres off the coast, as a place of refuge in times of danger. At some point after 1277 Bartholomew de Ravendel, lord of Maraclea, returned to the area and built a powerful new tower on this rock. Although Bartholomew's tower was demolished in 1285 as part of a peace treaty with Sultan Kalavun of Egypt,58 its design and location can be compared with that of the sea castle at Sidon, built by Frederick II's followers during the winter of 1227-28 on a small island in the harbour. This structure, which was connected to the shore via a stone and wooden bridge, was gradually altered and improved

MILITARY ARCHITECTURE

23

by its crusader occupants until it was finally evacuated shortly after the fall of Acre in i29i.59 Another tactic adopted by the Franks was to place urban citadels not at the strongest point of the defences but at the weakest, as they had done with the donjons of Montfort and Margat. The fortifications of such sites tended to be more regular in design, for they could not rely on the sea to protect them. Hence at Gibelet, situated between Nephin and Beirut, the citadel consisted of a rectangular enclosure with four corner towers built around an isolated central keep. It stood at some distance from the sea, so that it could guard the land approaches to the walled town which grew up between it and the shore.60 This fortress, which had originally been built at the beginning of the twelfth century, was lost to Saladin in 1188 but was then regained through negotiations in 1197 and held by the Genoese Embriaco family at least until 1289.6! The layout of Gibelet was also copied at Sidon, for in 1253 Louis IX constructed an irregular enclosure defended by several flanking towers on a small motte at the most landward extremity of the town. Excavation of this motte indicates that Louis' land castle merely replaced an earlier twelfth-century fortification, which must therefore have been the town's only refuge point until the sea castle was constructed in 1227-28.62 Contemporary descriptions and plans suggest that the citadels of Acre and Tyre were also rectangular in form, and perhaps resembled a partially preserved urban castle built at Famagusta around 1300.63 That of Tyre was placed 'near the Sidon gate', suggesting that it defended the northern side of the narrow causeway which connected the city with the mainland. It was apparently reconstructed from around 1212 onwards, for in that year John of Brienne, king of Jerusalem, bought land inside the city so that he could finish building a new castle.64 At Acre the citadel had originally fulfilled a similar purpose to that of Gibelet, but rapid urban growth during the crusader period eventually left it stranded behind newer ramparts constructed further inland. By the late thirteenth century it functioned more as a barracks than as a strongpoint, and appears to have played no significant role in the defence of Acre during the final siege of i29i.65 A more crucial role was played by the citadel of Jerusalem, where the massive Herodian structure known as the Tower of David was incorporated into a medieval stronghold built around a central courtyard and attached to the western wall of the city. The Tower of David, which was so strong that one contemporary described it as 'a single stone from its base up',66 clearly formed the core of this citadel, but it seems that other elements were added or repaired by the Latins until the late 12308. Most notably, archaeologists believe that parts of the sloped masonry revetments which form the

24

U N K N O W N C R U S A D E R CASTLES

base of the present Ottoman enclosure are Prankish in origin.67 Between 1229 and 1244 this structure withstood a number of important sieges.68 Finally, reference should be made to the citadels of Antioch and Tripoli, which differed from those of Acre, Tyre and Jerusalem in that they relied more on rugged terrain than man-made defences to protect them. At Tripoli the citadel had originally been built during the initial Prankish siege of the city between 1102 and 1109 and consequently it was not attached to the urban defences but stood on a rocky knoll some distance inland. Eventually it formed the nucleus of an entirely new suburb but was rarely even mentioned in contemporary descriptions of the city.69 By contrast, the tenth-century citadel of Antioch was situated at a point where the older Theodosian and Justinianic ramparts surrounding the city reached the summit of Mount Silpius. It was therefore easy for the local population to reach, yet elevated enough to be extremely difficult to besiege.70 Perhaps it is hardly surprising, therefore, that when Tripoli fell in 1289 no one appears to have sheltered inside its castle,71 whereas the citadel of Antioch contained thousands of people when Baybars obtained its surrender in 1268.72 The compact design of crusader castles and citadels forms a sharp contrast with the numerous urban fortifications built or inherited by the Franks. Unlike certain strongholds such as Montfort or Akkar, major settlements were without exception located in low-lying areas and could therefore require even larger defences than those at sites like Saphet and Crac des Chevaliers. In 1335 a German traveller visiting Acre wrote that 'this famous city situated on the coast is constructed using extraordinarily large blocks of stone, with high and strong towers standing scarcely a stone's throw away from each other. Each gate is flanked by two towers. The walls were, and still are, so thick that two chariots going in opposite directions could easily meet on them. On the landward side they were also extremely powerful, with very deep ditches, further protected by a series of bastions and outworks of various types'.73 A number of similar contemporary descriptions, plus medieval maps and modern surveys, confirm that Acre was indeed protected by vast double ramparts which were flanked by at least twenty towers and ran for about two kilometres across the headland occupied by the city.74 The most significant thirteenth-century improvements made to these structures were carried out on the section defending the northern suburb of Montmusard by Louis IX in the early i25os.75 Apart from Acre, there were major urban fortifications at Tripoli, Tyre, Jerusalem and Antioch. Those of the latter city, which stood on the slopes of Mount Silpius, formed a triangle whose apex was defended by its mountain-top citadel. These walls were eighteen kilometres long and were renowned for their strength, but had actually been built by Byzantine

MILITARY ARCHITECTURE

25

Acre in the thirteenth century. Town plan embodying attempted reconstruction of old walls, scale 1:10,000, based on the present-day town plan (modern fortifications indicated by broken lines) but showing quarters referred to in old maps of the fourteenth to eighteenth centuries and giving their original Latin names. The lay-out of streets is based partly on these maps and partly on traces still discernible in the modern town. Extant or recently excavated buildings are reproduced in black (e.g. remnants of the Hospitaller quarter and the tower, or so-called Burg es-Sultan, which still stands beside the harbour).

emperors during the fifth and sixth centuries.76 Consequently, any Prankish building work undertaken at Antioch could only have amounted to small repairs and improvements. It has been suggested, for example, that a gate on the south slopes of Mount Silpius, plus a rectangular enclosure on the south-west side of the citadel itself, were crusader additions.77 This was in marked contrast to the much more extensive crusader fortifications

26

U N K N O W N CRUSADER CASTLES

constructed at Tripoli and Tyre, whose towers, ditches and double ramparts must have been similar to those of Acre. Tyre was also said to have had at least three successive curtain walls and as many as five separate gateways, plus a vast moat occasionally rilled with sea water, protecting the neck of land which connected it with the mainland.78 Jerusalem, on the other hand, was only ringed by a single curtain wall which generally followed the same course as the present Ottoman defences.79 After the Muslims captured Jerusalem in 1187, they initially repaired and maintained its defences, but these were subsequently slighted by al-Muazzam, ruler of Damascus, in 1219, as part of the same scorched earth tactics which led to the destruction of Mount Tabor.80 The historical and archaeological evidence suggests that, after Frederick II regained Jerusalem in 1229, the Christians tried to rectify this situation by repairing Sion Gate and St Stephen's (or the Damascus) Gate, located in the south and north of the city respectively. At St Stephen's Gate remains have been found of an 'L' shaped entrance guarded by a fairly elaborate barbican incorporating a church and possibly a customs house, whilst at Sion Gate it seems a similar bent entrance originally constructed by the Ayyubids gave access to the city through the side wall of a massive guard tower.81 Post-1229 improvements to these structures may have been carried out by the Teutonic Knights, to whom Frederick gave several properties in the city and relied on to garrison his new acquisition.82 Presumably they also show that the rest of the city's gates and walls were rebuilt, or perhaps simply repaired if al-Muazzam had actually only caused damage in places rather than flattening the entire circuit. On the other hand, the apparent ease with which 15,000 Muslim peasants broke into and looted Jerusalem shortly after Frederick's departure in 1229 suggests that the walls were either inadequately restored or severely undergarrisoned (or both) after that date.83 If this was the case, then life must have been difficult for the few Latins who returned to the city between 1229 and 1244, when Jerusalem was finally lost to the Khwarizmians.84 It has been estimated that in the twelfth century the populations of Acre, Tyre, and Jerusalem stood at around 30,ooo.85 In the thirteenth century the former two cities probably continued to have at least this many inhabitants living within their walls, but the population of Jerusalem between 1229 and 1244 may have only been a fraction of the pre-nS/ total, and perhaps as low as 5000 to 10,000 people.86 In terms of population at least, it had therefore slipped into the second rank of walled cities alongside places like Caesarea (perhaps 4800), Sidon (perhaps 5600) and Arsuf (perhaps 36oo).87 It is clear that several other cities of a similar size were fortified in some way during part or all of the crusader period, including Beirut, Jaffa, Ascalon, Gibelet, Tortosa and Haifa along the coast, plus Tiberias in the interior.

MILITARY ARCHITECTURE

27

We have already seen that this latter site was returned to the Franks in 1241 and that a new citadel was subsequently built there, but it seems unlikely that the entire city was properly recolonised, let alone refortified, during the six years before it was lost again to the Egyptians in 1247. A similar situation probably prevailed at Ascalon, whose citadel was rebuilt and lost again during the very same period. By contrast, it has already been shown that many of the other coastal settlements just mentioned were inhabited for much longer periods in the thirteenth century, and that sporadic attempts were at least made to build or maintain walls around them.88 However, it is only at Caesarea that any major archaeological remains have survived, for here it is still possible to view the extensive new defences erected by Louis IX between 1251 and 1252. These consisted of a sturdy rampart which was supported by a sloping masonry talus, was flanked by sixteen rectangular towers and was protected by a moat at least seven metres wide. From the base of the ditch, the revetment was eight metres high and the rampart, which has not survived, was probably a further ten metres high. The city gates penetrating this rampart were similar to those of Jerusalem, in that they were incorporated into the sides of towers and then turned through ninety degrees before leading into the town itself.89

Caesarea: the citadel and Louis DCs city walls.

28

UNKNOWN CRUSADER

CASTLES

Hence Caesarea had fairly elaborate defences from 1252 onwards, and considering that Louis IX also carried out repairs and improvements at Jaffa, Sidon and Haifa during this period, they too may have had similar fortifications.90 However, it seems unlikely that the city walls of all crusader towns were as large or as powerful as those of Caesarea. In 1231, for example, imperialist forces attacking the Ibelin lordship of Beirut were said to have breached the town wall in the space of a single night, whereas the actual citadel successfully withstood a determined siege lasting several months.91 Beirut's urban defences were probably similar in scale to those of Prankish bourgs or castle towns.92 These were new or growing settlements which had sprung up next to most larger fortresses, including Beaufort, Saphet, Crac des Chevaliers, Margat and 'Atlit. These sites were normally protected by a much simpler curtain wall and ditch than that which surrounded Caesarea. The wall around 'Atlit, for example, had a total length of about 800 metres, but was only defended by four small towers, only one of which could provide flanking fire for a gateway. The other entrances to the town were far more exposed than those at Caesarea, because they led directly through the wall rather than being incorporated into the sides of towers. All these structures were built shortly after the completion of the castle itself in c. 1220.93 Along with these relatively minor urban defences, the Franks also held countless smaller castles which were dotted across the countryside. Although their shape and size could vary greatly according to location, many were either fortified enclosures or individual towers. A good example of the former type can be found at Goliath, a Hospitaller property situated about 20 kilometres north of Tripoli. Goliath formed a rectangle which measured 63 metres by 56 metres, had four small corner towers, and was also provided with a fifth salient guarding the actual gateway.94 Castles of this kind are sometimes referred to as castrum fortifications because of their close architectural links with Roman and Byzantine forts, although they could also be based on Arabic farmsteads of the type which had been constructed in the East for many centuries. However, a square or rectangular enclosure is such an obvious way of defending a low-lying site that it could just as well have been developed by the Franks themselves.95 The same could be said for what one scholar has termed 'courtyard buildings':96 even smaller fortified enclosures which were built as hospitals, monastic buildings, administrative centres or secular farmsteads. The latter category is well represented by the Castle of Roger the Lombard, which was situated a few miles south of Caesarea and was probably constructed by Roger shortly after the First Crusade. It consisted of several vaulted structures built around a central courtyard, whose outer walls were approximately 1.6 metres thick and had

MILITARY ARCHITECTURE

29

very few openings in them, creating a fortified farm house roughly 33 metres square.97 Most other minor strongpoints took the form of individual towers; a design which is far likelier to have come directly from the West than have been inspired by Roman or Arabic precedents. One such tower which has been investigated by archaeologists in recent years is Qaqun (Caco), located on the Sharon plain about twenty-five kilometres south east of Caesarea. The remains of this building indicate that it measured 14.53 metres by 17.65 metres, that its walls were 2.8 metres thick, and that it had two vaulted storeys with a crenellated terrace above. There were no openings below the upper floor, which contained numerous arrow slits, a doorway, and traces of clay piping feeding an underground cistern. This cistern, plus a small perimeter wall, would have enabled the defenders to withstand a limited siege.98 A virtually identical castle in the county of Tripoli has survived at Chastel Rouge, whose inner tower measured 14 metres by 16 metres.99 Other structures of this kind, such as the tower of Tukhlah in the vicinity of Chastel Blanc, may have stood in isolation, although it seems likelier that their surrounding perimeter wall has simply disappeared.100 This outline of crusader fortifications in the Holy Land is by no means exhaustive, but it does enable us to identify two distinct trends in military architecture. First, it is clear that smaller castles and urban fortifications changed relatively little in the course of the twelfth and thirteenth centuries.

Goliath: a typical courtyard or castrum castle.

30

UNKNOWN CRUSADER CASTLES

Chastel Rouge. The castle is virtually identical to Qaqun and many other such structures.

All of the minor towers and fortified enclosures already mentioned date from before 1187, and although some, such as Goliath, were repaired or even rebuilt after that date, it seems that no efforts were made to improve their defences or radically alter their overall design before they were lost to the Mamluks between the 12608 and 12905. Similarly, it has been suggested that simpler urban defences, such as those which surrounded 'Atlit, consisted of little more than a straightforward perimeter wall and outer ditch, and appear to have been almost totally unaffected by advances in military architecture. As we shall see, this confirms that these sites were never intended to withstand major Muslim incursions involving tens of thousands of troops. On the other hand, it was clearly hoped that much more powerful Prankish strongpoints, such as the city walls of Acre, the citadel of Tortosa or the fortress of Crac des Chevaliers, would be strong enough to deal with aggressors of this kind. As a result, such places had to be upgraded constantly in order to deal with the growing strength of large Muslim armies, and in particular the far more powerful trebuchets which became available to the Ayyubids and Mamluks from the late twelfth century onwards. As we have seen, these weapons could hurl rocks weighing hundreds of kilogrammes with great accuracy and over relatively long distances, and so castles had to be adapted to cope with the new threat.101

MILITARY ARCHITECTURE

31

The most obvious way to do this was to build fortifications on a vast scale and to make exposed walls as thick as possible. Indeed, it has already been noted that the outer walls of Tortosa and Saphet alone were more than twenty metres high, and that the keep of Margat had walls 5.5 metres thick. The square layout of this tower's internal rooms meant that in some areas this figure rose to ten metres.102 Overall strength could also be improved by ensuring that the individual blocks of stone within walls were as large as possible, and sometimes this could be achieved by reusing the Cyclopean masonry of classical or pre-classical times. At 'Atlit for example, stones quarried or recycled locally were so large that they 'could barely be pulled in a cart by two oxen'."13 The extent to which such blocks were cut and dressed depended on the hardness of the stone, although some masonry, such as that used at Tortosa, may deliberately have been left with a bossed outer face in order to strengthen it against catapults.i04 The damage caused by sappers and trebuchets could also be reduced by bonding together stones with molten lead or iron clamps; a method which is known to have been used at Beirut, Sidon, 'Atlit and Maraclea.105 In addition, numerous fortifications had Roman columns incorporated into their ramparts, including the citadels of Ascalon, Tiberias 106 and Caesarea, where they had been 'placed horizontally in the body of the wall, in such a way that they had nothing to fear from sapping, and could not fall, even if they were undermined'.107 Muslim sappers were also kept out by placing curtain walls on top of sloping revetments; a technique which has already been referred to at Caesarea, Crac des Chevaliers and elsewhere.108 Similarly, in 1285 the foundations of one of Margat's outer towers, which stood at the southern tip of the castle just below the actual keep, formed such a solid lump of masonry that even after it had been brought down by sapping it could not be circumvented. Indeed, one Muslim source wrote that as a result 'the night passed in great confusion, for the use of catapults was made impossible by what had happened, and everything that could be done with mines had been done'. The Hospitallers still surrendered shortly afterwards, for they realised that the keep itself was now in danger of collapsing.109 The only definite way of escaping such a fate was to build castles on solid rock, which must have been one reason why the barren promontory occupied by 'Atlit was so attractive to the builders of this fortress.110 The growing power of trebuchets affected the appearance of fortifications in other ways. In the thirteenth century round or horseshoe towers, whose shape was not as vulnerable to artillery fire as the sharp corners of square salients, grew in popularity. These were constructed at Margat and Crac des Chevaliers around 1200, and possibly at Saphet in the early 124OS.111 It

32

UNKNOWN CRUSADER

CASTLES

is extremely likely that the former two castles were heavily influenced by the work of Armenian military architects, who had already recognised the advantages of rounded towers long ago.112 It is also interesting to note that even when thirteenth-century Prankish castle builders did continue to use quadrangular towers, these often took on a rectangular rather than a square shape, so that walls exposed to artillery fire became more elongated with greater distances between vulnerable corners. This development can be seen very clearly in the early thirteenth-century fortifications erected by the Templars at Tortosa and 'Atlit.113 As the range of trebuchets increased, there was also greater emphasis on trying to prevent such weapons, or indeed other forms of attack such as mines, ladders and crossbows, from getting close enough to a strongpoint in the first place. This was the primary role of castle ditches. At Saphet, for example, the ditches were more than fifteen metres deep and more than thirteen metres wide.114 The moat at 'Atlit even seems to have had a stone counterscarp which rose above ground level, forming a parapet where archers could be posted to prevent attackers from even reaching its outer lip.115 Several coastal strongpoints such as Nephin also had water-filled ditches;116 a method which was occasionally used at inland fortresses too, most famously along the southern front of Crac des Chevaliers.117 Furthermore, an account of the siege of Acre in 1291 indicates that one tower along the city walls had wooden hoardings with huge spikes attached to its base, suggesting that even if most ditches were dry, those of larger strongpoints could still contain obstacles which were lethal to infantry and cavalry alike.118 It is also from the siege of Acre that we hear of complex outworks, largely constructed during the last twenty years of Prankish rule and composed of a succession of ditches, earthworks and stone or wooden barriers.119 Such structures tended to be most elaborate in the vicinity of gateways, which, as we have already seen at Jerusalem and elsewhere, were often incorporated into the side walls of towers so that they were not exposed to a direct assault or bombardment by trebuchets.120 These entrance towers could themselves be equipped with portcullises, arrow slits and murder holes, making it extremely hazardous for attackers to enter castles even after their outer doorways had been breached. In 1276, for example, Templar troops besieging Nephin managed to reach the main gate safely, only to find themselves trapped after their opponents inside the castle lowered the portcullis behind them.121 Bearing in mind that Nephin was still a relatively small castle, it is easy to understand why Mamluk armies often preferred to keep their distance and rely on mining or bombardment against considerably larger strongholds such as Crac des Chevaliers, whose labyrinthine

MILITARY ARCHITECTURE

33

main entrance passage was so heavily defended that a direct assault against it would have resulted in very heavy casualties.122 Strong walls and deep ditches were a passive way of dealing with besiegers, but archers and counter-artillery enabled garrisons to strike back at their opponents. In order to maximise the effectiveness of garrison troops who were firing bows or crossbows, inner ramparts were built much higher than outer ones, whilst inner and outer towers alternated so that one wall or tower did not block the field of fire of another. Most towers and some ramparts also had rooms or shooting galleries incorporated into them, so that defenders were not simply restricted to firing from the tops of walls. The majority of these, defenders were straightforward archers or crossbowmen, but some operated much larger crossbows or trebuchets intended to destroy the trebuchets of their attackers. Another explanation for the sheer size of some thirteenth-century Prankish towers, therefore, is that they needed to be large enough to accommodate trebuchets on their roofs, which acted as flat, spacious fighting platforms supported by massive vaults. Behind these towers, inside the inner courtyards of castles, further vaults and undercrofts were needed to provide shelter against incoming artillery, support the outer ramparts against enemy bombardment, and take the strain of the castle's own trebuchets being fired continuously. The effect of this type of warfare on the appearance of a stronghold is perhaps most apparent at 'Atlit, which, as we have seen, was protected by a ditch with its own parapet for archers, an outer wall flanked by three towers, and a higher inner rampart whose massive double towers were probably designed to accommodate artillery weapons. Any attackers approaching this Templar stronghold across the open land to the east would therefore have been met by a shower of lethal bolts, arrows or large rocks being hurled at them.123 Indeed, in 1220 al-Muazzam, ruler of Damascus, had to call off his attempted siege of 'Atlit because this was precisely the kind of reception which greeted his forces.124 Nor was this the only place where the Templars used such weapons, because they installed them at Jaffa after they acquired this lordship in 1266 and employed 'crossbowmen with large crossbows' to guard the outer moat at Saphet.125 It is also possible that in 1188 Saladin was prevented from capturing the citadel of Tortosa because of the Templar crossbowmen stationed within it, implying that this Order was particularly skilled in the use of catapults and archers126 It is also in this context that we should view many of the improvements carried out at major strongholds in the late twelfth or early thirteenth centuries. At Tortosa, for example, we have seen that a straightforward mid twelfth-century rectangular keep had two flanking towers and elaborate outer defences added to it around 1200; a project which was clearly intended

34

UNKNOWN CRUSADER CASTLES

to provide more flanking fire for the keep and to stop attackers armed with trebuchets from getting too close to it.127 During their occupation of Beaufort from 1260 onwards, the Templars were also responsible for the new citadel constructed on the plateau opposite the castle. This was done to ensure that attackers would not deploy siege weapons in this area, although it still failed to prevent Baybars from doing just that when he captured Beaufort in 1268.128 We have seen that large parts of the inner defences and the entire outer circuit walls at Crac des Chevaliers were also completed around 1200, so that the fortress could withstand all but the most determined sappers and powerful trebuchets.129 Parallel developments took place in the Muslim world, for in the early thirteenth century massive new towers which were better able both to withstand and to accommodate trebuchets were added to the citadel of Cairo and numerous other major strongpoints.130 Finally, the most active defence of all entailed leaving the safety of a castle and launching minor raids, or sorties, against besieging forces in order to throw them off guard and hopefully destroy their siege weapons. This tactic required doorways, or posterns, to be incorporated into walls and towers which were large enough to make surprise attacks but small enough to defend easily. At Crac des Chevaliers, for example, there were a number of posterns situated along the outer ramparts, plus another small gateway at the north-west corner of the inner ward. Consequently the garrison could conduct an active defence even after the outer walls had fallen, and could continue the fight until virtually every building in the fortress had been captured.131 If a siege reached this point, however, one disadvantage for the defenders of a major strongpoint with multiple or concentric fortifications was that outlying structures could actually provide shelter for attackers. In 1220 concerns that this would happen explain why the Templar defenders of 'Atlit, on hearing that a large enemy force was approaching, responded by partially demolishing the outlying watch tower of Destroit before they retreated inside their fortress.132 In 1265, when Baybars carried out an extensive raid in the vicinity of Acre, its garrison did the same thing and chose to destroy many neighbouring buildings rather than watch the Muslims use them in an attack on the city itself.133 In the same year something similar actually took place at Caesarea. Once Baybars had broken through the walls of this settlement, he used the city's cathedral as a useful vantage point from which to shoot at Franks still holding out in the citadel.134 In another very interesting incident relating to Acre, some sources noted that during the final siege of 1291 the Latins preferred to leave the city gates open. This was done so that early on in the campaign the Franks were free to launch surprise raids against the Muslims assembling outside and to

MILITARY ARCHITECTURE

35

prevent them from getting too close to the walls.135 By the late thirteenth century the very largest crusader castles and cities should no longer be regarded as passive refuge points, but rather as centres of a very active and flexible defence which could sometimes radiate for many hundreds of yards into the surrounding countryside. Whilst this type of warfare ultimately decided the fate of the Holy Land, it is important to note that only a few major strongpoints such as Acre, Tyre, Crac des Chevaliers and 'Atlit were intended to cope with it. The massive fortifications which have been described at these sites were very expensive to build and maintain, only constructed reluctantly in response to the growing size of Muslim armies and the greater effectiveness of enemy siege weapons. The vast majority of Prankish strongholds dotted across the countryside were not intended to withstand such forces, and generally consisted of much smaller towers or enclosures whose size and shape varied considerably according to their location and the wealth of their owners. Similarly, virtually all the castles built in Cyprus before 1291 or in Greece after the Fourth Crusade were relatively small and poorly constructed, because warfare in these regions did not involve tens of thousands of troops or highly sophisticated trebuchets.136 A determining factor was therefore economic. The Ayyubids and even more so the Mamluks could draw on the enormous wealth of their territories (and especially Egypt) to raise very large invasion forces and construct extremely powerful siege weapons. Fortunately for the Holy Land, the spending power of the Military Orders and the great wealth of crusaders such as Louis IX meant that it was possible for a small number of sites like 'Atlit to be constructed in response. To put the cost of building major fortifications into perspective, it should be remembered that the Templars supposedly spent 1,100,000 Saracen bezants rebuilding Saphet in the 12405, and then spent a further 40,000 bezants per year simply maintaining it.i37 This was at a time when a local knight's fee 'ranged in worth from 400 to looo Saracen bezants per annum'.™ In Prankish Greece the relative poverty and humble background of most Latin settlers simply made this kind of expenditure impossible. Indeed, this point even extended to the quality of the stonework used inside castles, for beautifully ornate structures, such as the gallery of the great hall at Crac des Chevaliers, which was built by highly skilled craftsmen in an ornate mid thirteenth-century gothic style, had no parallel in Greece apart from the royal residence of Chlemoutsi. Most castles in Greece were built from recycled or very poorly cut stones without any concession to luxury or decoration.139 Another reason for the impressive appearance of sites like 'Atlit was location. The Latins' reliance on the sea as a means of communication with

36

UNKNOWN CRUSADER CASTLES

western Europe meant that many key settlements were located in relatively exposed parts of the coastal plain, where massive man-made defences were needed to make up for the flat terrain. Once again, therefore, the Franks were driven by necessity rather than by choice, but some scholars have tended to equate size with strength, regardless of the circumstances. It has been argued that Montfort, for example, 'was hardly as impressive' as Saphet because it lacked the perfect concentric defences of the latter castle.140 This sentiment has been echoed by other scholars, who point to Montfort's relatively poor masonry as proof that it was architecturally inferior; an argument which can be backed up by looking at contemporary records showing that the Teutonic Knights needed considerable financial assistance to complete their castle.141 The fact that Baybars needed two attempts to capture it, however, but only one to take Saphet, suggests that in some ways Monfort was actually stronger. In other words, the Teutonic Knights did not need massive concentric defences to strengthen a site which was already relatively easy to defend because of its mountainous location.142 Further north, it should also be noted that Baybars found it easier to capture Crac des Chevaliers than the fairly simple but very inaccessible castle of Akkar. After he captured Akkar in 1271, Baybars boasted of his achievements in a mocking letter to Bohemond VI, where he described 'how we transported the mangonels there through mountains where the birds think it too difficult to nest; how patiently we hauled them, troubled by mud and struggling against rain'. By comparison the apparently much more sophisticated defences of Crac fell in less than a month because the garrison was under strength, lacking in food and unable to contact a relieving force which could stop the Muslims from undermining the walls.143 Military architecture cannot be studied in isolation, for numerous wider political, economic, geographical and military factors influenced the appearance and relative strength of castles. But there has often been a tendency to ignore these factors, with the result that Prankish military architecture has been regarded as a separate science which gradually evolved as the Latins discovered new methods of defence and rejected old ones. For example, one scholar has recently challenged William of Tyre's claim that the fortress of Montreal was already defended by 'walls, towers, forewalls and a moat' in 1115, because this would suggest that the Franks were constructing elaborate double defences early in the twelfth century, long before excavated concentric castles such as Belvoir appeared.144 This argument relies too much on the notion that the crusaders only 'discovered' concentric defences gradually. In reality, their experience of Roman, Byzantine and Arabic fortifications at sites like Antioch and Constantinople, combined with the self-evident advantages of providing a fortified summit

MILITARY ARCHITECTURE

37

with more than one line of defence, suggest that they were well aware of such concepts long before the First Crusade. Similarly, historians now agree that the castrum, or fortified quadrangular enclosure, was not 'discovered' in the Holy Land either, for although there were numerous Byzantine or Arabic structures of this kind for the crusaders to copy when they got there, they are just as likely to have been influenced by Roman fortifications in the West or simply have developed this form of castle themselves.145 Furthermore, the trebuchet itself was not a new weapon, for this technology had already reached the eastern Mediterranean from China during the sixth century.146 Hence the arms race of the late twelfth and thirteenth centuries was not caused by a rash of new scientific discoveries, but rather by an increase in wealth of states like Egypt and France, which began to enjoy the greater liquidity and administrative skill needed to construct stronger castles, build more powerful trebuchets and muster larger field armies. The Byzantine empire could only build massive fortifications such as the double land walls of Constantinople because of its enormous wealth and sophisticated bureaucracy, indicating that it was lack of money rather than knowledge which prevented other states from emulating such feats until the later middle ages. The same probably applied to siege weapons, for recently one scholar writing of the eleventh century has noted that 'the principles of siege-engines seem to have been well-known, but actually building them was difficult and subject to frequent failure. Improvements in siege-techniques seem to have sprung from better general organisation rather than advances in technology'.147

3

Fortifications and External Security The key defensive role of the major crusader strongpoints which have been described was to resist full-scale Muslim invasions involving thousands or even tens of thousands of troops. Between the n6os and 1187 the Franks living within the kingdom of Jerusalem built or strengthened a number of very large castles which were intended to protect the frontiers of the kingdom against a potential invasion by the powerful late twelfth-century ruler of Aleppo and Damascus, Nur al-Din, or by his even stronger successor, Saladin. These castles included Saphet, Belvoir, Montreal, Kerak, La Feve, Maldoim, Darum, Hunin and Chastelez (Jacob's Ford). Unlike many smaller castles built during the relatively peaceful mid twelfth century as centres of trade, agriculture and lordship, these new fortifications served a more overt military purpose, for they were very large and were often situated on the outskirts of the kingdom rather than at its heart. This is perhaps most striking at Belvoir, which was a heavily fortified base constructed from scratch after 1168 and situated on a frontier hill top far from any major population centres.1 It is clear that a similar process took place in the north, for the Templars started to build the powerful new rectangular donjons at Tortosa and Chastel Blanc during the second half of the twelfth century, at a time when Nur al-Din's power was at its greatest. Indeed, the Templars had probably acquired Tortosa in the first place because an assault by Nur al-Din had left the town in ruins. It consequently needed to be properly refortified in case of a similar attack in the future.2 When the expected Muslim offensive came it was led by Saladin rather than Nur al-Din, but the preparations made by the Templars proved very effective nevertheless. The new Templar donjons at Tortosa and Chastel Blanc, along, it seems, with the neighbouring fortress of al-Arimah, held out against Saladin's campaign during the summer of 1188, as did the castles of Crac des Chevaliers and Margat.3 The previous summer, when Saladin had overrun virtually the entire kingdom of Jerusalem, many of the castles built or upgraded by the Franks during the previous twenty years had also proved their worth. Kerak held out until November 1188, whilst Saphet fell the following month. In January 1189 the Muslims captured Belvoir, and in May they

F O R T I F I C A T I O N S AND E X T E R N A L S E C U R I T Y

39

took possession of Montreal. As a consequence the only inland fortress left was Beaufort, whose remote mountain location enabled it to withstand a blockade from May 1189 until April 1190. Although they were all lost eventually, the staunch resistance offered by these fortresses helped to dissipate Saladin's forces and sap his energy, thereby contributing to the ultimate survival of the kingdom beyond the iiSos.4 Apart from al-Arimah, Tortosa, Chastel Blanc, Crac des Chevaliers and Margat, the only other crusader possessions to survive Saladin's onslaught were the heavily fortified cities of Tyre, Tripoli and Antioch.5 This clearly illustrated to the Franks the wisdom of the policy which they had pursued between the n6os and the n8os, and illustrated the importance of maintaining a number of very powerful strongpoints which could resist any similar Muslim invasions in the future. Bearing in mind that the Hospitallers did not acquire Margat until 1186, this suggests that they carried out extensive improvements on this castle and on Crac des Chevaliers in direct response to Saladin's devastating Syrian campaign. Similar concerns clearly caused the Templars to strengthen Tortosa's defences even further in the early thirteenth century, and to construct 'Atlit on the massive scale that they did. The elaborate defences at many other sites, such as Acre, Saphet and Caesarea, must also have been constructed in the full knowledge of what another disaster like Hattin would entail. During the next century the Franks continued to rely on a select group of strongpoints to maintain their position in the east. These sites were far too powerful to fear the arrival of raiders, whose primary aim was to loot the countryside. Thus in 1205 Aleppine forces attacking the area around Margat and Crac des Chevaliers departed without inflicting any damage on either fortress,6 whilst in 1218 Acre successfully withstood another Ayyubid raid.7 In 1247 and 1250 the massive walls of Antioch protected local people against further incursions by Syrian Muslims,810,000 of whom returned to ravage the area around Tripoli in 1252.9 In the 12608 Baybars launched a series of devastating raids both against the county of Tripoli and against Acre and Tyre, which doubtless inflicted a lot of damage on the surrounding countryside but posed no serious threat to these cities.10 Over the decades, there were also a number of failed sieges which illustrated the difficulties the Muslims had in reducing larger Prankish strongholds. In 1262, before he made his more successful conquests a few years later, Baybars was thwarted in an effort to capture Antioch.11 Al-Muazzam's disastrous attempt to capture 'Atlit in 1220 has already been referred to.12 Two years previously, Oliver of Paderborn wrote that an Aleppine army, hoping to divert Prankish resources away from the Fifth Crusade in Egypt, attacked Chastel Blanc and 'destroyed its towers'. This must either have been an exaggeration or only

40

UNKNOWN CRUSADER CASTLES

have referred to the outer defences of the fortress, because the architectural evidence proves that Chastel Blanc's keep predates 1218 and therefore survived the Aleppine assault. Consequently the keep may have held out in the same way that the citadel of Tortosa defied the Muslim invasion force of 1188, even after Saladin had sacked the surrounding town.13 To the north, meanwhile, the fortress of Margat proved so powerful that in 1281 the outnumbered garrison successfully withstood a determined assault by 6000 Muslim besiegers.14 Eleven years earlier Baybars had been forced to call off another attempt to besiege this castle because of the atrocious weather.^ Perhaps the best example of the way in which major Latin strongpoints guaranteed the continued existence of the crusader states took place in 1244. In that year a Khwarizmian force supposedly containing up to 20,000 horsemen swept into Palestine and captured Jerusalem, implying that the city's walls were indeed in a poor state of repair. Prankish efforts to halt the Khwarizmians in the field ended in disaster at the battle of La Forbie, and the Latins were far too outnumbered to prevent them from occupying large areas of the Holy Land. In theory, this could only have been done by constructing a continuous barrier like Hadrian's Wall or the Great Wall of China; but, even if this had been physically possible, tens of thousands of men would have been needed to garrison such a structure. Consequently, for much of 1244 'the Christians only held the fortresses'.16 Inland, these included Saphet, Montfort, Beaufort, Crac des Chevaliers and Chastel Blanc, whilst along the coast the Franks successfully defended all their major strongholds except Ascalon. As a result the Khwarizmians, who were a nomadic people and lacked their own siege equipment, were able to inflict terrible damage on the rural economy but could not make any permanent conquests of their own. This enabled the Franks to wait securely inside their castles until lack of food and shelter, combined with a realisation that further progress would be almost impossible, forced their opponents to retreat.17 These tactics could work equally well against aggressors who were more disciplined and more capable of undertaking a siege that the Khwarizmians. In 1220, for example, al-Muazzam appears to have reached 'Atlit, which lay at the very heart of the kingdom of Jerusalem, without meeting any resistance. Once he got there, al-Muazzam failed to overcome the massive fortifications, whose 4000 Christian defenders inflicted serious casualties on his own troops.18 Consequently the Franks had prevented al-Muazzam from conquering areas around 'Atlit without having to guard lengthy frontiers or defend strategic roadways. During the Third Crusade, Saladin used a similar strategy against Richard I. By ravaging areas east of Jaffa, but at the same time strengthening Jerusalem's defences, he made it impossible for

F O R T I F I C A T I O N S AND E X T E R N A L S E C U R I T Y

41

the crusaders to win back their former capital.19 For his part, Richard knew that it would be suicidal to proceed into a desolate wasteland and besiege a city without adequate food, water or shelter, and so he was obliged to retreat. Thus it was far more important to defend individual strongholds than the countryside which surrounded them, for outlying areas could easily be reoccupied provided that larger castles and cities successfully withstood a temporary invasion.20 It will be seen that in Greece, the Catalans, Franks and Greeks all used exactly the same tactics to withstand hostile incursions.21 By maintaining just a handful of extremely powerful castles, the Franks could also force their opponents to abandon some campaigns before they had even begun. In 1221 a large Muslim army gathered at Horns in the hope of launching an attack against the county of Tripoli which would distract Christian efforts during the Fifth Crusade. However, eventually it was decided to cancel the offensive and march south to Egypt, because the Muslims 'reflected that the castles of the Hospitallers or the Templars could not easily be captured in a short time'.22 Similar worries may explain why the Mongols, having conquered all of Muslim Syria, did not invade the kingdom of Jerusalem in 1260. It has been argued that they did not do so because they hoped to forge an alliance with the Franks against the Egyptians, but it is just as possible that a fear of Prankish military might caused the Mongols to hesitate. If this is the case, then it suggests that the Mongols, despite supposedly numbering 20,000 men, were afraid of entering a region defended by a series of strongholds which would probably have taken them several years to capture. To some extent, the reign of Baybars subsequently proved them right, for although this sultan spent most of the 12605 and early 12705 campaigning against the Franks, even he failed to capture all of their castles and fortified cities.23 Baybars' inability to finish the Franks off is all the more remarkable when one bears in mind that some of his largest field armies contained thousands more troops than anything his Christian opponents could muster. The Latins relied on their strongest fortifications to make up for their lack of troops. Indeed, many Prankish castles were deliberately constructed in such a way that they could be defended by relatively few men against much bigger besieging armies. According to one chronicler, the Tower of David was so powerful that it only required a garrison of fifteen to twenty men.24 The anonymous author of De constructione castri Saphet considered it a great advantage that far more troops would be needed to attack this fortress than to defend it. This comment seems to have been borne out by subsequent events, for Saphet reputedly had a wartime garrison of 2200, but Baybars may well have brought more than 12,000 men with him when he captured this fortress in 1266.25 Although exact totals are hard to calculate, it is likely

42

UNKNOWN CRUSADER CASTLES

that other Prankish garrisons were just as outnumbered, for in 1281 the 6000 Muslim troops besieging Margat were driven off by a mere 600 Hospitaller horsemen.26 Similarly, when Baybars occupied the citadel of Antioch in 1268, he found that 'eight thousand fighting men' had taken shelter inside, which suggests that unless thousands of other defenders had already been captured elsewhere in the city, then the sultan's army could have outnumbered the entire Christian garrison.27 These figures indicate why the Franks rarely faced their opponents in open battle and preferred to concentrate their meagre forces inside strongholds rather than trying to defend the countryside. Similar problems help to explain why the Latins seem to have found it far harder to maintain and garrison sprawling city walls than compact castles and citadels. We have seen, for example, that the Byzantine ramparts around Antioch were more than eighteen kilometres long and covered an entire mountain side. In the thirteenth century, when the city's population is likely to have been much lower than it had been in Roman and Byzantine times, the princes of Antioch must have found it difficult to come up with enough troops to guard every tower, gate and postern.28 Indeed, descriptions of the fall of the city to Baybars in 1268 imply that the defenders were quite simply overwhelmed by the sheer amount of Muslim attackers, who were so numerous that they 'surrounded the whole city and the citadel' before scaling the walls with relative ease.29 Similarly, it has been suggested that the walls of Jerusalem were in a poor state of repair during the 12305 and 12405, and this appears to be borne out by the historical evidence, which indicates that only the citadel was kept in good condition. Thus in 1229 15,000 Muslim peasants broke into the city with ease but were unable to take the stronghold, which provided shelter for the beleaguered Christians until a relieving force arrived from Acre and drove the Muslims back into the hills.3» Around this time a Muslim source wrote that, having recently regained the city, 'the Franks constructed at Jerusalem a citadel which incorporated amongst its towers that of David'.31 This was confirmed ten years later by a Latin chronicler who commented that Jerusalem 'had not been fortified strongly except the keep ... which was called the Tower of David'.32 Indeed, in 1239 the citadel withstood another failed assault by al-Salih of Egypt,33 before eventually falling to Malik an-Nasir Dawud, ruler of Kerak, after a blockade lasting more than three weeks.34 During these two sieges, the Tower of David is reported to have been defended by very small garrisons numbering first twenty knights and then one knight plus seventy infantry. These campaigns also add to the impression that Jerusalem's city walls were no longer complete, and that the Christians had concentrated all their meagre

F O R T I F I C A T I O N S AND E X T E R N A L SECURITY

43

resources on making sure that the Tower of David was capable of withstanding a siege. If this was the case, then Dawud's campaign of 1239 may have proved decisive, for although the Christians subsequently regained Jerusalem by treaty, the Muslims did not return the city until they had demolished parts of the citadel and had even managed to shift some of the gigantic Herodian stones at the base of the Tower of David. It seems unlikely that the Franks had the time or the resources to rebuild this structure properly before 1244, leaving them wholly at the mercy of the dreaded Khwarizmians.35 In a sense, therefore, the Latins only lost Jerusalem once they had lost its citadel. Likewise, the new strongholds constructed at Ascalon and Tiberias during the early 12405 may have represented other, less successful, attempts to defend and even recolonise settlements without having to build extensive new city walls straight away.36 The relative ease with which compact citadels of this kind could be defended compared with long, sprawling circuit walls can be illustrated further by returning to the fall of Antioch in 1268. Whereas Baybars managed to storm the city's ramparts in a mere three days, the citadel atop Mount Silpius only surrendered when the thousands of people who had taken shelter there found that they lacked the supplies to survive a protracted siege.37 On the other hand, the fate of Antioch proved that even the most impregnable citadels could not provide adequate protection for settlements inhabited by thousands of people. The only way to do so and to succeed in repopulating places like Tiberias was to build vast urban fortifications. But the fact that Antioch was lost 'because there was not in it a force sufficient for its defence' proved that the Franks were not always capable of maintaining existing walls, let alone constructing new ones.38 This helps to explain why they never returned to Ascalon, Tiberias or Jerusalem after the mid 12408. By contrast, the number of people living in the three great coastal settlements of Tyre, Tripoli and Acre probably rose in the thirteenth century and the latter city is said to have contained 40,000 inhabitants by i2pi.39 This was partly because the gradual loss of inland territories led to a concentration of Christians along the coast, but it also reflected the great political and economic importance of these cities. In military terms it also meant that adequate troops and resources could be found to guard and maintain their walls properly, and this may explain why their citadels, whose strategic insignificance has already been referred to, are hardly ever mentioned in the contemporary sources.40 However, it is important to remember that even at these, the wealthiest and most densely populated of all the crusader cities, the task of defending the walls had to be shared out. The Hospitallers, Templars and Teutonic Knights were especially important

44

U N K N O W N CRUSADER CASTLES

in this respect, and their contribution will be discussed in more detail below.41 At Acre the much smaller Order of St Lazarus also guarded the northern tip of Montmusard's fortifications,42 whilst medieval maps and descriptions of the city indicate that other sections of the ramparts were entrusted to the Italian city states or to the patriarch of Jerusalem.43 Under these circumstances it is hardly surprising that at smaller, poorer or less well-populated coastal settlements the local inhabitants, like the citizens of Jerusalem, depended far more on their citadels than their town walls to shield them against a major invasion force. In 1265 Baybars easily sacked the walled bourg which had sprung up around 'Atlit but did not even attempt to storm the castle itself.44 We have also seen that in 1188 Saladin captured the town but not the citadel of Tortosa, and that in 1232 Lombard troops failed to take the castle of Beirut even thought they managed to storm the surrounding city in the space of a single night.45 These events imply that the urban defences of many smaller settlements were not just relatively modest in scale, but also lacked enough skilled garrison troops to defend them against a concerted assault by a major attacking army. Apart from being easier to defend than city walls, many citadels situated along the coast also enjoyed the added advantage of being accessible for seaborne reinforcements, and could therefore benefit from the considerable naval superiority of the Latins. This point can be illustrated by returning to the successful defence of 'Atlit against al-Muazzam in 1220. The core garrison of this fortress was gradually strengthened by the arrival of Prankish contingents from Acre and Cyprus, who 'brought a great supply of soldiers and funds' with them. Indeed, news that more reinforcements were being prepared by the lords of Gibelet and Tripoli contributed to the decision made by al-Muazzam, who had no means of stopping ships from arriving, to call off the siege. This suggests that 'Atlit would have been blockaded for much longer, and perhaps even forced to surrender, if it had not received a steady flow of seaborne assistance.46 No doubt Saladin would also have captured the newly rebuilt Prankish citadel of Jaffa, which he besieged in July 1192, if Richard I had not arrived to relieve its garrison with a hastily organised fleet from Acre. Like al-Muazzam, Saladin did not have the necessary naval strength to prevent Richard from mounting a rescue operation of this kind, and so he was obliged to retreat.47 But even if Saladin had had enough ships with him, he might still have encountered problems, for it could be extremely difficult to maintain an effective naval blockade over any length of time. In 1232, for example, John of Ibelin managed to send reinforcements into the castle of Beirut by ordering his troops to swim past the Lombard ships guarding the harbour under cover of darkness. Eventually, John's son also slipped through the blockade

F O R T I F I C A T I O N S AND E X T E R N A L S E C U R I T Y

45

aboard a small boat carrying one hundred further troops, and soon after the Lombards raised the siege and withdrew to Tyre.48 Help could also be sent to cities or castles which were under threat, even if they were not actually under siege. The arrival of 130 Cypriot knights at Acre in 1265 boosted this city's defences just as Baybars was besieging Caesarea and Arsuf a few miles down the coast.49 Many years earlier Saladin had also been deterred from attacking Tripoli after a contingent of Sicilian knights arrived to help the city. Indeed, both Tyre and Tripoli presented Saladin with a dilemma, for the longer he left them in Prankish hands, the more reinforcements they would receive from the west; yet he was reluctant to besiege these strongpoints as long as there were other, less powerful castles still to be taken elsewhere.50 Their naval power therefore enabled the Latins to protect coastal fortifications which would otherwise have been lost to the Muslims, or in the case of Beirut, a rival western faction. Even when a city or fortress could no longer be defended against land attackers, lives could still be saved if its garrison had some way of escaping by sea. In 1218 a Genoese fleet sent to relieve Caesarea, which was being besieged by a large Damascene army, could do nothing to save the city but did at least manage to rescue its defenders and take them to Acre.51 In 1291 the garrison of Sidon was also able to get away from Mamluk troops besieging the town by first withdrawing into the sea castle, and then sailing to Cyprus once any further resistance became pointless.52 Although the citadel of Tiberias was of course situated many miles from the Mediterranean, its location next to Lake Tiberias gave its occupants the same option, for in 1187 Raymond of Tripoli, lord of the city, advised his wife and her companions to take to their boats if Saladin, who was at that time campaigning close by, managed to break into the castle. It does not appear that this was done, for shortly afterwards, when the Christians had been defeated at Hattin, Saladin occupied Tiberias and granted Raymond's wife safe passage to Tripoli.53 At other times access to the sea really did make the difference between life and death; a point which can be illustrated by looking at the contrasting fate of Haifa and Arsuf, both taken by Baybars in 1265. Whereas almost all the inhabitants of Haifa managed to flee in boats just as the Muslims broke into the town, the Hospitallers defending Arsuf were prevented from making contact with Prankish ships trying to assist them, and were consequently all killed or captured.54 Such incidents confirm that proximity to the sea became an important element in the defensive strategy of all coastal cities and fortresses. One of the principal reasons why Richard of Cornwall decided to refortify Ascalon in the 12405 was that it could be reached by sea if it ever came under attack. Similarly, the site of 'Atlit was regarded as an ideal place to build a castle

46

UNKNOWN CRUSADER CASTLES

because it had 'a naturally good harbour'.55 This was not the only stronghold designed in such a way that retreating garrisons could virtually step straight into waiting vessels. At Sidon there was a small jetty right next to the sea castle, which was presumably used by the city's defenders when they fled to Cyprus in i29i.56 Further south, the citadel of Caesarea acted as the southern breakwater of the harbour, and must have been easy for the Genoese relieving force to reach in I2i8.57 The entrances to many larger crusader ports were also protected by flanking towers, usually with a chain between them which could be raised to prevent hostile ships from gaining access. Such structures existed at Beirut, and it is interesting to speculate whether they played any role in the Lombard siege of 1231-32.58 Similar defences also existed at Tyre and at Acre, where the Venetians and the Genoese periodically fought for control over the so-called Tower of the Flies guarding the harbour entrance.59 A combination of naval strength and powerful fortifications enabled the Latins to hold on to their coastal territories until the very end. But on land the situation was very different, for here it was the Muslims who were at an advantage, and sometimes outnumbered their Prankish opponents by as much as ten to one. This discrepancy prevented the Franks from relieving inland strongpoints in the same way as they had done at 'Atlit in 1220. Admittedly, there were incidents of field armies rather than seaborne forces rescuing besieged castles or cities. In 1262 Armenian-led troops prevented Baybars from capturing Antioch, and in 1229 Prankish knights from Acre drove Muslim attackers out of Jerusalem. However, it is significant that the relief of Antioch involved a large Mongol contingent rather than a purely Christian force, and that the Muslims expelled from Jerusalem were rebels rather than proper Ayyubid troops.60 In fact there is no evidence that Baybars ever had to abandon a siege because a purely Latin field army turned up and forced him to retreat. The Franks quite simply lacked the resources to mount a major expedition of this kind into the interior, where a Christian force of two or three thousand could easily be ambushed or even annihilated by a far larger Muslim army. Their overwhelming superiority on land enabled the Muslims, and especially the Mamluk sultans of the late thirteenth century, to besiege Prankish strongpoints of the interior almost at will. Although the presence of twentyone Muslim galleys at the siege of Ascalon in 1247 suggests that they did sometimes have enough ships at their disposal to blockade coastal sites, such incidents appear to have been rare, and Egyptian efforts to create an effective fleet were not very successful.61 But inland there was very little to stop the Muslims from surrounding a strongpoint, thereby preventing the garrison from escaping in the way that the remaining occupants of Sidon

F O R T I F I C A T I O N S AND EXTERNAL SECURITY

47

had done in 1291. As a result, inland sieges could end with appalling massacres and devastating looting sprees. Such atrocities occurred at Saphet and also at Antioch, where virtually the entire population was either killed or enslaved, and centuries of Byzantine culture were wiped out in a matter of hours.62 These incidents explain why Prankish control over inland regions crumbled decades earlier, and why the vast majority of fortifications built or repaired during the thirteenth century were located along the coast. In 1253, for example, Louis IX decided to rebuild Sidon's defences rather than construct a new fortress in the interior, because the local barons advised him that such a place would be too exposed to Muslim attacks without any access to the sea.63 In 1230 Gregory IX expressed similar concerns for the castle of Montfort, because it was located several miles inland and was consequently proving costly to build and difficult to defend properly.64 The fate of the Holy Land was ultimately decided by the fact that the Latins tended to dominate the sea and the Muslims found it easier to control the land. Under such circumstances, the absence of an adequate Christian field army proved decisive, for even the most powerful inland strongpoints could not be expected to hold out indefinitely without any external assistance. As the fall of Antioch plainly illustrated, the loss of a major Christian strongpoint represented a human tragedy as well as a territorial loss, for crusader fortifications were intended to protect people just as much as national frontiers. Indeed, when Baybars finally obtained the surrender of Antioch's citadel, he found that there were many 'women and children' who had tried to seek refuge inside it alongside the actual fighting garrison.65 But before this final catastrophe occurred, we have seen that the walls of Antioch protected the inhabitants of the city against many previous Muslim incursions, including Saladin's invasion of 1188. Similarly, both Acre and Tyre had such massive defences that their citizens had little to fear from the raids which Baybars carried out in the surrounding countryside during the i26os.66 The importance of ensuring that the walls of major settlements were properly maintained and garrisoned can be illustrated further by returning to the Khwarizmian invasion of 1244. Whilst Acre and Tyre do not appear to have been threatened by this event, Jerusalem was sacked by the invaders with the result that 7000 inhabitants were supposedly massacred.67 It has already been suggested that sites like Acre, whose city walls were able to withstand tens of thousands of Muslim troops, were few and far between, and that most smaller towns or bourgs were only protected by a single ditch and rampart. Consequently, the inhabitants of smaller settlements relied on their citadels more than their town walls to protect them from major attackers. Several incidents of this kind have already been referred to, such as the raid made by Baybars against 'Atlit in 1265 which resulted

48

UNKNOWN CRUSADER CASTLES

in the destruction of the town but caused no physical harm to the Christians sheltering inside the actual fortress.68 This point can be illustrated further by taking a closer look at the history of Sidon during the thirteenth century. Following the completion of Sidon's sea castle by members of Frederick II's crusade in 1228, Louis IX sent a contingent of his army to build a second fortress and new urban fortifications there in the summer of 1253. While this work was still going on the Muslims launched a surprise raid on the Franks, which resulted in the death of 2000 Christians, most of whom were killed because there was not enough room for them inside the sea castle. As a result Louis IX decided to supervise the rest of the project in person, so that it was completed before he returned to France the following year.69 Six years after Louis IX's departure, lulian, lord of Sidon, made a rash incursion into territories to the east of Beaufort, which were held by the Mongols at that time. The Mongols responded by launching a devastating counter-attack against Sidon but, thanks to Louis's new defences, Julian was able to hold the Mongols off at the gate just long enough for the Christian population to retreat. Meanwhile the presence of a new land castle to complement the older sea castle ensured that this time there was enough space for every one to find shelter. Consequently, when the Mongols finally broke into the empty town, all they could do was to carry out widespread looting and dismantle the walls, but they made no attempt to attack either citadel. Louis IX's wish that the massacre of 1253 should not be repeated had therefore been fulfilled.70 Apart from people who inhabited an adjacent town or bourg, strongpoints also provided shelter for those who lived in the surrounding countryside. Regardless of whether they were ultimately intending to besiege a Latin strongpoint or merely carry out a swift raid, Muslim forces invariably caused devastation to the Christian territories they passed through. Cattle and slaves were seized, crops destroyed and orchards burnt, partly to feed the army and to reward the soldiers, and partly to inflict economic damage on the Franks. As a result, Christians living in the countryside deliberately settled close to castles so that they could find shelter relatively quickly. The author of De constructione castri Saphet wrote that, once this fortress had been completed, 10,000 people living in 260 villages recolonised an area of central Galilee which had previously been considered too dangerous to inhabit.71 During the reign of Baybars, who frequently carried out large-scale raids in order to destroy crops needed to feed Prankish garrisons, we often hear of hundreds or even thousands of non-combatants taking shelter inside fortresses. When the Mamluks captured Saphet (1266), Beaufort (1268), Akkar (1271) and Chastel Blanc (1271), they found many such people who had arrived from an adjacent bourg or from neighbouring villages.72

F O R T I F I C A T I O N S AND EXTERNAL SECURITY

49

The thousands of people living in the countryside could only hope to reach the safety of sites like Saphet if they were given adequate warning of an imminent Muslim incursion. Consequently, one consideration for castle builders was to ensure that Prankish garrisons enjoyed good views over the surrounding area, so that they could spot approaching attackers and warn others by means of messengers, fire signals or carrier pigeons. There is considerable evidence that the Franks learnt this latter method from their Muslim opponents and relied on it throughout the twelfth and thirteenth centuries.73 Both the Hospitallers and the Templars kept carrier pigeons at Acre,74 and in 1217 the Hospitallers used this form of communication to inform the garrison at Crac des Chevaliers that the preacher Jacques de Vitry wanted to pay a visit. An armed guard could therefore be arranged to escort this important guest from Tripoli.75 Visibility could also have a profound effect on location. Belvoir, for example, was deliberately placed on a site which overlooked the crossing points of the River Jordan below Lake Tiberias, as we would expect bearing in mind that the fortress was intended to strengthen the eastern frontier of the kingdom of Jerusalem.76 Similarly, Cave de Tyron's location in the side of a mountain overlooking Sidon indicates that the sole function of this remote outpost was to keep an eye on the coastal roads linking Tyre with Beirut.77 Other castles were situated in such a way that their garrisons could see, and perhaps communicate with, numerous surrounding strongholds. The defenders of Beaufort could spot any Muslim troops moving along the Beqa valley toward the coast, and could keep watch over Subeibe, which was the closest Damascene fortress to the kingdom of Jerusalem. Subeibe was situated a mere twenty-one kilometres to the south east, and was considered a major threat by the Franks, who may have tried to retake the neighbouring town of Banyas during the 12505 in an attempt to re-establish their authority east of the River Jordan.78 In addition, Beaufort was intervisible with the friendly castles of Toron (Tibnin), which lay to the west along the main route to Tyre, and Chateauneuf (Hunin), situated eighteen kilometres to the south.79 Beaufort held so much strategic importance that it was one of the few castles in the region not demolished by Saladin or Baybars.so Whilst it was clearly advantageous for sites like Cave de Tyron and Beaufort to enjoy good views over the surrounding countryside, the notion that larger groups of crusader castles were deliberately placed so that they could communicate with each other has probably been exaggerated in the past. Around the plain of Akkar, for example, the defenders of Crac des Chevaliers, Tortosa, Chastel Blanc, Akkar, Chastel Rouge and al-Arimah, along with numerous smaller forts and towers, could all theoretically send each other fire signals

5O

U N K N O W N CRUSADER CASTLES

either directly or via other castles.81 In the 19505 the scholars R. Fedden and J. Thomson wrote that groups of fortifications like this 'were sited so that intercommunication should be possible over an extraordinarily wide area', implying that these strongpoints formed a kind of early warning system against possible Muslim attacks from the direction of Hama and Horns.82 Yet the possibility that the Latins deliberately created such a network seems remote bearing in mind that some of these fortifications, such as Crac des Chevaliers, had originally been constructed before the First Crusade when the frontiers of the county of Tripoli were still unknown.83 Consequently it is likely that the intervisibility enjoyed by most castles in this region was a coincidence caused by the fact that Crac des Chevaliers, Akkar, Chastel Blanc and al-Arimah, like most medieval strongpoints, were built on hill tops. Similar observations can be made about numerous smaller fortifications in the county of Tripoli. In the late nineteenth and early twentieth centuries Rey and Deschamps argued that local structures such as Tuhklah, a sturdy square tower near Chastel Blanc, were primarily intended as lookout posts. Yet more recently it has been pointed out that Tukhlah was basically a residential and agricultural complex which stood in the heart of a village and 'would have been very badly placed as a watchtower since it is overlooked by a much higher hill about half a kilometre to the east'.84 In later sections it will be shown that the same point applies to virtually all other fortifications of this type in the Holy Land, Greece and Cilician Armenia.85 Hence theories that Christian castle builders in the eastern Mediterranean deliberately created vast networks of intervisible strongpoints seem to stem from a tendency to look at maps rather than the actual terrain, to ignore the eclectic history of most fortifications, and to overlook the more day-to-day domestic functions of sites like Tukhlah. Returning to the role of fortifications as a means of sheltering civilians, during the final decades of Prankish rule the ability of sites like Beaufort to spot Muslim attackers or send fire signals to other castles made little difference anyway. As Latin field armies found themselves overwhelmingly outnumbered and could not come to the rescue of inland strongholds, sites like Beaufort became increasingly isolated and could be besieged by the Mamluks almost at will. Once an inland castle or city was broken into, people sheltering there were consequently doomed because they had nowhere to run and had no Prankish relieving forces to appeal to. Indeed, when the Muslims began to swarm into Antioch, many of its inhabitants found themselves trapped by their own walls, because Baybars had deliberately sealed off all the gates in order to prevent any loot from being carried away86 By contrast, superior Latin naval power meant that civilians sheltering

F O R T I F I C A T I O N S AND EXTERNAL S E C U R I T Y

51

behind coastal fortifications at least had some hope of escaping by sea. In 1291 those citizens of Acre who were fortunate enough to have access to ships were able to flee to Cyprus. The actions of others who were not so lucky and found themselves left behind shed further light on the strategic limitations of citadels situated inland. Contemporary sources do not mention anyone seeking shelter inside Acre's citadel, which would in a sense have become more of a trap than a refuge once the Muslims controlled the streets around it. Instead about 10,000 Christians made their way to the headquarters of the Templars, which stood at the water's edge in the south-west corner of the city. This structure had very strong walls and towers, as well as a postern giving access to the sea, and it is obvious that the Franks rated their chances of survival better by sheltering here than in a building which offered no obvious means of escape. In theory, the Temple could therefore have been used to organise a seaborne evacuation, if the Latins had not already been so utterly defeated that its garrison was finally overwhelmed after a blockade lasting a further ten days.87 During the fall of Tripoli in 1289 a similar set of circumstances led to a high number of casualties. Tripoli's citadel was located on a hill opposite the actual city, making it impossible for anyone to reach it once the Muslims had begun their assault. As a result, the Franks had nowhere to hide after their urban defences had been breached, and hundreds of citizens who sought refuge on a small island in the harbour were ruthlessly massacred. If this island had been fortified, these people could have been evacuated in the same way that the defenders of Sidon were two years later. Sites which had a powerful inner fortress with direct access to the sea were strategically superior to settlements protected by strong curtain walls and a relatively weak or isolated citadel.88 It is important to address the question of whether Latin strongpoints were constructed according to a centrally coordinated defensive strategy. In the past it was argued that during the twelfth century, when the monarchy was stronger, such a strategy did exist. Deschamps, Rey and others believed that once the First Crusade had come to an end 'the Franks proceeded to organise the various parts of the countryside' as though they were all of one accord, and built their fortifications as part of a national scheme of defence.89 In his work on crusader castles, Rey in particular spoke of 'lines' of strongholds, starting with coastal sites like Acre and emanating outwards in defensive rings which protected the kingdom of Jerusalem from external enemies.90 This theory of a vast network of fortifications specifically built to block all entry points into crusader lands was largely demolished by Smail. Smail pointed out that Prankish territories had been conquered in a series of campaigns usually undertaken by individual barons who were often driven by personal greed. As a result, castles built or captured by the

52

UNKNOWN CRUSADER CASTLES

crusaders were intended for local rather than national defence, and had little strategic value beyond their immediate surroundings. Moreover, the regular incursions made by Nur al-Din, Saladin and others into Christian territories showed that frontier castles stood little chance of blocking Muslim attacks anyway.91 At first glance, it would appear that these observations applied to the thirteenth century just as much, and perhaps even more so because of the frequent lack of central authority within the crusader states. It is clear that external pressure and internal squabbling sometimes made it virtually impossible for the Franks to pursue a common military strategy toward their Muslim enemies. The castle of Beirut, for example, largely remained in Christian hands for as long as it did because its lords negotiated a series of peace treaties with the Muslims, irrespective of what their Prankish neighbours were up to.92 Similarly, in 1266 the Templars made a treaty with Baybars whereby the latter recognised the Order's possession of Tortosa and Chastel Blanc in exchange for half of Jabala, a coastal settlement to the north which the Templars shared equally with the Hospitallers. This deal was clearly made without reference to the Hospitallers, who were so furious that their troops ended up fighting Muslim forces sent by the sultan to occupy his new share of the town.93 This general mood of 'every man for himself' has led scholars to argue that Smail's view of twelfth-century fortifications still holds true for the period after 1187. It has been written that 'few of the Latin strongpoints were of any genuine strategic value', apart from Margat, Crac des Chevaliers and to some extent 'Atlit.94 On a smaller scale, many Latin strongholds such as Goliath were normally too weak to withstand large Muslim invasion forces anyway, and were primarily intended as local centres of defence, farming and lordship. These less overtly military functions of crusader fortifications had hardly been looked at by historians before Smail's time. Although we should probably still reject Key's theory of nationwide castle networks, it would be an oversimplification to say that during the thirteenth century the Franks were incapable of implementing any kind of overall strategy at all, for some larger strongholds do appear to have been intended to interact with others. This was most notable along the southern frontier of the kingdom of Acre, where the Latins were concerned about possible offensives by the Egyptians but also hoped to reconquer Jerusalem or make counter-attacks against Egypt. 'Atlit, for example, was referred to as 'the breastwork of the city of Acre' at the time of its construction.95 This implies that the fortress had not just been built to administer local Templar properties but had deliberately been placed along the coastal route south of Acre. Any Egyptian troops attacking the Prankish capital would have to contend

F O R T I F I C A T I O N S AND E X T E R N A L S E C U R I T Y

53

with an impregnable fortress in their rear whose defenders could disrupt their supplies or even cut off their line of retreat. Likewise, Ascalon became the focus of much crusader activity because it was close to Egypt and Jerusalem rather than because it was particularly valuable in its own right. By building new defences here Richard I, Theobald of Champagne and Richard of Cornwall all hoped to deter the Muslims from attacking Galilee, whilst at the same time creating a potential starting point for Christian incursions toward Gaza or Jerusalem. These factors caused Matthew Paris to describe Ascalon as the 'key' to the kingdom, and help to explain why both Saladin and Baybars were so keen to demolish it.96 Here at least there is some sense that Ascalon and 'Atlit were fortified for the benefit of everyone living in the crusader states, not just for their actual owners. Moving north and east, reference has already been made to Ellenblum's argument that certain very large late twelfth-century castles such as Belvoir were also erected for national rather than local defence, and it has been suggested in this section that this building programme was extended into the thirteenth century at sites like Crac des Chevaliers, Margat, Tortosa, Chastel Blanc and, from the early 12408 onwards, Saphet in Galilee. These castles were intended to protect the crusader states (and not just themselves) against any future invasions like those launched by Saladin in 1187 and 1188. There may well have been more thirteenth-century fortresses than one might expect which were built or enlarged for the defence of the crusader states as a whole, and reflected the common desire of all Latins to preserve themselves against the growing Muslim threat. Indeed, as Muslim armies grew larger and larger, the need for massive and virtually impregnable strongholds such as 'Atlit must have grown, for there is little evidence that the Franks could rely on natural barriers such as mountain ranges to protect them. Although such features obviously made hostile incursions impossible in certain regions, they could always be bypassed relatively quickly and clearly did not deter Muslim conquerors in the long run. This was most apparent in the kingdom of Jerusalem, whose overall size fluctuated wildly between 1099 and 1291, but does not appear to have fared any better when a greater proportion of its frontiers passed through inhospitable areas. There seems little difference between the ease with which Saladin invaded the coastal plain in the twelfth century and that with which Baybars overran the same area eighty years later. Ultimately, it was crusader fortifications rather than mountains, deserts or rivers which had to be overcome before the Franks could be driven out of the Holy Land.

4

Fortifications and Internal Security Strongpoints the size of Acre, 'Atlit or Saphet were always in the minority. Most crusader fortifications were much smaller. Structures of this kind varied in size and design from isolated towers or minor courtyard buildings, such as Tuklah and the Castle of Roger the Lombard, to medium-sized fortifications like Qaqun or Goliath. In addition, some urban defences, such as the walls around the bourg of 'Atlit, were considerably weaker than the ramparts of sites like Caesarea, Tripoli and Acre.1 From studying the history of these smaller fortifications it is immediately apparent that they were not intended to withstand the kind of armies which attacked the Holy Land under Saladin in 1187/88, the Khwarizmians in 1244 or Baybars in 1271. This point can be illustrated further by looking at the known history of Goliath, a castrum-type castle held by the Hospitallers a few miles north of Tripoli.2 Although it is not specifically mentioned in the contemporary sources, it seems unlikely that this structure could have withstood Saladin when he ravaged the plain of Akkar in n88.3 However, it must have been reoccupied by the Hospitallers soon afterwards, for in 1207/8 al-Adil stormed the castle before demolishing most of its defences.4 Goliath had still not recovered from this blow when Willbrand of Oldenburg saw it four years later, and the fortress may well have suffered further damage as a result of incidents such as the Khwarizmian attack on the county of Tripoli in 1244.5 The bewildering mixture of masonry types still visible in its present remains confirms that it was rebuilt a second time between 1211 and 1266, when Baybars overran it and again left it in ruins.6 Even after this disaster the Hospitallers may well have returned to the site, or at least cultivated the lands around it, for this region was not lost conclusively until Baybars conquered the entire plain of Akkar in 12717 Numerous structures with a similar history were dotted across the countryside near the major Latin cities of Acre, Tyre and Tripoli. Whilst carrying out a raid against Tripoli in 1268, for example, Baybars captured a tower near the city 'in which a number of Franks had held out against him and these were beheaded'.8 On the outskirts of Tyre he may also have taken the tower known as La Tor de 1'Opital in 1266.9 Near Acre the fortified mills of Doc and Recordane, which were owned by the Templars

F O R T I F I C A T I O N S AND I N T E R N A L S E C U R I T Y

55

and the Hospitallers respectively, also had a chequered history in the thirteenth century. Both were destroyed during a Muslim raid on the area in 1253. They subsequently suffered farther damage in 1263, when Doc was demolished, and again in 1267, when the same fate befell neighbouring Recordane.10 These events, as well as the substantial remains of Recordane which still survive today, suggest that the Military Orders were able to reconstruct both mills again and again right up until the fall of Acre itself in 1291.n A number of important conclusions can be drawn from this brief historical sketch of sites like Goliath. First, it is clear that, even in the unstable climate of thirteenth-century Syria, major offensives of the kind launched by Saladin and Baybars only occurred every few decades. With the benefit of hindsight we know that from 1266 onwards Baybars virtually destroyed the entire county of Tripoli, and that during the late 12805 his successor Kalavun quickly wiped out what was left. By focusing on the famous victories of Saladin, Baybars and Kalavun we might therefore be tempted to regard the constant rebuilding of Goliath as a lost cause, but this attitude overlooks the fact that this region enjoyed relative peace and security during most of the early thirteenth century. Indeed, until the 12408 at least it was usually Tripoli's Muslim neighbours who felt threatened by the Franks rather than the other way round.12 A similar situation prevailed in the south, for until the battle of La Forbie (1244) the kingdom of Jerusalem was able to maintain or even expand its borders through a series of negotiated settlements with the Muslims of Egypt or Damascus (or both). The Khwarizmian and Egyptian conquests in southern Palestine between 1244 and 1247 were the first Muslim campaign of this kind since the fighting with Saladin came to an end in 1192. Even after these setbacks some areas were relatively unaffected by warfare until Baybars began to pick off Christian coastal settlements in the 12605.13 At 'Atlit, for example, the bourg must have suffered severe damage during al-Muazzam's failed siege of 1220 (if it already existed then) and again during Baybars' raid of 1265, but there is no record of any other attacks on this settlement during the entire thirteenth century, suggesting that its inhabitants lived in peace for forty-five years between these two dates.14 After 1265 archaeological evidence showing that key structures such as the church were left incomplete or not rebuilt suggests that the town was at least partially abandoned; a clear illustration of the escalating warfare and greater suffering which affected the area during the last thirty years of the crusader period.15 For certain periods and certain places, the same argument can therefore be applied to the thirteenth century as that which has been applied to the

56

UNKNOWN CRUSADER

CASTLES

period before 1188 by the scholar Ronnie Ellenblum. According to Ellenblum, there were long spells between the 11205 and n6os when extensive parts of the Kingdom of Jerusalem were almost untouched by external aggressors, yet this did not deter the Franks from building or repairing many smaller castles in just the same way as they later reconstructed Goliath. Indeed, more medium-sized castles, towers and fortified farmhouses were erected during this stable period than at any other time.16 Closely connected to this point is Ellenblum's argument that far greater numbers of Prankish settlers lived in the countryside than has previously been assumed. In the past crusader historians, relying on sources which perhaps inevitably concentrate on military crises and incidents of warfare, have tended to argue that the few westerners who remained in the Holy Land after the First Crusade lived inside the larger castles and walled cities. Here they were safe from the threat of Muslim incursions, and they rarely visited their rural estates, whose day-to-day running was left in the hands of native representatives. But by using a combination of written and archaeological evidence Ellenblum has shown this to be an oversimplification, for there were many small rural structures of a fortified, semi-fortified or even purely domestic nature which were clearly inhabited by the Franks. Far from being removed from each other, city dwellers and the rural population were actually closely connected, for in a society which put so much value on land holding it is inconceivable that westerners hardly ever strayed into the countryside.1? These observations help us to understand the defensive functions of smaller fortifications in the Latin East. The Franks knew that Goliath or the walls of 'Atlit's bourg were incapable of withstanding a concerted attack by thousands of Muslim troops, but during the many years or decades when a major incursion of this kind did not occur such places still served a number of functions as centres of farming, trade and administration. Castles, towers and enclosures could be used for accommodation, storage, food production or simply as a means of displaying a powerful lord's wealth.18 Furthermore, the role of fortifications as a means of asserting ownership and exploiting agricultural land explains why the Franks actually built more castles in peacetime. In this sense the Holy Land was no different from western Europe, for, as one scholar has noted recently, the nobility always 'established castles to guard their proprietorial rights'.19 Archaeologically speaking this was most clearly reflected in the countless Prankish towers erected in the crusader states, often very similar to other towers built at this time by lesser nobles in France and England.20 Stone structures of this kind were expensive for minor lords to build and maintain, but their presence actually attests to a period of relative prosperity rather than one of constant warfare, for there would have been no time or money to

F O R T I F I C A T I O N S AND I N T E R N A L S E C U R I T Y

57

erect so many towers if the surrounding countryside was always being devastated by Muslim incursions. The same argument can be applied to the Prankish states in Greece.21 The construction of minor crusader strongpoints had less to do with the massive campaigns of opponents like Saladin or Baybars than it did with the insecure nature of medieval society in general. The Holy Land differed from France or England in that most of the population was Muslim or at least non-catholic, but otherwise it suffered from the same universal problems of crime and theft which affected much of thirteenth-century Europe.22 Landowners needed to protect their property against small-scale raids and common criminals. Although some acts of violence were instigated by Muslims for religious or political reasons, others were committed by rival Prankish lords or simply motivated by selfish greed. Some idea of the general insecurity which existed at all levels of society can be gained from a clause in the Rule of the Templars dedicated to the dangers of travelling unescorted within the kingdom of Jerusalem. It relates how two brothers in the vicinity of Acre 'found Saracens who attacked them and killed one of the brothers and led away his horse; the other was badly wounded'.23 Similarly, when Joinville was given the task of escorting the French queen from Acre to Tyre, during Louis IX's crusade, it was considered safer to travel by night in order to avoid such attacks. Joinville observed that this mission was particularly dangerous because they had to stop twice along the way to feed the royal children.24 Numerous other incidents, such as the need for an armed guard to accompany Jacques de Vitry while he was preaching the Fifth Crusade,25 and the presence of a tower near 'Atlit 'because of bandits who threatened strangers ascending to Jerusalem', also suggest that internal security within the crusader states was poor.26 Another threat to social stability which occasionally affected most medieval states, but was potentially given an extra dimension in the Holy Land for religious reasons, was the possibility of an uprising by the peasantry. During the first half of this century, French scholars in particular argued that Muslim peasants living under Prankish rule were well treated and often fared better than they had done under the authority of their co-religionists.27 This view has since been criticised by historians like Mayer, who have suggested that at times the Muslims faced a very high burden of taxation. Arguments of this kind have tended to create an image of the Holy Land as a place inhabited by a large and resentful Muslim population ruled over by a small number of Franks who were reluctant to leave their walled cities because they feared local Muslim rebels just as much as external Muslim invaders.28 In recent years this is another topic which has been given an added

58

U N K N O W N C R U S A D E R CASTLES

dimension by the findings of Ronnie Ellenblum, who argues that the Muslim population was not actually as large as historians have previously assumed. Ellenblum gives several reasons for this. First, he argues that the massacres committed by crusaders during or shortly after the First Crusade meant that many Muslims had either been killed or had simply fled. Secondly, he points out that large parts of the Holy Land were inhabited by native Christians, and that they often actually outnumbered the remaining Muslims, who have themselves tended to be lumped into one indiscriminate Islamic grouping with little regard to regional, cultural or religious variations. Thirdly, Ellenblum's conclusion that historians have underestimated the number of Franks settled in the countryside enables him to show that the overall Latin population of the Holy Land, whilst still smaller than the combined number of Muslims and native Christians, cannot have constituted less than 25 per cent of the entire population even by the standards of older scholars.29 If we combine these findings with the arguments of Mayer and others, we are left with a picture of a Muslim population which was both smaller and more harshly treated than scholars assumed in the nineteenth and early twentieth centuries. This picture tends to fit the pattern of Muslim uprisings in the crusader states from 1187 onwards, for such rebellions did break out from time to time, implying that the peasantry felt oppressed by their Latin overlords, but they were rare, suggesting that the notion of a very large and threatening Muslim population has indeed been exaggerated in the past. Indeed, the Muslims normally only felt confident enough to rebel during periods of Latin weakness or military defeat. Thus uprisings occurred in 1187 and 1188, following the battle of Hatting In 1266 a large Prankish army campaigning near Tiberias lost around five hundred troops in a Muslim ambush, subsequently being turned upon by local Muslim peasants, who harassed the defeated Latins as they retreated back to Acre. At this time Prankish authority over the rural population throughout Galilee was waning as Baybars gradually advanced from the east and south.31 Even during the first half of the thirteenth century it was clearly the case that the more isolated a Prankish outpost became, the more likely its inhabitants were to face the wrath of local Muslims. We have seen that in 1229 Muslims broke into and ransacked Jerusalem, at a time when the city had only just been regained by the Christians and was probably inhabited by very few Latins living behind dilapidated or incomplete walls. This attack must have been prompted by a desire for booty, but it probably also reflected a widespread feeling that Jerusalem should not have been handed over to the Christians by al-Kamil.32 The value of castles as a means of dealing with this type of threat is illustrated by the outcome of the 1229 attack on Jerusalem. According to a

FORTIFICATIONS

AND I N T E R N A L SECURITY

59

western source this incident involved no less than 15,000 Muslim peasants, but it is unlikely that they possessed the organisation or equipment needed to besiege any well-defended strongpoints. As a result, the Prankish defenders were able to retreat into the Tower of David and wait until a relieving force of knights from Acre turned up and drove the Muslims out of the city.33 On a smaller scale, the sheer proliferation of minor fortifications throughout the crusader states helps to explain why uprisings were relatively rare, for even if the local Muslims wished to rebel, they must have understood that their chances of inflicting permanent damage or heavy loss of life on the Franks were remote. An isolated tower might not have had much chance against Saladin but could easily have resisted small groups of local Muslim troublemakers. Similarly, robbers, bandits and localised raiders had neither the will nor the means to launch attacks against Latin fortifications. By constructing towers, fortified farmsteads or small castles the Franks could protect themselves, their cattle and their movable property against such people, particularly at night. For years or even decades at a time, the vast majority of crusader strongholds were simply intended to maintain law and order, defend proprietorial rights and safeguard economic activities in much the same way that fortifications did throughout late medieval Europe. This makes the constant rebuilding of sites like Goliath much more logical than it initially appears. Whilst fortifications could increase security in the face of rebels or attackers, it could also diminish the amount of control which lords enjoyed over their own Latin vassals. The fact that crusader armies tended to be small, whilst at the same time Muslim pressure made it necessary for certain Latin strongholds such as Crac des Chevaliers to be extremely large, could make it difficult for a ruler to capture the major stronghold of a vassal who chose to rebel. In the south this dilemma is illustrated by Frederick IFs largely unsuccessful attempts to impose his authority over the kingdom of Jerusalem in the late 12205. In 1229 Frederick marched south from Acre to 'Atlit and ordered the garrison to hand it over. When the Templars, who, as we have seen, were opposed to Frederick's crusade because of his arguments with the papacy and status as an excommunicate, refused to open their gates, the emperor saw that he did not have enough troops to take 'Atlit by force and was obliged to retreat. Later, Frederick also tried to capture the Templar headquarters in Acre, but could not do so and eventually abandoned the city altogether.34 Indeed, although it was also a useful stepping stone in the drive to regain Jerusalem, Frederick's construction of a new citadel at Jaffa not long afterwards may have reflected a realisation that he was better off building his own castle rather than trying to occupy somebody else's, if he wanted to secure a strong base in the region.35 Such tactics did

ÓO

UNKNOWN CRUSADER CASTLES

little to improve political stability, for after Frederick's eventual departure to Europe sporadic clashes continued to take place between his followers and the supporters of a rival Ibelin faction until as late as 1242.36 Another claimant to the throne of Jerusalem whose ambitions were thwarted by the sheer strength of the kingdom's castles was Hugh III of Cyprus. Hugh had been declared king in 1269 but faced a powerful rival in Charles of Anjou, who had bought his claim to the crown from Maria of Antioch in the early 12705. In 1279 Hugh III had to admit defeat in this dispute because the French regiment occupying the citadel of Acre came out in favour of Charles. Hugh appears to have given up hope of controlling the city as long as this Angevin stronghold held out in the heart of his supposed capital, and he therefore withdrew to Cyprus. Even if Hugh had had the resources to besiege this castle, he was also opposed by the Templars, whose bases at Acre, Sidon and 'Atlit were extremely well defended. As a result, Lusignan power was not restored in Acre until 1286, when Hugh Ill's son Henry arranged a carefully negotiated truce with the French regiment. Only then could he regain control over the citadel and have himself crowned king.37 Further north, there were also a number of internal thirteenth-century disputes involving castles. For eighteen years after the death of Bohemond III in 1201 there was sporadic fighting for control of Antioch between the rival candidates Bohemond IV, who was allied to the Templars, and Raymond Roupen, who was supported by the Hospitallers and by his ambitious great uncle, Leon II of Cilician Armenia. The course of this conflict, like the disputes already mentioned at Acre, was heavily influenced by possession of fortifications. For years Raymond Roupen and Leon II struggled to seize Antioch, and when they finally did in 1216, they immediately made sure that the walls and citadel were strongly garrisoned with the help of their Hospitaller allies. In 1219 Bohemond IV was only able to regain control over Antioch with the aid of anti-Armenian factions within the city, and even then he only succeeded in occupying the virtually impregnable citadel because Raymond Roupen, who had taken shelter there, panicked and fled.38 Some years before these events Bohemond IV had to deal with further internal problems involving his vassal Renaud III, lord of Nephin. In 1203 Renaud married the heiress to the lordship of Akkar without Bohemond's permission. Bohemond therefore ordered Renaud to come to Tripoli to explain himself, but when the latter failed to appear Bohemond declared war on him and attacked the castle of Nephin. Renaud responded by launching a raid on Tripoli, but this did not deter Bohemond, who subsequently captured Nephin and Akkar with the help of the Genoese and

F O R T I F I C A T I O N S AND INTERNAL SECURITY

6l

the lord of Gibelet39 During the second half of the thirteenth century these places became involved in a far more serious baronial rebellion, although this time the lords of Nephin were allied with the counts of Tripoli against the Embriaco rulers of Gibelet. In the course of this conflict, the Embriacos and their Templar allies attacked Tripoli no less than three times (1258, 1276, 1282), only being kept at bay by the city's massive defences. In 1276 the Templars also made a failed assault against Nephin, and it was not until 1282 that Bohemond VII finally captured Bertrand II of Gibelet, executed him and occupied his castle.40 At times the apparent weakness of rulers like Frederick II and Bohemond IV also led to internecine struggles between other rival factions. The most famous clashes of this kind probably took place in Acre, where the Military Orders and the Italian city states had all constructed fortified compounds or towers in order to intimidate their rivals and protect their own interests. The Templars' headquarters, an imposing quadrilateral citadel with four corner towers, may have been the strongest such structure, but both the Hospitallers and the Teutonic Knights probably held similar compounds.41 The Genoese, the Pisans and the Venetians also held their own quarters dominated by the kind of large, isolated towers still visible in some Italian towns.42 These structures divided Acre into numerous walled enclosures or distinct quarters, and the type of fighting which sporadically took place there can be illustrated by giving a brief description of the war of St Sabas, fought between Venice and Genoa in the late 12505. This conflict, initially sparked off by a property dispute, eventually led to widespread street fighting between the Venetians, who were aided by the Pisans, and the Genoese, whose principal allies were the Hospitallers. The two sides bombarded each other with crossbows and catapults mounted on top of towers and other tall buildings, and some of these weapons were so powerful that they could hurl rocks weighing 200 kilogrammes over considerable distances. In these circumstances, it is hardly surprising that many houses were destroyed and countless lives lost in the space of just a few months. It is equally understandable that the first action taken by the Venetians after they finally emerged victorious was to raze the Genoese quarter to the ground.43 The war of St Sabas also had wider implications which went beyond the city of Acre itself. Most notably, the Embriaco lords of Gibelet were Genoese, and their long struggle with the rulers of Tripoli began in 1258, when the fighting at Acre was at its height. These clashes may therefore have represented a wider campaign to augment Genoa's power in the east, and even turn Tripoli into a fortified trading post controlled solely by the Genoese. They later established such a base at Famagusta, and after the

62

U N K N O W N CRUSADER CASTLES

death of Bohemond VII in 1287, they finally managed to set up an Embriaco-led commune at Tripoli in the final years before the fall of the city.44 They also continued to clash with the Venetians from time to time, particularly during the mid 12605, by which time they were based at Tyre rather than at Acre. The Italian city states were often prepared to pursue their own aggressive policies, even if this meant that large amounts of men and resources were diverted away from the struggle with the Muslims.45 These events reflect the military and political dilemma confronting Latin settlers living in the Holy Land. Clearly, strongholds could undermine royal power and encourage warfare amongst the nobility. The Embriaco lords of Gibelet were able to defy their overlords on and off for a period of more than twenty years, simply because the rulers of Tripoli proved incapable of capturing their castle until 1282. In order to avoid such problems, there were two obvious solutions. First, fortifications could have been demolished in the same way that the Genoese quarter at Acre was in the late 12508, bringing peace to the city for the first time in years. Secondly, the creation of much larger royal armies might have enabled rulers such as the counts of Tripoli to subdue rebellious strongpoints like Gibelet much more swiftly than they did. However, the constant lack of troops made this latter solution unfeasible, whilst the former would have been politically impossible and would have exposed Christian territories to a swift and decisive Muslim invasion. Hence the Latins had to put up with periods of internal fighting rather than risk being wiped out by their common enemy. Viewed within the context of late medieval Europe, it could also be argued that the Holy Land was no more violent or self-destructive than many other regions. Most notably, thirteenth-century Acre, with its numerous towers and sporadic clashes involving Frederick II or other local powers, seems strongly reminiscent of contemporary Italy. There were 150 urban castles in Florence alone, and 'in central Italy, the papacy tried to impose its control over a multitude of communes and lordships, almost all fortified, by a mixture of force, diplomacy and bribery'.46 Therefore, whilst it may be true to say that internal fighting hastened the fall of the Holy Land, it is unfair to suggest that during the last decades of their existence the crusader states slipped into a state of total anarchy, or that this was unparalleled anywhere else in the Mediterranean world. Nor was the decline of central authority steady and inevitable, but rather sporadic and localised. Early in the thirteenth century Bohemond IV successfully captured the castle of Nephin, and in 1282 Bohemond VII did finally manage to subdue his Embriaco opponents. Some moments of royal weakness can also be explained, at least in part, by wider political factors. In 1229, for example, Frederick II may not have pursued his arguments with the Templars because of his desire

F O R T I F I C A T I O N S AND I N T E R N A L S E C U R I T Y

63

to get back to Europe and avoid getting bogged down in the Latin East. Perhaps, therefore, there has been too much of a tendency in the past to treat crusader fortifications as a distinct group, whereas in reality their role in internal conflicts, like their role in the maintenance of law and order, was virtually identical to that of countless other strongpoints in Italy and many other parts of western Europe.

5

The Offensive Functions of Fortifications When the Franks hoped to strengthen or extend their territorial power, they could again rely on strongholds to make up for their lack of troops in the field. By constructing new fortifications in areas which had been abandoned by the Muslims or acquired by treaty, the Christians could consolidate their borders without having to face their opponents in open battle. In the short term, this tactic was relied upon to re-establish Latin control over areas which had been lost temporarily. Goliath, a castrum-style fortification in the county of Tripoli, probably fell to Saladin in 1188, was demolished by al-Adil twenty years later, and was again overrun by Baybars during a Mamluk raid carried out in 1266. There is no evidence that the Franks tried to prevent these attacks by confronting the Muslims, and in 1266 at least Goliath's garrison quite simply fled without even attempting to hold such an exposed and relatively small castle against a far larger Mamluk army.1 As long as more powerful neighbouring strongpoints such as Chastel Blanc, Crac des Chevaliers and Tripoli held out, it was equally pointless for the Muslims to install their own troops at Goliath, for these forces would themselves have been terribly exposed to a Christian counter-attack once the rest of the Muslim army had withdrawn. Consequently, successive Muslim aggressors decided to sack Goliath, but none of them could actually prevent the Franks from reoccupying the site and rebuilding it at least twice. Goliath was probably not lost for good until Baybars conquered the entire plain of Akkar in 1271.2 Many other smaller sites situated close to major strongpoints, such as the fortified mills of Doc and Recordane, had no wider strategic value, and were therefore abandoned or even destroyed occasionally, but held great importance in terms of local security and prosperity, and were therefore rebuilt again and again. The larger fortified centres such as Acre or Crac des Chevaliers were not only capable of withstanding major incursions, but also acted as focal points for Christian reconquest and reconstruction once temporary invasions had come to an end.3 So far this strategy has been described in fairly localised terms, but it was also used to consolidate Prankish rule over more extensive new acquisitions. The key strongholds erected or reoccupied in the early 12408, most notably

THE O F F E N S I V E F U N C T I O N S OF F O R T I F I C A T I O N S

65

Beaufort and Saphet, ensured that many of the inland areas gained at this time remained under Christian rule until the 12605. Initial efforts to recolonise numerous urban sites lost or destroyed after the battle of Hattin centred around the construction of new castles, especially at Caesarea, Sidon, Ascalon, Tiberias and Jerusalem. Further north, the tower which Bartholomew de Ravendel constructed at Maraclea after 1277 represented another, unusually late, attempt to .re-establish a Prankish lordship previously overrun by the Muslims.4 Just along the coast, Bohemond VI also retook Latakia around 1261 'and built a strong new tower' there, so that the town, which had belonged to the Muslims since 1188, returned to Christian control for the next twenty-six years.5 It is interesting to note that out of all these gains this last example appears to have been the only settlement restored to Christian rule after a direct confrontation with Muslim defenders. This suggests that the Latins were normally too outnumbered to make conquests by force, yet castles still enabled them to hold on to areas gained through diplomacy. Although it did not involve any territorial expansion as such, Louis IX's visit to the Holy Land between the summer of 1250 and the spring of 1254 is also instructive. According to Joinville, Louis arrived in Acre with barely one hundred knights still available to him out of an original force of 2800 before his defeat in Egypt.6 Yet he was still able to consolidate Latin rule in the Holy Land and help to extend the life of the crusader states by building new fortifications or repairing older ones at Acre, Jaffa, Sidon, Caesarea, and Haifa.7 During this building programme Louis IX's meagre forces were not entirely safe from enemy incursions, for in 1253 a Muslim raid against Sidon resulted in substantial Christian casualties.8 This probably helped motivate a Latin counter-raid against the Muslim town of Banyas soon afterwards.9 Some Christians at least survived the Muslim attack by sheltering inside Sidon's sea castle, constructed during the winter of 1227-28 by members of Frederick II's crusade. Similarly, when Louis IX's forces rebuilt the walls of Jaffa, Joinville noted that the army made camp on the fields around the much older fortress already built there by Frederick; a location which was not only convenient but was close to a source of shelter. Both Frederick and Louis slowly but steadily consolidated Latin power through the construction of fortifications, with the latter being aided by the work of his predecessor, even though Louis no longer had any hope of actually confronting the Muslims in a pitched battle.10 There are other examples dating from this period of fortifications being built or repaired by a crusader army as it moved deeper and deeper into enemy territory. On such occasions castles were still being erected to secure territorial gains for the future, but in the short term they were also needed to

66

U N K N O W N CRUSADER CASTLES

protect the army itself from potential counter-attack. Richard I's campaigns in Palestine during the Third Crusade provide us with a classic example of such tactics. Following the recapture of Acre in July 1191, the crusaders immediately set about refortifying its defences so that they were no longer as exposed to Saladin's incursions as they had been when camped outside the city.11 Later, Acre became a springboard for Richard Fs march southwards, but the further he moved away from the city, the more vulnerable he became to enemy attacks. In a sense, self-preservation had as much to do with Richard's hasty refbrtification of Jaffa and Ascalon as the desire to reconquer these cities, for his troops needed regular places of shelter where they could rest and receive supplies from the crusader fleet.12 Moving inland from Jaffa during the autumn of 1191, Richard used similar methods to protect his army in his unsuccessful bid to recapture Jerusalem. In October 1191 the Templars accompanying Richard quickly rebuilt Casal des Plains, a relatively small fortification consisting of a central keep surrounded by a walled outer enclosure. This castle, which had probably belonged to the Templars earlier in the twelfth century, had been destroyed recently by Saladin, but its location along the road to Jerusalem made it a useful base for Richard. As was the case at Ascalon and Jaffa, there was a dual incentive for the reconstruction of Casal des Plains: in the short term it protected the members of the Third Crusade, but in the long term the Templars were clearly staking their claim to land which had been theirs before the battle of Hattin. These hopes did not materialise, for after Richard retreated from the area it fell under Saladin's control, and there is no evidence that the Templars ever returned to Casal des Plains.13 Despite these setbacks, the Templars were closely involved in thirteenthcentury efforts to regain Jerusalem and the land which connected it with the coast, an area where they had held other properties before 1187 and had carried out their self-appointed task of protecting pilgrims.14 Once again, it was clearly hoped that the construction of castles would enable the Christians to bring this region back under their control and to defend it against the Muslims even without the presence of a large Prankish field army. When the massive Templar fortress of 'Atlit was built from 1217 onwards along the coastal route toward Jerusalem, Oliver of Paderborn wrote that 'the primary advantage of this building is that the assembly of Templars ... will remain in the garrison of this fort up until the restoration of the walls of Jerusalem'.15 Toward the end of 1243, by which time Jerusalem had been regained through peace treaties, the Templars were planning to reconstruct Latrun (Toron des Chevaliers), another ruined fortress they had once held along the route to the city which originally consisted of a square keep defended by two outer ramparts. Although this would clearly have

THE O F F E N S I V E F U N C T I O N S OF F O R T I F I C A T I O N S

67

strengthened the life-line between Jerusalem and the coast, there is no evidence that the project was even begun before the city fell to the Khwarizmians in 1244 and the area came under Egyptian rule.16 In 1253, as we have seen, Louis IX toyed with the idea of reconstructing an inland fortress, before the local nobility persuaded him that such a castle would be too isolated and he therefore refortified Sidon instead. It would appear that Latrun was the inland fortress which Louis had in mind, suggesting that even at this late stage some Franks were still interested in the possibility of inching their way toward Jerusalem by constructing castles along the way, in the manner of Richard Lionheart.17 So far we have only looked at fortifications situated on land which already belonged to the Latins, had been acquired through diplomacy, had been very weakly defended by the Muslims or had even been abandoned by them. However, at certain times, and particularly when a major expedition arrived from western Europe, the Latins did attempt to capture enemy fortifications through direct force, and on such occasions the strongpoints which they already held acted as useful supply bases and starting points for offensives. The earliest and most obvious example from this period was the role of Tripoli and Tyre during the dark days following Saladin's victory at the battle of Hattin. Despite a determined siege during the autumn of 1187, Tyre was the only port in the kingdom of Jerusalem which Saladin failed to capture. It therefore became a vital refuge for many beleaguered Christians.18 Meanwhile King William II of Sicily sent fifty galleys and 500 knights to Tripoli, which was therefore able to withstand Saladin's forces the following year. If these two cities had not successfully resisted the Muslim onslaught, it seems unlikely that the Christian counter-attack which culminated in the recapture of Acre three years later could ever have taken place.19 Later Acre itself played a similar role, for although Richard I was not obliged to besiege Caesarea, Arsuf, Jaffa and Ascalon, all of which had been slighted and abandoned by the Muslims, he still had to defeat Saladin at the famous battle of Arsuf before he could reach them, and constantly relied on seaborne supplies sent from Acre during this risky phase of the campaign.20 After the end of the Third Crusade in 1192 numerous other expeditions were launched from Acre, as the Latins continued to try to reconstruct the kingdom of Jerusalem. In 1197 German troops belonging to Henry VFs crusade arrived there before marching north and capturing Beirut, whose Muslim garrison fled in panic after attempting to make a stand in front of their castle.21 Other, less successful campaigns which relied on Acre as their starting point included the failed siege of Mount Tabor undertaken by members of the Fifth Crusade in 1217,22 and the Lord Edward's somewhat

68

U N K N O W N CRUSADER CASTLES

confused attack on Qaqun in 1271, which may have been an extensive raid rather than a siege.23 During the German crusade of 1197-98, Acre also became the initial base for an attack on the isolated inland stronghold of Toron (Tibnin), but this had to be abandoned at the approach of a Muslim relieving force.24 Further north, Prankish efforts to regain Muslim-held territories had equally mixed results. In 1191 Bohemond III failed to reconquer Latakia and the neighbouring port of Jabala in an offensive which was presumably launched from Tripoli itself.25 In 1207 Raymond Roupen, an Armenian claimant to the throne of Antioch, granted Jabala to the Hospitallers in order to gain their support against his political opponents, and in 1210 this was followed by Bikisrail, another former crusader castle situated to the north east of Margat.26 However, the Hospitallers only ever managed to recover half of Jabala, which they were forced to share with the Templars, who had been granted the very same settlement by Raymond's rival Bohemond IV.27 Meanwhile, their strong presence in Syria meant that both Orders were at the forefront of many campaigns against the Muslims in the county of Tripoli, such as the failed siege of Horns which the Hospitallers of Crac des Chevaliers undertook in 12O7.28 To the north of Antioch, the Templars also spent much of the thirteenth century trying to reconquer the vast estates which they had held there before 1188, and in particular the strategic fortresses of Baghras and Darbsak. These castles guarded two of the most important mountain passes connecting Antioch with the Cilician plain, but the former was occupied by the Armenians between 1190 and 1216, whilst the latter had been garrisoned by Aleppine forces in the wake of Saladin's invasion. As a result, the Templars only retained the neighbouring stronghold of Roche Roussel (Hadjar Shoglan/Chilvan Kale) during the early years of the thirteenth century, and it was from here that they made an unsuccessful attack on Darbsak in 1237.29 The last phase of Christian expansion in Syria occurred after the Mongol destruction of Aleppo in 1260, which enabled Bohemond VI to commit 'many acts of aggression against the lands of Islam' so that 'he took a number of villages in Muslim territory', as well as the town of Latakia itself. All these conquests appear to have been made from the city of Tripoli.30 It was also at this time that Bohemond acquired Darkoush, Kafr Dubbin and other castles situated to the east of Antioch, which had been captured by the Muslims in 1188 but could now be reoccupied by the Franks after the Mongol invasion.31 Unfortunately for the Templars, a similar set of circumstances enabled the Armenians to occupy the fortress of Darbsak. In 1268, however, all these strongholds were lost to the Mamluks along with Antioch.32 These events again highlight the Franks' chronic lack of

THE O F F E N S I V E F U N C T I O N S OF F O R T I F I C A T I O N S

69

manpower, for it is clear that Bohemond's (and the Armenians') successes during the early 1260$ began and ended with the Mongol invasion of eastern Syria. Moreover, most attempted Christian sieges of this period, including those of Horns, Qaqun and Toron, failed because the Franks were too outnumbered to face the Muslim relieving forces sent to assist these places. The Franks were consequently obliged to retreat. Such setbacks probably explain why sustained, large-scale sieges undertaken by the Latins were extremely rare during the thirteenth century, often being rejected in favour of smaller raiding expeditions which could be called off as soon as the Muslims tried to retaliate. Offensives of this category were also more manageable because they were intended to gain booty rather than permanent territorial conquests and could consequently be undertaken by fewer troops. Many campaigns of this kind still represented a relatively large commitment of men and resources for the Latins, and were therefore equally dependent upon castles or fortified cities as starting points and supply bases. In 1271, for example, the lord Edward's troops joined forces with Templars, Hospitallers and other soldiers from Acre in a raid against St George, situated just a few miles inland. The expedition, undertaken by 1500 horsemen, succeeded in causing widespread destruction of Muslim crops and property, and at the same time bringing its participants substantial booty in the form of grain and cattle.33 Other raids were also launched from the coast against Sidon, Beaufort and central Galilee during the Fifth Crusade,34 and in 1253 some of Louis IX's followers carried out an attack on the Muslim town of Banyas while they were based at Sidon.35 It has already been noted that three years later Julian of Sidon also raided Mongol territories to the east of Beaufort,36 whilst in 1269 the arrival of a small Aragonese crusade enabled the Franks to attack several Muslim villages near Montfort with a force led by around 130 knights.37 In the county of Tripoli, several similar offensives were carried out from Margat and Crac des Chevaliers during the first years of the thirteenth century. In 1203 an army consisting of 400 knights, 1400 footsoldiers, numerous turcopoles and archers, many of them drawn from the garrisons of these strongholds, suffered a crushing defeat near Montferrand, only a month after another Hospitaller force had been routed while trying to attack Kama.38 Nevertheless, the very next year the Order made more successful incursions against both Horns and Hama, and may well have participated in another expedition toward Jabala.39 In 1265 the Templars and Hospitallers also joined forces with the bellicose Bohemond VI on a raid which was eventually repulsed by Muslim troops from Horns.40 Fourteen years later, the garrison of Margat organised its last expedition against former Hospitaller estates around Crac des Chevaliers and Chastel Blanc,

7O

UNKNOWN CRUSADER CASTLES

and clearly maintained an aggressive stance toward the Muslims right up until the very end.« It has been suggested that, despite the risks involved and the occasional defeats suffered, these offensives were an essential way of protecting Latin lands, for by using attack as the best means of defence, the Franks were masking their true vulnerability from their opponents and maintaining their 'reputation for ferocity'.42 Although it is clear that the vast majority of offensive campaigns organised by the Franks during the thirteenth century either failed to make any territorial gains or were not in fact intended to do so, the sheer amount of destruction which these expeditions inflicted on exposed towns and villages was such that they could also be used to keep large areas of the countryside in thrall. Sometimes the mere threat of punitive raids enabled the garrison of a single Latin fortress to extend its authority over many estates and settlements previously held by the Muslims. For example, the security provided by Saphet encouraged more than 10,000 peasants to recolonise neighbouring territories, while the castle's strategic location above the River Jordan also made it possible for its defenders to launch raids on Muslim lands as far as Damascus. As a result, the rulers of this city were forced to relinquish many of their properties in this area.43 Once it had been completed, 'Atlit served a similar purpose, for 'between Acre and Jerusalem there is no fortification which the Saracens hold, and therefore the unbelievers are harmed greatly by that new fortress; and with the fear of God pursuing them, they are forced to abandon these cultivated regions'. Oliver of Paderborn, the author of these words, added that 'the construction of this castle is presumed to have been the cause of the destruction of [Mount Tabor], because in .the long wide plain, which lies between the mountainous districts of this camp and of Mount Tabor, no one could safely plough or sow or reap because of fear of those who lived in it'.44 The Franks finally got rid of Mount Tabor not by launching a direct military assault but by building their own stronghold, whose Templar garrison could harass the Muslims and deprive them of their food supplies. These tactics resulted in the deliberate demolition of Mount Tabor by the Muslims in 1218.45 Almost half a century later Baybars was obliged to repair the neighbouring castle of Qaqun because 'his subjects living in those parts needed a place of protection', suggesting that the Templars were still using their fortress to dominate large areas of central Galilee.46 Ironically, by the 12808 the situation had reversed, for by then Qaqun was used to intimidate 'Atlit; a clear indication of declining Prankish power in the final years before 1291.47 Although there is no direct evidence to prove it, a parallel set of events may have taken place at Maraclea during exactly the same period. As we have seen, Bartholomew de Ravendel constructed a new tower here

THE O F F E N S I V E F U N C T I O N S OF F O R T I F I C A T I O N S

71

during the late 12705, but in 1285 Kalavun insisted on the destruction of this stronghold as part of a new peace treaty with the Latins. It is possible that Kalavun, who had just captured the neighbouring Hospitaller fortress of Margat, feared that Maraclea would be used by the Franks to raid his newly acquired lands or even attempt to recapture Margat itself.48 To the south of Maraclea, the Templars and the Hospitallers also relied on intimidatory raids to maintain their power in the county of Tripoli and to keep their Muslim neighbours in check. One such neighbour was the tribe of the Assassins, schismatic Muslims who controlled the extremely mountainous district to the east of Latakia. The rough terrain in this area enabled the Assassins to retain their independence for much of the thirteenth century, but their close proximity to Margat, whose garrison could easily launch punitive raids against them, forced them to pay the Hospitallers an annual tribute of 1200 gold pieces and 100 bushels of wheat and barley. The threat posed by Crac des Chevaliers toward the emir of Hama and the ruler of Bokebais (Or Abu Qubais, a Muslim castle between Hama and the territory of the Assassins) meant that they too had to make similar annual payments to the Order worth 4000 and 800 gold pieces respectively.49 If they did not pay there could be grave consequences, for in 1229 the Hospitallers carried out an extensive raid around Montferrand, a castle they themselves had hoped to control since the twelfth century, because the emir of Hama had not paid up.50 The following year a force of 500 horsemen and 2700 footsoldiers, composed of both Hospitallers and Templars, launched another attack toward Hama, but this time the expedition ended iii defeat at the hands of the emir's army.51 The Hospitallers were not deterred. Angered by the emir's persistent refusal to pay, they organised a third offensive in 1233. This expedition was far larger than those of 1229 and 1230 and involved 100 knights, 400 mounted sergeants and 1500 footsoldiers led by the Hospitaller grand master, as well as twenty-five Templar knights, eighty knights from the kingdom of Jerusalem, thirty knights led by Bohemond V's brother Henry, and 100 knights from Cyprus. This impressive force marched overnight toward Montferrand, enabling the Franks to make a surprise attack on its castle town the following morning. Having sacked this settlement, they continued to ravage the surrounding countryside before returning toward the coast, without having encountered any Muslim resistance. Subsequently the sultan of Damascus advised the emir of Hama to pay the money he owed, and peace was re-established with the Hospitallers.52 Indeed, the Order appears to have collected tribute for a further thirty years, until Baybars finally obliged it to renounce these payments as part of a peace treaty established in 1266.53 Until that time, the Hospitallers' arrangement with Hama can

72

U N K N O W N CRUSADER CASTLES

be compared with the more common thirteenth-century method of using Latin castles to collect taxes from local Greeks or Muslims.54 Despite the fact that the Hospitallers' aggressive policies were generally successful until the 12605, the sporadic losses which they suffered indicate that it was still a risky business for relatively small Latin field armies to set off into regions where they might encounter numerically superior Muslim forces. One way to minimise this risk was to carry out raids swiftly and without warning; but another was to try to operate as close as possible to friendly castles, which therefore played an important role as refuge and supply points for armies on campaign. In 1253, for example, French companions of Louis IX who were participating in a raid on Banyas narrowly avoided a potentially disastrous confrontation with the Muslims by withdrawing to Sidon at the first sign of trouble.55 It is also interesting to note that the first major raids which the Hospitallers carried out from Crac des Chevaliers and Margat occurred in 1203 and 1204, almost twenty years after the Order had acquired the latter castle. This raises the possibility that the Hospitallers completed their rebuilding programme about the same time; an accomplishment which would not only have released more funds for warfare, but would also have enabled the brothers to go on the offensive safe in the knowledge that their own fortresses were now strong enough to resist possible counter-attacks.56 Similarly, Templar forces carrying out raids around Mount Tabor and the River Jordan must have had relatively little to fear because they knew that, even if the Muslims attempted to pursue them, they could easily retreat to Saphet or 'Atlit, both of which were virtually impregnable. In 1271 Baybars is reported to have besieged Akkar because 'brigands could come down from it and they would fortify themselves there'. According to Ibn al-Furat, these Franks could attack the surrounding countryside with impunity, for Akkar itself lay 'in difficult hill country, far from supplies of water', and was therefore almost totally immune to Muslim counter-offensives.57 On other occasions, the Franks were persuaded to stop and fight rather than run away, particularly if they were close enough to a castle to disengage from a battle if the need arose. In 1279, 200 horsemen from the garrison of Margat used this tactic against the 5000 Muslims who were trying to prevent them from ravaging the neighbourhood of Crac des Chevaliers. The Hospitallers, knowing that it would be suicidal for them to confront this force in the open, therefore allowed themselves to be chased until they had almost reached Margat itself before turning on the Muslims and routing them with the loss of only one mounted sergeant. This encounter's close proximity to a fortress made it very similar to the kind of sorties or surprise attacks sometimes launched by Latin garrisons when they were being

THE O F F E N S I V E F U N C T I O N S OF F O R T I F I C A T I O N S

73

besieged. Indeed, when Margat came under siege a mere two years later, 600 Hospitaller cavalry troops burst out and successfully drove off the 6000 startled Muslims who were attempting to capture the castle.5» In 1256 a comparable series of events took place at Jaffa, which was initially used as the starting point for a successful raid against inland territories south of Ascalon. When this raid came to an end, its participants found shelter behind the fortifications of Jaffa, which came under siege from a Muslim counter-attack. Shortly afterwards, the blockade was brought to a swift end when a force of 200 knights plus 300 mounted sergeants, archers and crossbowmen, most of whom had presumably just returned from the original raid, rushed out and defeated the Muslims.59 Similarly, during the final siege of Acre in 1291 the Franks initially kept their city gates open, so that they could launch surprise raids against the enemy troops arriving on the plain outside the city.60 Clearly, it was sometimes worth risking a direct encounter with the Muslims, provided that the Franks had some means of protecting themselves should the battle start to turn against them. The Rule of the Templars advised members of the Order that, if they were defeated in the field, they had to try to reach the nearest fortress in order to keep their casualties to a minimum.61 Likewise, the castle of Ascalon, which had been used as a collection point for Prankish troops on the eve of the battle of La Forbie, became a refuge site for those few Christians fortunate enough to escape the subsequent carnage.62 Those who did not suffered the same fate as many of the Templars involved in the failed assault on Darbsak in 1237. This campaign turned out to be one of the worst military disasters in the Order's history, with most of its participants being cut down by an Aleppine counter-offensive long before they could reach the safety of Roche Roussel (Hadjar Choglan/Chilvan Kale).63 In 1266 Hospitallers, Templars, Teutonic Knights and other secular troops campaigning near Tiberias also found themselves trapped in a Muslim ambush which resulted in the loss of around 500 troops; a total which subsequently appears to have increased because the Franks were then forced to march all the way back to Acre, a journey of approximately fifty kilometres, while being constantly harassed by local Muslim peasants. This humiliating defeat would surely not have been so costly, and perhaps could have been avoided altogether, if the Franks had been able to take refuge in a nearby castle.64 The numerous raids or sorties which took place during the thirteenth century remind us that the Franks did not simply regard their fortifications as passive refuge points but used the garrisons of such sites to raise additional troops for their field armies. This strategy had backfired disastrously in the months after the battle of Hattin, when Saladin, having annihilated a Latin

74

UNKNOWN CRUSADER CASTLES

field army containing virtually every Christian soldier who could be found in the kingdom of Jerusalem, overran several Prankish castles which were virtually empty. But their lack of troops, combined with the obvious practical advantages of relying on soldiers stationed nearby to conduct local warfare, meant that the Latins continued to rely on such forces in the thirteenth century.65 In times of defeat this policy clearly still carried a heavy penalty. The Templars, Hospitallers and Teutonic Knights, who may have contributed as many as 600 knights to the battle of La Forbie, must have found it difficult to defend their local fortresses properly after they suffered severe casualties at the hands of the Egyptians and Khwarizmians. This applied most to the Teutonic Knights, who are reputed to have lost all but three of their contingent at La Forbie, presumably putting a severe strain on their garrisons at Montfort, Acre and elsewhere.66 Perhaps losses sustained by the Hospitallers also contributed to the fall of Ascalon in 1247.67 Sixteen years after the battle of La Forbie, a somewhat smaller Christian army suffered another crushing defeat in southern Galilee. This expedition, which was led by the lord of Beirut, was partly made up of Templars sent from Acre, 'Atlit, Saphet and Beaufort.68 As a result, all these sites may well have been seriously undermanned after the battle, although it is extremely difficult to estimate exactly how many troops had been drawn from each individual fortress. However, it has been calculated that Saphet alone had a peace time garrison of 1650 men, fifty of whom were mounted knights,6^ and if similar numbers applied to their other major castles, the Templars could have represented a substantial percentage of the 900 knights and additional turcopoles who are said to have taken part in the battle. Contemporary sources suggest that the Order suffered particularly heavy casualties and had to pay a considerable ransom for the release of their captured commander.?0 By contrast, it was probably less risky for the Latins to take troops from strongholds which were not particularly important strategically. It has been noted, for example, that the citadel of Acre contributed relatively little to the overall strength of this city, and it could probably be left virtually empty provided that surrounding ramparts were still being guarded. From 1254 onwards, this structure acted as the barracks for the French regiment, a standing force which had been established by Louis IX and normally contained about 100 knights, plus additional crossbowmen and infantry. During the 12505 and 12605 this regiment participated in several large-scale raids, including the campaign which ended in the successful defence of Jaffa in 1256.71 Although the French regiment clearly boosted Latin troop numbers, its contribution was not as significant as that of the Military Orders, who participated in virtually every Christian campaign of the thirteenth century.

i. Chastel Blanc: the keep. (Jonathan Phillips}

2. Antioch: the city walls and the citadel. (Denys Pringle)

3. Caesarea: the city walls built by Louis IX. (Denys Pringle)

4- Sidon: the sea castle. (Denys Pringle)

5. Gibelet: the citadel. (Denys Pringle)

6. Tortosa. Note the extremely large and well cut stones used in the Templar defences. (Denys Pringle)

7. Chastel Rouge: gateway through perimeter wall into the castle enclosure, (Denys Pringle)

8. Chastel Rouge: the central keep and perimeter walls. (Denys Pringle)

9. Markopoulo: a typical Latin tower in Greece. (Coutauld Institute)

10. The Red Tower: a typical Latin tower in the Holy Land. (Denys Pringle)

11. Qaqun: the central keep, with traces of the perimeter defences on the right. (Denys Pringle)

12. Chastel Blanc: the chapel inside the keep. (Jonathan Phillips)

13. Crac des Chevaliers: the great hall. (Denys Pringle)

14. Crac des Chevaliers: distant view showing the rich farmland of the surrounding hills. (Jonathan Phillips)

15. Crac des Chevaliers: the central courtyard. (Jonathan Phillips)

16. Recordane: the fortified mill. (Denys Pringle) ,

i/. Recordane: generai view of the mill. (Denys Pringk)

THE O F F E N S I V E F U N C T I O N S OF F O R T I F I C A T I O N S

75

The dominance of these Orders was particularly apparent in Syria, where the Templars and the Hospitallers did not simply contribute to a wider Prankish war effort but were pursuing their own aggressive policy toward the Assassins and the Muslims of Horns, Hama and Aleppo. As we have seen, this policy was maintained through a series of punitive raids, most of which were launched from the Hospitaller castles of Margat and Crac des Chevaliers. To a large extent these fortresses were used for such campaigns because of their strength and strategic location, but they must also have been useful because of their unusually large garrisons. In 1212, for example, Willbrand of Oldenburg claimed that there were 2000 men stationed at Crac des Chevaliers and a further 1000 troops defending Margat.72 No similar figures survive for Chastel Blanc or Tortosa, the two largest Templar castles in the region, although they presumably also had hundreds of soldiers living within their walls. These totals would have enabled the Templars and Hospitallers to use garrison troops for the many expeditions which have already been mentioned, including, for example, the joint campaign against Montferrand of 1233.73 Furthermore, an incident dating from 1266, when the Templar castellan of Chastel Blanc sent 'fifty crossbowmen and arbalasters' to reinforce Crac des Chevaliers against the Mamluks, indicates that soldiers stationed within castles could sometimes bolster other fortresses just as much as Latin armies in the field.74 A number of problems make it difficult for us to add any precise details to these general observations. For a start, Willbrand of Oldenburg's figures for Crac des Chevaliers and Margat sound suspiciously like vague approximations, whilst it has already been noted that no figures at all exist for Tortosa or Chastel Blanc. When Baybars captured the latter stronghold in 1271, he found 700 people inside it. While this total may reflect the actual size of the garrison, it could also be another rough estimate, or simply refer to non-combatants who were seeking shelter rather than participating in the fighting.75 Even if Willbrand's totals for Crac des Chevaliers and Margat are accurate, they do not indicate what percentage of each castle's garrison consisted of infantry, archers, mounted troops or actual brothers of the Hospital. Nor is it likely that these numbers still applied during the increasingly desperate years of the late thirteenth century, for in 1268 the Master of the Order wrote that there were only 300 brother knights left to defend all Hospitaller properties in Syria. This suggests that when archers had to be sent from Chastel Blanc to Crac des Chevaliers in 1266, the latter fortress no longer had the same number of defenders as it had done in I2i2.76 By using more detailed thirteenth-century figures given for the Templar garrison of Saphet, and combining them with a papal letter of 1255 which states that sixty Hospitaller knights should ideally be stationed at Crac des

j6

UNKNOWN CRUSADER CASTLES

Chevaliers, the military historian Christopher Marshall has estimated that Crac and Margat probably had a total mounted force of 160 and 80 horsemen respectively.77 If these calculations are correct, the 600 cavalry troops who defeated Muslim besiegers outside Margat in 1281, as well as the 200 horsemen from the same fortress who ravaged the plain of Akkar two years earlier, cannot all have come from this castle's garrison.78 Clearly, any castle which was chosen as the starting point for a raid, or alternatively came under direct Muslim attack, would have had its troop numbers boosted by mercenaries or soldiers brought in from elsewhere. While the basic point still applies that castle troops were often used to augment Latin field armies, the quantity and quality of the men involved could vary from stronghold to stronghold and must have changed with time according to the broader military situation. For much of the time, Latin strongholds were consequently intended to provide troops, supplies and shelter for Christian field armies without becoming directly involved in any fighting. But occasionally the Franks also constructed fortifications which were specifically designed to blockade enemy castles. The imperialist siege of Beirut, for example, was largely conducted from a small fort which Frederick II's Lombard troops erected on a hill opposite the citadel in the autumn of 1231. This fort, which had been built out of 'stones with wood above', was used by Richard Filangieri and his men to bombard the Ibelin castle with catapults.79 On a much larger scale, the Christian forces who besieged Acre between 1189 and 1191 based themselves within fortified encampments which were situated on a hill to the east of the city and were defended by successive lines of wooden palisades and ditches. By the beginning of 1190, this camp had virtually become a walled town in its own right, containing numerous churches and other residential or functional buildings.80 Another interesting aspect of this siege was Richard Fs use of a 'portable' castle. This large, wooden structure, initially erected by Richard in Sicily during his journey to the Holy Land, was subsequently taken to Acre, where it was used to attack the city walls.81 Similar siege tactics were also used in Prankish Greece.82 Such fortifications were almost entirely made of wood, but it should also be noted that much more substantial stone defences could be built with remarkable speed. We should not make the mistake of assuming that the appearance of crusader castles remained unaltered for decades or even centuries. 'Atlit, for example, was begun in 1217/18, but was already strong enough to resist a major siege two years later.83 On another occasion, Richard I was in such a hurry to repair the citadel of Jaffa that he did not use any mortar but simply built dry stone walls which could be strengthened later.84 During the spring of 1192 Richard also spent a mere four months

THE O F F E N S I V E F U N C T I O N S OF F O R T I F I C A T I O N S

77

rebuilding Ascalon's walls, only for them to be demolished again soon after as part of his peace treaty with Saladin.85 Indeed, both Saladin and Baybars systematically dismantled many strongholds to prevent their recapture, and even as late as 1270 the latter sultan virtually flattened the last remaining defences at Ascalon. This was presumably done to deter Louis IX or the Lord Edward from devoting their crusades to refortifying these sites and then attacking Jerusalem or Egypt.86 Such incidents suggest that castles were seen as far more temporary structures than they are today, and are proof indeed of William of Tyre's famous saying that 'a castle destroyed is a castle half-rebuilt'.»? Finally, it is important to remember that many Prankish strongpoints along the coast were used for naval as well as land based offensives. Attacks of this kind were most common during the frequent clashes between the Italian city states, which tried to dominate local trade by deploying enormous war fleets against each other. In 1264, for example, the Venetians attacked the pro-Genoese city of Tyre with no less than fifty galleys equipped with special boarding towers designed to overwhelm defenders positioned along the sea walls.88 This assault proved less successful than an earlier, more discreet raid carried out in 1242, when Venetian galleys were allowed to slip into the harbour after their Ibelin allies had secretly lowered the chain across its entrance.89 Both these offensives were planned and carried out from Acre, but in 1232 Tyre itself became the starting point for another naval raid, this time undertaken by twenty-two imperialist galleys, which attacked Ibelin forces camped a few miles to the south at Casal Imbert.90 The previous year Lombard troops had also arrived by sea at Beirut, enabling them to bring with them many of the building materials needed for their fort.91 Similar episodes took place during internecine fighting in the north, such as the fleet of fifteen galleys sent by Bohemond VII from Tripoli against Templar Sidon in 1278. The provocation for this assault had been the Templars' involvement with the lord of Gibelet, who was plotting to seize Tripoli from Bohemond.92 Numerous seaborne attacks were also organised against the Muslims during the thirteenth century. In the build up to the Fifth Crusade, for example, the Christian fleet gathered in or near the harbour of 'Atlit during preparations for the forthcoming expedition to Egypt.93 Although it is not recorded where they set out from, a number of Frankish ships launched another damaging attack on Alexandria in 1270, and made off with two Muslim vessels.94 These incidents, as well as references to a Templar shipyard at Acre, make it clear that coastal fortifications provided secure bases for the maintenance and construction of ships and galleys, essential if the Latins were to maintain their dominance over the sea.95

78

U N K N O W N C R U S A D E R CASTLES

Virtually all aggressive military activities undertaken by the Franks involved fortifications in some way. Such structures were used to extend or consolidate Latin territories despite the numerical inferiority of Christian field armies. At the same time, they could provide land based or seaborne Christian forces with supplies, shelter or additional soldiers. Although some of the strongholds mentioned in this chapter, such as Casal des Plains, Maraclea or the temporary Lombard fort erected opposite Beirut, were smaller in size, most were extremely large. Fortified cities like Acre, Tyre and Tripoli, plus major castles such as 'Atlit, Crac des Chevaliers, Margal and Saphet, have been referred to again and again. The same massive fortifications which played such a key defensive role in preserving the crusader states against full-scale Muslim invasions also played a prominent part in any Latin offensives. This is hardly surprising considering that these sites had the most troops and resources concentrated within them, and that some, most notably Acre, provided plenty of shelter for crusaders arriving from Europe on the eve of a campaign. Their great size and strength meant that such strongpoints were large and secure enough for Christian armies to retreat inside whenever they had been defeated or were being pursued by the Muslims. Once again these observations indicate that we need to distinguish between the much wider military role of some of the largest crusader fortifications, which contributed to the security of the crusader states as a whole, and most smaller strongpoints, which were normally only built with local defence in mind.

6

The Military Orders By the late thirteenth century the Military Orders, whose presence had already been extremely important to the survival of the crusader states for one hundred years, had acquired most, and in some areas virtually all, key Latin fortifications. This process speeded up considerably after 1250, when constant Muslim incursions and a chronic lack of resources forced many barons to hand their castles over to the Hospitallers, Templars and Teutonic Knights. During this period some Mamluk armies contained tens of thousands of troops. If these forces were intent upon capturing a crusader strongpoint, it only stood any realistic chance of survival if it had the kind of enormous and elaborate defences which have already been referred to at sites like 'Atlit. However, it is clear that eventually most secular lords did not even have the means to maintain their existing defences, let alone upgrade their fortresses to the standard of Saphet or Crac des Chevaliers. In 1257, for example, Julian, lord of Sidon, sold numerous estates between Sidon and Beirut to the Teutonic Knights for 23,500 Saracen bezants, suggesting that he was in serious financial difficulties and needed to raise cash quickly. In the same year the German Order also purchased the rights to Cave de Tyron, which may no longer have been in Prankish hands,1 and three years later Julian even sold Beaufort and Sidon to the Templars. This last transaction was necessitated by the Mongol destruction of Sidon in 1260, which appears to have bankrupted Julian and left him unable to pay for the town's reconstruction.2 Similar financial pressures explain why the Military Orders acquired many other territories and fortresses during the thirteenth century. The Teutonic Knights in particular profited from baronial poverty by buying up estates in the vicinity of Acre. These properties included the site of the Order's headquarters at Montfort, formally gained in 1228, as well as other lands nearer the coast which were acquired in 1220 and were protected by the smaller castles of Mhalia (Castrum Regis) and Judin (Jiddin).3 This latter stronghold, which appears to have been constructed by the Teutonic Knights at some point after the mid 12205, is also architecturally interesting in that its defences were dominated by not one but two approximately square keep towers. It has already been noted that the Knights built a

80

UNKNOWN CRUSADER CASTLES

huge D-shaped donjon at Montfort to guard the approaches to this fortress, whilst in a later chapter we shall see that they constructed yet another massive keep of this kind at Amuda, their principal stronghold in Cuidan Armenia. This has led to the suggestion that the Order constructed its fortifications according to a distinct German style which was not dissimilar to that of some Rhineland castles. It also indicates that the Military Orders were still prepared to build castles centred around an isolated keep well into the thirteenth century, despite the absence of such a feature at the more famous concentric strongholds of the period like 'Atlit and Crac des Chevaliers.4 To the north of Jüdin, Frederick II also helped the Teutonic Knights to gain control over several manors near Toron, whilst Julian of Sidon continued to sell parts of his lordship to the Order until as late as i26i.5 In 1209 Bohemond IV also entrusted three towers along the curtain walls of Tripoli to the Teutonic Knights, but these appear to have been the only fortifications guarded by the Order in all of northern Syria. This grant is particularly surprising considering Bohemond's appalling relations with the Armenian monarch Leon II, a staunch ally of both the German emperor and the Teutonic Knights. Perhaps, therefore, it was an isolated attempt to get the Order to abandon its support for Leon in exchange for property grants. If this was the case, it must have failed, for the close links between the Germans and the'Armenians may explain why the Teutonic Knights never expanded any further in the northern crusader states.6 Although the Teutonic Knights rapidly became powerful landholders in the first half of the thirteenth century, their properties were not as extensive or as significant in military terms as those of the Templars. Perhaps the most important Templar possessions in the south were Saphet and Beaufort, because of their close proximity to the frontier with Damascus. Both these castles had originally been held by individual lords, but they were granted to the Templars in 1168 and 1260 respectively, again reflecting the steady erosion of baronial power in the Latin East.7 The other most important Templar fortresses in the kingdom of Jerusalem were Sidon and 'Atlit, the latter having been built on a site which already belonged to the Order before nS/.8 In the county of Tripoli the Templars had also acquired the castles of Tortosa and Chastel Blanc in the middle years of the twelfth century, and had immediately set about rebuilding their defences. Tortosa in particular had probably been handed over by the local count because, like Julian of Sidon in 1260, he could not afford to repair the terrible damage inflicted on the site by Nur ai-Din. It seems that many other neighbouring estates and properties, including the castle of al-Arimah, were sold to the Templars at about the same time.9 In addition, the Templars held another

THE M I L I T A R Y O R D E R S

8l

large territorial block to the north of Antioch, which may have come into their possession as early as the 11305. As far as the Hospitallers were concerned, the most important fortresses in the kingdom of Jerusalem were Arsuf, Ascalon and Mount Tabor. Arsuf was rented out to them in 1261 by Balian of Ibelin, whilst Ascalon was entrusted to the Order soon after its reconstruction in 1241. Mount Tabor, on the other hand, was granted to the Hospitallers by Alexander IV, acting on behalf of the monks who lived there. This transaction is an interesting example of growing papal involvement in the military affairs of the Latin East, and suggests that the church as well as the nobility had problems defending its territories. Further north, however, the key Hospitaller strongholds of Crac des Chevaliers and Margat had both been sold to the Order by laymen rather than the church. These purchases had occured in 1144 and 1186 respectively, so that like the Templars, the Hospitallers were already well established in the county of Tripoli before the battle of Hattin.10 In addition to these major fortresses, the Hospitallers held countless smaller strongholds whose size and shape varied considerably. On and around the plain of Akkar they possessed Goliath, Chastel Rouge and numerous other sites which were not permanently lost until 1271.11 After the fall of Chastel Blanc, Baybars occupied 'its territory, together with the forts and towers in the neighbourhood of Hisn al-Akrad' (Crac des Chevaliers). Once Baybars had conquered Crac, the Hospitallers abandoned several further towers, 'burning all their property that they could not remove'.12 Most of these structures had probably been built or acquired in the twelfth century, and many had been sold along with Margat and Crac des Chevaliers. Indeed, when they purchased the former stronghold, the Hospitallers received the entire town of Banyas (Valania), which lay below the slopes of Margat.13 In the kingdom of Jerusalem, it appears that many twelfth-century Hospitaller castles were either lost for good in the period 1187-92, or were left in ruins even if neighbouring areas were subsequently regained by treaty. Excavations at the castle of Belmont, for example, confirm that this site was never restored after Saladin demolished it in 1191, and the area was later occupied by a Muslim village. No attempt was made to rebuild Belmont even after Frederick II negotiated the return of Jerusalem itself in 1229.14 Furthermore, archaeological investigation suggests that the Hospitallers never regarrisoned Belvoir or the neighbouring castle of Forbelet either. In theory this region of Galilee came under Christian rule again during the early 12405, and until Belvoir was actually excavated in the 19605 some historians therefore argued that this site had been reoccupied by the Order briefly.15 By contrast, it seems that the Hospitallers did return to some properties regained at this time which were not so remote, such

82

UNKNOWN CRUSADER CASTLES

as Bethgibelin, located inland from Ascalon and held for a short spell between 1240 and 1244.16 Nearer the coast, the Hospitallers also possessed a number of other small strongpoints which were similar in design to the tower of Tukhlah. These included La Tor de l'Opital, on the outskirts of Tyre, and Turns Salinarum, situated near Caesarea. Buildings of this kind were primarily intended as local watch towers and places from which to administer neighbouring estates.17 Bearing in mind that identifying smaller Hospitaller properties can sometimes be difficult, it is hardly surprising that trying to do the same for the Templars, whose records no longer survive, can be virtually impossible. Some information can be gleaned, however, from contemporary descriptions of property disputes between the Templars and the Hospitallers. It is in this way, for example, that we know the two Orders agreed to divide the town of Jábala between them, after the Templars disputed the way in which it had been granted to the Hospitallers by Raymond Roupen.18 This deal was finalised in 1233, but in 1266 the Templars relinquished their half of the town to Baybars in exchange for a peace treaty covering Chastel Blanc and Tortosa. Ibn al-Furat's account of this arrangement is interesting in that it mentions the presence of a tower at Jábala, which could either have been similar to that erected by Bohemond VI after he captured neighbouring Latakia around 1261, or could have been built by the Templars and the Hospitallers themselves.19 Alternatively, it may have been Jabala's fortified Roman theatre, which was presumably the 'strong castle' Willbrand of Oldenburg referred to when he passed by more than half a century earlier.20 In the kingdom of Jerusalem, the fortified mill at Doc represents another site whose Templar ownership can be confirmed by studying Hospitaller documents. Doc was located along the same river as Recordane, which belonged to the Hospitallers, and as a result both mills became the object of a fierce argument between the two Orders when it was proposed to build a new dam across this vital waterway. This dispute lasted for several years until it was finally settled in 1235.21 Hospitaller records also suggest that the Templars occupied the Red Tower, situated to the south east of Caesarea, even though the abbey of St Mary of the Latins (the actual owner of the site) had originally rented out this small agricultural centre to the Hospitallers in an agreement dating from iiSp.22 This situation had probably come about at the time of the Third Crusade, when Templars accompanying Richard I may have garrisoned the tower without even consulting its official tenants. Subsequent references to this stronghold indicate that it may not have been returned to the Hospitallers until as late as 1248. Without these scraps of evidence we would not have known about a Templar presence on the Sharon plain lasting more than half a century!23

THE M I L I T A R Y O R D E R S

83

Elsewhere in the south the Templars seem to have been just as badly affected by Saladin's conquests of 1187 as the Hospitallers. In the twelfth century they held a number of castles between Jerusalem and the coast whose role it was to secure this region for pilgrims travelling to and from the holy city. But we have already seen that, with the exception of Casal des Plains, which was only reconstructed briefly during the Third Crusade itself, this chain of strongholds was never rebuilt.24 It is hardly surprising that other Templar properties to the east of Jerusalem which had once protected pilgrims visiting Jericho and the River Jordan were also lost for good in nS/.25 To the south, the twelfth-century Templar fortress of Gaza had a similar history to that of Casal des Plains, being lost after the battle of Hattin, restored briefly by Richard I in 1192 but demolished under the terms of the peace treaty with Saladin later that year.26 During the early 12405, at the same time that plans were afoot to rebuild the Templar fortress of Latrun near Jerusalem, the Order also hoped to restore Gaza. Clearly there was a strong desire at this time to recreate the Order's former territories in the south and strengthen crusader rule in Palestine, but the battle of La Forbie in 1244 and the subsequent reconquest of the area by the Egyptians quickly put paid to the entire project27 Like the Hospitallers, the Templars seem to have had more luck in areas which were not as exposed as Gaza or as far inland as Jerusalem. Around 'Adit, for example, Templar lands were actually expanding during the thirteenth century as former secular lordships such as Caymont, a castle and village which had passed to the Order by 1262, were gradually bought up.28 Many of these lands had to be given up to the Mamluks in a peace treaty arranged in 1283, again reflecting the rapid decline of 'Atlit's status during the final decade of its existence.2« Apart from holding their own castles, the three Orders also helped to garrison the major Christian cities located along the coast. At Acre, the Teutonic Knights were expected to maintain and garrison a section of the ramparts near the Accursed Tower, including the gate of St Nicholas. These defences had been granted to the Order in H93.30 A very similar arrangement was made with the Hospitallers soon after the Third Crusade, and in 1291 all three Orders fought valiantly in defence of the city.31 Outside Acre, it has already been noted that the Teutonic Knights helped garrison Tripoli, whilst another section of this city's ramparts which was badly damaged during the siege of 1289 was held by the Hospitallers.32 Elsewhere along the coast, the Hospitallers shared the burden of defending Sidon, long before this lordship had passed out of baronial control, and guarded another tower at Jaffa, which was likewise held by an individual lord.33 In addition, the Teutonic Knights held two towers flanking Caesarea's town wall, and from 1229 onwards Frederick II relied on this Order to contribute to the defence

84

UNKNOWN CRUSADER

CASTLES

of Jerusalem.34 Even though they were royal cities, all three Military Orders also carried out repairs on the fortifications of Acre and Tyre when an imminent Mongol attack was feared in I26o.35 Such incidents highlight the importance of the Military Orders, and the extent to which even the most powerful secular rulers relied on their help. By the second half of the thirteenth century they were virtually alone in having the financial resources needed to maintain large fortifications. The fact that they often repaired or strengthened castles shortly after they had acquired them suggests that many secular lords could no longer prevent their fortifications from falling into disrepair. At Beaufort, for example, the Templars constructed a whole new citadel opposite the older castle in the space of just eight years, and at Arsuf urban fortifications erected by the Hospitallers soon after they took over the site were said to have enraged Baybars.36 Even after they had completed Saphet, the Templars spent 40,000 bezants each year on the upkeep of this fortress,37 whilst the vast new defences erected at Tortosa and Chastel Blanc from the mid twelfth century onwards, plus those built by the Hospitallers at Crac des Chevaliers and Margat until c. 1204, must have been just as expensive to maintain. The money required for such projects was largely drawn from the Orders' extensive estates in Europe and the Holy Land, and clauses in the Rule of the Templars suggest that it was then channelled into central funds specifically set aside for castles. Many strongholds, including Saphet, were also paid for by pilgrims and crusaders.38 Although there were greater signs of strain from the 12605 onwards, it also seems that the Military Orders often had more garrison troops at their disposal than the nobility, particularly in early thirteenth-century Syria. The remains of vast cisterns, undercrofts and storerooms at sites like Crac, combined with Willbrand of Oldenburg's claim that Margat contained enough supplies to withstand a five-year blockade, indeed suggest that the Orders had the resources to prepare their major Syrian castles for almost any emergency that might arise.39 Both the Templars and the Hospitallers also enforced strict rules regarding the use and defence of castle entrances to prevent spies and traitors from gaining access.40 This kind of meticulous planning and discipline, as well as the obstinate refusal of the Orders to give up the fight even as they were being driven out of Acre, suggests that the Hospitallers, Templars and Teutonic Knights were more skilled, disciplined and better equipped than any other Latin troops in the Holy Land. The forward planning and sheer commitment of the Military Orders can be contrasted with the inefficiency and poverty of many lay rulers or lords. In 1268, for example, troops and citizens from Antioch sheltering in the citadel were quickly forced to surrender to Baybars because, unlike Margat,

THE M I L I T A R Y O R D E R S

85

this stronghold was badly stocked with provisions and 'had neither enough water nor enough mills' to feed everybody.41 Eighty years earlier, Saladin had also been able to capture the secular fortress of Saone, situated along the land route between Latakia and Antioch, in a matter of three days, even though Margat, Tortosa, Chastel Blanc and Crac des Chevaliers had all held out. Saone's defences were just as impressive as (or perhaps even stronger than) those of Margat and its other neighbours to the south, but the fact that it was held by an individual lord seems to have contributed to the weak and demoralised state of its garrison.42 This incident proves that the successful defence of a fortress ultimately had more to do with the preparedness and mental state of its garrison that it did with the scale of its fortifications. It also shows how the Templars and the Hospitallers became virtually the only Franks to hold any inland castles in the county of Tripoli after the battle of Hattin. Indeed, it would not be an exaggeration to say that from 1188 onwards this crusader state owed its very existence to these two Orders. In the kingdom of Jerusalem the situation did not become quite so extreme until the mid thirteenth century, when entire lordships like Sidon were regularly handed over to the Orders. However, it would be misleading to assume that barons here were in a stronger position than their neighbours further north. The lords of Jaffa, for example, were only able to hold on to their castle by making regular appeals to the papacy for more money.43 We have also seen that during the early 12505 Louis IX refortified Jaffa whilst staying in the Holy Land. According to Joinville the papal legate Odo of Chàteauroux, who was helping the king with this project, spent a staggering 30,000 livres rebuilding just one particular gateway and adjoining curtain wall. This implies that Louis himself, who erected a further two gates and all the other ramparts, must have spent at least three times as much money; an amount which would have been well beyond the means of the lord of Jaffa himself.4* Similarly, it has been shown in recent years that, by the time Caesarea fell to the Muslims in 1265, this barony 'had disintegrated so far that the lord's influence in many parts of his own lordship must have been minimal'. Again, this was because many lands and properties had been sold off to the Military Orders or other institutions in order to raise revenues and pay off ever increasing bills.45 The pattern of baronial ownership in the east also confirms that the further inland castles lay, the more exposed they became to Muslim incursions. As a result, they were costlier to defend and more likely to get sold off to one of the Military Orders. Jaffa, Beirut and Caesarea were the only major strongholds in the kingdom of Jerusalem which were still in secular hands when they were captured from the Muslims. Likewise coastal Gibelet,

86

UNKNOWN CRUSADER CASTLES

Nephin and Maraclea were all baronial castles throughout the thirteenth century, whilst Bohemond VI kept Latakia for himself after he recaptured it in 1261. By contrast, the only inland stronghold held by the counts of Tripoli during this period was Akkar.46 It is also important to remember that virtually all the new defences built by individual lords during the thirteenth century were situated along the coast. These included Bohemond VI's towers at Latakia and (possibly) Jábala, plus Bartholomew's tower at Maraclea. Early in the thirteenth century John of Ibelin had also been granted Beirut, and subsequently carried out extensive improvements there. These probably concentrated on the building of a new outer wall, for although the castle itself no doubt needed repairs, there is no evidence that Saladin had demolished anything other than the urban defences of the town.47 Inland the only comparable building project undertaken by a secular lord was Odo of Montbéliard's completion of a new citadel at Tiberias, although it has already been noted that this too was a coastal fortress in the sense that it stood on the shores of Lake Tiberias.48 To conclude this chapter, it is interesting to ask whether fortifications belonging to the Military Orders ultimately helped or hindered the crusader states in their struggle with the Muslims. The famous military historian R. C. Smail wrote that central authority in the Holy Land 'was progressively weakened ... because lands, castles, powers and rights over men had continually to be conceded to Orders, which were not wholly part of the feudal structure'.4^ This certainly seems to have been the case in the county of Tripoli, where we have seen that the Templars and Hospitallers built or acquired massive fortifications which were then used to administer local lands and wage war with the Muslims with hardly any reference to outsiders. In short, these lordships became virtually independent enclaves. It has also been shown that both these Orders were prominent in some of the internecine struggles which took place during the thirteenth century, such as the Templars' defiant behaviour toward Frederick II at Acre and 'Atlit. It could certainly be argued that episodes of this kind undermined the power of local rulers and squandered military resources which could otherwise have been directed against the Muslims. But we have also seen that the Military Orders often played a vital role in the defence of the crusader states. The strength of castles like Margat and Tortosa prevented Saladin from wiping out the Latin East in 1188, and thereafter they ensured that the Holy Land survived for another hundred years. Ultimately, it seems fair to conclude that the fortifications of the Military Orders did far more to help the Christian cause than to hinder it.

PART TWO

The Kingdom of Cyprus 1191-1374

This page intentionally left blank

7

Warfare Thanks to its physical isolation, the kingdom of Cyprus was by far the most successful of all the crusader states in the eastern Mediterranean, enjoying long periods of peace and tranquillity at a time when most Latin territories on the mainland were gradually being reconquered by the Greeks, Muslims or Bulgars. This notwithstanding, two distinct types of warfare threatened to undermine Cypriot stability from time to time. First, there was always the danger that external forces would attack the island. Most incursions of this kind were carried out by pirates, who appeared sporadically from the 11903 onwards, and were particularly active during the mid fourteenth century. Other, more politically motivated raids were also organised by the Muslims of Egypt in 1220 and 1271, by the Genoese in the early fourteenth century, and by the Turks in the 13605. These were all temporary attacks, however, and although they could be damaging on a local scale, they were not as alarming as the threat of a full-scale foreign invasion.1 It was of course just such an attack which had brought Cyprus under Latin rule in the first place, for in the summer of 1191 King Richard I of England landed at Limassol whilst travelling east to join the Third Crusade besieging Acre. After he arrived, he ousted the self-declared emperor Isaac Komnenos, who had broken away from Byzantine authority in the mid iiSos. Richard probably only saw Cyprus as a means of financing his campaign in the Holy Land, and once he reached Acre he sold it first to the Templars, then to Guy of Lusignan, the former king of Jerusalem, who received the island as compensation for losing his old title to his rival Conrad of Montferrat. As a result, Cyprus came under the control of the Lusignans, who also completed the task of suppressing the native population, encouraging more Prankish settlers and setting up a Latin church.2 After these dramatic changes, the likeliest threat of a new external invasion came from the Byzantine empire, which had already made an unsuccessful attempt to reconquer Cyprus back from Isaac Komnenos in 1187. This possibility was quickly eradicated by the political turmoil which afflicted the empire both before and during the Fourth Crusade, not emerging again as a realistic threat until the reign of the Nicaean emperor Michael VIII

90

UNKNOWN CRUSADER CASTLES

Palaiologos, who recaptured Constantinople from the Latins in 1261 and was said to have been planning an attack on Cyprus two years later. In the long run these plans came to nothing, for the Byzantine authorities were always far too busy dealing with enemies closer to home, and in particular the Turks.3 By the late thirteenth century the Lusignans were less concerned with the Greeks than events in the Holy Land, where Prankish territories which had for so long acted as a buffer zone between Cyprus and the Muslims were rapidly being overrun by the Mamluks of Egypt. This process reached a crisis point in 1291, when the last Christian outposts on the mainland were lost, and the ruler of Egypt, al-Ashraf Khalil, was said to have prepared a huge fleet with which to invade Cyprus. In response, Pope Nicholas IV hastily sent galleys to help defend the island, which was only saved by the sudden assassination of the sultan and a subsequent power struggle in Cairo.4 A combination of similar political crises, various economic setbacks and inadequate Muslim naval forces, which had proved so helpful for the Latins during the twelfth and thirteenth centuries, proved equally beneficial to Cyprus after 1291, and the island was not in fact attacked by the Mamluks until the early fifteenth century, and not conquered permanently until the Ottoman invasion of 1570-71.5 The events of 1291 also brought Cyprus to the attention of another potential invader, for the fall of Acre and Tyre encouraged the Italian city states to transfer their business activities, as well as their almost constant struggles for commercial dominance, from the Levantine coast to Famagusta. During most of these clashes the Cypriot kings tended to favour the Venetians, for it was felt that Genoa was abusing its privileges in order to cover up the corrupt activities of its merchants. The situation deteriorated steadily in the course of the fourteenth century, resulting in Genoese raids on the coast around Paphos in 1312 and 1316, as well as a more serious incursion in 1373. This latter attack was merely a prelude to a full-scale invasion later that year intended to give Genoa total control over Cyprus, and with it all trade in the eastern Mediterranean. In the end it failed to achieve its objective, for although the Genoese captured Famagusta, they could not take the coastal stronghold of Kyrenia and only occupied Nicosia temporarily. They still managed to retain Famagusta until 1464, and therefore posed at least as great a threat as Egypt to the Lusignan monarchs of the fourteenth and fifteenth centuries.6 The second category of warfare which affected Cyprus during this period involved internal rather than external aggressors. Clashes of this kind were instigated either by the native Greeks, who rebelled twice shortly after Richard I's invasion and periodically threatened to do so again,7 or by

WARFARE

91

disaffected members of the Latin nobility, who challenged royal authority on a number of occasions. The first and most serious conflict of this latter kind can be traced to the death of King Hugh I in 1218, for Hugh's son, the future Henry I, was only eight months old at the time. As a result, a power struggle over who should act as Henry's regent quickly developed between the powerful local Ibelin family and those nobles who supported the claims of the German Emperor Frederick II, the suzerain of Cyprus.8 During Frederick's crusade to the Holy Land between 1228 and 1229, the emperor's personal presence in the East ensured that his five representatives (or baillis) on Cyprus gained the upper hand in this conflict. Their leading rival John of Ibelin was forced to submit to imperial authority after attempting to defy Frederick briefly by sheltering inside the mountain fortress of St Hilarion.9 Once the emperor had departed for Europe in May 1229, however, the Ibelins quickly defeated the five baillis outside Nicosia, and eventually regained control over Cyprus after besieging their enemies for almost a year inside St Hilarión and the neighbouring castle of Kantara.10 In the autumn of 1231 Frederick II responded to these setbacks by sending Lombard troops east under his marshal, Richard Filangieri, who attacked Ibelin-held Beirut on the mainland and therefore enabled the imperialists to regain the initiative on Cyprus. These successes were short-lived, for Filangieri failed to capture Beirut and retreated to Tyre, giving John of Ibelin the chance to defeat the imperialists once and for all at the battle of Agridi, near Nicosia, in June 1232. This encounter effectively ended Hohenstaufen rule on Cyprus, although many supporters of Frederick II retreated to the fortress of Kyrenia, where they continued to hold out for a further twelve months before hostilities finally came to a close.11 After 1233 there were only a couple of other serious internal disputes involving any bloodshed. In 1306 King Henry II (1285-1324) was deposed by his brother Amaury of Tyre, who ruled as 'governor' until Henry's return from exile four years later. Although Amaury was eventually murdered because of his increasingly brutal regime, he had initially enjoyed widespread support from the nobility, many of whom apparently felt that Henry II had not done enough to defend Cyprus against the Mamluk threat or to try to regain the Holy Land after the disasters of 1291.12 But after the violent death of Amaury and the triumphant return of Henry in 1310, some fifty years elapsed before the final serious internal crisis of the period emerged during the reign of Peter I (1359-69). Ironically, Peter's popularity amongst the nobility disintegrated because he did too much rather than too little. This ruler spent most of his reign undertaking ambitious naval offensives against the Muslims of Asia Minor and Egypt. Although these campaigns aided

MAP 3. Cyprus

WARFARE

93

Cyprus in the sense that they disrupted Muslim trade and may therefore have brought a temporary halt to the mid fourteenth-century economic decline of Famagusta, Peter paid for them by imposing crippling taxes on his own subjects. As a result he was eventually assassinated in 1369 by a faction of Cypriot nobles, who then placed his young son Peter II (1369-82) on the throne. This crisis had still not been fully resolved at the time of the Genoese invasion in 1373, when Peter's widow Eleanor of Aragon initially allied herself with the invaders against his brothers John and James of Lusignan, who had almost certainly been involved in Peter's assassination.13 These then were the principal clashes which threatened to undermine Cypriot stability, and it will be shown that virtually all castles built or occupied on the island during this period were intended to deal with at least one of them. Some types of warfare, however, were likelier to break out than others. The Cypriot uprisings of 1191 and 1192, for example, indicate that the threat of internal rebellion was at its greatest during the earliest years of Prankish rule. On the other hand, the wide chronological distribution of clashes involving fellow westerners, which broke out in 1228, 1306 and 1369, suggests that the risk of such confrontations did not alter dramatically one way or the other. Much the same can be said for attacks by pirates, although raids of this kind did increase markedly during the middle years of the fourteenth century. The most dramatic change of all in Cypriot fortunes clearly took place after 1291, when the threat of a major external invasion, either by the Mamluks or later by the Genoese, grew enormously. The risk of such an attack had barely existed during the first hundred years of Prankish rule, particularly once the Fourth Crusade had eliminated any chance of a Byzantine counter-offensive. After 1291, the Lusignans had to face the possibility of an attack by the kind of vast Mamluk armies which had wiped out Latin territories on the mainland. Indeed, it has already been noted that from the reign of Baybars (1260-77) onwards Egyptian invasion forces in the Holy Land had rarely numbered fewer than 12,000 men. At this time Mamluk armies also improved dramatically in terms of quality as well as quantity, for they increasingly relied on professional soldiers, and in particular highly skilled siege engineers. These men had at their disposal far larger and more accurate catapults (or trebuchets) than their twelfth-century predecessors, and by using these powerful weapons, in combination with mines and ladders, they took much of the credit for reducing even the greatest Prankish strongpoints on the mainland.14 By the late fourteenth century the Genoese were scarcely less intimidating, particularly because of their great naval strength, the only significant weakness in the Mamluk war machine. According to one source, the invasion fleet of 1373 was made up of 'forty-three galleys plus many

94

UNKNOWN CRUSADER CASTLES

other vessels', and 'there were brought together in the Genoese army 14,000 troops and more ...'15 These impressive totals can be contrasted with the kind of troop numbers available to the kings of Cyprus. Although exact figures are hard to establish, it seems that the 200 knights and 500 infantry brought to Acre by Henry II in 1291 were roughly equivalent to the maximum field army which could be raised without stripping Cypriot castles of their core garrisons.16 These forces would have been adequate for dealing with pirates who had landed on Cyprus. They would also have sufficed against any possible uprisings by the local population, for accounts of the Greek rebellions which took place in 1191 and 1192 suggest that their participants were badly organised and lacking in sophisticated weaponry such as siege engines or heavy cavalry troops.17 It is also clear that internal clashes between rival western factions normally involved just a few hundred troops. At the battle of Agridi (1232), for example, the Ibelins only had 233 mounted troops. Although they were supposedly confronted by 2000 imperialist horsemen, this latter figure may have been an exaggerated approximation given by a source favourable to the Ibelins in order to play up the scale of their victory.18 After 1291 the Lusignan kings suddenly found themselves having to deal with a potential invasion by more than 10,000 Mamluk (or later Genoese) soldiers. Inevitably, these factors had a profound effect on castles. Although the Franks built new fortifications on Cyprus before 1291, these were not particularly extensive, and added little to the already existing network of Byzantine strongpoints. After the fall of Acre, however, a massive building programme, later renewed in response to the growing Genoese threat, resulted in the construction of many new fortifications and urban defences. This process continued well into the sixteenth century, for from 1489 onwards Cyprus came under the control of the Venetians, who upgraded many medieval strongpoints with new ramparts designed to withstand the artillery of the Ottoman Turks.19 As far as Cypriot fortifications were concerned, 1291 was actually a far more significant date than 1191.

8

Military Architecture During the crusader period, the key fortifications on Cyprus were to be found at Paphos, Kyrenia, Limassol and Famagusta, situated on the coast, at Nicosia, the inland capital of the kingdom, and at St Hilarión, Buffavento and Kantara, three mountain strongholds in the north. This unusually small total, plus the fact that Cyprus often remained at peace for many decades at a time, makes it relatively easy to trace the few changes in military architecture which took place there between 1191 and 1374. However, thanks to the major alterations carried out by the Venetians after they gained control over Cyprus in 1489, it is no longer always possible to study the actual physical remains of these developments. As a result, a combination of both archaeological and historical sources must be used in order to assess what the island's defences were like before the fifteenth century. Inevitably, this applies most to the very earliest period of Prankish rule, for at several sites it is only by studying contemporary descriptions of Richard I's invasion that we can ascertain what fortifications existed there in the late twelfth century. This is certainly the case at Limassol, which was the first Cypriot settlement occupied by Richard in May 1191. It seems that the king's initial intention in stopping here had merely been to recover some sailors shipwrecked nearby in a violent storm. However, Isaac Komnenos's refusal to free these people, combined with Richard's realisation that Cyprus could prove a useful source of wealth and supplies, quickly led to confrontation between the two men, and on 6 May the English defeated the Greeks along the shore and occupied their castle. These events prove that a Byzantine stronghold existed at Limassol before 1191, although the apparent ease with which the English took control of it suggests that it cannot have been particularly large.1 After the fall of Limassol, Isaac Komnenos fled north and east away from Richard's advancing troops. Along the way he sought refuge briefly at Nicosia, suggesting that it too was defended by some sort of strongpoint; a conclusion which can be confirmed by looking at the Cypriot rebellion of 1192, when the Templar garrison of the city also took shelter there.2 Thereafter Isaac retreated to the castle of Famagusta, but he quickly realised that it 'would not be safe to withstand a siege', and so he 'hid himself in

96

UNKNOWN CRUSADER

CASTLES

the woods'.3 Meanwhile, as Isaac's resistance crumbled in the east, Richard sent another contingent to occupy 'the castle which is called Paphos' in the west, implying that its defenders did not even attempt to put up any resistance against their Latin opponents.4 Consequently it seems that the citadels at Nicosia, Famagusta and Paphos were all so small that the emperor had no confidence in them as a means of protecting himself, let alone halting Richard's advance. To some extent the lack of archaeological evidence at these sites appears to strengthen this theory. The absence of any such remains within the large urban settlements of Nicosia and Famagusta is understandable. But it is surprising that the extensive and recent excavations carried out at Paphos have failed to uncover any fortifications which can definitely be attributed to the twelfth century. This investigation has concentrated on a classical and medieval site located just inland from the town. Although it has led to the discovery of major defences from the early Prankish period, there are few signs of Greek occupation after a much older Byzantine strongpoint was destroyed by the Arabs in 654 AD. 5 An alternative theory is that the castle of Paphos occupied by Richard I's troops should not be sought so far inland, but was actually situated next to the harbour, where there may have been 'at least one [fort] in the Byzantine period'.6 However, this possibility can only be explored further by excavating the much later Prankish, Venetian and Ottoman defences located in this area.7 A similar search inside the present ruins of Limassol Castle, the oldest visible parts of which have been dated to the thirteenth century, might likewise reveal traces of a Byzantine predecessor, and could perhaps also establish whether there are any links between this structure and the small fort or fortified church said to have been occupied by Richard I's shipwrecked sailors in the spring of iipi.8 It seems accurate to conclude that the Byzantine castles at all the sites mentioned so far were small, weak and perhaps even dilapidated. Indeed, according to one source the citadel at Nicosia was subsequently 'destroyed and demolished' by the Franks and replaced by a church.9 Not all Byzantine strongholds on Cyprus were treated in the same way, for it is clear that the much larger castles of Kyrenia, St Hilarión, Buffavento and Kantara were not demolished but were preserved intact, or even enlarged and strengthened. At Kyrenia, for example, the Franks had retained the original Byzantine layout of the fortress, which consisted of a substantial rectangular enclosure situated on a coastal promontory and defended by four flanking corner towers. At the time of Richard I's invasion these towers were relatively small, but along the more exposed landward rampart to the south they were strengthened by the presence of an additional horseshoe

Kyrenia: the fortress and walled town.

98

U N K N O W N CRUSADER CASTLES

tower, and beyond it an outer wall flanked by several pentagonal salients. This was a far larger strongpoint than the Byzantine defences at Famagusta, Limassol, Nicosia and Paphos, and the archaeological evidence suggests that it subsequently underwent relatively few alterations until the late thirteenth century.10 Kyrenia also differed from its neighbours in that the settlement next to the castle probably already had some sort of urban defences at this time. In 1212 the German pilgrim Willbrand of Oldenburg described Kyrenia as small but fortified, whilst accounts of the castle's surrender by the Lombards to the Ibelins in 1233 also imply that it was fortified.11 It is possible that two towers incorporated into the present urban defences, built at a much later stage by the Venetians, contain fourteenth- or even thirteenth-century masonry and are the successors of an even older circuit wall dating back to Byzantine times. Hence the alternative theory that Kyrenia did not gain any town walls until after 1300 should probably be rejected.12 To the south of Kyrenia, the three mountain fortresses of Buffavento, St Hilarión and Kantara were likewise left almost completely unaltered by the Franks during the thirteenth century. Perhaps the most famous of these was St Hilarion, situated on an isolated peak overlooking the strategic valley which carried the road between Kyrenia and Nicosia. Its defences were divided into three baileys along the castle's eastern side, for its north, west and south slopes were so steep that they barely required any man-made fortifications at all. The lower and middle baileys contained various residential quarters and service buildings, whilst beyond them a steep path gave access to the upper bailey housing further royal apartments. To the east of and even higher than this level, two towers also occupied the ridge of the mountain top itself, and one of these in particular probably acted as a final refuge for the castle garrison.^ Virtually all these buildings, and especially those of the lower two baileys, are Byzantine constructions, or at least have Byzantine foundations. Indeed, apart from repairs and minor alterations, the only significant Prankish addition to the castle seems to have been the thirteenth-century hall located in the upper bailey.14 This was a purely domestic structure which added nothing to the military strength of the site. It may have been from near here that the Constable John of Antioch, who guarded St Hilarion during the Genoese attack on neighbouring Kyrenia in the spring of 1374, threw the Bulgarian members of his garrison to their deaths. John is supposed to have done this because he thought the Bulgars had been instructed to kill him by Eleanor of Aragon, who believed that he had been responsible for the death of her husband Peter I. While it does at least illustrate the great height of the castle, this incident may well be fictitious.15

St Hilarión

100

UNKNOWN CRUSADER CASTLES

The appearance of St Hilarión is very similar to that of Buffavento, situated on another mountain top nearby. Again, this summit did not need complete circuit walls, its two baileys being located on the castle's slightly less precipitous southern side. The lower bailey contained a series of buildings, occupying a long, narrow plateau, and was presumably used to store supplies and to house the garrison. Dominating this area, and rising more than twenty-five metres above it, were the structures of the upper bailey, which, as at St Hilarión, probably served as royal apartments. To the west of these apartments, and separated from them by a deep cleft in the rock, stood a remote tower which could only be reached via a wooden drawbridge. Despite later repairs, it seems that these structures should also be attributed to Byzantine rather than Prankish architects.16 Buffavento also shares many characteristics with Kantara, the third of the great mountain fortresses. Even though the site occupied by this castle is very remote, it is not quite as steep as those of its two neighbours. As a result, it is composed of a single and fully enclosed bailey occupying a reasonably flat plateau, rather than a series of two or more staggered levels. The most elaborate defences of this bailey were placed to the south and south east, where it was slightly easier for an attacker to approach. These defences included two successive gateways, each flanked by a pair of horseshoe towers, plus a prolific number of arrowslits guarding the southern slope. Inside Kantara's ramparts, and at its highest point, there also stood a small rectangular keep with its own arrowslits and drawbridge, which can clearly be compared with the isolated towers already mentioned at St Hilarión and Buffavento. This entire layout again appears to be Byzantine in origin, although the present ruins of Kantara actually contain far more fourteenth-century repairs than either of its neighbours.17 The impressive size and strength of all three mountain castles can be highlighted further by looking at the way in which they were captured from the Greeks in 1191. Toward the end of his campaign on Cyprus Richard I fell ill briefly at Nicosia, so Guy of Lusignan, who had recently arrived from the mainland in order to ally himself with the English king, began to lay siege to St Hilarión. However, Guy failed to make any progress. Instead both this stronghold and Buffavento appear to have surrendered because the men guarding them felt little real loyalty toward Isaac Komnenos.18 By this point Isaac is said to have relied on brute force to keep his subjects in check, so that 'the natives regarded him with great hatred', and he 'was tolerated rather than loved' even by his own troops.19 This explains why Richard and Guy found it so easy to capture these strongholds even though one source wrote that they 'could never have been captured by the machines of any enemy, unless by deceit or famine'.2« No doubt similar circumstances

MILITARY ARCHITECTURE

IDI

Kantara

led to the surrender of Kantara soon afterwards, as well as that of Kyrenia, which was occupied without a struggle by Guy of Lusignan. After the loss of these four castles, Isaac finally gave up and was taken prisoner by Richard.21 At Kyrenia, St Hilarión, Buffavento and Kantara, therefore, there were extensive and well-maintained Byzantine fortifications already in existence, and these were subsequently retained by the Franks. Elsewhere, it is clear that a new building programme began shortly after 1191, to replace the much smaller Byzantine castles of Limassol, Nicosia, Paphos and Famagusta. At Nicosia, for example, Willbrand of Oldenburg referred to the construction of a new citadel in 1212, whilst at Limassol, Guy of Lusignan is said to have established another stronghold as early as ii93.22 Presumably this structure was a predecessor to the approximately square two-storey tower located at the heart of Limassol's present castle. About a hundred years ago the architectural scholar Camille Enlart dated this tower to the thirteenth century, noting that attached to its east wall there had also once stood a small hall or chapel.23 It is possible that this was the same church or fort used as a temporary refuge by Richard I's sailors in 1191.24 Intriguing archaeological remains of towers similar to that at Limassol have also been unearthed at both Paphos and Famagusta. At the former site, recent excavations have revealed 'a walled encampment with a watchtower' above the level of the much older Byzantine stronghold destroyed in 654 AD. Again, this may have been built by Guy of Lusignan, and contained Latin coins dating from the first ten or fifteen years of Prankish rule.25

102

UNKNOWN CRUSADER CASTLES

Similarly, at Famagusta one corner tower of the late thirteenth-century citadel, which stood near the harbour entrance, appears to be a much older, isolated building subsequently incorporated into the surrounding defences. This structure has tentatively been attributed to Guy of Lusignan and must have been the 'sea tower' captured by Ibelin troops in 1232, shortly after they reoccupied Famagusta in the wake of Richard Filangieri's failed siege of Beirut.26 The ease with which the Ibelins entered Famagusta before they surrounded the 'sea tower' also proves that at that time this was an open city, lacking any form of urban defences. Similar conclusions can probably be reached about Limassol, Paphos and Nicosia, which were, according to Willbrand of Oldenburg, 'not well fortified' in I2i2.27 Indeed, Enlart's efforts to find traces of any such fortifications at Limassol, which one might expect could still have existed in the late nineteenth century, were completely unsuccessful.28 Hence it is clear that from 1192 onwards the Franks were happy to provide these settlements with relatively simple defences, which, judging by the remains at Limassol, Paphos and Famagusta, amounted to little more than sturdy two-storey towers surrounded by outer enclosures. The date of Willbrand of Oldenburg's visit to Cyprus proves that these towers were probably completed by King Hugh I (1205-1218), and had been started by his uncle Guy of Lusignan (1192-94) or his father Aimery (1194-1205). The apparent lack of earlier Greek remains and (if it can be trusted) the written evidence claiming that the Franks demolished the old castle at Nicosia suggest that the Byzantine predecessors to these towers were considered too small or dilapidated to bother retaining after 1192. Following the completion of these towers, there were a number of other important construction phases in the course of the thirteenth century. The earliest of these occurred at Paphos, where Guy of Lusignan's watchtower was quickly replaced by a much more substantial concentric fortress. This stronghold consisted of an almost square inner castle, built around a central courtyard and defended by four corner towers, as well as an outer curtain flanked by a further eight towers. In addition, it housed workshops, stables, a chapel, soldiers' barracks and luxurious residential apartments. In short, this was a far larger structure than the tower it replaced, although it was never rebuilt after being flattened by a devastating earthquake in 1222.29 After the destruction of Paphos castle, there were no further efforts to improve Cypriot fortifications until the reign of Henry II (1285-1324). Henry is said to have begun 'the castle wall of Lefkosia' (Nicosia), which was presumably intended to strengthen the early thirteenth-century tower there.30 Perhaps this statement also implies that Nicosia had limited urban defences at this time. More substantial alterations took place at both Kyrenia and

MILITARY ARCHITECTURE

1O3

Famagusta towards the very end of the thirteenth century. At Famagusta the older Prankish sea tower was converted into one of four corner towers attached to a new rectangular citadel, which was equipped with arrowslits, a postern and a central courtyard surrounded by numerous vaulted buildings. This fortress may have resembled the citadel of Tyre, which was constructed by King John of Brienne around 1212. In addition, town walls were built for the first time, then hastily extended during Amaury of Tyre's brief rule over Cyprus.31 By the time Amaury was murdered and Henry returned from exile in 1310, these must have been virtually complete, for in that year the citizens of Famagusta declared themselves staunchly on the king's side, walled up the town gates and demolished their drawbridges to prevent Amaury's supporters from regaining control of the city.32 During the same period extensive improvements were also made at Kyrenia. The north and east curtain walls of the original Byzantine castle were replaced by new defences built with ashlar masonry similar to that used in thirteenth-century Syria. These structures incorporated two floors of shooting galleries, with a rampart above, whose crenellated parapet was still perfectly intact when Enlart saw it around a century ago. To the south, the Franks retained some of the Byzantine fortifications, but again strengthened them with a new curtain wall and shooting gallery. There are further traces of Prankish works along the west curtain of the castle, including the L-shaped gateway, which was built using massive blocks of masonry and was shielded by the castle's north-west tower. This tower also had a postern gate facing the sea, and originally had a barbican attached to it which defended the main gate against potential attacks from the town.33 Over the next fifty years, the new defences at Kyrenia and Famagusta were no doubt maintained and perhaps even extended, for by the mid 13405 an anonymous traveller wrote that the latter was 'a city strongly built ... on the rock. It is surrounded by deep and broad ditches cut out of the rock, and has high walls and towers subtly built of squared and cut stone'.34 This description is very similar to that given by another pilgrim about fifty years later.35 Moreover, when the Genoese invasion force of 1373 arrived at Famagusta, a number of inconclusive skirmishes broke out outside the city before the attackers suggested that the citadel should be evacuated, and that seventeen men from each side should meet there to discuss a truce. Peter II (1369-82) agreed to this, but once the Genoese negotiators had entered the citadel, they overpowered their Cypriot counterparts, let more troops in, and eventually occupied the whole city. The use of such a stratagem by the Genoese implies that by this date Famagusta's defences were indeed very strong.36 This impression may be misleading, however, for shortly afterwards the

104

UNKNOWN CRUSADER CASTLES

Genoese admiral Peter Campofregoso consulted his men and 'gave orders that they should raise the height of the walls wherever they were low, and he tried to bring the sea all around the place [i.e. the citadel] so as to make it an island'.37 In 1380, during one of the many clashes between Genoa and Venice, wooden towers were also added to Famagusta's sea walls in order to make them taller.38 Today, it is only possible to identify one tower along the city's more recent Venetian ramparts which seems to date from this period, and it may therefore represent the sole visible architectural evidence of new Genoese defences at Famagusta during the late fourteenth century.39 While these changes were taking place at Famagusta, the Franks were also busy constructing new urban defences at Nicosia. Some earlier fortifications of this kind may have been erected under Henry II after 1291, but more substantial construction work began during the reign of Peter I (1359-69).4° This had clearly not been completed properly before the accession of his son Peter II (1369-82), who found that 'the walls were very strong; but they were low, and he sent word to the country round, and men came together and built them up with earth and stones; and they dug out the ditch and constructed 133 platforms to fight from in addition to the towers'.41 These improvements were continued during the late fourteenth century by Peter II's successor, James I (i382-98).42 The overall sequence of events at Nicosia may have been that sporadic efforts were made to build city walls between 1291 and 1359, but because of the city's inland location this project was not pursued as vigorously as the more urgent strengthening of exposed coastal sites like Kyrenia, and was not finally completed until the period of growing Genoese aggression and Turkish raiding after the 13505. Nicosia's citadel also underwent many radical changes during this period. By the 13505 it is difficult to ascertain what state the original castle built between 1192 and 1212 (and possibly extended by Henry II) was in. If this stronghold still existed, it may have been demolished by Peter I, who was responsible for establishing a new citadel on a hill on the outskirts of the city. This replacement basically consisted of a 'strong and impressive' tower, whose different floors, including a dungeon in the cellar, were connected by ladders. It was also surrounded by a large moat, which had been excavated by slaves who 'dug the earth all day, and carried it out on their backs'.43 The Margarita Tower, as this citadel was known, never even reached completion before Peter II ordered it to be demolished, along with several other buildings which were in the way. A new, larger castle replaced it. This fortress lay near the Paphos gate, in the west of the city, and must have been far larger than its predecessor, for it contained royal apartments. It was built by Genoese prisoners of war in the space of just ten months, using recycled masonry from the Margarita Tower and 'walls in the town which were of

MILITARY ARCHITECTURE

105

no use'.44 Its ramparts were subsequently strengthened by James I, who 'made it a famous work', and by his successor Janus (1398-1432).45 James I's building activities were not limited to Nicosia, for he also undertook extensive repairs at several other sites in the course of the 13905. At Kantara, for example, many of the defences may owe their present appearance to James,46 whilst at Limassol the internal layout of the early Prankish keep was drastically altered, and the hall or chapel attached to it was replaced by a new structure and a second tower.47 In addition, James established a new castle at Sigouri, on the coastal plain near Famagusta, and rebuilt the old stronghold of La Cava, overlooking the main route from Nicosia to Larnaca. Sigouri was a classic example of a castrum, consisting of an almost square enclosure flanked by four small corner towers and surrounded by a water-filled moat.48 La Cava, on the other hand, occupied an elevated plateau, and its defences were therefore less regular, simply incorporating two massive towers linked by a curtain wall.49 Some chroniclers also claim that James built two new towers near the harbour of Paphos,50 but he should probably only be credited with repairing these structures rather than building them from scratch, for in 1373 they had already been captured by the Genoese.51 It may therefore prove impossible to date these towers without detailed excavation, although the failure of the sources to mention them during an earlier Genoese raid on Paphos in 1316 does at least suggest that they were completed after that date.52 After the frantic activity of King James's reign, another inactive period followed until the late fifteenth century, when the Venetians undertook the last great construction programme on Cyprus before the Ottoman invasion of 1570-71. At this time the tall and relatively thin curtain walls of the medieval period gave way to the extremely thick and low defences introduced after the arrival of gunpowder. Such fortifications were similar to those at Famagusta, where the Venetians encased or replaced both Henry II's late thirteenth-century citadel and the medieval city walls with much sturdier ramparts.53 Similar additions were also made to the citadel at Limassol during this period, whilst at Nicosia, whose location in a valley rendered it extremely vulnerable to artillery attack, new urban fortifications completely replaced those of the fourteenth century.54 At Kyrenia, meanwhile, many of the Prankish defences, particularly along the castle's south and west sides, were buried under new ramparts, and a vast artillery bastion added to the south-west corner.55 In addition, there are traces of Venetian work to be found at Paphos,56 whilst the three mountain castles of Buffavento, St Hilarión and Kantara were also dramatically altered in a quite different sense, for they were all demolished at this time in order to prevent their falling into enemy hands.57

1O6

U N K N O W N CRUSADER CASTLES

Major alterations have, unfortunately, made it impossible to study the actual appearance of many of these defences, but enough historical and archaeological evidence nevertheless survives to enable us to draw some general conclusions about military architecture on Cyprus between 1191 and 1374. First, it is clear that, at Kantara, Buffavento and St Hilarión, the Franks were happy to occupy and carry out relatively few changes to older Byzantine fortifications. Both the design and the history of these sites can be compared with similar Greek fortresses held by Latins elsewhere, such as Arcadia in the Peloponnese. After 1204 the Latins often lacked the troops and resources needed to subjugate the Peloponnese (or indeed the Byzantine empire as a whole), and would certainly have been unable to construct large numbers of new castles from scratch. As a result, they were more than happy to garrison Byzantine or even classical sites with their own soldiers, and no doubt a similar situation prevailed on Cyprus immediately after 1191. It is interesting to note that the Franks believed Arcadia had been built by 'giants', because of the enormous classical blocks used in its construction,58 whilst St Hilarión was thought to date from the time of 'the gods and the pagans'. These traditions reflect the great age of such strongholds, and suggest that the Franks were grateful to inherit powerful mountain fortresses whose original construction must have seemed like an almost superhuman achievement.59 The fact that the Lusignans apparently chose not to retain the existing Byzantine twelfth-century defences at Paphos, Famagusta, Limassol and Nicosia seems to confirm that these structures were much smaller and weaker than Kyrenia or the mountain castles. In order to replace them, the Franks adopted another tactic which was also used widely outside Cyprus: the construction of square towers standing either in isolation or surrounded by an outer circuit wall. Such structures have already been referred to at numerous sites in the Holy Land, including Qaqun and Chastel Rouge, and they were also common in Prankish Greece. Some medieval towers in Greece, most notably Kalamata, Arcadia, Corinth and Neopatras, were even built within much older Byzantine castles.60 They therefore provide a direct parallel with their Cypriot counterparts, which presumably stood on or near the original Greek strongpoints captured by Richard I in 1191, and may well have been constructed using masonry taken from them. Many other Cypriot fortifications built or altered by the Franks shared certain characteristics with strongpoints on the mainland. The concentric castle at Paphos, for example, has been compared with the twelfth-century Hospitaller fortress of Belvoir, although this does not necessarily make it a Hospitaller construction.61 Once the extensive building programme undertaken by Henry II had been completed, the citadels at Famagusta and

MILITARY ARCHITECTURE

lOJ

Kyrenia must also have had more in common with Syrian strongpoints than they did with their immediate neighbours on Cyprus. Most notably, the elaborate gateway, new towers and extensive shooting galleries which Henry built at Kyrenia, using high quality ashlar masonry, must have turned this fortress into a Cypriot version of 'Atlit or Crac des Chevaliers. However, the fact that the walls of both Nicosia and Famagusta later had to be heightened using earth, stones or even wooden stockades suggests that none of the urban fortifications on Cyprus ever attained quite the same strength as the vast multiple ramparts of cities like Acre and Tyre.62 Finally, and on a smaller scale, the castrum castle of Sigouri can be compared with that of Goliath in the county of Tripoli. Goliath was probably built by the Hospitallers after they acquired the site in 1127, whereas Sigouri was constructed by James I during the i39os.63 This confirms the suggestion that military architecture evolved according to financial and geographical necessity rather than scientific discovery. It has been written that 'the castrum form of the castle was exclusive to the twelfth century'.64 But the construction of Sigouri almost two centuries later indicates that it would be wrong to assume that this extremely simple design was immediately rejected by the Latins as soon as they started to build more complex castles such as 'Atlit. In reality Sigouri's design was eminently suited to its exposed surroundings, whilst its size must have made it large enough to withstand minor Genoese incursions, yet small enough to be affordable for a relatively poor ruler like James I. Cypriot fortifications changed according to specific military, political or financial needs rather than any clearly defined architectural discoveries.

9

Fortifications and Internal Security Having looked at the history and appearance of fortifications on Cyprus during this period, we can now turn to the various functions which these structures were expected to fulfil. It is logical to begin such a discussion by looking at the question of internal security, a primary concern for Latin newcomers in the wake of Richard I's conquest. Almost as soon as Richard left Cyprus and sailed to Acre, the royal officials he had left behind had to deal with a popular uprising whose leader appears to have been one of Isaac Komnenos's relatives. Although he was eventually tracked down and hanged by the English and a modicum of order was restored,1 large parts of the island may have remained out of control, for in 1192, shortly after they had bought Cyprus from Richard I, the Templars were confronted by another rebellion orchestrated by local people living in and around Nicosia. The Templar garrison of the city, which consisted of fourteen knights, another twenty-nine mounted troops and seventy-four foot soldiers, was therefore forced to retire into the Byzantine castle before making a desperate sortie which succeeded in dispersing the hostile crowds outside.2 Thereafter the Templars, whose desire merely 'to rob, beat and ill-treat' local people can only have enraged the population even further, decided that they lacked the resources to control Cyprus, and returned it to Richard I.3 When he acquired it shortly afterwards, Guy of Lusignan may have found the island in a state of virtual anarchy. By this point some of the Greek noble families on Cyprus had probably fled. In order to bring the anarchy to an end and obtain land which could be granted to his followers, Guy confiscated the estates of virtually all those who still remained. This may have done much to stabilise the situation and reduce the threat of future rebellions, for it appears that the uprisings of 1191 and 1192 had been organised by Greek lords who had never come to terms with either Richard Lionheart or the Templars.4 But Cyprus, whose population at this time has been estimated at around 100,000, must nevertheless have been a dangerous place for the first Latin settlers arriving there during the ngos.5 Therefore, Guy's primary objectives in 1192 were to deter the Greeks from taking up arms again, and, failing this, to protect newcomers should any violence

F O R T I F I C A T I O N S AND I N T E R N A L S E C U R I T Y

1O9

break out. Clearly, the new Prankish towers erected at Limassol, Paphos, Famagusta and Nicosia between 1192 and 1212 were intended to achieve this goal. These structures would have been perfectly suited for the task, for they were sturdy enough to withstand attacking rebels, but small enough to be built quickly and economically, and to be garrisoned by relatively few men. Larger strongholds would have been a waste of resources against an enemy lacking any kind of siege weapons or organised structure. This incidentally suggests that the Byzantine predecessors to these towers really had been dilapidated, if they could not even be relied on to keep out crowds of rioters. Lastly, the fact that (with the probable exception of Kyrenia) no urban fortifications existed on Cyprus prior to 1291 also tends to confirm that at this stage the Latins were far more concerned about internal than external security. Returning to the Prankish towers themselves, further evidence of their intended function may be gleaned from their location, for with the exception of Paphos they all appear to have been situated at the centre of their respective settlements. Although it dates from the late sixteenth century and may therefore be unreliable, one source even claims that at Nicosia the Byzantine castle stood near the market place. If the Franks built their new tower on or near such a central site it must have dominated the entire city.6 It would have become a powerful symbol of the new regime, acting as a constant reminder to the Greeks that resistance against their new overlords was futile. Hence the function as well as the appearance of these structures closely resembled that of other Prankish towers on the mainland, where many small fortifications of this type, dotted across Syria, Palestine and later Greece, protected Latin settlers and their agricultural interests against local criminals or rebels.7 The subsequent history of Cyprus can be used to highlight the success of the new towers built by Guy of Lusignan, his brother and his nephew, for there were no further widespread rebellions on the island until the fifteenth century. Although the risk of such incidents must have receded considerably after the 11905, it did not disappear completely. In 1359, for example, the papal legate Peter Thomas attempted to convert Orthodox Greeks at Nicosia to Catholicism but the meeting ended in a riot, with many locals shouting 'death to the legate', who was only saved by the swift intervention of royal troops.8 At about the same time the Venetians also warned Peter I that Greeks rebelling on Crete could try to contact their co-religionists and supporters on Cyprus.9 As had been the case in the Holy Land following Saladin's victory at Hattin, the native population also took advantage of Prankish military setbacks. During the Genoese invasion of 1373, for example, the Franks were

HO

UNKNOWN CRUSADER CASTLES

so preoccupied elsewhere that Greeks west of Kyrenia simply slipped out of their control. This situation did not change until John of Antioch, Constable of Cyprus, arrived in person, for 'when the villagers saw the Constable they did homage to him, as they had been in rebellion and had been refusing obedience both to the castellan [of Kyrenia] and to their lord; nor were they performing their forced labour, nor were they paying their taxes'.10 Even more serious problems of this kind occurred in 1426, when a large Egyptian incursion led to the abduction of King Janus by Mamluk forces and the widespread breakdown of central authority. As a result 'many of the poor folk in their dwellings rose in rebellion and pillaged the Christians [Latins], and also killed many of them ... and at Lefkosia [Nicosia] they set up King Alexis, and all the peasants submitted to his rule'.11 These events, which bear a remarkable resemblance to the very first Cypriot rebellion following Richard I's invasion, prove that there still remained hostility between many Franks and Greeks even after 230 years of Latin rule. As long as this hostility existed, there was a potential need for strongpoints which could protect western settlers and intimidate the native population. Apart from Greek rebellions, the second threat to internal security at this time came from a number of violent clashes involving fellow westerners. During such incidents castles were again relied on to try to impose order, either by the Lusignan kings, or by their Latin opponents. Hence in 1228, when the German Emperor Frederick II arrived on Cyprus to stake his claim over the regency of young King Henry I, he garrisoned the island's strongholds with his own supporters. Thereafter he took John of Ibelin, who was his strongest local rival and had attempted to defy him by sheltering inside St Hilarión, with him to the Holy Land. The following year, Frederick again made sure that all Cypriot castles were well stocked with supplies and imperialist troops before ending his crusade and returning to the West.12 These measures were clearly intended to prevent the Ibelins from regaining the upper hand even after the emperor's departure. Much of the subsequent fighting which took place on Cyprus also revolved around the control of individual fortresses. After Frederick had gone home, his supporters were eventually defeated by the Ibelins outside Nicosia in the summer of 1229 and therefore lost control over Cyprus. However, many of them managed to escape the battle and retreat to the castles of Kantara, St Hilarión and Kyrenia. The garrison of the last stronghold gave in relatively quickly, honouring an agreement whereby it would surrender if imperialist reinforcements had not arrived by an agreed date.13 But the two mountain castles held out for about ten months, even though inadequate food supplies forced the defenders of St Hilarión to eat a donkey for their Easter feast

F O R T I F I C A T I O N S AND I N T E R N A L S E C U R I T Y

111

in the spring of 1230. This problem was temporarily solved by a surprise raid on the Ibelin camp outside the castle, but in the end such tactics could not prevent the starving garrison from capitulating shortly afterwards.14 A similar combination of hunger and lack of reinforcements made Kantara surrender at about the same time, despite the fact that all efforts by the Ibelins to take it by storm had ended in failure. At one point the pro-Ibelin knight Anceau de Brie had attempted to bombard Kantara with a large trebuchet, which 'battered down nearly all the walls, but the rock was so strong that it could not be scaled'.i5 The island's mountain castles proved equally difficult to capture in the opening months of 1232, when the situation was reversed. The arrival in the East of Frederick Ifs marshal Richard Filangieri enabled the imperial baiili Aimery Barlais to regain power over virtually all of Cyprus, but this time the Ibelins rather than the imperialists looked to St Hilarión and Buffavento as a means of continuing the struggle. These two strongholds provided shelter for various Ibelin supporters, including John of Ibelin's daughter-in-law Eschive de Montbéliard, who fled to Buffavento from Nicosia disguised as a Franciscan. Despite being short of supplies, Eschive and her companions held out successfully until other Ibelin forces arriving from the mainland defeated Richard Filangieri at the battle of Agridi in June 1232, forcing him to call off the siege and retreat inside Kyrenia.16 Yet another lengthy blockade now ensued, for Kyrenia itself was so well defended that its garrison of fifty knights and 1000 other troops held out for almost a year. During this time John of Ibelin and his Genoese allies blockaded the castle by land and sea, and attacked it with numerous siege engines, many of which were set alight in the bitter hand-to-hand fighting. Once again, however, it was despair and hunger rather than military defeat which finally caused the Lombards to surrender and withdraw from Cyprus permanently.17 The outcome of the civil war was therefore determined by the fate of Kyrenia and the three mountain strongholds, for neither side could hope to control Cyprus fully until it held all of these castles. It is interesting to note, for example, that in 1228 Frederick II was extremely concerned not to leave the island before John of Ibelin had been removed from St Hilarión. That this was achieved through negotiation rather than confrontation indicates that even the emperor of Germany did not rate his chances against such a powerful fortress.1« Similarly, the following year it can be said that imperialists lost their grip on Cyprus not so much because they were defeated in battle but because they failed to defend Kyrenia and Buffavento. They eventually also had to give up Kantara and St Hilarión. The defenders of these latter sites only surrendered through hunger, and could otherwise, it

112

U N K N O W N CRUSADER CASTLES

seems, have held out almost indefinitely. If this had been possible they could have waited until supply problems, a sense of despair or the arrival of Richard Filangieri's reinforcements had forced their opponents to raise the blockade. During the first six months of 1232 these were the exact circumstances which led to the end of Hohenstaufen power on Cyprus. Aimery Barlais and Richard Filangieri's Lombard troops failed to capture St Hilarión and Buffavento from the Ibelins, so that when John of Ibelin himself arrived from the mainland with an army of supporters gathered at Acre, the imperialists found themselves fighting on two fronts. If they had not already committed vital troops and resources to besieging St Hilarión they would perhaps have stood a better chance at the battle of Agridi in June. In the wake of the battle, the imperialists had no choice but to retreat to Kyrenia, leaving their garrison at Kantara completely cut off and unable to do anything but surrender. Once it became apparent that the Lombards at Kyrenia were not going to receive any more help from Frederick II, they were forced to do the same.19 These events confirm that two key factors favoured the Ibelins during the civil war of 1228-33. First, the imperialists relied too much on foreign troops sent by Frederick II whereas the Ibelins raised their armies locally, so that although they did have trouble finding enough troops, they could at least gather reinforcements more swiftly than Frederick's baillis. This factor proved decisive in 1230 when St Hilarión and Kantara were starved into submission, and again between 1232 and 1233, when the Ibelins succeeded in relieving St Hilarión but the imperialists failed to assist Kyrenia in time. Secondly, it has already been noted that the troop numbers involved in this conflict were relatively small, so that the battle of Agridi, for example, was only fought between a few hundred mounted troops. These figures inevitably resulted in swift battles and lengthy sieges, for it was far easier for a small army to defend Kyrenia and the mountain castles than it was for one to attack them. The fact that John of Ibelin enjoyed much greater political power and popular support on Cyprus made him the clear favourite in a war of attrition of this kind.20 The importance of Cypriot castles to internal stability came to light once more during Amaury of Tyre's reign as 'governor' between 1306 and 1310. As soon as Amaury, aided by the vast majority of the nobility, had deposed King Henry II, he immediately 'sent castellans and bailies to all the towns and castles of the kingdom of Cyprus', to prevent them from falling into royalist hands and to guarantee the loyalty of their garrisons.21 Thereafter Amaury's harsh rule and shaky legitimacy caused his popularity to go into decline, and eventually brought about his murder in 1310. It was not just

FORTIFICATIONS

AND I N T E R N A L S E C U R I T Y

113

Amaur/s death, however, but the failure of his supporters to hang on to the kingdom's strongholds, which finally led to the defeat of his faction. Shortly after Amaury had been killed, the people of Limassol, Paphos and Famagusta, plus the garrison of Kyrenia, all came out in favour of Henry II. The defection of the first two settlements, which were still only defended by individual towers at this time, may not have represented such a great blow to Amaury's followers, but the loss of Kyrenia and Famagusta proved decisive. At Famagusta the recently constructed town walls were manned by royalists, and all the city gates were walled up. As we have seen, the Franks had also recently built new defences at Kyrenia, and so Henry II's enemies suddenly found themselves isolated at Nicosia, with little chance of retaking either of these extremely well fortified strongpoints. Consequently the king himself, who had been exiled in Cilician Armenia since early 1310, was able to return to Famagusta unhindered, and the usurpers were forced to surrender shortly afterwards.22 It is hardly surprising, therefore, that throughout the medieval period all Cypriot strongholds belonged to the king. In Prankish Greece and the Holy Land constant incursions by the Greeks or the Muslims forced Latin rulers either to transfer royal castles to the nobility and the Military Orders, or to allow these groups to build their own fortresses This was done in order to spread the huge military and financial burden of defending Latin territories.23 But in Cyprus physical isolation made this unnecessary, so the Lusignans were able to keep all cities and major castles within the royal domain. This in turn guaranteed their income, for it enabled them to retain many of the judicial and financial privileges, such as the minting of coins, which had been lost in other crusader states. Their overwhelming wealth and power also allowed the Lusignans to keep a close eye on the transfer of estates between fellow vassals, so that gifts of land to the Military Orders, for example, could be strictly monitored.24 Apart from the Military Orders, which were in fact allowed to hold small fortified properties on their estates, no private castles were constructed by the nobility. In this way, both the local Cypriot population and the king's Latin vassals could be kept in check, for neither group had their own strongholds where they could take refuge or challenge royal authority. This was in marked contrast to the Holy Land, where it has been shown that Latin rulers such as Frederick II and Bohemond VII of Tripoli experienced great difficulty in suppressing defiant vassals once they had retreated inside their own powerful castles.25 But on Cyprus it was the king rather than rebellious lords who could use strongholds in this way. These structures helped the Lusignans to eliminate the anarchy which prevailed on Cyprus in 1191 and 1192, and they also explain why outbreaks of warfare between

114

U N K N O W N CRUSADER CASTLES

fellow Latins were so unusual. The only two clashes of this kind which did occur ended in victory for the Lusignans thanks to the strength of their castles. Thus Henry I, greatly aided by his powerful vassal John of Ibelin, triumphed over the imperialists because they lacked the resources to hold on to the island's largest fortresses after Frederick II's departure in 1229, whilst Henry II ultimately defeated Amaury of Tyre's faction because it could not hope to recapture Kyrenia and Famagusta in 1310. Whilst all Cypriot castles clearly enhanced royal power, their role varied somewhat depending on their size and location. The isolated mountain fortresses of St Hilarión, Buffavento and Kantara served as final refuge points which could be relied on to tie down large numbers of attacking troops and resources. Although they may also have helped to suppress the native population during the 11905, they were far too remote to intimidate any Greek settlements directly and therefore had more to do with national defence than the Prankish towers constructed at Limassol, Paphos, Nicosia and Famagusta. These latter strongpoints could not hope to withstand the kind of large catapult deployed by Anceau de Brie against Kantara in 1230, but they could maintain law and order on a local scale and acted as highly visible symbols of the new Prankish regime. Kyrenia, on the other hand, enhanced both local and national defence, for it overlooked a relatively large town but also proved too powerful even for the most determined besiegers. Similarly, Famagusta played a relatively small role in the civil war of 1228—33, but in 1310, after it had been provided with extensive new urban fortifications, the support of its citizens for Henry II helped determine the course of Cypriot history.

10

Fortifications and External Security It has already been shown that two external aggressors threatened Cyprus at this time. First, there were pirates or other sporadic raiders who wreaked havoc on a local scale, and secondly there were those who intended to launch more permanent invasions of the entire island, particularly the Mamluks of Egypt and the Genoese. The destruction of the Byzantine Empire from 1204 onwards, the naval weakness of the Muslims, and the priority given to the Holy Land itself meant that this latter threat did not emerge until after the fall of Acre in 1291, an event which brought more Italian traders to Famagusta and made Cyprus the next natural Latin target for the Mamluks.1 Before 1291, Cyprus only suffered a relatively small number of piratical attacks. Those who carried out such raids hoped to gain booty or inflict damage locally before escaping again as quickly as possible, and had neither the equipment nor the inclination to get embroiled in lengthy sieges. Consequently, there was no need for the Lusignans to build extensive coastal fortifications at this time, and it may indeed be that they considered the new isolated towers erected at Paphos, Limassol and Famagusta during the late twelfth and early thirteenth centuries as sufficient shelters against pirates as well as potential Cypriot rebels. This possibility is strengthened by the position of the tower at Paphos, built on an eminence above the town and recently described as a 'watchtower'.2 In the early 11905, a Greek pirate known as Cannaqui descended on the island and made off with the family of Aimery (1194-1205), Guy of Lusignan's heir and brother.^ The timing of this incident and the fact that it had such a dramatic effect on Aimery himself are significant, for it has already been shown that the Franks probably began to construct their new towers at about the same time, and that both Guy and Aimery must have been directly involved in this project. The period between 1192 and 1195 therefore gave the Lusignans a dual incentive for building their new towers, for it witnessed a serious piratical attack against Cyprus as well as a major Greek uprising at Nicosia. After 1192 there were a number of further seaborne raids. In 1220, for example, Muslim ships raided Limassol harbour in order to disrupt Christian supply lines between the West, Acre and the members of the Fifth Crusade

Il6

UNKNOWN CRUSADER CASTLES

who were at that time campaigning in Egypt. This attack supposedly resulted in the death, capture or dispersal of at least 13,000 crusaders anchored there.4 More than fifty years later, in 1271, Baybars made another attempt to raid the town so that the Franks' attention would be drawn away from his campaign against the castle of Montfort in Galilee. This latter incursion ended in disaster when the Muslim fleet was shipwrecked on the coast; a vivid illustration of the incompetence of Egyptian naval forces at this time.5 Even if the attack had been successful, it is clear that both this raid and its predecessor were isolated incidents which had far more to do with events on the mainland than on Cyprus itself. Incursions by pirates were equally sporadic, occurring in 1302, when corsairs from Rhodes appeared,6 and more seriously during the reign of King Hugh IV (1324-59), who had over a hundred prisoners hanged because 'these men were thieves, pirates, robbers, who went pillaging and murdering, and did great damage'.7 In the early fourteenth century more deliberate attacks were carried out by the Genoese, who raided the coast around Paphos in 1312 and 1316 because a loan they had made to the usurper Amaury of Tyre was not fully repaid by Henry II after he regained the throne in 1310.8 Shortly before their conquest of Famagusta, the Genoese also targeted both Paphos and Limassol, where 'they landed and burned the houses ... and did much damage'.9 Some years earlier the naval campaigns of Peter I (1359-69) against the coast of Asia Minor also led to retaliatory raids by the Turks, who 'went and pillaged the country' in the 13605.10 During such attacks, Cypriot fortifications were expected to fulfil two basic functions. First, strongpoints located near the coast or on high ground could act as look-out posts. Consequently their garrisons could give early warning both to civilians wishing to take evasive action and to naval or other troops hoping to confront the intruder. Inevitably the three mountain castles, which enjoyed extensive views over the northern, eastern and southern coasts of Cyprus, were most significant in this respect. Some of these strongholds were also intervisible both with each other and with surrounding settlements such as Kyrenia. They could therefore communicate using fire or smoke signals. Buffavento lay at the heart of this network, and here 'they kept a look-out every night, and as soon as they spotted ships at sea, they would signal with fire or torches to the town of Nicosia and the castle of Kyrenia'.11 This raises the possibility that the isolated uppermost towers already described at Buffavento, Kantara and St Hilarión were normally used as watch towers rather than as final refuge points for the garrisons of these castles.12 On a less spectacular scale, the harbour location of the Prankish towers

FORTIFICATIONS AND EXTERNAL SECURITY

ll/

at Limassol and Famagusta, as well as the elevated position of the Paphos tower, indicate that these structures could fulfil a similar function. In addition they must have been useful landmarks (and perhaps even lighthouses) for ships approaching these ports. Numerous examples can be cited from the history of Famagusta to illustrate how these structures acted as look-out posts which enhanced the security of the island as a whole. During the reign of Henry II (1285-1324) the city's tower was incorporated into a much larger rectangular citadel, and it was probably at this time that a chain tower was constructed on a reef opposite the harbour entrance.13 In 1368 the garrison of these defences spotted two pirate galleys from Morocco approaching Famagusta. The chain across the harbour was raised to prevent these vessels from gaining access, while the admiral was informed and prepared two vessels for a counter-attack. These precautions were seen by the pirates, who quickly lost courage and withdrew. 1* Similarly, five years later an attempt by seven Genoese galleys to raid Famagusta in preparation for their coming invasion of Cyprus ended in failure, because 'the watch on the [chain] tower perceived them, and were able to shoot at them with their arrows as much as they chose, and with stones, nor were [the Genoese] able to fix a gangway to the shore'.15 If such measures still failed to deter attackers, the second function of Cypriot fortifications was to provide temporary shelter for those living close by. This would not have been a problem for the large urban defences of Kyrenia and Famagusta, although it seems unlikely that the Franks ever allowed the Greek population inside the actual citadels of these sites. Similarly, the small size of the original Prankish towers at Famagusta, Paphos and Limassol, which were after all intended to deter the native population from rebelling, suggests that these structures were never designed to house anyone apart from Latin garrisons and (in emergency) civilians. When the Genoese raided Limassol in 1373 'the inhabitants took flight', but there is no indication that the castle was capable of protecting large numbers of local Greeks.16 Apart from being too small to house the native people in general, some coastal defences were also too weak to withstand concerted enemy attacks. They clearly relied on the fact that most raiders, wishing to leave before royal troops arrived, would never even attempt to besiege them. This thinking was most evident at Paphos, where two new towers were constructed in the course of the fourteenth century. Their location near the harbour suggests that they were built in response to the Genoese raids of 1312 and 1316; but, like their neighbour at Limassol, they cannot possibly have been large enough to shelter anyone apart from western troops and settlers. When the Genoese returned in 1373, with a small fleet sent ahead

Il8

UNKNOWN CRUSADER

CASTLES

of their main invasion force, an army of 2000 mercenaries landed near Paphos and occupied both towers with ease. They then 'set to work and heightened them, and cut a trench, so that the sea flowed in and surrounded them with water'. As a consequence, 'when the Cypriots brought up fightingtowers and soldiers in them, they resisted the attack without anxiety for the result'. Thereafter Genoese troops may have stayed at Paphos until the unsuccessful campaign against Kyrenia came to an end in the spring of 1374-17 The damage caused by the Genoese no doubt explains why James I carried out such extensive repairs at Paphos in the 13905. Claims by some sources that he also built a 'citadel' there, no trace of which can be seen today, imply that he may have tried to remedy the problems which Paphos experienced at the hands of the Genoese.18 Similar concerns, combined with the general escalation in piracy and in Turkish raiding which took place during the mid fourteenth century, help to explain why lames also carried out the new building work at Limassol, La Cava and Sigouri described in Chapter 8. Claims that during the 13605 Turks frequently 'landed and went and pillaged the country, and took away gear and many people' suggest that by this point it was not only coastal areas but many of the island's inland regions which were subject to attack.1« But piracy still remained sporadic, so the Lusignans made no efforts to erect vast fortifications intended to protect the entire population and, with the exception of the Military Orders, a strict royal monopoly was maintained on wider castle building. The countryside never became dotted with minor fortifications and baronial strongpoints in the way that Palestine and Syria had been before 1291. The second, and ultimately far more serious, external threat to Cypriot security came from full-scale invaders. Whereas individual towers such as those erected at Paphos after 1316 were the only new defences built specifically to deal with raiders, vast resources were committed to protecting Cyprus against first the Mamluks, then the Genoese, and finally the Ottoman Turks. The large building projects of the thirteenth and fourteenth centuries can all be dated to one or more of the major invasion scares which gripped Cyprus during this period. The new city walls and castle of Famagusta, the alterations at Kyrenia and the possible improvements to Nicosia's defences carried out by Henry II were a direct response to the fall of Acre and the other remaining Latin outposts in the Holy Land in 1291. These events caused great alarm, raising concerns both on Cyprus and in the West that the Mamluks were about to invade. If in the end these fears were not realised, the Lusignans could not have predicted this in 1291 and had to prepare for the worst.20

F O R T I F I C A T I O N S AND E X T E R N A L S E C U R I T Y

119

The massive new defences built at Famagusta and Kyrenia were expected to fulfil the same defensive role as earlier Prankish strongpoints on the mainland, such as Acre, Tyre, Saphet or Tortosa. The various shooting galleries, posterns, towers and ramparts added to Kyrenia meant that this fortress came to resemble many of the larger Prankish castles on the mainland.21 Such sites were constructed on a vast scale so that they could resist Muslim invasion forces numbering 10,000 or even 20,000 troops. Similarly, the large ashlar blocks used during the rebuilding of Kyrenia were clearly chosen in response to the advanced Mamluk artillery and sapping techniques which the Franks had come to dread on the mainland.22 The sheer scale of the enemy which now confronted Cyprus also made it futile to defend the island as a whole, or even prevent attackers from landing in the first place, as was sometimes possible with pirates. The maximum field army which could normally be raised by Cypriot kings probably only contained around 200 knights and 500 infantry.23 Like their predecessors on the mainland, the Lusignans therefore made no effort to defend their borders but concentrated their outnumbered forces inside a small number of extremely powerful strongpoints, even if this meant abandoning control over the countryside temporarily. Famagusta and Kyrenia were deliberately strengthened to create such strongpoints, alongside the already impregnable mountain fortresses of Kantara, St Hilarión and Buffavento. During the civil war of 1228-33, as already shown, these castles could be relied on to hold out almost indefinitely provided that they were well supplied and garrisoned, forcing an attacker to devote large numbers of men and resources to a siege which could last for many months or even years. The relatively small size of Cyprus, combined with the long lines of communication between it and the mainland, would have put a severe strain on a large Mamluk army living off the land for any length of time. The Lusignans therefore hoped to withstand an Egyptian attack in much the same way that the kingdom of Jerusalem had survived the Khwarizmian invasion of 1244, when lack of equipment, supplies and patience prevented the Khwarizmians from capturing key strongholds and eventually obliged them to withdraw.24 During the mid fourteenth century similar thinking probably induced Peter I to carry out considerable improvements on the urban defences of Nicosia. The Margarita Tower, Peter's replacement for Nicosia's original Prankish tower, was built on a strategic hill on the outskirts of the city rather than at its centre. It therefore had much more to do with guarding the capital against external attackers than suppressing the local Greeks.25 These defences were constructed because Peter's aggressive foreign policies had renewed the threat of a serious Muslim attack on Cyprus. His naval

12O

U N K N O W N CRUSADER CASTLES

offensives against the coastal settlements of Asia Minor help to explain why there were so many retaliatory Turkish raids at this time, whilst Peter's crusade against Egypt in 1365, which resulted in the temporary occupation and partial destruction of Alexandria, revived fear of a retaliatory Mamluk invasion. These events probably also contributed to Peter's assassination in 1369, for the king had already spent vast amounts of money on his ambitious naval campaigns, and now had to raise yet more taxes from his resentful vassals in order to pay for new defences against the very enemies which he himself had provoked.26 In the end it was not the Mamluks but the Genoese who put the Lusignans' elaborate defence plans to the test. In 1373 they invaded Cyprus in the hope of turning the island into an exclusive Genoese domain capable of controlling the lucrative trade routes of the eastern Mediterranean. At first Cypriot hopes that an invasion of this kind would be defeated in a series of lengthy sieges failed miserably, for in October the Genoese captured Famagusta through stealth, so that their force of 14,000 men and more than forty galleys, far from getting bogged down, remained relatively unscathed.27 Not long afterwards a Genoese contingent also succeeded in capturing Nicosia, where Peter I's new walls were not yet tall enough to withstand a siege.28 Once inside Nicosia the Genoese used these defences for their own benefit. Having realised that they lacked the resources to occupy the entire city, they only garrisoned a stretch of the ramparts 'from the Market Gate to the Tower of St Andrew, and they made the walls higher and held the place in great force. And the tower which stands near the Market Gate they filled with earth and stones and made it like a castle'.29 These fortifications protected the Genoese against two unsuccessful counter-attacks by Prankish forces from Kyrenia. They also enabled the invaders to intimidate the entire city, which succumbed to a long period of looting and street fighting involving both the Genoese and the local citizens.30 The Genoese intended to conquer the rest of the island and proceeded to fight their way north through the Kyrenia mountains, whose passes were guarded by Bulgarian mercenaries from St Hilarión and Kyrenia. It was at this stage that Peter Fs widow, Eleanor of Aragon, having realised that the Genoese were far more interested in taking control over all of Cyprus than in helping her to avenge her husband's murder, fled from their army and rejoined the Cypriot cause.31 Once the Genoese reached Kyrenia itself, 'the bridges were raised and the gates nailed up', and the attackers' advance was halted in exactly the way that had been hoped.32 Despite several weeks of intense fighting, the Genoese failed to make any progress against such a massive obstacle, whose defenders successfully bombarded or set fire to their opponents' siege weapons in the same way that

F O R T I F I C A T I O N S AND E X T E R N A L S E C U R I T Y

121

the Lombards had done in 1232-33. At one stage many Genoese troops were killed because they charged onto a drawbridge which had deliberately been left open, and was counterpoised in such a way that anyone standing on it could be tipped into the moat.33 This contraption shows just how carefully the defences had been prepared at Kyrenia and is reminiscent of the kind of elaborate devices used to protect Prankish castles on the mainland, such as the labyrinthine main entrance at Crac des Chevaliers.34 At Kyrenia fighting continued after a brief period of inconclusive talks. But still the Genoese were repelled, despite constructing a vast wooden platform which was lashed between two galleys and used to fire into the castle from off shore. As a result, most of the attackers withdrew to Nicosia, and a ceasefire was agreed which effectively ended the siege in March 1374-35 These events marked a turning point in the war, for the Genoese knew that 'if they did not have the castle of Kyrenia, they would not be able to hold Nicosia, or the rest of the island'.36 They could not afford to bypass Kyrenia, which was defended by the constable, John of Lusignan, and a large contingent of foreign mercenaries, all of whom were capable of harassing or even severing Genoese supply lines from Famagusta and Nicosia. Indeed, in the course of the siege Bulgarian troops hiding in the Kyrenia Mountains did just that, capturing food and siege equipment being transported from the capital.37 This would not, of course, have been a problem if Kyrenia had been captured, for then the Genoese would have held another secure port on the Cypriot coast to complement their newly acquired base at Famagusta. Indeed, bearing in mind that Genoa was one of the greatest naval powers in the Mediterranean, and that her initial intention in attacking Cyprus had been to dominate seaborne trade there, it made sense to attack Kyrenia before any of the neighbouring mountain castles. By the same token, their failure to take Kyrenia persuaded the Genoese to relinquish Nicosia, whose low walls made it difficult to defend, and whose inland location gave it little strategic or economic importance in its own right. A maritime power like Genoa was just as uninterested in Nicosia as the Italian cities had once been in Jerusalem. After their withdrawal from Kyrenia and Nicosia, the Genoese were limited to the city of Famagusta and its immediate surroundings, which they held until 14Ó4-38 But the threat of further Genoese attempts to complete the conquest of Cyprus remained, and it is in this context that the numerous building projects by Peter II (1369-82), James I (1385-98) and Janus (13981432) should be considered. Hence the new citadel and additional urban defences erected at Nicosia after 1374 were clearly designed to spare the capital from another Genoese occupation, as the first one had resulted in

122

UNKNOWN CRUSADER

CASTLES

much destruction and had illustrated the dangers of leaving fortifications only half-completed. Further afield, the alterations carried out by James I at Limassol, Paphos and Kantara, as well as the new castles he built at Sigouri and La Cava, created a ring of fortifications around western Cyprus. These were intended to deter the Genoese from attacking in the first place and, failing this, could provide shelter against land or sea raids launched from Famagusta. The completion of several new fortifications, and in particular the major new defences at Nicosia, also made it even less likely that the Genoese would ever be able to complete the conquest of Cyprus. In 1374 they had failed to capture Kyrenia, and now this impregnable fortress had been joined by numerous other strongpoints which would also need to be taken before the entire island could be subjugated.39 In many ways the most important castles in this network were Kantara and Sigouri. They lay closest to Famagusta, and the name of the latter fortress may indeed derive from the word securus, or secure.40 According to one source James I built this castle 'so that he could resist the Genoese, and halt the raids and incursions which they made from Famagusta'.41 Kantara, meanwhile, was not only extremely well defended against a potential siege but was also perfectly positioned in the mountains above Famagusta for its garrison to keep an eye on the Genoese and to warn other Cypriot strongholds of an imminent attack. Their proximity to Famagusta also gave Sigouri and Kantara a potential offensive function, for they could be used to launch counter-attacks either against the city itself or against Genoese forces operating close by. Their use can therefore be compared with that of other strongpoints built by the Franks elsewhere. As will be shown, the Latin conquerors of Greece built a number of small castles to blockade Corinth shortly after the Fourth Crusade.42 In addition, Sigouri was very similar to the castrum-style castles of Bethgibelin, Blanchegarde, Gaza and Ibelin, constructed by the twelfth-century kings of Jerusalem to contain the Muslim-held coastal outpost of Ascalon.43 The Genoese invasion of 1373-74 provided parallels with the civil war of 1228-33 between Frederick IFs supporters and the Ibelins. Although this conflict concerned internal rather than external disputes, it did involve a high proportion of foreign troops sent by Frederick. During both these wars the most important strongholds of all were the three mountain castles, plus Kyrenia, which 'was never breached or taken by storm, despite having been attacked in so many wars'.44 The sheer strength of these places inevitably reduced both campaigns to wars of attrition which favoured local rulers rather than foreign attackers. The fact that Kyrenia, St Hilarión, Buffavento and Kantara were all situated in the same region of Cyprus made it possible for their garrisons to put

FORTIFICATIONS

AND E X T E R N A L S E C U R I T Y

123

up a coordinated defence against invasion forces. In the winter of 1373-74 the Genoese had already seized both Famagusta and Nicosia and were moving north in order to attack Kyrenia. To reach this castle they had to march through a narrow mountain pass, but this vital route was defended by Bulgarian mercenaries in Cypriot pay and it was not until January 1374 that the Genoese finally broke through to the coast. After they had been defeated, the Bulgars quickly retreated to St Hilarión and Kyrenia,45 both of which had already been carefully provisioned by James of Lusignan in expectation of the coming blockade.46 This indicates that troops from these two castles were working in close harmony and according to a carefully prepared plan. While unable to prevent the Genoese from reaching the coast, the plan slowed them up, enabling the defenders of Kyrenia itself to prepare for a lengthy siege. This siege ended in defeat for the Genoese and effectively destroyed their hopes of conquering Cyprus. The Genoese invasion of 1373-74, as well as the underlying dispute between Eleanor of Aragon and the alleged killers of her husband Peter I, left Cyprus in a sorry state. Yet the outcome would have been even worse if it had not been for Kyrenia and the mountain castles. These strongholds halted the Genoese in exactly the same way that sites like Crac des Chevaliers had prevented Saladin from wiping out the Holy Land almost two centuries earlier. This cluster of fortresses therefore represented the cornerstone of Lusignan power on Cyprus, as is confirmed by their subsequent history, for they remained in royal hands throughout the Lusignan period. More than a century later the Venetians, who wished to concentrate their resources on fortifying Nicosia and the coastal settlements, preferred to demolish such formidable castles rather than see them fall into the wrong hands.47

11

Fortifications, Naval Warfare and the Crusades Cyprus's close proximity to Asia Minor, Cilician Armenia, Egypt and the Holy Land meant that its coastal strongholds could act as secure starting points for various campaigns against the Muslims, whilst at the same time protecting lines of communication with other Latin states. During the thirteenth century, the most important such base appears to have been Limassol. One admittedly late source even states that Guy of Lusignan 'built a castle there, because this was the closest place in Cyprus to the Holy Land, which was no further than about 150 Italian miles'.1 This may imply that it was the first Prankish tower constructed on Cyprus; a logical starting point when one considers that the interior was still politically unstable in 1192 and that Limassol harbour represented the quickest way of gaining help from the Holy Land, or alternatively escaping to it. The fact that Willbrand of Oldenburg described the citadel of Nicosia as new in 1212 also suggests that the Lusignans built the coastal, and therefore less isolated, towers of Limassol, Paphos and Famagusta before they finally suppressed Nicosia itself.2 No doubt they had learnt the dangers of entering the interior unprepared, and becoming stranded there, from the Nicosian uprising which had forced the Templars to abandon Cyprus in 1192. Once the Franks had gained full control over Cyprus, Limassol was used to wage war against the Muslims rather than the island's native population, for it provided an ideal stopping-off point on the sea voyage between western Europe and the Holy Land. In 1228, for example, Frederick II stayed there briefly before sailing on to Acre.3 This must have been his intention for some time, for the previous year many Latin nobles had already gathered in the city until the emperor's departure from Europe was postponed.4 Exactly twenty years later Louis IX made even greater use of Limassol, spending the entire winter of 1248-49 in the city before attacking the Egyptian port of Damietta, and supplying his troops from vast storerooms, granaries and wine cellars built up by agents sent out ahead of the main army.5 Other western expeditions which relied either on Limassol itself or on neighbouring Cypriot ports included the Fifth Crusade (1217-21)6 and the crusade of Edward, the future king of England (1271-72).7 Cyprus became a logical port of call for these crusades from a

F O R T I F I C A T I O N S , N A V A L W A R F A R E AND THE C R U S A D E S

125

political as well as a geographical point of view, for Cypriot troops subsequently participated in all of them.8 It should also be noted that for much of the thirteenth century, and particularly when Cyprus and Acre were ruled by the same man, the island's resources could still be used to boost the weakening defences of the Holy Land, regardless of whether a full crusade was present. As late as 1291 Henry II took an army of 200 knights and 500 footsoldiers to Acre in a final bid to save it from al-Ashraf Khalil's besieging army.9 During these expeditions, the coastal fortifications of Cyprus were expected to provide a modicum of security and comfort for crusaders. Both Louis IX and Frederick II, for example, may have used Limassol's castle as a residence or as a safe place to store funds and documents. Despite the small size of this stronghold and the naval inferiority of the Muslims, the risk of surprise raids made it desirable for them to remain as close as possible to this look-out post and potential shelter. During Frederick's crusade there must have been fresh memories of the Muslim raid on Limassol harbour which had taken place only eight years earlier.10 Similarly, it is possible that Baybars's failed raid on Limassol in 1271, primarily launched to distract the Latins from his campaigns in Galilee, was also intended to disrupt the Lord Edward's crusade to the Holy Land. Baybars's raid took place in the summer of 1271, by which time Edward had already been in Acre for some time, but it should be noted that his brother Edmund did not arrive in the East until September. In the meantime King Hugh III of Cyprus (1267-84), relying on men and supplies brought from his island, had joined Edward in the Holy Land. Baybars may therefore have hoped to intercept Edmund or, more plausibly, sever supply lines between Cyprus and the mainland, at the same time punishing Hugh for his involvement in the crusade. In these circumstances the castle of Limassol, however small, must have been a convenient stepping-stone between Cyprus, Acre and the West, and a potential depot for vital supplies, u Returning to the first half of the thirteenth century, it is also possible that Paphos, the closest Cypriot harbour to Egypt, was either involved in or had at one stage been intended as a starting point for the Fifth Crusade against Egypt (1217-21). It may have featured in Honorius Ill's initial plans to use Cyprus as an assembly point, in the same way that Louis IX later did.12 In the end, these plans were shelved. Although Limassol was subsequently used to ship some supplies to Damietta from the Holy Land, it was Acre rather than Cyprus which eventually acted as the starting point for the campaign. But if Paphos never played any role in the Fifth Crusade, it is difficult to understand what function the major new castle built there by the Franks between c. 1205 and 1222 was intended to serve.

126

UNKNOWN CRUSADER CASTLES

This fortress was situated on a hilltop overlooking the town and stood directly above the earlier tower probably built by Guy of Lusignan around 1192. It was defended by two successive curtain walls arranged around a central courtyard, making it far stronger than any other Prankish fortification built on Cyprus before 1291.13 But if simple towers were deemed sufficient to deter pirates and suppress the native population elsewhere, why did the Franks need to construct such a large stronghold at Paphos? It has been suggested that it was probably built by the Hospitallers between 1200 and 1204, because of its close architectural links with Belvoir, the Order's famous concentric castle overlooking the River Jordan. Its supposed role was to defend Cyprus against an expected Byzantine invasion, and it therefore complemented the Templars' new coastal base at Gastria, which lay north of Famagusta and was completed at some point before mo.14 It seems unlikely, however, that the Templars and the Hospitallers were building major coastal fortifications at this time, partly because Gastria amounted to little more than a walled enclosure incapable of withstanding a major siege, and partly because the Byzantine empire itself was in such a state of crisis during the final years leading up to the Fourth Crusade that it was no longer able to launch a major assault on Cyprus.15 Finally, it is odd that Willbrand of Oldenburg, who had a strong interest in crusader fortifications and always provided descriptions of other strongpoints he saw in the Latin East, failed to mention a large, prominent castle when he visited Paphos in 1212.16 This makes it possible that the castle was actually built during the Fifth Crusade, perhaps after the Muslim raid on Limassol harbour in 1220, an event which may have led to fears of more raids, or even a full-scale Ayyubid counter-attack against the whole of Cyprus.17 Alternatively the castle may have been constructed during the opening years of the Crusade between its proclamation by Innocent III in 1215 and its actual start in 1217. This is implied by the discovery of numerous west European coins at the site, along with a papal seal attributable to Honorius III (1216-27), a stockpile of arrowheads and a large store of catapult ammunition. These finds have been used to suggest that the castle was a Hospitaller base frequented by European pilgrims, but they could just as well indicate that Paphos acted as a stepping-stone for troops and supplies destined for Egypt.1» Its construction should therefore perhaps be attributed to Hugh I (1205-18), or representatives of his infant son Henry I (1218-53), rather than to the Hospitallers, although their participation in the Egyptian campaign does not rule out the possibility that they took part in the project.^ This would also explain why the fortress was never rebuilt after the devastating earthquake of 1222, for by then the Fifth Crusade had come to an end. It may

F O R T I F I C A T I O N S , N A V A L W A R F A R E AND THE C R U S A D E S

12/

even account for the incomplete nature of the castle's rock-cut ditches; a perfectly understandable omission if construction of the castle had only begun a few years before it was destroyed.20 It is interesting to speculate whether material taken from these ditches was subsequently used in Egypt, for according to Jacques de Vitry the Franks imported Cypriot stone to repair the walls of Damietta.21 After the Fifth Crusade, it is unlikely that Paphos played any part in subsequent Christian campaigns. Indeed, after Louis DCs crusade Limassol itself gradually came to be overshadowed by Famagusta, whose economic importance and volume of shipping grew steadily during the late thirteenth century.22 In 1291 it was from here rather than Limassol that Henry II set out in his final bid to save Acre.23 His failure to do so merely encouraged the growth of Famagusta, which emerged as the richest port in the eastern Mediterranean still in Latin hands, and therefore gained new urban defences in preference to Limassol.24 In addition, the fall of Acre had a profound effect on the strategic importance of Cyprus as a whole, for it now became the new front line in a battle zone where galleys and warships were more important than knights and horses. In the fourteenth century Famagusta acted as a starting point for a number of campaigns undertaken by local naval powers, including the Hospitallers of Rhodes and the Cypriot Franks themselves. The first major expedition of this kind took place in 1300, when the Templars, Hospitallers and Cypriots set off from Famagusta on an offensive aimed against Alexandria. They sailed northwards along the Syrian coastline until they reached Maraclea, which they sacked before heading back to Cyprus. In the same year troops from Cyprus and the Military Orders even went ashore briefly at Tortosa, in a failed attempt to meet up with Mongol and Armenian forces.25 Thereafter the Templars also garrisoned a small castle on the island of Rouad, just opposite Tortosa, with a force of 120 knights and 500 archers. This stronghold must have relied on Cyprus almost entirely for its food supply, for shortly before it surrendered to a Mamluk besieging fleet in 1302, the Franks tried in vain to relieve it by sending ships from Famagusta. The Franks therefore used Cyprus as a springboard for raids against the mainland but were prevented from making any permanent reconquests there because of their usual lack of troops and resources.26 A similar fate befell Peter Fs renewed efforts against the Muslims in the mid fourteenth century. Peter launched several expeditions against the mainland. These have traditionally been linked with his pious and adventurous personality, but were probably also intended to prevent Muslim ports from killing off Famagusta's declining international trade.27 The most famous of these was his crusade against Alexandria in 1365, which brought

128

UNKNOWN CRUSADER CASTLES

the Latins considerable booty but they again lacked the strength to advance into the Egyptian interioras Undeterred, Peter continued to garrison Corycus and Satalia (Antalya), two ports along the coast of Cilician Armenia, which he had occupied in 1360 and 1361 respectively. In addition, he attacked several adjoining Turkish settlements such as Anamur, forcing local rulers to pay him tribute.29 The defences of Corycus and Satalia were also strengthened, so that at the latter site the Cypriots 'pulled down the old lodgings, and cut down the trees and filled up the ditches, because the Turks used to hide in them and leap out to attack people, and did much damage. And by the captain's orders they built up the wall of the castle and also heightened the towers'.30 Thereafter these outposts were supplied from Cyprus in the same way that the island of Rouad had been, and in 1367 Corycus, which had been blockaded by 'a great host of Turks', was successfully relieved by a Cypriot fleet.31 This port subsequently held out until 1448, although Satalia returned to Turkish rule in 1373. In occupying these sites Peter may again have been motivated by a combination of crusading and economic interests, for they enabled him to attack the Turks but were also useful ports in the lucrative trade network between Cyprus, Rhodes and the mainland.32 During this period, as well as during the earlier post-npi efforts to regain the Holy Land, Famagusta's defences were intended to provide a secure naval base for Cypriot ships. By the mid fourteenth century the city was protected by a harbour chain and chain tower as well as its landward urban defences.33 These structures made Famagusta the safest anchorage in Cyprus, where major expeditions against the mainland could be prepared. In addition, they facilitated the creation of smaller regular patrols such as that, consisting of two Cypriot galleys, which captured two pirate vessels in i325.34 Famagusta now became virtually immune to the kind of Muslim counterraids which had devastated the Christian fleet at Limassol in 1220 and had been attempted by Baybars in 1271. It is even possible that the first major Christian raid on the Holy Land after the loss of Acre, carried out in 1300, coincided with the completion of at least some of Henry IPs new defences at Famagusta. This would have released more funds for galleys and have given Prankish vessels a definite means of escape in case of Muslim counterattacks. If this was the case, it is reminiscent of the Hospitallers' earlier reluctance to raid Muslim targets in Syria before their own castles of Margat and Crac des Chevaliers had been made strong enough to withstand potential retaliatory sieges.35 Famagusta may not, however, have been as well fortified as the Franks would have liked at this stage, for Amaury of Tyre was still improving the city's defences ten years later.36 Beyond Famagusta, accounts of the Genoese siege of Kyrenia in 1374

F O R T I F I C A T I O N S , NAVAL W A R F A R E AND THE CRUSADES

129

suggest that here too a chain existed across the harbour entrance.37 Further afield, it is clear that the Prankish outposts of Rouad, Satalia and Corycus were themselves heavily dependant upon fortifications to prevent their immediate recapture by the Muslims. Just as they had done in the Holy Land, the Franks relied on castles and city walls rather than superior troop numbers to gain new territory and intimidate their Muslim neighbours. As late as the 13605 the defences of Famagusta, Satalia and Corycus enabled Peter I to dominate local sea routes and to maintain the crusading tradition of his predecessors.

12

The Military Orders The role of Cypriot strongholds belonging to the Military Orders was very different from that of similar castles in the Holy Land. In Cyprus the strict royal monopoly on castle-building meant that the monarchy had far more control over national security than in other crusader states. Not only did the king garrison the defences of Famagusta, Nicosia, Paphos and Limassol, he also held the key strongholds of Kyrenia, St Hilarión, Buffavento and Kantara, which had been so instrumental in defeating the Lombards, containing the Genoese at Famagusta, and overcoming Amaury of Tyre's supporters in 1310. This was made possible by Cypriot tranquillity in general, for the Lusignans never needed to encourage the Military Orders to build vast fortresses like Crac des Chevaliers in order to defend the kingdom's frontiers. To have done so would merely have undermined the strategic value of Kyrenia and the mountain castles, for it would have led to the kind of political fragmentation which had often affected the mainland before 1291. Here we have seen that the Military Orders created virtually independent enclaves centred around castles like 'Atlit, whose defences were so powerful that their garrisons could simply ignore the will of local rulers.1 After the loss of Acre, this explains why Henry II feared the Military Orders, whose members fled to Cyprus in large numbers, almost as much as the Mamluks. In order to stop the Orders from slipping out of his control, he went to great lengths to prevent them from building up new estates. In 1299, for example, he was rebuked by Boniface VIII for imposing taxes on the Templars and the Hospitallers, and for refusing to allow the transfer of any lands to them without the express permission of the king or the pope.2 These tactics appear to have succeeded in blocking the rapid expansion of the Orders on Cyprus, and may well explain why the Templars supported Amaury of Tyre's take-over of power in 1306.3 Although the Hospitallers eventually favoured the king in this dispute, their campaign to conquer Rhodes, begun in 1306, may also have reflected their frustration and lack of independence on Cyprus. Indeed, the tension which existed between the Orders and the monarchy did not come to an end until 1307-8, by which time the Hospitallers had almost completed their conquest of Rhodes, and the Templars had been arrested for alleged heresy.4

THE M I L I T A R Y O R D E R S

131

The Military Orders therefore never gained the kind of power in Cyprus which they had once enjoyed on the mainland. This can be seen by giving a brief summary of their Cypriot castles and estates, starting with the Templars.5 The Templars had been the first Latins to rule Cyprus after Richard I's invasion; but, although Guy of Lusignan was later buried in a church established by them at Nicosia, it seems unlikely that they had had time to construct any new fortifications before their departure in 1192.0 Six years later, however, both this Order and the Hospitallers were asked by Innocent III to help defend Cyprus against external attackers, and in 1210 we find the first reference to the new Templar castle at Gastria, suggesting that the Templars were quick to re-establish themselves on the island during the 1190S.7 The fact that Gastria's true age cannot be fixed more accurately than this confirms that it is just as difficult to reconstruct Templar properties on Cyprus as it is everywhere else in the Latin East. In general this stems from the dissolution of the Templars and the transfer of their lands and castles to the Hospitallers, but in the case of Cyprus it may have been compounded by the events of 1279, when King Hugh III confiscated or even destroyed the Order's belongings because of its support for Charles of Anjou, Hugh's rival for the throne of Jerusalem.8 Thanks to these disasters, archaeological traces of Templar fortifications on Cyprus are rare and the records of which Order held which estates prior to 1307 have remained unclear ever since the sixteenth century. Hence Florio Bustron's list of Templar estates gained by the Hospitallers probably contains numerous properties which already belonged to them before 1307, whilst in 1580 it was even claimed that Gastria had been constructed by the Egyptian invasion force of 142.6.9 Despite these difficulties, some remains of the Templar defences at Gastria have survived to the present day. These were constructed on a small coastal promontory to the north of Famagusta, and consisted of a single curtain wall and rock-cut ditch protecting the site from the interior. To the north of this promontory, it seems that a small inlet served as a harbour for the castle. 10 Beyond Gastria, the Templars also held minor fortifications at Khirokitia and Yermasoia, two of their estates near Limassol. At Khirokitia, below the fifteenth-century ruins of a Hospitaller building, the remains of what appears to have been an older Templar tower were still visible in Enlart's day. It may have been here that the Marshal of the Order was held in 1307, when all the Templars on Cyprus were arrested. There are no such remains at nearby Yermasoia, although Florio Bustron noted that the Templar Commander was similarly incarcerated here, presumably in another small tower or fortified building, n Descriptions of Hugh Ill's seizure of Templar properties in 1279 also confirm that the Order had towers or houses

132

U N K N O W N C R U S A D E R CASTLES

at Limassol and Paphos, whilst in the spring of 1232 the Lombards are said to have occupied a similar building in Nicosia.12 A document dating from 1264 records that the Order later sold this place to two secular knights for 2000 bezants blanch The Hospitallers were granted many of these properties after 1307,14 including Gastria, Yermasoia and Khirokitia, where a fifteenth-century tower replaced its Templar predecessor.^ The Hospitallers also had their own properties at Limassol and Nicosia, which played a small role in the war between the Lombards and the Ibelins. Iri 1228 Frederick II had John of Ibelin's sons imprisoned in the Hospitallers' compound at Limassol, because it 'was strong and nearer his ships', implying that it may have been larger and more reliable as a prison than the royal castle itself.16 The following year the Ibelin supporter Philip of Novara described how he managed to avoid being captured by Frederick's five baillis and slipped away to the hospital at Nicosia along with 150 troops, plus many women and children of the Ibelin faction. This building cannot have been very strong, because Philip hastily equipped it with a cistern and a wooden palisade, making sure it was stocked with biscuits and other food suitable for a long siege. In the end these precautions were not needed, for shortly afterwards John of Ibelin arrived from Syria and defeated Frederick's baillis in battle. This enabled Philip to launch a sortie against his enemies, some of whom presumably still occupied Nicosia's royal citadel.17 The principal Hospitaller castle on Cyprus was located at Kolossi, to the west of Limassol. This estate had been granted to the Hospitallers by Hugh I in 1210 and subsequently became the Order's Grand Commandery, suggesting that although the impressive tower which stands there today was built in the mid fifteenth century, it was preceded by a similar older structure. Florio Bustron's claim that in 1307 the Hospitallers also inherited another tower at Kolossi from the Templars should probably be dismissed as a further example of the confusion which surrounds the latter Order's properties.18 On the other hand, it may be that the Hospitallers did own a second tower at Episkopi, another of their estates near Limassol. According to Estienne of Lusignan Episkopi 'was given by the king of Lusignan [Hugh I] to the Knights of St John of Jerusalem, who built and fortified the castle, before the island fell into the hands of the Turks'.19 The Teutonic Knights also held estates in the vicinity of Limassol, plus a couple of houses at Nicosia. These had been granted to the Order either during the reign of King Aimery (1196-1205), who received his royal crown from the German Emperor Henry VI, or during the late 12208, when Frederick II's influence was at its height.20 Although the Teutonic Knights held on to these properties after the defeat of the Lombards in 1233, their

THE M I L I T A R Y O R D E R S

133

strong imperialist links, as well as their later involvement in the Baltic, always restricted their influence on Cyprus. Moreover, there seems to be no firm evidence in the contemporary sources to confirm the claim that the Order held a castle at St George (near Limassol), although it may have possessed fortified houses on its estates.21 It is therefore clear that none of the Military Orders held any major fortifications in Cyprus. Even Kolossi and Gastria, which may have been the largest such strongholds, should really be described as fortified enclosures, towers or houses rather than large castles. The fact that these structures were not intended to withstand lengthy sieges is further highlighted by the hasty improvements which Philip of Novara had to carry out at the Hospitaller base in Nicosia in 1229. Similarly, the closest Gastria ever came to being directly involved in a major conflict occurred in June 1232, when Lombard troops defeated at the battle of Agridi sheltered briefly in the ditch before being rounded up by the Ibelins. The Templars, whose headquarters in Acre had been besieged by Frederick II three years previously, made no attempt to help them.22 Gastria's small size also makes it extremely unlikely that it had originally been constructed to stop the Byzantine empire from reconquering Cyprus during the iipos.23 It is possible, however, that Gastria and the Hospitaller compound in Limassol, both of which were situated near the sea, later played some role in the naval raids launched from Cyprus after the fall of Acre. We have seen that the Templars and the Hospitallers both took part in attacks on the Syrian coastline in 1300. Thereafter the Templars maintained a garrison on the island of Rouad until 1302, while the Hospitallers sent two expeditions into Armenia between 1300 and i3O5.24 Just as they were transferring their headquarters to Rhodes, the Hospitallers were also closely involved in a plan by Pope Clement V to launch a major new crusade.25 During this period it may be that Limassol and Gastria were used to collect troops and supplies destined for overseas expeditions or the Templar castle on Rouad. Indeed, following the loss of Acre, the Hospitallers decided that 'as long as the Master and the Convent are to be in the kingdom of Cyprus, the General Chapter should be held at Limassol', confirming that it was from here that many of their military campaigns were planned in the years immediately after i29i.26 The suitability of these sites for large-scale campaigns must, however, have been limited. Limassol was no longer being used as much as Famagusta for naval warfare, while the inlet at Gastria would have been too small to accommodate large war galleys. Other strongpoints belonging to the Orders, such as Khirokitia and Yermasoia, were not situated near the sea, and would have been useless even as coastal look-out posts. Instead, they were primarily

134

UNKNOWN CRUSADER

CASTLES

used as administrative and agricultural centres, and therefore only needed to be strong enough to protect crops and revenues against local bandits, rebels or pirates. Kolossi, for example, lay at the heart of a large agricultural complex and sugar plantation, and had a fortified barn and several other farm buildings situated close to it.2? Similarly, Khirokitia and Yermasoia, along with many estates belonging to the Teutonic Knights, were located in the rich sugar and cotton producing district around Limassol rather than an area of strategic military importance. A late fourteenth-century reference to salt still being kept in 'the warehouse of the Templars' suggests that this Order too had relied on Gastria and various other small fortified structures to protect its agricultural produce against thieves, troublemakers and the weather.28 The kings of Cyprus therefore achieved their goal of keeping the Military Orders in check, for strongholds which were only built for local defence, administration and farming could never hope to withstand royal besieging forces, and could not, therefore, undermine the authority of the monarchy.

PART THREE

Cilician Armenia 1187-1375

This page intentionally left blank

13

Warfare By the late twelfth century the Armenians living in Ciucia, which had originally been part of the Byzantine empire, had effectively become independent. This had come about through a combination of local geography and dramatic political change. By the iiSos military pressure applied by the Turks had caused Greek influence in eastern Asia Minor to disappear, and during the next two decades the Byzantine empire was racked by a series of political crises which culminated in the capture of Constantinople by the Fourth Crusade (1204). Meanwhile, the Armenians had nothing to fear from their Latin neighbours in the Holy Land, who were far too preoccupied with the struggle against Saladin to have any kind of real influence over Cuidan affairs. In addition, Cilicia's physical isolation gave it a certain amount of protection from the Seljuk Turks of the Anatolian interior, as well as Saladin's Ayyubid descendants in Syria. The Cilician interior, which was dominated by a large, fertile plain, was protected by the Mediterranean sea to the south and by a series of rugged mountain chains to the north, east and west. These chains could only be penetrated by a limited number of mountain passes, such as the Amanus Gates in the east or the famous Cilician Gates which led toward Constantinople. At times during the thirteenth century the Armenians enjoyed a certain amount of political control beyond these passes, particularly in the coastal regions to the west of Silifke and in the territories around Antioch. However, the Cilician Plain and the mountains around it always formed the core of their lands, and it was the castles situated within this region that were of central importance. For much of the twelfth century politics within Ciucia itself had revolved around the struggle between various powerful Armenian families trying to achieve local dominance. The two key dynasties involved in these clashes were the Hethoumids, who frequently sought aid from the Byzantine empire, and the Roupenids, who normally favoured greater independence from the Greeks. This goal was finally achieved in 1198 when Leon II, head of the Roupenid dynasty, overcame Hethoumid resistance and had himself crowned as the first king of the Armenians. Thereafter Leon's new kingdom increasingly resembled a feudal western state which prospered through

138

UNKNOWN CRUSADER CASTLES

international trade with Europe and Asia. Below the king, the Armenian Church flourished at cultural centres such as Lampron1 whilst the local nobility, whose many strongholds and estates were dotted across the countryside, grew rich on the agricultural wealth of the Cilician Plain.2 Throughout the thirteenth and fourteenth centuries, however, external attackers did sporadically penetrate Cilician Armenia and inflict serious damage. From the time of Leon II's rise to power until the mid thirteenth century, the most dangerous such opponents were the Seljuk Turks of the interior. In 1187, for example, they made a failed attack on Sis (the Armenian capital) which was successfully repelled near the Amamis Gates.3 About a quarter of a century later there was further fighting to the north-west of Marash involving Kay-Khushraw I, the Seljuk sultan of Rum (1204-11), and in 1216 his successor Kay-Kavus I (1211-20) attempted to capture the Armenian castle of Gabán, which was probably located somewhere in the same region. Although Kay-Kavus was eventually obliged to abandon this campaign, he still managed to inflict heavy losses on Armenian relieving forces sent to help Gabán, indicating that at this time the Seljuks of Caesarea (Kayseri) posed a serious threat to the Armenians' northern frontier.4 This period also marked the end of Armenian influence to the west of Silifke and the Cilician Gates, for in 1218 Leon II was forced to cede the fortresses of Loulon and Lausada, along with many neighbouring lands, in order to secure the release of Armenian nobles captured during the fighting for Gabán.5 Two years earlier the Seljuks had even launched an unsuccessful attack on Silifke itself,6 and they continued to keep up the pressure under Kay-Qubad I, whose sultanate (1220-37) witnessed the prosperous growth of key Turkish centres such as Antalya, Anamur and Alanya.? In the second half of the thirteenth century the chroniclers continued to record sporadic Seljuk raids, such as that which was halted to the north west of the Cilician Gates in 1259.8 Turkish influence over the region declined with the arrival of the Mongols, however, who defeated the Seljuks in 1243 and turned the sultanate of Rum into a protectorate.9 Thereafter this threat receded until the fourteenth century, when renewed fighting on the Cilician Plain heralded a period of Turkish resurgence which eventually culminated in the birth of the Ottoman empire.10 Until then, however, the rise of the Mamluks ensured that from the 12605 onwards the principal danger came not from the north but from the south. Starting with the reign of Baybars (1260-77), successive Mamluk sultans of Egypt gradually overran the crusader states in the Holy Land, and during this period the same process began to take place in Cilician Armenia. Thus in 1266 Baybars sent an army onto the Cilician Plain via the Amanus Gates, where an Armenian force led by Hethoum I (1226-69) tried in vain to halt it.11 Subsequently the Egyptians

WARFARE

139

MAP 4. Cuidan Armenia

launched regular attacks against the area, with particularly destructive raids taking place in 1275,1298 and 1322.12 These were followed by even more extensive campaigns in 1337 and 1375, the latter resulting in the permanent conquest of Cuidan Armenia and its incorporation into the Mamluk empire.13 During much of this period the Armenians could theoretically look for support from the Muslims' Mongol opponents, with whom they had wisely fostered friendly relations since the late 1240S.14 In 1260, for example, the Mongol conquest of Aleppo enabled them to extend their power east of the Amanus mountains and to occupy sites like Darbsak, a former Muslim stronghold which the Armenians continued to garrison until i268.15 But

140

UNKNOWN CRUSADER CASTLES

success stories of this kind should be contrasted with tales of Mongol unpredictability. In 1266, when the Armenians asked for help against Egyptian forces ravaging the kingdom, the Mongols arrived too late to confront the Mamluks and decided to go on their own looting spree instead.16 The Mongol conversion to Islam in 1304 did not help matters, for not long afterwards they treacherously murdered the Armenian king Leon IV (1301-7) at a supposedly friendly meeting below the Cilician fortress of Anavarza.17 These incidents suggest that the Mongols were often either too far away or too unreliable to provide consistent support, particularly during the fourteenth century. Indeed, in 1320-21 they carried out a substantial raid in the vicinity of Sis.18 In general terms, the Armenians found themselves increasingly outnumbered by their opponents in the course of the thirteenth century, and eventually proved powerless to stop the final Mamluk campaigns of 1337 and 1375. It is extremely difficult, however, to add detailed figures to these very general observations. The size of Seljuk armies, for example, were not usually recorded by the Armenians, whilst in other sources they tended to vary from 5000 to 800,000 troops.19 One thirteenth-century western source described a Seljuk army containing 11,000 soldiers, which certainly sounds more feasible than 800,000 even if it was still an exaggeration.20 Much more than 10,000 troops seems improbable when one bears in mind that the Armenians, although sometimes under severe pressure from the Seljuks, were never actually driven from the Cilician Plain by them. Indeed, in 1211 Leon II (1198-1219) organised a failed attempt to capture the Seljuk city of Caesarea to the north,21 whilst on another occasion he led a major punitive raid against his Turkish neighbours.22 In 1223 his successor Philip (1223—25) also reconquered some lands and castles which had been lost to the Seljuks shortly after Leon's death.23 These campaigns suggest that, while the Armenians may have been outnumbered by the Seljuk Turks, they were not necessarily overwhelmed by them. The same cannot be said for the Mamluks or the Mongols. It has been claimed that the Mongol conquest of Aleppo and Damascus undertaken in 1260 involved 120,000 troops, of whom 20,000 supposedly stayed behind to garrison Syria.24 Even if these figures were exaggerated, the ease with which the Mongols captured so many Muslim lands and cities makes it clear that they posed a far greater potential threat to the Armenians than the Seljuks. The cavalier fashion in which the Mongols acted when they ravaged Cilicia in 1266 and murdered Leon IV in 1307 also indicates that they regarded Cilician Armenia as a client state rather than an equal ally. But at least attacks by the Mongols were relatively rare; by contrast, the Mamluks remained consistently hostile throughout the late thirteenth and

WARFARE

141

fourteenth centuries. Historians have estimated that from the reign of Baybars (1260-77) onwards Mamluk armies sent against the Franks often contained more than 12,000 men, and the same numbers applied to the Egyptian forces which operated in Ciucia.25 For example, Hethoum I's attempt to halt a Mamluk advance near the Amanus Gates in 1266 appears to have been swept aside with relative ease, and resulted in the death of one of the king's sons and the capture of another.26 When the Mamluks returned in 1275, the Armenians actually managed to defeat the first wave of Muslim troops before being overwhelmed by another, which alone contained 8000 men.27 In 1298 the Mamluks were also said to have attacked the region with 20,000 troops drawn from Egypt, Aleppo and other parts of Syria.2« By the fourteenth century the situation had deteriorated even further, for in 1337 one chronicler wrote that 'the troops of the sultan of Egypt ... entered Cilicia with 60,000 cavalry troops'.29 This last figure looks impossibly high, but the fact that the Armenians lost many of their castles for good at this point proves that they must still have been hopelessly outnumbered. The effectiveness of an army depended on the quality as well as the quantity of its soldiers, and varied according to the intentions of its leaders. The Mongol raid of 1266, along with the Mamluk campaigns of 1266, 1275 and 1298, all seem to have been loosely organised expeditions primarily intended to gather booty. Thus in 1298 the Muslims did not press home their attack on Sis, the Armenian capital, because some sections of the army considered it more profitable to go off on an extensive looting spree.3° On the other hand, when they did wish to make permanent territorial conquests, both the Mamluks and the Seljuks could deploy plenty of skilled engineers and siege engines. During their attack on Gabán (1216), for example, the Seljuks used several siege weapons against the castle which the garrison only managed to destroy by launching a daring sortie.31 Similarly, when the Mamluks campaigned in Cilicia in 1298, they captured one fortress with the help of siege engineers who dug mines and erected wooden palisades to protect the attackers.32 We have already seen that by this date such tactics had already been well rehearsed in the Holy Land, where increasingly powerful trebuchets and greater use of professional siege engineers had enabled the Muslims to capture the last Prankish strongholds in the years before 1291. During the fourteenth century the same process brought about the demise of the Armenian kingdom.33 In response to such tactics, the Armenian kings could probably only muster relatively small armies. In 1275, for example, 500 horsemen were involved in the failed attempt to prevent Mamluk forces from penetrating the interior via the Amanus Gates.34 Bearing in mind that there were

142

UNKNOWN CRUSADER CASTLES

supposed to have been at least 8000 attackers confronting them, these men were presumably accompanied by other soldiers who were not necessarily on horseback. Similarly, when Oschin, lord of Corycus, defeated a Turkish incursion in 1318-19, he may have had more troops at his disposal than the 300 cavalry said to have been under his command.35 This implies that Armenian chroniclers, like many of their Latin counterparts, tended only to include noblemen or troops mounted on horseback, without necessarily giving specific figures for socially inferior foot-soldiers, archers and mercenaries. Sometimes enemy troop numbers could also be exaggerated in order to enhance the bravery of the Armenians themselves. This was clearly the case in 1246-47 when the chronicler Constable Sempad stated that 260,000 Turks were held at bay by the defenders of Tarsus.36 Such claims make it impossible to give an accurate figure for the total number of troops available to Armenian rulers. During the winter of 1264-65 Hethoum I, however, responding to an appeal from the Mongols, was said to have contributed 200 knights to the failed Mongol siege of al-Bira on the Euphrates.37 If we place this figure alongside that of the 300 cavalry referred to previously, and if we assume that Armenian chroniclers were at least slightly more trustworthy when they spoke about their own knights, we are left with the impression that there were rarely more than 500 mounted troops present on Armenian campaigns. The fact that the death of a mere twenty-one knights during the Muslim raid of 1322 was treated as a disaster may confirm this.38 But even reaching this rather vague estimate is easy compared with calculating the amount of infantry soldiers who accompanied the mounted troops. If Armenian forces were anything like their Byzantine counterparts, the infantry component of the army was largely made up of ill-disciplined recruits whose exact numbers were not even known to the commanders themselves.39 Under such circumstances it is difficult to know whether we can trust the claim that 5000 troops were involved in an internal conflict which broke out in 1220. The fact that this rebellion involved a substantial element of the Armenian nobility implies that kings could normally hope to gather at least this many soldiers together, although the total may actually have been much smaller.40 The figure of 18,000 men (including 4000 cavalry, 1000 regular infantry, 4000 crossbowmen and numerous siege engineers) given for Hethoum II's army in 1302 also sounds suspiciously high, particularly as it comes immediately after a romanticised account of Armenians effortlessly killing thousands of Muslim horsemen in battle.41 The absence of any trustworthy references to the size and composition of the royal Armenian army means that these figures should only be regarded as suggestions. Some additional information can be gleaned from other

WARFARE

143

campaigns. In 1265 Hethoum I (1226-70) managed to deter Baybars from attacking Lesser Armenia by gathering his forces together in the Belén Pass, a key route between Cilicia and Antioch. Not only does this incident provide us with a rare example of Armenians successfully defending their own frontiers against the Mamluks, it also sheds some light on the composition of the army. After the crisis was over, 'the troops of the king were able to return to their villages and settlements (bourgs}', whilst others who had responded to the 'general levy' also headed for home.42 This indicates that the army was in large part made up of various contingents fulfilling their military obligations, and that the different communities scattered across Cilicia were all expected to provide quotas of soldiers. Another incident shortly earlier gives further credence to this theory: in the early 12605 the Seljuks carried out a raid on the western frontier of the kingdom, and consequently 'the king, having set off, reached Tarsus, then arrived at Silifke, where mounted troops, foot-soldiers and porters were gathering'.43 Sadly no precise figures were given for this expedition, although it does confirm that the royal army included both cavalry and infantry, as well as ancillary non-combatants who were on this occasion required to carry wheat to the castle of Silifke. During the 12605 both Tarsus and Silifke belonged to the king,44 but apart from troops gathered within the royal domain, other contingents were led by individual barons. It has already been noted that in 1218 the Seljuks accepted lands and strongholds to the west of Silifke as a ransom for nobles who had been captured fighting alongside Leon II at Gabán two years earlier.45 The army which Hethoum I used to block the Belén Pass in 1265 also contained both royal and baronial forces,46 whilst a year later Hethoum's own sons helped to co-lead the disastrous campaign near the Amanus Gates. This latter encounter with the Muslims also involved Constable Sempad, older brother of Hethoum I and an important adviser to the king.47 Indeed, according to one chronicler, Sempad eventually met his own end in similar circumstances, for when the Muslim raiders of 1275 clashed with the Armenians, 'Baron Sempad himself, the uncle of the king [Leon III, 1270-89], was killed, and other nobles ... and thirteen other famous men, and 300 horsemen'.48 These incidents indicate that Armenian kings were heavily dependent upon soldiers provided by individual barons or other members of the royal family. Apart from mustering regular troops, it is clear that the Armenians were just as capable of training siege engineers and constructing catapults as their Seljuk or Mamluk rivals. At the beginning of the fourteenth century, one chronicler described the army of Hethoum II as containing 'numerous engineers for conducting the sieges' which were expected to take place during a forthcoming campaign with the Mongols.49 The Armenians also

144

UNKNOWN CRUSADER CASTLES

enjoyed a reputation for their expertise in siege warfare dating back to the twelfth century, when the Franks had employed one of their countrymen to oversee the Latin siege works at Tyre (1124) .5° A quarter of a century earlier another Armenian had been used by the Muslims to design an elaborate new gateway for the citadel of Cairo, presumably because it was felt that anyone who was good at constructing siege weapons would be just as good at building castles.si This knowledge was put to good use in Ciucia itself, where the Armenians held many impregnable strongholds on the rocky outcrops dotted around the Cilician Plain. The sheer strength of many local castles, aided by the ruggedness of the terrain, enabled the Armenians to retain their independence until as late as 1375, long after their field armies had become too outnumbered to put up any effective resistance against the Mamluk empire. This proved to be a burden as well as a blessing for many Armenian rulers, for it encouraged some nobles to rebel, safe in the knowledge that even the king would not be able to capture their remote mountain strongholds. Problems of this kind reached chronic levels in the fourteenth century but were evident from the very beginning of Leon II's reign. In many ways the military situation in Cilician Armenia resembled that of the Latin states further south. It has been suggested the Armenians could perhaps muster a field army of 5000 men, including a core of around 500 mounted troops. They were also renowned for their skills in siege warfare and capable of building extremely powerful fortifications. Even allowing for the extravagance of some contemporary chroniclers, however, they were invariably outnumbered by their Muslim opponents, and in particular by the Mamluk armies which eventually brought about their downfall in 1375. Like the Franks, they needed impregnable fortresses to withstand the Muslim onslaught, even though such structures inevitably undermined central authority. This dilemma ensured that castles played a very prominent role in the history of Cilician Armenia, both in terms of internal politics and external security.

14

Military Architecture Any discussion devoted to military architecture in Cilician Armenia is immediately hampered by two difficulties. First, the history of the kingdom was extremely badly recorded by contemporaries, particularly during the fourteenth century, when external invasions and internal rebellions led to almost constant political instability. Secondly, successive Roman, Byzantine, Arab, Crusader, Mamluk and Ottoman rulers controlled this region both before and after the lifespan of the Armenian kingdom. As a result, local fortifications present the historian with numerous problems of dating, attribution and interpretation. Although ruined strongholds still occupy countless local hilltops, many have no written history, whilst others contain such a bewildering mixture of architectural elements that it is impossible to establish their exact age. Despite these difficulties archaeologists studying the area have shown that several castles do contain substantial remains of what can be termed a distinct Armenian style. One such stronghold is Servantikar, which was built near the Amanus Gates on a roughly triangular plateau approximately 500 metres above sea level. The castle's close proximity to such an important route into Cilicia ensured that it witnessed a lot of warfare, for it was here that Egyptian armies attacked in 1266, 1275 and 1299.! On the last occasion the Muslims even sacked Servantikar, and in 1337 it appears to have been incorporated hito the Mamluk empire as part of a peace treaty negotiated with King Leon V (i32o-4i).2 Today the castle lies in ruins, but a contemporary source shows that even in its heyday it needed few man-made defences: '[Servantikar] is a strong citadel on an outcrop in a valley. Many of its sides do not have walls, as they are naturally defended by the cliff's edge'.3 Indeed, fortifications were only really necessary along the gently sloping east front, which was protected by a long, sinuous curtain wall flanked by numerous round or horseshoe towers. Any attackers were obliged to pass below these defences and expose themselves to lethal flanking fire in order to reach the entrance in the north-east corner. Here they would be confronted by the outer gateway, an 'L' shaped structure incorporating two further towers and a machicolated ceiling. Even if they overcame this obstacle, attackers would still only have

UNKNOWN CRUSADER CASTLES

Servantikar: a typical Armenian mountain castle. Note the horseshoe towers and absence of a central keep.

reached the lower castle. In order to capture the upper bailey, they needed to repeat the whole process by penetrating another curtain wall and gateway. The inner bailey itself contained numerous residential buildings and was dominated by a cluster of four powerful towers situated at the southern

MILITARY ARCHITECTURE

147

tip of the site. These towers were attached to adjoining curtain walls but could also act as a kind of keep or final refuge. Any efforts to capture Servantikar by storm would therefore have proved costly, whilst the generous provision of water cisterns within the walls ruled out the possibility of gaining entry through a swift blockade. Many of the features present at Servantikar were typical of Armenian fortifications throughout this region. First, there was the castle's location on a rocky promontory, which made it difficult to approach and impossible to breach using mines. The castle was so inaccessible that complete ramparts were not always needed. Where the Armenians did build curtain walls, these consistently followed the cliff's edge and therefore enhanced the natural strength of the site. This layout could then be repeated, for Servantikar was only one of many Armenian sites where two or more walled baileys created successive lines of defence. This in turn meant that towers could be used sparingly, for the irregular shape of the fortress created plenty of natural angles for flanking fire. Any towers which the Armenians did add were invariably round or horseshoe shaped, thereby making them better able to withstand earthquakes,4 catapults or battering rams. By attaching all towers to surrounding walls, rather than providing Servantikar with a free-standing keep, the castle's builders also strengthened its ramparts whilst at the same time creating even more vantage points for archers stationed on the wall walks. Numerous other building techniques used at Servantikar confirm that the Armenians were highly skilled in the art of siege warfare. Not only were those stretches of the curtain wall most likely to be hit by catapults extremely thick, their outer faces were also covered with small, square blocks of bossed stone which would be more resilient against bombardment. Servantikar also only had one gateway, which was flanked by two towers and could not be attacked head on because it was placed at a right angle to the actual rampart. The entrance was defended from above by slot machicolation, enabling troops in the upper chamber of the gatehouse to shoot at intruders through holes incorporated into the vaulting. All these features indicate that Servantikar had been carefully designed to maximise the defensive potential of the site and to make the use of siege weapons against the castle virtually impossible.5 The general appearance of Servantikar was very similar to that of Lampron, situated near the Cuidan Gates, and Vagha, another Armenian stronghold located in the mountains to the north of Sis. Both these castles were built on remote spurs hundreds of metres above sea level, and were defended by successive baileys rising toward impregnable inner citadels. Indeed, the upper parts of Lampron could only be reached via a circuitous ramp, which

148

UNKNOWN CRUSADER

CASTLES

had been cut out of the solid rock and was barely two metres wide. As a result, it could not be used by more than one attacker at a time.6 The approaches to Vagha were almost equally inaccessible, being additionally protected by several gateways, at least one of which incorporated the same kind of slot machicolation as that used at Servantikar.7 Many of these defensive features can also be seen in strongholds occupying outcrops on the Cilician Plain. The best preserved such sites are Gökvelioglou,8 Tumlu9 and Yuan, all of which lie to the south and south-west of Sis. The most famous of these was probably Yuan (Ilan Kale/The Castle of the Snakes), whose complex inner gateway and stout horse shoe towers represent some of the most impressive Armenian structures of the entire region.10 Through careful study and comparison scholars have concluded that these castles were predominantly built by the Armenians. Many of the outcrops they occupied, however, were so ideally suited to fortification that few could have been virgin sites before the Armenians arrived. It is interesting to speculate, for example, whether the twenty metre wide moat dividing the castle of Lampron from neighbouring mountains had initially been dug by the Greeks, for it bears a close resemblance to the famous rock hewn ditch at Saone, whose original excavation has been attributed to Byzantine rather than Latin engineers.11 Servantikar can also be compared with Kantara, St Hilarión and Buffavento, the three great mountain fortresses built by the Greeks on Cyprus using a similar combination of successive curtain walls.12 Indeed, the difficulties in trying to identify the original occupants of some sites have led different scholars to attribute Azgit, a remote mountain castle between Servantikar and Vagha, to both the Greeks and the Armenians.13 It should also be noted that castles like Yuan and Servantikar represented a tiny minority, for most strongholds in Cilician Armenia were far more complex architecturally and probably only contained a few Armenian repairs. Between Yuan and Sis, for example, the Armenians garrisoned the classical acropolis of Anavarza (Anazarbus), a Roman settlement which also shows evidence of early twelfth-century Prankish occupation. 1* Other Roman or Byzantine acropolis sites reoccupied by the Armenians included Bodrum,15 which lay a few miles north west of Servantikar, and Sis, the capital of the kingdom, is At other castles, such as nearby Toprak (Til Hamdoun), the paucity of Armenian fortifications was caused by later as well as earlier occupants. In 1212 the German pilgrim Willbrand of Oldenburg wrote that Toprak was a 'good strong fortress held by an [Armenian] nobleman',17 yet the present defences generally appear to be a Mamluk construction dating from the fourteenth century.18 Toprak was probably largely rebuilt after 1337, when according to the chronicler Nerses Balientz

MILITARY ARCHITECTURE

149

Yuan the Muslims acquired much of eastern Cilicia by treaty.19 This in turn suggests that the site had been severely damaged during previous Mamluk raids on the area, particularly those of 1266, 1298 and 1304.2° The original castle seen by Willbrand of Oldenburg may, however, not necessarily have been any more Armenian than the Mamluk one which replaced it, for the entire site occupies a much older partially man-made tell, and had probably been fortified in some way for centuries.

150

UNKNOWN CRUSADER CASTLES

The eclectic nature and sheer age of many fortified sites in the region can also make it difficult to establish whether the Armenians copied their predecessors or invented their own architectural style, which then influenced others. For example, while the Armenians often excavated deep ditches or built successive curtain walls to defend weak spots, this had already been done by Byzantine engineers both locally and elsewhere. These tactics were indeed so obvious and so commonly used to protect castles (including Latin strongholds such as Montfort) that it would be misleading to attribute them to any one group or period.21 It is more useful to concentrate on specific details such as masonry types and vaulting, an approach which immediately reveals interesting links between Armenian fortifications and crusader castles elsewhere. The Hospitaller strongholds of Crac des Chevaliers, Margat and Silifke are particularly significant in this respect, for all three incorporated the kind of complex gateways and horseshoe towers built by the Armenians at Servantikar. Silifke, which lay on the south-western fringes of the Cilician Plain, was granted to the Hospitallers by Leon II in 1210 and remained under their control for the next sixteen years. During that period the Order effectively rebuilt the entire castle; a project which coincided with major repairs at Crac and Margat.22 This link seems to have been limited to these three castles, for in general horseshoe towers were not at all popular with the Franks. On the other hand, other individual features such as slot machicolation, which 'appears to be an Armenian invention', did catch on and were used extensively in Latin castles of the period.23 It has also been noted that the Franks sometimes employed Armenian siege engineers, implying that they were highly regarded as military architects beyond Ciucia itself and that some of the features incorporated into their castles were genuine innovations.24 In other cases the Franks rejected Armenian precedents in favour of their own designs. It has been shown that many crusader castles were less irregular and relied more on square or rectangular towers, incorporating several gateways or posterns rather than a single inaccessible entrance. Several Prankish strongholds, including Tortosa, Chastel Blanc and Montfort, were built around a large, free-standing keep. This type of structure had been brought to the Holy Land from Europe and was not used by the Armenians. During the thirteenth century the Latins also used smooth as well as bossed masonry, and often built their walls on large, sloping revetments to give them extra strength.25 These differences between Armenian and Prankish military architecture can partly be explained in terms of local terrain. The need to maintain sea links with the West forced the Latins to build many castles and city ramparts on the coastal plain, where straight walls were more logical and man-made features such as revetments compensated for

MILITARY ARCHITECTURE

151

the absence of sheer cliffs. This does not rule out the possibility that the Armenians occupied, repaired or even constructed their own urban fortifications. There has been a tendency to assume that the Armenians only built remote mountain castles because of the spectacular remains at sites like Servantikar, Yuan and Vagha, but it is their very remoteness which has helped these strongholds to survive. It would be equally mistaken to deny that the Franks built any urban defences simply because the massive walls of Acre and Tyre have since disappeared. While it is true that the inaccessible location, sinuous ramparts and elaborate gateways of Servantikar represented a distinct style, this was not the only type of fortification constructed and used by the Armenians.

15

Fortifications and External Security All warfare between the Armenians and their external opponents was heavily influenced by the local terrain, for the Cilician Plain, which represented the heart of the Armenian kingdom, was protected on all sides by natural defensive barriers. In the south there was the gulf of Alexandretta, whilst to the west, north and east there lay the Taurus, Anti-Taurus and Amanus mountains. These peaks were bisected by a limited number of mountain passes which inevitably took on great significance as the only practical land routes into and out of the area. In the west, the most important of these routes was that situated along the coast in the vicinity of Silifke. To the north and north west there lay the Cilician Gates, the most famous means of reaching Constantinople, as well as a number of valleys which led toward the inland cities of Marash and Caesarea. Further east the Amanus Gates, close to Servantikar, represented an important route toward the Euphrates. Finally, anyone travelling from Antioch would use the Belén Pass, or possibly another neighbouring valley, in order to reach the south-eastern fringes of the Cilician Plain close to Alexandretta.i Knowing that external invaders would always have to use one or more of these routes, the Armenians made it a key element of their defensive strategy to meet their opponents in the mountains where attacking armies could be blocked before they spilled onto the Cilician Plain itself. In 1265, for example, King Hethoum I (1226-69) is said to have deterred Baybars from attacking the region by gathering troops together from various baronial castles as quickly as he could and stationing them in the Belén Pass, near the castle of Baghras.2 Similarly, when the Mamluks returned via the Amanus Gates the following year, King Hethoum's soldiers were waiting for them in the neighbouring forest of Mani. Just under a decade later the Armenians used the same tactics, for in 1275 they confronted the Egyptians in this region, not far from the castle of Servantikar. These later encounters ended in terrible defeat for the Armenians,3 but Hethoum's initial success in deterring Baybars from attacking the Belén Pass suggests that there was nothing wrong with the basic plan. Hethoum and his successors realised that it would be suicidal to meet numerically superior invasion forces in the open, so they tried to even the odds and make up for their lack of

F O R T I F I C A T I O N S AND E X T E R N A L S E C U R I T Y

153

troops by confronting their enemies in narrow, wooded mountain passes, where the terrain was ideally suited to ambushes. Whilst it was impossible to block valleys like the Amanus Gates and the Belén Pass completely, castles formed an important part of this defensive strategy. Outlying strongholds could act as look outposts, giving rulers early warning of hostile advances and enabling them to mobilise royal troops and other baronial units quickly. The same castles were then used to assemble and supply troops near the frontier, providing safe bases for soldiers who were advancing or in retreat. During the Mamluk attacks through the Amanus Gates in 1266,1275 and 1298 Servantikar was probably used in this way, and it must have provided some refuge for Armenians defeated in the forest of Mani. Similarly, in the early 12605 Hethoum I responded to Turkish raiding around the western frontier outpost of Silifke by bringing a large body of troops and supplies to the castle.4 Enemy forces might well be reluctant to bypass this type of stronghold, whose garrison could threaten their supply lines, and they would therefore be obliged to halt their advance and undertake a costly and time-consuming siege. These factors proved decisive during the fighting which took place between Leon II and the Seljuk Turks early in the thirteenth century. In 1211 Leon II found it impossible to capture the fortified Seljuk city of Caesarea, situated beyond his northern frontier.5 Five years later, however, the Seljuk advance on Ciucia got no further than the neighbouring Armenian castle of Gabán. The Seljuks quickly got bogged down in front of Gabán; and, even though they defeated a relieving army sent by Leon II, lack of progress still forced them to retreat back to the north.6 Hence it was the sheer strength of the local fortifications rather than any mutually agreed borders which defined the frontier between the Armenians and the Seljuks in this area. The ability of frontier castles to prevent or at least provide early warning of enemy offensives no doubt explains why Armenian rulers placed so much importance on their maintenance and construction. According to the Muslim chronicler al-Makrizi, Hethoum I built numerous towers to guard the mountains around Cilicia, suggesting that his efforts to halt Baybars' attacks in the Belén Pass and Amanus Gates inspired a more widespread scheme for national defence.7 In 1269 Hethoum retired to a monastery, but his successor Leon III (1270-89) also 'ordered the construction of a strong castle at the foot of Mount Taurus ... to defend this district and the famous route of Xoz Jor'. This route penetrated the mountains a few miles west of Vagha, therefore representing a potential entry point for invaders attacking from the north.8 As the massive incursions of the late thirteenth and fourteenth centuries show, all these precautions ultimately failed to stop the Muslims from

154

U N K N O W N CRUSADER CASTLES

ravaging the Cilician Plain. But, even after the soldiers and castles which defended the frontiers had been overcome, the Armenians could still use the strongholds of the interior as refuge points. Castles built on outcrops protruding from the plain, such as Yuan or the citadel of Sis, were so strong and so widely spread out that they could not all be captured by attackers, who were therefore prevented from making any permanent territorial conquests. As a result, the Armenians were able to wait inside their castles until lack of supplies forced their opponents to retreat, at which point temporarily lost areas could be reoccupied. When describing the Mamluk invasion of 1266, Vahram of Edessa noted that Muslim forces quickly occupied the countryside, but 'in front of the fortresses which they attacked, they failed miserably'.9 When the Muslims returned a few years later, Leon III found himself so outnumbered that he did not even attempt to confront them in the open, but even so 'those who occupied the fortified places, or had retreated to the fortresses, escaped the carnage'.10 There can be little doubt that the strength and proliferation of Cilician castles helped the Armenians to retain their independence until 1375 despite being so outnumbered. During attacks castles did not simply preserve the frontiers of the kingdom, they also sheltered hundreds of civilians living in surrounding areas. This is confirmed by accounts of what happened when attackers actually managed to break into Armenian fortresses. In 1298 one stronghold taken in the vicinity of Toprak was found to contain 'a large throng, composed of peasants, farmers' wives and children'.11 After this castle had surrendered, al-Makrizi wrote that a further 'eleven places of the Armenians similarly fell under the control of the victors'. Although these sites were later reoccupied with the help of the Mongols, this disaster must have inflicted heavy suffering on the local population.12 Muslim sources also contradict Vahram of Edessa's claim that the Mamluk raid of 1266 failed to capture any local strongholds, for according to Ibn al-Furat the castle of Amouda, which lay on a small hilltop between Toprak and Sis, was forced to surrender at this time. Inside the Muslims found '2200 people, both fighting men and others, of whom the men were killed and the captives distributed amongst the troops'.13 Willbrand of Oldenburg's remarks that Amouda provided a place of shelter for local people and that the nearby river yielded good supplies of fish give further weight to the impression that the castle acted as a refuge point for many neighbouring communities.14 Interesting comparisons can be made between these events and those which were taking place in the crusader states during the thirteenth century. Like the Armenians, the Franks relied on their strongest castles to protect themselves and to prevent permanent territorial losses, even if their Muslim opponents sometimes occupied the land. In 1244, for example, we have

F O R T I F I C A T I O N S AND E X T E R N A L S E C U R I T Y

155

seen that the Latins survived the Khwarizmian invasion of the Holy Land because massive fortresses like Saphet and Pilgrims' Castle held out. More poorly defended sites, however, were overrun, most notably Jerusalem, whose undergarrisoned and dilapidated walls could not prevent the massacre of thousands of Christians inside the city.15 Jerusalem shared the same fate as Amouda because neither site proved capable of withstanding major external invasion forces. Their tragic history can be contrasted with that of the remote and virtually impregnable mountain strongholds situated in less accessible locations. Vagha, for example, continued to act as the residence of the Catholicos, or head of the Armenian Church, well into the fifteenth century.16 The lords of Gabán, whose exact location somewhere to the north of Sis has never been established, seem to have retained a certain amount of independence until the same period.17 Indeed, Sis itself was not permanently conquered by the Mamluks until 1375, when the citadel was captured and the surrounding city destroyed.18 Consequently, whilst all Armenian castles contributed to the overall security of the region, a distinction needs to be made between the more exposed sites such as Amouda and strongholds situated along the mountainous northern fringes of the Cilician Plain. These latter castles proved better able to withstand the Muslims, in the same way that formidable crusader strongpoints like Pilgrims' Castle held out long after more vulnerable sites like Jerusalem had been lost. In their long struggle to survive the Muslim onslaught, the Armenians were also aided by wider political factors which had nothing to do with the strength of their mountain fortresses. Many of the thirteenth-century attacks launched by the Seljuks or by the Mamluks were probably not in fact intended to conquer the Cilician Plain. The relatively rapid destruction of Leon II's power to the west of Silifke between 1200 and 1220 suggests that the Seljuks certainly had the military capacity to capture Armenian fortresses given the right circumstances. However, later setbacks, such as the arrival of the Mongols, usually ensured that subsequent Turkish attacks amounted to little more than temporary raids, and they did not pose a serious threat again until the fourteenth century. The Mamluk campaigns of 1266, 1275 and 1298 were all primarily designed to secure loot rather than new lands.19 By contrast, Baybars captured several Armenian-held castles to the east of the Amanus mountains and incorporated them into the Mamluk empire as early as 1268, shortly after he had taken the city of Antioch from the Franks. Many of these castles, abandoned by their former Muslim owners during the devastating Mongol invasion of 1260, had then passed under the control of Bohemond VI of Antioch or the Armenians.20 Perhaps, therefore, Baybars was keen to regain these sites in order to make up for the humiliation of 1260 and to ensure that the formerly Christian principality of Antioch

156

U N K N O W N CRUSADER CASTLES

could now be turned into a strong buffer zone against any future Mongol aggression. All these events prove that, when it suited him, Baybars was perfectly capable of making permanent territorial gains at the expense of the Armenians, but instead he chose to treat the Cuidan Plain itself as a source of tribute and booty which could be exploited from time to time. The desire for quick booty at the expense of few casualties helps to explain a number of late thirteenth-century incidents relating to the Cuidan Plain. In 1265 Baybars may have called off his expedition to the area because Hethoum I had discovered the plan and had mustered a field army in response, thereby removing all hope of a profitable surprise raid without any casualties. The following year Mamluk troops returned to the region, but the fact that they only attacked relatively easy targets such as Amouda and appear to have ignored tougher castles in the mountains suggests that they had come to plunder rather than to conquer. In 1298 an attempt to capture Sis was abandoned in favour of a much easier looting spree across the Cilician countryside.21 A peace treaty between Leon III and Kalavun dating from 1285 also indicates that for much of this period the sultans of Egypt relied on the mere threat of punitive raids to extract large amounts of tribute.22 Non-payment by the Armenians resulted in further Mamluk incursions in the opening years of the fourteenth century. According to al-Makrizi, the Muslims pillaged the slopes below the citadel of Sis in 1302 and sacked Toprak in i304.23 Two years later 2000 soldiers were sent from Aleppo because of further Armenian refusals to pay up, but these troops were defeated with the aid of the Mongols, who had themselves come to Sis to claim tribute as the price of their friendship. The Egyptian authorities reacted angrily, dispatching a further 4000 men who were finally able to collect the outstanding payment, presumably because the Mongols had now departed.24 These raids reflect a continuing interest in temporary plunder rather than permanent conquest, for they involved relatively few troops and continued to avoid difficult targets, such as the actual citadel of Sis, in favour of somewhat easier prey like Toprak. Any Armenian castles captured by the Muslims in this period were then either abandoned deliberately or later given up in the face of Mongol pressure. Despite being overrun in 1298, Toprak had been regarrisoned by Armenian troops by 1304, for the Mamluks had not actually held on to the castle after they had looted it. Strongholds do not seem to have been captured for the sake of territorial gain rather than booty until later in the fourteenth century. It is clear that 1337 was a turning point in this respect, for in that year 'the troops of the sultan of Egypt ... entered Cilicia with 60,000 cavalry troops and besieged Ayas ... they would not leave until the town had been delivered to them, along with all the land between the

F O R T I F I C A T I O N S AND E X T E R N A L S E C U R I T Y

157

Ceyhan river and the territory of the Arabs, land where there lay forty castles and fortresses, each with its own lord. These were abandoned to the Arabs voluntarily and by treaty'.25 The Ceyhan river ran through the heart of Cilicia, from the mountainous interior around Marash to the Mediterranean coast near Ayas. Consequently, if this treaty was carried out to the letter, the Armenians lost many of their most important castles, including Servantikar, Yuan and Toprak. It has already been noted that the current remains of the last fortress are predominantly Mamluk, suggesting that after 1337 Muslim policy changed from sacking Toprak to repairing it and placing it under permanent Egyptian rule.26 Once this shift in Mamluk thinking had taken place, the Armenians were doomed, for they could no longer pursue a policy of waiting inside their mountain castles and then reoccupying the land once temporary Muslim raiding expeditions had withdrawn. When the Mamluks made their final assault on Sis in 1375, they were no longer content to sack the lower city but besieged the citadel itself, refusing to give up until it had been captured and the last Armenian king had been led away in captivity.27 Returning to the treaty of 1337, it is clear that during these final decades of Armenian independence Mamluk incursions were beginning to inflict such widespread damage on the economy that the Muslims could demand the surrender of castles which they had not even attacked directly. By this point one chronicler noted that Muslim tactics had 'made a desert of the land of the Armenians',28 implying that many local garrisons no longer had the will or the resources to resist even if they did occupy virtually impregnable mountain strongholds. Once again, the end of Cilician Armenia as an independent kingdom can be compared with the demise of the crusader states before 1291. These had survived temporary looting sprees such as the Khwarizmian occupation of Palestine but could not prevent Baybars and his successors from picking off even the strongest crusader strongholds through a combination of devastating raids and systematic sieges.29 The ability of the Armenians to hold out for almost ninety years after the fall of Acre should still be regarded as a great achievement. It can be attributed to the ruggedness of the terrain and the sheer strength of castles like Sis or Lampron. Indeed, the impressive remains of such fortresses have led some scholars to argue that there was another factor which contributed to the remarkable resilience of the Armenians: visual communication. The archaeologist R. W. Edwards, for example, has written that 'most of the forts have intervisibility which allows for rapid communication and the efficient mustering of troops'.30 Edwards adds that 'every garrison fort in Cilicia Pedias has intervisibility with at least two other forts in the plain and most can communicate directly with the highland valleys'.31 This theory has

158

U N K N O W N CRUSADER CASTLES

been taken even further by the scholar Hansgerd Hellenkemper, who has concluded that in the early thirteenth century Leon II deliberately constructed a large number of Cilician strongpoints to fit into a vast network that eventually included Ayas, Misis, Yuan, Gökvelioglu, Tumlu, Anavarza, Toprak, Amouda, Harunia, Ak Kale and Bodrum. These, it is argued by Edwards and Hellenkemper, could all communicate with each other using fire or smoke signals, thus giving early warning of an imminent Muslim invasion.32 This conclusion can be backed up by a number of written and archaeological sources. The deliberate creation of an intervisible network required the presence of a strong central ruler and Leon II seems the most likely candidate. Not only did his reign (1198-1219) mark the zenith of Armenian power in many ways,33 some chroniclers also made direct reference to his interest in castles. Michael the Syrian wrote that 'the valorous Leon extended his domination over seventy-two fortresses',34 whilst Vahram of Edessa claimed that 'he constructed a number of castles and strongholds, with which he surrounded Cilicia'.35 Even more significantly, the twelfth-century keep at Anavarza actually has a surviving inscription which commemorates repair work carried out there by Leon II.36 A similar feature dating from 1206 has been discovered on the walls of the sea castle at Corycus,37 whilst some have argued that a carved relief above the inner gateway at Yuan represents an enthroned Leon after his coronation in 1198.38 The close architectural links between this latter fortress and several other mountain castles, all of which have the familiar Armenian combination of horseshoe towers and sinuous curtain walls, could also suggest that they were erected by the same man. Edwards himself, however, has sounded a note of caution about dating so many strongholds to such a brief period of time. He has pointed out that the inscription at Anavarza dates from 1188, ten years before Leon II became king, and that both this inscription and that at Corycus only commemorate repair work rather than new foundations. The idea that the relief at Yilan depicts Leon II is also more of a tradition than a fact, whilst the building style used in this and many other castles is common to all Armenian military architecture between the tenth and fourteenth centuries. Therefore Edwards concludes that it is an oversimplification to attribute so many Cilician fortresses to one king, but he still argues in favour of a deliberate intervisible network of castles created by the Armenians over a longer period of time.39 Even if such a network had been built up over the entire two-hundred-year period between the growth of Armenian independence in the late twelfth century and the fall of the kingdom in 1375, numerous difficulties remain.

F O R T I F I C A T I O N S AND E X T E R N A L S E C U R I T Y

159

For a start, the architectural history of most local fortifications is extremely complex, with many sites such as Anavarza dating back to Roman times. Even castles which today contain an overwhelming amount of Armenian remains, such as Yilan and Servantikar, probably have foundations which are considerably older.4" Another problem is the nature of castle ownership, for the few written sources available on this subject make it clear that many fortresses were owned by individual families for decades or even centuries. The list of forty-five barons who attended Leon II's coronation in 1198, for example, indicates that Servantikar, Harunia, Gabán, Amouda, Toprak, Lampron, Corycus and Silifke all belonged to separate lords at that tune, as did numerous other sites whose exact location cannot be established.41 It is evident that baronial ownership of castles was equally widespread one hundred and forty years later: when the Mamluks laid claim to much of eastern Ciucia in 1337, this region contained 'forty castles and fortresses, each with its own lord'.42 It will also be shown below that one or two of the castles supposedly belonging to the network, such as Amouda and Harunia, were owned and partially rebuilt by the Military Orders in the thirteenth century.43 Consequently these strongholds had been built or repaired by a wide range of people at many different times, and most occupied sites which had been chosen long before the twelfth century. Under these circumstances it is difficult to see how Armenian kings could have forced their subjects to alter or construct castles so that they fitted into a complex early warning system covering much of Cilician Armenia. The proliferation of castles from many different ages did not simply rule out the possibility of an intervisible network for practical reasons; it also put strict political limitations on Leon II and his successors. The network could only have worked if there had been close cooperation between the various royal and non-royal garrisons stationed across the Cilician Plain, but in reality the sheer strength and isolation of many baronial castles encouraged their owners to disobey rather than help their rulers. The causes and consequences of these rebellions will be discussed below, but their frequency throughout the period, and in particular the fourteenth century, is another argument against the existence of a widespread early warning system.44 On the other hand, a more limited network may have existed involving only royal castles, whose garrisons could be expected to remain loyal at all times. References made by chroniclers to the castle-building activities of both Leon II (1187-1219) and his successor Hethoum I (1226-69) certainly suggest that these rulers tried to construct watchtowers or similar structures to guard important mountain passes.45 Even these activities cannot be investigated in any depth, because the true extent of the royal domain remains uncertain. Some information can be gleaned from the baronial list

l60

U N K N O W N CRUSADER CASTLES

of 1198, for several major sites which are missing clearly belonged to the royal domain, including the capital Sis, the port of Ayas and the cities of Tarsus, Misis and Adana. Other strongpoints omitted from the list, such as Tumlu, Yuan and Vagha, presumably also belonged to the king, but the possibility that these were actually baronial castles which had not yet been constructed in 1198 cannot be ruled out completely. These difficulties make it virtually impossible to prove whether or not royal castles formed their own intervisible network.46 It is also unclear whether such a network would have been particularly practical anyway. The purpose of a nationwide network would have been to provide an early warning system, so that the garrison of a fortress like Toprak, which was situated relatively close to the Belén Pass and the Amanus Gates, could send fire or smoke signals to warn the king and his barons of an imminent Muslim attack. For this system to have worked, part of each castle garrison would have been required to keep a constant look out and be prepared to light some form of large beacon at short notice. Even in ideal weather conditions, this would not always have been easy. At Anavarza, for example, one scholar has noted that it is possible to see both Yuan and Tumlu from the parapet of the keep, but the roof of this structure can only be reached via three cramped and partially blocked storeys, making it unsuitable for building bonfires.47 The keep was constructed by the Franks early in the twelfth century, possibly not long after the First Crusade, indicating that it was not intended to fit into any subsequent Armenian defence schemes.48 The usefulness of warning systems which rely on simple beacons is also limited because it is difficult to describe the nature of an invading army or the direction in which it is travelling. It is equally problematic for anyone who receives a fire signal to send anything more than a prearranged reply, whilst the cost of maintaining the whole network throughout the year may have been prohibitive. All these problems appear to have persuaded the Byzantine authorities to disband a similar system many centuries earlier. Their chain of beacons had stretched all the way across Asia Minor to Constantinople and had as its starting point the fortress of Loulon, which was situated near the Cilician Gates and could therefore be used to observe Arab raids in the region. Its abandonment long before the Greeks actually lost control over this area suggests that it proved unworkable.49 A variety of practical and political difficulties therefore make it unlikely that the Armenian kings were able to create a nationwide network of intervisible strongholds. This need not rule out the possibility that individual castles could sometimes light beacons to warn local people working in the countryside, or that fire signals were used between a very small number of

FORTIFICATIONS

AND E X T E R N A L S E C U R I T Y

101

castles. We have already seen that a limited network of this kind proved perfectly satisfactory on Cyprus, where the defenders of Buffavento, an isolated mountain fortress which enjoyed extensive views over the island's coastal areas, could apparently warn other, less elevated castles of an imminent pirate attack. In this case it is worth remembering that all the strongholds concerned were held by the king, which presumably made it easier to supervise the system.50 Perhaps individual Armenian castles were occasionally constructed with intervisibility in mind, especially the new strongholds said to have been built by Leon II and Hethoum I for the defence of the mountain passes.51 Recent research carried out on the Aegean island of Euboea indicates that intervisibility between two or more strongpoints need not necessarily have been deliberate. It has been shown that the numerous Latin towers dotted across the Euboean countryside were not intended for fire signals, as was once thought, but were used as fortified farmsteads, trading posts and status symbols. Their location was not determined by strategic thinking but by more mundane needs such as water supply and proximity to good agricultural land. Any intervisibility between towers was purely coincidental, and perhaps inevitable when one considers that people tend to build fortifications on hill tops.52 Similarly, the many small strongpoints built by the Franks in areas such as the plain of Akkar were simply fortified agricultural centres rather than clusters of look-out posts which had deliberately been placed so that they could communicate with each other.53 Returning to Cilician Armenia, this point is epitomised by the fortress of Amouda. Amouda sheltered thousands of civilians during the Mamluk raid of 1266, whilst its location next to the Ceyhan river provided it with plenty of fish and probably made it a useful place to collect tolls from people travelling between Sis and the Amanus Gates. All these factors suggest that the outcrop occupied by the castle was chosen because it lay at the heart of a well-populated area good for fishing and farming. The keep at Amouda was also clearly built by its thirteenth-century owners, the Teutonic Knights, with little regard to local building techniques and on a site which had already been fortified in some way for many centuries. In other words, the fact that other castles can be seen from the ramparts of Amouda is probably a coincidence, having nothing to do with the primary functions of the stronghold as a centre for local trade, farming and defence. As was the case in the Holy Land, most castles therefore had no wider strategic importance beyond their immediate surroundings54 Another explanation sometimes given for the ability of the Armenians to withstand the Muslims for so long was their dislike of urban life. This theory has probably again been influenced by the survival of so many Armenian

102

UNKNOWN CRUSADER CASTLES

fortifications at remote mountain sites such as Yilan or Lampron, which can be contrasted with the lack of impressive archaeological remains at most low-lying urban settlements. It has been claimed that 'there is no evidence that city walls were ever constructed by the Armenians. Nor is there any mention that the Roman-Byzantine-Arab walls around cities like Tarsus and Adana were ever repaired'.55 This was because the Armenians were a mountain people who supposedly had a natural aversion to city life. Instead much emphasis has been placed on 'the rural nature of Armenian society', which largely consisted of small villages dotted across the countryside.56 Willbrand of Oldenburg, who travelled across Cilician Armenia in 1212, also reported that, although Tarsus, Adana and Misis were still inhabited, their city walls were in a state of decay.57 This implies that these cities were gradually dying, their Armenian rulers deliberately letting their defences fall into disrepair in order to concentrate on the mountain castles which proved so effective against the Muslim incursions. The archaeological evidence appears to confirm this at sites like Anavarza, a Roman settlement which was located in a valley overlooked by an ancient acropolis. During the Armenian period the walls which surrounded the actual town were left in ruins but the defences on top of the rugged, inaccessible acropolis were extensively repaired. These repairs were commemorated in a plaque dating from ii88.58 This policy has been described as 'the triumph of a non-urban strategy' and has been contrasted with the Latin tendency of fortifying low-lying coastal cities like Acre, which were so vulnerable that they had all been overrun by I29i.59 Whilst it is certainly true that a stronghold like Anavarza was more difficult for attackers to approach than a city like Acre, this argument can be criticised on several grounds. First, it has already been suggested that other factors such as the sheer distance between Egypt and the Cilician Plain, plus the long-standing Muslim policy of raiding rather than conquering this area, help to explain why Cilician Armenia outlived the crusader states.60 Considering that it was so much closer to Egypt and was therefore exposed to far more Mamluk raids, Acre's ability to survive until 1291 seems just as impressive.61 The Armenians may also not have differed from the Latins quite so dramatically: although many lived in the countryside, it is clear that others did inhabit a variety of urban communities. At the smaller end of the urban scale, some lived in settlements which had sprung up next to individual castles. For example, when Hethoum I gave the Teutonic Knights the fortress of Harunia (situated to the north of Servantikar) in 1236, the relevant charter specifically stated that the Knights also received 'civitatem Haroniam'.62 This indicates that under the previous owner, named as an Armenian baron in Leon II's coronation list of upS,63 a small town

F O R T I F I C A T I O N S AND E X T E R N A L S E C U R I T Y

163

had already been attached to the fortress. This settlement must have been very similar to the bourgs or castle towns in the Holy Land, such as that which stood next to the Templar fortress of 'Atlit.64 It is clear that communities of this kind existed elsewhere in Cilicia, and that even remote mountain castles like Vagha, far from existing in complete isolation, usually had some form of settlement located on the slopes below them.65 Despite the evidence for ruined walls at places like Tarsus, the Armenians may not have been hostile to the idea of living in larger cities. The impression given by Willbrand of Oldenburg that Tarsus, Misis and Adana were past their prime may not have been caused by deliberate royal policy but could simply have reflected the general decline in urban numbers which had taken place since classical times. By comparison, the Greek and Armenian population of Antioch had shrunk considerably since the fifth century, even though the Franks who ruled the city during the crusader period were supposedly far keener on the idea of urban living. We have seen that the vast ramparts built by Theodosius and Justinian around Antioch were a mixed blessing to the Franks: they were extremely powerful but must have been difficult and expensive to maintain for a city whose population may have dropped by as much as two thirds.66 Perhaps the Roman-Byzantine walls of cities on the Cilician Plain had been allowed to decay because it was a waste of effort to maintain them, even though many Greeks and Armenians still lived there. It is also possible that the ruins seen by Willbrand of Oldenburg only represented a temporary state of affairs, for other sources indicate that Misis, Tarsus and Adana were all properly fortified both before and after Wulbrand's visit of 1212. In 1151, for example, the Greek general Andronikos Komnenos, sent into the region to try to re-establish waning Byzantine control there, found his progress halted by the defiant Armenian citizens of Misis, whose leader Thoros eventually defeated the Greeks outside the city. An Armenian account of these events leave us in no doubt that the walls of Misis were intact at this point.67 Similar evidence has survived from 1221, when Leon IFs death two years earlier led to a violent succession dispute between his great nephew Raymond Roupen and his daughter Isabelle, who was aided in the struggle by the powerful Hethoumid baron Constantine.68 According to Constable Sempad this conflict came to an end when Constantine trapped Raymond Roupen's supporters inside Tarsus. Constable Sempad's description of this incident is worth quoting in full: So the lords of Tarsus, having returned, re-entered their city and then, having shut the gates, mounted the ramparts in order to fight off their pursuers. But a man in the city whose name was Vasil made contact with [Constantine], who

104

U N K N O W N CRUSADER CASTLES

promised him whatever he wanted. The man opened the gates fully during the night; [Constantine] and his troops penetrated the city where they pillaged the properties of the Greeks. But the rebellious lords fled the city and climbed up to the citadel which was strong and impregnable thanks to its defences; eventually the skilled ... Constantine overcame them without bloodshed by means of negotiation .. ,69

From these events it is clear that within eight years of Willbrand of Oldenburg's departure from Tarsus its walls were perfectly intact, with gates, walk-ways and a powerful citadel. Indeed, this latter structure had been mentioned by Willbrand himself, who noted that the city had 'a strong and good castle at its highest point'.70 Perhaps the walls described by Willbrand only had a dilapidated appearance but were still capable of keeping attackers out. Alternatively, they may have been damaged fairly recently and had not been in a permanent state of decay since classical times. One explanation for this may have been a local earthquake such as those which damaged many Syrian castles in 1170,1200/01 and i202/03.71 Another cause may have been warfare, for during a much later Muslim incursion in 1246 certain stretches of the ramparts were partially demolished.72 If either of these mishaps had taken place before Willbrand's visit, the citadel which he saw may have been the first stage in a major restoration programme undertaken at this time and completed before the civil war of 1220. At any rate, the history of Tarsus after 1212 clearly indicates that the Armenians did in fact build and maintain urban fortifications.73 The known history of other settlements also suggests that, far from turning their backs on urban life, the Armenians were determined not to give it up. The capital of Sis, for example, was the target of repeated Muslim raids in 1266, 1275, 1278, 1298, 1302, 1337 and 1375, as well as a Mongol attack in 132O.74 The frequency of these offensives suggests that after each raid the city was restored and regained enough prosperity for attackers to be lured back again. In 1212 Willbrand of Oldenburg wrote that this was another settlement without complete walls, yet the citadel which occupied the outcrop above it was so powerful that the citizens could shelter there with impunity.75 In 1275 the Muslims 'made their way as far as Sis, but they found no man therein, for they had all sought protection for themselves in the citadel, and piece by piece [the Muslims] burnt [the city]'.76 A century later, during the final siege of Sis, numerous early Mamluk attacks on this stronghold failed 'because the castle was very strong, and well manned, and well equipped with stones and trebuchets on the walls, which the king [Leon VI, 1374-75] had placed there, and so there was nowhere the castle could be attacked except in front of the gate'.77 It would be wrong to conclude that no provision had been made for the defence of Sis, for its

FORTIFICATIONS

AND E X T E R N A L S E C U R I T Y

165

citadel was clearly well prepared for a long siege even as late as 1375, and had enabled the city's inhabitants to emerge unscathed from numerous earlier hostile incursions. The present remains of Sis indicate that the slopes below the citadel were also protected by outer walls and defences, many of which must presumably have been built after Willbrand of Oldenburg's visit.78 The tenacity with which the Armenians rebuilt their cities after Muslim attacks is also evident at the ports of Ayas (Lajazzo) and Corycus. The limited number of natural harbours along the Cilician coast gave these settlements great economic importance as the key trade links between the West, Lesser Armenia and the Mongol territories to the east of Ciucia, particularly after the fall of Acre in ligi.79 These factors also made them attractive to hostile raiders, so the Armenians were careful to maintain the Roman, Greek and Arab defences which protected them. At both sites these defences consisted of a land castle overlooking the town and a sea castle built on an island in the harbour. The eclectic history of many of these structures can best be illustrated by describing the land castle of Corycus, which consisted of two square enclosures built in concentric fashion and flanked by a variety of square or polygonal towers. These fortifications had been constructed using anything that came to hand, including a Roman triumphal arch incorporated into the inner ramparti It is clear that both this fortress and its neighbour in the harbour were garrisoned by the Armenians, for a plaque has survived on the latter stronghold which commemorates repairs carried out there by Leon II in 1206.81 Indeed, at Ayas it is possible that the entire sea castle was rebuilt by the Armenians, for the chronicler Bar Hebraeus wrote that when the Muslims attacked the city in 1282 'they did not find in it one of its inhabitants, for they had all fled to sea, and had gone into a new fortress which they had built out in the sea'.82 This incident shows that the citizens of Ayas and Corycus were able to survive enemy attacks by retreating into powerful citadels, just like the inhabitants of Sis did. Only seven years before the raid described by Bar Hebraeus 8000 Mamluk troops had entered Ayas 'and they killed those whom they found inside' but were unable to reach those inhabitants who had fled to the sea castle. These people were therefore able to escape with their lives, although they were subsequently turned upon by Prankish pirates who 'even carried off their cloaks' in a tragic example of Christian disunity.83 Nevertheless, Ayas as a whole recovered from both this incident and the 1282 raid; and between 1321 and 1322 the city even survived consecutive attacks by the Mongols and the Mamluks. The latter offensive resulted in the capture of the sea castle 'by various means and bombardment with

l66

U N K N O W N CRUSADER CASTLES

catapult',84 suggesting that the Muslims had come better prepared and were now thinking in terms of more permanent territorial expansion. However, according to a peace treaty with the Egyptians negotiated in 1323, the sultan agreed to make good the damage he had caused both at Ayas and elsewhere.85 A document dating from June 1328 also records the arrangements made by Pope John XXII to contribute 30,000 florins 'toward the rebuilding and construction and fortification of the sea and land castles of Ayas and other places in the kingdom of Armenia'.86 Whether or not these combined efforts led to the total restoration of all the affected strongholds cannot be ascertained, although Ayas did not finally come under Egyptian rule until the Mamluk campaign of 1337.87 Throughout this period the Armenians did not simply try to keep Ayas alive by repopulating its houses and rebuilding its defences. Successive kings also granted privileges to western merchants, who were allowed to trade there in return for paying tolls to royal officials. Such privileges were granted to the Genoese in 1201, 1215, 1216, 1288 and 1289, to the Venetians in 1201, 1245,1271, 1307,1321 and 1333, and to the Pisans in m6.88 Further privileges were granted to merchants from Sicily (1331), Montpelier (1314, 1321) and Catalonia (1293), as well as to the Florentine banking company, the Bardi (i335)-89 Bearing in mind that these documents continued to be issued both before and after enemy incursions, and that the last one dates from 1335, a mere two years before the final Mamluk conquest of Ayas, it is clear that the Armenians did their best to maintain and expand the economic life of the city until the very end. That this policy was highly successful can be seen from Marco Polo's description of Ayas in 1295 'as the market for all the riches of the East. All the spicery, and the cloths of silk and gold, and other valuable products that come from the interior are brought to that city ... Whoever would travel to the interior takes his way by this city of Layas'.90 This does not fit in with the idea that the Armenians ignored the cities on the plain, preferring to live only in the mountains and the countryside. It has been argued throughout this chapter that they actually built or garrisoned a wide range of fortifications, including city defences, and were prepared to live in both rural and urban settlements. The purpose of all their major strongpoints, regardless of whether they were walled towns, citadels or mountain castles, was to provide shelter for local troops and civilians whilst at the same time preventing external attackers from making permanent territorial conquests. It is true, however, that mountain castles like Vagha and Servantikar proved most successful in achieving this, and were therefore largely responsible for the ability of Armenian kings to retain their independence until as late as 1375. During the first half of the thirteenth

FORTIFICATIONS AND EXTERNAL SECURITY

l6j

century Leon II and Hethoum I may well have constructed new fortresses which were specifically intended to watch and defend the key mountain passes. The impressive remains of some of these strongholds should not, however, lead us to believe that the Armenians only ever built mountain castles, and that these were all specifically designed to fit into an intervisible network. Instead Cilician Armenia seems to have had more in common with the Holy Land, where a variety of urban defences and individual castles continued to protect Prankish lands until Muslim armies finally overcame them through sheer weight of numbers.

lò

Fortifications and Internal Security Whilst their fortifications helped the Armenians to keep their external enemies at bay, such structures tended to undermine central authority. It has already been shown that several key strongpoints formed part of the royal domain, including Adana, Misis, Tarsus, Sis and Ayas. Other major castles not described as baronial properties in Leon II's coronation list of 1198, most notably Anavarza, Vagha, Yuan and Tumlu, presumably also belonged to the king.i The vast majority of other Armenian fortresses were held by individual nobles, however, who had originally been able to take possession of these sites because of the ruggedness of the local terrain and the weakness of neighbouring powers. The crushing defeat inflicted by the Seljuk Turks on the Greeks at the battle of Manzikert (1071), for example, had caused Byzantine rule over most of Anatolia to collapse and meant that the Armenians were effectively left to their own devices for most of the late eleventh century. During the next hundred years occasional expeditions from Constantinople restored order, but for much of the time Byzantine emperors could not prevent Armenian warlords from seizing local castles and pursuing their own policies. Some of these men, such as the Hethoumids of Lampron, recognised Byzantine overlordship, whilst many others did not and hoped to shake off Greek rule completely. This led to periods of bitter in-fighting during the twelfth century which were often intensified by the religious and cultural differences between the Armenians and those Greeks who actually lived in Ciucia itself. Meanwhile the years leading up to the Fourth Crusade (1204) and its aftermath witnessed an even greater decline in Byzantine authority, whilst Prankish influence over the area eroded first through the loss of Edessa (1144) and later because of the threat posed by Saladin. While these factors enabled Leon II to achieve independence, they also meant that by 1200 many local castles had become the permanent hereditary possessions of powerful Armenian families.2 Perhaps the best example of this type of stronghold was Lampron. This fortress was situated high up in the Taurus mountains near the Cilician Gates, its isolated location and powerful defences making it almost invulnerable to enemy attack.3 Following the battle of Manzikert, the Armenian Hethoumids had taken possession of the castle with Byzantine support; an

FORTIFICATIONS

AND I N T E R N A L SECURITY

169

alliance which made them the arch-enemies of Leon It's own anti-Greek Roupenid family to the east and south.4 During the twelfth century Roupenid efforts to overcome the Hethoumids were repeatedly thwarted by the sheer strength of Lampron, which was unsuccessfully besieged in the early 11705 and again about ten years later.5 It was even claimed that 'all the lords of this castle rebelled because they knew it was impregnable'.6 Bearing in mind that Leon II's coronation list mentions another forty-four local lords, all of whom held at least one fortress,7 it is easy to see why Leon II and his successors feared the possibility of baronial rebellion, even if all castles were not as strong as Lampron or their owners as fiercely independent as the Hethoumids. Leon II pursued a number of policies intended to strengthen his own authority over the nobility. The royal domain itself, built up through conquest by Leon's twelfth-century Roupenid predecessors, represented a good starting point, for sites like Sis and Ayas gave the king far greater economic (and therefore military) power than any other individual lord. However, the history of Lampron showed Leon that it was not enough to try to intimidate difficult barons; their castles had to be confiscated. In 1201 he came up with a plan to bring down the Hethoumids which relied on subtlety rather than brute force: With this aim in mind, he sent to Hethoum, son of Oschin, a message to trick him: 'I want to establish a bond of friendship with you', he told him, 'and give Phillipa, daughter of my brother Roupen, in marriage to your oldest son Oschin', Hethoum accepted this suggestion. The celebrations for the marriage were to take place at Tarsus. When the Hethoumids had arrived in this city with all their relatives and children, King Leon seized them and occupied Lampron without any bloodshed. After having imprisoned Hethoum for a while, he released him, gave him a number of villages, and treated him with good will from then on. As for Hethoum, he showed himself to be a loyal vassal.8

In some ways it is surprising that Hethoum fell for this trick, for as recently as 1185 the Hethoumids, aided by their ally Bohemond III, had themselves ensnared Leon's predecessor, Roupen III (1175-87), by inviting him to 'orgies with women of ill repute' in Antioch.9 The ruse clearly worked, however, and while the Hethoumids later regained Lampron, Leon's acquisition of this one castle considerably increased his own power. Leon's treatment of Hethoum after Lampron had been taken is also significant, for it shows that the king was trying to score a symbolic as well as a strategic victory. This was more than a straightforward dispute between two rival Armenian clans, as might have been the case a century earlier; it was an opportunity for Leon to emphasise the superiority which his new royal status had given him. Hethoum was made to understand that he could

I/O

U N K N O W N CRUSADER CASTLES

no longer act independently, but only received his lands in return for service and loyalty toward the king. Similarly, in 1198 the forty-five barons present at Leon II's coronation were encouraged to attend because the new king 'attracted them with his promises, and made them his men with his grants'.10 Clearly Leon wanted these lords to understand that they did not just hold their castles and estates by conquest or birthright but because he allowed them to do so. When Vasil of Vaner, one of the barons who had witnessed Leon's coronation, died without an heir in 1214, the king quickly placed his lands under royal control so that Leon rather than anyone else could decide the fate of these important estates.11 Such actions indicate that Armenian rulers tried to prevent baronial uprisings by confiscating or distributing castles and fiefs in the manner of west European kings. The 'westernising' policies pursued by Leon II extended to other areas of government. It was Leon, for example, who granted many privileges to Italian merchants, thereby attracting more trade to Ayas, increasing royal income and improving links between Cuidan Armenian and Europe.12 More significantly in terms of castles, Leon also transferred numerous estates and strongholds to the Hospitallers and the Teutonic Knights.13 These arrangements were primarily intended to bring military assistance against the Seljuk Turks, but when it came to the western frontier castle of Silifke they may also have provided Leon with a means of removing an important stronghold from another independently-minded baronial family. Silifke was granted to the Hospitallers in 1210 along with considerable surrounding estates which included the strongholds of Goumardias (Camardias) and Norpert (Castellum Novum).14 Three years earlier, these properties had belonged to a local nobleman called Henry and his three sons Constantine, Joscelin and Baldwin. They had conspired against Leon who had arrested them and confiscated their lands. The exact cause of this rift is not known, although the fact that Henry was married to the sister of John, the Armenian Catholicos, must have been a contributing factor as Leon and John were in the midst of a violent personal feud at this time.15 By handing Silifke over to the Hospitallers Leon was placing a key frontier castle in the care of more trustworthy occupants who were likelier to remain loyal to their patron. Two important considerations indicate that this argument should be used with caution. First, there is no evidence to suggest that other lands and castles transferred to the Military Orders had originally belonged to rebellious nobles. The estates of Vasil of Vaner, for example, were eventually sold to the Hospitallers, but these properties were acquired by Leon II because Vasil had died without an heir rather than because he had misbehaved in some way.16 Similarly, the castles of Amouda and Harunia,

i8. Margat: the east double walls of the inner castle, with round keep at the far end. (Jonathan Phillips)

19- St Hilarión: general view of the castle. (Courtauld Institute)

20. St Hilarión: the Byzantine defences of the lower bailey. (Denys Pringlé)

2l. Buffavento: the lower structures of the castle. (C. Enlart)

22. Kantara: view of the castle from the south east. (C. Enlart)

23. Kantara: aerial view. (C. Enlart)

24. Baghras: view of the castle from the east. (Courtaula Institute)

25. Silifke: the southern defences. (Courtauld Institute)

26. Servantikar: the east front of the castle. (R. W. Edwards)

27. Servantikar: the main entrance. (R. W. Edwards)

28. Karytaina: view of the castle from the east. (Courtaula Institute)

29. Arcadia from the south. The rounded corner tower to the right is thought to be a Prankish addition to an otherwise Byzantine fortress. (Courtaula Institute}

30. Androusa: the keep. Note the poor quality of the masonry compared to fortifications in the Holy Land such as Tortosa. ( Courtaula Institute)

31. Thessaloniki: the city walls. Similar Byzantine defences were occupied and repaired by the Latins at Constantinople and elsewhere. (Courtaulá Institute]

32. Yuan: general view of the fortress. (R. W. Edwards)

33. Corycus: view of the sea castle from the land castle. (R. W. Edwards)

F O R T I F I C A T I O N S AND I N T E R N A L S E C U R I T Y

1/1

which were transferred to the Teutonic Knights in 1212 and in 1236 respectively,17 both belonged to Armenian barons at the time of Leon's coronation in 119818 but need not have been seized by the king because these lords had plotted against him. Secondly, it is important to remember that the Military Orders were not necessarily any more loyal to a ruler than individual nobles. During this period military pressure forced many lords in the Holy Land to grant castles and privileges to the Templars, Hospitallers and Teutonic Knights. Although this policy did bring much needed assistance against the Muslims, it also enabled these Orders to build up vast, semi-independent enclaves which were sometimes beyond the authority of local rulers. The granting of strongholds to the Military Orders can normally be regarded as a sign of weakness rather than strength.19 Efforts to remove castles from potential trouble makers failed anyway, for although Leon II (1187-1219) and Hethoum I (1226-69) generally had successful reigns, dangerous underlying tensions still existed within Armenian society. Rather than eliminating the causes of rebellion, royal policies sometimes encouraged discontent to grow just beneath the surface. The closer links with the West fostered by Leon II, for example, seem to have caused resentment amongst some elements of the nobility, particularly because of fears that this would lead to interference with the Armenian Church. These fears grew as time went by, for the deteriorating military situation after the 12ÓOS led to increasingly urgent appeals for assistance from Europe, which the pope would only countenance in return for promises that the Armenian clergy would be brought into line with Rome. These religious tensions appear to have been exasperated by more local disputes between the Armenian and Greek Orthodox Churches. Finally, even allowing for the colourful prose of the chroniclers, it is clear that the aristocracy often indulged in murderous vendettas which could result in members of the same family imprisoning, mutilating or even killing each other with alarming frequency. The aggressive, independent tendencies which Leon II had tried to stamp out had clearly not disappeared, for the subsequent history of the kingdom was riddled with plots and rebellions motivated by a bewildering combination of religious conflict, political ambition and personal animosity.20 These inherent weaknesses came to the surface almost as soon as Leon II died in 1219. Leon nominated his daughter Isabelle (Zabel) as his successor, but she faced a rival in the form of Leon's great nephew Raymond Roupen. In 1221 he used the port of Corycus, a baronial possession whose lord must therefore have been involved in the plot, as a base for an attempt to seize the throne. This crisis came to an end not long afterwards when Constantine, leader of the Hethoumid clan, besieged and captured Raymond Roupen inside Tarsus.21 Thereafter Constantine and the other barons arranged a

1/2

U N K N O W N CRUSADER CASTLES

marriage between Isabelle and Philip, son of Bohemond IV of Antioch. This match was intended to bring a formal end to the bitter power struggle for Antioch which had raged between Bohemond and the Armenians for most of the previous twenty years.22 The reconciliation ended in bloodshed when Philip was murdered after a reign which had only lasted two years (1223-25), apparently because Constantine and the other barons began to resent Philip's pro-Frankish policies and his refusal to join the Armenian Church. Isabelle responded to her husband's death by taking refuge inside the Hospitaller fortress of Silifke, until pressure from Constantine forced the Order to hand over their guest and relinquish their castle. These events suggest that anti-western hatred had built up amongst elements of the Armenian nobility, although it should be noted that Constantine was just as motivated by personal advancement, for he subsequently forced Isabelle to marry his own son Hethoum, who was crowned king in 1226.23 Throughout the crisis it also appears that some western elements, including the Teutonic Knights, actually supported Constantine rather than Philip. The Order was prepared to condone the murder of Philip despite the obvious distress it caused Isabelle, the daughter of the man who had granted the Knights their first properties in Cilicia, which confirms that the Military Orders were not necessarily any more reliable or selfless in their motives than individual barons.24 An inscription which commemorates alterations carried out on Silifke by the king in 1236 proves that Hethoum I kept this castle after it was given up by the Hospitallers, and suggests that he was reluctant to grant the fortress to another Order or Armenian baron.25 The only way for a ruler to prevent castle garrisons from defying his will was to keep them under his personal supervision. In some ways the turbulent events of the 12205 also set the tone for the future, for internal conflicts tended to coincide with the end of an old reign or the beginning of a new one. During these periods of weakness rebellious lords used their castles as centres of resistance in the same way that Raymond Roupen had based himself at Corycus in 1221. In 1271, for example, the new king Leon III (1270-89) had to deal with an uprising by a faction of Greek Orthodox nobles which was only put down once the rebels' castles had been captured and their ringleader had been killed in 'the fortress of the city of Anavarza'.26 Even worse problems confronted Leon Ill's successor Hethoum II (1289-1307), who spent much of his reign trying to control his three brothers Sempad, Thoros and Constantine. During the late 1290S Sempad even seized power and had Thoros strangled and Hethoum partially blinded, but Hethoum was rescued by Constantine, whose possession of the castle of Gabán enabled him to stay beyond the grasp of Sempad. Hethoum subsequently regained the throne, only to be murdered

F O R T I F I C A T I O N S AND I N T E R N A L S E C U R I T Y

1/3

by the Mongols, alongside his nephew and nominated successor King Leon IV, in 1307.27 In the fourteenth century conflicts of this kind continued to break out with depressing regularity. In 1329 Oschin, lord of Corycus and father-in-law of King Leon V (1320-41), was murdered, perhaps because the king felt he had gained too much power, or because he believed that this nobleman had been responsible for the deaths of his predecessor King Oschin (1307-20) and the latter's sister.^ By this point religious disputes were also escalating, for in 1309-10 King Oschin had been forced to break up a large demonstration at Sis against a plan to accept papal authority and introduce Catholic practices within the church.29 This plan would have attracted more western support against the Muslims, but the hostility it aroused indicates that it was risky for any ruler to foster too close links with the Latins. Leon V married Constance of Aragon, widow of Henry II of Cyprus (1285-1324), in an effort to gain help from this quarter, but he was apparently murdered as a consequence. A similar fate befell Henry II's nephew Guy of Lusignan, who succeeded Leon V as Constantine IV (1342-44). Even though Guy's mother was a member of the royal Armenian family, he was quickly assassinated and had 'an ephemeral reign, because the troops rebelled'.30 These actions were short-sighted, for when Constantine VI (1367-73) eventually succeeded in establishing an alliance with Peter I of Cyprus (1359-69), the latter was able to halt an attack on Corycus by the nearby Turkish emirate of Karamania.31 Instead, the chronicler Jean Dardel claimed that the middle decades of the fourteenth century witnessed such a constant succession of internal feuds that some barons were still plotting to overthrow the last king, Leon VI (1374-5), even as the Mamluks were conducting the final siege of Sis.32 Although Jean Dardel was doubtless prone to exaggeration, the Armenian rulers generally failed to overcome the religious and political rifts which divided their kingdom. Leon II's attempts to try to reduce the power of noble families like the Hethoumids had only achieved relative success for as long as Leon himself was alive. The anarchy which followed his death and the ease with which Constantine was able to place first Philip and then Hethoum on the throne suggest that Leon's authority had only been based on his own dynamic personality. Indeed, Constantine's ruthless treatment of Isabelle and Philip shows that Leon had not instilled the barons with a sense of respect for the institution of monarchy, and that he had failed to prevent the Hethoumids from continuing to act as though they were above the law. While these problems were kept in check during the long and relatively stable reign of Hethoum I (1226-69), they reappeared once regular Muslim attacks began to sap the economic and military power of Hethoum's

1/4

UNKNOWN CRUSADER CASTLES

successors. By the fourteenth century it is clear that Armenian kings were often incapable of keeping their own families in check, let alone bringing recalcitrant nobles to book. Wider divisions within the aristocracy were also compounded by constant bickering over the nature of political and religious links with the West. These circumstances make it extremely unlikely that Leon II or any of his successors could have maintained a network of intervisible castles which relied on close cooperation between the king and all his barons. In many ways the Armenians faced the same dilemma as their Prankish neighbours to the south. In order to withstand the armies of the Seljuks, the Mamluks and even the Mongols, the Armenians needed to build and garrison powerful fortifications. Castles like Lampron, however, were so well defended that they enabled their owners to disregard royal authority, thereby weakening the ability of kings to maintain order and resist external invaders. By comparison, Frederick II's efforts to assert his authority over the kingdom of Jerusalem during the crusade of 1228-29 failed in part because opponents such as the Templars refused to deliver their powerful castles to him. In contrast, on Cyprus the Lusignan kings imposed a strict royal monopoly on castle building in order to avoid this problem. But in the Holy Land the scale of their strongpoints sometimes enabled the Templars, Hospitallers, Teutonic Knights and the Italian city states to pursue totally independent policies and to fight amongst themselves.33 A similar process clearly took place in Cilician Armenia, although here the vast majority of castles still belonged to individual nobles rather than the Military Orders. The overall effect was the same, for internal fighting undermined central authority and meant that precious resources were squandered locally rather than being used against the Muslims. This process only proved fatal later on, but even as early as the 12205 the clashes which followed Leon II's death were said to have made the Seljuk Turks masters of 'many of the fortresses of Cilicia'.34 We are left with the irony that the very same castles which protected the Armenians for so long also contributed to the disintegration of their kingdom and the final destruction of their capital in 1375.

17

The Military Orders An important development during the reign of Leon II (1187-1219) was the increasing involvement of the Military Orders in Cilician Armenia. All three Orders held castles inside or on the fringes of the kingdom. As we shall see, the Templars spent most of their time fighting Leon, whereas the Hospitallers and the Teutonic Knights became his close allies. In the case of the last Order, this was the result of Leon's own late twelfth-century efforts to turn his domain into a kingdom under the suzerainty of the German empire. The Emperor Henry VI (1190-97), who was about to go on crusade, was favourable to the idea of crowning both Leon II and Aimery of Lusignan, the new Prankish ruler of Cyprus. This project would clearly enhance the status of all three men. Even though Henry himself died before reaching the East, Leon was crowned in 1198 in the presence of the papal legate Conrad of Mainz and the imperial chancellor Conrad of Hildesheim.1 At this stage the Teutonic Knights were still in their infancy, but as early as 1212 Leon II confirmed their possession of the castle of Amouda, which lay on a small outcrop at the centre of the Cilician Plain.2 This grant reflected the continuing links between Leon and the Emperor Otto IV (1198-1218), and was a direct response to visits in that year by Hermann of Salza, the master of the Teutonic Knights, and Willbrand of Oldenburg, the leader of an imperial delegation.3 The Teutonic Knights subsequently remodelled Amouda by building a square three-storey tower with very thick walls and a solitary doorway at the highest point of the castle. The general design of this keep was very similar to that of other towers built by the Order at Montfort and Judin, two of their castles in Palestine, and can even be compared with numerous medieval strongholds in Germany itself.4 The construction of such a keep at Amouda is significant in two ways: first, it indicates that the Teutonic Knights intended this castle to be their local headquarters and they therefore spent a lot of time and money on it; secondly, the use of a typically German (or at any rate non-Armenian) design strengthens the argument that local strongholds did not belong to a vast network of intervisible castles controlled by the Armenian monarchy. To the east of Amouda and situated relatively close to the Amanus Gates, the Teutonic Knights also possessed Harunia (Haruniye/Haroun), which

l/O

UNKNOWN C R U S A D E R CASTLES

had belonged to an individual baron in 1198 but was given to the Order by Hethoum I and Queen Isabella in 1236.5 Once again, the timing of this grant probably had a political dimension to it, for Hethoum had been supported by the Teutonic Knights during the succession dispute of the 12203 and wished to remain on friendly terms with the imperial forces of Frederick II active in the Holy Land at that time.6 Architecturally, Harunia was smaller than Amouda and amounted to little more than an elongated keep with a cramped central courtyard, two floors of shooting galleries and a large rounded tower in the north-west corner. The only possible traces of Latin occupation are to be found in this tower, which may have been repaired by the Teutonic Knights and perhaps even used by them as a chapel.7 In addition, the Order probably held even smaller fortified houses or towers in the Cilician countryside, for both Amouda and Harunia had been granted to the brothers along with considerable surrounding estates. In 1212, for example, Willbrand of Oldenburg noted that the Teutonic Knights were in possession of Cumbethefort, a village situated somewhere between Misis and Tarsus which had been formally handed over by Leon II at the same time as Amouda.8 According to a document dating from 1271, the Order also possessed a toll station in the vicinity of Servantikar known as the Black Tower.9 This minor stronghold has recently been identified as Hasanbeyli, a medieval tower located near the eastern approaches to the Amanus Gates. This identification implies that the Teutonic Knights held properties quite far beyond the Cilician Plain and cut off from estates around Harunia itself by the Amanus mountains and by lands belonging to the lords of Servantikar. It may therefore be that the Black Tower lay to the west of Servantikar nearer Harunia in a spot which has not yet been identified.10 This brief description of properties held by the Teutonic Knights gives us some idea of what functions their Armenian castles were expected to fulfil. It is clear that on one level they held no great military significance in their own right, being simply a means of maintaining good relations with the German emperors and thereby legitimising first the coronation of Leon II (1198) and then the reign of Hethoum I, who only acquired the throne by forcing Leon's daughter to marry him (1226).n This applies most of all to Amouda, which lay too far away from any frontiers for its garrison to take rapid action against hostile invaders. It may therefore have been granted to the Teutonic Knights in return for more general political or military help rather than any local strategic support. On the other hand, Harunia's location relatively close to the Amanus Gates indicates that this castle did have an immediate military role to play, something which Hethoum I no doubt had in mind when he referred to the Knights' constant struggle with 'the enemies of Christ' in his donation charter.12 Seljuk troops

THE MILITARY ORDERS

177

Amouda. Note the keep (H) added by the Teutonic Knights.

had already used the Amanus Gates to attack the region in the past,13 and during the 12205 and 12305 there were more raids which put the Armenians under severe pressure.14 Later attacks through the Amanus Gates by the Mamluks were to show just how common it was for Muslim armies to use this route when approaching from the south east. Presumably Hethoum I hoped that the Teutonic Knights would help him to stop hostile incursions of this kind either by drawing the Muslims into a lengthy siege of Harunia itself, or by sending troops to defend the Amanus Gates in the same way that he himself did against Baybars in 1266.15 If the intention had been to secure the eastern frontier of Cilicia against future Muslim attacks, Hethoum and his successors must have been sorely disappointed. As we have already seen, in 1266,1275 and 1298 large Mamluk armies could not be prevented from passing through the Amanus Gates and then ravaging the Cuidan Plain. Indeed, in 1266 Amouda itself was taken and sacked by the Mamluks. Bearing in mind that the fortifications of Harunia were no stronger, these castles must have been too small to halt attackers by forcing them to stop and undertake lengthy sieges.16 To some extent this problem could not have been foreseen by Leon II and Hethoum I, for the size and professionalism of Muslim armies grew steadily in the

r/8

UNKNOWN CRUSADER CASTLES

course of the thirteenth century. The steady increase in Muslim power also had wider implications, for the Teutonic Knights had castles to defend in the Holy Land, which lay closer to Egypt and presumably took precedence in terms of troops and resources. They may therefore have been prevented from constructing more fortifications in Cilicia or sending more soldiers there. The loss of Acre in 1291 ultimately persuaded them to concentrate on their Baltic activities.17 These problems should not obscure the fact that Amouda and Harunia were important in other ways. First, it is clear that these strongholds gave the Teutonic Knights a certain amount of influence over Armenian politics, for Leon II and his successors did not wish to offend an Order which gave them a link with the German emperors and was so closely associated with their elevation to royal status. The Teutonic Knights received lands and enjoyed royal protection under both Leon II and Hethoum I, even though the latter ruler gained the throne against the wishes of Leon's designated heiress Isabelle. Hethoum appears to have been far more concerned about remaining on friendly terms with the Teutonic Knights than the Hospitallers, who were forced to deliver the castle of Silifke to him because they supported Isabelle during the upheavals which followed the death of Leon II.18 Amouda, Harunia and their surrounding estates also provided the Order with a major source of income. Harunia, for example, was granted along with its own small town, numerous surrounding villages and abbeys, plus many estates, orchards, woods and waterways.19 Amouda also received a steady supply of fish from the adjacent Ceyhan river, whilst the claim that over 2OOO people were sheltering there during the Mamluk raid of 1266 implies that this was a well-populated area providing plenty of scope for agriculture.20 Indeed, when the Teutonic Knights were given Amouda in 1212, they had only been in existence as a Military Order since the 11905, and during their early years it must consequently have represented one of their most lucrative possessions. The value of their Armenian properties was also increased by the fact that the Teutonic Knights were exempt from paying taxes when buying or selling various goods, even though they themselves could impose tolls on others travelling through their lands.21 Amouda may have been used for this purpose because of its close proximity to the Ceyhan river and its location between Sis and the Amanus Gates.22 At the toll station known as the Black Tower the Knights who garrisoned Harunia made a good income from people who had emerged from the Amanus Gates and were travelling further north.23 As time went by the Armenian fortifications belonging to the Teutonic Knights doubtless had more significance as centres of local farming, trade and defence, and as a way of maintaining good relations with the German emperors, than they did as a

THE M I L I T A R Y

ORDERS

1J9

means of stopping a major invasion force or significantly enhancing the national security of the kingdom. In some ways the rapid expansion of the Teutonic Knights during the early thirteenth century was very similar to that of the Hospitallers, whose involvement in the region increased dramatically when Leon II granted them the fortress of Silifke (Seleucia) in i2io.24 This stronghold, which lay on the coastal route leading westwards out of Cilicia, consisted of an inner curtain wall flanked by several horseshoe towers, and a wide outer moat whose inner revetment was so large that it effectively formed a second rampart. Inside the castle the various service buildings and living quarters were arranged around an open courtyard rather than being incorporated into a free-standing central keep. This layout is architecturally significant, for although Silifke strongly resembles Armenian fortresses such as Yuan and Servantikar, it also looks like the famous Hospitaller castles of Crac des Chevaliers and Margat in Syria. Many of the features which can be seen at Silifke, such as horseshoe towers and the absence of an isolated keep, were copied at these latter strongholds in the early years of the thirteenth century, proving that the Armenians had some impact on Prankish military architecture.25 The fact that the Hospitallers virtually rebuilt Silifke from scratch also suggests that the site held great strategic significance. Apart from Silifke the Order had been granted numerous surrounding estates and properties, including the castles of Goumardias (Camardias) and Norpert (Castellum Novum). Although the exact location of these smaller strongholds remains uncertain, it is clear that they formed part of a powerful lordship occupying the south-western corner of the Cilician Plain.26 During the early thirteenth century this area was rapidly turning into the new frontier between Leon II and the Turks, as the Seljuks managed to take control over former Armenian strongholds to the west and north of Silifke.27 By granting the Hospitallers so many local lands, Leon hoped that they would help him to halt any further Muslim expansion, a policy which came to fruition in 1216 when the brothers fought alongside Armenian troops in the successful defence of Silifke against a Seljuk attack.28 This encounter implies that within a mere six years of acquiring the castle the Hospitallers had already completed their ambitious reconstruction of it. When one bears in mind that they were simultaneously carrying out major repairs at both Margat and Crac des Chevaliers, the Order must have had considerable funds at its disposal during this period.29 In order to maintain good relations with Leon II's successor Hethoum I, the Hospitallers also agreed to make an annual contribution of 400 horsemen to the royal army, indicating that their financial and military resources far outstripped anything that could have been offered by an individual Armenian baron.30

l8o

UNKNOWN CRUSADER

CASTLES

The history of Silifke during this period can be used to illustrate the various ways in which Leon hoped that closer links with the Military Orders would benefit him. It has already been shown that until 1207 this castle, along with Goumardias and Norpert, belonged to the Armenian baron Henry and his three sons Constantine, Joscelin and Baldwin, but that they were then deprived of their lands for plotting against the king.31 Leon may have felt that granting these castles to the Hospitallers provided him with an opportunity to present them to someone who did not belong to a strong local clan, thereby reducing the threat of baronial unrest. However, Leon was probably far more concerned about the Seljuks, and as early as 1207 he presumably realised that Henry and his sons were not only disloyal but also lacked the resources to deal with the growing Muslim threat. The fact that the Hospitallers effectively rebuilt Silifke implies that under baronial control this castle had been too weak to withstand a siege and was in desperate need of improvement. Henry's rebellion may indeed not have been entirely unwelcome, for it gave Leon an excuse to transfer Silifke to a body which could rectify the situation and also contribute urgently needed troops to the army. Other transactions between Leon II and the Hospitallers during this period confirm that the long regarded them as an important source of military and financial support. In 1214 the Order provided Leon with 20,000 Saracen bezants in return for the right to collect revenues from several other estates, including the port of Canamella, which lay along the coast between Alexandretta and Misis. These estates were to be handed back to Leon, provided he could repay his loan within two years; otherwise the Order could keep them.32 At this time the Hospitallers also paid a sizeable amount of cash for the lordship of Vaner, whose original Armenian lord, Vasil, had died without an heir. It has been suggested that these estates were centred around the fortress of Gökvelioglu, an impressive ruin situated on an outcrop south of Misis, but this cannot be verified as the relevant documents do not mention fortifications.33 The scale of Leon IFs dealings with the Hospitallers suggests that he was experiencing financial difficulties. Partly caused by his struggle with the Seljuks, they were compounded by Leon's long and costly campaign to gain control over Antioch. Until 1201 this city had normally been controlled by Bohemond III, but his death in that year caused a succession dispute to break out between his son Bohemond IV and his grandson RaymondRoupen, who was the product of a marriage between Bohemond Ill's deceased older son Raymond and Leon II's niece Alice. Inevitably Leon took the side of Raymond Roupen, who as a small boy became little more than a pawn in his great uncle's plans to extend Armenian power into Syria. Bohemond IV, on the other hand, was supported by the Templars, which

THE M I L I T A R Y O R D E R S

l8l

doubtless accounts for Leon II's growing involvement with the Hospitallers.34 Originally they had supported Bohemond IV, but in 1204 they changed sides.35 The need to preserve their support must have influenced Leon's decision to grant them Silifke six years later. In 1207 Raymond Roupen also promised them Jábala, a coastal town within the principality of Antioch which was actually held by the Muslims at that time.36 As a result the Hospitallers eventually occupied half of Jábala, although there is no evidence that they ever recaptured Bikisrail, another ex-crusader stronghold to the north east of Margat granted to them by Raymond Roupen in 1210.37 These grants indicate that Raymond Roupen saw himself as the legitimate ruler of the principality, but he was unable to take control over Antioch itself until 1216, when the Hospitallers helped him to overcome local opposition by garrisoning the ramparts and the citadel. Three years later anti-Armenian factions within the city joined forces with Bohemond IV and drove Raymond Roupen and his Hospitaller allies back to Ciucia, thereby ending any hopes of incorporating Antioch into Leon II's kingdom.38 Leon II's motives for fostering friendly relations with the Hospitallers were therefore similar to those which drove him and his successor to establish links with the Teutonic Knights. Both these Orders proved to be useful political allies, for one helped bring Leon a royal crown from the German empire, whilst the other gave him a strong local ally in his struggles with Bohemond IV. Like their German counterparts at Harunia, the Hospitallers were also granted Silifke and its adjoining estates in order to defend a vulnerable frontier of the Cilician Plain against the Seljuks. The impressive archaeological remains of Silifke, its successful defence in 1216, and the Hospitallers' offer of 400 horsemen to the royal army all suggest that they initially fulfilled this objective with more success than the Teutonic Knights. This was partly because they had greater economic and military resources to draw on, and partly because they had much wider experience of managing such territories from the Holy Land, where Margat and Crac des Chevaliers both formed centres of vast frontier lordships which the Hospitallers used to wage war against neighbouring Muslims and to protect Prankish lands nearer the coast.39 It is clear that after 1210 the new castle of Silifke began to take on a similar role, for it became the residence of the local preceptor.40 Like the estates around Amouda and Harunia, the lands which were attached to Silifke must have generated very considerable income, thereby offsetting the cost of rebuilding the fortress and providing troops for the royal army. Although there were clear benefits both for the Armenians and for the Hospitallers, Leon's death in 1219 quickly caused all these arrangements to unravel. In that year Raymond Roupen's control of Antioch ended, and his subsequent attempt to seize the Armenian throne, carried out 'with the

l82

UNKNOWN CRUSADER

CASTLES

help of Guérin, Master of the Hospital of St John', was thwarted by Hethoum I's father Constantine.41 In 1226 Constantine forced Leon's daughter Isabelle, who had been sheltering inside Siiifke, to marry Hethoum. The Hospitallers' close links with both Raymond Roupen and Isabelle now proved their undoing, for Constantine forced them to relinquish Silifke immediately afterwards.42 The Hospitallers do not seem to have been entirely unhappy about this because 'the defence of this place was a heavy burden', indicating that the Order had underestimated the Seljuk threat and consequently no longer enjoyed the same level of financial or military strength as it had done under Leon II.43 A combination of political instability and financial worry therefore ended the development of what could have been a new Crac des Chevaliers after a mere sixteen years. Whilst Leon forged close ties with the Hospitallers and the Teutonic Knights, he spent almost his entire reign in conflict with the Templars. During the twelfth century, and possibly even as early as 1131, this latter Order had been granted an extensive march which protected the northern frontier of Antioch against external aggressors. Perhaps the most significant castle within this march was Baghras (Gaston), which dominated the important mountain route between Ciucia and Antioch known as the Belén Pass (or Syrian Gates). To the north, another valley which bisected the Amanus mountains was also guarded by the Templar strongholds of Darbsak (Trapesac) in the west and Roche Roussel (Hadjar Shoglan/Chilvan Kale) in the east. A fourth castle in this region which the Latin chroniclers called Roche Guillaume has never been located, although it is clear that to the north west of Baghras the Templars also held the small harbour of Arsuz, known to them as Port Bonnel.44 In 1188, following his decisive victory at Hattin, Saladin was able to tear this impressive territorial block apart by capturing Baghras and Darbsak, but although the latter fortress remained under Muslim control, the former was subsequently abandoned. In 1191 this enabled Leon II to occupy Baghras, along with Port Bonnel for a time. As a result the Templars fell out with him and became the natural allies of Bohemond IV during the subsequent Antiochene succession dispute.45 The archaeological remains of Baghras are almost as complex as its chequered history, for the castle is located on a steep outcrop and incorporates so many different building styles that scholars have argued over its exact origins.46 One archaeologist has concluded that the lower bailey which guarded the eastern approaches to the fortress, as well as the huge revetment supporting the shooting galleries of the south-west corner, were both constructed by the Armenians shortly after they occupied Baghras in iipi.47 This view has been criticised on the grounds that the Armenians rarely, if ever, constructed revetments, and that there are close parallels

THE M I L I T A R Y O R D E R S

183

between the masonry of the lower bailey and numerous other Templar sites, including Darbsak. Similarly, the various defensive and residential structures of the upper bailey are arranged in a compact, typically Armenian fashion, but a more detailed study of individual architectural features within these buildings reveal certain similarities with the Templar citadel of Tortosa.48 It would be wrong, therefore, to argue that Baghras is largely an Armenian construction simply because it is a mountain castle situated in a remote area. Indeed, it has been claimed that 'the Armenian presence here is no more than a flirtation' and that the importance of the site to Leon II has been exaggerated.49 This conclusion does not, however, concur with the historical evidence. The Arab historian Ibn al-Athir, for example, wrote that Leon 'marched on this place which was near his territory. He rebuilt Baghras carefully and stationed a garrison there to carry out raids on the surroundings'.50 According to Willbrand of Oldenburg, Baghras was 'a very powerful castle, with three strong walls and towers around it, situated in the last mountains of Armenia. It carefully guards the entrances to that land, whose ruler, the king of Armenia, holds it'.51 These descriptions suggest that Leon's presence at Baghras from 1191 onwards was more significant than a 'flirtation', and that contemporaries considered the castle to be strategically very important. This is confirmed by Leon's own efforts to hold on to Baghras for as long as possible rather than to return it to its original Templar owners. From the 11905 onwards Leon was often prepared to discuss the issue of Baghras but frequently used stalling tactics, such as asking the pope to arbitrate in the matter, rather than dealing with the Templars directly. In 1199, for example, Leon tried to persuade Innocent III that Baghras was his by right of conquest, but the pope quickly rejected this.52 During the ensuing years Innocent's representatives were kept almost constantly busy trying to solve both this problem and the parallel Antiochene succession dispute between Bohemond IV and Raymond-Roupen.53 In 1203 the situation worsened when Leon tried to capture Antioch itself, but he was pushed back by a combined force of Templars defending their headquarters and supporters of Bohemond IV.54 Subsequent fighting provoked Leon into taking temporary possession of Roche Roussel and Roche Guillaume, the remaining Templar castles in the Amanus march, and then justifying his actions by portraying the Order's efforts to retrieve Baghras as unprovoked aggression against the Armenians.55 Two years later Leon also made a failed attempt to recapture Darbsak from Aleppine forces, whose friendship with Bohemond IV at this time provides us with an interesting example of a Franco-Muslim alliance against a fellow Christian.56 Needless to say, these activities did not please the Templars, and in 1211 another period of failed negotiations and sporadic warfare culminated in a major Templar attack

184

UNKNOWN CRUSADER CASTLES

Baghras. Ground-plans of the castle, scale 1:10,000. A. Lower fortress and lower floor of upper fortress. B. Upper floor of upper fortress (ground-floor premises stippled) with the earlier bunding phase (Byzantine?) in black and the portions dating from the twelfth to thirteenth centuires cross-hatched, i. Lower fortress. 2. Forecourt and gatehouse. 3. Lower gallery. 4. Palace premises. 5. Great Tower. 6. Magazine chambers.

on the area, including perhaps Baghras itself.57 By this point the pressure applied by the papacy, the Templars and Bohemond IV, combined with the Seljuk threat in the west and the financial shortages implied by Leon's transactions with the Hospitallers, all conspired to bring about a change in Armenian policy. In 1212 Leon agreed to restore all Templar properties, and in 1213 the excommunication which had previously been imposed upon him was lifted by Innocent III. Even then Leon still managed to stall the return of Baghras until 1216, twenty-eight years after Saladin's invasion.58 Leon's determination to hold on to Baghras for such a long time, and his ultimate decision to give it back to the Templars, reveal, much about

THE M I L I T A R Y O R D E R S

185

the strategic importance of the castle. This stronghold's original twelfthcentury purpose had been to protect Prankish Syria against incursions from the north, but Willbrand of Oldenburg noted that 'it overlooks Antioch directly1.59 It could therefore be used to spot attacks aimed at Ciucia by Bohemond IV. Alternatively, it could act as a starting point for Armenian incursions toward Antioch, and it seems likely that Leon utilised Baghras in this way during his failed attack on the city in 1203. Ibn al-Athir's claim that the Armenian defenders of the fortress carried out frequent local raids also suggests that Leon regarded Baghras as an important military asset in his long campaign to gain control over Antioch.60 On the other hand, in 1226 the reinstalled Templar garrison successfully withstood an Aleppine besieging army arriving from the east. Although this force had actually been encouraged to attack Baghras by the Armenians and therefore posed no threat to them, its failure to capture the fortress indicates that it could also be used to defend the Cilician Plain against the Muslims.01 Depending on who occupied it, Baghras could be used to protect or attack both Antioch and Ciucia. It seems reasonable to conclude that whilst some scholars have overestimated the extent of its Armenian remains, others have underestimated its strategic importance. Leon's actions also help us to understand the military role of other Templar fortifications near Baghras. In the same way that Baghras guarded the Belén Pass, so the Templar castle of Darbsak controlled the entrance to a more northerly defile through the Amanus mountains during the twelfth century, but it was never recaptured by the Order after 1188.62 Nevertheless Leon's attempt to retake the fortress in 1205, another failed crusader attack in 1237 and Hethoum I's eagerness to reoccupy the site after the Mongol invasion of 1260 all indicate that Darbsak had once been a key Templar stronghold whose loss potentially threatened both Prankish and Armenian security.63 On the Cilician side of this valley, meanwhile, the Templars still retained Roche Roussel (Hadjar Shoglan/Chilvan Kale) in the thirteenth century. This castle was located on yet another steep outcrop which required few manmade defences, although some of the remains there, most notably a square keep and a chapel, may date from the Templar occupation.64 It was from here that the Templars launched their failed expedition against Darbsak in 1237, illustrating that the role of Roche Roussel had changed dramatically since 1188. Before that date Rouche Roussel, Darbsak and Baghras had all worked together to create a network protecting Antioch from the north, but in the thirteenth century Roche Roussel found itself caught between the Muslims in the east and the Armenians in the west. Presumably the unidentified neighbouring castle of Roche Guillaume suffered a similar fate after 1188, when it continued to be guarded by the Templars.65

l86

UNKNOWN CRUSADER CASTLES

The Templar lordship also extended beyond the Amanus mountains to the coast, where the Order held Port Bonnel (Arsouz).66 Moving further north, it seems unlikely that the Templars owned any properties on the Cilician Plain itself during the reign of Leon II, or indeed his successor Hethoum I, who became involved in further clashes with the Order during the 1230S.67 It has sometimes been argued that the Templars possessed Canamella, a coastal port to the north of Alexandretta. This theory may be based on Ibn al-Furat's claim that during the Mamluk raid of 1266 'a Templar fortress known as al-Tina, or according to another version al-Tinat (Canamella), was destroyed, and a large number of fortresses and towns of theirs were burned and destroyed'.68 But this source should not be accepted at face value, for we have seen that Canamella was held by the Hospitallers, whilst Amouda, another fortress which the chronicler attributed to the Templars, actually belonged to the Teutonic Knights. Ibn al-Furat appears to have confused all three Orders, and there is no reliable evidence that the Templars held any major properties north of Roche Roussel. So far Baghras and its neighbours have been discussed in purely military and strategic terms, but some observations also need to be made about the economic and political significance of these castles. First, during the years of failed negotiations between Leon II and the Templars the king had clearly attempted to strike a deal with the Order whereby it would regain Baghras if it agreed to support Raymond Roupen's claims to Antioch. These efforts explain why Leon waited until 1216 before he finally returned the site to the Templars, for in that year Raymond Roupen succeeded in taking control of Antioch, thereby removing the need to keep Baghras as a bargaining counter.69 Meanwhile Leon had attempted to seek Templar support against the Seljuk Turks,70 suggesting that his dispute with the Order should not simply be regarded as a straightforward territorial conflict, but was also a failed attempt by Leon to build up an alliance with the Templars in the same way that he had done with the Hospitallers. Leon had managed to convince the Hospitallers to abandon Bohemond IV in favour of Raymond Roupen and to provide valuable military support against the Turks in return for numerous castles and estates in Ciucia. Perhaps Leon hoped for similar success with the Templars, but he made a serious misjudgement by thinking that they would support him in exchange for Baghras, which they already regarded as their own property. Financial concerns must also have come into this, for by occupying Baghras and damaging neighbouring properties Leon threatened to undermine local Templar resources to such an extent that the Order would be forced to come to terms. It has been pointed out that Roche Roussel, Roche Guillaume and (possibly) Port Bonnel alone raised annual revenues of 20,000 besants for the Templars. Like the local

THE M I L I T A R Y ORDERS

l8j

strongholds of the Hospitallers and the Teutonic Knights, Templar castles consequently fulfilled an important function as agricultural centres for valuable surrounding estates.71 In some ways the involvement of the Military Orders in Cilician Armenia during the first half of the thirteenth century can be summed up as a case of lost opportunity. All three Orders held frontier lordships whose functions were theoretically the same as the great Hospitaller and Templar domains which shielded the county of Tripoli from the Muslims. In reality, the Teutonic Knights of Harunia lacked the resources to take an active role against the Seljuks and ultimately failed even to defend their own castles against the Mamluks. The Hospitallers, meanwhile, were deprived of Silifke for internal political reasons, whilst the Templars refused to be bullied into joining Leon II against either the Seljuks or Bohemond IV. Indeed, after Baghras had finally been restored to them, they continued to have such bad relations with Ciucia that 'far from being the frontier either of Antioch or Armenia, the Amanus was a semi-independent territory in which the Templars went their own way'.72 Here they conducted separate campaigns against the Muslims with little regard to their Armenian neighbours.73 Despite these setbacks all three Military Orders still had some involvement in Cilician Armenia beyond the first half of the thirteenth century. A document dating from 1271 shows that the Teutonic Knights were still holding Harunia in that year,7* although it may have succumbed to the Mamluk invasion which sacked neighbouring Servantikar in 1299.75 Amouda may not have been repaired after the Mamluks destroyed it in 1266,76 for in 1298 another Egyptian army used the castle as a meeting point for its campaign on the Cilician Plain.77 When Roche Guillaume fell the following year, the Templars probably lost their last major castle on the mainland.78 Yet the source which records this event does not specifically mention the Templars. As they had already lost neighbouring Roche Roussel, Baghras and Port Bonnel shortly after Baybars captured Antioch in 1268,79 it is conceivable that they no longer owned Roche Guillaume in i299.80 The Templars had abandoned their other castles without even attempting to resist Baybars; a decision which reflects the increasing power of the Mamluks and the deteriorating military situation throughout the area.81 These factors, combined with the eventual dissolution of the Templars and the growing commitments of the Teutonic Knights in the Baltic, ruled out any further involvement by these two Orders in Cilician Armenia after the early fourteenth century. The Hospitallers, in contrast, continued to play an active role in the defence of the kingdom almost until the very end. After they were forced to relinquish Silifke in 1226 they probably did not hold any major castles in the area, but this did not prevent them from owning other estates or

l88

U N K N O W N CRUSADER CASTLES

participating in local campaigns involving the Armenians, the Mongols and the Cypriot Franks. The transfer of lands to the Order by the lord of Lampron as early as 1233 suggests that the succession dispute of the previous decade had been a temporary argument which did not lead to a lasting rift with the Hethoumids.82 If the Muslim chronicler Ibn al-Furat confused the Templars with the Hospitallers, as has been suggested, the latter Order also still controlled the port of Canamella in 1266, when it was sacked by the Mamluks.83 Thereafter the Hospitallers are mentioned regularly by the sources: in 1281, for example, they probably contributed troops to the Mongol-Armenian force which clashed with the Mamluks near Horns.84 After the fall of Acre in 1291 the Hospitallers continued to send aid to the region, which was often regarded as a potential starting point for the reconquest of the Holy Land in a joint campaign with the Mongols. Between 1300 and 1305 the Order sent two major expeditions to Cilician Armenia.85 During the 13605 they also participated in the attacks made against the Cilician coast by King Peter I of Cyprus (1359-69), including the capture of Satalia (Antalya) in i36i.86 Two factors prevented the Hospitallers from becoming more involved in Cilicia, let alone participating in the hoped for recapture of Jerusalem. First, their conquest of Rhodes between 1306 and 1310 led to a shift away from the eastern Mediterranean to the Aegean, where numerous projects, such as the occupation of Smyrna and the defence of Latin states in the Peloponnese, usually absorbed all of their already overstretched resources. Secondly, the move to Rhodes coincided with an internal conflict on Cyprus between King Henry II (1285-1324) and his brother Amaury. During this dispute the Hospitallers supported Henry but King Oschin of Armenia (1307-20) supported Amaury. Consequently, Oschin confiscated the Hospitallers' Cilician estates, which subsequently stayed under royal control until the 132OS.87 By the 13303 these properties 'still existed on paper', but their extent remains uncertain and eventually they were presumably all lost to the Turks or to the Mamluks.88 Meanwhile the volatile fourteenth-century history of the Armenian monarchy, combined with the hostility felt toward westerners by some local nobles and churchmen, added to the reluctance of the Hospitallers or other Latins to get too closely involved. These problems are summed up by the suggestion made in 1332 by Pope John XXII that the Hospitallers should garrison the two local castles of Siquinum and Antiochetta. These coastal fortresses were both situated to the west of Silifke in the vicinity of Anamur and were therefore important in the struggle to keep the Turks at bay. Nothing came of this proposal, confirming that financial problems and mistrust of the Armenians prevented the Hospitallers from holding any major Cilician fortifications during the fourteenth century.89

PART FOUR

Prankish Greece 1204-1380

This page intentionally left blank

18

Warfare The Fourth Crusade (1202-04), consisting of a land army composed of French and Italian troops and a powerful Venetian naval fleet, had originally been planned as an offensive against Egypt. Through a combination of greed, political intrigue and mutual distrust, the expedition ended up attacking the Greeks instead.1 After the Crusade captured Constantinople in 1204, those members of the expedition who chose to stay in Greece undertook the conquest of what remained of the Byzantine empire. This resulted in the creation of several new Latin states around the Aegean. The Latin empire itself encompassed Thrace and the northern fringes of Asia Minor, while Macedonia and parts of Thessaly formed the kingdom of Thessaloniki. Further south, Byzantine Attica and Boeotia were turned into the Prankish duchy of Athens, and below it the principality of Achaia covered the Peloponnese. Meanwhile the Venetians established a number of important trading colonies around the Aegean, particularly at Modon, Coron, Constantinople and Crete. They also dominated numerous islands which lay either directly within the Venetian sphere of influence or were colonised by individual Venetian citizens. These included Corfu, Cephalonia and Euboea, as well as the Cyclades, which formed the duchy of the Archipelago under the Sañudo dukes of Naxos. Eventually other islands came to be controlled by rival powers, most notably the Genoese, who held Chios and several neighbouring trading posts between the fourteenth and sixteenth centuries. In addition, from 1306 onwards the Hospitallers subjugated Rhodes. Leading Greeks who had fled from Constantinople during the Fourth Crusade also created three new states situated on the fringes of the Byzantine world. The capital cities of these states lay at Trebizond along the Black Sea, at Nicaea in Asia Minor and at Arta in Epiros. All three of these territories, and in particular the latter two, saw themselves as the natural heirs to the Byzantine empire and consequently often clashed both with each other and with their unwelcome new Latin neighbours. To the north of Constantinople, the Franks were also threatened by the Bulgars, who had been opponents of the Byzantine empire before 1204 and now continued to launch frequent attacks into northern Greece in conjunction with an aggressive tribe of horsemen known as the Cumans.2

192

UNKNOWN CRUSADER CASTLES

This highly splintered political situation inevitably led to almost constant fighting in the Aegean region and ultimately resulted in the destruction of virtually all the Latin states which had been set up after 1204. The first to disappear was the ephemeral kingdom of Thessaly, whose conquest by the despots of Epiros was effectively completed with the fall of Thessaloniki in 1224. At this time the Epirote Greeks also made some significant advances against the duchy of Athens to their south. Meanwhile, the Greeks of Nicaea made good progress in the east. By the end of the 12205 the Franks had lost their possessions in Asia Minor and the Latin empire itself had been reduced to the area immediately around Constantinople, leaving the rulers of Nicaea and Epiros to fight it out for control of northern Greece. Indeed, as the Nicaean Greeks crossed the Sea of Marmara and began to advance westwards, those of Epiros became so alarmed that they formed an alliance with their former enemies, the Franks of Achaia. In 1259 this alliance was ended by the Nicaean ruler Michael VIII Palaiologos (1259-82), who inflicted a crushing defeat on William II of Achaia (1246-78) at the battle of Pelagonia in southern Macedonia. As a result of this encounter Epirote power in northern Greece was weakened, and Prince William found himself in captivity until 1262, when Michael finally released him in exchange for the Achaian fortresses of Mistra, Old Mania and Monemvasia. These powerful castles were located in the south-eastern corner of the Peloponnese and provided Michael and his successors with a perfect bridgehead from which to reconquer Achaia from the Franks. This process continued sporadically until the 14203, when the last Latin remnants of the principality were swallowed up and incorporated into the Greek province of Mistra. Two years after the battle of Pelagonia, the Nicaean Greeks also managed to recapture Constantinople, thereby wiping out the Latin empire. This made them masters of a newly recreated Byzantine empire stretching from western Asia Minor to Thessaly, and enabled Michael VIII and his son Andronikos II (1282-1328) to apply further pressure on the despotate of Epiros, the principality of Achaia and the duchy of Athens. While Franks in the Peloponnese struggled to contain attacks from Mistra, the 12708 witnessed the temporary conquest of Euboea by Licario of Karystos, a Veronese adventurer in Michael VIIFs employ. Meanwhile, the duchy of Athens came under attack from the north, and in order to halt these advances and to strengthen his position in relation to other Greek lords in Epiros and Thessaly, Duke Gautier I of Athens (1308-11) eventually decided to seek external assistance. This led to the arrival of a ferocious band of Catalan mercenaries in central Greece. These soldiers of fortune had previously been used by Andronikos II to fight the Turks in Asia Minor, but after falling out with their former employer they had gradually moved west

WARFARE

193

through the Byzantine empire, pillaging and looting as they went, before they eventually found themselves in the pay of Duke Gautier. However, after some initial successes an argument broke out between the duke and the volatile Catalans over pay, and as a result they took to the field against Gautier and his Achaian allies. This decisive encounter, traditionally known as the battle of Cephissus but actually fought at Halmyros (southern Thessaly), took place in March 1311 and led to the death of the unfortunate Gautier along with virtually all of his Athenian knights. In its wake Gautier's lands were conquered by the Catalans, who then colonised and ruled the duchy of Athens until the late fourteenth century. During this period the Catalans acknowledged the overlordship of the kings of Aragon, who held the ducal title and ruled Athens through representatives whom they regularly sent out to Greece.3 The dramatic advances made by the Greeks against the Latins had an important effect on the internal history of Achaia. Originally the Prankish rulers of this crusader state had recognised the Latin emperors as their overlords, but after Michael VIII recaptured Constantinople in 1261, and gained control over Mistra, Old Mania and Monemvasia the following year, it became essential for the Villehardouin princes of Achaia to find a powerful new suzerain who could aid them against the Greeks. During the 12605, therefore, William II of Villehardouin (1246-78) allied himself with Charles of Anjou, brother of Louis IX of France and ruler of Naples (1266-85). As part of this alliance it was agreed that, if William did not produce a male heir, Achaia would eventually pass under Charles's control, which was exactly what happened following William's death in 1278. These events turned out to be a mixed blessing for the Peoloponnese, for on the one hand Charles and his Angevin descendants did provide sporadic military support for the Peloponnese (as William had hoped), but on the other relatively few of them ever visited the region, preferring instead to send royal representatives from Italy or to grant parts of Achaia to their own followers. Perhaps the most important such figure was the wealthy Florentine lord Niccolo Acciajuoli, who acquired large parts of northern Morea, including the castellany of Athens, between the 13308 and 13505. Toward the end of the fourteenth century Niccolo's descendants also made significant advances against the Catalans of Athens, who were detested in France and Italy because of their close links with the Aragonese, arch-enemies of the Angevins and the Avignon papacy. The fact that the Angevin rulers of Achaia resided in Naples rather than Greece gradually had the effect of weakening central authority in the Peloponnese, so that the region's late fourteenth- and early fifteenth-century history was reduced to a series of internal clashes between various local

194

UNKNOWN CRUSADER

CASTLES

factions. In the north the Florentine lords of Corinth, allied with a new company of Navarrese mercenaries who had arrived in the 13808, fought against the Catalans. In the south east the Greeks of Mistra continued to advance, whilst Achaia itself sporadically found itself disputed between rival Angevin claimants. As the Turkish threat grew, the Venetians and the Hospitallers also became more involved in the politics of Latin Greece, sometimes placing entire areas under their own protection. Indeed, from the 13905 onwards, local Christian squabbles became little more than an irrelevance as the Ottoman Turks began to overrun the entire region regardless of whether it was controlled by Greeks, Italians, Franks or Iberians. By 1460 the Byzantine empire had disappeared and most of the Greek mainland had been incorporated into the Ottoman empire. During the next two centuries all remaining Christian islands in the eastern Mediterranean also fell to the Turks.4 During this period the military strength of the Latins and their opponents varied considerably. The relatively short life span of the kingdom of Thessaly and the Latin empire, for example, indicates that here western settlers found themselves under extreme pressure from the very beginning. It has been estimated that the total number of cavalry, including both knights and mounted sergeants, available to the first Latin emperor, Baldwin of Flanders (1204-5), only came to between 500 and 1000 horsemen.5 Contemporary accounts bear this out, for in late 1204 a mere 120 knights left Constantinople, crossed the Bosphorus and began to conquer as many lands as they could from the Greeks of Nicaea.6 Similarly, Baldwin's successor Henry (1206-16) campaigned against the Bulgars with a total of 400 knights in 1206, and was accompanied by 260 knights when he advanced against the Nicaean Greeks in mi.7 Initially these figures do not seem small when compared with Greek forces, for the largest reliable total given for a contemporary Nicaean army is that of 2000 cavalry (including 800 Latin mercenaries) at the battle of Antioch on the Meander, an encounter with the Seljuk Turks which took place in i2ii.8 It should also be noted that the advances of 1204 and Henry's campaign of 1211 both ended in Prankish success, suggesting that Latin and Nicaean forces were evenly balanced at this time, particularly when the Greeks were also struggling with their Seljuk neighbours to the east. To some extent the same may have applied to the Cuman and Bulgar troops who threatened the northern border of the Latin empire and the kingdom of Thessaly. In 1208, for example, a mere 2000 Latins allegedly confronted 33,000 Bulgars in northern Greece, but the fact that this encounter ended in a Prankish victory certainly casts doubt on the latter figure and suggests that the Bulgars could not possibly have enjoyed such an

MAP 5. The Agean

196

U N K N O W N CRUSADER CASTLES

overwhelming numerical superiority.9 Similarly, the combined armies of Nicaea and Bulgaria which reached the walls of Constantinople in 1235 must have been smaller than western sources imply, otherwise they would have annihilated the 140 Latin knights who emerged from the city and successfully drove them off under the leadership of John of Brienne.10 Here we are confronted with the usual problem of trying to establish accurate troop numbers from medieval chroniclers who tended to exaggerate (or simply invent) the size of hostile armies. Whilst the number of western knights mentioned by these sources appears to be realistic, they also fail to provide us with any detailed information about the quality and quantity of other troops such as sergeants, infantry and archers, even though these soldiers presumably made up the bulk of Latin armies. We can only assume that an apparently miraculous victory inflicted on tens of thousands of opponents by a few hundred western knights actually represented a more equally matched contest between a Prankish army which was really somewhat larger and a Greek, Bulgar or Cuman force which must have been considerably smaller. It still appears that Latin armies in northern Greece were generally outnumbered, and certainly too thinly spread out to defend both this region and the frontier with Nicaea at the same time. When in 1205 Kalojan of Bulgaria (1197-1207) invaded Prankish territories and encouraged the Greek population of Thrace to rebel, the Latin emperor Baldwin had to persuade his brother Henry to abandon 'all that he had conquered' in Asia Minor to cross the Bosphoros and help to defend the empire's European borders.11 Even these drastic measures failed to prevent Kalojan and his Greek allies from occupying virtually all of Thrace. Consequently, even if we dismiss Villehardouin's exaggerated claim that there were an additional 14,000 Cumans alongside all the other troops in Kalojan's army, we must surely agree with his general conclusion that the Latin emperor 'could not raise enough troops to defend his territories'.12 These events also confirm that, apart from having to cope with their external enemies, Latin settlers were heavily outnumbered by a potentially hostile native population. Indeed, during Kalojan's invasion some Franks continued to fear for their lives even after they had found shelter inside Tchorlu, a fortified city to the west of Constantinople, 'because they doubted the people of the town'.13 It was not the overwhelming superiority of any one opponent but the number of different enemies, the unreliability of the local people and the sheer size of the territories which needed to be defended that contributed to the early demise of the Latin states in northern Greece. It is perhaps hardly surprising that the crusader states set up in south-western Greece and on the Aegean islands survived much longer. The duchy of Athens,

WARFARE

197

the principality of Achaia and the Venetian colonies were all smaller in size and were protected by seas, mountains or other natural features such as the isthmus of Corinth. Moreover, the further Latin conquerors travelled from Constantinople, the more they benefited from the increasing regionalism which had been apparent within the Byzantine empire since before the Fourth Crusade. When Boniface of Montferrat, the man who established the kingdom of Thessaly, undertook his conquest of northern Greece in 1204, he received a warm welcome from some locals who regarded Constantinople as a distant and corrupt absorber of taxes, and consequently felt that it was best simply to come to terms with the Latins.14 Similarly, when the Franks reached the Peloponnese local Greek lords ('archons') were often allowed to keep their incomes provided that they recognised their new Prankish masters. By contrast, the volatile political situation in Constantinople itself before and during the Fourth Crusade, followed by the flight of many leading Greeks to Nicaea or elsewhere, left little room for a quiet transition of power in northern Greece.15 The relative stability which existed in south-western Greece during the first half of the thirteenth century ensured that the rulers of Achaia were more prosperous and enjoyed greater military power than their Latin neighbours in Thessaloniki and Constantinople. The total number of knights settled in the region has been estimated at between five and six hundred, and according to one rather romanticised source Prince Geoffrey II of Villehardouin (1228-46) kept eighty knights just at his own court.16 These figures are confirmed by the Chronicle of Morea, if we can believe its claim that, at the battle of Tagliacozzo (1268), Geoffrey's successor William II (1246-78) fought alongside his overlord Charles of Anjou with 400 cavalry brought from Achaia.17 This implies that, even though it was smaller, the military strength of Achaia was almost equal to that of the original Latin empire, and that its rulers had enough resources to intervene in conflicts beyond the Peloponnese. Indeed, in 1236 Geoffrey II of Villehardouin sent help to Constantinople whilst it was being besieged by the Bulgarians and the Greeks of Nicaea.18 In 1249 Geoffrey's brother William II joined Louis IX's crusade against Egypt with a fleet of twenty-four Achaian ships. This episode took place at roughly the same time as the Latin empress of Constantinople made a desperate plea to Louis for 300 knights to help her and her husband defend their capital, and therefore highlights the contrast in military strength between the Franks of Achaia and those of northern Greece.1? The military strength of the principality of Achaia should not be overestimated, however, for it still took the Villehardouins many decades to subjugate the Peloponnese fully, and they did not capture the impregnable

198

UNKNOWN CRUSADER CASTLES

Greek outpost of Monemvasia, situated in the south-east corner of the principality, until 1249.2° Moreover, the days of sending troops to foreign conflicts came to an end after the 12605, when Monemvasia, Mistra and Old Mania were returned to the Greeks and subsequently used by them to reconquer the entire Prankish Morea. It is difficult to calculate the number of troops involved in this conflict, but the fact that it took the Greeks until the early fifteenth century to recover all of the Peloponnese suggests that local armies were fairly equally balanced. It has been estimated that even during the reign of Michael VIII (1259-82), which marked the zenith of late Byzantine power, major Greek campaign armies never contained more than 10,000 soldiers, and that most were much smaller.21 Bearing in mind that there never seem to have been more than five or six hundred western knights settled in Achaia, it seems unlikely that local clashes between Latin forces and the Greeks of Mistra ever involved more than a few hundred horsemen. When the Chronicle of Morea claimed that in 1262 300 Franks defeated a Byzantine force of 15,000 men, it may have given a reasonably accurate estimate for the Prankish army but must surely have exaggerated the scale of the Greek expedition.22 The Franks must have been less outnumbered than this, but the fact that the Greeks probably had a slight rather than an overwhelming advantage meant that the conflict dragged on for many decades. As the Franks gradually retreated toward the north and west of the principality, the land which they conceded was often devastated by regular yet indecisive fighting.23 The inability of any one ruler to deal a swift death blow to his opponents also seems to have characterised the fighting further north. During the first half of the thirteenth century the Greeks of Epiros quickly recaptured Thessaloniki and wiped out many Latin conquests which had been made in Thessaly, but they were unable to overrun the duchy of Athens. Epiros eventually became a frequent ally of the Latins against the Greeks of Nicaea, who were themselves only successful in weakening rather than destroying Epiros and Athens. The exact details of this fighting are either poorly recorded or heavily exaggerated. The papal correspondence of Honorius III (1216-27) and Gregory IX (1227-41) indicates that during the 12205 the Greeks of Epiros attacked Boudonitza and Salona on the northern fringes of the duchy of Athens, and that by 1235 they had even reached Thebes.24 No details are given regarding the strength and nature of the opposing forces, although the concern caused by these events at the papal court is probably in itself indicative of Epirote superiority on the battlefield. By the second half of the thirteenth century, it is clear that this superiority had largely disappeared in the face of Nicaean pressure, which forced Epiros to seek the friendship rather than the enmity of the Latins. Accurate details

WARFARE

199

of the fighting are still impossible to come by. Thus in 1292 the Chronicle of Morea stated that the Epirote city of loannina was besieged by a Byzantine (formerly Nicaean) force of 30,000 infantry and 14,000 cavalry, but that this entire expedition retreated in panic when Florent of Hainault, prince of Achaia (1289-97) and ally of Epiros, approached with an army which probably only contained four or five hundred men!25 Once again, the Byzantine figures should clearly be disregarded, for the number of western knights living in this region suggests that most clashes between Greeks and Latins cannot have involved more than a few hundred horsemen, and perhaps only a couple of thousand men in total. Indeed, it has been suggested that during the entire late Byzantine period most Greek campaigns 'probably involved hundreds rather than thousands of troops'.26 Whilst the Nicaean Greeks eventually succeeded in conquering Achaia, we have seen that it was the Catalans who finished off the Prankish duchy of Athens. The ruthlessness with which they did so suggests that they were unusually well-trained and aggressive soldiers. According to the chronicler Muntaner the Catalan company consisted of 2500 cavalry plus 4000 Almogavers (light infantry) and 1000 other footsoldiers27 These figures seem high, although the total of around 2500 horsemen was also given by another source.28 On the other hand, Muntaner claimed that, when they defeated Duke Gautier I of Athens in March 1311, the Catalans overcame a Prankish army containing 30,000 infantry, plus 700 knights of whom all but two were killed. Even if all the lords of Athens and Achaia had turned up, these are impossibly high totals and unfortunately cast doubt on the other troop numbers given by this chronicler.29 The decisiveness of the Prankish defeat does at least suggest that Gautier had underestimated both the fighting skill and the size of the Catalan foe, even if the total number of combatants involved in this encounter was greatly exaggerated. The Catalans also participated in another form of warfare which affected much of the Aegean area during the thirteenth and fourteenth centuries: piracy and raiding. Even before they arrived in the duchy of Athens the Catalans had a terrible reputation for such activities in Byzantine Greece, where 'there was not a town or city that was not looted and burnt' by them.30 From the mid fourteenth century onwards the Turks became an even greater threat, causing so much destruction that by the 13805 some Aegean islands had been completely abandoned by their original inhabitants. Even as early as 1358 a document recording the transfer of Corinth castle to the Florentine lord Niccolo Acciajuoli makes it clear that the Turks were having a devastating effect on the region: one of Niccolo's primary duties was to persuade people to return to the area around Corinth, which had been overrun so many times by Greek, Catalan or Turkish raiders that

200

U N K N O W N CRUSADER CASTLES

homelessness and starvation had forced entire communities to flee.31 In the thirteenth century there may not have been quite as much destruction, but piratical attacks, such as that carried out during the i28os against Bartolomeo Ghisi, the Venetian lord of Tinos and Mikonos, nevertheless represented a common threat for the inhabitants of islands and coastal regions.32 There were also many clashes between Venetian and Genoese naval forces which had either been sent out by their respective cities or were operating independently in a semi-piratical manner.33 When the Venetians occupied Corfu and Crete in the years after the Fourth Crusade, they first had to get rid of Genoese privateers based on these islands.34 The behaviour of the Catalans and the constant rivalry between Venice and Genoa also remind us that far from being united in their struggles against the Greeks or Turks, Latin settlers in the Aegean area were frequently embroiled in their own internal disputes. Warfare of this kind squandered precious resources and resulted in yet more suffering for those who already had to cope with the threat of piracy, looting and brigandage. Between 1255 and 1258, for example, Duke Guy de la Roche of Athens (1225-63), aided by his Venetian allies, fought a bitter civil war against William II of Villehardouin in a dispute over land on Euboea. William eventually won this conflict by systematically ravaging Guy's lands until he surrendered; but, although this tactic worked, it must have caused untold suffering for the peasants and farmers who lived in the affected areas.35 Considering that most of these unfortunate rural inhabitants were Greek, it is hardly surprising to find that fighting also broke out occasionally between western newcomers and local natives, who sometimes had to put up with other forms of maltreatment such as unbearably high taxes.36 Incidents of this kind must have been particularly unnerving as Latin settlers formed such a tiny section of the overall population. On Chios, for example, it has been estimated that there were 10,000 Greeks being ruled by a mere one to two thousand Genoese toward the end of the fourteenth century.37 Villehardouin claimed that, during the initial Prankish siege of Constantinople in 1203, the city's inhabitants outnumbered their attackers by as much as two hundred to one, and that the crusaders were consequently reluctant to enter Constantinople even after they had seized the city walls.38 Such statistics help to explain why Latin conquerors in Achaia and elsewhere often tried to be as conciliatory as possible toward local Greek landholders. One final form of combat which needs to be mentioned is that of naval warfare. Generally speaking, the Latins possessed far greater naval power than the Greeks; a factor which proved vital to the continued existence of some of the crusader states. In 1235 and 1236, when Constantinople was besieged by the combined forces of John Asen II of Bulgaria (1218-41) and

WARFARE

201

John III Vatatzes of Nicaea (1222-54), the Venetians were able to break the naval blockades around the city and effectively save the Latin empire from destruction.39 Apart from the Italian city states, other Latin powers also had warships available to them. In 1236 many vessels from Achaia assisted in the relief of Constantinople and William II of Villehardouin later contributed a fleet of twenty-four ships to the crusade of St Louis. At the beginning of the fourteenth century the Catalans also brought their own fleet to the Aegean, using it in countless raids along the coasts of northern Greece. In 1305, for instance, after the Catalans had fallen out with their former employer Andronikos II, they sent five galleys out from their temporary headquarters at Gallipoli to attack neighbouring Byzantine targets.40 The financial, administrative and political problems of the late Byzantine state often prevented the Greeks from raising adequate naval forces to deal with such aggressors. During the 12808, for example, Andronikos II reduced the size of the Byzantine fleet in order to cut costs.41 However, it would be misleading to assume that the Greeks were entirely overawed at sea, for at times, and in particular during the reign of Michael VIII Palaiologos (1259-82), the Byzantine authorities were able to build up naval forces by relying on mercenaries. Thus in the 12705 Michael VIII employed the Latin pirate Giovanni de lo Cavo to act as his admiral in the Aegean.42 Latin seamen also did not have matters entirely their own way because of the growing threat of the Turks. Even the ruthless Catalans had to be aware of this danger, for during the winter of 1303-4 they sent their fleet to a secure winter anchorage at Chios 'because the Turks, with barques, ravage all [the Aegean] islands'.43 This summary of thirteenth- and fourteenth-century warfare around the Aegean makes it clear that there were many motives for the Latins who lived there to construct or occupy castles. Although exact troop numbers are almost impossible to establish, most of the Greek or Latin armies active in the region only contained between one and five hundred horsemen. Even when additional foot soldiers accompanied them, campaigns were rarely undertaken by more than one or two thousand men. Consequently the Latins who settled in Greece after the Fourth Crusade were not as heavily outnumbered by individual opponents as some of the contemporary chroniclers would lead us to believe. There were, however, ultimately too few Latins living in northern Greece to halt the combined attacks of many different Nicaean, Epirote, Cuman and Bulgarian armies. The fall of the duchy of Athens in 1311 suggests that Prankish settlement was so fragile that a single defeat in a pitched battle could seal the fate of an entire crusader state. Eventually the Franks proved equally incapable of halting the gradual loss of Achaia after the Greek acquisition of Mistra in 1262. In addition,

202

UNKNOWN CRUSADER CASTLES

the Aegean was almost constantly affected by some form of localised raiding, rebellion or piracy. While this did not necessarily threaten national security in the short term, it ultimately ground down the economic and military strength of states such as Achaia. In short, this was an extremely insecure world, and in order to protect themselves against it Latins needed fortifications to make up for their lack of troops and to defend their property against the constant threat of enemy attack.

19 Military Architecture The sheer quantity of medieval ruins still standing in Greece and Turkey bears witness to the turbulent history of the region. This should make the task of identifying and describing crusader fortifications relatively simple, but this is not the case. There are a number of problems which make the study of Prankish castles around the Aegean fraught with difficulties, not least the lack of historical sources. Bearing in mind that there are certain periods, particularly in the fourteenth century, when 'we cannot reconstruct completely' the history of southern and central Greece in general, it is hardly surprising that the fate of individual castles frequently remains obscure for decades at a time.1 Moreover, when the Latins who settled in this area constructed or repaired fortifications, they adopted the same tactic of recycling older masonry as their Byzantine predecessors. As a result, it is often impossible to tell whether a wall composed of classical masonry blocks robbed from a much older structure is the work of Greeks or Latins. Trying to establish if a strongpoint is Prankish, Catalan, Navarrese or attributable to another western dynasty (such as the Acciajuoli lords of Corinth) is even harder. Inevitably, these problems have led to arguments about dating and origin amongst the few archaeologists and historians who have studied the subject. Such arguments can perhaps best be summed up by looking at the castle of Androusa, located at the heart of the principality of Achaia in the south-western Peloponnese. The present remains of Androusa include sections of a single curtain wall, flanked by several rounded, square and polygonal salients, as well as a large tower whose design and unusually thick walls suggest that it acted as a keep. The castle's masonry is typical for medieval Greece: a mixture of small, uncut stones quarried locally, shards of pottery and other recycled fragments found at the site, and larger, presumably classical, masonry blocks used to strengthen corners, doorways and other weak spots.2 In his monumental work, on the castles of Morea, Antoine Bon dates this stronghold to the mid-thirteenth century, with possible additions in the fourteenth century. The Aragonese version of the Chronicle of Morea appears to confirm this when it states that Androusa was built by William

204

UNKNOWN CRUSADER CASTLES

II of Villehardouin, who reigned from 1245 until 12/8.3 But there are problems with this conclusion. A number of features at Androusa, and in particular the use of decorative brickwork and the presence of a pentagonal open gorge tower along the north curtain, are typically Byzantine, but extremely rare in Prankish military architecture.4 Bon accounts for these anomalies by suggesting that the crusaders used local craftsmen, which may well be true, but he never seems to consider the possibility that some of the defences at Androusa were constructed either before 1205 or during the brief period of Greek domination in the area between the 14205 and 14505.5 Likewise, when discussing the date of the oldest fortifications at Mistra, Bon again lets the historical evidence take precedence, basing his conclusion that the castle must have been constructed 'in its entirety' by William II on a statement to that effect in the Chronicle of Morea.6 In a clear reference to Bon's work, the Byzantine scholar David Winfield challenges these findings on archaeological and architectural grounds, asserting that at Mistra 'there is nothing in the hill top citadel and not very much on the site as a whole to indicate that Villehardouin did more than reoccupy a typical Byzantine hill town'. Winfield then goes on to claim that in Greece 'a number of fortifications that are probably Byzantine have been assigned without question to the Franks'.7 Both scholars therefore take somewhat extreme views, with Bon preferring to believe the Chronicle of Morea even if this contradicts the archaeological evidence, and Winfield taking a more Byzantine stance which immediately assumes that the written sources must be wrong. In trying to resolve this argument we immediately stumble upon two further problems. First, it should be noted that many of the key fortifications occupied by the Latins between 1204 and 1380 were subsequently taken over, repaired or even totally rebuilt by the Venetians. By 1715 the medieval defences of sites like Corinth, Nauplia, Monemvasia and Patras had all been partially or totally hidden below new ramparts and artillery bastions.8 From around 1500 onwards the Turks likewise demolished some older crusader defences, as appears to have been the case at Boudonitza,9 or carried out their own alterations and improvements, as can be seen at Modon.10 Secondly, and perhaps as a result of all these other difficulties, subsurface archaeological exploration of medieval strongpoints in Greece has barely even begun. It cannot yet compete with the kind of detailed studies which have been carried out on Crac des Chevaliers and other sites in the Holy Land. As a result it would be pointless, and in most cases virtually impossible, to describe the appearance and history of every stronghold in Prankish Greece. It is more useful to give specific examples of medieval fortifications in order to shed light on their various functions.

MILITARY

ARCHITECTURE

205

Misura. Note the keep (A), possibly added by the Franks.

All crusader fortifications in Greece can be placed in three general categories: those which the Latins built from scratch on sites showing very little, if any, evidence of previous occupation; those which were constructed incorporating older Byzantine or classical structures; and, finally, those which were already in good condition at the time of the Fourth Crusade, and were simply reoccupied by the Franks. In Thrace, Macedonia, Asia Minor and eastern Thessaly most strongholds fall into this latter category, mainly because the crusaders were too poor or in control too briefly to carry out their own repairs and building programmes. Thus Appolonia, located on the furthest outskirts of Prankish territory facing Nicaea, was referred to as 'one of the strongest and most imposing castles to be found' as early as 1204, and clearly needed few improvements before the crusaders could garrison it with their own troops.u North of Apollonia, a series of equally well defended coastal strongholds formed the backbone of Prankish power in Asia Minor. Perhaps the most important of these was Spiga (Pigae), one of the last Latin outposts to fall to the Nicaean Greeks in 1225. Located on a promontory forming a good natural harbour, Spiga was protected by a powerful Byzantine rampart flanked by a series of closely set pentagonal towers, which must have been maintained by the Franks and are still largely preserved to this day.12 On the European side of the Bosphorus, the crusaders inherited some even more spectacular Byzantine fortifications, particularly at Constantinople itself. This city's landward side was protected by the famous double walls of Theodosius II (408-50), as well as a number of later structures, including the rampart of Manuel Komnenos (1143-80), which was located near the

2O6

UNKNOWN CRUSADER CASTLES

Blachernae palace and bore the brunt of the Prankish land offensive in i2O3.13 That these defences had been kept in good condition is attested to by the events of the Fourth Crusade, when the Latins failed to take Constantinople from the west, being forced to concede that 'never was a city so well fortified'.14 Interestingly, the Venetians had far more success when they launched a naval attack on the seaward defences and managed to capture twenty-five towers along the Golden Horn by lowering wooden platforms from their ships onto the ramparts.15 In preparation for the second Prankish siege of Constantinople in April 1204, the Greeks were forced to heighten these towers 'with two or three wooden storeys', but were still unable to prevent the Venetians from gaining access to the city in almost exactly the same spot. This suggests that the single sea wall may have been lower, weaker and possibly in a far worse state of repair than the land defences. It also illustrates the huge advantage which Venetian naval power gave the crusaders over their Greek opponents.16 If the thirteenth-century sea defences of Constantinople were not in a particularly good state, the Franks were fortunate that the Venetians could patrol the Bosphorus for them, and that most of their opponents attacked the city by land. Baldwin of Flanders and his successors were all so lacking in troops and resources that 'no repairs are attested for the period of the Latin empire', and it seems that until 1261 the sheer size of the Theodosian walls alone saved the city from early capture by the Greeks or the Bulgars.17 These observations are confirmed by the fact that Michael VIII Palaiologos carried out extensive repairs on the ramparts between the 12605 and 12805.18 Contemporary accounts of the Prankish conquest of territories to the west of Constantinople suggest that many Byzantine settlements in Thrace, Macedonia and Thessaly were similarly well protected at the beginning of the thirteenth century. Demotika, for example, was described as 'a very fine, strong and wealthy castle' at this time, while the fortress of Christopoli (Kavala) was, according to Villehardouin, 'one of the strongest in the world'.19 Descriptions of the Thracian rebellion of 1205 and Kalojan's subsequent campaigns in the area also reveal that Arcadiopolis, Stenimaka, Philippopolis, Rousion and Rodosto all had urban fortifications of some kind. Meanwhile Adrianople was so well defended that Prankish attempts to capture the city during the initial Latin conquest of Greece failed despite a sustained siege involving ladders, catapults and sapping.20 Special reference should be made to Thessaloniki, whose vast defences, begun in the fourth and fifth centuries, formed a triangle around the city, with an impregnable hill top citadel at its apex.21 The history of these defences is obscure during the crusader period, although the few facts available to us suggest that they were in good repair in 1204, and that they

MILITARY ARCHITECTURE

2O/

subsequently proved just as essential to Latin settlers as the walls of Constantinople. Indeed, as early as 1210 they successfully halted an attempt to recapture the city by the despot of Epiros, Michael Komnenos Doukas (1204-14). Fourteen years later Michael's successor Theodore (1214-30) only managed to finish the job after a siege which lasted several months. In 1209 Thessaloniki's walls had also proved an insurmountable obstacle for the Latin Emperor Henry, who used cunning rather than force to enter the city and deal with a rebellion staged there by local Lombard nobles.22 Accounts of the Norman attack on Thessaloniki which took place in 1185 suggest that the sea wall had also been kept in good condition by the Greeks.23 Moving further west and south of Thessaloniki, the citadel of Larissa represented another perfectly intact stronghold at the time of the Fourth Crusade. This is implied by the events of 1209, when the emperor Henry arrived in the city and was obliged to besiege its rebellious Lombard defenders, who only withdrew from the citadel as a result of protracted negotiations rather than a successful assault by imperial forces. It is interesting to speculate whether the rapid conquest of northern Greece undertaken five years earlier by Boniface of Montferrat and Baldwin of Flanders, the first western rulers of Thessaloniki and Constantinople, would have succeeded if the Greeks had put up this much resistance. After the fall of Constantinople, virtually all the strongholds between the capital and Larissa surrendered to the Franks without any real struggle. Those places which did try to defend themselves, most notably Adrianople, caused the Latins considerable trouble. Hence the conquerors of 1204 were lucky, first because they were spared a series of lengthy sieges as they moved through Thrace and Thessaly, and secondly because they were able to reoccupy a large number of intact strongholds rather than having to build or repair brand new castles. It is extremely doubtful whether the Latins, with their limited numbers and resources, could have conquered Greece if the circumstances had not been so accommodating.24 The poverty and brief life span of the Latin states in northern Greece, combined with the good condition of existing Byzantine fortifications, either made it unnecessary or impossible for the Franks to build many new defences of their own. As one would expect, written records indicate that repairs were occasionally undertaken. For example, in 1206-7 Boniface of Montferrat restored the walls of Serres in response to Kalojan's attack on the city the previous year.25 The total lack of archaeological evidence for such projects, even at Constantinople, however, merely reinforces the impression that the Franks relied almost entirely on the works of their Greek predecessors. To the south of Larissa, where Latin rule was stronger and more permanent,

208

UNKNOWN CRUSADER CASTLES

it is likely that one of the few sites which remained untouched by the newcomers was Monemvasia, located on the east coast of the Laconian peninsula. Its situation on a vast, sheer-sided rock separated from the mainland by a narrow stretch of water made it so inaccessible that it was not captured by the Franks until 1249, following a bitter three-year siege. By 1262, however, the castle had already been handed back to the Greeks as part of William II's ransom following the battle of Pelagonia (1259).26 As a result, the Franks barely had time to integrate Monemvasia into the principality of Achaia, and it is extremely unlikely that they ever built any new fortifications there. This is confirmed by the archaeological evidence, which suggests that the Byzantine citadel is the only pre-Venetian structure on the site.27 The second category of castles and fortifications covers the vast number of strongpoints which incorporated a combination of both Prankish and older Byzantine or classical defences. Structures of this type varied enormously. As one would expect, they occurred most commonly in Achaia and the duchy of Athens, where Latin rule lasted much longer, but there were also a few such fortifications further east. At Nicomedia, for example, the Franks fortified a large Byzantine church and surrounded it with outer defences in a simple but effective arrangement which withstood at least one Nicaean siege.28 In 1206 the crusaders also reached Cyzicus, a narrow headland not far from Spiga, and found that 'there had been in ancient times a fortress with walls, towers and ditches; and they were nearly in ruins. And the army of the Franks entered it, and Peter of Bracieux, to whom the land had been assigned, began to rebuild it, and to construct two castles and two entrances'.2' Other castles in Asia Minor which were built or repaired using similar methods included Charax, Civetot and Panormos, all of which lay along the coast.30 Within the European half of the Latin empire, the chronicler Henry of Valenciennes wrote that in 1208 the emperor Henry decided to reconstruct the ruined castle of Pamphilon in Thrace. Having defeated a vast Bulgar and Curtían invasion force, Henry went to this spot, where he and his marshal swore not to leave 'until the walls had been rebuilt and repaired'. Local labourers were subsequently recruited and the work was completed with such speed and determination that a new Frankish garrison had been installed before the onset of winter. Although this castle must have been situated somewhere along the main road between Constantinople and Philippopolis, its exact location has never been established.31 Its successful reconstruction in such a determined fashion suggests that the reign of Henry (1206-16) represented the high point of the Latin empire. However, when reading about the fortification of places like Cyzicus and Pamphilon, one

MILITARY ARCHITECTURE

2O9

is still struck by the impression of great haste, and the lack of money, troops and resources. Clearly the ephemeral nature of Prankish rule within the Latin empire meant that it was not only convenient but essential to reoccupy older Byzantine or classical ruins, where the immediate supply of ready-cut stone facilitated the construction of new defences in a matter of weeks. In south-western Greece and the Peloponnese, greater political stability ensured that former Byzantine and classical sites were repaired, improved or even rebuilt on a far larger scale. This did not necessarily mean that crusader fortifications were constructed to a higher standard. At Patras, for example, William Alemán, the first Prankish lord of the city, considerably improved the defences of the lower bailey of the Byzantine citadel, but this was done using masonry robbed from the neighbouring archbishop's residence. Admittedly William was embroiled in a dispute with the archbishop at this time, and was therefore motivated by political as well as practical concerns, but his desperate actions also reflect the poverty and relative insecurity of many Prankish adventurers shortly after they arrived in Greece32 At other sites, the amount of alteration undertaken by the crusaders varied considerably, depending on whether these places were in ruins or were still occupied at the time of the Prankish invasion. A good example of the latter type of fortification is Kalamata, located on the coastal plain of south-western Morea. In 1205 the western newcomer William of Champlitte only obtained its surrender by promising the defenders that he would

Patras: one of the many Byzantine strongholds occupied and repaired by the Latins after 1204.

210

UNKNOWN CRUSADER CASTLES

respect their land and property.33 This proves that the stout double baileys of the fortress, which protected its more vulnerable eastern side, predate the thirteenth century. It seems that the only structure at Kalamata built by the crusaders was the keep, a huge, slightly rectangular building located at the highest point of the inner bailey and constructed using the same combination of classical ashlar blocks and small uncut stones as that already described at Androusa.34 Another stronghold which was probably occupied almost continuously from classical times up to the thirteenth century and beyond was Arcadia, situated near the Ionian coast on the other side of the Messenian peninsula from Kalamata. Like its neighbour, Arcadia also relied on its isolated position at the top of a steep, narrow hill for much of its defensive strength. This meant that the design of the castle was almost totally dictated by the shape of the summit. It enabled the original builders to leave particularly inaccessible areas, such as the sheer north side of the hill, virtually devoid of man-made defences, and to concentrate their efforts on the weaker southern and eastern approaches. Here two successive baileys and curtain walls, flanked by numerous towers and salients, defended the main access route to the summit itself.35 Small sections of classical masonry, some of it still in situ, suggest that this defensive arrangement is extremely old and was merely improved upon by the Byzantines. When the crusaders first arrived in 1205 they were immediately impressed by the sheer strength of Arcadia's fortifications, and in particular the large tower at the summit of the castle, which they believed had been built by 'giants'; a clear reference to the vast classical blocks used in its construction. Indeed, Arcadia did not surrender to the Franks until they had begun to bombard this tower with several catapults.36 Many of the upper levels of Arcadia's walls, however, were built using far smaller, uncut stones, whose appearance is so generic that it is very difficult to date them accurately. It does nevertheless seem that the large, round tower at the eastern corner of the site was constructed by the Franks, like the keep at Kalamata.37 Numerous other fortifications in southern and central Greece can be placed in the same group as Arcadia and Kalamata. These include Coron, Patras, Athens and Thebes, taken by the crusaders in i205,38 and Corinth, Nauplia and Argos, captured in 1210,1211-12 and 1212 respectively.39 At about the same time the crusaders also occupied Neopatras, which they held until its recapture by Theodore of Epiros in 1219.4° Almost all these places had to be taken by force, a clear indication that they were fully functioning military strongholds at the time of the Fourth Crusade. In terms of design and location they also shared a number of important characteristics both with each other and with Arcadia and Kalamata. Corinth, Argos, Neopatras

MILITARY ARCHITECTURE

211

and Thebes, for example, were all mountain castles, and were defended by successive walls arranged around an ancient acropolis. All four sites had either been occupied continuously since classical times or had been refortified during the political upheavals of the eighth and ninth centuries.41 Although Patras, Nauplia and Coron were located in less precipitous coastal areas, they too were built on hilltops or promontories which had already acted as refuge sites for many centuries.42 The great age of these sites can be seen from their masonry, which still contains the same mixture of high quality ancient blocks and more recent poorly cut stones as that used at Arcadia.43 From the early years of the thirteenth century onwards, the Latins carried out their own repairs and improvements on these castles. It is remarkable how often these alterations involved the addition of a strong, central tower or keep. Such structures have already been mentioned at Kalamata and Arcadia, and good examples have also survived at Corinth and Neopatras.44 It may have been the former structure that the Chronicle of Morea was referring to when it claimed that repairs were carried out on 'the castle of Corinth' by William II of Villehardouin (i246-78).45 Reference should also be made to the famous medieval tower on the acropolis in Athens, which was probably built by the Acciajuoli in the fourteenth century but was unfortunately demolished in 1874. From photographs it seems that this tower was unusually sturdy because it had been constructed almost entirely from smooth marble slabs, without the usual filling of smaller stones or rubbled mortar. Standing to a height of at least twenty-five metres, it must have been one of the most impressive medieval structures anywhere in Greece.46 It is also important to think of the acropolis itself not as a cluster of classical ruins but as a formidable refuge point which was still being used in a military capacity. As recently as 1205 it had successfully withstood a major attack by Leon Sgouros, the Greek ruler of Corinth, during the brief period of anarchy on the eve of the Prankish invasion.4? Other improvements carried out by the Latins normally amounted to the restoration or rebuilding of outer defences and curtain walls. Such work was undertaken by William Alemán at Patras, whilst at Corinth the Acciajuoli may have strengthened the ancient ramparts in the fourteenth century.48 One contemporary source also implies that the Venetians improved the Byzantine defences of Coron considerably, although the proliferation of post-medieval artillery defences makes this difficult to prove.4^ The Chronicle of Morea also noted that in the thirteenth century there were no less than two fortresses at Nauplia, implying that the lower slopes of the promontory were just as well protected as the summit itself.so It is possible that a short stretch of wall, a triangular bastion and two rounded towers visible beneath a collapsed fifteenth-century talus represent a small section of these defences.

212

UNKNOWN CRUSADER

CASTLES

Certainly a postern gate incorporated into the wall and built using recycled classical stone resembles later fourteenth-century Catalan work at Boudonitza and Salöna (Amphissa), where doorways were framed with large, antique masonry slabs.51 Similar building techniques were also used in the construction of a tower at Thebes, the only remaining structure from a larger castle erected by the Prankish lord of the city, Nicholas II of Saint Omer, in 1287. According to the Chronicle of Morea this was a palatial building in its heyday,52 but the recycled appearance of its surviving masonry, combined with the fact that Lombard rebels in Thebes had withstood a major attack by the emperor Henry in 1209, indicate that this was yet another site whose defences were originally Byzantine and had then been extended or repaired by Prankish newcomers rather than built from scratch.53 Many of these observations also apply to the ruined castles or acropolis sites which the Latins reoccupied in southern and central Greece. Unlike Arcadia, Corinth, Thebes and all the other strongholds just mentioned, these places had not been maintained in recent times, and therefore lay abandoned or were only inhabited as open settlements at the time of the Fourth Crusade. Nevertheless, their natural strength and ready supply of high quality stone made such ruins extremely attractive to the Latins, who exploited them in the same way that the Prankish rebuilders of Cyzicus had done in I2o6.54 Inevitably fortresses belonging to this group contained far more new structures dating from the thirteenth and fourteenth centuries, although their reliance on ancient ruins both for their location and for their building materials meant that their overall appearance did not differ very much from intact Byzantine strongholds such as Arcadia. At Boudonitza, an important castle on the northern frontiers of the Athenian duchy, we find the familiar arrangement of successive baileys dominated by a hill-top citadel, even though this stronghold was probably built almost exclusively by the Franks and the Catalans. The masonry at Boudonitza also confirms that the Latins, like their Byzantine predecessors, tended to rely on rather low-quality uncut stones quarried locally if it was found that there were not sufficient classical remains left on the site which could be recycled.55 Other interesting examples of this type of fortification include Salona and Zeitoun (Lamia) in central Greece, and Modon and Akova in the principality of Achaia.5« At Modon the crusaders came across an open settlement 'which had been without walls for a long time', even though this city was apparently still inhabited by Greeks living inside the ruins of much older urban defences. It seems that these remains were subsequently used by the Latins when they refortified the site in 1205, and they may have been recycled once more by the Venetians after they acquired Modon in 120p.57

MILITARY

ARCHITECTURE

213

Chlemoutsi: an unusually well-built fortress intended as a royal residence for the Villehardouin rulers of Achaia.

At other sites whose ancient ramparts had been abandoned by 1204, the Franks appear to have begun the process of restoration by constructing large, isolated towers out of the surrounding ruins. This was certainly the case at Akova, situated in the mountainous interior of the Peloponnese. Here it seems that a robust rectangular keep dominated the castle, which otherwise relied on a combination of sheer cliffs and thin, poorly constructed curtain walls for its defence.58 Similarly, the central donjon at Boudonitza had walls two metres thick and an entrance positioned 2.2 metres above the ground, whereas the outer ramparts of the fortress were barely half as thick and had flanking towers situated as much as 80 metres aparté At Salona, another medieval stronghold built on an ancient acropolis, a large tower located at the summit of the inner ward may belong to the same category, although the fact that this structure was round, an almost unique feature in central Greece, raises the possibility that it was built by the Turks rather than the Franks or the Catalans.60 What is clear, is that castles like Salona were generally dominated by central towers which were far stronger than any surrounding fortifications. They are therefore very similar to the more intact Byzantine strongholds already referred to which had Latin donjons added to them, such as Corinth, Kalamata and Neopatras. Having looked at castles which were either intact in 1204, or were constructed on the ruins of far older fortifications, it is now time to discuss

214

UNKNOWN CRUSADER CASTLES

the third category of strongholds referred to at the beginning of this chapter: those castles which the Latins built from scratch on previously unoccupied sites. Such castles were extremely rare, and it seems that none were erected outside south-western Greece. Within this area, one of the most important fortifications of this kind was Chlemoutsi (Clermont), situated on a gently sloping hill in the north-west corner of the Peloponnese. Chlemoutsi's Prankish origins can be proved beyond doubt, for the circumstances surrounding its construction are uniquely well documented. During the first two decades of crusader rule in Achaia, relations between the Latin church and the Prankish nobility deteriorated steadily, partly because of the scandalous behaviour of men like William Alemán, and partly because the church refused to provide military or financial help for the defence and conquest of Morea. These problems came to a head in the period between 122O and 1223, when Geoffrey I of Villehardouin confiscated all church land in Achaia. This amounted to a third of the entire Peloponnese, and consequently provided Geoffrey with the revenues he needed to build Chlemoutsi. These events are clearly recorded in the Chronicle of Morea and the correspondence of Honorius III,61 and are borne out by the total absence of any pre-thirteenth-century remains at the castle itself. The considerable amounts of money raised by Geoffrey from the church enabled him to build an unusually impressive fortress intended to act as a royal residence just as much as a military strongpoint. This is immediately apparent from the design of the inner citadel, a hexagonal structure superior to all other crusader fortifications in Achaia in terms of its strength, size and the quality of its masonry. Both the residential buildings inside the citadel and the defences of the large outer bailey were almost unique because of the high standard of vaulting and stone work employed in their construction. This splendour also makes Chlemoutsi unusually easy to date on architectural grounds, for many of the features added to the castle for decoration, such as the styling of the fireplaces, confirm that it does indeed date from the first half of the thirteenth century.62 No other local strongpoint can be dated as accurately as this, although it does seem that Glarentza (Clarence), located on a headland not far from Chlemoutsi, can also be attributed to the Franks in its entirety. Glarentza was built on a rocky plateau overlooking a small harbour which was once protected by a number of reefs and jetties. The town itself was defended by a single curtain wall roughly two metres thick and built from irregular, uncut stones which were only replaced by proper courses of masonry along the parapet and in other more exposed areas. Along the length of the wall there were at least two gates, which gave direct access to the town and were not protected by any additional measures such as machicolation or

MILITARY ARCHITECTURE

215

L-shaped passageways. They were originally separated from the surrounding plain by a twenty metre wide moat running the length of the landward rampart, while a small rectangular citadel located halfway along the west curtain acted as a final refuge for the town's population.63 Little is known about the history of these fortifications or the exact date of their construction, but the total absence of any recycled classical stone at the site suggests that they were all built in the thirteenth century on a previously unused site.64 They can therefore be compared with the fortress of Livadia in southern Thessaly, whose defences consisted of a central keep approached via an elaborate outer barbican and protected by a combination of sheer cliffs and poorly built curtain walls. Although classical masonry was used in the construction of this castle, it had not been found at the site but had been robbed from an ancient temple some distance away. According to some scholars, this suggests that Livadia itself was a purely western construction, built on a virgin site by Catalan or possibly Prankish settlers, and later also occupied by the Turks.65 Relying on the evidence of sites like Glarentza and Livadia, Antonie Bon has argued that features such as low-quality masonry, inaccessible location and the presence of a keep can be used to identify other Prankish castles in the principality of Achia. Bon is even more certain that such fortresses were built by westerners if their foundation was mentioned in the Chronicle of Morea, the best written source available to us on thirteenth-century Achia. A good example of such a castle is Karytaina, located in the mountainous interior of the Peloponnese and specifically described by the Chronicle as a Prankish construction.66 Built on a rocky outcrop far above the surrounding valleys, Karytaina's north, south and west sides were so steep that they required few defences and were not even protected by complete circuit walls. On the relatively gentle eastern slope, however, a large outer bailey and a fairly elaborate barbican guarded the main approach to the castle from the village below.67 Within the inner fortress itself, there were a number of buildings ranged around a lower and an upper courtyard, and two of these are worth mentioning individually. First, there was a great hall situated along the south curtain, whose doorway, windows and fireplace were all built in an unmistakably French style. Secondly, there appears to have been a tower, or keep, which stood in isolation between the two courtyards. The presence of these buildings, as well as a number of other features such as the inferiority of the masonry and the weakness of the walls (only 90 centimetres thick in the barbican), have led Bon to conclude that Karytaina was indeed built by the Franks from scratch in the thirteenth century.68 Using a similar combination of archaeological evidence and statements

216

U N K N O W N CRUSADER CASTLES

Karytaina. This castle is similar to many other mountain castles bunt or occupied by the Franks, including Mistra and Arcadia. Note the possible remains of an isolated square keep in the upper courtyard.

MILITARY ARCHITECTURE

2l/

to that effect in the Chronicle of Morea, Bon has argued that the Achaian fortresses of Androusa, Old Navarino, Mistra, Beaufort (Leutron), Old Mania and Géraki were also founded by the Franks after 12O4.69 But is this apparently simple method of identifying new Prankish castles actually as reliable as it seems? Architecturally, there are a number of problems with this approach, for it has already been noted that Androusa may well have existed in some form before 1204. Similarly, many of the features which Bon regards as typically Prankish, such as low quality masonry or the use of successive baileys to defend an isolated hilltop, have already been described at numerous castles which were quite clearly Byzantine, including Kalamata, Arcadia, Argos and Corinth. This suggests that even if Glarenza and Livadia were built from scratch by the Latins, as appears to have been the case, then their present remains are so similar to neighbouring Byzantine ruins that it is pointless trying to isolate a typical 'Prankish style' from them. Even within the last five years, archaeological exploration carried out at Vardounia, a castle situated in the Taygetos range to the south of Mistra, has only led to the conclusion that the site is of 'Prankish or Byzantine origin' because of the close similarities between all medieval fortifications in the Peloponnese.70 It has also been shown that similar uncertainties have led other scholars to challenge Bon's argument that Mistra was founded by William II of Villehardouin. The upper castle of this site was dominated by a large rectangular keep, which, as we have seen, was an extremely common feature in local Prankish military architecture. Most of the surrounding ramparts and gateways, however, show strong Byzantine influence and have therefore caused David Winfield to reject the traditional claims made by Bon and others.71 These problems suggest that Bon has been too concerned to make the archaeological evidence fit in with the Chronicle of Morea, rather than allowing the medieval ruins to speak for themselves. At Karytaina, for example, he acknowledges the presence of what appears to be much older classical masonry incorporated into the keep, but the thirteenth-century claims of the Chronicle seem to prevent him from seeing this as a possible sign that parts of the castle already existed before 1204 and were therefore Byzantine.72 Indeed, in the case of Kalavryta, another fortress in northern Achaia, Bon has attributed the defences to the Latins even though the Chronicle only listed this as an important Prankish lordship rather than claiming that it had actually been built by the newcomers.73 Another difficulty with Bon's argument arises from the Chronicle of Morea itself. This source survives in French, Greek and Aragonese versions, plus a much shorter Italian translation dating from the sixteenth century. Although these versions were all based on an older (probably French) original, they

2l8

UNKNOWN CRUSADER

CASTLES

only survive in manuscripts dating from the fourteenth century, and are not always reliable when describing events which took place more than one hundred years earlier.74 According to the Chronicle, for example, the Franks who invaded Greece after the Fourth Crusade did not capture Corinth, Argos and Nauplion until the 12408, even though the papal correspondence of Innocent III indicates that these strongpoints actually fell between 1210 and 1212.75 In other cases it is clear that when the Chronicle speaks of French settlers constructing new fortifications, it does not necessarily mean that they were building castles from scratch on previously unoccupied sites. We have already seen that the citadel of Thebes erected by Nicholas of Saint Omer around 1287 cannot possibly have been anything other than an extension or reconstruction of a much older Byzantine strongpoint.76 Finally, it is important to note that some versions of the Chronicle occasionally attributed the construction of a single fortress to more than one man. Karytaina, for example, was said to have been built by both Rugues de Bruyere, the first lord of the castle during the early years of Prankish rule, and by his later successor Geoffrey around the time of the fall of Monemvasia (1249). Similarly, Gerald was attributed to two entirely different generations of the same family.77 These anomalies suggest that we should not always accept the Chronicle of Morea at face value when attempting to date Achaian fortifications. In a sense, we have come full circle and have returned to the architectural and historical dilemmas outlined at the beginning of this chapter. Medieval remains in Greece are often so similar in appearance that they can be attributed to the Greeks, as Winfield has done, or to the Franks, as Bon preferred, with equal validity. This situation probably sterns from the fact that both Latins and Greeks used the same local materials and craftsmen to build their fortifications. This certainly appears to have been done on Cyprus, where the castle of Paphos has been dated to the early Prankish period even though it contains a number of polygonal towers built in a Byzantine rather than a Latin style.78 Returning to Greece, it is clear that western settlers sometimes also built or repaired churches 'using Greek craftsmen working in a Greek style'.79 In reality, therefore, it may not actually be possible to identify a 'pure' thirteenth-century Prankish castle simply by looking at its masonry and location, and then turning to the Chronicle of Morea for confirmation. One element of these castles which can probably be attributed to the Franks with far more certainty is the keep. In other crusader states scholars have long argued that square or rectangular stone towers of this kind represented the earliest and most common structures erected by the first Latin settlers. Speaking of the kingdom of Jerusalem, historians have

MILITARY ARCHITECTURE

21p

noted that 'the conquest was followed by the construction of large numbers of keep-and-bailey castles',80 and that strongholds built by western lords 'almost always took the form of stone towers'.81 We have seen that in the Holy Land this tradition of constructing towers continued into the thirteenth century, both at small sites such as Tuklah and at much larger castles such as Montfort. There are also examples from other areas of Latins building such structures inside much older ruined or intact Byzantine strongholds. Following Richard I's conquest of Cyprus, for example, it has been argued that the first Lusignan rulers of the island built a series of new towers where small Greek castles had once stood.82 In Cuidan Armenia a Prankish keep was also added to the Roman and Byzantine acropolis of Anavarza not long after the First Crusade.83 In Greece itself we have already seen that the Franks built numerous similar towers at the intact Byzantine castles of Athens, Corinth, Kalamata, Neopatras and Arcadia, as well as the ruined fortresses of Akova, Boudonitza and Salona. Perhaps, therefore, when the Chronicle of Morea spoke of Franks bunding new castles it was actually referring to keep towers being added to much older Byzantine or ancient defences. This theory would help to explain why supposedly 'pure' crusader sites like Karytaina contain recycled ancient stone, and would reconcile Bon's argument that many Achaian strongholds are Prankish with Winfield's conclusion that they already existed before 1204. In other words, it may be that Androusa, Mistra, Karytaina, Beaufort (Leutron), Old Navarino and Gerald were not built from scratch by the Franks, as Bon and others have assumed, but were older Greek sites which were remodelled and had donjons added to them. This suggestion may also explain the claims made by the Chronicle of Morea that Karytaina and Geraki were built by more than one Prankish lord, for these statements no longer seem illogical if they are interpreted as referring to continuing repairs and alterations rather than the completion of entire new fortresses. Hence at Karytaina, for example, some features which can probably be attributed to the Franks with more certainty, such as the gothic hall or the central keep, may have been completed by different people at different times in the course of the thirteenth century. The popularity of stone towers is also confirmed by the fact that such structures made up the vast majority of smaller fortifications built by Latin settlers. Strongholds of this kind were either left to stand in isolation or were surrounded by an outer curtain wall. More than twenty towers of the isolated variety have been identified in mainland Greece, and particularly in the former duchy of Athens. Most of these structures are extremely ruinous and are fast disappearing, but the tower of Markopoulo, located about twenty-five kilometres south east of Athens, still stands to its original

220

U N K N O W N CRUSADER CASTLES

Haliartos: a typical Latin tower of the eastern Mediterranean.

crenellated height of 18-20 metres.84 This tower's external measurements (5.4 metres by 8.2 metres) are slightly smaller than those of three similar buildings at Haliartos (Moulki), Dadi (Amphikleia) and Thurion, along the main route between Athens and Lamia, although all four structures were built using the familiar combination of recycled masonry and smaller, uncut stones.85 At Haliartos (Moulki), in particular, it is still possible to see four arrow slits in each wall of the tower, divided equally between the first and second floors.86 The original entrance was also located well above ground level; an arrangement which was copied at Thurion and Dadi, but not at Markopoulo, where there was direct access to the ground floor.87 Most of the towers in central Greece, including those mentioned above, were situated in low-lying areas. Others located on more isolated outcrops or hilltops occasionally had curtain walls around them, making them less accessible to the enemy. At Aetos in Messenia, for example, there are

MILITARY ARCHITECTURE

221

traces of a square keep at the summit, surrounded by an outer curtain wall enclosing an area measuring c. 45 metres by 90 metres. To the north and east in particular, a further wall also protected the main approaches to the keep, but all these outer defences were weak and very badly constructed.88 A more robust example of such a fortification was the acropolis at Athens, whose medieval tower shared certain similarities with Markopoulo, including some kind of ground floor entrance. This raises the possibility that the towers at Athens and Markopoulo were built at roughly the same time, and possibly even by the same architect.89 If this was the case, it may help us to date other medieval towers in central Greece, for it is most commonly claimed that the tower at Athens was erected by the Acciajuoli lords of the city during the late fourteenth or early fifteenth centuries.90 This may be a dangerous generalisation, however, because the first floor entrances at Moulki, Thurion and Dadi can be compared with a similar doorway in the central keep at Boudonitza, which appears to be a Catalan rather than an Acciajuoli construction.91 To complicate matters further, the tower at Athens had a first floor entrance as well as a ground floor one, and it was framed with reused classical blocks just like the inner gate at Boudonitza.92 A similar technique was used in the construction of a postern at Nauplia, normally attributed to the Franks.93 Under these circumstances it is hardly surprising that both the thirteenth-century Prankish rulers of Athens and their fourteenth-century Catalan successors have also been credited with building the tower on the acropolis.94 These problems illustrate the many difficulties involved in trying to date the smaller fortifications of medieval Greece. Indeed, the constant recycling of masonry and the uniformity of local building techniques has meant that even the Macedonian tower of Mara Brankowicz, the only such site to be properly excavated, cannot be dated more precisely than the medieval period in general.95 Many medieval towers which are equally difficult to date have survived on various Greek islands, including Euboea, where more than fifty such strongholds have been identified. A particularly well-preserved example is Politika, a square three-storey tower which still stands to its original height of fifteen metres and has a perfectly preserved doorway at first floor level.96 It is impossible to calculate the exact age of this building, although the traditional view that Euboean towers were constructed by the Venetian authorities according to a standard design has been challenged in recent years, for it is likelier that such structures were erected by individual Lombard settlers and 'were probably well-established by the fourteenth century'. On Chios the Genoese also built towers well into the sixteenth century, some of which were of a more sophisticated round design and

222

U N K N O W N CRUSADER CASTLES

strengthened by a talus.97 During the earliest years of the thirteenth century they relied on similar defences on Crete in an attempt to hold the island against the Venetians. Many of these fortifications were erected by the Genoese, but it seems that others were merely repaired, suggesting that older Byzantine sites were often reoccupied on islands just as much as they were on the mainland.98 A good example of a Greek stronghold being taken over by the Latins in this way dates from the invasion of Naxos by Marco Sañudo in 1207. In order to subjugate this island Sañudo first had to capture the remote stronghold of Apalire, situated on a steep inland mountain whose slopes were defended by two, and in places even three, successive curtain walls dating from antiquity. The fact that it took the Venetians more than five weeks to take the castle suggests that Apalire was an ancient acropolis whose fortifications had been maintained and improved by the islanders over the centuries. After Sañudo had conquered the rest of Naxos, he quickly made sure that this important stronghold was repaired and garrisoned by his own men." The only new fortification constructed by Sañudo appears to have been yet another tower, situated in the coastal town of Naxos and surrounded by its own curtain wall. Although the exact appearance of this structure remains unclear, it may well have resembled the towers already referred to at Markopoulo and Moulki, as well as those erected by other Italians on Euboea.100 Having looked at a variety of fortifications built, occupied or altered by Latin settlers in Greece, it is possible to draw a few general conclusions. First, it is clear that virtually all medieval strongpoints in Greece, regardless of who built them, were often poorly constructed, using a combination of recycled older materials and uncut stones, rubble or mortar. Apart from Chlemoutsi, and perhaps also the gothic hall at Karytaina, Latin fortifications were built by relatively unskilled local craftsmen rather than well-trained masons working in a more west European style. This explains why the same type of stone work appears again and again in strongholds built for different lords over many decades or even centuries. Meanwhile the historical sources tend to confirm that the vast majority of fortresses held by western newcomers had already existed in some form in 1204, either as intact Byzantine castles or as much older classical ruins. It has also been suggested that the Chronicle of Morea, which has often been accepted without question as a means of identifying new crusader castles, was only referring to repairs and additions when it stated that many Achaian strongholds were built by the Franks in the course of the thirteenth century. These factors indicate that Bon, along with several other scholars, may have overestimated the number of sites which were constructed from scratch by Latin settlers. In reality this group may have been limited to Glarentza, Chlemoutsi and possibly

MILITARY ARCHITECTURE

223

Livadia. The almost total reliance of the Latins on local workmen and existing fortifications also makes it virtually meaningless to try to identify a distinct Prankish style of military architecture. It has been suggested that the only exception to this general observation was the keep or isolated tower, which was so popular with westerners both on the mainland and on the Aegean islands that it can probably be attributed to Latin newcomers with more confidence than certain types of masonry or particular stretches of curtain wall. A number of solutions can be put forward to explain the unimpressive appearance of most medieval fortifications in Greece. One possibility is that poverty prevented the completion of more elaborate defences; a point which can be illustrated by taking a closer look at the history of some Achaian fortresses. Akova, Géraki, Karytaina, Patras and Kalavryta, for example, were some of the most important lordships in Achaia, containing as many as twenty-four fiefs each (Patras and Akova),101 and in certain cases possessing special rights of jurisdiction.102 Most of these castles were held by the same Prankish families until the late thirteenth or early fourteenth centuries, when they were either lost to the Greeks or returned to the royal domain because their owners had begun to die out.103 Others, such as the archbishop's fortress at Patras, were sold to or sought the protection of Venice, which increasingly became the only western power strong enough to defend Latin territories against the Greeks and the Turks.104 Yet even before this decline in Latin power, which did not set in properly until the last third of the thirteenth century, many Prankish castle owners do not seem to have had the resources to improve the fortifications which they occupied. We have already seen that at Karytaina some of the curtain walls were less than one metre thick and so badly constructed that they have long since collapsed, even though this lordship owed the service of twenty-two knights and was therefore the third largest barony in all of Achaia.1Q5 The same contradictions can be found in central and northern Greece. For many years during both the Prankish and the Catalan periods, the castle of Boudonitza retained its virtual independence as a frontier lordship on the northern outskirts of the duchy of Athens. It seems that it successfully withstood the Epirote conquerors of Thessaly in the I22os,106 resisted the aggressive Catalan conquerors of Athens,107 and even held up the Turks for a while in the fifteenth century, yet its defences were relatively simple in design, and so little money seems to have been spent on them that good or even mediocre masonry was only used in the most vulnerable parts of the castle.108 These observations also apply to other baronial castles in central Greece, such as Salona, which became an important Prankish lordship between 1205 and 1311.109

224

UNKNOWN CRUSADER

CASTLES

Even more surprising is the fact that many royal strongholds were just as badly constructed. Androusa, for example, belonged to the princes of Achaia throughout the medieval period until it was finally lost to the Greeks in the 1420S, yet its walls were built using a slapdash mixture of broken pottery, bricks and uncut stones.110 The castle of Livadia in the duchy of Athens first belonged to the de la Roche dukes in the thirteenth century, then rose further in importance under the Catalans, who made it the chief residence and administrative headquarters of the vicar-general, the official representative of the Aragonese dukes. As a result Livadia outshone Thebes, and often rivalled both Athens and Neopatras, but one would hardly think so by looking at the actual remains of the fortress.111 The most striking example of this irony must surely be that of Constantinople itself, whose rulers bore the imperial title yet lacked the money to repair the walls of their own capital.112 Needless to say there were some exceptions, most notably Chlemoutsi, which had been constructed with the help of a fortuitous windfall, and the luxurious thirteenth-century castle of Saint Omer in Thebes.113 But, on the whole, it seems that many medieval fortresses in Prankish Greece owed their appearance to the poverty of their occupants. In the case of Baldwin of Flanders and his successors, this can be explained in terms of the sheer number of castles and amount of territory they were expected to govern. As far as the nobility was concerned, it reflected the relatively humble background of the settlers who came to Greece in the first place. William Alemán, lord of Patras, used stone robbed from church properties to repair his castle out of necessity as much as ruthlessness, even though on paper his lordship was one of the most important in Greece, owing the service of twenty-four knights. Even more significantly, William's ancestry in Europe remains a mystery, and it seems that he was only one of many crusaders who came to the east with little to lose and much to gain.U4 Architecturally speaking, lack of adequate resources could also explain why so many of William's fellow newcomers chose to add donjons or free-standing towers to the former Byzantine castles which they took over. Structures of this kind could be built relatively quickly and cheaply, whilst their compact shape meant that they could be garrisoned by very few men even if older outer defences were left in ruins or had to be abandoned in the face of a numerically superior enemy. This situation can be contrasted with that of the other crusader states, and in particular the Holy Land itself. Here the majority of larger castles, and certainly those which were in royal hands or were held by the greatest landholders, were constructed on a far grander scale by skilled craftsmen using well-dressed stone. This difference in quality reflected the contrast in wealth between men like William Alemán and powerful European

MILITARY ARCHITECTURE

225

crusaders such as Louis IX, who repaired the town walls of Caesarea, Jaffa, Sidon, Haifa and Acre. Most of the other great thirteenth-century fortifications in the Holy Land, such as Crac des Chevaliers and 'Atlit, were built by the Military Orders relying on large donations and land revenues collected in western Europe.115 Another obvious reason why castles like Crac des Chevaliers were never built in medieval Greece was that they were not actually needed. To some extent the local terrain helps to account for this. The fortress of Karytaina, for example, occupied a very steep hill top more than 150 metres above the surrounding valleys. It must have been extremely difficult for any attacker who was under fire to reach it safely, let alone deploy siege engines against its ramparts. The Prankish lords of the fortress may well have considered walls which were only ninety centimetres thick to be perfectly adequate, us Karytaina was only one of countless other castles in Thessaly, Boeotia and the Peloponnese situated in mountainous areas, where massively thick walls or perfectly dressed stones would often have been a waste of time and money. We have already seen that mountain strongholds in Cyprus or Cilician Armenia often relied on sheer cliffs rather than sophisticated man-made defences to protect them against attackers.117 But this does not explain why more exposed strongpoints were sometimes just as poorly defended. Notably, the crusader town of Glarentza was only protected by a single ditch and curtain wall even though it was situated on a relatively flat and open coastal plain.118 This was a far cry from the elaborate double ramparts, overlapping towers and complex gateways which protected coastal cities like Acre in the Holy Land. The obvious implication is that warfare in medieval Greece, and in particular siege warfare, had not yet reached the same level of sophistication as in the Holy Land. There late thirteenth-century Mamluk armies regularly contained up to 12,000 (or more) men, whilst major crusader castles needed to be capable of withstanding an array of Muslim sappers, catapults and other siege weapons. By contrast, most of the warfare which took place in the Peloponnese, and indeed the Aegean area in general, probably involved hundreds rather than thousands of troops, and there was often a far greater risk of attacks by raiders or pirates than by vast invasion forces fully equipped with siege engines. Lack of adequate troops and siege engineers may also explain the sporadic and indecisive nature of the fighting both in northern Greece and in the Peloponnese.119 On at least one occasion the Greeks of Mistra failed to conquer neighbouring Prankish lands because they had not brought enough siege weapons with them.120 Some castles in this region, including Karytaina, were also captured through bribery or stealth rather than a direct assault; tactics which the Greeks may have fallen back on because they were

226

U N K N O W N CRUSADER CASTLES

not able to undertake difficult sieges.121 The political, geographical and military situation in medieval Greece therefore differed considerably from that of the Holy Land. This inevitably had an effect on the appearance of local fortifications.

20

Fortifications, Naval Warfare and the Crusades As Latin conquerors spread across the Aegean in the wake of the Fourth Crusade, their most urgent task was to construct or occupy strongholds near the coast. This is hardly surprising considering their great naval strength, which in turn reflected their almost total reliance on the West for military assistance, particularly during the very earliest years of the conquest. According to Villehardouin, setbacks such as the loss of over 200 war horses during fighting in 1203 or the departure of 7000 Latins in 1205 threatened the very existence of the Latin empire, for replacements could only be obtained from western Europe in a process which often took many months.1 This explains why Cyzicus, Nicomedia and other places on or near the coast were the first sites to be conquered by the crusaders in Asia Minor. The Franks hastily fortified Modon in 1205 so that they could leave 'their baggage and their servants there' and, if necessary, receive supplies and reinforcements from Constantinople or the West quickly and in relative safety.2 Once bridgeheads of this kind had been established, they acted as starting points for further raids or conquests inland. At Modon, the safety provided by its new defences enabled the Prankish knights to leave behind the slow and unarmed sections of their army and push on into the interior unhindered. In Asia Minor, the Franks used the fortified peninsula of Cyzicus in the same way, 'and from there they began to ravage the lands of Lascaris, and took much booty and many head of cattle, and took the booty and the cattle to their island' (Cyzicus).3 Similarly, during the autumn of 1205 the emperor Henry was able to launch a campaign against the Bulgar leader Kalojan in the vicinity of Demotika because he received a constant supply of food and reinforcements from the port of Rodosto.4 Three years later, when he had to march west along the Thracian coast in order to deal with a rebellion by Lombard settlers in Thessaly, Henry relied on a fleet sailing parallel to him to feed his army during the bitterly cold winter. A surprise attack on his fleet by pirates, combined with the fact that the Lombards still held Christopoli (the most important harbour along the route), explains why this expedition ran into so much trouble and almost had to be abandoned through starvation outside the walls of Thessaloniki.5 Well-defended harbours and anchorages were therefore essential to

228

U N K N O W N CRUSADER CASTLES

any aggressive campaigns conducted by the Franks, particularly during the early years of Latin rule. Once the crusader states around the Aegean had been established, such places still acted as important supply points whenever the newcomers were threatened in any way. This can be illustrated by taking a closer look at the thirteenth-century history of Glarentza. After the loss of Constantinople in 1261, William II of Villehardouin made a treaty with Charles of Anjou, ruler of Naples (1266-85), whereby the latter would replace the Latin emperor as suzerain of Achaia in exchange for military aid against the Greeks.6 Glarentza therefore became the obvious destination for ships sent across the Ionian Sea from Brindisi, and it may not be too fanciful to suggest that the city's Prankish defences were constructed at this time to protect a vital new life line.7 In 1270, for example, Charles sent a fleet to Achaia to counter the growing threat posed by Michael VIII Palaiologos, and during the next twenty years countless troops and supplies from Italy flowed through Glarentza in response to the increasingly pugnacious activities of the Greeks at Mistra.8 According to the Chronicle of Morea, it was also during this period that Nicholas II, co-ruler of Thebes, built (or, as has been suggested, rebuilt) the castle of Old Navarino (Port de Jone) just north of Modon.9 Bearing in mind that in 1287 Nicholas completed another castle at Thebes which was largely intended as a luxurious residence, it may be that Old Navarino was never designed to fulfil any specific military function.10 However, it is interesting to note that Nicholas chose to construct his fortress on an inaccessible outcrop overlooking a large, natural harbour. Presumably this was done so that Old Navarino could be reinforced by sea quickly and conveniently, particularly if it ever came under attack from the Greeks, or indeed became the starting point for an incursion against Mistra. The castle of Old Navarino can also be compared with that of Naxos, which Marco Sañudo is said to have built on a hill top situated very near the coast. Here Sañudo could easily receive help from other Venetian states and colonies, or even Venice itself, making it an ideal centre from which to rule both Naxos and the entire archipelago. In comparison, the Greeks were only able to establish, and eventually expand, the despotate of Mistra because of the invincibility of Monemvasia, whose natural strength made it almost impossible to capture but relatively easy to supply by sea.11 Coastal strongholds were needed to establish and maintain Latin rule over the interior. If and when the political situation deteriorated, access to the sea was just as essential, either to get help or to escape. The Latin empire narrowly avoided almost certain destruction in 1235 and 1236 because Venetian ships were able to break through the Greek and Bulgar blockades established around Constantinople.12 Similarly, when the Prankish garrison

F O R T I F I C A T I O N S , NAVAL W A R F A R E AND THE CRUSADES

229

of Cyzicus came under siege in 1207, the emperor Henry organised a relieving fleet which sailed across the Sea of Marmara, forcing Theodore Lascaris's land and naval forces to retreat. If Cyzicus had been located inland, it is extremely unlikely that Henry would have reached it in time or have had enough troops at his disposal to face Lascaris in a land battle.13 These factors explain why heavily fortified coastal strongholds, such as Modon and Coron in the Peloponnese, were invariably the last western outposts to fall, sometimes outliving neighbouring inland castles by many decades.14 When it was no longer possible to defend even the most powerful fortifications, Latin garrisons could still use the sea as their final means of escape. Many westerners, including the Latin emperor Baldwin II, managed to flee from Constantinople in 1261 in much the same way that the defenders of Acre, Sidon and other coastal outposts in the Holy Land were to do exactly thirty years later.15 Many decades earlier Prankish troops had also managed to make an orderly retreat from Kibotos in the vicinity of Nicomedia by sailing away across the Sea of Marmara to Constantinople (1207). Interestingly, Kitobos had already withstood a Nicaean attack successfully, but it was still decided to evacuate the castle because the Franks knew its garrison was too weak to hold out for much longer. Kitobos provides us with another example of how inadequate troops and resources rather than a lack of military skill prevented the Franks from holding on to their territories within the Latin empire.16 The fate of Kibotos merely confirms that the Latins could only hope to sustain their fragile position in the former Byzantine empire by protecting their seaborne links with each other and the West. In practice this meant that they were heavily dependent upon the Venetians. It was Venice which saved Constantinople in 1235 and 1236, and it was their desperate need for Venetian assistance that forced the Franks to give up their claims to Modon and Coron in 1209, even though these cities had originally been captured by French rather than Italian crusaders.17 Other coastal fortresses would never have surrendered in the first place if it had not been for the Venetians. Nauplia and Monemvasia, for example, were both captured because Venetian ships enabled the Latins to blockade these strongholds by sea as well as by land. In order to obtain Venetian support for his campaign against Monemvasia, William II of Villehardouin confirmed the republic's possession of Coron and granted it certain privileges. In return, Venice provided the Franks with four galleys for the campaign, and undertook to maintain a further two galleys for the permanent defence of Achaia18 This agreement illustrates how Venice dominated all naval activities in Prankish Greece, in the same way that it had dominated the Fourth Crusade itself. This can be contrasted with the situation in Cyprus and the Holy Land, where the

230

UNKNOWN CRUSADER CASTLES

crusading movement remained more international in character, and the naval strength of Venice, Genoa and Pisa was more equally divided. To some extent Venetian naval power also contributed to the rapid fragmentation of crusader states in Greece during the fourteenth century. The archbishops of Patras, for example, managed to retain the status which they had gained as independent lords answerable only to the pope by allying themselves to the Venetians. In 1366 it was a Venetian commander who led the defence of Patras against the forces of Hugh of Lusignan, a claimant to the principality by virtue of his mother's marriage to Robert of Taranto, ruler of Achaia from 1346 to 1364.19 Hugh of Lusignan's failure to take Patras effectively ended his hopes of controlling the Morea, but Venice still continued to send reinforcements to the city during the 13708 to prevent Hugh's rival and Robert's successor, Philip II of Taranto (1364-73), from launching a similar attack to that of 1366. The Venetians were able to help Patras in the same way that they had once assisted Constantinople, only now they were fighting fellow west Europeans rather than the Greeks.20 Like the Prankish troops at Cyzicus in the early thirteenth century, rival claimants and pretenders to the principality of Achaia used coastal strongpoints as bases from which to make territorial conquests inland. Perhaps the best example of such a campaign dates from 1315. In that year Ferdinand of Majorca, who was married to the granddaughter of William II of Villehardouin, invaded the Morea because he felt he had a legitimate claim to the principality. He therefore landed near Glarentza, occupied the city, and used it as a bridgehead for further conquests inland. The following year Ferdinand's Catalan reinforcements arrived at the northern port of Vostitza, presumably hoping to march south and meet up with Ferdinand so that all of Achaia and Elis would eventually be overrun. By this point, however, Ferdinand had already been killed in a pitched battle with the Angevins, but his failed expedition nevertheless illustrates the important role played by fortified harbours in all invasion attempts, regardless of whether they were undertaken by crusaders or by rebellious usurpers.21 Returning to the earliest days of Prankish rule in the Peloponnese, one last example can be cited to illustrate this point. During the campaign of 1204-5, the crusaders advised their commander, William of Champlitte, that 'you should try to take the fortresses which are by the sea; for, if you have the ports and the entry points into the country, then you can get reinforcements of troops and supplies when you need them'. In this way, they argued, 'you can easily have the rest of the countryside'.22 By the fourteenth century coastal strongholds were not simply expected to maintain seaborne links with the West, thereby underpinning Latin rule in the East. After a period of relative stability following Venetian expansion

F O R T I F I C A T I O N S , NAVAL W A R F A R E AND THE C R U S A D E S

231

in the Aegean, piracy was making a comeback, particularly at the instigation of the Turks and the Catalans. The wholesale devastation which these groups inflicted on the Aegean caused entire islands and coastal districts to be abandoned for years at a time.23 That the Palaiologi emperors of Constantinople periodically both employed and clashed with entire fleets of pirate vessels gives some idea of the scale of the problem.214 At the time of the Fourth Crusade, it seems that most islanders who had remained in the eastern Mediterranean had retreated inland, to places such as the fortress of Apalire on Naxos, captured by Marco Sañudo in I2o/.25 Many people clearly continued to rely on this tactic for centuries, for as late as the 13903, the travelling pilgrim Niccolo da Martoni described how the population of Thermia (Fermia) in the Cyclades lived in an isolated settlement in the mountains, to which he himself had to flee for a few days because of the threat of pirates.26 According to one historian, the Sañudo dukes of the Archipelago also built the castle of Apano-Castro, located in the interior of Naxos, in direct response to the Turkish raiders operating in the Aegean at the beginning of the 13905. This stronghold enabled both Greeks and Latins living on Naxos to abandon the coast temporarily when the Turkish threat was at its greatest.27 This type of incursion also helps to explain the presence of so many isolated towers on the island of Euboea, for even if these structures were far too small to withstand major campaigns of siege and conquest, they could shelter local people from raiders or pirates hastily passing through in search of moveable booty.28 The Latins, and in particular the Venetians, relied on the sea far too much simply to turn their backs on it. Whereas the native inhabitants of most islands only regarded the Aegean as a source of food, and could therefore afford to resettle permanently in the interior, the Latins needed fortified bases on the coast to maintain their military and political power, and to protect their lucrative international trade routes. One such settlement was Naxos, where it seems that the Sañudo dukes built a new harbour as well as a castle.29 Another stronghold probably situated near the coast was that of Siros, a small island to the north west of Naxos. In 1286 its lord, William Sañudo, purchased a branded ass stolen by pirates during a raid on neighbouring Tinos and Mikonos. As a result, Bartolomeo Ghisi, ruler of these latter islands, attacked William and besieged him in the castle of Siros using trebuchets and other siege engines. This assault continued until William was relieved by a fleet sent by Charles of Anjou's admiral and Duke Marco II of the Archipelago, forcing Bartolomeo Ghisi to withdraw and come to terms.30 Although this incident only involved piracy indirectly, it is indicative of the kind of localised raiding and counter-raiding which affected many Aegean islands, illustrating the widespread need for

232

U N K N O W N C R U S A D E R CASTLES

strongholds which were situated near the sea. Similarly, the many isolated towers constructed in harbours could protect people and revenues against opportunistic raiders whilst at the same time helping to guide approaching ships toward their destination. Towers of this kind have survived on Euboea at Aliberi-Matsoukela and Rovies, and the latter may even date back to the time of William of Villehardouin's campaigns on the island during his civil war with Duke Guy I of Athens and the Venetians (1255-58)31 Many castles on the mainland also acted as refuge points, for Greek coastal areas were just as likely to be attacked by pirates as the Aegean islands. At Monemvasia, for example, the harbour was situated on a vulnerable strip of land facing the sea but was protected by the impregnable citadel above it. Even though a devastating Catalan raid during the 12908 resulted in the destruction of the town, its inhabitants escaped unhurt because 'they climbed up to the castle on the great rock which stands there'. In this way an important anchorage and its civilian population were prevented from falling into the hands of the dreaded Catalans.32 Another mainland settlement which was often targeted by pirates was Corinth. From 1311 onwards this city and its immediate surroundings must have been particularly exposed to raids, for both the Catalans of Athens and the Greeks of Mistra could easily reach it by land or sea. As a result, living conditions in this area got so bad that by the middle of the fourteenth century large parts of the countryside had been completely abandoned. The Angevin rulers of Achaia responded by granting the castellany of Corinth to Niccolo Acciajuoli, a powerful Florentine lord who was instructed to repair the citadel's defences and was granted special privileges to help him meet the huge cost of maintaining several smaller castles in the neighbourhood (1358). These included that of St George, situated near the frontier with Mistra and 'valiantly' defended by its Latin garrison during the I35os.33 Several years earlier Acciajuoli had also been granted Messenian lands in the southern Peloponnese, equally devastated by years of Greek, Turkish and Catalan pillaging. To halt this process, Acciajouli had used the same tactic as at Corinth, erecting a new castle in the vicinity of Kalamata which could protect the area against external raiders.34 Elsewhere fortifications situated in coastal regions played a more active role in the fight against piracy, for they provided safe anchorages from which to organise counter-strikes against seaborne raiders. By the mid fourteenth century the most formidable intruders of this kind were the Turks. In order to deal with this threat the papacy, the Venetians, the Cypriots and the Hospitallers of Rhodes formed a powerful naval alliance known as the Holy League. This was not simply regarded as a military force but as a continuation of the crusading movement once land-based offensives against Jerusalem

F O R T I F I C A T I O N S , NAVAL W A R F A R E AND THE CRUSADES

233

itself had become unrealistic.35 The important contribution coastal fortifications made to the continuing struggle against Islam is shown by the League's first and greatest victory: the capture of Smyrna, on the coast of Asia Minor, in October 1344. This was achieved by a fleet of Cypriot vessels sent from the fortified harbour of Famagusta, Hospitaller galleys stationed at the Order's heavily defended base on Rhodes, and a large contingent from the formidable Venetian navy. These forces all gathered at the Venetian port of Negroponte on Euboea before attacking Smyrna itself.36 Although they have long since disappeared, Negroponte's medieval defences at this time were extensive and included a fortified bridge connecting Euboea with the mainland. This structure was divided by a drawbridge 'no larger than to let a galley pass through', so that Venetian ships anchored opposite the city would be protected from Turkish, Genoese or Catalan raiders.37 The security provided by sites like Negroponte, Rhodes and Famagusta enabled the Holy League to launch its successful offensive on Smyrna and to maintain pressure on the Turks throughout the eastern Mediterranean.38 Smyrna and other Christian outposts in Asia Minor, such as Cypriot-held Corycus,39 were themselves further examples of the kind of coastal strongholds already mentioned at Cyzicus, Naxos and Monemvasia. All these sites owed their very existence to a combination of good sea links and powerful fortifications. Between them, these places created a network of safe anchorages which were not only used by major expeditions, such as the crusade against Smyrna, but could also protect much smaller naval forces. During the latter half of the thirteenth century, for example, the Venetians constantly patrolled the waters around the Peloponnese with two of their own galleys to reduce the general level of piracy rather than to deal with a specific threat. These vessels no doubt operated between Venetian bases at Modon, Coron, Negroponte and in the Cyclades, never straying too far from a friendly port.40 Pirates themselves normally used coastal strongholds as bases from which to launch their raids. One group which relied on this tactic were the Catalans. After clashing with their former Byzantine employers, but before they established themselves at Athens, they withdrew to Gallipoli for a few years at the very beginning of the fourteenth century. They refortified this site by digging huge ditches and erecting new wooden stockades, making it a perfect headquarters from which to carry out naval attacks on neighbouring harbours such as Panidos and Rodosto. On one occasion so many Catalans left Gallipoli on a seaborne raiding party that the Greeks came close to recapturing the town, which was almost empty.41 But the Catalans survived this crisis, and Muntaner wrote that in the following months 'except the cities of Constantinople and Adrianople and Christopoli and Salónica, there

234

UNKNOWN CRUSADER CASTLES

was not a town or city that was not pillaged and burnt by us, nor any place, unless it was a castle in the mountains'.42 This programme of systematic looting was only made possible by the strength of the defences at Gallipoli itself, which enabled the Catalans to act without fear of reprisal from the Byzantine authorities On the other hand, Muntaner's statement indicates that the Greeks had been able to maintain the defences around their larger cities, whilst his reference to the impregnability of mountain castles explains why elevated strongholds like Monemvasia were so popular with those who lived in exposed areas.43 Many other raiders and pirates were stationed on islands. Corfu, for example, had to be cleared of Genoese pirates before the Venetians could take control of the island in the wake of the Fourth Crusade.44 Elsewhere, other Genoese forces, acting in closer cooperation with their native city used similar tactics against their Venetian rivals. During the 13405 and 13505 Genoa tried to use its new colony on Chios as a springboard for further expansion in the eastern Mediterranean, including the conquest of Euboea from Venice. This project came to an abrupt end after the Genoese failed to capture the port of Oreus along the north coast of the island in i35i.45 Almost 150 years earlier, the Genoese were equally unsuccessful in their efforts to subjugate Crete and to defend it against the Venetians, who had bought the island from Boniface of Montferrat shortly after the Fourth Crusade. It was only by capturing the heavily fortified castle and harbour of Palaiocastro, situated on the north coast of Crete, that Venice finally ousted the Genoese and prevented them from establishing a base which could have posed a serious threat to Venetian power in the Cyclades.46 The Genoese were very similar to their Venetian, Prankish and Catalan rivals in that their military activities in the Mediterranean were almost entirely centred around the construction, capture or defence of castles. Such structures were often the only means by which the various Latin powers operating in the eastern Mediterranean could hope to maintain control over the islands and coastal regions of the former Byzantine empire. Offensively, castles were an ideal way of creating bridgeheads which could be reinforced by sea and used to make further conquests inland. In addition, they protected important harbours which were needed for naval campaigns such as the crusade against Smyrna. Defensively, powerful strongholds like Monemvasia provided shelter for local people whilst preventing strategic anchorages from falling into the wrong hands, for the activities of the Catalans at Gallipoli proved that raiders could use such sites to wreak havoc over vast areas. This last point also highlights the immense importance of coastal defences as a means of protecting the sea routes between East and West, for fortified settlements such as Glarentza were needed to maintain the steady flow of

F O R T I F I C A T I O N S , NAVAL W A R F A R E AND THE CRUSADES

235

arms, troops and supplies upon which Prankish Greece depended, Coastal strongpoints were invariably the first to be captured by the Latins, and also the last to fall, not only in the Aegean but throughout the crusader states.

21

The Offensive Functions of Inland Fortifications The Latins relied on inland strongpoints to perform many of the same military functions as their neighbours nearer the sea. Crusader fortifications were often used as starting points for land-based as well as seaborne incursions into enemy territory. This was particularly true during the earliest years of the conquest, when the expansion of Prankish power in the interior often overlapped with the process of establishing coastal bridgeheads. The fall of Constantinople in 1204 provided the crusaders with an important and heavily fortified harbour, but it also enabled them to regroup and start conquering inland territories to the north and west. In the summer of 1204 Boniface of Montferrat left the new Latin capital, gradually seizing or occupying castles as he marched through Thrace. This in turn enabled him to obtain the surrender of Adrianople and Demotika, two of the most important strongholds to the west of Constantinople, opening up the route toward Christopoli, Salonika and the whole of Thessaly.1 After the Greek uprisings and Bulgar invasions of 1205—6, when virtually all of Thrace was lost, this process had to be repeated by the emperor Henry. On this latter occasion the fortified city of Tchorlu, located relatively near Constantinople, also acted as an assembly point for Prankish troops hastily collected from both sides of the Bosphorus.2 Over the next century, there were many similar examples of how the acquisition of one strongpoint could lead to the capture of others. During the resurgence of Byzantine power in the second half of the thirteenth century, the Greeks used such tactics against their Prankish enemies in the Morea. In 1271, for example, Michael VIII sent 'a great company of men-atarms from the Levant, Turks, Cumans and Greeks' to Monemvasia in order to invade the principality of Achaia.3 Although this particular campaign was eventually cancelled because of a Latin counter-attack, many similar expeditions were launched from this area, using Byzantine troops first brought to Monemvasia and then assembled further inland at Mistra. In the 13205 these attacks culminated in the fall of several Prankish strongholds to the north of Mistra, including Karytaina, Akova (Mathegriffon) and St George. According to the Chronicle of Morea, these places were captured through bribery as well as warfare, suggesting that by this period both the

THE O F F E N S I V E F U N C T I O N S OF I N L A N D F O R T I F I C A T I O N S 237

morale and the fighting strength of the Franks were beginning to wane.* A comparable case of territorial conquest being based around castles dates from the period of Catalan rule at Gallipoli. Once they had established themselves here, the Catalans were not content to launch naval attacks against neighbouring coastal settlements, but also expanded inland until they dominated the entire surrounding peninsula. This meant having to capture local Byzantine fortresses such as Maditos, situated to the south of Gallipoli. Maditos only fell after an eight month siege, when the Catalans managed to climb the walls during the garrison's afternoon siesta. Such a prolonged campaign would not have been possible without the use of Gallipoli as a base from which to bring up food, supplies and troop reinforcements, and it highlights the need for armies to have some kind of secure starting point before they could take to the field.5 These considerations also applied to the many less ambitious campaigns which were launched by Prankish forces to collect booty rather than gain territory. As part of the Latin emperor Henry's efforts to re-establish Prankish control in Thrace after the Bulgar invasion of 1205, no less than 120 western knights were stationed in the walled town of Rousion, to the south of Demotika. During the winter of 1205-06, these troops carried out a raid against hostile Greeks in neighbouring territories, killing many and capturing forty horses. The purpose of this attack was not to retake lost fortresses or land but rather to intimidate the local population and force them to accept Prankish rule.6 The following summer Henry used Adrianople as a base for a similar incursion into Bulgar territories to the north, even though the Franks clearly lacked the numbers to occupy such a large area permanently. Instead, Henry simply hoped to gain some booty, inflict economic damage on the region, and thereby deter the Bulgars from attacking the Latin empire again. These tactics are virtually identical to those used by the Hospitallers of Crac des Chevaliers and the Templars of Saphet to keep their Muslim neighbours in the Holy Land at bay.7 In 1271 Michael VIII Palaiologos hoped to launch a major attack on the Morea using troops he had assembled at Monemvasia. On hearing this news, William II of Villehardouin immediately prepared for a pre-emptive strike against the area to the north of Mistra, which was dominated by Byzantine troops and rebellious local Greeks. William's forces therefore marched to Glarentza, where they were joined by other troops sent by Charles of Anjou, before moving east to Karytaina. Along the way, other barons met them with further contingents and supplies for two months, and at Karytaina itself the lords of both this castle and nearby Akova also joined the expedition. Once everybody had arrived, William called a meeting to discuss tactics, before setting off on a raid which not only halted Michael

238

UNKNOWN CRUSADER CASTLES

VIII's own advance but also brought the Franks much booty and cattle. During the preliminary meeting William remained camped along the river below the castle of Karytaina, so that his army had a ready supply of food and water but could still find shelter quickly in case of a Greek counterattack. The success of the entire campaign relied upon the food, supplies and protection provided by Glarentza and Karytaina.8 In describing William's preparations for this expedition, the Chronicle of Morea implies that those barons who joined the prince brought troop contingents drawn from their own castle garrisons. The same source provides us with more evidence that this tactic was used by Greeks and Franks alike. In 1259, on the eve of the battle of Pelagonia, the Franks decided to confront their Nicaean opponents in open battle, thereby hoping to destroy the Greek field army in one blow. 'Those who were most experienced in warfare' argued that this would leave the Greek castles in Thrace and Macedonia completely unprotected, and that their outnumbered defenders would be forced either to flee or to surrender. Similar circumstances had of course enabled Saladin to overrun the kingdom of Jerusalem in 1187, but the crushing defeat at the battle of Pelagonia ended any Prankish hopes of emulating this success in Greece.^ Nor do the westerners seem to have learnt from this lesson, for in 1311 they lost so many troops at the battle of Halmyros that the Catalans were able to do precisely what the Franks had planned to do against the Nicaeans. At Livadia, for example, there were not enough Latins left to prevent the local Greeks from rebelling and opening the gates to the Catalans. Elsewhere the invaders simply occupied former Prankish castles without a struggle, and some even married the widows of those barons who had fallen only a few weeks earlier.10 The relatively small number of Latins settled in Greece made it risky to commit too many garrison troops to major Prankish field armies. If such forces were defeated, reinforcements were hard to come by and strongholds were extremely vulnerable to enemy counter-attacks. In 1187 and 1311 these factors led to the destruction of entire crusader states. It was more prudent to avoid confrontations in the field altogether, and to rely on castles to strengthen control over territory. In the Morea, William II of Villehardouin used this strategy against the Melings, an aggressive Slavic tribe whose homeland in the Mani peninsula was so mountainous and inhospitable that it still refused to acknowledge Prankish overlordship as late as the 12403. William realised that it would be useless to send a force of knights into such an area, so he constructed (or reoccupied) the castles of Mistra and Old Mania (Grande Maigne) at the northern and southern ends of the peninsula. From here his troops could observe the Melings, launch punitive raids against them and shelter local people from their depredations without

THE O F F E N S I V E F U N C T I O N S OF I N L A N D F O R T I F I C A T I O N S 239

having to carry out a systematic conquest of the entire region. For their part, the Melings realised that 'pressured between these two castles, it was impossible for them to resist the prince', and so they made peace with him.11 William then built a third castle to the east called Beaufort (Leutron), 'all the better to contain [the local Slavs] and place them under his rule'.12 The construction of this latter fortress suggests that Mistra and Old Maina had been successful in suppressing the Melings, and that castlebuilding enabled William to keep his Slavic subjects in check without having to occupy their homeland in its entirety. By constructing castles, William was able to make up for the fact that he lacked the troops to suppress the Melings through sheer weight of numbers. At other times, however, the Franks did manage to defeat their opponents in open battle, but still needed castles to make sure that the advantage gained from a victory was not lost. This problem most affected the Latin empire, where the emperor Henry had to keep campaigning almost continuously because 'he could not raise enough troops to defend his territories'. As a result, the Franks were often obliged to rely on castles rather than soldiers to maintain their borders.13 After victory over the Cumans in 1208, for example, Henry immediately occupied the castle of Crucemont, about thirty kilometres west of Philippopolis. He then rode south to Pamphilon, where he constructed a new castle as quickly as possible before the onset of winter. In so doing, Henry hoped to take full advantage of his victory, secure his northern border, and prevent the Cumans and Bulgars from simply reoccupying this region as soon as he had withdrawn to Constantinople.14 Similarly, during the 12603 Michael VIII Palaiologos enjoyed considerable success against the Latins because, once he had harried Prankish Greece with his land and sea forces, he 'occupied many places and built powerful castles on mountains and in very strong passes'. Michael knew that this would make it far harder for the Franks to regain this territory, as they would only be able to do so by undertaking a series of lengthy sieges.15 One way of responding to such tactics was to adopt a scorched earth policy and to demolish all strongholds likely to be occupied by an advancing enemy. In 1331, for example, the Catalans responded to news of Gautier II of Brienne's imminent invasion by demolishing the fortress of Saint Omer at Thebes 'to make sure that the duke of Athens would not take it in any way and recover the duchy by using this castle'.16 In 1207 the emperor Henry agreed to demolish both Cyzicus and the fortified church at Nicomedia in his peace treaty with Theodore Lascaris of Nicaea. By insisting on the destruction of these two castles, Theodore hoped to prevent the Franks from continuing to use them as bases for raids on the surrounding

240

U N K N O W N C R U S A D E R CASTLES

countryside.17 Although he was probably mostly interested in booty, the destruction of fortified sites in Thrace by Kalojan may have been a comparable attempt to prevent the Franks and Venetians from reoccupying the area, and to create a kind of no man's land between Constantinople and the Bulgars.18 Beyond Prankish Greece, Kalojan's actions can be compared with the systematic demolition of former crusader strongpoints undertaken by Saladin and Baybars in the Holy Land.19 That rulers sometimes preferred to destroy perfectly good fortifications in this way shows us just how difficult it was to remove an opponent from an area once he had occupied its castles. If an army had no alternative but to attack an opponent barricaded inside a major strongpoint, one way of sustaining a lengthy siege was to build temporary castles in the immediate vicinity. At Corinth, the Franks erected two structures of this kind on the eastern and south-western sides of the citadel. Some traces have survived of the south-western fort, known as Pendeskouphi, showing that it amounted to little more than an isolated keep of traditional Prankish design. It clearly dates from the time of the original Latin siege between 1205 and mo.20 An equally small fortification, built out of bricks, or possibly even just earth embankments, was used by the crusaders during the campaign against Patras in i2O5,21 whilst at Constantinople the land forces of the Fourth Crusade established their camp around 'the castle of Bohemond, which was an abbey surrounded by walls'.22 These so-called castles were temporary structures which were probably abandoned at the end of a siege. During a campaign, however, they fulfilled several important military functions. At Corinth, for example, early Prankish hopes of seizing the citadel were dashed when its Greek defenders made a daring night time raid against the crusaders stationed in the town below. This sortie inflicted 'great damage' on the Franks, and persuaded William of Champlitte to scale down his activities there and concentrate on the conquest of the Morea instead.23 Likewise, during the first Latin siege of Constantinople in 1203, the crusaders to the west of the city were prevented from gathering enough supplies, let alone attacking the ramparts, because of a series of Byzantine sorties launched from within the capital.24 One of the primary functions of fortifications like 'the castle of Bohemond' was to protect the besiegers against the besieged, and provide them with a base where food, water and other supplies could be stored in safety. The crusaders' fort at Patras may be particularly significant in this respect because it was located right at the water's edge, suggesting that it was being reinforced by sea.25 Offensively, these forts were also used to blockade besieged garrisons and prevent them from gathering supplies, or worse still, contacting relieving

T H E O F F E N S I V E F U N C T I O N S O F I N L A N D F O R T I F I C A T I O N S 24!

forces in other areas. The two strongholds erected by the Franks at Corinth were specifically intended to ensure that the besieged 'could not rush out from the walls to collect water or any other supplies to sustain them'. If they attempted to do so, the Franks could easily launch a counter-raid from either of their forts in order to force the Greeks back inside the citadel.26 At Constantinople Villehardouin wrote that 'never had so many been besieged by so few in any city', and this explains why the crusaders established their fortified camp opposite one particular stretch of the defences, rather than trying to spread out around the whole city.2? It would have been impossible for the Latins to surround vast sites such as Corinth or Constantinople entirely; but, by gathering all their forces together inside fortified camps, they made the most of the troops available to them, and could still blockade far larger castles for months or even years. Once again castles took the place of troops in the field. These factors also explain why it took the crusaders five years to capture Corinth: with so few soldiers available, an all out assault on the walls was usually out of the question; if they could not bombard a fortress into submission, the Franks simply had to wait until hunger or low morale persuaded its garrison to surrender.28 The methods used by the crusaders at the siege of Corinth also reflect their military fragility in general. Every Latin soldier, and in particular every knight, was a precious asset for the Franks, who had enough trouble raising adequate forces for their campaigns even at the best of times. Hence one of the primary functions of all castles in Prankish Greece was to protect the Latin field army, especially during a military crisis. In the spring of 1205, for example, the first Latin emperor of Constantinople, Baldwin of Flanders, was captured outside Adrianople, while his army suffered a heavy defeat at the hands of the Greeks and the Bulgars. This disaster forced the crusaders to flee south, past Pamphilon and toward the coast. It was not until they reached the fortified city of Rodosto, situated along the Sea of Marmara, that they finally found refuge from the pursuing forces of Kalojan.29 The walls of Rodosto effectively saved the Prankish and Venetian army from total annihilation, which may explain why the emperor Henry chose to fortify Pamphilon three years later, by which date Adrianople had returned to Prankish control. Pamphilon must have acted as a useful stopping off point for Latin troops travelling from Rodosto or Constantinople toward the northern frontier of the empire. Alternatively, Henry knew that, if Adrianople were ever recaptured by his opponents, the new defences at Pamphilon could protect westerners retreating south, who would otherwise have to march all the way to Rodosto without any hope of shelter along the route.30 William II of Villehardouin's campaign against the Greeks, undertaken

242

UNKNOWN CRUSADER CASTLES

in 1271, was organised around the strongholds of Glarentza and Karytaina. This latter castle was chosen as the starting point for the raid on Mistra itself because it lay along the main route between the west coast and the mountainous interior. To the south east, the narrow valley below Karytaina also continued toward Mistra, Monemvasia and the east coast. Karytaina, situated roughly halfway between Glarentza and Monemvasia, guarded this strategic line of communication through the heart of the Peloponnese and could protect any Prankish soldiers operating in the area. This was particularly important in a region where the terrain made it easy for Byzantine troops or rebellious Greeks to ambush heavily armed western knights. In 1302 Greeks attacking Prankish lands from the direction of Mistra advanced through the valley below Karytaina because this route was 'easier and safer than all the others'^ Whenever it was possible, therefore, Latin field armies tried to stay within easy reach of friendly castles. The further they strayed into enemy territory, the greater the risk of heavy losses, particularly if an opponent defeated them in open battle. This explains why the Prankish campaign against the Bulgars and Cumans undertaken by the emperor Henry in 1208 was so fraught with danger. On the eve of his famous victory, Henry advised his knights to put their faith in their horses, shields, lances and above all God, for 'y°u are assembled here in alien territory, and do not have a castle or refuge where you can hope to find shelter'.32 Henry knew that if his followers were defeated there would be no means of escape for them, and that he himself would probably suffer the same fate as his brother Baldwin, who was never seen again after the Bulgars captured him outside the walls of Adrianople. A comparable incident took place in 1224, when Latin troops sent from Constantinople to recapture the Thracian city of Serres from Epirote Greeks had to be recalled because other Franks in Asia Minor had been severely defeated by a Nicene army at the battle of Poimanenon. The Serres expedition sustained many losses as it hastily retreated all the way back to Constantinople, again highlighting the risks involved in sending Prankish armies into hostile territories far from any reliable refuge points.33 Inland castles were therefore needed to protect troops in the field and the roads they travelled on in the same way that fortified harbours sheltered Latin warships and kept important sea routes open.

22

The Defensive Functions of Inland Fortifications The unstable situation around the Aegean during the thirteenth and fourteenth centuries ensured that Latin strongholds were called upon to fulfil a variety of defensive roles. Their most important function was probably as heavily defended strongpoints to stop major invasion forces from making permanent conquests. When in 1235 the Bulgar leader John Asen II (1218-41) overran all of Thrace, allied himself with John III Vatatzes of Nicaea (122254), and attacked Constantinople itself with a huge besieging army, the city's vast ramparts held up the assault long enough for 160 Prankish knights to organise a sortie and drive off their Greek and Bulgar opponents. At the same time the Greek fleet blockading Constantinople from the east was decisively defeated by the Venetians,1 who returned to save the capital during a second siege the following year, when the Latins were also assisted by warships sent from Achaia.2 During the next twenty-five years the Franks rarely enjoyed any authority beyond the immediate vicinity of Constantinople; but, as long as they were protected by its walls, their enemies proved incapable of wiping out the Latin empire completely. Beyond Constantinople, the Franks also guarded their northern frontier by stationing large numbers of knights inside carefully selected fortified settlements rather than trying to spread them out along the entire border. In 1205 these strongpoints included Bizöe and Rousion, garrisoned by 120 and 140 knights respectively, as well as Selymbria, a port along the Sea of Marmara which was defended by a further fifty knights. As such troops were presumably accompanied by socially inferior infantry or cavalry forces which the chroniclers did not bother to mention, they must have formed unusually large garrisons for the period. In addition, the Venetians held Arcadiopolis, situated roughly halfway between Rousion and Bizöe, so that these sites formed a kind of arc along the northern fringes of the empire. By concentrating his forces in this way, the Latin emperor Henry knew that they would stand a better chance of survival and of being able to defend their respective strongholds than if they had been spread out in a ratio of five or ten knights per city.3 When the Bulgar invasion which Henry had been expecting came only a few months later, Kalojan's army was so powerful that both Rousion and

244

UNKNOWN CRUSADER CASTLES

Arcadiopolis had to be abandoned. Indeed, as Kalojan swept south toward Constantinople, destroying everything in his path, things began to look ominous for the Franks. But despite the appalling damage which the invaders were inflicting on the countryside, the Prankish defenders of Selymbria, Bizöe and Constantinople stayed put, refusing either to retreat or face Kalojan in the field. Meanwhile, the Bulgars and their nomadic Cuman allies were causing so much destruction that cracks began to appear in their alliance with the local Greeks, who refused to let Kalujan's troops inside Adrianople. This in turn gave Henry the chance he had been waiting for, and he finally gained the nerve to leave Constantinople with an army supposedly containing 400 knights and to march north toward Bizöe. This latter stronghold now became the starting point for a Prankish revival. As the Bulgar advance faltered before Adrianople, Henry gradually reoccupied Thrace and forced Kalojan to retreat. Kalojan's invasion collapsed because he had failed to capture Selymbria, Bizöe or Constantinople. By withdrawing into these three strongholds, the Franks avoided a pitched battle which they would almost certainly have lost. They also knew that they had time on their side, for Kalojan's large army, potentially deadly in the short term, was politically unwieldy and difficult to keep together over a longer period. Relations between Kalojan's Greek and Cuman troops must have been particularly strained, because the former hoped to reconquer Thrace intact whilst the latter were nomadic horsemen only interested in short-term pillaging.4 All the Franks had to do was wait behind their defences until Kalujan's army fell apart of its own accord. On other occasions, the Franks were themselves prevented from conquering new territories because of similar circumstances. In 1304, for example, the principality of Achaia and the neighbouring county of Cephalonia became involved in a dispute between King Charles II of Naples (1285-1309) and Anna Palaiologina, ruler of Epiros, over who should inherit Anna's despotate after her death. As a result, Charles sent a combined Italian and Moreot army into Epiros to besiege the capital city of Arta and settle the crisis by force. When they arrived, Charles's men found that Arta had been abandoned and that its defenders had retreated into the citadel taking as many arms and supplies with them as possible. They had even demolished all the houses which lay close to the citadel 'to have space to fight' and prevent the Latins from using these structures for shelter.5 Consequently the Latins withdrew and made a brief attempt to capture another castle in the vicinity of Arta, but it was situated on a steep hill surrounded by water and proved impossible even to approach, let alone besiege properly. Meanwhile the army was running out of food, but wherever they went the Latins found that 'the people from the villages had escaped to the mountains and

THE D E F E N S I V E F U N C T I O N S OF I N L A N D F O R T I F I C A T I O N S 245

the fortresses with all their supplies, so that our people could not find anything to eat'. Those who tried to follow the locals were simply attacked by Epirote troops hiding in the forests and valleys of the interior. As autumn set in and the threat of starvation began to loom, Charles's forces had no choice but to retreat. In this way the most powerful warriors of the Morea and Cephalonia were defeated not by a rival army but by the strength of the despotate's castles.e Almost twenty years later the Catalans used the same strategy to thwart an attempt by Gautier II of Brienne to recapture the duchy of Athens, which his father had lost following the battle of Halmyros in 1311. In 1331 Gautier sailed from Brindisi with an army allegedly containing 800 French knights which he had gathered together at great cost. During the ensuing weeks Gautier led this force into Attica and waited for the Catalans to meet him in battle, confident that they would be wiped out by his Prankish warriors. But the Catalans 'did not want to come out and fight', and simply waited inside their castles until Gautier, who had not come prepared for siege warfare, ran out of money.7 The Catalans emerged victorious because they understood that it would be more difficult for Gautier to hold his army together in the field than it would be for them to wait inside their strongholds. The invasions of 1205 and 1304 also collapsed through a combination of political and logistical problems long before any real fighting had taken place. This 'wait and see' strategy meant that a region could be defended effectively without the use of thousands of troops stationed along its frontiers, for in 1304 Charles II failed to conquer Epiros even though his forces had been allowed to swarm across its borders and reach Arta totally unhindered. In the Holy Land, similar tactics were used during crises such as the Khwarizmian invasion of 1244, showing that political boundaries in the modern sense were meaningless during the crusader period.8 It was castles and urban fortifications rather than lines on a map which defined the limits of a ruler's authority. The sheer amount of damage which invading armies could inflict on an area still made it desirable to stop such forces as quickly as possible, and this was the primary function of frontier castles. One of the most important such strongholds was Corinth, which acted as a kind of buffer against any invasion forces from the north because of its location at the entrance to the Peloponnese. In 1205, for example, it brought Boniface of Montferrat's advance through Byzantine Greece to a halt. Although the subsequent campaign in the Morea by William of Champlitte and Geoffrey of Villehardouin showed that this castle could be circumvented, the crusaders knew that it was dangerous to leave such a powerful citadel in the hands of the enemy. For as long as it remained uncaptured, its garrison could cut off

246

U N K N O W N CRUSADER CASTLES

the Franks' retreat and leave them stranded in the Peloponnese. This also explains why Corinth retained its strategic importance after the Catalans settled in Athens, and again during the period of Turkish expansion, when the Acciajuoli spent vast sums on maintaining its defences. Corinth did not finally fall to the Turks until 1458, by which time it had been retaken by the Greeks.9 Far to the north of Corinth, other Latin-held castles guarded the strategic mountain passes which connected northern Greece with the duchy of Athens. Perhaps the most famous of these was Boudonitza, near Thermopylae, where Leon Sgouros tried to prevent Boniface of Montferrat from entering Boeotia in 1205. It is possible that some of Boudonitza's oldest medieval fortifications date from this period and from the 12205, when Honorius III demanded that the castle should be strengthened in response to Theodore Komnenos's Epirote invasion of Latin Thessaly. Like Corinth, Boudonitza could not in itself prevent a hostile force from moving south; but anyone wishing to control the area permanently would sooner or later have to return and capture the castle, or risk being cut off.10 The importance of frontier castles as a means of delaying invaders and preventing them from making permanent conquests also gave these fortifications a special standing in the feudal structure of the principality of Achaia. This is made clear in the Assizes of Romania, the law code of Prankish Greece, which stated that any such strongholds held by the prince could not be destroyed or handed over to the enemy without the consent of his most powerful vassals. This rule was strictly adhered to in 1262, when the noble ladies of Achaia, along with the few local lords who had not been killed or captured at the battle of Pelagonia, held lengthy debates on whether to hand over Monemvasia, Mistra and Old Mania as a ransom for Prince William. In the end they agreed to do so, although those who had argued that this would enable the Greeks to 'throw us out of the country' were ultimately proved right.11 Frontier strongholds could only delay or intimidate an invader rather than stopping him in his tracks, but there were a few examples of medieval fortifications in Greece which were comparable to more elaborate continuous barriers such as Hadrian's Wall or the Great Wall of China. Most famously, the ancient Hexamilion Wall across the Isthmus of Corinth was rebuilt several times from 1415 onwards in an effort to halt Turkish attacks on the Morea; something which was never achieved because there were not enough troops or resources available to defend the wall properly.12 Less well known were the efforts made by Andronikos II in about 1307 to prevent the Catalans, who had by that time left Gallipoli and moved further west toward Thessaloniki, from returning to Constantinople. As a result Andronikos blocked

THE D E F E N S I V E F U N C T I O N S OF I N L A N D F O R T I F I C A T I O N S 247

off the narrow valley situated between the steep hills rising up from the sea and the small fortified promontory occupied by the city of Christopoli. Without access to this vital coastal route the Catalans would have found it almost impossible to lead their horses back into Thrace through the woods and mountains of the interior. Thus Andronikos may unwittingly have contributed to the downfall of Prankish Athens by encouraging the Catalans to keep moving west rather than returning to Byzantine territory.13 Decisive events such as the Catalan conquest of Athens or the Greek acquisition of Mistra were in reality rare. For most of this period strongholds were simply expected to protect people against localised raids rather than to defend the frontiers of entire Christian states. These raids could be perpetrated by a whole variety of Greek, Catalan, Turkish or even Prankish aggressors, and they ensured that most communities were situated within easy reach of some form of fortification. At Monemvasia, for example, people living in the town managed to survive a Catalan raid by climbing up to the citadel above.14 Likewise, in 1262-63 a large Greek force from Mistra entered the principality of Achaia along the same route which William II later used in 1271, past the castle of Karytaina and towards the city of Andreville. Along the way the Greeks stopped at the fortress of Veligosti, where 'they destroyed the market and left the castle intact'. This suggests that local people could at least rely on this fortress to save their lives, even if their houses were still exposed to the Greek attackers.15 Thirty-five years later Isabelle of Villehardouin (1289-1307) constructed another castle in this area which served a similar purpose, for it provided a place of refuge for rural communities whilst at the same time ensuring that local peasants paid their taxes to the Franks rather than the neighbouring Greek garrison of Gardiki.16 Similar tactics were still being used in the fourteenth century, when Niccolo Acciajuoli constructed a new Messenian stronghold to the south of Veligosti to 'provide safety for the province of Kalamata'.17 There are many Prankish towers dotted across southern Greece and the Aegean islands. Recent field work carried out in south-western Boeotia, an area which belonged to the duchy of Athens during the thirteenth and fourteenth centuries, indicates that minor Latin lords built such structures close to or inside small rural communities so that they could provide emergency shelter for Latin settlers and a place to store precious belongings. The construction of Haliartos, an unusually well-preserved tower, can be explained by the fact that a neighbouring 'large village or small medieval town' was thriving during the Prankish period. Not far away the droughts of 1989 and 1990 lowered the levels of a lake and revealed the remains of another tower and a number of surrounding structures whose outer walls formed a kind of 'protective outer facade for defensive purposes'.18 While

248

UNKNOWN CRUSADER CASTLES

towers and buildings of this kind could never have withstood a prolonged siege, they must have been an ideal way of protecting small numbers of people, along with their crops and cattle, from robbers or pirates. The presence of so many towers in certain parts of Greece also raises some interesting questions about the nature of rural settlement and the state of the local economy. On the one hand, these structures suggest that areas like Boeotia were extremely insecure and affected by constant localised warfare, yet it can also be argued that it was both expensive and timeconsuming to build stone towers of this kind, and that they could only have been erected during relatively long periods of peace and prosperity. Indeed, the field work carried out in south-western Boeotia suggests that the population of this region still continued to grow between 1300 and 1460 despite the political upheavals of the period.19 It is also worth noting that some of the better preserved Latin structures on the island of Euboea contain decorative elements such as the circular brick feature still visible on a tower at Phylla. Embellishments of this kind would surely not have been added to defences constructed at haste and with few resources in the face of overwhelming Turkish raids.20 Whilst it is true that the towers of Prankish Greece were primarily intended for defensive purposes, they were not necessarily built during the most intensive periods of warfare. It follows that the more devastated by fighting a region became the fewer strongpoints it contained, because there were not enough people or resources available locally to facilitate the construction of castles in the first place. As was noted in Chapter 4, a similar conclusion has been made in recent years with regard to the Holy Land.2i Personal touches of the kind mentioned at Phylla indicate that for many smaller western landholders the construction of a new tower represented a major financial commitment. Once it was completed it therefore became a source of pride and a status symbol within the community. It is likely that towers were intended to be inhabited continuously rather than left empty during peace time and only occupied temporarily whenever raiders appeared. If this is the case, Latin settlers may have spent far more time in the Greek countryside than one might have thought. It is often argued that Latins lived almost exclusively in cities like Glarentza, and that they rarely visited the estates and villages which they held in the countryside.22 Yet it seems odd for so many western newcomers to have constructed expensive (and sometimes even personalised) stone towers if these buildings were only used for fleeting visits, handed over to local Greek representatives or simply allowed to stand empty for most of the year. Recent research carried out in the Holy Land suggests that, far from being reluctant to leave their castles and walled cities, the Franks often lived and worked in the countryside

THE D E F E N S I V E F U N C T I O N S OF INLAND F O R T I F I C A T I O N S 249

throughout the crusader period. The construction of so many rural towers may well indicate that the same attitude prevailed in Prankish Greece.23 Other examples of the ways in which isolated towers were used to protect rural communities can be cited from the history of Mount Athos. In 1307, for example, Russian monks living there saved themselves from a Catalan incursion by seeking refuge in a tower normally used for storing wine.24 The narrow neck of land connecting Mount Athos with the mainland usually represented the first point of attack for such raiders. Consequently, in 1325, the monastic authorities constructed a tower there whose defenders were expected to give surrounding peasants early warning of any hostile incursions. They could then take shelter within the perimeter walls of the tower until the crisis had passed.25 A letter dating from the pontificate of Innocent III shows that in 1214 the unscrupulous Latin bishop of Samaria had erected an earlier strongpoint in this very same region, but that it was being used to intimidate the outraged Orthodox communities of Mount Athos rather than to protect them.26 Depending on the circumstances, towers provided an ideal way for relatively small groups of people to protect or bully those who lived around them. For many, the presence of a castle or tower somewhere close by was not enough, and they sought to protect themselves even further by actually living inside city walls or the outer ramparts of individual fortresses. The ramparts of sites such as Glarentza, Modon and Coron protected communities numbering several hundred or even thousand, whilst most larger castles incorporated entire villages situated within their outer defences. In 1391, for example, a survey of Latin settlements in the Peloponnese found that the north Achaian castle of Saint Omer had the greatest population in Morea, with 500 hearths. This figure may reflect the inaccessibility of the site, which would have offered far more protection against the Greeks and the Turks than low-lying areas nearer the coast. Today it is still possible to see the remains of this town, which was located to the north of the fortified summit and surrounded by a long curtain wall.27 The outer walls of Boudonitza and Salona in central Greece probably defended similar communities in the fourteenth century.28 The actual inhabitants of these fortified enclosures (or bourgs) must have been a mixture of Franks, Greeks and Italians. In some cases, the conquered were segregated from the conquerors. At Nauplia, the Chronicle of Morea recorded that there were two castles, one of which was granted to the Greeks by the crusaders. This suggests that the Franks occupied the upper citadel but allowed the locals to live within the lower defences, for if they had expelled them they would have been left with a ghost town.2' The Sañudo dukes of the Archipelago apparently encouraged the Greeks to live within

250

U N K N O W N CRUSADER CASTLES

the lower bourg at Naxos, whose outer curtain wall could still be seen early this century. The dukes themselves, along with their Venetian followers, probably lived at a higher level inside the actual castled In this way the Sañudos were able to minimise the threat of violence between the two peoples, and also perhaps stress the symbolic overlordship of the Venetians over the locals. That settlements of this kind were protected both by an outer wall and by an inner citadel must also have made them attractive places to live, for their outer fortifications acted as a deterrent against pirates and minor raiders, whilst their central defences could still halt the progress of more determined invaders. A similar arrangement protected most Latin-held cities on the mainland. When Kalojan attacked the Macedonian stronghold of Serres in 1205, the defenders eventually had to abandon the outer town but could still retreat into the castle, 'which was very strong'.31 In the end, the citadel of Serres could not hold out against Kalojan indefinitely, and its garrison suffered the same fate as hundreds of other Macedonians who were robbed or killed by the Cumans and the Bulgars.32 During the Prankish conquest of Achaia the spirit of the local Greeks was finally broken because they ran out of castles in which to take shelter. Like the defenders of Serres, they realised that further resistance was useless once 'they could not get help from anywhere else'. The fate of these people illustrates the importance of strongholds as refuge sites in times of war. Without such places, the inhabitants of bourgs, towns and villages could not escape being robbed, enslaved or killed.33 As in the Holy Land, the population relied on fortified houses, castles and cities to protect itself against anything ranging from full-scale invasions right down to common thieves and burglars.34 Castles could not always be expected to withstand enemy attacks in isolation. No fortress was powerful enough to resist a determined besieging army indefinitely. It was just a matter of time before Serres had to capitulate once it became clear that there were no Prankish troops on their way to relieve the castle. This realisation must have had a negative effect on the morale of the garrison, thereby bringing the fall of the citadel even closer. In a comparable incident which took place during the summer of 1292, the same fate nearly befell the Epirote strongpoint of loannina when it came under attack from Byzantine (formerly Nicaean) forces hoping to conquer the despotate of Epiros. When they heard that Florent of Hainault, ruler of Achaia, count Richard of Cephalonia and Thomas the despot were all marching toward them, the Byzantine Greeks decided, however, to raise the siege rather than 'wait for battle and be defeated and flee dishonourably'. In this way loannina, and ultimately perhaps the entire despotate, were saved from a Byzantine invasion.35

THE D E F E N S I V E F U N C T I O N S OF I N L A N D F O R T I F I C A T I O N S

25!

In certain circumstances, castles relied on field armies just as much as field armies relied on castles. Neither could survive indefinitely without the protection of the other. This interdependence between troops and fortifications could only be maintained if there were enough soldiers available to raise new armies quickly whenever castles came under threat. In order to do so, besieged lords were even allowed to call upon the assistance of their vassals without issuing the customary fifteen-day warning.36 In the case of Serres such precautions were useless, because the Franks lacked the men and resources to confront Kalojan in the open and force him to retreat. The same reason explains why the rest of Macedonia, Thessaly and indeed much of the Latin empire eventually proved powerless to resist the relentless invasions of the Bulgars and the Greeks. From the early 12605 onwards the same fate befell the hopelessly outnumbered Franks of Syria and Palestine, who could only watch and wait as successive Mamluk sultans picked off one Latin castle or city after another.37 Whenever castles were situated in the mountains, their Latin garrisons stood a much better chance of defending themselves even if they could not match their opponents man for man. Inaccessible sites such as Karytaina were also better suited to defensive warfare because they had good visibility. This was particularly important for elevated frontier castles like Corinth, whose defenders could spot an invasion force crossing the nearby isthmus when it was still several miles away. On a clear day the garrison of Monemvasia could see as far as Crete, and enjoyed such extensive views over the surrounding coastline that a surprise attack on the site must have been out of the question. This no doubt explains why the inhabitants of the town below had plenty of time to escape the Catalan raid on their community in 1292. Further north, both Thebes and Salona were also built on outcrops which dominated roads linking Boeotia, Thessaly and the Gulf of Corinth. Soldiers stationed in these strongholds could therefore use messengers or fire signals to warn people living nearby about imminent hostile incursions, thereby giving them time to reach the nearest castle or tower.38 At other sites garrisons were able to send messages to each other, not just neighbouring settlements. During the 13305, for example, the German pilgrim Ludoph of Sudheim referred to a network of fire signals between the Hospitaller islands of Rhodes, Cos and Castellorizzo, which would have acted as an early warning system against Turks or pirates. Ludolph wrote that it was manned by brothers of the Order, who used 'smoke by day and torches by night'.39 It is also possible that the citadel of Corinth and the Prankish tower on the Athenian acropolis could communicate with each other, in much the same way that Cypriot mountain castles such as Buffavento could further east. Indeed, the small projecting turret visible in

252

UNKNOWN CRUSADER

CASTLES

nineteenth-century photographs of the Athenian tower may have been specifically designed for sending fire signals.40 In the past it has also often been argued that many of the medieval towers in Greece formed even larger intervisible networks covering entire regions or islands. Towers such as that at Haliartos in Boeotia, it has been claimed, guarded strategic roads, whilst the fifty-odd towers built on Euboea were thought to have been look-out posts constructed by the Venetian authorities to guard against seaborne raiders.41 However, we have already seen that the Euboean towers were actually built by individual Lombard settlers; a conclusion which is confirmed by the fact that very few of them were situated in strategic coastal locations or were even intervisible with each other. Instead, it has been shown that these strongpoints, along with similar structures in the duchy of Athens such as the Haliartos tower, were intended as local refuge points, agricultural centres or status symbols, having no wider strategic function beyond their immediate surroundings. Even though they must have been useful places from which to observe the countryside around them, there is no evidence that they belonged to larger, centrally planned networks of towers capable of communicating with each other by means of fire signals.42 In any case, a whole variety of financial, logistical and political difficulties ruled out the creation of early warning systems involving more than just a handful of castles.43

23

Fortifications and Internal Security As in other parts of the eastern Mediterranean, the Latins who settled in Greece occasionally had to deal with the threat of internal violence or rebellion. In order to minimise the risk of traitors or rebels suddenly taking control over individual Latin strongholds, it was essential that such places were always properly defended by loyal soldiers regardless of whether the surrounding countryside was actually in a state of war or not. This was stressed in the Assizes of Romania, which stipulated that liegemen who performed annual service to their lord should spend at least four months a year simply guarding castles.1 To some extent the Franks also reduced the threat of rebellion by adopting a more tolerant attitude toward the Greeks and recognising the administrative practices of the former Byzantine empire. In Achaia, for example, 'the Greek noblemen who held fiefs and estates and the villages of the land' were normally allowed to keep enough earnings to maintain their social status. It was also agreed that 'the people should pay and serve in the same way that they had done during the overlordship of the emperor of Constantinople'.2 The Franks, realising that there was enough land available for everyone, were happy to make such arrangements because they were far too outnumbered to impose their will on the Greeks through brute force. Rather than expelling or even killing local people, which would have brought the rural economy to a standstill, the Franks allowed the Greeks to hold property in exchange for their taxes, obedience and military service. This latter obligation was sometimes offered willingly, for the Greeks feared other aggressors, such as pirates, just as much as the Latins.3 Although the Prankish conquest of Greece did not bring with it any of the horrors associated with the fall of Jerusalem and the early years of crusader rule in the Holy Land, this did not alter the fact that the Latins were still regarded as unwelcome invaders, and that they consequently needed places of shelter which could protect them against potential rebels. It has already been suggested that this was why Latin newcomers very often constructed isolated donjons or towers in the areas which they occupied, for such structures were relatively quick and easy to build yet provided plenty of protection for small groups of settlers arriving in a foreign land.4

254

UNKNOWN CRUSADER CASTLES

The historical evidence also suggests that the Greeks, lacking the equipment needed to besiege castles and noting that others might already have come to terms with the Franks, tended to stop fighting as soon as local strongholds were occupied by the crusaders. In Messenia, for example, the Franks initially experienced problems imposing their authority over the local Greeks, but once they captured the strategic fortress of Kalamata this situation changed dramatically. 'Afterwards', wrote Villehardouin, 'more Greeks from the country submitted to them than ever before.'5 Kalamata became one of the many Byzantine castles strengthened by the construction of new Latin fortifications, and a similar process was going on all over the Morea during the early years of the thirteenth century. The men responsible for these changes were figures like William Alemán of Patras, newly created barons who began 'to change their surnames and take the names of the fortresses they were building'.6 In so doing, they were actively encouraged by the rulers of Achaia, who specifically stated in the Assizes of Romania that the twelve most powerful vassals of the principality could build their own strongholds unhindered. This policy ensured that the countryside was subjugated as quickly as possible: the more Prankish castles there were, the more pointless it became for the Greeks to rebel. It may also have been intended to encourage Prankish nobles to stay in Greece and consolidate their new baronies, thereby stabilising the political situation between fellow newcomers, not just between Greeks and Franks.7 The need to create strong and viable new lordships as quickly as possible meant that other westerners relied on the same tactics as those used by the Achaian Franks. At Coron, the ever expanding thirteenth-century fortifications of the citadel enabled the Venetians to dominate much of the surrounding countryside, not just the city itself.8 During the first three decades of Venetian rule on Crete, castles and towers were also used to protect Italian settlers against the locals, who resented having to share their land with the newcomers and were less well treated than the Greeks in the Morea.9 The mountainous areas of western Crete proved particularly troublesome in this respect, so the Venetians eventually built the fortress of Suda on the north-west coast of the island to try to bring order to this region. While this gradually wore down the resistance of the Greeks, it is interesting to note that, by placing Suda by the sea, the Venetians created an escape route for themselves in case it ever became necessary to withdraw from the area completely. Suda's location also suggests that the Venetians were happy to contain rather than conquer western Crete, in the same way that William II of Villehardouin decided to surround the Melings with a ring of frontier castles.10 Despite such tactics the Greeks still rebelled openly from time to time,

F O R T I F I C A T I O N S AND I N T E R N A L SECURITY

255

particularly in areas where Latin rule was already relatively weak. After Michael VIII Palaiologos acquired Maina, Mistra and Monemvasia in 1262, for example, a combination of local grievances and encouragement from the Byzantine authorities often led to uprisings in the south-eastern Peloponnese. In 1296, for example, an obscure argument between a Greek lord and a Prankish knight was used as the excuse for a massive insurrection in northern Laconia. This in turn resulted in the fall of St George, an important Latin frontier castle, after a traitor within the ramparts lowered a ladder to Greeks waiting below. In response, Florent of Hainault organised a lengthy campaign against St George, which incidentally involved the use of small forts similar to those built by the Franks at Corinth between 1205 and mo.11 This must have been successful, for according to the Chronicle of Morea St George belonged to the Latins again about ten years later.12 Most other rebellions mentioned by contemporary sources were caused by more specific grievances. Heavy taxation, combined with a general feeling that the Latins were not maintaining established Byzantine customs, seem to have sparked off a second uprising around St George in 1302.13 Similar complaints probably led to such a large number of disturbances on Crete.14 On Chios heavy-handed tactics by the Genoese later led to so much local resentment that between 1329 and 1346 Genoa was forced to abandon the island altogether. No doubt the Genoese were encouraged to do so because of the strong political support, and perhaps even direct military assistance, given to the inhabitants of Chios by Andronikos II.15 Like the Muslim peasants who rebelled after the battle of Hattin, the Greeks also took their chances whenever the Latins had been defeated by an external foe and were in serious military trouble.16 The Thracian rebellion of 1205 was clearly timed to coincide with Kalojan's invasion of the area, whilst in 1207 the Greek capture of the local Prankish lord precipitated a more spontaneous uprising near Nicomedia.17 Insurrections also broke out in the wake of both the battles of Halmyros (1311) and Pelagonia (1259), devastating encounters which inflicted such heavy losses on the Franks that it proved impossible for them to maintain law and order.18 Even if castles sometimes failed to prevent rebellions, such structures could still minimise their impact. This point is perhaps best illustrated by returning to the uprising of 1302, supposedly caused by the heavy taxes imposed on the Greeks by the Achaian prince Philip of Savoy, but clearly also backed by the Byzantine authorities at Mistra. This rebellion started well for the Greeks, who captured and destroyed the castles of Ste Hélène and Crevecoeur located near Karytaina, before the Franks even had time to react. But by the time the rebels reached Beaufort (another fortress near Karytaina, which should not be confused with its namesake in the Mani

256

U N K N O W N C R U S A D E R CASTLES

peninsula),19 they had already lost the element of surprise and realised that it would be impossible to storm the castle without suffering heavy losses at the hands of the Prankish crossbowmen stationed along the ramparts. They therefore changed their minds and marched to St George, only to find that they lacked the equipment to besiege it. To solve the problem, they asked for a trebuchet to be brought from Monemvasia, but by now they had lost so much time that the initial impetus of the rebellion was slipping away. Meanwhile, Philip of Savoy and his barons were able to raise an army and march south toward St George, forcing the rebels to retreat into the mountains or back to Mistra. As a result, all resistance crumbled, and Philip was able to rebuild or strengthen those castles which had been attacked, collect any outstanding taxes and deal with the leaders of the uprising.20 Despite managing to destroy two entire castles, the Greek rebels had failed because they lacked the resources to undertake a more extensive campaign involving lengthy sieges. They needed the element of surprise to succeed. Once this had been taken away from them, they did not have their own strongholds nearby in which to regroup or take shelter. The Prankish defenders of St George and Beaufort, on the other hand, knew that they had time on their side, and were happy to adopt the same tactic of wait and see as that used by the Catalans against Gautier II of Brienne in 1331. In doing so they prevented the rebellion from spreading any further, and ultimately caused it to collapse entirely. By retreating inside their castles rather than trying to confront rebels in the field, the Franks could also keep their casualties to a minimum. In 1302 this was only a temporary measure until Philip of Savoy turned up, but on other occasions it was less clear if and when a relieving army would arrive. This was certainly the case for the unfortunate Renier of Trit, lord of Philippopolis and Stenimaka, a town and castle situated on the northern fringes of the Latin empire. When Kalojan invaded this area and the local Greeks rose up in arms against the Latins (1205), Renier found himself cut off at Philippopolis with 120 knights. During the ensuing weeks, small groups from this force tried to leave the city and make the dangerous nine-day journey to Constantinople despite Renier's efforts to persuade them to stay. Most of them probably suffered the same fate as Renier's son and brother, who were captured and beheaded along with at least thirty other knights soon after departing from Philippopolis. As if this were not bad enough, Renier also began to hear rumours that the inhabitants of Philippopolis itself were planning to rebel and deliver the city to Kalojan. He therefore retreated to the castle of Stenimaka with his fifteen remaining knights and stayed there, cut off from the outside world, until a relieving force from Constantinople finally reached him in June 1206. By this point Renier had

F O R T I F I C A T I O N S AND I N T E R N A L SECURITY

257

been stranded for thirteen months, but by refusing to leave his stronghold he had avoided the fate of most of his companions, and had survived one of the worst rebellions of the entire crusader period.21 Somewhat less spectacular tactics enabled the Genoese to withstand an uprising on Chios in 1347. This rebellion had been organised by a local nobleman, who had gathered together a force of Greeks and mercenaries which was so large that it may well have outnumbered the total number of westerners on the island. Rather than trying to confront them, the Genoese responded by withdrawing inside the powerful citadel of Chios. Here they were besieged for a while, until the arrival of reinforcements caused the entire insurrection to fizzle out. Once again, fortifications rather than superior troop numbers had saved both the lives and the territorial claims of the Latin newcomers, enabling them to wait in safety until further assistance arrived from elsewhere.22 For the actual rulers of crusader states there was always the added threat that their Latin as well as their Greek vassals would rebel. The obvious way to prevent this was to make sure that the most powerful castles in any given area belonged to the local ruler rather than his barons. In the Catalan duchy of Athens, the key centres Thebes, Livadia, Siderokastron, Neopatras and Athens all belonged to the royal domain of the Aragonese dukes.23 These sites were guarded by inaccessible hill top citadels whose capture would have been beyond the means of any local baron. In 1275 not even the Byzantine army of Michael VIII proved capable of taking Neopatras.24 In Achaia the strategic fortresses of Corinth, Kalamata and Androusa, as well as Chlemoutsi and Glarentza, were all held by the rulers of the principality. Between 1249 and 1262 these were joined briefly by Mistra and Monemvasia.25 Further east, the fortified cities of Constantinople and Thessaloniki were of course also royal properties, although Constantinople in particular contained a very large Venetian quarter.26 Having gained possession of key strongholds like these, it was essential for individual rulers to maintain their defences. When Kalojan sacked the royal city of Serres during the Bulgar invasion of 1205, he deprived Boniface of Montferrat of an important economic and military centre and therefore undermined the latter's political domination over the surrounding area. As soon as Kalujan retreated, Boniface refortified the site, thereby re-establishing his own authority over the local Greeks and Latins, and at the same time protecting these people against future Bulgar invasions.27 Even if ramparts were kept in perfect condition, this was only good enough if the troops who manned them could be trusted. As a result, garrisons were often changed on the accession of new sovereigns. In 1301 Philip of Savoy, ruler of Achaia by virtue of his marriage with Isabelle of Villehardouin, 'had the

258

U N K N O W N C R U S A D E R CASTLES

castellans and constables and some of the sergeants changed in all the castles of his principality of Morea, and placed in them some of the people he had brought from Piedmont and Savoy'.28 The purpose of this policy was to avoid treachery and foster loyalty toward individual rulers, but it did not always work. The traitor who let Greek rebels inside St George in 1296, for example, had the suspiciously Latin sounding name of Boniface.29 Likewise, individual Franks at Karytaina and neighbouring castles were prepared to hand these places over to the Greeks in return for personal financial gain.30 Most rebellions or acts of treachery which involved fellow westerners were caused by wider political disagreements. During the earliest days of Latin rule at Constantinople arguments arose between Boniface of Montferrat and his overlord, Baldwin of Flanders, regarding the kingdom of Thessaly. Thessaloniki in particular became a sore point between the two men. Boniface hoped to make it his new capital, whereas Baldwin was well aware of the city's traditional imperial status.31 Matters deteriorated even further in the autumn of 1204, when Baldwin managed to seize Thessaloniki before Boniface could get to it. As a result, the latter decided to rebel openly against his lord and hastily laid siege to Adrianople, which had also been captured recently by Baldwin. The emperor's followers inside Adrianople therefore appealed to Constantinople for help, and a relieving army was dispatched from the capital which forced Boniface to raise his siege and come to terms. Boniface was subsequently allowed to keep Thessaloniki in exchange for recognising the emperor's overlordship and territorial claims in Thrace.32 Boniface had been prevented from conquering Thrace, and perhaps even overthrowing Baldwin, because he had failed to capture Adrianople, whose role in the affair can be compared with that of Beaufort or St George during the Greek rebellion of 1302. Sometimes, fortifications tended to erode rather than consolidate central authority. This is hardly surprising when one considers that some Latin nobles had inherited sprawling classical or Byzantine strongholds originally intended to withstand far larger besieging armies than anything the Franks could muster. If vassals who held such sites were able to resist external invasion forces numbering thousands of men, it must have been tempting at times to defy a royal army composed of just a few dozen knights. This certainly seems to have been the case during the winter of 1208-9, when a powerful group of Lombard nobles who had settled in northern Greece rebelled against the Latin emperor Henry. This rebellion was centred around a plot to replace Demetrius, the infant son of Boniface of Montferrat (died 1207), with a rival claimant to the kingdom of Thessaly.33 In order to protect Demetrius Henry was obliged to lead an expedition into the kingdom of

F O R T I F I C A T I O N S AND I N T E R N A L SECURITY

259

Thessaly in midwinter, passing the important harbour of Christopoli (Kavala) before reaching Thessaloniki itself. The Lombards refused to allow Henry's troops inside either of these walled cities, thereby preventing his army from collecting supplies from the imperial fleet sailing along the coast, or indeed finding adequate shelter from the terrible winter weather. The Lombards must have known that in these circumstances Henry had no chance of storming Thessaloniki's ramparts. As a result he eventually had to use guile to gain entry to the city.34 Even then, Lombard troops continued to defy Henry at Serres and Christopoli, and later also at Larissa and Thebes. In the end, Henry failed to capture any of these strongholds by storm and had to fall back on a combination of threats, diplomacy and victories in the field to obtain their surrender.35 When one considers the problems Henry had in dealing with the Lombards, it is easy to understand why the Aragonese dukes of Athens, who did not even reside in Greece, found it increasingly difficult to impose their rule over their Catalan vassals. Indeed, by the early 13605 it appears that vicars-general sent to Athens by King Frederick III, the Aragonese ruler of Sicily (1355-77) who also held the ducal title, were no more than pawns in the political power games of local Catalan barons. One of the most important such figures was Roger de Lluria, who seized and ruled Thebes between 1362 and 1366. Lacking the means to remove Roger from this powerful citadel, Frederick III eventually had no alternative but to acknowledge his status as the de facto vicar-general. Roger continued to rule until his death in c. 1370.36 During the long and violent history of Prankish Greece there were many similar instances of baronial disloyalty involving castles. During the mid 12505 Thebes had been at the centre of another internal conflict, when its lord Guy de la Roche, aided by his Venetian allies, clashed with his overlord William II of Villehardouin in a dispute over land on Euboea. William must have thought that he had brought this struggle to an end when he finally defeated Guy in 1258, but the latter managed to escape and seek refuge at Thebes, which proved too strong for William's besieging forces. As a result, Guy only surrendered after Achaian troops systematically began to ravage his lands and thereby threaten him with financial ruin.37 In a similar but much later episode Hugh of Lusignan's inability to overcome the fortifications of Patras dashed his hopes of seizing control over Achaia in 1366. Hugh's greatest rival in this dispute was Philip of Taranto, younger brother of Robert, who had ruled the principality between 1346 and 1364. Shortly before the siege of Patras Philip launched his own attack against Hugh's headquarters at Old Navarino, but he too failed to make any progress. For a while Hugh and Philip were deadlocked, because neither

2ÓO

UNKNOWN CRUSADER CASTLES

claimant succeeded in taking the other man's castles. Philip only emerged victorious after Hugh promised to leave Greece in return for a substantial pay-off.38 Incidents like these illustrate how castles could prevent lords from imposing their will on troublesome vassals or rival claimants. But the chronology and frequency of rebellions can also be used to shed more light on the decline of central authority in general. Many of the rebellions which have been mentioned occurred in the latter half of the fourteenth century, when the rulers of Latin states in Greece were increasingly absent in western Europe. By contrast, the fact that there were relatively few disturbances in Achaia before the loss of Mistra, Monemvasia and Old Mania (1262) suggests that the principality enjoyed better internal security during this period. To some extent, this is confirmed by contemporary descriptions of Andravida (Andreville), the thirteenth-century capital of Achaia, which was situated in the middle of an open plain, 'without any walls or a citadel'. Instead of trying to rectify this situation, the earliest rulers of Achaia preferred to reside at Chlemoutsi, which provided them with shelter but still lay conveniently close to their centre of government.39 This situation would have been unthinkable in most other crusader states. If Constantinople, Salonika, Acre, Antioch or Tripoli had not been protected by major urban fortifications these cities, along with all the territories around them, would have been overrun in a matter of weeks. Indeed, the only Prankish settlement in the entire eastern Mediterranean which was at all comparable with Andravida was Nicosia, the capital of Cyprus. Before 1291, and possibly until as late as the 13505, Nicosia was only defended by a citadel and probably did not have any urban fortifications to protect it against a major invasion force. Cyprus was also similar to the early thirteenth-century principality of Achaia in that both were geographically isolated and therefore less exposed to external invaders than Christian territories which shared long borders with hostile Greek, Bulgar or Muslim states. As a result Cyprus and pre-i26os Achaia frequently enjoyed greater internal stability than their neighbours. This may explain why the Franks did not make an effort to fortify Andravida until the Greeks had taken over Mistra and were beginning to attack the principality regularly. These attacks encouraged local people to rebel, thereby increasing the need for urban defences even further.40 A similar process took place on Cyprus, whose capital may not have had adequate urban defences around it until improvements were carried out following the Genoese occupation of Famagusta in the i37os.41 Returning to the Aegean, it would be a mistake to conclude that Latin settlers and Greek natives all hated each other, and that the two communities

F O R T I F I C A T I O N S AND I N T E R N A L S E C U R I T Y

201

existed in total isolation. The existence of so many Prankish towers dotted across the landscape indicates that many newcomers from western Europe actually lived amongst (or at least close to) the Greek rural population. On the other hand, we have seen that rebellions did break out from time to time, particularly in frontier regions which were vulnerable to external attack. Even in times of peace cultural and religious differences could sometimes spill over into violence. For example, in 1212 it was reported that a Latin priest had been assaulted by Greeks near Gravia in the duchy of Athens.42 Under such circumstances the Latins doubtless felt reassured by the presence of castles, towers and fortified farmhouses even if they did not necessarily live inside them all the time. The situation seems to have been very similar to that which prevailed in the Holy Land, where Frankish settlers were perfectly prepared to live in the countryside yet still relied on strongpoints to shield them against the crime and violence which were so commonplace in the middle ages. On a broader level, it has also been shown that the Latins themselves often clashed over important political or territorial issues, and that on such occasions possession of a key fortress or city could decide the ultimate outcome. Thus fortifications played a vital, if sometimes destructive, role in the internal politics of Frankish Greece.43

24

The Military Orders Unlike the Holy Land, where the Hospitallers, Templars and Teutonic Knights held numerous powerful strongpoints such as Crac des Chevaliers, in Greece the contemporary sources suggest that these Orders kept a much lower profile. According to the Chronicle of Morea, for example, they each owed the prince of Achaia four knights' service annually; a tiny amount when one considers that in Galilee the castle of Saphet alone had a peace time garrison of 1700 troops, including fifty Templar knights.1 The relative insignificance of the Military Orders in Prankish Greece makes it difficult to establish precisely which castles and estates they held there. The headquarters of the Teutonic Knights, for example, appear to have been situated at Mostenitsa, which, according to a papal letter of Gregory IX, lay somewhere in the diocese of Coron.2 As a result, attempts by historians to link Mostenitsa with a ruined Prankish tower in southern Elis have been rejected by those scholars who argue that this site lies too far north of the Venetian colony. Whether one believes that Mostenitsa was an isolated tower in Elis, or indeed a totally different fortification which has long since disappeared, the lack of firm evidence is in itself indicative of how small this castle must have been compared with German strongholds in the Holy Land such as Montfort or Judin.3 From a military point of view, we should probably attach even less importance to the grant made to the Teutonic Knights in 1237 of a house inside the walls of Chlemoutsi. Although this indicates that the brothers helped to garrison the fortress, Chlemoutsi lay far away from any frontiers at this stage and was designed more as an impressive fortified residence than as a strategic military outpost.4 The grant at Chlemoutsi nevertheless reflects the close links between the Teutonic Knights and the rulers of Achaia; a relationship which took on much greater significance after 1300. According to the Aragonese version of the Chronicle of Morea, for example, the Order was granted the Moreote fortress of Castiello Nuevo toward the end of the fourteenth century, about a hundred years after it had been constructed by the Achaian baili Nicholas III of Saint Omer.5 This seems to have formed part of a wider fourteenthcentury effort to use the Order in the struggle against the Greeks of Mistra, for around 1320 it was claimed that the Knights' preceptor was killed

THE M I L I T A R Y

ORDERS

263

defending the castle of St George. This fortress was situated not far from Karytaina, which fell to Mistra at about the same time and was located so far inland that by this point the Franks clearly needed all the help they could get to stem the Greek advance.6 During the 13305 the German traveller Ludolph of Sudheim also reported that 'in Achaia, or Morea, there are brothers of the Teutonic house guarding very strong castles [and] constantly fighting against the dukes of Athens and the Greeks'.7 This indicates that as the fourteenth century progressed the Order was kept even busier by Catalan as well as Byzantine incursions against the remaining Prankish lands in the north and west of the Peloponnese. Presumably the Teutonic Knights continued to participate in this struggle until the very end, for the last known references to Mostenitsa date from the early fifteenth century, by which time most of Messenia already belonged to the Greeks of Mistra.8 Beyond the Peloponnese it is impossible even to gather this amount of evidence about the Knights. If they did hold any strongholds in the Latin empire before 1261, or the in the duchy of Athens until 1311, all historical and archaeological traces of their existence have long since disappeared.9 Due to their arrest and disappearance at the beginning of the fourteenth century, the evidence concerning the Templars is even sketchier. This problem can be illustrated by looking at the history of Palaiopolis, a small village situated in central Elis. In 1210 this and two other settlements were granted to the Templars by leading members of the Prankish invasion force, including William of Champlitte.10 After the fall of the Templars, the site was taken over by the Hospitallers, who either inherited or constructed the medieval tower situated on a hill near the village. The present remains of this structure indicate that it was rectangular, incorporated reused classical masonry and had been equipped with a large vaulted cellar or cistern. In short, this was a typical Prankish tower of the kind erected all over Greece, although the presence of a cistern suggests that it was slightly more substantial than the majority of such fortifications.11 Sixty years after the dissolution of the Templars, 'the castle of Palaiopolis' still belonged to the Hospitallers.12 At this time the Order also held a castle, or, more likely a fortified tower, at Laffustan (Phostena) in northern Achaia, another site originally granted to the Templars. Presumably, Laffustan evolved from a Templar village into a Hospitaller stronghold in the same way that Palaiopolis did.13 Beyond the Peloponnese, papal documents dating from 1210 reveal that the Templars had held Lamia (Zeitoun) and the neighbouring (but unidentified) Byzantine stronghold of Ravennika until the previous year. These sites had been granted to the Order by Boniface of Montferrat and other Latin newcomers shortly after their initial arrival in Thessaly. Thereafter the Order

264

UNKNOWN CRUSADER

CASTLES

claimed that it had been responsible for the reconstruction of Lamia, an ancient acropolis whose location on the borders of Boeotia and Thessaly gave it great strategic importance throughout the Prankish and Catalan periods. In 1209, however, both this castle and Ravennika had been taken from the Templars by the Latin emperor Henry because of their support for the Lombard rebellion of the previous winter.14 It is even likely that Lombard rebels fleeing from Henry had taken shelter inside Lamia at the beginning of 1209, indicating that the Templars had become friendly with the newly arrived north Italian settlers of eastern Thessaly. However, Henry II's punitive actions against the Order, combined with the steady advances of the Epirote Greeks into the region over the next few years, make it uncertain whether the Templars ever returned to Lamia or Ravennika later in the thirteenth century, i s During the same early period that they were given these two castles, the Templars probably also acquired Sykaminon, situated near the east coast of Attica and referred to in the papal documents of Innocent III as àe Rupo.16 By 1314 this castle, like those already mentioned in Achaia, had passed to the Hospitallers as a result of the fall of the Templars.17 It is clear that the Hospitallers subsequently retained Sykaminon throughout the Catalan period, for in 1395 the pilgrim Niccolo da Martoni, who had just arrived in the area by ship from Euboea, described a harrowing three-mile journey which he undertook from the coast in order to visit them. According to Niccolo, the entire region around the castle was permanently threatened by the Turks, as well as a group of Albanian robbers who occupied a neighbouring stronghold, and as a result he was greatly relieved both to reach and to get away from Sykaminon without being attacked. It is clear that by the late fourteenth century the castle acted as an isolated refuge in a sea of lawlessness and piracy.18 Another important fortress held briefly by the Hospitallers was that of Gardiki, situated in the mountains to the east of Lamia. Unlike many other Hospitaller properties, which were acquired from the Templars at a much later date, Gardiki was seized by the Order from the local Latin bishop in 1210.19 Before this dispute could be resolved, Gardiki fell to the Greeks of Epiros, and it was not returned to Prankish control until the 12705, when William de la Roche married the daughter of the Greek ruler of Neopatras and received both Gardiki and neighbouring Lamia as her dowry.20 Almost half a century later these strongholds were captured by the Catalans, ruling out any possibility that the Military Orders ever returned to them.21 Indeed, both the Templars and the Hospitallers did little in the region throughout the thirteenth century, presumably because they wished to concentrate most of their resources on the Holy Land. Nevertheless, occasional references to

THE M I L I T A R Y O R D E R S

205

these Orders, such as their contribution to William II of Villehardouin's campaign against Guy de la Roche in 1258, show that they were still involved in some of the military activities of the region even after they no longer held the major castles of Lamia and Gardiki.22 Events which took place during the late thirteenth and early fourteenth centuries were to change this situation dramatically. After the fall of Acre in 1291 the Templars and the Hospitallers retreated to Cyprus, but its ruler, Henry II, feared that they would gain too much power on the island and therefore prevented them from acquiring any new properties there. Then, in 1312, the Templars were suppressed and their properties in Greece were gradually transferred to the Hospitallers. Even before they received these estates the Hospitallers, frustrated at the situation in Cyprus, were expressing a growing interest in the Aegean. This area offered them greater freedom to expand and the opportunity to continue the struggle against Islam by fighting the Turks rather than the Mamluks. Between 1306 and 1310 they therefore invaded Rhodes, an island which still formally belonged to the Byzantine empire but which had increasingly fallen under Turkish and Genoese influence. In a drawn out campaign, whose chronology remains unclear, the Hospitallers gradually wore down Greek resistance on the island, probably securing the actual city of Rhodes in mid 1308. By 1310 the conquest was complete, although it still took several years to establish full control over surrounding islands; a process which was hampered by lack of money and the resistance of Greeks, Turks and Italian merchants angered by the Hospitallers' condemnation of their Muslim trade links.23 Despite these problems, the Hospitallers succeeded in turning Rhodes into a heavily fortified base for their crusading activities. The exact appearance of the island's defences at the time of the Order's invasion in 1306 is impossible to establish, although the amount of time it took the Hospitallers to conquer Rhodes suggests that some of them were formidable. During the fourteenth century the brothers built or repaired numerous castles such as Lindos, the remains of which can still be seen today. Outlying islands controlled by the Order, including Castellorizzo, Carpamos and Telos, were also defended by smaller strongpoints. In the fifteenth century these were joined by a new castle at Bodrum, the Hospitallers' foothold in mainland Turkey between about 1408 and 1522.24 The most extensive fortifications, however, were erected around the city of Rhodes, protected by a powerful rampart which formed an arc around the town and the harbour. In the north west a separate walled enclosure overlooking the port formed the inner citadel and contained the living quarters, administrative buildings and hospital of the Order. The harbour was protected by two fortified moles and, until its destruction in the nineteenth century,

206

UNKNOWN CRUSADER CASTLES

was dominated by an enormous tower which stood to a height of almost fifty metres. In the fourteenth century the land defences consisted of a single wall, outwork and moat, but later, and particularly in response to the failed Turkish siege of 1480, another rampart and ditch were added and older structures were remodelled to withstand the impact of gunfire. Many pre-i38o fortifications have therefore been hidden beneath alterations carried out between the fifteenth century and the final loss of Rhodes to the Ottomans in 1522. Contemporary descriptions of Rhodes by travellers such as Ludolph of Sudheim nevertheless indicate that by the 13305 extensive Hospitaller fortifications already existed both here and at other sites in the Dodecanese.25 These fortifications turned Rhodes into an ideal centre for operations against the Turks. Within fifteen years of their arrival on the island, the Hospitallers were involved in sporadic naval campaigns throughout the Aegean. Between 1344 and its loss in 1402 the Order took most of the responsibility for the defence of Smyrna, a Christian outpost on the coast of Turkey whose capture has already been referred to.26 As the military situation in mainland Greece deteriorated, the Hospitallers were also encouraged by the papacy to take a more active role in the defence of the Peloponnese. In 1356 Innocent VI considered giving the Order more land in Achaia to bolster Latin resistance against the Turks.27 Nothing came of this, but twenty years later Queen Joanna of Naples granted the entire principality to the Hospitallers for an annual rent of 4000 ducats over a five year period.28 In 1381 this arrangement came to an end, but the Order was prominent in efforts to defend Corinth and the nearby isthmus at the beginning of the fifteenth century. This rather patchy involvement in the struggle against the Ottoman Turks reflected the financial and political problems experienced by the Hospitallers after 1306. Although Rhodes itself was an impressive strongpoint, its running costs turned out to be enormous. The Hospitallers often found themselves heavily in debt and lacking the troops needed to garrison their existing castles, let alone any new fortresses granted to them. Their military effectiveness was also undermined by a combination of corruption, internal squabbles at Rhodes and wider political issues such as the papal schism. Defeat in battle and mounting debts explain why the Order did not extend its lease over Achaia after 1381. The financial burden of defending strongholds like Smyrna and Bodrum made it impossible to exploit the full potential of these bridgeheads by undertaking further conquests inland. Hence the Hospitallers' efforts against the Turks, though often very effective on a local scale, lacked the strength to alter the final outcome.29 Returning to mainland Greece before 1380, castles belonging to the

THE M I L I T A R Y O R D E R S

267

Military Orders fulfilled two basic functions, one military and the other economic. Turning first to the military role, it is clear that Ravennika and Lamia had originally been granted to the Templars to protect the duchy of Athens and the eastern fringes of the kingdom of Thessaly from attacks by the Greeks of Epiros. These castles were needed 'to defend the land' and presumably the Hospitallers occupied neighbouring Gardiki for the same reasons.30 The emperor Henry's confiscation of Lamia and Ravennika from the Templars may have been foolish, for by the 12203 the Epirote Greeks had overrun this area and were even threatening Thebes.31 In broader terms, the Templars' links with these castles so soon after the Fourth Crusade indicate that the Order was involved with the initial Latin conquest of Greece. In 1206 the Latin emperor granted the Byzantine settlements of Satalia and Pergamum, both of which lay in areas of Asia Minor which had not yet been conquered by the Franks, to the Templars and the Hospitallers respectively.3^ This suggests that the first western settlers in Greece were attempting to pursue the same policies as their Christian neighbours in Cuidan Armenian and the Holy Land, where frontier strongholds were granted to the Military Orders in order to encourage expansion, spread the financial burden of defence and protect vulnerable territories against external aggressors. In the former Byzantine empire, these policies ultimately failed through a combination of internal conflict with the Latin emperor Henry and overwhelming external pressure from the Greeks of Epiros and Nicaea. These two neighbours had already overrun virtually all Latin territories in northern Greece and Asia Minor by the mid 12205. Clearly a similar process took place in southern Morea about a century later, when the Teutonic Knights were called upon to garrison castles like St George in a valiant but ultimately failed attempt to stop the Greeks of Mistra from advancing into the principality of Achaia. The Hospitaller garrison of Sykaminon seems to have performed a similar role by the time Niccolo da Martoni visited it in 1395, but this cannot have been the castle's original function before the arrival of the Ottoman Turks, when it had still been located deep inside Latin territory. The presence of other Hospitaller properties between this area and Athens, combined with Niccolo da Martoni's reference to 'the port of Sykaminon', suggests that the site had originally been intended as an economic and administrative centre which could communicate with Rhodes or western Europe via the sea.33 Similar observations can be made about Achaian strongholds such as Mostenitsa and Palaiopolis, mere towers which were not located anywhere near important frontiers before the loss of Mistra in 1262. On the other hand, Mostenitsa lay very close to other German estates in Messenia which became the subject of a land dispute with nearby Modon during the early

268

UNKNOWN CRUSADER

CASTLES

fifteenth century. The fact that they were later granted Castiello Nuevo, and had also acquired a house inside the walls of Chlemoutsi in 1237, implies that the Teutonic Knights gradually became relatively powerful landowners in the area. It seems that Mostenitsa must have acted as the focal point for a large agricultural domain, and that it was primarily designed to protect produce, cattle and farm revenues rather than any major settlements or strategic frontiers. The castle's fortifications only needed to be strong enough to deter local criminals, rebels and pirates.34 This arrangement had its parallels further east, where towers such as Hospitaller Kolossi on Cyprus were often used to house regional commanders and safeguard local farming against thieves or pirates.35 The changing role of the Military Orders in Prankish Greece, and in particular the Hospitallers, who started off as fairly minor landowners in the years immediately after the Fourth Crusade but eventually controlled all of Achaia between 1376 and 1381, reflected the wider political history of the region. The insignificance of the Orders in thirteenth-century Achaia implies that the principality enjoyed relatively good internal stability at this time, for no ruler would actively encourage the Hospitallers, Templars and Teutonic Knights to build up vast castellanies and estates unless it was absolutely necessary. Muslim pressure had already created a situation in Palestine and Syria where the three Orders held virtually independent lordships whose presence safeguarded Christian territories but eroded central authority.36 Similarly, the appearance of the Hospitallers and Templars at Gardiki, Lamia and Ravennika in the early thirteenth century, followed by the much later activities of the Hospitallers and Teutonic Knights in Achaia, indicate that the power of the Military Orders in Greece rose whenever the military situation deteriorated. During more peaceful periods, their local castles and estates could still provide the Orders with revenues which helped to fund military activities in other areas.

PART FIVE

The Non-Military Functions of Fortifications

This page intentionally left blank

25

Residences It would be misleading to assume that crusader fortifications were constantly involved in some form of warfare. Although the Christian strongpoints built or occupied in the eastern Mediterranean, like all medieval fortifications, were primarily military structures, it does not follow that their occupants were always engaged in fighting. Clearly there were periods of intense warfare, such as the Third Crusade (1189-92) or the sieges undertaken by Baybars (1260-77), but we should not allow these famous episodes to overshadow the fact that most castle garrisons enjoyed years or even decades of tranquillity between relatively short bursts of conflict. During these more peaceful times, castles had only to cope with the threat of minor raids, piratical attacks or the activities of local criminals. In this sense, crusader castles were no different from similar strongholds all over medieval Europe. Recent research indicates that the more stable and prosperous the twelfthcentury kingdom of Jerusalem became, the more castles, towers and fortified residences were built there by its Latin inhabitants.1 This is clear evidence that strongpoints were not simply built to withstand large-scale invasion forces but were intended to act as agricultural and administrative centres whose owners could protect themselves from the dangers of medieval society. The same argument can be applied to areas beyond the kingdom of Jerusalem and to the period after 1187. For example, in Syria, Cyprus, Cilician Armenia and Greece, many of the castles held by the Military Orders were little more than fortified farmhouses used to administer local estates. We will now take a closer look at day-to-day activities of this kind. For the sake of convenience these have been described as 'non-military' functions, although it will quickly become apparent that they were virtually inseparable from the more overt military activities which have already been discussed. Thanks to the security provided by such structures, castles frequently acted as the permanent homes of local lords. They were therefore provided with numerous features intended to make daily Ufe as pleasant and comfortable as possible. Inevitably these were most common within strongholds belonging to the actual rulers of Latin states in the East, such as the princes of Achaia, whose residence at Chlemoutsi in the north-western Peloponnese

272

UNKNOWN CRUSADER

CASTLES

contained an unusually large number of state rooms equipped with fireplaces, latrines and cisterns, and was built on a far grander scale than any other Prankish castle in Greece.2 To the north, the Prankish, Catalan and Florentine lords of Athens also converted the classical ruins of the acropolis into a palace which was described as 'a great and wondrous work' by a European traveller who saw it in the fourteenth century.3 Similarly, the kings of Cyprus enjoyed palatial accommodation inside the citadels of Kyrenia and Nicosia, whilst in the summer they retreated to the much cooler and healthier mountain fortress of St Hilarión, whose upper baileys housed extensive royal apartments.4 St Hilarion's role as a summer retreat also highlights the need for its living quarters to provide as much respite from the hot climate as possible, for during the crusader period a large gothic vault was added to this fortress, situated at the cliff's edge and deliberately left open to the elements. The sole purpose of this vault appears to have been to give the Lusignan kings a shaded spot where they could sit and enjoy the view, s Similar needs no doubt help explain why some castles were provided with gardens, which supposedly existed on top of the acropolis in Athens and possibly also at Montfort.6 These were no doubt used for relaxation and recreation as well as the cultivation of herbs and vegetables. Finally, it should be noted that Turkish baths, remains of which have been unearthed at Paphos, Belvoir and 'Atlit, were also adopted by the Latins,7 whilst at Beirut, one room within the citadel even contained a marble fountain in the shape of a dragon. This would have provided a decorative centrepiece to the room and at the same time cooled the surrounding air.8 According to Willbrand of Oldenburg, the residential rooms at Beirut also enjoyed extensive views over the Mediterranean and the surrounding countryside, and had marble floors and ceilings designed to represent gently lapping waves and the symbols of the Zodiac.9 In Greece, the mid thirteenth-century castle of Thebes was likewise famous for its magnificent murals depicting the Latin conquest of Syria.10 Although these have long since disappeared, traces of contemporary frescoes have been discovered at the Hospitaller strongholds of Margat and Crac des Chevaliers, as well as the transjordan castle of Kerak.11 These remains suggest that most of the religious or communal rooms of larger crusader fortifications were in fact decorated in this way. References to Latin lords actually living in such rooms on a daily basis are rare and usually coincidental. It only becomes clear, for example, that leading members of the Teutonic Knights used Montfort as a regular residence from incidents such as the resignation of the Master there in 1244 following a lengthy dispute within the Order.12 Similarly, a round chamber

RESIDENCES

2/3

within one of the highest inner towers at Crac des Chevaliers, which was supported by a beautifully decorated gothic vault and lit by two relatively large windows, was probably reserved for the Master of the Hospitallers, even though there are no specific references to it being used for this purpose.13 Such buildings were usually only referred to on special occasions, for castles did not merely act as the residences of those permanently settled in the East, but were also used to accommodate important guests and crusaders. During Louis IX's crusade to Egypt (1249-50), for example, his wife Queen Margaret spent much of her time at 'Atlit, which, according to Oliver of Paderborn, contained an entire 'palace' within its inner bailey.14 Presumably it was here that Queen Margaret stayed, and even gave birth to a new son.15 Following his return from Egypt, Louis himself also visited several key strongholds in the kingdom of Jerusalem, including Jaffa and Sidon, whose outer defences were repaired at this time.16 There are numerous other thirteenth-century examples of important travellers being put up in castles. At the time of the Fifth Crusade, Andrew II of Hungary visited Margat and Crac des Chevaliers, and was so impressed by these strongholds that he gave the Hospitallers certain estates in his homeland to express his gratitude for their generosity.17 About eighty years later, during negotiations over a possible marriage alliance between the principality of Achaia and the duchy of Athens, Isabella of Villehardouin deliberately chose to meet Duke Guy de la Roche and his entourage in an area of Achaia where there were enough castles to accommodate all the various nobles attending this meeting.18 Elsewhere it seems that particularly significant fortresses, including Montfort, whose role as the headquarters of an international Military Order probably guaranteed a steady flow of visitors from Acre and beyond, even had specific accommodation set aside for guests. In order to house these visitors comfortably, the Teutonic Knights constructed a guest hall immediately below the castle at some point between 1229 and 1260. This structure measured approximately forty metres by ten metres, and its vaulting, windows and doorways were clearly the work of highly skilled craftsmen following a typically gothic style. It was impressive and spacious enough to accommodate an entire household within easy reach of Montfort itself.19 The arrival of important visitors and other special occasions such as weddings and coronation ceremonies were often accompanied by much feasting and celebrating, and large, rectangular halls similar to that which has survived at Montfort provided a fitting backdrop for these events. Extensive remains of such halls can still be seen at numerous castles in the eastern Mediterranean, including St Hilarión and Kyrenia on Cyprus, Karytaina and Chlemoutsi in Greece, and Beaufort, Sidon and Crac des Chevaliers in

274

U N K N O W N C R U S A D E R CASTLES

Montfort: the domestic hall near the fortress.

the Holy Land. These generally date from the middle years of the thirteenth century and were all built in a gothic style similar to that used for such structures in the West.20 It is possible to gain some idea of how these halls were used from contemporary descriptions of celebrations following the coronation of Henry II as king of Jerusalem in 1286. These took place in the auberge, a hall owned by the Hospitallers in Acre's suburb of Montmusard, and included re-enactments of Arthurian legends and other stories over a period of fifteen days.21 No doubt the castle of Corinth witnessed similar scenes about twenty years later, when the Achaian ruler Philip of Savoy (1301-6) organised a famous tournament there which lasted for about three weeks and involved virtually all the lords and knights of Prankish Greece.22 Philip clearly enjoyed this form of entertainment, for in 1303 he also met the duke of Athens at the neighbouring castle of Vostitza, so that the latter could do homage to him before both men spent eight days 'feasting and celebrating with their barons and the knights who were with them'.23 Thirteenth- and fourteenth-century castles were increasingly designed to enhance the romantic and heroic image of knights, and formed part of the growing interest in chivalry/The magnificent state rooms which have already been mentioned at Chlemoutsi, for example, were intended to impress visitors and probably helped to establish the reputation of the castle's builder, Geoffrey II of Villehardouin (1228-46), as the epitome of knightly virtue. According to a later source Geoffrey 'constantly maintained eighty knights with golden spurs' at his court, 'whom he gave all that they required besides

RESIDENCES

2J5

their pay'.24 Similarly, in 1252 Joinville wrote that on hearing of Louis IX's imminent arrival, the count of Jaffa adorned his castle with a pennon and a shield bearing his arms at each of the five hundred openings in the battlements, in order to present the king and his French followers with an almost idealised image of a rich and powerful fortress.25 Its castles contributed to the reputation which the eastern Mediterranean enjoyed in the West as an exotic and romantic destination for knights wishing to prove their courage as well as their piety. At other times, castles provided convenient meeting points for more solemn occasions. In 1262, for example, the castle of Nildi, whose location somewhere in Achaia has not yet been identified, played host to the famous Parlement des Dames. At this gathering the wives of lords captured by the Greeks at the battle of Pelagonia (1259), including Prince William II of Villehardouin (1246-78), discussed the ransom terms demanded by the Byzantines for the release of their husbands.26 In 1209 the Latin emperor Henry (1206-16) called a similar assembly of Prankish and Lombard lords at the citadel of Thessaloniki, in order to negotiate a settlement to the succession dispute affecting that city.27 Once again, the large size and impressive architecture of castle halls made them ideally suited to meetings of this kind. Spacious halls and communal rooms were limited to the largest and most important strongholds, but further down the social scale many smaller castles and isolated towers were clearly also regarded as permanent homes. At the Red Tower, a relatively simple twelfth-century building in the vicinity of Caesarea, traces of mosaics and painted plaster have been found in the upper parts of the structure.28 Similar features are virtually unheard of in the isolated towers of Prankish Greece, most of which seem to have been far more primitive in terms of their internal appearance. However, at least one such building is known to have contained a number of residential rooms arranged over several floors, and to have had an oven and a wine press attached to it. Clearly this was a fortified farm house rather than a mere refuge site, and it has justifiably been compared with the Hospitaller complex at Kolossi, another isolated keep from which the Order administered its sugar plantations on Cyprus.29 Returning to Greece, we have seen that many similar towers there were built in open, fertile countryside, implying that they belonged to, and sometimes acted as the day-to-day residences of, Latin settlers. They were therefore situated near rural estates rather than strategic hill tops or lines of communication. Recent field work in south-western Boeotia has confirmed this, and has shown that local Prankish towers 'previously believed to exist in isolation from contemporary settlement' did in fact stand within or close to small farming communities.30

2/6

UNKNOWN CRUSADER

CASTLES

In recent years it has also been suggested that these towers were status symbols, and that their height did not merely give their occupants greater protection, but reflected the wealth of the men who built them.31 The best examples of such structures are to be found on the island of Euboea, where it has been shown that well over fifty towers were constructed by wealthy Lombard, Greek and (perhaps) Venetian landholders, partly in response to late fourteenth- and early fifteenth-century Turkish incursions, but partly also to impress neighbours.32 One of these towers situated in the village of Politika measures approximately six metres square at its base, contains three floors with a crenellated parapet above, and stands to a height of fifteen metres. Its beautifully constructed first-floor entrance, round-headed windows and plaque above the doorway all reflect the aspirations of a wealthy local landholder.33 Politika should also be compared with other medieval towers built in an urban rather than a rural setting. Acre's skyline appears to have been dominated by similar buildings which belonged to different Orders and trading nations constantly trying to construct slightly taller towers than their rivals.34 This suggests that the destruction of Genoa's tower following her defeat by the Venetians and the Pisans in 1258 was intended as an insult to civic pride just as much as a military precaution against future Genoese aggression.35 Beyond the Holy Land itself, reference can also be made to the tower which originally stood on the acropolis in Athens, as well as the sole remaining urban tower in Negroponte, both of which were probably once very similar to surviving structures of this kind in Tuscany.36 Returning to Politika itself, one opening in the upper parts of the tower may originally have led to a wooden balcony.37 This feature may simply have been intended for light and ventilation, but it could also have been used for lordly display, giving physical reality to a landholder's economic and social superiority over his dependants. On a far grander scale, the great hall at Kyrenia had large balconies overlooking the central courtyard of the fortress, where the kings of Cyprus may have presented themselves to their vassals below.3« If this was the case, Politika and Kyrenia may represent eastern Mediterranean equivalents of similar openings in English castles such as Newark and Sherborne, whose role has been compared to that of 'the balcony at Buckingham Palace'.39 The presence of such balconies, particularly in the exterior wall of the tower at Politika, does at any rate suggest that fashion, symbolism and comfort could at times take precedence over military considerations. Hence the strongpoints of the Latin east were often intended to display the wealth and power of local lords, not only to those invited inside castles but to those who lived and worked around them every day.

26

Prisons Strong walls and deep ditches meant that it could be just as difficult to get out of fortresses as it was to break into them, making them ideal for use as prisons. 'Atlit, for example, appears to have been the principal gaol in the East for the entire Templar Order, and the Rule of the Templars records several cases of violent or dishonest brothers being locked up there.1 The Hospitallers likewise used part of their heavily fortified headquarters at Acre for this purpose,2 whilst at Beirut Willbrand of Oldenburg wrote that troublesome citizens were placed in the castle moat.3 This is hardly surprising when one considers that ditches of this kind could be well over ten metres deep, and would therefore have been impossible to climb out of.4 It seems unlikely that prisoners were allowed to roam around freely. They were probably either shackled or locked up in cells similar to that cut into the side of the famous rock-hewn ditch at Saone in Syria.5 At other times, those kept in castles were political prisoners rather than common criminals. Hence the long internal conflict between Bohemond VII of Tripoli and Guy II, lord of Gibelet, came to a dramatic end in 1282 when the latter was finally captured and supposedly left to starve to death in the moat at Nephin near Tripoli.6 Good examples of similar incidents abound from the period following the restoration of King Henry II of Cyprus (1285-1324), who was deposed briefly by his rebellious brother Amaury between 1306 and 1310. After Amaury himself had been murdered, many of his followers were held in the dungeons at Kyrenia, where they were only fed a small amount of bread and water each day, and were forced to share cells barely two metres across. They eventually starved to death.7 Such episodes may have been unusually grim, but the fact that Henry's successors Hugh IV (1324-59) and Peter I (1359-69) both imprisoned their enemies at Kyrenia suggests that this was in fact the most important gaol on Cyprus during the crusader period.8 At other times, neighbouring Buffavento, whose location on a rugged mountain top made it ideal for such a role, was also used to house political opponents, including those followers of Amaury of Tyre who had been lucky enough not to get sent to Kyrenia.9 During the 13805 a knight imprisoned here by King James I (1382-98) even managed to escape by resorting to the classic trick of using a sheet as a

2/8

UNKNOWN CRUSADER

CASTLES

makeshift rope.10 Some years earlier, it seems that Peter I of Cyprus had also intended his new citadel at Nicosia, known as the Margarita Tower, to replace Kyrenia as the largest royal prison in the kingdom. Toward the end of his reign, Peter even forced one of his disobedient nobles to work alongside the slaves excavating the moat of this tower, but this so enraged his other barons that it may well have contributed to the king's subsequent murder. It seems that the Margarita Tower soon came to symbolise Peter's oppressive rule, implying that he had built it to intimidate his Nicosian vassals as much as to defend his capital against the Mamluks and the Genoese.11 Shortly after these events, Leontios Makhairas also referred to one of Hugh IV's enemies being 'cast into the prison of Famagusta', although it is unclear whether this was actually inside the royal citadel or not.12 In Cuidan Armenia, castles were occasionally used to imprison religious rather than political dissenters, for there were frequent arguments over the merits of forging closer links between the Catholic and Armenian Churches in exchange for more western aid against the Muslims. In 1307 or 1308, following the Armenian Catholicos Constantine's decision to accept papal supremacy, riots broke out in Sis orchestrated by 'monks and religious men, priests and deacons, as well as doctors and bishops and many people, both men and women, who refused to accept the use of water in the chalice at mass, and other innovations. The king Oschin, with the agreement of the patriarch and the barons, arrested all of them, imprisoned the doctors in the citadel, and put to death many men and women, and some religious men ,..'13 Apart from criminals or political and religious opponents, prisoners of war were frequently held in castles. In 1262, for example, William II of Villehardouin inflicted such a heavy defeat on Greeks attacking central Achaia that they later had to be distributed to several different neighbouring strongholds.14 After the fall of Acre in 1191, all 500 Christians who had been held in the city were released, while the most important Muslim captives were locked up in one of the towers.15 This tower may have been situated in or near Acre's citadel, for according to Ibn al-Furat this was where Muslims were later imprisoned during the reign of Sultan Baybars (1260-77).16 Prisoners of this kind were subsequently often reduced to the status of slaves, being obliged to work in order to survive. From 1240 onwards about looo Muslim captives helped to reconstruct Saphet in Galilee,17 whilst in the late fourteenth century the Cypriots were happy to employ Genoese prisoners during the fortification of Nicosia, even though they were fellow Christians.18 This suggests that the Latins relied on captured soldiers to make up for their desperate shortage in manpower, although

PRISONS

2/9

Baybars' use of Christian slaves to rebuild the walls of Damascus, as well as his insistence that the Prankish defenders of Arsuf should demolish their own citadel after it had been captured in 1265, indicate that in the thirteenth century this was regarded as normal military practice throughout the eastern Mediterranean.19 Prisoners of noble birth were usually treated with far more respect. Isaac Komnenos, the Greek emperor of Cyprus overthrown by Richard I in 1191, ended up at Margat, where he may have spent his time in one of the residential rooms of the fortress overlooking the Mediterranean.20 Important captives were considered an asset because they could be released in exchange for Latin prisoners or large amounts of money. The Byzantine commander defeated by William of Villehardouin in 1262 did not join his troops in some gloomy dungeon but was sent to Chlemoutsi, where he stayed for a while before being swapped for a Prankish knight captured by the Greeks.21 Some years later Thomas, heir to the despotate of Epiros, was also held hostage at this castle to ensure that his father would honour his alliance with the Franks against the Greeks of Constantinople. Once again, the Chronicle of Morea makes it clear that Thomas was regarded more as a guest than as a prisoner, and no doubt made use of some of the extensive residential quarters already mentioned at Chlemoutsi.22

27

Taxation, Justice and Administration The fact that many castles were either used as prisons, residences or both meant that they often became centres of justice and venues for local courts. In Achaia, article 43 of the Assizes of Romania, the law code of Prankish Greece, suggests that lesser barons all had such courts to deal with day-to-day cases of fighting and stealing. Within the princely domain minor offences of this kind were normally sorted out at the royal cities of Glarentza or Androusa, 'where the lord has a captain to dispense justice'.1 More serious cases were brought before one of a select group of the most powerful barons in Prankish Greece, for only they had the right to exercise 'blood justice', or justice over life and limb. These men also formed the nucleus of the high court, although this institution probably met at Glarentza or the unfortified city of Andravida, rather than a specific castle.2 On Cyprus, it is clear that the royal castellan of Kyrenia presided over the local court of burgesses which served the town below the fortress.3 In the Holy Land a similar court was also run by the Templars at 'Atlit.4 By establishing courts in or near castles, the Franks could store fines collected from criminals securely. Taxes imposed on Greeks, Muslims or western settlers were also normally brought to the nearest Latin stronghold. Indeed, tax collecting appears to have been the primary role of Chàteauneuf, a late thirteenth-century castle situated in the south-eastern Peloponnese, in the frontier region between Prankish-held Achaia and the Byzantine outpost of Mistra. The Chronicle of Morea recorded that this stronghold's estates included: all the villages as far as Arcadia and Old Navarino, which were accustomed to paying taxes to the Greeks of Mistra ... for the Greeks did not hold any other castles in this area at the time. And after Chàteauneuf had been completed, it was agreed by general consent of the barons and nobles and fiefholders who held land in this castellany and had paid taxes to the Greeks, that all the taxes which the Greeks had collected should be given and paid to Chàteauneuf for seven years.5 Chàteauneuf 's military and administrative functions overlapped, for its construction enabled the rulers of Achaia to re-establish control over an exposed frontier region, bring local people whose loyalty had been wavering

TAXATION, JUSTICE AND A D M I N I S T R A T I O N

l8l

back into line and regain a vital source of income which had been lost to the Greeks of Mistra. This example has a close parallel in the Holy Land, for the rebuilding of Saphet in the early 12405 quickly brought the Templar garrison of this castle considerable amounts of money from the 10,000 people living in the surrounding countryside. Many of these people had previously fallen under the control of Damascus.6 Elsewhere in the Holy Land, similar links between castles, internal stability and taxation are revealed in a Hospitaller document dating from 1263, which stated that a Muslim settlement in lower Galilee was refusing to pay its taxes to the Order, because of waning Prankish control in the area. This example can be used to illustrate the fragility of a system which relied on isolated strongpoints rather than superior numbers to suppress a hostile population. However, it also implies that until the reign of Baybars, who systematically pushed the Latins out of the region during the 12605, local people had been paying up on time for decades without ever voicing complaints.7 There are numerous cases of castles in the Holy Land being used to impose order over local people and to collect revenues from them dating back to the twelfth century, such as the fortress built on Mons Glavianus above Beirut during the reign of Baldwin II (1118-31) .8 Thanks to their international structure and far-reaching commitments, the Military Orders provide us with several examples of even greater amounts of money being stored in castles. Most famously, the Templars safeguarded their many financial activities by keeping their main treasury inside their powerful headquarters at Acre, whilst the European wing of their banking system was centred around an equally impregnable tower at the Temple in Paris.9 Judging by John of Joinville's account of Louis IX's stay at Acre in the summer of 1250, the former building was used almost like a modern bank by crusaders who could deposit or withdraw money there as required.10 The funds held in Acre were also relied on by the Order itself, for the Rule of the Templars stipulated that 'when the Master wants to go to Tripoli or Antioch, he can take 3000 bezants or more from the treasury if necessary, to aid the houses there'.11 More locally, the Rule also indicates that other Templar castles and houses had their own stores of money for various military and domestic expenses.12 Similar arrangements clearly existed within major castles belonging to other Orders, such as the Teutonic Knights' headquarters at Montfort.13 Apart from fines, taxes and stores of money, farm revenues and important administrative documents were kept inside castles. At Saphet, for example, the seven towers of the inner bailey housed 'numerous offices for all necessary requirements' and it was presumably from here that the Templars administered neighbouring estates and organised the daily running of their

282

UNKNOWN C R U S A D E R CASTLES

castle.14 Beyond major strongholds such as Saphet, the cultivation of farmland also centred around smaller towers and fortified structures. During the twelfth century, and possibly again after 1192, the tower of Qaqun in the lordship of Caesarea was 'used to enforce the lord of Caesarea's authority over his seigneur/, even though he himself probably rarely visited this site.15 Qaqun was only one of countless minor fortifications which had been built between Acre and the southern fringes of the kingdom of Jerusalem during the twelfth century in order to manage local estates and farmlands.16 Many smaller strongholds beyond the Holy Land itself, such as Hospitaller Kolossi on Cyprus or the numerous Prankish towers in Greece, fulfilled the same function. On a far larger scale, the administrative institutions of entire crusader states could be protected by fortifications. The principal mint of Achaia was situated inside the walls of Glarentza, close to the Villehardouin centres of government at Chlemoutsi and Andravida in the north-western Peloponnese.17 On Cyprus, the secrete, an archive recording royal debts, privileges, rents and other earnings, was incorporated into the partially fortified palace at Nicosia. During the 13905 this important office was moved into the new castle built by King James I, making it far more secure against potential Genoese or Mamluk incursions, and ensuring that James's citadel protected the infrastructure of his kingdom as well as the inhabitants of his capital.18 During the thirteenth century, strongholds in the east belonging to the Military Orders performed the same function, for they were used to administer vast estates both in Europe and the Holy Land. Between 1204 and 1206 Margat played host to a General Chapter of the entire Hospitaller Order, no doubt making use of the spacious chapel and large communal rooms of the fortress.19

28

Agriculture and Daily Life As was the case with taxation and finance, the administrative and agricultural functions of castles had much to do with their wider military role, for the protection afforded by such structures encouraged people to live and work close to them, safe in the knowledge that they could find shelter relatively quickly in times of war. Hence the rebuilding of Saphet brought more than 10,000 people living in 260 surrounding villages under Christian protection. These villages had either been uninhabited or dominated by the Muslims ever since Saladin's conquest of Galilee in 1187. ! The threat posed by Mount Tabor, a Muslim fortress constructed in 1211 in order to harry Christian territories as far west as Acre, was largely eliminated by the completion of 'Atlit five years later. The Templar garrison at this latter fortress brought Muslim raids to an end, opening the entire surrounding region up to Christian farming.2 Mount Tabor itself eventually acted as the focal point of Hospitaller estates stretching as far east as the river Jordan, after the Muslims abandoned the site and the Order acquired it in the mid 12505.3 Travelling pilgrims such as Burchard of Mount Sion and Willbrand of Oldenburg frequently noted how lush and fertile the areas around all these strongholds were, and how intensively they were farmed by local peasants. This applied most to regions nearest the coast, many of which were famous for their wines and partially irrigated by old Roman and Byzantine aqueducts.4 Both Saphet and 'Atlit also lay in areas abundant with woods, meadows, rivers and streams, all of which could be cultivated in safety, and were relied upon to provide fruit, vegetables, herbs, oils, fish, meat and dairy products.5 The construction of a single fortress could therefore revitalise the rural economy of an entire region, although it should be noted that Saphet and 'Atlit were relatively unusual, for lack of men and resources meant that the more ambitious colonisation and castle-building projects of the twelfth century, such as the completion of Montreal beyond the Dead Sea in 1115, were not repeated after nSj.6 Agricultural activities also relied heavily on fortified sites because they could protect crops and produce just as much as farmers and peasants. A document dating from 1257 reveals that the Teutonic Knights used the castle of Mhalia, which lay at the heart of their estates to the north east of Acre, as

284

UNKNOWN CRUSADER CASTLES

a collection point for surrounding farmlands, whilst on Cyprus the Hospitallers may have stored sugar cane in a fortified building next to the tower of Kolossi.7 Another important industry which was often organised around fortifications was the extraction of salt. Salt mined near 'Atlit, for example, was probably brought inside the walls of this fortress for storage, export or consumption.8 Fourteenth-century salt works owned by the Venetians on Corfu and the Hospitallers on Castellorizzo were supervised from nearby towers, where both the salt and the men who extracted it could be sheltered from the ravages of Turks and pirates.9 This arrangement can be compared with that of Recordane, a fortified Hospitaller mill situated on the plain of Acre, defended by a sturdy two-storey tower.10 Whenever a conflict erupted, strongpoints could shelter many other valuable belongings which would normally have remained in the countryside. During one of several mid thirteenth-century clashes between the Greeks of Mistra and the Franks of central Achaia, local peasants took their cattle, produce and anything else they could carry with them inside the nearest stronghold.11 This incident illustrates why frontier castles and strategic lookout posts were so important, for if farmers did not receive adequate warning of an imminent attack they were forced to leave many of their possessions behind. Hence the Cuman and Bulgar invasion of Latin-held northern Greece in 1205 caused so much devastation because the attackers did not simply overrun the area, but 'took the cattle in the countryside' with them when they left.12 Likewise, in January 1374 Genoese invasion forces on Cyprus reached the key northern stronghold of Kyrenia so quickly that they managed to capture the cattle grazing in nearby fields before they could be brought inside the fortress.13 If peasants were not alerted early enough, or were subsequently unable to find shelter for their livestock, the impact on the local economy could be disastrous, and the fact that farmers had themselves escaped death or enslavement became immaterial if their only source of food and income had been destroyed. Castle garrisons were also keen to protect and participate in agricultural activities as they relied on food produced locally just as much as farmers and peasants did. The Hospitallers of Margat, for example, collected more than 500 wagon loads of crops annually from the fertile slopes below the castle, whilst at Saphet fresh fish was delivered daily from the River Jordan and the Sea of Galilee.14 Surplus crops which were not needed by the garrison itself were sold off at market; a policy which provided the castle of Areas near Tripoli with 'considerable revenues, the annual income of its lands coming from imposts, cane and cultivated fields, and amounting to a large sum'.15 Other food which was not needed immediately could be stored for use

A G R I C U L T U R E AND D A I L Y L I F E

285

during the winter or a protracted siege. Margat was supposedly capable of withstanding a five-year blockade, during which time its defenders would presumably have relied on supplies stored in the kind of vast grain silos still extant at Cave de Tyron in the lordship of Sidon.16 Certain types of hardened bread were particularly useful in this respect, for in 1232 Ibelin supporters who had been besieged by their Lombard enemies inside the Hospitaller compound at Nicosia baked a large stock pile of 'biscuits' which they hoped would sustain them through a lengthy blockade.17 Similarly, when the Muslims captured the city of Latakia in 1287, the inhabitants of Tripoli were so afraid of an imminent Mamluk attack that Bohemond VII commandeered all forty bakeries in the city, thereby guaranteeing food supplies for the defenders.18 Other essential items supplied to strongholds from surrounding territories included iron, steel and leather, used to make clothing and armour, plus fodder, needed to feed warhorses and livestock.19 In the Assizes of Romania, one particular clause also stipulated that certain forests were set aside 'to supply the castles', so that their garrisons never ran short of timber or firewood.20 Indeed, the remains of aqueducts at some sites, most notably Crac des Chevaliers and the northern stronghold of Baghras, indicate that even water had to be channelled into castles from nearby streams and springs.21 Once crops, animals and other supplies had been brought inside strongpoints, fortifications were not merely relied on for shelter but actually became integrated with daily agricultural activities. The remains of a feeding trough in the moat surrounding the outer town of 'Atlit confirms that this ditch was used as a corral for cattle during peace time.22 At the neighbouring watchtower of Destroit, mangers hewn out of a nearby rock face also indicate that the Templars normally relied on the walled enclosure surrounding this structure to accommodate livestock.23 According to William of Tyre, the Franks likewise tried to shelter cattle in the moat of Kerak when Saladin besieged the fortress in ii83.24 These examples, combined with instances of prisoners being held in castle ditches, indicate that such features were frequently seen as useful places to keep animals or people who needed to be contained. This did not of course apply to water filled ditches, but such defences could also be adapted for various non-military functions and were often used as open cisterns. A cistern of this kind existed between the inner and outer ramparts along the south face of Crac des Chevaliers, providing the Hospitaller garrison with plenty of water for washing, cooking and (perhaps) drinking.25 Similar reservoirs have been found at several other strongholds, including one at Beaufort which may have been used for washing horses.26

286

UNKNOWN CRUSADER CASTLES

The towers and ramparts of strongpoints were expected to fulfil secondary, non-military functions just as much as their ditches. The elevated position of one of Crac des Chevaliers's outer towers made it an ideal location for the castle's windmill.27 Another mill said to have been located inside the sea castle at Sidon was no doubt driven by stiff sea breezes coming in from the Mediterranean.28 Remains found at Paphos and Montfort suggest that castles which were not so well placed to catch the wind were nevertheless equipped with water or animal-powered mills.29 To strengthen walls and minimise the impact of siege weapons or earthquakes, fortifications were often provided with large vaulted chambers or undercrofts positioned behind ramparts and towers. From excavations carried out at Crac des Chevaliers, it is clear that these and other internal structures served as storerooms, bakeries, and kitchens in peace time.30 The large scale of these facilities is hardly surprising when one considers that the garrison of Crac des Chevaliers, along with similar fortresses such as Margat and Saphet, probably numbered between one and two thousand people, making food preparation a virtually constant occupation.3i This is confirmed by extensive archaeological remains at other castles, including Belvoir, where a large kitchen building equipped with three ovens was unearthed in the ipöos.32 De constructione castri Saphet likewise mentions several ovens, plus a variety of wind-, water- and animal-powered mills, being used to feed the occupants of Saphet, who annually consumed '12,000 mule-loads of barley and corn as well as other food'.33 Mentioned in the Rule of the Templars are several related facilities needed to keep the kitchens themselves running, including gardens which were intended to meet culinary rather than recreational needs.34 Adhering to the Cistercian practices which influenced their Order, the Templars also expected sergeants wishing to become full brothers to perform manual labour 'at the oven, or the mill, or in the kitchen, or with the camels, or in the pigsty'.35 Hearths, pottery remains and various cooking utensils have also been found by archaeologists at numerous smaller strongpoints, such as the early thirteenth-century Prankish castle of Paphos in Cyprus.36 Other rooms within castles were used as workshops rather than kitchens. The Rule of the Templars makes several references to the people who used these facilities, including cobblers at Saphet, and masons, who presumably undertook constant repairs and improvements.37 Saphet also housed workshops producing arms and siege engines, whilst archaeological evidence from Paphos suggests that the dead ground between the different ramparts of castles could be used to store stone ammunition for catapults.38 Documents referring to those parts of Acre's defences held by the Teutonic Knights suggest that these were strengthened by a variety of towers,

A G R I C U L T U R E AND DAILY LIFE

28/

undercrofts and other buildings which provided shelter and flanking fire in wartime but which were normally used for storage and accommodation.39 There are also numerous cases of Baybars keeping prefabricated catapults inside fortresses, which could be brought out, along with specially trained siege engineers and sappers, whenever the sultan intended to attack a Prankish stronghold.40 For most of their existence, therefore, fortified structures were actually used as arsenals and storage points or for other mundane domestic activities, rather than the waging of war. If months or even years could pass between incidents of warfare, it follows that many of those who lived inside strongholds rarely, if ever, got involved in any fighting. Figures which have survived for the castle of Saphet tend to confirm this, for they record that the peace time garrison of approximately 1700 people consisted of 430 knights, sergeants, turcopoles and crossbowmen, plus another 820 staff and 400 slaves who were needed 'for labour and other offices' within the castle.41 In other words, more than 70 per cent of the garrison was primarily employed to carry out repairs, and to staff the kitchens, workshops and various other amenities which have already been referred to, even if they could doubtless assist in the defence of Saphet in an emergency. This would not of course have applied to the slaves living at the castle, some of whom may well have been the same Muslim prisoners of war used in its original construction.42 The rest of the ancillary staff may have been native Christians, for judging by contemporary accounts of the Muslim capture of Saphet in 1266, they also made up most, if not all, of the sergeants and crossbowmen.43 Bearing in mind that virtually all the turcopoles may likewise have been native horsemen employed as mercenaries, it becomes evident that, even if we include the full-time soldiers stationed there, Saphet's garrison only contained a tiny minority who actually came from western Europe.44 Saphet provides us with incomparably detailed figures regarding its garrison, but it seems that a similar situation prevailed inside many other Latin castles. In 1263, for example, the Hospitallers and the Templars rejected a proposed prisoner exchange between the Latins and the Mamluks because 'their [Muslim] slaves were very useful to them, for they were all craftsmen and it would have been very expensive to hire other craftsmen'.45 Although the sources do not explicitly say that these unfortunate people were kept in castles, it seems probable that Saphet was not the only fortress in the Latin east which was constructed and repaired using slave labour. Plenty of circumstantial evidence also exists for the presence of paid staff in castles, who, like their colleagues at Saphet, were probably only expected to fight if they came under direct attack. Indeed, it has been estimated that Crac des Chevaliers, Margat and Mount Tabor had total garrisons of 2000,

288

U N K N O W N CRUSADER CASTLES

looo and 1300 people respectively, but only housed 520, 260 and 340 fighting soldiers each.46 When in 1302 the Muslims captured the island of Rouad, which lay opposite Tortosa and had been occupied by the Templars as a forward base for attacks on the mainland, they also found 120 knights, 500 archers and 400 'men and women of the population'.4? Unlike Saphet, the servants did not, therefore, outnumber the full-time troops, but they were nevertheless extremely numerous for such a small and perhaps otherwise uninhabited island. They clearly also included a large number of women, disproving the impression given by some military historians that castle garrisons were completely male dominated. Similar arrangements appear to have been made in Greece, where the number of western settlers was often as low, if not lower, than in the Holy Land. In the Peloponnese, for example, the fortified town of Nauplion was said to have contained two castles, one of which was granted to the Greeks, whilst the other was occupied by the Franks. This implies that western settlers garrisoned the upper citadel, but relied on locals to man the lower defences of the city.48 Moreover, the Assizes of Romania suggests that Greeks did, at times at least, perform military service to Latin lords.49 When the Greeks of Epiros overran Thessaly in the early i22os, their blockade of Thessaloniki appears to have lasted for several weeks or months. Unless this only applied to the citadel, it seems impossible that Latins living there could have defended the city's vast ramparts for so long without the aid of native citizens or mercenaries.so Incidents of this kind suggest that, like their neighbours further east, many Latins living around the Aegean often had to boost their castle garrisons with locals who acted as servants or part-time soldiers. It is clear that peasants, farmers and castle garrisons were all closely dependant upon each other. Those who lived and worked inside strongpoints relied on neighbouring estates for food, supplies and revenues, but these goods could only be provided if peasants in the countryside felt safe enough to go about their work. This interdependence between peasants and garrisons was extremely important to the Latins; if it broke down, their control over any given region could collapse remarkably quickly. During Sultan Baybars' raid on the county of Tripoli in 1270, his soldiers' horses 'grazed on the meadows and crops of Hisn al-Akrad (Crac des Chevaliers), and this was one of the reasons why it was captured, since its only provision came from its crops and these were all used for pasture by the Muslim troops at this time'.51 Moreover, previous attacks carried out by Baybars during the 12608 had clearly already taken a heavy toll on the area, for in 1268 the Hospitaller Master complained that a lack of troops and resources was making the strain of maintaining this and surrounding fortresses almost unbearable for

A G R I C U L T U R E AND DAILY LIFE

289

the Ordert Another document dating from 1263 also noted that the region around Rafania, which lay just to the north east of Crac des Chevaliers, 'had once been the seat of a bishopric, but is now desolate and abandoned by the Christians'.53 These reports imply that Crac des Chevaliers's storerooms were virtually empty when Baybars returned in 1271 and took the fortress in a mere three weeks.34 Further south, similar circumstances led to the fall of Montfort, whose territories were so eroded by enemy incursions that by 1268 all but ten of its surrounding villages had been lost to the Muslims. As a result, in 1270 Montfort's defenders made a temporary arrangement with the Hospitallers, whereby they would be allowed to grow crops for the coming year on land belonging to this latter Order. Before the year was through, however, Montfort fell to Baybars, who had effectively starved its garrison into submission without even needing to undertake a lengthy sieged5 Beyond the Holy Land, constant raids by the Greeks of Nicaea, who from 1261 onwards were able to use Mistra as a starting point for attacks on Prankish Achaia, left large parts of the principality in a similar state. As a result, William II of Villehardouin made himself a vassal of Charles of Anjou, in the hope that the latter could bring military aid from southern Italy.56 Charles responded by sending arms and supplies from Brindisi, including grain, to various castles in Albania and Greece.57 Clearly, enemy incursions had damaged the economy to such an extent that Latin garrisons could no longer rely on a steady supply of local food, and were therefore incapable of defending themselves properly. About eighty years later further raids by Greeks, Turks and Catalans persuaded the rulers of Naples to grant the castellany of Corinth to Niccolo Acciajuoli, who was likewise expected to repair local castles and bring desperately needed external aid to this highly exposed area of the Peloponnese.58

29

Trade The problems experienced by Crac des Chevaliers, Montfort and several royal castles in Greece help to explain why Latin fortifications with direct access to the sea survived far longer than their neighbours further inland. Thanks to the naval strength of the Italian city states, the Latins remained unrivalled at sea for most of the thirteenth and fourteenth centuries, but on land they found themselves hopelessly outnumbered by the Greeks and the Muslims. Hence inland strongpoints which relied solely on locally grown produce for their income and food supply suffered far more economic damage from land-based enemy incursions than coastal sites such as Acre and Tripoli, whose defences protected lucrative seaborne trade routes, not just vulnerable rural estates. This can be illustrated further by looking at the growth of Famagusta, a Cypriot port which quickly replaced Acre as the most important Latin trading centre in the eastern Mediterranean after the latter city fell to the Muslims in 1291. Famagusta acquired its first real urban defences during this period. According to the sixteenth-century historian Estienne de Lusignan, Henry II of Cyprus 'fortified it and made it resemble Acre in every way, creating a market there, where all the strangers of the Orient arrived, and traded in all sorts of merchandise'.1 This view has been criticised in recent years on the grounds that Henry was simply responding to the growing Mamluk threat, and was not consciously trying to create a replacement for Acre. There seems to be no reason to think, however, that Henry did not have both these concerns in mind, for Famagusta's walls prevented the city from being captured until 1374 but clearly also protected a booming economic centred This is given further credence by the fact that Famagusta's post-1291 expansion occurred at exactly the same time as the decline of Limassol, situated along the southern coast of Cyprus. During the thirteenth century these two cities had not differed much in terms of commercial importance, for Limassol had been used as a staging post for the crusades of Frederick II (1228-29) and Louis IX (1248-49).3 After the fall of Acre, however, when Cyprus became the new front line between Christianity and Islam, Limassol remained unfortified. As a result it was considered far less secure, making it

TRADE

291

unpopular with Italian merchants and causing it to go into a steady economic decline.4 Other examples show that, in order to be prosperous, coastal settlements needed to be fortified. In particular, the walls of Acre and Tyre clearly safeguarded the trading activities as well as the inhabitants of these cities against countless Muslim incursions.5 Further afield, the Achaian city of Glarentza, whose defences protected the principal sea route between Greece and Brindisi, became another boom town from the mid thirteenth century onwards; a period which, significantly, seems to have coincided with the construction of the city's walls. A document dating from 1350 records that by this point Glarentza even had its own set of weights and measures, suggesting that thanks to its defences the city's status as an international trading centre was growing, even though the political situation in the Peloponnese as a whole was deteriorating rapidly at that time.6 This also applies to the heavily fortified Venetian colonies of Modon and Coron, situated at the southern tip of mainland Greece. Both were still important trading centres on the very eve of their capture in 1500, long after the rest of the Morea had been overrun by the Turks.7 Other well-defended ports of commercial importance in this area included Patras and the Greek outpost of Monemvasia, which was even granted special trading privileges by the Byzantine Emperor Andronikos II (1282-1328) .8 The prosperity of these sites was only made possible by the presence of powerful defences, for the history of Limassol after 1291 makes it clear that settlements which were not defended by walls or castles did not do as well. Apart from their far larger city walls, many coastal centres contained numerous smaller fortifications which protected the maritime trading interests of individual nations. The most famous such structures were built at Acre, where the Genoese, Pisans and Venetians all held their own fortified quarters protected by walls and towers dominating the city skyline.9 These defences gave the Italians almost total autonomy and enabled them to pursue their own economic, military and political goals. By the mid thirteenth century royal authority in the kingdom of Jerusalem was so weak that, when war broke out between Venice and Genoa in 1258, the two protagonists could not be prevented from destroying large parts of Acre during several months of intense street fighting.10 This conflict clearly illustrates the dangers of allowing too many privately-held fortifications to be constructed, and it is interesting to note that, more than a century later, the Genoese were still demanding their own fortified base on Cyprus (1373). This implies that the Lusignan kings, having seen how compounds of this kind had eroded central authority in Acre before 1291, had prohibited their construction at Famagusta. Firm action of this kind helps to explain why

292

U N K N O W N CRUSADER CASTLES

Cyprus remained the most peaceful of all the crusader states throughout the thirteenth and fourteenth centuries.11 Large, heavily defended quarters in close proximity to each other, which had only emerged at Acre thanks to its commercial importance and inadequate central government, were not constructed elsewhere. Such bases were either actively discouraged by local rulers, as was the case on Cyprus, or were simply not needed in ports where the volume of trade, and the competition for that trade, was far less intense. In lesser ports, the Italians normally only held much smaller towers or fortified houses, which were simply used to store goods, revenues and administrative records. In 1294, for example, one chronicle referred to a Genoese tower at Limassol, which incidentally confirms that until the late thirteenth century this city still played host to a fair proportion of Cypriot trade.12 Further west, contemporary records indicate that, during the following century, the Venetians (or those living under Venetian protection) held many similar towers on Aegean islands and the coasts of Prankish Greece.13 A well-preserved example of such a tower, which has three floors, roof top battlements and stands to a height of fifteen metres, can be seen near the Euboean coastal settlement of Aliberi, w In Cilician Armenia maritime trade during this period was almost all limited to Corycus and Ayas, whose economic importance rose dramatically during the second half of the thirteenth century, partly because of their trade links with the Mongol Il-Khanate of Persia, and partly because of their status as the only mainland ports still in Christian hands after the fall of Acre. The numerous customs dues and tolls imposed on Italian merchants using Ayas were collected by a representative of the Armenian kings, the captain of customs, whose administration may well have been located in the land castle of this city. The fact that this stronghold was pillaged by Venetian sailors in 1307 also suggests that tolls were stored in it and that it played a major role in the running of the port.15 At some sites, fortifications were also designed in such a way that they controlled the arrival and departure of individual merchants and vessels. During the 12603 Venice and Genoa regularly fought over the Tower of the Flies: its position on a reef in Acre's harbour meant that it dominated the 85 metre wide entrance to the port, and whoever occupied it could therefore control much of the economic life of the city.16 Similarly, the fortified bridge connecting the Aegean island of Euboea with the Greek mainland was divided by a drawbridge, which the Venetians of Negroponte no doubt used to impose tolls on certain vessels, whilst at the same time preventing Genoese or Turkish ships from getting through.17 These observations also apply to land-based trade, for tolls could easily be collected from merchants

TRADE

293

Margat: tower used as a toll station and situated between the fortress and the Mediterranean sea.

as they passed through city gates. In 1266 John of Ibelin, lord of Beirut, decided to give the Order of St Lazarus ten bezants a year from the customs dues imposed on tradesmen entering or leaving his city.18 At Acre access to the royal Court of Chain, which was the main customs house, was likewise guarded by an iron gate regulating the flow of goods between the port and the rest of the city. Rulers of Acre jealously guarded their monopoly over this lucrative area, whilst other trading nations competed fiercely for access to the harbour front. The walls built in and around cities were sometimes intended to keep trading rivals out just as much as political opponents.19 Once merchants had left the safety of the fortified cities and were travelling across the countryside, smaller Prankish strongholds offered them protection against bandits and highwaymen. Possibly the best example of such a structure is that of Destroit, a watch tower in the vicinity of 'Atlit. Destroit was built at a point where the rocky terrain forced the main coastal path into a narrow defile, which could easily be used to ambush travellers.20 The restricted nature of the site made it an ideal place for levying tolls from tradesmen, and the Templars may well have done so in the same way that the Hospitallers did near Margat. Here a wall had been constructed running from the fortress itself 'toward the sea near the port of Margat', so that people travelling between the county of Tripoli and the principality of Antioch were obliged to pass through a small gate and pay a fee before they could continue their journey. This operation was supervised from a tower near the gate, whose occupants carried out the dual task of protecting travellers against robbers, and at the same time raising revenues for the

294

UNKNOWN CRUSADER CASTLES

Hospitallers.21 Further north it is possible that Hasanbeyli, another look-out post which lay along an important route between Antioch, Marash and the Cilician Plain, performed a similar function. An Armenian document dating from 1271 implies that this structure was the Black Tower, a toll station which the Teutonic Knights held near the Cilician fortress of Servantikar.22 This tower was probably lost at the same time as Servantikar, captured by the Muslims in 1299.23

30

The Churchy Pilgrims and Crusaders Like merchants, pilgrims were another group of vulnerable travellers who needed to be protected but who could also be exploited financially. Once again, reference can be made to Destroit in this respect, which had originally been built in the twelfth century 'because of bandits who threatened strangers travelling to Jerusalem'.1 Once it had been completed, 'Atlit itself became a popular spot for visitors to spend the night, for it housed numerous important relics which had been brought from Constantinople in the wake of the Fourth Crusade.2 In the county of Tripoli, the urban defences which the Templars had reconstructed around the coastal town of Tortosa also sheltered many visitors, including John of Joinville, who believed that this was the scene of St Peter's first mass.3 Further inland, the author of De constructione castri Saphet noted that the castle of Saphet enabled pilgrims to visit a number of holy sites near Lake Tiberias, including the spot where the feeding of the 5000 had taken place, indicating that this area had been too dangerous to travel in before Saphet was reconstructed in the early i24os.4 All these strongholds made it possible for the Templars to continue their traditional role of protecting pilgrims travelling in the Holy Land. Apart from sheltering visitors and providing them with accommodation, many strongholds like Saphet could offer medical care for foreigners who fell ill in the harsh local conditions. The Hospitallers, whose original purpose had been to carry out such work, were probably most famous for their infirmaries, which incorporated many of the most up-to-date medical discoveries from the Muslim world. Both the Teutonic Knights and the Templars ran similar hospitals for sick and needy travellers, whilst all three Orders were expected to feed and clothe local paupers as often as they could. Hence castles became focal points for much charitable work involving both native and visiting Christians.5 Urban fortifications and individual strongholds were also expected to safeguard the buildings, assets and infrastructure of the local church. This was perhaps most strikingly evident at the small Prankish coastal town of Banyas (Valania), which remained so exposed to further Muslim incursions following Saladin's invasion of 1188 that its bishop transferred his seat to

296

U N K N O W N CRUSADER CASTLES

the neighbouring castle of Margat. The chapel of this fortress may have become the bishop's new cathedral, serving the needs of the surrounding diocese.6 Alternatively, it is possible that services conducted for the benefit of those who did not belong to the actual garrison took place in another building located in the castle town, for at Crac des Chevaliers it appears that local peasants used a small church situated just outside the main entrance rather than the much more grandiose chapel of the fortress itself.7 In Syria, the Latin patriarchs of Antioch also managed to outlive the Mamluk conquest of the principality in 1268 by taking shelter inside the fortress of Cursat, which was situated in the rugged interior and was not finally lost to Baybars until 1275.8 This castle had presumably been chosen as the principal residence of the patriarchs, and the depository of their treasure, because of its strength and inaccessible location. In the very same year that Cursat fell, Mamluk pressure forced the catholicos, or head of the Armenian Church, to retreat from the exposed Cuidan Plain to the impregnable stronghold of Vagha, located high in the Anti-Taurus mountains and eventually the home of all the Armenians' most important relics.9 In the neighbouring Taurus chain, the equally inaccessible fortress of Lampron became an important centre for Armenian miniaturist painting in the medieval period, and therefore helped to preserve local culture despite the deteriorating military climate.10 By taking such precautions, both the Franks and the Armenians hoped that their holiest possessions would not share the same fate as the undefended church of Nazareth, which was demolished by Baybars in 1263 in an effort to undermine Christian morale.11 Several Prankish castles in the Holy Land were themselves thought to be extremely important in the fight against Islam. By garrisoning Mount Tabor, the Hospitallers were not merely occupying a strategic vantage point in central Galilee but were also defending the supposed scene of Christ's Transfiguration.12 At Saphet it was considered equally important that the new Templar castle stood on the ruins of a mosque and a synagogue.13 Gregory IX was referring to the religious as well as the military importance of Montfort when he spoke of its proximity to the Muslims and its vital contribution to the defence of the Holy Land. Crusader fortifications could take on great spiritual significance in this way as the furthest outposts of Christendom, particularly when they were guarded by the Hospitallers, Templars or Teutonic Knights.14 The members of these Orders were all monks as well as warriors, and their fortresses acted as monasteries, not just places of war. When 'Atlit was built at the time of the Fifth Crusade, for example, it was seen as a perfect retreat for the Templars from the 'sinful' city of Acre. It must have been regarded as a most suitable base

THE C H U R C H , P I L G R I M S AND CRUSADERS

297

for an Order which had originally been heavily influenced by the reclusive Cistercians.15 The monastic role of castles belonging to the Orders was also reflected in their architecture, for they were normally provided with beautiful chapels which the brethren could use for their daily services. Well-preserved examples of such buildings can be found at Margat, Crac des Chevaliers and Chastel Blanc, where the thirty metre long chapel of the Templars formed the lower floor of the keep itself.16 At 'Atlit, and possibly Saphet, there were also round churches of the type normally associated with the Templars, although these are likelier to have been inspired by the Constantinian rotunda within the Holy Sepulchre rather than the Temple of the Lord, as was once thought.17 At Crac des Chevaliers the Hospitallers also added an intricate gothic arcade to the central hall of the castle to act as a monastic cloister. It has even been suggested that by designing several of their earlier castles, including Belmont in Judaea and Belvoir near Lake Tiberias, around a cloister-type central courtyard, the Hospitallers contributed to the development of concentric fortifications almost by accident, for an isolated central keep clearly obstructed the traditional monastic layout which this Order hoped to achieve.18 Extensive archaeological research carried out at 'Atlit has confirmed that the many vaulted halls built around the inner bailey of this stronghold also corresponded 'to the dormitory, refectory, infirmary, and guest apartments of a monastery'.19 After the Templars acquired Beaufort and Sidon, they constructed large, communal halls at both sites, so that these formerly baronial castles were better suited to the daily life of monastic communities.20 By building their own chapels within their strongholds, the Military Orders also excluded themselves from the authority of the local clergy. The papacy encouraged this process by granting them a series of privileges, including the right to appoint chaplain brothers for the knights' spiritual needs, and sweeping exemptions from the payment of tithes. As a result, their castles gave the Hospitallers, Templars and Teutonic Knights ecclesiastical as well as military independence, for in theory at least they were answerable only to the pope. It is hardly surprising that Latin churchmen in the East, whose income from tithes could drop sharply as a consequence of the growth in estates belonging to the Military Orders, were vocal in calling for their privileges to be withdrawn.21 This must have been a complex issue, as it has already been shown that some members of the clergy, such as the bishops of Banyas, relied on fortifications held by the Orders to protect them against the Muslims. By the late thirteenth century this bishopric had become almost entirely dominated by the Hospitallers, who effectively treated it as part of the lordship of Margat.22 A few miles down the coast,

298

U N K N O W N C R U S A D E R CASTLES

the same dilemma confronted the bishops of Tortosa, for vast Templar and Hospitaller estates around Crac des Chevaliers, Chastel Blanc and Tortosa itself had left them in charge of a mere handful of local churches, yet their very own cathedral lay behind ramparts garrisoned by the Order of the Temple.23 Although the Hospitallers, Templars and Teutonic Knights were envied for their wealth and power, it was clear to everybody that their castles were sorely needed to defend Christian territories. Such structures were so important that it became popular for crusaders who lacked the troops and resources to attack the Muslims to spend their time in the East constructing and financing new Latin fortifications. The thirteenth-century sea castle at Sidon was built in the late 12205 by European crusaders waiting for Frederick II to arrive from the west, whilst 'Atlit was also known as Pilgrims' Castle because members of the Fifth Crusade largely paid for its construction.24 Both Louis IX and Richard I also took part in castle building, as it was another way of expressing their piety once further campaigns against the Muslims had become impracticable. In 1252 Louis IX specifically helped his men to construct a new citadel at Jaffa 'to earn his indulgence'.25 Ultimately, Latin fortifications became linked with the religious ideas which underpinned the entire crusading movement, for contemporaries did not distinguish between the more down-to-earth military functions of such buildings, and their highly symbolic role as fortified monasteries and outposts of Christianity.

3l

Conclusion We have seen that each of the four regions which have been considered all had certain architectural characteristics. In Cilician Armenia, the horseshoe towers and sinuous curtain walls of mountain castles like Servantikar and Yilan represented a distinct Armenian style, which was closely copied by the Hospitallers at Silifke, Margat and Crac des Chevaliers, but which differed from that of most other Prankish castles in certain respects. In Greece the ready availability of earlier Byzantine or classical fortifications meant that there were few truly Latin strongpoints; and even, when western settlers did build new fortifications, these tended to be of a low quality and difficult to distinguish from pre-1204 structures. A similar situation appears to have prevailed on Cyprus, although it has been shown that major new defences which were probably built to a much higher standard were erected at Famagusta, Kyrenia and Nicosia after 1291. Meanwhile, in the Holy Land the most substantial and sophisticated Latin fortifications of the entire period were built at sites like Acre, Tyre, Tortosa and 'Atlit. From this it might seem that the Franks who lived in the Holy Land were technologically superior to Latins in other regions. Indeed, as castles like Crac des Chevaliers grew in complexity, it is tempting to suggest that a steady chronological advancement took place, whereby older, simpler designs were gradually rejected in favour of more sophisticated ones which were better able to withstand Muslim siege weapons. This argument can be applied to major Syrian strongholds such as Crac des Chevaliers, but to view military architecture in general as a precise science which unswervingly moved forward because of new technological discoveries tends to oversimplify the evidence. For example, it has been shown that even though the castrum-type castle consisting of a straightforward rectangular enclosure was an extremely simple design, it was not abandoned after the twelfth century, as some have suggested, but turned up again in the late fourteenth century, when lames I of Cyprus built such a structure at Sigouri. Similarly, it would be wrong to assume that, after they started to build massive concentric fortifications such as 'Atlit, the Military Orders no longer constructed castles with an isolated central keep, for the Teutonic Knights did just that when they erected their new headquarters at Montfort during the late 12205. The

300

U N K N O W N CRUSADER CASTLES

notion of steady technological advance seems even more misplaced when applied to smaller fortifications. Countless towers were erected throughout the eastern Mediterranean, including Greece, Cyprus and the Holy Land. Although their masonry and internal accommodation were often poorer in Greece, the general size and layout of these towers seems to have been remarkably similar in all areas, for they were usually square, with each wall measuring approximately ten to fifteen metres, and they normally contained two vaulted storeys. There was no dramatic change or progress in their design, even though they continued to be built from the twelfth to the fifteenth centuries and possibly beyond. In reality, military architecture should not be regarded as a separate science removed from the rest of medieval society, for the appearance of fortifications was greatly influenced by wider political and economic factors. People could already recognise the benefits of building concentric strongholds such as Crac des Chevaliers or using powerful siege weapons such as counterweight trebuchets long before 1187, but inadequate resources prevented them from doing so. Both the Latins and the Muslims had to wait until they possessed the political authority and the administrative institutions needed to raise the money which could fund the construction of better castles and war machines. In other words, it was a lack of readily available cash rather than any kind of scientific ignorance which held up progress, for complex fortifications such as the double land walls of Constantinople had already been in existence for centuries and must have been well known both in western Europe and in the Muslim world. The appearance of sites like 'Atlit or Crac des Chevaliers owed more to the rapidly increasing income of institutions such as the French monarchy, the Egyptian sultanate and the Military Orders than it did to any new technological discoveries. The reverse of this argument can be applied to Greece (and to some extent to Cyprus before 1291). None of the local political powers, including the revived Byzantine empire after 1261, had as much money to spend on catapults or as many soldiers to place in field armies as the thirteenth-century sultans of Egypt did, and as a result there was no need to build any castles like Crac des Chevaliers. In short, the Christians and Muslims who confronted each other in the Holy Land were not technologically superior to the protagonists in Greece, they just had more money. Numerous other factors determined the appearance of Christian strongholds. Castle builders also had to consider the various day-to-day functions of such structures. The concentric appearance of fortresses like 'Atlit and Belvoir, for example, reflected the need for castles belonging to the Military Orders to have monastic cloisters within them. Elsewhere the local terrain played a significant part. Numerous castles in Greece, Cilicia and Cyprus

CONCLUSION

3O1

did not even need complete walls around them because they were situated on precipitous hill tops, whereas coastal sites such as Tortosa required huge ramparts to protect them from land attackers. Once again these considerations make it dangerous to try to suggest that military architecture progressed in a straight chronological line. For example, some scholars have implied that the castle of Montfort, completed in the late 12205, was not as sophisticated as Saphet, rebuilt in the early 12405. Whilst this may be true on purely architectural grounds, it does not prove that the Franks had become technologically more advanced because fifteen years had passed since the completion of Montfort. In fact, Montfort may have been stronger that Saphet in some ways, and simply required fewer man-made defences because it was situated on a remote mountain spur. Thus the military architecture of the eastern Mediterranean could vary considerably according to local conditions and to local needs. The Latins and the Armenians often found themselves heavily outnumbered by their opponents. In the Holy Land, internal squabbles amongst the Muslims and the peace treaties with the Christians which resulted from them meant that this did not become a serious problem until the 12405, and even more so after 1260. Between that date and 1291 the crusader states were wiped out by Mamluk armies which could be more than ten times as large as those of the Franks. Until 1375 a similar process took place in Cilician Armenia, where the Armenians had managed to hold their own against the Seljuks but were far too weak to stop the Mamluks (and sometimes the Mongols). After 1291 the meagre Christian forces on Cyprus also had to be on the alert for a Mamluk invasion; a threat which had not really existed before the fall of Acre. Meanwhile around the Aegean none of the Greek or Bulgar armies which confronted the Latins seem to have been as large or as skilled in siege warfare as those of the Egyptians, but they could still outnumber the western newcomers if they took up arms at the same time. The kingdom of Thessaly and the Latin empire were quickly wiped out by the simultaneous efforts of the Greeks in Nicaea and Epiros, plus the Bulgars and the Cumans. In south-western Greece the protagonists were more evenly matched, but here too the Latins were eventually ground down first by their Greek enemies and then more decisively by the Ottoman Turks. Close comparisons can be made between Cyprus, Greece, Cilician Armenia and the Holy Land. In this last area we have seen that major strongpoints such as Acre and Saphet were so powerful that even though they only had garrisons of one or two thousand Christians, they could successfully withstand tens of thousands of Muslim attackers. Although the defenders of such sites had to cope with the destruction of the surrounding countryside,

3O2

UNKNOWN CRUSADER

CASTLES

they could wait behind their defences until inadequate supplies, lack of progress or bad weather forced the Muslims to retreat. In this way fortifications made up for the lack of soldiers in the field and enabled the heavily outnumbered Christians to remain in the Holy Land for most of the thirteenth century. Exactly the same tactics were used in many other parts of the eastern Mediterranean. For example, even though the number of combatants involved was far smaller, the Catalans successfully defended the duchy of Athens against Gautier II of Brienne's invasion of 1331 by retreating inside their castles. Strong fortifications (especially around Constantinople) also explain why the Latin empire survived for as long as it did. Elsewhere they enabled the Armenians to retain their independence until 1375 and prevented the Genoese from conquering any Cypriot territories beyond Famagusta. Famous episodes such as the successful Christian defence of Tyre against Saladin in 1187 had many less well-known parallels throughout the thirteenth and fourteenth centuries. The written sources also reveal two other important facts regarding the role of Christian fortifications during hostile incursions. First, it is clear that there were no nationwide networks of castles which could all communicate with each other using fire or smoke signals. Other scholars have rejected this argument when discussing the Latin towers built in Greece, and it should also be rejected with regard to the county of Tripoli and Cilician Armenia. This does not rule out the possibility that much smaller groups of strongpoints were linked in this way, as was clearly the case on Rhodes and on Cyprus. But in other regions, and particularly Cilician Armenia, a combination of political, logistical and architectural factors would have made it impossible for an early warning system of this kind to have extended over many hundreds of miles. Secondly, and also regarding Cilician Armenia, it has been shown that the Armenians were not radically different from the Latins when it came to constructing urban fortifications. Far from turning their backs on city life and preferring to dwell in or near mountain castles, they continued to inhabit a number of urban settlements which were defended by walls or citadels. Throughout the eastern Mediterranean, therefore, there were many Christian communities which relied on city ramparts to protect them against external attackers, regardless of whether they were Latin or Armenian. Similarities can also be identified between the Holy Land and other areas when it comes to the attacking role of fortifications. This is most apparent in Prankish Greece, where we have seen that the Latins used strongpoints to consolidate new territorial gains, to launch punitive raids against opponents such as the Bulgars and to blockade Greek strongholds like Corinth. These activities were very similar to those undertaken by the Franks in the

CONCLUSION

3O3

Holy Land, where Saphet, Crac des Chevaliers and numerous similar fortresses enabled the outnumbered Christians to extend their authority over large tracts of the surrounding countryside. In both these regions the Latins' lack of troops meant that there was also a close interdependence between castles and field armies. Castles needed external assistance if they came under siege, but at the same time Christian soldiers operating in the field relied on neighbouring strongholds for food, water and shelter. During the last one hundred years scholars' thoughts on the defensive role of crusader fortifications in the Holy Land have changed dramatically. Once it was argued that virtually every local stronghold, ranging in size from Crac des Chevaliers to isolated watch towers, formed part of a vast network of castles intended to protect Christian lands from external invaders. But more recent research has shown that we need to distinguish between very large fortresses, like Belvoir or Saphet, and the vast majority of smaller Prankish castles, towers or fortified farmsteads. These latter structures were simply intended to protect the local agricultural activities of individual Christian lords. They had nothing to do with external security on a nationwide scale and were no different from the thousands of strongholds erected all over Europe during the middle ages. Most current scholarship has focused on the kingdom of Jerusalem in the twelfth century, but in this book it has been suggested that the same arguments can be applied to the other crusader states in the Holy Land and to the period after 1187. We have seen, for example, that the fortress of Goliath in the county of Tripoli was sporadically destroyed by major Muslim incursions, but that for most of its thirteenthcentury existence it lay in a relatively peaceful area where it was only expected to safeguard neighbouring Hospitaller estates against much smaller Muslim raids or the activities of local criminals. For certain parts of their history it is clear that many fortifications beyond the Holy Land were equally unaffected by full-scale warfare. Before 1291 Cyprus was so untroubled by the threat of an external invasion that its Prankish rulers did not bother to construct any large new castles. Nor was there any need for them to allow their leading vassals to do so. After 1291 fear of a Mamluk invasion led to a change in policy, but as it turned out no major attackers turned up until the Genoese invasion of 1374. Consequently, with the exception of sporadic internal clashes such as the civil war of 1229-33, Cypriot fortifications hardly ever faced anything more serious than occasional raids by pirates. Furthermore, if we apply the argument that more castles were actually built in peace time to Prankish Greece, it becomes apparent that this region also enjoyed certain periods of calm and prosperity. Western settlers would not have had the time or the money to erect the numerous towers which have been mentioned on Euboea, in the duchy of

3O4

UNKNOWN CRUSADER CASTLES

Athens and elsewhere if these areas had been affected by constant warfare. Although the evidence is not so compelling for Cilician Armenia, certain features, such as the local castles, estates and toll stations belonging to the Teutonic Knights, suggest that this was another prosperous agricultural region before the devastating Mamluk incursions of the late thirteenth and fourteenth centuries. Indeed, in an effort to dispel the misconception that castles were only ever used for fighting or that the crusader states were in a permanent state of war, the last part of this book was devoted in its entirety to the many non-military functions of fortifications. Having looked at strongpoints throughout the eastern Mediterranean, it is also possible to identify certain common strands regarding their role in suppressing local populations. Greek or Muslim rebels normally lacked the strength to capture even relatively small Latin strongholds. Structures of this kind enabled western settlers to survive numerous native uprisings, including the Cypriot rebellion of 1192 and the failed Muslim attack on the Tower of David in 1229. As a result, the mere presence of Latin fortifications often prevented the outbreak of violence. Insurrections were usually limited to periods when the Latins had been defeated in battle or when rebels were able to gain external assistance. Local Muslims took up arms after the battle of Hattin, and the Greeks living around Mistra were frequently in a state of war with the Franks of Achaia after this castle had passed under Byzantine control in 1262. When the Latins were stronger, however, the threat of major rebellions receded. It should again be stressed that some crusader states were not nearly as violent as one might think: the notion that Franks living in the Holy Land hardly ever dared to venture outside their fortified strongpoints for fear of being attacked by Muslim peasants has been exaggerated in the past. Although they helped to maintain law and order, it is clear that fortifications also played a prominent role in the many internal clashes which took place between fellow westerners. The problems which Frederick II had in trying (and failing) to get the Templars of 'Atlit to submit to his rule indicate that large Christian strongholds could be used by their owners to defy the will of local rulers. Beyond the Holy Land comparable incidents took place on Cyprus during the civil war of 1229-33 and in fourteenthcentury Greece. The most chronic difficulties of this kind affected the Armenian kings of Ciucia, who never really succeeded in imposing their full authority over the lords of impregnable castles like Lampron. On the other hand, the fact that most rulers possessed far more lands and castles than any of their vassals could also strengthen their rule, a situation which was most apparent on Cyprus. Indeed, the relatively calm history of Cyprus, which was of course physically isolated, suggests that central

CONCLUSION

3O5

authority was weakest in those areas which were most exposed to external attackers. In the Holy Land Christian rulers had to allow the barons and the Military Orders to build massive castles so that they could protect themselves against the Muslims, despite the threat that this posed to royal power. Similarly, by the late fourteenth century Achaia was under so much pressure from Greek, Catalan or Turkish aggression that royal government disappeared completely for a while and the entire region was leased to the Hospitallers. Fortifications could either enhance or undermine internal stability, depending on local conditions. In Cyprus, Greece and the Holy Land there was a very strong interdependence between coastal fortifications and Christian shipping. In the same way that field armies needed friendly castles to act as refuge points, so fortified ports like Acre, Famagusta and Glarentza provided Latin vessels with safe anchorages where they could conduct their trade and replenish their stocks of food and water. It has been written that in the Holy Land maritime links with western Europe were 'absolutely essential for the continued survival of the crusader states' because they provided the constant stream of troops and resources needed to sustain the war with the Muslims.1 Clearly this was also the case for Latin territories in Greece and on Cyprus, whilst the Armenians' growing links with the West were equally dependent upon the sea. Cyprus acted as a safe base for numerous crusades of the period, including those of Frederick II and Louis IX. In Greece we have seen that Glarentza provided a life line between the Peloponnese and Brindisi, and that for much of its existence the Latin empire was effectively reduced to Constantinople, a fortified haven which could only be reached safely by sea. Throughout the eastern Mediterranean virtually all Latin territories were eventually reduced to coastal outposts of this kind, and after the fall of Acre in 1291 it was only by fortifying ports like Smyrna that the crusading movement could sustain any kind of presence at all on the Muslim mainland. As far as the Christians were concerned, the only negative aspect of this naval activity was that rival maritime powers such as the Genoese, the Venetians and the Catalans relied on heavily defended ports like Tyre and Gallipoli to wreak havoc or wage war against each other. In the period before 1291, all three Military Orders were most active in the Holy Land as an inevitable consequence of their spiritual goals. They had the resources needed to build and guard fortifications which were large enough to keep up with the growing Muslim threat. From the 12505 onwards they took over more and more former baronial strongpoints, such as Sidon and Arsuf. Castles which were situated on the frontiers and were therefore most exposed to Muslim attacks were likeliest to pass to the Hospitallers, Templars or Teutonic Knights. These included Ascalon, Saphet, Beaufort,

3O6

U N K N O W N C R U S A D E R CASTLES

Crac des Chevaliers and Baghras. During the desperate final years of Christian rule coastal strongholds such as Margat, Tortosa and 'Atlit effectively also became frontier outposts. Even though this high level of castle ownership could sometimes put the Orders beyond the reach of local rulers and therefore undermine central authority, it is clear that their contribution considerably lengthened the life span of the crusader states. Despite the huge commitment which the Orders made to the Holy Land, they were already prepared to undertake a limited number of similar projects elsewhere in the Mediterranean before 1291. In Cilician Armenia the castles of Silifke and Harunia, held by the Hospitallers and the Teutonic Knights respectively, were both frontier strongholds intended to bolster this region's defences against the Seljuk Turks. Silifke was particularly interesting, because at one time this little-known fortress must have rivalled the much more famous Hospitaller castle of Crac des Chevaliers, both in terms of its appearance and its local status. However, a combination of financial, military and political pressures forced the Order to give up Silifke, whilst in the end Harunia probably contributed very little to the struggle with the Turks. In Greece the Templar frontier castles of Lamia and Ravennika, along with the neighbouring Hospitaller-occupied fortress of Gardiki, had a very ephemeral existence before they were swallowed up by the Greeks of Epiros. Furthermore, it has been argued in this book that the Military Orders were not granted any Cypriot castles after 1192 which were specifically intended to guard the island's frontiers. It has been suggested that the castle of Paphos belonged to the king rather than to the Hospitallers, and that the Templars simply used Gastria to administer their local estates rather than to protect eastern Cyprus against a threatened Byzantine invasion. The year 1291 clearly marked a dramatic turning point for all three Orders. Cyprus became a base for the late thirteenth- and early fourteenth-century efforts by the Hospitallers and the Templars to regain a toe hold in the Holy Land. During the next hundred years the Hospitallers were involved in the sporadic but ultimately failed projects to use Cilician Armenia as a starting point for the reconquest of Syria. The fall of Acre also changed the Military Orders' attitude toward the Peloponnese. Their presence there had been minimal during the previous eighty years, but in the fourteenth century both the Teutonic Knights and the Hospitallers participated in the doomed struggle against the Greeks, the Catalans and the Ottoman Turks. During these clashes the Orders' local castles, which had originally been situated far away from any exposed frontiers, suddenly found themselves in the firing line and therefore took on a similar role to that of earlier frontier castles like Crac des Chevaliers, Harunia or Lamia. The Hospitallers' acquisition of the entire principality of Achaia between 1376 and 1381 can be

CONCLUSION

307

compared with their dominance in the county of Tripoli after 1188. On both occasions a single Military Order virtually guaranteed the continued existence of an entire crusader state. Many castles belonging to the Military Orders, and in particular those which were much smaller than places like Crac des Chevaliers, actually spent most of their existence being used as administrative or agricultural centres rather than as strategic military outposts. Kolossi and Gastria in Cyprus, Amouda and the Black Tower in Cilicia, Goliath in Syria or Palaiopolis and Laffustan in Greece may all have been attacked by pirates or raiders occasionally, but for the rest of the time their primary function was to administer surrounding estates. These estates were an extremely important source of income, which could be used to maintain larger castles or pay the wages of mercenary troops. We have seen that the Templars were just as anxious to get Baghras back from Leon II because of the value of its agricultural estates as they were because of its strategic significance. It is a mistake to isolate a stronghold's more obvious military functions from its wider role as a focal point for the local rural economy. Taking the history of the eastern Mediterranean between 1187 and 1380 as a whole, the overriding pattern was one of gradual retreat. By the late fourteenth century Cuidan Armenian and the Holy Land had been lost and most Latin territories in Greece had been reconquered by the Byzantine empire, which was itself being overrun by the Ottoman Turks. Eventually the Latins only controlled the islands, until they too were gradually captured by the Turks between the fifteenth and seventeenth centuries. The basic reason for this failure was a lack of troops, but for many centuries the Latins were able to compensate for this shortfall by building or occupying fortifications. There can be little doubt that powerful castles like 'Atlit enabled the Christians to extend the lifespan of their territories without having to match their opponents man for man. As a consequence, many of the strongpoints which have been discussed in this book became essential to the continued existence of the entire crusading movement. In the long run, this problem could only have been solved by encouraging more westerners to settle in the eastern Mediterranean permanently, for, as Baybars observed, 'towns are not guarded by walls, nor are citizens protected by trenches, but by swords together with resolution'.2

Abbreviations ABSA BZ CHSB EHR HC MGH SS PEQ PG PL

QDAP RHC RHCArm RHCLois RHCOc RHCOr RHGF RIS RIS n.s. ROL RRH RS

Annual of the British School at Athens Byzantinische Zeitschrift Corpus scriptorum historiae Byzantinae, ed. B. G. Niebuhr, I. Bekker et al., 50 vols (Bonn, 1828-97) English Historical Review A History of the Crusades, general ed. K. M. Setton, 6 vols (Madison, Wisconsin, 1955-89) Monumenta Germaniae historica: scriptores, ed. G. H. Pertz et al., 34 vols (Hanover, 1826-1980) Palestine Exploration Quarterly Patrologiae cursus completus: series graeco-latina, ed. J. P. Migne, 161 vols in 166 parts (Paris, 1857-66) Patrologiae cursus completus: series latina, ed. J. P. Migne, 221 vols (Paris, 1844-64) Quarterly of the Department of Antiquities in Palestine Recueil des historiens des croisades, ed. Académie des Inscriptions et Belles-Lettres (Paris, 1841-1906) RHC Documents arméniens, 2 vols (Paris, 1869-1906) RHC Lois: Les Assises de Jerusalem, 2 vols (Paris, 1841-43) RHC Historiens occidentaux, 5 vols (Paris, 1844-95) RHC Historiens orientaux, 5 vols (Paris, 1872-1906) Recueil des historiens des Gaules et de la France, ed. M. Bouquet et al., 24 vols (Paris, 1737-1904) Rerum Italicarum scriptores, ed. L. A. Muratori, 25 vols (Milan, 1723-51) RIS nova series, ed. G. Carducci et al. (Bologna, 1900- ) Revue de l'Orient latin Regesta regni Hierosolymitani, 1097-1291, ed. R. Röhricht (Innsbruck, 1893) Rolls Series: Rerum Britannicarum Medii Aevi Scriptores (London, 1858-96)

Notes Notes to Introduction 1. R. C. Smail, Crusading Warfare, 1097-1193, second edition with an introduction by C. Marshall (Cambridge, 1995). 2. C. Marshall, Warfare in the Latin East, 1192-1291 (Cambridge, 1992). 3. D. Pringle, Secular Buildings in the Crusader Kingdom of Jerusalem: An Archaeological Gazetteer (Cambridge, 1997). 4. Cyprus: C. Enlart, L'art gothique et de la Renaissance en Chypre, 2 vols (Paris, 1899). Cilician Armenia: R. W. Edwards, The Fortifications of Armenian Cilicio, Dumbarton Oaks Studies, 23 (Washington, DC, 1987); H. Hellenkemper, Burgen der Kreuzritterzeit in der Grafschaft Edessa und im Königreich Kleinarmenien (Bonn, 1976). Greece: A.Bon, La Morée franque: recherches historiques, topographiques et archéologiques sur la principante d'Acaïe, 1204-1430, with a volume of plans and plates (Paris, 1969). See also W. Müller-Wiener, Castles of the Crusaders, trans. J. Brownjohn (London, 1966). 5. See below, pp. 161-67.

Notes to Chapter i, The Holy Land: Warfare 1. For an outline of Saladin's conquests of 1187-90 and the Third Crusade of 1189-92, see M. C. Lyons and D. E. P. Jackson, Saladin: The Politics of Holy War (Cambridge, 1982), pp. 255-363; H. E. Mayer, The Crusades, second edition (Oxford, 1988), pp. 134-51; S. Painter, 'The Third Crusade: Richard the Lionhearted and Philip Augustus', in HC, vol. 2, pp. 45-86; E. N. Johnson, 'The Crusades of Frederick Barbarossa and Henry VI', in HC, vol. 2, pp. 87-116. 2. 'L'estoire de Eracles empereur et la conqueste de la Terre d'Outremer', in RHCOc, voi. 2, pp. 227-28 (hereafter 'Eracles'); Arnold of Lübeck, 'Chronica Slavorum', in MGH SS, voi. 21, pp. 205—6; Abu Shama, 'Le livre des deux jardins', in RHCOr, vol. 5, pp. 116-17; ïbn al-Athir, 'Kamel-Altevarykh', in RHCOr, vol. 2, pp. 86-87. 3. 'Eracles', pp. 224-26. 4. Ibn al-Athir, 'Kamel Altevarykh', RHCOr, vol. 2, p. 96; Abu'1-Fida, 'Ármales', in RHCOr, vol. i, p. 83. 5. La continuation de Guillaume de Tyr (1184—1197), ed. M. R. Morgan (Paris, 1982), pp. 191-93 (hereafter Coni. WT); Arnold of Lübeck, 'Chronica Slavorum',

310

NOTES TO PAGES 3-4

p. 204; Abu Shama, 'Deux jardins', RHCOr, vol. 5, pp. 116, 152; Ibn al-Athir, 'Kamel Altevarykh', RHCOr, vol. 2, pp. 84-86; Abu'1-Fida, 'Annales', p. 74. 6. See below, pp. 13-14, 69. 7. 'Eracles', pp. 369-74; Matthew Paris, Chronica malora, ed. H. R. Luard, RS, 57, 7 vols (London, 1872-83), vol. 3, pp. 159-61,172-77; Abu Shama, 'Deux jardins', RHCOr, vol. 5, p. 186; Ibn al-Athir, 'Kamel Altevarykh', RHCOr, vol. 2, pp. 17576; al-'Ayni, 'Le collier de perles', in RHCOr, vol. 2, pp. 187-94; al-Makin Ibn al-'Amid, Chronique des Ayyubides, 602-658/1205-6-1259-60, trans. A-M. Eddé and F. Micheau (Paris, 1994), p. 41. 8. 'Eracles', pp. 416-19; 'Continuation de Guillaume de Tyr de 1229 a 1261, dite du manuscrit de Rothelin', in RHC Oc, vol. 2, p. 554 (hereafter 'Rothelin'); Matthew Paris, Chronica malora, vol. 4, pp. 64-65, 140-43; Abu'1-Fida, 'Annales', pp. 120, 122; Abu Shama, 'Deux jardins', RHCOr, vol. 5, p. 193 (mentions Beaufort). See also P. Thorau, The Lion of Egypt: Sultan Baybars I and the Near East in the Thirteenth Century, trans. P. M. Holt (London, 1992), pp. 18-19. 9. For an outline of these crusades and the rebuilding of Ascalon, see P. Jackson, 'The Crusades of 1239—41 and their Aftermath', Bulletin of the School of Oriental and African Studies, 50 (1987), pp. 32-60. See also 'Eracles', pp. 413-22; 'Rothelin', pp. 526-56; 'Les gestes des Chiprois', in RHCArm, vol. 2, pp. 725—26 (hereafter 'Gestes'); Matthew Paris, Chronica malora, vol. 4, pp. 25-30, 71, 78-80, 138-45, 166. Jerusalem: Abu'1-Fida, 'Annales', pp. 117-18; al-Ayni, 'Le collier de perles', pp. 196-97; 'Rothelin', pp. 529-31. 10. Chronica de Mallros, ed. J. Stevenson (Edinburgh, 1835), pp. 156-62; Matthew Paris, Chronica malora, vol. 4, pp. 299-311, 337-40; 'Rothelin', pp. 561-63; 'Eracles', pp. 427—28; al-'Ayni, 'Le collier de perles', p. 198. 11. Matthew Paris, Chronica malora, vol. 4, pp. 341—43; 'Rothelin', pp. 564-65; 'Eracles', pp. 429-33; 'Gestes', p. 740; John of Joinville, Hlstoire de Saint Louis, ed. N. de Wailly (Paris, 1874), pp. 288-90; Abu Shama, 'Deux jardins', RHCOr, vol. 5, pp. 193-94; Abu'1-Fida, 'Annales', pp. 122-25; Ibn al-Furat, 'Selections from the Tarikh al-Duwad wa'1-Muluk', in Ayyublds, Mameluks and Crusaders, ed. and trans. U. and M. C. Lyons, with notes and an introduction by J. S. C. Riley-Smith, 2 vols (Cambridge, 1971), vol. 2, pp. 4-8, 10-11, 46. 12. For an outline of Louis IX's Crusade, see J. Richard, Saint Louis: Crusader King of France, ed. S. Lloyd and trans. J. Birrell (Cambridge, 1992), pp. 85-152. 13. For an outline of Baybars' career, see Thorau, The Lion of Egypt, especially at pp. 142-210. 14. For an outline of these events, see Mayer, The Crusades, pp. 269-74; J- RileySmith, The Crusades: A Short History (London, 1987), pp. 203-7; P- M. Holt, The Age of the Crusades: The Near East from the Eleventh Century to 1517 (London, 1986), pp. 99-106; R. Irwin, The Middle East In the Middle Ages: The Early Years of the Mamluk Sultanate, 1250-1382 (London, 1986), pp. 62—84. 15. For an outline of Muslim divisions during the first half of the thirteenth century, see R. S. Humphreys, From Saladin to the Mongols: The Ayyublds of

NOTES TO PAGES 4-8

311

Damascus, 1193-1260 (New York, 1977), pp. 41-307; Holt, The Age of the Crusades, pp. 60-67. 16. For an outline of the fall of the Ayyubids and the rise of the Mamluks in Egypt, the Mongol invasion of Muslim Syria in 1260 and the accession of Baybars, see Humphreys, From Saladin to the Mongols, pp. 309-63; Holt, The Age of the Crusades, pp. 82-89; Thorau, The Lion of Egypt, pp. 27-88. For the Mongols, see D. Morgan, The Mongols (Oxford, 1986), pp. 147-58. 17. R. S. Humphreys, 'The Emergence of the Mamluk Army', Studia Islamica, 45 (1977), PP- 74-76. 18. Oliver of Paderborn, 'Historia Damiatina', ed. H. Hoogeweg, in Bibliothek des Litterarischen Vereins in Stuttgart, vol. 202 (1894), P- 24519. For references, see above, note 11. 20. Lyons and Jackson, Saladin, pp. 279-361; R. C. Smail, Crusading Warfare, second edition with an introduction by C. Marshall (Cambridge, 1995), pp. 64-75. 21. A. S. Ehrenkreutz, ' The Place of Saladin in the Naval History of the Mediterranean Sea in the Middle Ages', Journal of the American Oriental Society, 75 (i955), PP-100-16. 22. 1271: 'Annales de Terre Sainte', ed. R. Röhricht and G. Raynaud, Archives de rOrient Latin, 2 (1884), pp. 427-61, at p. 455 (text B); 'Gestes', p. 778; 'Eracles', p. 460; al-Makrizi, Histoire des sultans mamlouks de l'Egypte, trans. M. E. Quatremère, 2 vols in 4 parts (Paris, 1845), vol. i(b), p. 87; Ibn al-Furat, 'Selections', pp. 152-54; Marino Sañudo, 'Liber secretorum fidelium crucis', in Gesta Dei per Francos, ed. J. Bongars (Hanover, 1611), p. 224. See also C. Hillenbrand, The Crusades: Islamic Perspectives (Edinburgh, 1999), pp. 571-72; J. H. Pryor, Geography, Technology and War: Studies in the Maritime History of the Mediterranean, 649-1572 (Cambridge, 1988), pp. 112-34. 23. C. Marshall, Warfare in the Latin East (Cambridge, 1992), pp. 32-37; D. Ayalon, 'Studies on the Structure of the Mamluk Army', Bulletin of Oriental and African Studies, 15 (1953), pp. 203-28, 448-76 at p. 223 and 16 (1954), pp. 57-90 at pp. 67-70; Thorau, The Lion of Egypt, pp. 98-100, 175, 196. 24. 'Gestes', pp. 804, 806-7. 25. Ayalon,'The Mamluk Army', particularly at 15 (1953), p. 222,16 (1954), pp. 70-71; Marshall, Warfare in the Latin East, pp. 32-33. 26. al-Makrizi, Histoire des sultans, vol. i(b), pp. 27-31; Ibn al-Furat, 'Selections', pp. 88-96. See also 'Gestes', pp. 764-66; 'Eracles', pp. 454-55; Marino Sañudo, 'Liber secretorum', pp. 222-23. 27. Ayalon, 'The Mamluk Army', 15 (1953), pp. 222-28. See also Marshall, Warfare in the Latin East, pp. 33-34, and for more details on the composition, equipment and training of thirteenth-century Muslim armies in general, see Hillenbrand, The Crusades: Islamic Perspectives, pp. 439-67. 28. al-Makrizi, Histoire des sultans, vol. i(b), p. 78. 29. al-Makrizi, Histoire des sultans, vol. 2(a), pp. 120-31; Abu'1-Fida, 'Annales', pp. 163-64; 'Gestes', pp. 805-18. See also Francesco Amadi in Chroniques d'Amadi et de Strambaldi, ed. R. de Mas Latrie (Paris, 1891), pp. 219-26; Florio

312

NOTES TO PAGES

8-11

Bustron, Chronique de l'ile de Chypre, ed. R. de Mas Latrie (Paris, 1886), pp. 119-2530. 'Gestes', p. 792. 31. P. Chevedden, L. Eigenbrod, V. Foley and W. Soedel, 'The Trebuchet', Scientific American, vol. 273 (July 1995), pp. 58-63, especially at p. 59. 32. P. Chevedden, 'Crusader Warfare Revisited: A Revisionist Look at R. C. Smail's Thesis. Fortifications and Development of Defensive Planning', Journal of the Rocky Mountain Medieval and Renaissance Association, 14 (1993), pp. 13-31, at p. 22. 33. Chevedden, 'Crusader Warfare Revisited', pp. 22-23. 34. Chevedden, 'The Trebuchet', p. 59. 35. al-Makrizi, Histoire des sultans, vol. i(b), pp. 27-31 (Saphet), 51 (Beaufort), 85 (Crac des Chevaliers). See also Ibn al-Furat, 'Selections', pp. 89, 91-93 (Saphet), 110-12 (Beaufort), 145-46 (Crac des Chevaliers). For more details on Muslim siege weapons, see also Hillenbrand, The Crusades: Islamic Perspectives, pp. 521-3336. For an outline of the internal politics of the crusader states in the thirteenth century see J. S. C. Riley-Smith, The Feudal Nobility and the Kingdom of Jerusalem, 1174-1277 (London, 1973), especially pp. 101—230; P. W. Edbury, The Kingdom of Cyprus and the Crusades, 1191-1374 (Cambridge, 1991), especially pp. 74-100; Mayer, The Crusades, pp. 231-50, 262-74. 37. For more details on these disputes, see below, pp. 59-63. 38. For an outline of this conflict, see Edbury, The Kingdom of Cyprus, pp. 55-69, 81-82. For Cyprus, see below, pp. 90-91. 39. See Riley-Smith in Ibn al-Furat, 'Selections', p. 173 n. 2. See also Marshall, Warfare in the Latin East, pp. 32, 51-52. 40. 'Gestes', pp. 806-7; Marino Sañudo, 'Liber secretorum', p. 231. Contemporary figures for the number of Muslim besiegers at Acre vary considerably. See Marshall, Warfare in the Latin East, p. 219. 41. Andrew's departure: Oliver of Paderborn, 'Historia Damiatina', p. 168; 'Bracks', p. 325. For an outline of the Fifth Crusade see J. M. Powell, Anatomy of a Crusade, 1213—1221 (Philadelphia, 1986), especially pp. 123-91. 42. 'Eracles', p. 461; Ibn al-Furat, 'Selections', p. 155; 'Annales de Terre Sainte', p. 455 (text B). For an outline of the Lord Edward's crusade, see S. Runciman, 'The Crusader States, 1243-1291', in HC, vol. 2 (1962), pp. 582-83. 43. M. Barber, The New Knighthood: A History of the Order of the Temple (Cambridge, 1994), p. 231, and see pp. 230-31; P. Contamine, War in the Middle Ages, trans. M. Jones (Oxford, 1984), pp. 94-97, 116-17. 44. RRH, no. 1347, p. 352; Barber, The New Knighthood, pp. 232-33; Marshall, Warfare in the Latin East, p. 79, and for more details on this French regiment, pp. 77-83; C. Marshall, 'The French Regiment in the Latin East, 1254-1291', Journal of Medieval History, 15 (1989), pp. 301-7. 45. For more details on the problems of supplying adequate horses and other necessities to the Holy Land, see also M. Barber, 'Supplying the Crusader

NOTES TO PAGES 11-14

313

States: The Role of the Templars', in B.Z.Kedar (ed.), The Horns of Hattin (Jerusalem and London, 1992), pp.314-26. 46. 'Gestes', p. 767. For a brief outline of the Aragonese crusade, see Thorau, The Lion of Egypt, pp. 199-201. 47. 'Gestes', pp. 767-68; Ibn al-Furat, 'Selections', pp. 137-39; al-Makrizi, Histoire des sultans, vol. i(b), p. 77. The figure of 15,000 Muslims was reported by a Prankish knight held prisoner at Saphet, who later escaped to Acre. See 'Gestes', p. 768. Notes to Chapter 2, The Holy Land: Military Architecture 1. 1186: Cartulaire general de l'orare des Hospitaliers de St Jean de Jerusalem, 1100-1310, ed. J. Delaville Le Roulx, 4 vols (Paris, 1894-1906), vol. i, no. 783, pp. 491-96, no. 809, p. 505. Description based on P. Deschamps, Les chateaux des croisés en Terre Sainte, vol. 3, La defense du comté de Tripoli (Paris, 1973), pp. 272-84; H. Kennedy, Crusader Castles (Cambridge, 1994), pp. 167—79; E. G. Rey, Etude sur les monuments de ^architecture militaire des croisés en Syrië et dans l'ile de Chypre (Paris, 1871), pp. 19-38; W. Müller-Wiener, Castles of the Crusaders, trans. J.M.Brownjohn (London, 1966), pp.57-58; T.E.Lawrence, Crusader Castles, ed. D. Pringle (Oxford, 1988), p. 88; C. Cahen, La Syrie du Nord a l'époque des croisades et la principauté franque d'Antioche (Paris, 1940), pp. 171-72; A. Boas, Crusader Archaeology: The Material Culture of the Latin East (London, 1999), pp. 113-15. 2. Cartulaire, vol. i, no. 144, pp. 116-18, no. 391, pp. 266-68. 3. P. Deschamps, Les chateaux des croisés en Terre Sainte, vol i, Le Crac des Chevaliers (Paris, 1934), pp. 142-305; Kennedy, Crusader Castles, pp. 150-63; Rey, Etude, pp. 39-67; Müller-Wiener, Castles of the Crusaders, pp. 59-62; Lawrence, Crusader Castles, pp. 77—88; Boas, Crusader Archaeology, pp. 112—13. 4. Deschamps, Le Crac des Chevaliers, pp. 279-83; idem, La defense du comté de Tripoli, pp. 283-84. Earthquakes: Ibn al-Athir, 'Kamel Altevarykh', RHCOr, vol. i, p. 572 (1170); Abu Shama, 'Deux jardins', RHCOr, vol. 4, p. 154 (1170); Ibn al-Athir, 'History of the Atabegs of Mosul', hi RHCOr, vol. 2, pp. 261-62 (1170); Abu'1-Fida, 'Annales', pp. 40 (1170), 79 (1200/01), 83 (1202/3). 5. Abu Shama, 'Deux jardins', RHCOr, vol. 4, p. 154. 6. Silifke: see below, pp. 179-80. 7. 'The Templars and the Castle of Tortosa in Syria: An Unknown Document Concerning the Acquisition of the Fortress', ed. J. Riley-Smith, EHR, 84 (1969), pp. 278-88, at pp. 284-86. 8. M. Braune, 'Die mittelalterlichen Befestigungen der Stadt Tortosa/Tartus', Damaszener Mitteilungen, 2 (1985), pp. 44-54; Kennedy, Crusader Castles, pp. 134-44; Rey, Etude, pp. 69-83, 211-14; Deschamps, La defense du comté de Tripoli, pp. 289-91. 9. Rey, Etude, pp. 85-92; Deschamps, La defense du comté de Tripoli, pp. 252-58; Müller-Wiener, Castles of the Crusaders, pp. 51-52; Kennedy, Crusader Castles,

314

NOTES TO PAGES

14-17

pp. 138-41; Lawrence, Crusader Castles, pp. 138-41; Boas, Crusader Archaeology, p. 96. 10. 1218: Oliver of Paderborn, 'Historia Damiatina', p. 235. 11. Deschamps, La defense du comté de Tripoli, pp. 313-16; Barber, The New Knighthood, pp. 81-82, 85; Kennedy, Crusader Castles, pp. 68-73; Müller-Wiener, Castles of the Crusaders, p. 53. 12. al-Makrizi, Histoire des sultans, vol. i(b), pp. 84-85; Ibn al-Furat, 'Selections', pp. 143-49; 'Gestes', pp. 768, 777; 'Eracles', p. 460; Marino Sañudo, 'Liber secretorum', p. 224; 'Annales de Terre Sainte', pp. 454 (text A), 455 (text B). 13. al-Makrizi, Histoire des sultans, vol. 2(a), p. 8o; 'Gestes', p. 792; Marino Sañudo, 'Liber secretorum', p. 229. 14. al-Makrizi, Histoire des sultans, vol. 2(a), p. 126; Marino Sañudo, 'Liber secretorum', p. 232. 15. RRH, no. 1447, p. 377; al-Makrizi, Histoire des sultans, vol. 2(a), p. 222. 16. Oliver of Paderborn, 'Historia Damiatina', pp. 169-71. See also Jacques de Vitry, Lettres de Jacques de Vitry, ed. R. B. C. Huygens (Leiden, 1960), pp. 99100; Matthew Paris, Chronica maiora, vol. 3, p. 14; Burchard of Mount Sion, 'Descriptie Terrae Sanctae', in Peregrinatores medii aevi quatuor, ed. J. C. M. Laurent (Leipzig, 1873), p. 83; C. N. Johns, 'Excavations at Pilgrims' Castle ('Atlit): The Ancient Tell and the Outer Defences of the Castle', QDAP, 3 (1933-34) > PP-145-64, at pp. 152-64; C. N. Johns, Guide to 'Atlit: The Crusader Castle, Town and Surroundings (Jerusalem, 1947), pp.36-67. See also M. Benvenisti, The Crusaders in the Holy Land ( Jerusalem, 1970), pp. 179-82; Kennedy, Crusader Castles, pp. 124-27; D. Pringle, Secular Buildings in the Crusader Kingdom of Jerusalem: An Archaeological Gazetteer (Cambridge, 1997), nos 21, 22, pp. 22-24; D. Pringle, 'Town Defences in the Crusader Kingdom of Jerusalem', in I. A. Corns and M. Wolfe (eds), The Medieval City under Siege (Woodbridge, i995)> PP- 69-121, at pp. 91-92; P. Deschamps, Les chateaux des croisés en Terre Sainte, vol. 2, La defense du royaume de Jerusalem (Paris, 1939), pp. 32-33; Rey, Etude, pp. 95-100; D. Pringle, The Churches of the Crusader Kingdom of Jerusalem, 3 vols (Cambridge, 1993 onwards), vol. i, pp. 69-71; Boas, Crusader Archaeology, pp. 110-12. 17. 'Gestes', p. 818; Marino Sañudo, 'Liber secretorum', p. 232. 18. 1168: RRH, no. 447, p. 116.1240: 'Eracles', pp. 417-18, 435; 'Rothelin', pp. 551-53. 19. 'Un noveau traite du texte De constructione castri Saphet', ed. R. B. C. Huygens, Studi Medievali, series 3, vol. 4, part i (1965), pp. 335-87, at pp. 383-84, lines 160-93; D. Pringle, 'Review Article: Reconstructing the Castle of Safad', PEQ, 117 (1985), pp. 139-49. Another description of Saphet is given by Huygens in his introduction to De constructione castri Saphet (pp. 370-77), but Pringle (pp. 141, 142, 145) disagrees with many of his conclusions. See also Pringle, Gazetteer, no. 191, pp. 91-92; Deschamps, La defense du royanme de Jerusalem, pp. 140-42; Benvenisti, The Crusaders in the Holy Land, pp. 199-201; Kennedy, Crusader Castles, pp. 128-29; Pringle, Churches, vol. 2 (1998), pp. 206-9; Boas, Crusader Archaeology, p. 118.

NOTES TO PAGES 17-19

315

20. Ihn al-Furat, 'Selections', p. 87; al-Makrizi, Histoire des sultans, vol. i(b), p. 27; 'Annales de Terre Sainte', p. 452 (text B). 21. 1189: Ihn al-Athir, 'Kamel Altevarykh', RHCOr, vol. i, pp. 712-13, 737; Imad ad-Din al-Isfahani, Conquête de la Syrië et de la Palestine par Saladin, trans. H. Masse (Paris, 1972), pp. 150-54; Cont. WT, p. 75. For a description of ßelvoir, see J. Prawer, The Latin Kingdom of Jerusalem (London, 1972), pp. 300-7; Pringle, Gazetteer, no. 46, pp. 32-33; Pringle, Churches, vol. i, pp. 120-21; Boas, Crusader Archaeology, pp. 106-8. 22. Cartulaire, vol. 2, no. 2726, p. 777, no. 2811, pp. 815-17. 23. Abu'1-Fida, 'Annales', pp. 86-87; 'Eracles', p. 317; Marino Sañudo, 'Liber secretorum', p. 206; Benvenisti, The Crusaders in the Holy Land, pp. 360-61; Pringle, Churches, vol. 2, pp. 66-67. 24. Oliver of Paderborn, 'Historia Damiatina', pp. 165-67; 'Eracles', p. 324; Jacques de Vitry, Lettres, p. 98; Abu Shama, 'Deux jardins', RHCOr, vol. 5, pp. 163-64. 25. Oliver of Paderborn, 'Historia Damiatina', pp. 171-72; Jacques de Vitry, Lettres, p. 108; 'Eracles', pp. 330-31; Matthew Paris, Chronica maiora, vol. 3, p. 14; Abu Shama, 'Deux jardins', RHCOr, vol. 5, pp. 165-66. 26. Benvenisti, The Crusaders in the Holy Land, pp. 361-62; Pringle, Churches, vol. 2, p. 68. 1263: Ibn al-Furat, 'Selections', p. 56 n. i. 27. 1190: Abu Shama, 'Deux jardins', RHCOr, vol. 4, pp. 395-400,441; Imad ad-Din, Conquête de la Syrië, pp. 159-62, 210; Ibn al-Athir, 'Kamel Altevarykh', RHCOr, vol. i, pp. 738-39; Abu'1-Fida, 'Annales', p. 61; Cont. WT, pp. 79-81. 1268: Ibn al-Furat, 'Selections', pp. 108-12; al-Makrizi, Histoire des sultans, vol. i(b), p. 51; 'Gestes', p. 771; 'Eracles', p. 456. Descriptions: Deschamps, La defense du royanme de Jerusalem, pp. 198-208; Pringle, Gazetteer, no. 44, p. 31; MüllerWiener, Castles of the Crusaders, pp. 62-63; Rey, Etude, pp. 127-32; Boas, Crusader Archaeology, p. no. 28. Tabulae ordinis Theutonid, ed. E. Strehlke (Berlin, 1869), no. 63, pp. 51-53; RRH, no. 1002, p. 263. 29. D. Pringle, 'A Thirteenth-Century Hall at Montfort Castle in Western Galilee', Antiquaries Journal, 66 (1986), pp. 52-82, at pp. 54-56; Pringle, Gazetteer, no. 156, pp. 73-75; Benvenisti, The Crusaders in the Holy Land, pp. 335-37; Rey, Etude, pp. 146-51; Marshall, Warfare in the Latin East, pp. 108-11; Kennedy, Crusader Castles, pp. 129-31; Pringle, Churches, vol. 2, pp. 40-42; Boas, Crusader Archaeology, pp. 109-10. 30. Marshall, Warfare in the Latin East, pp. no—n. 31. Deschamps, La defense du comté de Tripoli, p. 309; Kennedy, Crusader Castles, pp. 68-70. 32. Deschamps, La defense du royanme de Jerusalem, pp. 211-13, 217-20; Pringle, Gazetteer, no. 225, p. 103; Pringle, Churches, vol. 2, p. 372. 33. Akkar: Ibn al-Furat, 'Selections', pp. 147-49 (1271); al-Makrizi, Histoire des sultans, vol. i(b), p. 85; 'Gestes', p. 777. Cave de Tyron: Pringle, Churches, vol. 2, p. 372; Pringle, Gazetteer, no. 225, p. 103. 34. Saladin captured Beirut in 1187 and demolished the town walls in 1190. See

3l6

NOTES TO PAGES lp-20

Cont. WT, pp. 78, 98. 1197: Arnold of Lübeck, 'Chronica Slavorum', pp. 205-6. Ibelins: 'Gestes', pp. 678-79. 35. Willbrand of Oldenburg, 'Itinerarium Terrae Sanctae', ed. S. de Sandoli, in Itinera Hierosolymitana Crucesignatorum, 4 vols (Jerusalem, 1978-84), vol. 3, p. 204. 36. R. du Mesnil du Buisson, 'Les anciennes defenses de Beyrouth', Syria, 2 (1921), pp. 235-57, SV-V; I- France, Western Warfare in the Age of the Crusades, 10001300 (London, 1999), p. 268 n. 33. See also Pringle, 'Town Defences', pp. 87-88; Pringle, Gazetteer, no. 45, p. 32; Pringle, Churches, vol. i, pp. 111-12; Boas, Crusader Archaeology, pp. 45-46. 37. For a full list of these and other castles whose destruction was ordered by Saladin, see Itinerarium peregrinorum et gesta Regis Ricardi, ed. W. Stubbs, RS, 38, 2 vols (London, 1864-65), vol i, pp. 280-82; Ambroise, L'estoire de la Guerre Samte, ed. G. Paris (Paris, 1897), lines 6840-69, col. 183; Imad ad-Din, Conquête de la Syrië, pp. 231, 345-46. The extent to which Saladin's orders were actually carried out seems to have varied. 38. Caesarea: Itinerarium, pp. 254-56; Benvenisti, The Crusaders in the Holy Land, p. 138. Arsuf: Itinerarium, p. 282; Matthew Paris, Chronica maiora, vol. 2, p. 376. Jaffa: Itinerarium, p. 284; Cont. WT, p. 135; Ambroise, L'estoire, lines 6941-7053, cols 185-88. Ascalon: ibid., lines 7778-8086, cols 208-16; Itinerarium, pp. 315-17; D. Pringle, 'King Richard I and the Walls of Ascalon', PEQ, 116 (1984), pp. 13347, at pp. 136-42. See also Pringle, 'Town Defences', pp. 84-85 (Ascalon), 89-90 (Caesarea), 93 (Jaffa), 94 (Arsuf). 39. Itinerarium, pp. 396-423; Matthew Paris, Chronica maiora, vol. 2, pp. 387-90; Cont. WT, pp. 145-47; Imad ad-Din, Conquête de la Syrië, pp. 384-87; Abu Shama, 'Deux jardins', RHCOr, vol. 5, pp. 67-71; Ibn al-Athir, 'Kamel Altevarykh', RHCOr, vol. 2, pp. 64-65. 40. Rey, Etude, pp. 205-10; Benvenisti, The Crusaders in the Holy Land, pp. 121-25; P. Deschamps, 'Les entrees des chateaux des croisés en Syrië et leurs defenses', Syria, 13 (1932), pp. 369-87, at p. 386; Boas, Crusader Archaeology, pp. 44-45. 41. Ascalon: Itinerarium, pp. 428-29; Cont. WT, p. 153. Jaffa: Cont. WT, pp. 191-93; Arnold of Lübeck, 'Chronica Slavorum', p. 204; Abu Shama, 'Deux jardins', RHCOr, vol. 5, pp. 116.152; Abu'1-Fida, 'Annales', p. 74; Ibn al-Athir, 'Kamel Altevarykh', RHCOr, vol. 2, pp. 84-86. 42. Caesarea: Oliver of Paderborn, 'Historia Damiatina', pp. 168-70; 'Eracles', p. 325. Arsuf: Willbrand of Oldenburg, 'Itinerarium', p. 232. 43. 'Rothelin', pp. 531-32, 553, 555-56; 'Eracles', pp. 413-14, 421; Matthew Paris, Chronica maiora, vol. 4, p. 143; Pringle, 'King Richard I and the Walls of Ascalon', pp. 143-46; Benvenisti, The Crusaders in the Holy Land, pp. 119-20, 125-26. Pringle and Benvenisti both argue that no attempts were made to rebuild Ascalon's town walls after 1192, but Pringle challenges Benvenisti's earlier view on the exact location of the citadel. See also Pringle, Gazetteer, no. 20, p. 21; Pringle, 'Town Defences', p. 85. 44. Cartulaire, vol. 2, no. 2320, p. 615. Egyptians: 'Rothelin', p. 565; 'Eracles', p. 741;

NOTES TO PAGES 20-22

3V

Matthew Paris, Chronica malora, vol. 4, p. 343; 'Annales de Terre Sainte', p. 442; Ihn al-Furat, 'Selections', pp. lo-ii; Abu Shama, 'Deux jardins', RHCOr, vol. 5, p. 194; Abu'1-Fida, 'Annales', p. 125. 45. 'Eracles', p. 334; Oliver of Paderborn, 'Historia Damiatina', p. 244; Jacques de Vitiy, Lettres, pp. 101-2. 46. Chronique d'Ernoul et de Bernard le Trésorier, ed. L. de Mas Latrie (Paris, 1871), pp. 458-61; 'Eracles', p. 373; Matthew Paris, Chronica malora, vol. 3, p. 175. 47. Jaffa: Ibn al-Furat, 'Selections', p. 107; Joinville, Histoire de Saint Louis, pp. 282— 84, 306-8. Caesarea: ibid., pp. 256, 258, 336; Benvenisti, The Crusaders in the Holy Land, p. 141. 48. 'Gestes', p. 728; 'Annales de Terre Sainte', p. 440. Quote: Joinville, Histoire de Saint Louis, p. 308. 49. Cartulaire, vol. 3, no. 2972, p. i, no. 2985, p. 6; RRH, no. 1302, p. 341, no. 1313, P- 343! 'Gestes', pp. 758-59; J. S. C. Riley-Smith, The Knights of St John in Jerusalem and Cyprus, 1050-1310 (London, 1967), pp. 133-34. 50. Arsuf and Caesarea: 'Gestes', p. 758; 'Eracles', p. 450; Marino Sañudo, 'Liber secretorum', p. 222; 'Annales de Terre Sainte', pp. 451-52; al-Makrizi, Histoire des sultans, vol. i(b), pp. 6-15; Ibn al-Furat, 'Selections', pp. 70-71, 73-82. Jaffa: 'Gestes', p. 771; 'Eracles', p. 456; Marino Sañudo, 'Liber secretorum', p. 223; al-Makrizi, Histoire des sultans, vol. i(b), p. 51; Ibn al-Furat, 'Selections', pp. 1068. 51. Pringle, 'King Richard I and the Walls of Ascalon', p. 145; Pringle, 'Town Defences', p. 85; Pringle, Gazetteer, no, 20, p. 21; Pringle, Churches, vol. i, no. 18, pp. 66-67. 52. I. Roll, 'Medieval Apollonia-Arsuf: A Fortified Coastal Town in the Levant of the Early Muslim and Crusader Periods', in M. Baiard (ed.), Autour de la Première Croisade (Paris, 1996), pp. 595-606, at pp. 602-6; Benvenisti, The Crusaders in the Holy Land, pp. 134—35; Pringle, Churches, vol. i, pp. 59—61; Pringle, Gazetteer, no. 19, pp. 20-21; Pringle, 'Town Defences', p. 94; Boas, Crusader Archaeology, pp. 43—44. 53. Pringle, Churches, vol. \, pp. 264-67, and see no. 109, pp. 267-68, nos 111-12, p. 270; Pringle, Gazetteer, no. no, p. 52; Marshall, "Warfare in the Latin East, pp. 139-44, especially p. 141; Pringle, 'Town Defences', pp. 93-94; Boas, Crusader Archaeology, pp. 49-50. 54. 'Eracles', pp. 432-33; Joinville, Histoire de Saint Louis, pp. 288, 290.1241: Matthew Paris, Chronica malora, vol. 4, p. 142.1247: see above, p. 3. Description: Z. Razi and E. Braun, 'The Lost Crusader Castle of Tiberias', in Kedar, The Horns of Hattin, pp. 216-27, especially at pp. 221-27; Pringle, Gazetteer, no. 222, pp. 101-2; Pringle, 'Town Defences', pp. 94-95. See also Pringle, Churches, vol. 2, pp. 351-53 and no. 267, p. 365; Boas, Crusader Archaeology, p. 57. 55. K. G. Holum, R. L. Hohlfelder, R. J. Bull and A. Raban, King Herod's Dream: Caesarea on the Sea (New Haven and London, 1988), pp. 231-32; Benvenisti, The Crusaders in the Holy Land, pp. 143-44; Pringle, Gazetteer, no. 76, p. 43; Pringle, 'Town Defences', pp. 89-90.

3l8

NOTES TO PAGES 22-23

56. Quote: Burchard of Mount Sion, 'Descriptie', p. 28, and see pp. 27-28; Willbrand of Oldenburg, 'Itinerarium', p. 206; Deschamps, La defense du comté de Trípoli, pp. 300-1. 57. 1188: Imad ad-Din, Conquête de la Syrië, p. 125; Ibn al-Athir, 'Kamel Altevarykh', RHCOr, vol. i, p. 718.1271: According to Muslim chroniclers, Baybars held Maraclea by this date. See Ibn al-Furat, 'Selections', pp. 150, 166; al-Makrizi, Histoire des sultans, vol. i(b), p. 100. For a general history of Nephin and Maraclea during the crusader period, see Deschamps, La defense du comté de Tripoli, pp. 297-300, 323-26. 58. Ibn 'Abd al-Rahim, 'Vie de Kalavun', extract cited in J. F. Michaud, Bibliothèque des Croisades, vol. 4 (Paris, 1829), pp. 551-52; Deschamps, La defense du comté de Tripoli, pp. 325-26. The exact history of Maraclea during the 12705 and 1280S remains unclear. See R. Irwin, 'The Mamluk Conquest of Tripoli', in P. Edbury (ed.), Crusade and Settlement (Cardiff, 1984), pp. 246-49, at pp. 2484959. 1227-28: Chronique d'Ernoul et de Bernard le Trésorier, pp. 458-59; 'Bracks', p. 365; 'Gestes', p. 676.1291: 'Gestes', p. 817; al-Makrizi, Histoire des sultans, vol. 2(a), pp. 130-31. Descriptions: H. Kalayan, 'The Sea Castle of Sidon', Bulletin du Musée de Beyrouth, 26 (1973), pp. 81-89 and plates 1-10; Pringle, Churches, vol. 2, no. 242, pp. 323-28; Pringle, Gazetteer, no. 201, pp. 94-95; Deschamps, La defense du royanme de Jerusalem, pp. 229-33; Rey, Etude, pp. 154-59; Kennedy, Crusader Castles, pp. 122-24; Pringle, 'Town Defences', p. 88; Müller-Wiener, Castles of the Crusaders, p. 70. 60. Deschamps, La defense du comté de Trìpoli, pp. 208-15; Müller-Wiener, Castles of the Crusaders, pp. 64-65; Kennedy, Crusader Castles, pp. 65-66. 61. 1188-97: 'Eracles', pp. 72, 227, 228; Cont WT, pp. 58, 98. It is possible that the Embriaco family was allowed to cultivate lands in this region beyond the fall of Tripoli in 1289 and Acre in 1291. See Irwin, 'The Mamluk Conquest of the County of Tripoli', p. 249, and for a general history of Gibelet see Deschamps, La defense du comté de Trìpoli, pp. 203-8. 62. Joinville, Histoire de Saint Louis, pp. 302, 336; Pringle, Gazetteer, no. 201, pp. 94-95; Deschamps, La defense du royaume de Jerusalem, pp. 227-29; Kennedy, Crusader Castles, p. 121; Müller-Wiener, Castles of the Crusaders, p. 70; D. Pringle, 'A Castle in the Sand: Mottes in the Crusader East', in Chateau Gaillard, 18 (Caen, 1998), pp. 187-91, at p. 188. 63. See below, p. 103. 64. Charles de Terre Sainte provenant de l'abbaye de Josaphat, ed. H. F. Delaborde (Paris, 1880), no. 46, pp. 95—96 (quote p. 96). See also Pringle, 'Town Defences', pp. 86-87. 65. E. G. Rey, 'Etude sur la topographic de la ville d'Acre au XlIIe siècle', Mémoires de la société nationale des antiquaries de France, 39 (1874), pp. 115-45, at p. 131; Benvenisti, The Crusaders in the Holy Land, p. 95. 1291: See below, pp. 50-51. 66. The pilgrim abbot Daniel, cited in C. N. Johns, 'The Citadel, Jerusalem: A

NOTES TO PAGES 23-26

319

Summary of Work Since 1934', QDAP, 14 (1950), pp. 121-90, at p. 164. See also ibid., pp. 140-44, 165. 67. Johns, 'The Citadel, Jerusalem', pp. 169-88; A. J. Boas, Crusader Archaeology, pp. 17-19; Pringle, Gazetteer, no. 115, pp. 54-55; Benvenisti, The Crusaders in the Holy Land, pp. 52-53. 68. See below, pp. 42-43. 69. William of Tyre, Chronicon, ed. R. B. C. Huygens, 2 vols (Turnhout, 1986), vol. i, bk. io, c. 26, p. 486, lines 1-18. For descriptions, see Deschamps, La defense du comté de Tripoli, pp. 293-95; Müller-Wiener, Castles of the Crusaders, pp. 4243; Kennedy, Crusader Castles, p. 63. The Mamluks rebuilt the citadel early in the fourteenth century; see al-Makrizi, Histoire des sultans, vol. 2(b), p. 281. 70. T. A. Sinclair, Eastern Turkey: An Architectural and Archaeological Survey, 4 vols (London, 1987-90), vol. 4 (1990), pp. 244-48 and plate 52; Cahen, La Syrie du Nord, p. 129; Rey, Etude, p. 190. 71. Abu'1-Fida, Annales, pp. 162-63; al-Makrizi, Histoire des sultans, vol. 2(a), pp. 102-3; 'Annales de Terre Sainte', p. 460 (text B). 72. Ibn al-Furat, 'Selections', p. 122. 73. Ludolph of Sudheim, 'De itinere Terrae Sanctae', ed. F. Deycks, in Bibliothek des Litterarischen Vereins in Stuttgart, vol. 25 (Stuttgart, 1851), pp. 39-40. 74. Willbrand of Oldenburg, Ttinerarium', p. 200; Burchard of Mount Sion, 'Descriptio', p. 23. The famous fourteenth-century Genoese map of Acre is reproduced at the back of Marino Sañudo, 'Liber secretorum', and also in D. Jacoby, 'Crusader Acre in the Thirteenth Century: Urban Layout and Topography', in D. Jacoby, Studies on the Crusader States and on Venetian Expansion (Northampton, 1989), c. 5, pp. 1-45, at figs 1-4. For more details of these and other maps and the known topography of crusader Acre, see A. Kesten, The Old City of Acre: Re-Examination Report, 1993 (Acre, 1993). For other modern studies on Acre's defences, see Rey, 'Etude sur la topographic', pp. 118-32; R. Frankel, 'The North-Western Corner of Crusader Acre', Israel Exploration Journal, 37, part 4 (1987), pp. 256-61; Boas, Crusader Archaeology, pp. 34-35; Pringle, 'Town Defences', pp. 81-84; Pringle, Gazetteer, no. 5, pp. 15— 16; Müller-Wiener, Castles of the Crusaders, pp. 72-74; Benvenisti, The Crusaders in the Holy Land, pp. 93-95. 75. Joinviüe, Histoire de Saint Louis, pp. 334, 336; Marino Sañudo, 'Liber secretorum', p. 219. It is clear that Montmusard was already fortified in some way by the 12505. For more details, see D. Jacoby, 'Montmusard, Suburb of Crusader Acre', in Jacoby, Studies on the Crusader States, c. 6, pp. 205-17, particularly at p. 213; Pringle, 'Town Defences', pp. 83-84. 76. Sinclair, Eastern Turkey, vol. 4, pp. 242-48 and plates 47, 51-54; Rey, Etude, pp. 185-204; Cahen, La Syrie du Nord, pp. 127-33; Deschamps, La defense du comté de Tripoli, pp. 46-47; Boas, Crusader Archaeology, p. 42. 77. Sinclair, Eastern Turkey, vol. 4, pp. 243-48. 78. Tripoli: Willbrand of Oldenburg, 'Itinerarium', p. 210. Tyre: William of Tyre, Chronicon, vol. i, bk. 13, c. 5, pp. 591-92, lines 1-17; Willbrand of Oldenburg,

32O

NOTES TO PAGES 20-28

'Itinerarium', p. 202; Burchard of Mount Sion, 'Descriptio', p. 25; Ihn Jubayr, 'Extrait du voyage d'Ibn Djobeïr', in RHCOr, vol. 3, pp. 451-52. For further descriptions see Deschamps, La defense du royanme de Jerusalem, pp. 135-37; Pringle, Gazetteer, no. 227, pp. 103-4; Pringle, 'Town Defences', pp. 85-87. 79. Boas, Crusader Archaeology, pp. 15-17; Pringle, 'Town Defences', pp. 78-81; Pringle, Gazetteer, no. 115, pp. 53-54; Benvenisti, The Crusaders in the Holy Land, pp. 49-52. 80. 1187 onwards: Imad ad-Din, Conquête de la Syrie, pp. 371-73; Abu Shama, 'Deux jardins', RHCOr, vol. 5, pp. 49-51. 1219: ibid., pp. 173-75; Abu'1-Fida, 'Annales', p. 91; 'Eracles', p. 339; Matthew Paris, Chronica maiora, vol. 3, p. 39; Marino Sañudo, 'Liber secretorum', p. 208. 81. 'Rothelin', p. 529; Matthew Paris, Chronica maiora, vol. 3, p. 177; Boas, Crusader Archaeology, pp. 19-21; Benvenisti, The Crusaders in the Holy Land, p. 51; Pringle, 'Town Defences', p. 80. 82. Tabulae ordinis Theutonici, no. 69, p. 55; Johns, 'The Citadel, Jerusalem', p. 167; I. Sterns, 'The Teutonic Knights in the Crusader States', in HC, vol. 5 (1985), PP- 315-78, at p. 365. 83. 'Eracles', pp. 384-85. 84. Chronica de Mauros, pp. 156-62; Matthew Paris, Chronica maiora, vol. 4, pp. 299-311, 337-40; 'Rothelin', pp. 561-63; 'Eracles', pp. 427-28; al-'Ayni, 'Le collier de perles', p. 198. 85. Benvenisti, The Crusaders in the Holy Land, p. 27. 86. This estimate is based on the contemporary claim that the Khwarizmians killed around 7000 people at Jerusalem in 1244. See Matthew Paris, Chronica maiora, vol. 4, p. 309. 87. Benvenisti, The Crusaders in the Holy Land, p. 27. 88. For an outline of the urban defences at Beirut, Jaffa, Ascalon, Arsuf, Caesarea, Sidon, Haifa and Tiberias, see Pringle, 'Town Defences', pp. 84-95, and see above, pp. 19-22. For Tortosa and Gibelet, see above, pp. 16, 19, 23. 89. Holum, Hohlfelder, Bull and Raban, King Herod's Dream, pp. 226-31; Benvenisti, The Crusaders in the Holy Land, pp. 140-45; Pringle, 'Town Defences', pp. 90-91;. Rey, Etude, pp. 221-27; Pringle, Gazetteer, no. 76, pp. 43-44; Müller-Wiener, Castles of the Crusaders, p. 74; Boas, Crusader Archaeology, pp. 46-48. 90. Joinville, Histoire de Saint Louis, pp. 282-84, 306-8 (Jaffa), 302, 336 (Sidon). Haifa: William of Nangis, 'Gesta Ludovici IX', RHGF, vol. 20, p. 384. 91. 'Gestes', p. 701; Amadi, Chroniaue, pp. 148-49; Bustron, Chronique, pp. 81, 8288. 92. Du Mesnil du Buisson, 'Les anciennes defenses de Beyrouth', pp. 235, 251, 317-26; Pringle, 'Town Defences', pp. 87-88; Pringle, Gazetteer, no. 45, p. 32. 93. 'Atlit: C. N. Johns, 'Excavations at Pilgrims' Castle ('Atlit): The Faubourg and its Defences', QDAP, i (1932), pp. 111-29, at pp. 112-24; Johns, Guide to 'Atlit, pp. 74-76, 81-85; Pringle, 'Town Defences', pp. 91-92; Benvenisti, The Crusaders in the Holy Land, pp. 178-79. Crac des Chevaliers: see above, pp. 13-14, and

NOTES TO PAGES 28-32

94. 95. 96. 97.

98. 99.

100.

101. 102. 103. 104. 105.

106. 107. 108.

109.

no. 111. 112. 113.

321

Rey, Etude, p. 40. Margat: see above, p. 13. Beaufort: see above, p. 18, Saphet: see above, pp. 16—17 and De constructione castri Saphet, line 255, p. 386. Deschamps, La defense du comté de Trípoli, pp. 311-12; Kennedy, Crusader Castles, p. 78; Boas, Crusader Archaeology, p. 102. Boas, Crusader Archaeology, p, 101; Marshall, Warfare in the Latin East, p. 100; Pringle, Gazetteer, p. 10. D. Pringle, 'Aqua Bella: The Interpretation of a Crusader Courtyard Building', in Kedar, The Horns of Hattin, pp. 147-67, at p. 147. D. Pringle, The Red Tower (London, 1986), pp. 15, 18-19, 73-75 (and for more examples of similar structures see p. 20); S. Tibbie, Monarchy and Lordships in the Latin Kingdom of Jerusalem, 1099-1291 (Oxford, 1989), p. 139; Pringle, Gazetteer, no. 229, p. 104. Pringle, The Red Tower, pp. 15, 63-68, 70; Pringle, Gazetteer, no. 168, pp. 83-84; Benvenisti, The Crusaders in the Holy Land, pp. 198-99. Pringle, The Red Tower, pp. 16-18; Deschamps, La defense du comté de Trípoli, pp. 317-19; Kennedy, Crusader Castles, pp. 73-75; Müller-Wiener, Castles of the Crusaders, p. 52; Boas, Crusader Archaeology, p. 95. Kennedy, Crusader Castles, pp. 75-77; Rey, Etude, pp. 101-2. For further examples, see Boas, Crusader Archaeology, pp. 93-100; D. Pringle, 'Towers in Crusader Palestine', in Chateau Gaillard, 16 (Caen, 1994), pp. 335-50. See above, pp. 8-9. Kennedy, Crusader Castles, p. 175. Oliver of Paderborn, 'Historia Damiatina', p. 170. See, for example, Kennedy, Crusader Castles, p. 134 and plate 46; Braune, 'Die mittelalterlichen Befestigungen der Stadt Tortosa', plates 17, 21, 21. Beirut: Willbrand of Oldenburg, 'Itinerarium', p. 204. Sidon: Deschamps, La defense du royaume de Jerusalem, p. 232. 'Atlit: Johns, 'The Faubourg and its Defences', p. 123. Maraclea: Deschamps, La defense du comté de Trípoli, p. 326. Ascalon: Matthew Paris, Chronica maiora, vol. 4, p. 243. Tiberias: Razi and Braun, 'The Lost Crusader Castle of Tiberias', p. 224. al-Makrizi, Histoire des sultans, vol. i(b), p. 7. Caesarea: Benvenisti, The Crusaders in the Holy Land, p. 141. Crac des Chevaliers: Deschamps, Le Crac des Chevaliers, pp. 189-90. We have seen that this was one of many Syrian castles damaged by earthquakes in the crusader period; hence sloping revetments were also intended to strengthen walls against this threat. Extract from the anonymous 'Tashrif', trans, and ed. F. Gabrielli in Arab Historians of the Crusades, trans, from Italian by E. J. Costello (London, 1969), p. 336, and see pp. 334-38. Oliver of Paderborn, 'Historia Damiatina', pp. 169-70. See above, pp. 13-14, 16-17. Saphet may have had a combination of rounded and rectangular towers; see Pringle, Gazetteer, no, 191, p. 91. See below, pp. 145-47. Kennedy, Crusader Castles, pp. 134, 137.

322

114. 115. 116. 117. 118. 119.

120. 121. 122. 123. 124. 125.

126. 127. 128.

129. 130. 131. 132. 133. 134. 135.

136. 137. 138. 139. 140.

NOTES TO PAGES 32-36

De constructions castri Saphet, lines 170-76, p. 383. Johns, 'The Ancient Tell and the Outer Defences of the Castle', pp. 162-63. Deschamps, La defense du comté de Tripoli, p. 300. Kennedy, Crusader Castles, pp. 99-100; Deschamps, Le Crac des Chevaliers, p. 189. 'Gestes', p. 814. 'Gestes', pp. 808-9; Marino Sañudo, 'Liber secretorum', pp. 229, 230; 'Ármales de Terre Sainte', p. 459; Rey, 'Etude sur la topographic', pp. 127-29; Deschamps, 'Les entrees des chateaux', pp. 386-87. Jerusalem: Boas, Crusader Archaeology, pp. 19-21; Pringle, 'Town Defences', pp. 80-81; Benvenisti, The Crusaders in the Holy Land, p. 51. 'Gestes', p. 782. For more details on defensive features of this kind see Benvenisti, The Crusaders in the Holy Land, pp. 286-89. Deschamps, Le Crac des Chevaliers, pp. 177-82. Johns, 'The Ancient Tell and the Outer Defences of the Castle', pp. 152-64, and especially pp. 162-63; Chevedden, 'Crusader Warfare Revisited', p. 27; Oliver of Paderborn, 'Historia Damiatina', pp. 255-56. Jaffa: Ibn al-Furat, 'Selections', pp. 90-91; Thorau, The Lion of Egypt, p. 169; Marshall, Warfare in the Latin East, p. 142. Saphet: De constructìone castri Saphet, lines 177—78, p. 383. Abu Shama, 'Deux jardins', RHCOr, vol. 4, pp. 353-55; Imad ad-Din, Conquête de la Syne, pp. 124-25. See also Cont. WT, pp. 86-87. Braune, 'Die mittelalterlichen Befestigungen der Stadt Tortosa', p. 53. Deschamps, La defense du royaume de Jerusalem, pp. 198, 208.1268: Ibn al-Furat, 'Selections', pp. 108-12. See also al-Makrizi, Histoire des sultans, vol. i(b), p. 51; 'Gestes', p. 771; 'Eracles', p. 456. Chevedden, 'Crusader Warfare Revisited', p. 27. Chevedden, 'The Trebuchet', pp. 60-61, 62; Chevedden, 'Crusader Warfare Revisited', pp. 24-27. Deschamps, Le Crac des Chevaliers, pp. 147-50, 155-56, 183, 185-87. Oliver of Paderborn, 'Historia Damiatina', p. 254. 'Annales de Terre Sainte', p. 452 (text B). Ibn al-Furat, 'Selections', p. 70; al-Makrizi, Histoire des sultans, vol. i(b), p. 7. Abu'1-Fida, 'Annales', p. 164; Bar Hebraeus, The Chronography of Gregory Abü'l Faraj, the Son of Aaron, the Hebrew Physician Commonly Known as Bar Hebraeus, trans. E. A. Wallis Budge (Oxford, 1932), pp. 492-93. Cyprus: see below, pp. 95-103. Greece: see below, pp. 203-26. De constructìone castri Saphet, lines 200-204, P- 384- F°r th£ growing cost of fortifications in Europe, see Contamine, War in the Middle Ages, pp. 110-17. Riley-Smith, The Crusades: A Short History, p. 60. Deschamps, Le Crac des Chevaliers, pp. 216-24. For Chlemoutsi and other Greek castles, see below, pp. 203-26. A. Forey, The Military Orders from the Twelfth to the Early Fourteenth Centuries (London, 1992), p. 63.

NOTES TO PAGES 36-39

323

141. Tabulae ordinis Theutonici, no. 64, p. 53, no. 66, p. 54, no. 72, pp. 56-57; Prawer, The Latin Kingdom, pp. 308-9; Pringle, 'A Thirteenth-Century Hall', p. 53; Kennedy, Crusader Castles-, pp. 129-31. 142. Montfort withstood an attack in 1266 before falling to Baybars in 1271. See Ibn al-Furat, 'Selections', pp. 87 (1266), 106-12 (1271); al-Makrizi, Histoire des sultans, vol. i(b), pp. 27 (1266), 87 (1271); 'Annales de Terre Sainte', p. 452 (text B) (1266); 'Eracles', p. 460 (1271); 'Gestes', p. 778 (1271). 143. Quote: Ibn al-Furat, 'Selections', p. 148, and see pp. 143-49; al-Makrizi, Histoire des sultans, vol. i(b), pp. 84-85; 'Gestes', pp. 768, 777; Marino Sañudo, 'Liber secretorum', p. 224; 'Annales de Terre Sainte', pp. 454 (text A), 455 (text B). 144. William of Tyre, Chronicon, vol. i, bk. 11, c. 26, p. 535, lines 26-27; Boas, Crusader Archaeology, p. 117. 145. Smail, Crusading Warfare, pp. 232-33; Marshall, Warfare in the Latin East, p. 100; Pringle, Gazetteer, p. 10. 146. Chevedden, 'The Trebuchet', p. 59. 147. France, Western Warfare in the Age of the Crusades, p. 3. For the walls of Constantinople, see below, pp. 205-6. Notes to Chapter 3, The Holy Land: Fortifications and External Security 1. R. Ellenblum, 'Three Generations of Prankish Castle-Building in the Latin Kingdom of Jerusalem', in Balard, Autour de la Première Croisade, pp. 517-51, at pp. 520-29. For Belvoir, see also Prawer, The Latin Kingdom of Jerusalem, pp. 300-7; Pringle, Gazetteer, no. 46, pp. 32-33. 2. Riley-Smith, 'The Templars and the Castle of Tortosa', pp. 278-79, 284-85. 3. Imad ad-Din, Conquête de la Syrie, pp. 121-22,124-26; Baha'-al-Din, 'Anecdotes et beaux traits de la vie de sultan Youssof', in RHCOr, vol. 3, pp. 106-7; Abu Shama, 'Deux jardins, RHCOr, vol. 4, pp. 349-63; Cont. WT, pp. 86-87; 'Eracles', pp. 119-22. 4. Abu Shama, 'Deux jardins', RHCOr, vol. 4, pp. 346, 381-86 (Saphet, Kerak, Montreal), 344-49, 386-89 (Belvoir), 395-400, 441 (Beaufort); Baha'-al-Din, 'Anecdotes', pp. 104-5,117-20 (Belvoir), 118-20,122 (Kerak, Montreal, Saphet), 121-22,132,151 (Beaufort); Imad ad-Din, Conquête de la Syrië, pp. 104-7, i47~5° (Kerak and Saphet), 104,152-53 (Belvoir), 159-62, 210 (Beaufort); Ibn al-Athir, 'Kamel Altevarykh', RHCOr, vol. i, pp. 712-13, 734-35, 737; Cont. WT, pp. 75-76 (Belvoir and Saphet), 76, 86 (Montreal and Kerak), 79-81 (Beaufort). 5. Tyre survived a Muslim siege in the winter of 1187-88. Saladin avoided direct attacks against both Tripoli and Antioch, preferring to loot and occupy castles around them. See Abu Shama, 'Deux jardins', RHCOr, vol. 4, pp. 341-45 (Tyre), 349-51 (Tripoli), 364-81 (Antioch); Baha'-al-Din, 'Anecdotes', pp. 98-99,102-3 (Tyre), 108-18 (Antioch); Ibn al-Athir, 'Kamel Altevarykh', RHCOr, vol. i, pp. 694-96 (Tyre); Imad ad-Din, Conquête de la Syrië, pp. 63-80 (Tyre), 121-22 (Tripoli), 144-45 (Antioch); Cont. WT, pp. 76-79 (Tyre), 85-86 (Tripoli), 87 (Antioch).

324

NOTES TO PAGES

39-41

6. al-Makrizi, 'Histoire d'Egypte', trans. E. Blochet, ROL, 9 (1902), pp. 6-163, at p. 139 n- i. 7. Oliver of Paderborn, 'Historia Damiatina', p. 245; Jacques de Vitry, Lettres, pp. 101-2.

8. 'Eracles', p. 435; 'Rothelin', p. 624. 9. Cartulaire, vol. 2, no. 2605, pp. 726-28. 10. al-Makrizi, Histoire des sultans, vol. i(a), pp. 199-200, vol. i(b), pp. 7, 27-28, 42, 52, 68-69, 77Í Ibn al-Furat, 'Selections', pp. 57-59, 71, 85-86, 87, 102-3, 116-18,132-34, 137-39; 'Eracles', pp. 446-47, 454-55! 'Gestes', pp. 764, 766, 76768. 11. 'Gestes', p. 755; 'Annales de Terre Sainte', p. 450 (text B); Marino Sañudo, 'Liber secretorum', p. 221; al-Makrizi, Histoire des sultans, vol. i(a), pp. 177—78; Ibn al-Furat, 'Selections', p. 50. 12. Oliver of Paderborn, 'Historia Damiatina', pp. 254-56; Jacques de Vitry, Lettres, p. 138. 13. Quote: Oliver of Paderborn, 'Historia Damiatina', p. 235. See also Abu Shama, 'Deux jardins', RHCOr, vol. 5, p. 166; Kamal-ad-Din, 'L'histoire d'Alep', trans E. Blochet, ROL, 5 (1897), pp. 37-107, at p. 55; Deschamps, La defense du comté de Tripoli, p. 257. 14. 'Gestes', p. 786; 'Annales de Terre Sainte', p. 457 (text A); Marino Sañudo, 'Liber secretorum', p. 228. 15. al-Makrizi, Histoire des sultans, vol. i(b), pp. 78-79; Ibn al-Furat, 'Selections', pp. 139^-40. 16. 'Rothelin', p. 565. 17. See above, p. 3. 18. Oliver of Paderborn, 'Historia Damiatina', pp. 254-56; Jacques de Vitry, Lettres, p.138. 19. Imad ad-Din, Consuète de la Syrië, pp. 356-57, 371-73; Abu Shama, 'Deux jardins', RHCOr, vol. 4, p. 462, vol. 5, pp. 49-51; Itinerarium, pp. 280-82; Ambroise, L'estoire de la Guerre Sainte, lines 6840-68, col. 185. 20. Itinerarium, pp. 380,394. For an alternative view that Richard could and should have attacked Jerusalem at this point, see R. Markowski, 'Richard Lionheart: Bad King, Bad Crusader?', Journal of Medieval History, 23 (1997), pp. 351-65. 21. See below, pp. 243-45. 22. Oliver of Paderborn, 'Historia Damiatina', p. 268. 23. P. Jackson, 'The Crisis in the Holy Land in 1260', EHR, 95 (1980) pp. 481-514, particularly at pp. 496-99. It has, however, been argued that Baybars deliberately allowed the Franks to keep the coastal areas he failed to conquer, and especially Acre, for economic reasons. See Riley-Smith in Ibn al-Furat, 'Selections', pp. xi-xii; Thorau, The Lion of Egypt, p. 148. 24. Fulcher of Chartres, Historia Hierosolymitana, ed. H. Hagenmeyer (Heidelberg, 1913), p. 285. 25. De constructione castri Saphet, lines 205-6, p. 384, lines 232-39, p. 385; Ayalon, 'The Mamluk Army', 15 (1953), p. 222, 16 (1954), pp. 70-71.

NOTES TO PAGES 41~45

325

26. 'Gestes', p. 786. See also 'Annales de Terre Sainte', p. 457 (text A); Marino Sañudo, 'Liber secretorum', p. 228. 27. Ibn al-Furat, 'Selections', p. 122. See also al-Makrizi, Histoire des sultans, vol. i(b), p. 5328. Sinclair, Eastern Turkey, vol. 4, pp. 242-48; Rey, Etude, pp. 185-204; Cahen, La Syrie du Nord, pp. 127-33. 29. Ibn al-Furat, 'Selections', p. 122. See also pp. 121-26; al-Makrizi, Histoire des sultans, vol. i(b), pp. 52-54. 30. 'Eracles', pp. 384-85. 31. al-'Ayni, 'Le collier de perles', p. 196. 32. 'Rothelin', p. 529. 33. Annales prioratus de Dunstaplia, ed. H. R. Luard, in Annales monastiti, vol. 3, RS, 36 (London, 1866), p. 150. This siege was recorded in few of the sources, and remains something of a mystery. See Marshall, Warfare in the Latin East, pp. 243-45; Jackson, 'The Crusades of 1239-41 and their Aftermath', p. 38. 34. 'Rothelin', pp. 529-30; Ibn al-Furat, 'Selections', p. 62; al-'Ayni, 'Le collier de perles', pp. 196-97; Abu'1-Fida, 'Annales', pp. 117-18. 35. Johns, 'The Citadel, Jerusalem', p. 169. 1244: Matthew Paris, Chronica maiora, vol. 4, pp. 299-311, 337-40; 'Rothelin', pp. 561-63; 'Eracles', pp. 427-28; al-'Ayni, 'Le collier de perles', p. 198. 36. See above, p. 22. 37. Ibn al-Furat, 'Selections', p. 122; al-Makrizi, Histoire des sultans, vol. i(b), p. 53. 38. Bar Hebraeus, Chronography, p. 448. 39. 'Gestes', p. 807. Benvenisti, The Crusaders in the Holy Land, p. 27, suggests 30,000 for Acre and Tyre. 40. See above, pp. 23-24. 41. See below, pp. 83—84. 42. Rey, 'Etude sur la topographic', p. 132; Marshall, Warfare in the Latin East, p. 67. 43. See maps of Acre at the back of Marino Sañudo, 'Liber secretorum'; in Jacoby, 'Crusader Acre', figs 1-4; Müller-Wiener, Castles of the Crusaders, p. 73. 44. al-Makrizi, Histoire des sultans, vol. i(b), p. 8; Ibn al-Furat, 'Selections', p. 72. 45. 1188: Abu Shama, 'Deux jardins', RHCOr, vol. 4, pp. 353-55; Cant. WT, p. 87. 1232: 'Gestes', p. 701. 46. Oliver of Paderborn, 'Historia Damiatina', p. 255. 47. Itinerarium, pp. 396-423; Matthew Paris, Chronica maiora, vol. 2, pp. 387-90; Cont. WT, pp. 145-47; Imad ad-Din, Conquête de la Syrie, pp. 384-87; Baha'al-Din, 'Anecdotes', pp. 323-33; Abu Shama, 'Deux jardins', RHCOr, vol. 5, pp. 67-71; Ibn al-Athir, 'Kamel Altevarykh', RHCOr, vol. 2, pp. 64-65. 48. 'Gestes', pp. 704-8; Amadi, Chronique, p. 155; Bustron, Chronique, p. 86. 49. 'Gestes', p. 758. 50. Abu Shama, 'Deux jardins', RHCOr, vol. 4, pp. 349-51; 'Eracles', pp. 119-21. 51. 'Eracles', p. 334; Oliver of Paderborn, 'Historia Damiatina', p. 244; Jacques de Vitry, Lettres, pp. 101-2.

326

NOTES TO PAGES 45~49

52. 'Gestes', p. 817; al-Makrizi, Histoire des sultans, vol. 2,(a), pp. 130-31. 53. Cont. WT, pp. 44-45, 56. 54. Ibn al-Furat, 'Selections', pp. 72, 75, 77; al-Makrizi, Histoire des sultans, vol. i(b), pp. 8, 10; 'Gestes', p. 758; 'Eracles', p. 450 (mentions Arsuf only). 55. Oliver of Paderborn, 'Historia Damiatina', p. 171. See also Matthew Paris, Chronica maiora, vol. 3, p. 14; Johns, Guide to 'Atlit, pp. 49-50. 56. Deschamps, La defense du royanme de Jerusalem, pp. 229-31; Rey, Etude, p. 157. 57. The northern breakwater of this harbour was built out of Roman columns; another example of crusader recycling. See Rey, Etude, pp. 222-23; Benvenisti, The Crusaders in the Holy Land, pp. 143-44; Holum, Hohlfelder, Bull and Raban, King Herod's Dream, pp. 231-34. 58. Rey, Etude, pp. 173-74; Du Mesnil du Buisson, 'Les anciennes defenses de Beyrouth', p. 244. 59. Tyre: In 1242 Ibelin opponents of Frederick II lowered the harbour chain at Tyre in order to let their Venetian allies into the city. See 'Gestes', pp. 732-35. See also Deschamps, La defense du royanme de Jerusalem-, p. 136; Rey, Etude, pp. 167-69. Acre: Jacoby, 'Crusader Acre in the Thirteenth Century', pp. 8-10; R. Gertwagen, 'The Crusader Port of Acre: Layout and Problems of Maintenance', in Balard, Autour de la Première Croisade, pp. 553-82, at pp. 559-60. Gertwagen argues that the layout of Acre's port was quite different from that envisaged by many previous scholars. 60. 1262: 'Gestes', p. 755; 'Annales de Terre Sainte', p. 450 (text B); Marino Sañudo, 'Liber secretorum', p. 221; al-Makrizi, Histoire des sultans, vol. i(a), pp. 177-78; Ibn al-Furat 'Selections', p. 50. 1229: 'Eracles', pp. 384-85. 61. 'Gestes', p. 741; 'Eracles', p. 433 says twenty-two galleys. For Egyptian naval weakness, see above, pp. 6-7. 62. Ibn al-Furat, 'Selections', pp. 93-96 (Saphet), 122-26 (Antioch); al-Makrizi, Histoire des sultans, vol. i(b), pp. 30 (Saphet), 52-54 (Antioch); 'Gestes', pp. 76466 (Saphet). 63. Joinville, Histoire de Saint Louis, p. 302. 64. Tabulae ordinis Theutonid, no. 72, pp. 56-57. 65. Ibn al-Furat, 'Selections', p. 122. See also al-Makrizi, Histoire des sultans, vol. i(b), p. 5366. See above, p. 39. 67. Matthew Paris, Chronica maiora, vol. 4, p. 309. 68. al-Makrizi, Histoire des sultans, vol. i(b), p. 8; Ibn al-Furat, 'Selections', p. 72. 69. Joinville, Histoire de Saint Louis, pp. 302, 336. 70. 'Gestes', p. 752; 'Eracles', p. 444; Jackson, 'The Crisis', pp. 499-500. 71. De construction castri Saphet, lines 256-58, p. 386. 72. al-Makrizi, Histoire des sultans, vol. i(b), pp. 30 (Saphet), 51 (Beaufort), 84-85 (Chastel Blanc and Akkar); Ibn al-Furat, 'Selections', pp. 94-95 (Saphet), 112 (Beaufort), 143 (Chastel Blanc), 148-49 (Akkar). 73. S. B. Edgington, 'The Doves of War: The Part Played by Carrier Pigeons in the Crusades', in Balard, Autour de la Première Croisade, pp. 167-75.

NOTES TO PAGES 49~52

327

74. La regle du Temple, ed. H. de Curzon (Paris, 1886), cl. 591. 75. Jacques de Vitry, Lettres, p. 93. 76. Ellenblum, 'Three Generations of Prankish Castle-Building', pp. 524-26, 528; Prawer, The Latin Kingdom of Jerusalem, pp. 300—7; Pringle, Gazetteer, no. 46, pp. 32-33, pi. xx. 77. Deschamps, La defense du royanme de Jerusalem, pp. 211-13, 217-20. 78. Joinville, Histoire de Saint Louis, pp. 310-18. For more details on Banyas and Subeibe, see A. Grabo'is, 'La cité de Baniyas et le chateau de Subeibeh pendant les croisades', Cahiers de civilisation medievale, 13 (1970), pp. 43-62. 79. Deschamps, La defense du royanme de Jerusalem, p. 178; R. Fedden and J. Thomson, Crusader Castles (London, 1957), p. 12. For more details on Toron, see Deschamps, La defense du royanme de Jerusalem, pp. 117-18; Pringle, Churches, vol. 2, pp. 367-68; Pringle, Gazetteer, no. 223, p. 102. For Cháteauneuf, see Benvenisti, The Crusaders in the Holy Land, pp. 300, 303; Pringle, Gazetteer, no. 164, pp. 79-80. These castles were restored to the Franks in 1229 (al-Makin, Chronique, p. 41) and 1241 (Matthew Paris, Chronica maiora, vol. 4, pp. 141-42) respectively and lost to Baybars in 1266 (Ibn al-Furat, 'Selections', p. 97). 80. This is confirmed by the archaeological evidence, including an inscription from the reign of Baybars. See Deschamps, La defense du royanme de Jerusalem, p. 208. 81. Deschamps, Le Crac des Chevaliers, pp. 105-7; Fedden and Thomson, Crusader Castles, p. 12. 82. Fedden and Thomson, Crusader Castles, p. 53, and see p. 12. See also Deschamps, Le Crac des Chevaliers, pp. 106-7. 83. Deschamps, Le Crac des Chevaliers, pp. 110-15. 84. Quote: Kennedy, Crusader Castles, p. 77, and see pp. 75-77; Rey, Etude, pp. 1012; Deschamps, La defense du comté de Tripoli, 327-29. 85. See below, pp. 56-57. Cilician Armenia: see below, pp. 157-61. Greece: see below, pp. 251-52. 86. Ibn al-Furat, 'Selections', p. 122. 87. 'Gestes', pp. 814, 816; 'Ármales de Terre Sainte', p. 461 (text A); Marino Sañudo, 'Liber secretorum', pp. 231-32. 88. Abu'1-Fida, 'Annales', pp. 162-63; al-Makrizi, Histoire des sultans, vol. 2(3), pp. 102-3; 'Annales de Terre Sainte', p. 460 (text B). 89. Rey, Etude, p. 2. 90. Rey, Etude, p. 4. 91. Smail, Crusading Warfare, pp. 204-15; R. C. Smail, 'Crusaders' Castles of the Twelfth Century', Cambridge Historical Journal, 10 (1951), pp. 133-49, at pp. 13545, 14992. Ibn al-Furat, 'Selections', pp. 50, 103, 104-5, U3> i35> 164; al-Makrizi, Histoire des sultans, vol. i(b), pp. 28, 42, 51, 70. Attempts to maintain these treaties even after the fall of Acre in 1291 failed. See ibid., vol. 2(a), p. 131; 'Gestes', p. 817. See also P. Holt, 'Baybars's Treaty with the Lady of Beirut in 667/1269', in Edbury, Crusade and Settlement, pp. 242-48.

328 93. 94. 95. 96.

NOTES TO PAGES 52-56 Ibn al-Furat, 'Selections', p. 128. Marshall, Warfare in the Latin East, p. 129, and see pp. 12,9-31. Oliver of Paderborn, 'Historia Damiatina', p. 256. Matthew Paris, Chronica maiora, vol. 4, p. 144. Saladin: Coni. WT, p. 153. Baybars: Ibn al-Furat, 'Selections', p. 142; al-Makrizi, Histoire des sultans, vol. i(b), p. 84.

Notes to Chapter 4, The Holy Land: Fortifications and Internal Security 1. See above, pp. 28-29. 2. Deschamps, La defense du comté de Tripoli, pp. 311-12; Kennedy, Crusader Castles, p. 78; Boas, Crusader Archaeology, p. 102. 3. Abu Shama, 'Deux jardins', RHCOr, vol. 4, pp. 352-53; Imad ad-Din, Conquête de la Syrië, pp. 121-22. 4. Ibn al-Athir, 'Kamel Altevarykh', RHCOr, vol. 2, p. 106. 5. Willbrand of Oldenburg, 'Itinerarium', p. 208. 1244: 'Eracles', pp. 427-28; Marino Sañudo, 'Liber secretorum', p. 217. 6. Ibn al-Furat, 'Selections', pp. 85-86; al-Makrizi, Histoire des sultans, vol. i(b), p. 27. See also Deschamps, La defense du comté de Tripoli, p. 312. 7. Ibn al-Furat, 'Selections', p. 150. 8. Ibn al-Furat, 'Selections', p. 116. 9. Ibn al-Furat, 'Selections', p. 87; Deschamps, La defense du royaume de Jerusalem, p. 119; Riley-Smith, The Knights of St John, p. 135; Pringle, Gazetteer, no. 69, p. 41. 10. 1253: 'Eracles', pp. 440-41; Matthew Paris, Chronica maiora, vol. 5, p. 398; Marino Sañudo, 'Liber secretorum', p. 220; 'Annales de Terre Sainte', p. 445 (text B). 1263: ibid., p. 450 (text A); Ibn al-Furat, 'Selections', pp. 57, 57 n. 4, 59.1267: ibid., p. 103 n. 2. 11. Recordane was defended by a sturdy two-storey tower. See Benvenisti, The Crusaders in the Holy Land, p. 251; Deschamps, La defense du royaume de Jerusalem, p. 124; D. Pringle, 'Survey of Castles in the Crusader Kingdom of Jerusalem, 1989: Preliminary Report', Levant, 23 (1991), pp. 87-91, at p. 89; Pringle, Gazetteer, no. 133, pp. 62-64. 12. See below, pp. 69, 71-72. 13. See above, pp. 3-4. 14. 1220: Oliver of Paderborn, 'Historia Damiatina', p. 255.1265: al-Makrizi, Histoire des sultans, vol. i(b), p. 8; Ibn al-Furat, 'Selections', p. 72. 15. C.N.Johns, 'Excavations at Pilgrims' Castle ('Atlit): An Unfinished Church in the Suburb', QDAP 4 (1934-35), pp. 122-37; Johns, Guide to 'Atlit, pp. 77-81; Pringle, Churches, vol. i, no. 2.7, pp. 75-80. 16. Ellenblum, 'Three Generations of Prankish Castle-Building', pp. 519-23; R. Ellenblum, Prankish Rural Settlement in the Latin Kingdom of Jerusalem (Cambridge, 1998), especially at pp. 12—19. 17. Ellenblum, Prankish Rural Settlement, especially pp. 3-6, 9, 12-19, 31-38-

NOTES TO PAGES 56-61

329

18. See below, pp. 271-98. 19. France, Western Warfare in the Age of the Crusades, p. 214. 20. Pringle, 'Towers in Crusader Palestine', pp. 341-42. 21. See below, pp. 247-49. 22. Ellenblum, Prankish Rural Settlement, p. 19. 23. La regle du Temple, cl. 616. 24. Joinville, Histoire de Saint Louis, p. 336. 25. Jacques de Vitry, Lettres, pp. 91-92. 26. Oliver of Paderborn, 'Historia Damiatina', p. 169. 27. See, for example, Deschamps, La defense du royanme de Jerusalem, p. 123; Ellenblum, Prankish Rural Settlement, pp. 10-11. 28. H. E. Mayer, 'Latins, Muslims and Greeks in the Latin Kingdom of Jerusalem', History, 63 (1978), pp. 175-92, especially at pp. 175—87. See also Smail, Crusading Warfare, pp. 62-63; B. Z. Kedar, 'The Subjected Muslims of the Prankish Levant', in J. M. Powell (ed.), Muslims under Latin Rule, 1100—1300 (Princeton, 1990), pp. 135-74. especially at p. 167. 29. Ellenblum, Prankish Rural Settlement, pp. 7-9, 20-36. 30. Abu Shama, 'Deux jardins', RHCOr, vol. 4, pp. 301-2; Smail, 'Crusaders' Castles of the Twelfth Century', p. 142; Kedar, 'The Subjected Muslims', p. 155. 31. 'Eracles', p. 455; Marino Sañudo, 'Liber secretorum', p. 222, and see below, p. 281. 32. 'Eracles', pp. 384-85. On Muslim resentment against Frederick IFs acquisition of Jerusalem from al-Kamil in 1229, see al-'Ayni, 'Le collier de perles', pp. 18794; Hillenbrand, The Crusades: Islamic Perspectives, pp. 216-23. 33. 'Eracles', pp. 384-85. 34. 'Eracles', pp. 373~7435. 'Eracles', p. 373; Marino Sañudo, 'Liber secretorum', p. 213. 36. For an outline of this conflict, see Edbury, The Kingdom of Cyprus, pp. 55-69, 81-82, and see above, pp. 9-10. 37. 'Gestes', pp. 783-84, 789-90; 'Annales de Terre Sainte', p. 456. For more details on this dispute, see Edbury, The Kingdom of Cyprus, pp. 90-97; Riley-Smith, The Feudal Nobility, pp. 224-27. Truce of 1286: RRH, nos 1465-66, pp. 382-83. 38. 'Eracles', p. 318; 'Annales de Terre Sainte', pp. 436-37; 'Gestes', p. 664; Hethoum the Historian, 'Table chronologique de Héthoum, comte de Gongos', in RHCArm, vol. i, pp. 483-84; Constable Sempad, La chronique attribuée au connétable Smbat, ed. and trans. G. Dédéyan (Paris, 1980), pp. 89-90; Cahen, La Syrie du Nord, pp. 621-22, 630-31. For more details of this conflict, see below, pp. 180-81. 39. 'Eracles', pp. 314-15; 'Annales de Terre Sainte', p. 435 (although text A confuses Gibelcar/ Akkar with Gibelet); Marino Sanudo,'Liber secretorum', p. 205; Cahen, La Syrie du Nord, pp. 608-9. 40. 'Gestes', pp. 748-50 (1258), 781-83 (1276), 787-88 (1282). See also Deschamps, La defense du comté de Trìpoli, pp. 206-8, 299-300; Irwin, 'The Mamluk Conquest of the County of Tripoli', p. 247.

33O

NOTES TO PAGES 6l-66

41. 'Gestes', pp. 814-16; Kesten, The Old City of Acre, pp. 69-81; Benvenisti, The Crusaders in the Holy Land, pp. 104-9; Barber, The New Knighthood, pp. 241-43; Riley-Smith, The Knights of St John, pp. 248-49; Rey, 'Etude sur la topographic', pp. 135—36, 139-40. The Hospitaller compound is in the process of being excavated. For a summary of recent discoveries, see Boas, Crusader Archaeology, pp. 37-41. 42. 'Gestes', p. 815; Kesten, The Old City of Acre, pp. 29-39 (Genoa), 40-45 (Venice), 59-67 (Pisa); Benvenisti, The Crusaders in the Holy Land, pp. 98—100 (Pisa), 100-2 (Genoa), 102-4 (Venice); Rey, 'Etude sur la topographic', pp. 137-39; Jacoby, 'Crusader Acre in the Thirteenth Century3, pp. 19-26 (Pisa), 26-30 (Genoa), 30-36 (Venice). 43. 'Gestes', pp. 742-48. See also 'Eracles', p. 443; 'Rothelin', pp. 633-35; 'Annales de Terre Sainte', pp. 447-48; Marino Sañudo, 'Liber secretorum', pp. 220-21; Riley-Smith, The Feudal Nobility, pp. 215-17. 44. 1258: 'Gestes', pp. 742-50.1287-89: 'Gestes', pp. 800-2; 'Annales de Terre Sainte', pp. 459-60. 45. 'Gestes', pp. 756-57, 768-69. 46. France, Western Warfare in the Age of the Crusades, p. 86.

Notes to Chapter 5, The Holy Land: The Offensive Functions of Fortifications 1. 1188: Abu Shama, 'Deux jardins', RHCOr, vol. 4, pp. 352-53; Imad ad-Din, Conquête de la Syrië, pp. 121-22.1207-8: Ibn al-Athir, 'Kamel Altevarykh', RHCOr, vol. 2, p. 106; Ibn al-Furat, 'Selections', pp. 85-86. 2. Ibn al-Furat, 'Selections', p. 150. 3. See above, pp. 54-55. 4. See above, pp. 16-18, 19-23, 42-43. 5. 1188: Imad ad-Din, Conquête de la Syrië, pp. 128-31; Abu Shama, 'Deux jardins', RHCOr, vol. 4, pp. 356, 359-63. c. 1261: Ibn al-Furat, 'Selections', p. 115. 6. Joinville, Histoire de Saint Louis, p. 230. 7. Joinville, Histoire de Saint Louis, pp. 282-84, 306-8 (Jaffa), 302, 336 (Sidon), 256, 258, 336 (Caesarea), 334, 336 (Acre). Haifa: William of Nangis, 'Gesta Ludovici IX', RHGF, vol. 20, p. 384. 8. Joinville, Histoire de Saint Louis, pp. 302, 336. 9. Joinville, Histoire de Saint Louis, pp. 310-18. 10. Joinville, Histoire de Saint Louis, p. 284. For the construction of new castles at Sidon and Jaffa in the 12205, see Chronique d'Ernoul et de Bernard le Trésorier, pp. 458-62; 'Gestes', p. 676; 'Eracles', p. 365, 372. 11. Itinerarium, p. 240; Matthew Paris, Chronica maiora, vol. 2, p. 376. 12. Itinerarium, pp.284 (Jaffa), 315-17 (Ascalon); Ambroise, L'estoire de la guerre sainte, lines 6941-7053, cols 185-88 (Jaffa), lines 7778-8086, cols 208-16 (Ascalon). 13. Itinerarium, pp. 280, 289-90, 296; Ambroise, L'estoire de la guerre sainte, line

NOTES TO PAGES 66-68

331

6854, col. 183, lines 7177-214, col. 193. Description: D. Pringle, 'Templar Castles between Jaffa and Jerusalem', in H. Nicholson (ed.), The Military Orders 2: Welfare and Warfare (Aldershot, 1998), pp. 89-109, at pp. 92-94. 14. Pringle, 'Templar Castles', pp. 89—109. 15. Oliver of Paderborn, 'Historia Damiatina', p. 171. 16. Matthew Paris, Chronica malora, vol. 4, p. 290; Pringle, 'Templar Castles', pp. 94-102. 1244: Chronica de Mauros, pp. 156-62. 17. Joinville, Histoire de Saint Louis, p. 302; Pringle, 'Templar Castles', p. 96. 18. Coni. WT, pp. 76-79; Imad ad-Din, Conquête de la Syne, pp. 63-80; Baha'-alDin, 'Anecdotes', RHCOr, pp. 98-99, 102-3; Abu Shama, 'Deux jardins', RHCOr, vol. 4, pp. 341-45; Ibn al-Athir, 'Kamel Altevarykh', RHCOr, vol. i, pp. 694-96. 19. Itinerarium, pp. 27-28; Coni. WT, pp. 83, 85-86; Abu Shama, 'Deux jardins', RHCOr, vol. 4, pp. 352-53> 356-5720. Battle of Arsuf: Itinerarium, pp. 262-80; Ambroise, L'estoire de la guerre sainte, lines 6090-630, cols 163-77; Abu Shama, 'Deux jardins', RHCOr, vol. 5, pp. 3840; Ibn al-Athir, 'Kamel Altevarykh', RHCOr, vol. 2, pp. 49-50; Baha'-al-Din, 'Anecdotes', pp. 258-61. 21. 'Eracles', pp. 224-26; Abu Shama, 'Deux jardins', RHCOr, vol. 5, pp. 116-17; Ibn al-Athir, 'Kamel Altevarykh', RHCOr, vol. 2, pp. 86-87; Arnold of Lübeck, 'Chronica Slavorum', pp. 205-6. 22. Oliver of Paderborn, 'Historia Damiatina', pp. 165-67; 'Eracles', p. 324; Jacques de Vitry, Lettres, p. 98; Abu Shama, 'Deux jardins', RHCOr, vol. 5, pp. 163-64. 23. 'Eracles', p. 461; 'Gestes', pp. 778-79; al-Makrizi, Histoire des sultans, vol. i(b), p. 101; Ibn al-Furat, 'Selections', p. 155; al-'Ayni, 'Le collier de perles', p. 246. See also Pringle, The Red Tower, pp. 60, 62; Marshall, Warfare in the Latin East, p. 206. 24. Arnold of Lübeck, 'Chronica Slavorum', pp. 207-10; Cont. WT, pp. 195-97; Abu'1-Fida, 'Annales', p. 74; Ibn al-Athir, 'Kamel Altevarykh', RHCOr, vol. 2, pp. 87-88. For a description of Toron, see Deschamps, La defense du royanme de Jerusalem, pp. 117-18. 25. Baha'-al-Din, 'Anecdotes', p. 274. 26. Le trésor des diartes d'Armenie ou cartulaire de la chancélleme royale des Roupéniens, ed. V. Langlois (Venice, 1863), no. u, pp. 130-31 ( Jábala), no. 12, pp. 132-33 (Bikisrail); Cartulaire, vol. 2, nos 1262-63, PP-70-71 (Jábala), no. 1355, PP-122-23 (Bikisrail). 27. RRH, no. 949, pp. 251-52, no. 1043, p. 272. 28. Ibn al-Athir, 'Kamel Altevarykh', RHCOr, vol. 2, pp. 105-6. 29. 1237: Matthew Paris, Chronica malora, vol. 3, pp. 404-6; Abu'1-Fida, 'Annales', pp. 112-13; Izz al-Din Ibn Saddad, Description de la Syrie du Nord, trans. A-M. Eddé-Terrasse (Damascus, 1984), pp. 265-66. The importance of Templar castles north of Antioch will be discussed in more detail below. See pp. 182-86. 30. Ibn al-Furat, 'Selections', p. 115. 31. Ibn al-Furat, 'Selections', pp. 115,126. See also Jackson ('The Crisis', pp. 494-96),

332

NOTES TO PAGES 6S-/2

who argues that Bohemond made these conquests through sheer force rather than with the assistance of the Mongols. 1188: Imad ad-Din, Conquète de la Syne, pp. 133-40; Baha'-al-Din, 'Anecdotes', pp. 112-115; Abu Shama, 'Deux jardins', RHCOr, vol. 4, pp. 368-74. 32. Darbsak: Cahen, La Syrie du Nord, p. 705. Darbsak is listed as one of the castles Baybars acquired from the Armenians. See Abu'1-Fida, 'Ármales', p. 152; Ibn al-Furat, 'Selections', p. 166. 1268: ibid., p. 126. 33. 'Gestes', p. 778; Ibn al-Furat, 'Selections', p. 155. 34. Matthew Paris, Chronica malora, vol. 3, pp. 10-11; 'Eracles', pp. 323-25; Jacques de Vitry, Lettres, p. 99; Oliver of Paderborn, 'Historia Damiatina', pp. 167-68. 35. Joinville, Histoire de Saint Louis, pp. 310-18. 36. 'Gestes', p. 752; 'Eracles', p. 444; Jackson, 'The Crisis', pp. 499-500. 37. 'Gestes', p. 767. 38. al-Makrizi, 'Histoire d'Egypte', pp. 126-28, 126 n. 3-4, 128 n. i; Abu'1-Fida, 'Ármales', p. 81. 39. Homs and Hama: al-Makrizi, 'Histoire d'Egypte', p. 135; Ibn al-Athir, 'Kamel Altevarykh', RHCOr, vol. 2, p. 105; Kamal-ad-Din, 'L'histoire d'Alep', pp. 44-45. Jábala: al-Makrizi, 'Histoire d'Egypte', pp. 135, 127m; 'Eracles', pp. 247-48. 40. Ibn al-Furat, 'Selections', pp. 83-84. 41. 'Gestes', p. 784; Marino Sañudo, 'Liber secretorum', p. 228; 'Ármales de Terre Sainte', p. 457 (text A). Hospitallers and Templars from Margat and surrounding strongholds may also have participated in an attack on Homs and Hama undertaken by Prankish, Armenian, Mongol and Georgian troops in 1282. See Hethoum the Historian, 'Table chronologique', p. 487. 42. France, Western Warfare in the Age of the Crusades, p. 218. 43. De constructione castri Saphet, line 251, p. 385. 44. Oliver of Paderborn, 'Historia Damiatina', pp. 171-72. 45. Oliver of Paderborn, 'Historia Damiatina', pp. 171-72; Jacques de Vitry, Lettres, p. 108; 'Eracles', pp. 330-31; Matthew Paris, Chronica maiora, vol. 3, p. 14; Abu Shama, 'Deux jardins', RHCOr, vol. 5, pp. 165-66. 46. Ibn al-Furat, 'Selections', p. 101. 47. Burchard of Mount Sion, 'Descriptio', pp. 83-84. 48. See above, p. 22. 49. Ibn al-Furat, 'Selections', p. 98; al-Makrizi, Histoire des sultans, vol. i(b), p. 32; Willbrand of Oldenburg, 'Itinerarium', p. 210. 50. Ibn al-Athir, 'Kamel Altevarykh', RHCOr, vol. 2, p. 180. 51. al'-Ayni, 'Le collier de perles', p. 194; Deschamps, Le Crac des Chevaliers, p. 128. 52. 'Eracles', pp. 403-5; Abu'1-Fida, 'Annales', p. 115. 53. Ibn al-Furat, 'Selections', p. 98; al-Makrizi, Histoire des sultans, vol. i(b), p. 32. 54. See below, pp. 280-81. 55. Joinville, Histoire de Saint Louis, pp. 310-18. 56. Deschamps, La defense du corate de Tripoli, pp. 283-84. 57. Ibn al-Furat, 'Selections', p. 148.

NOTES TO PAGES 72-76

333

58. 'Gestes', p. 784 (1279), 786 (1281); Marino Sañudo, 'Liber secretorum', p. 228; 'Annales de Terre Sainte', p. 457 (text A). 59. 'Rothelin', pp. 630-32; Marshall, Warfare in the Latin East, p. 143. 60. Bar Hebraeus, Chronography, pp. 492-93; Abu'1-Pida, 'Annales', p. 194. 61. La regle du Temple, cl. 168. 62. Matthew Paris, Chronica malora, vol. 4, p. 342. 63. Matthew Paris, Chronica maiora, vol. 3, pp. 404-6; Abu'1-Fida, 'Annales', pp. 112-13. 64. 'Eracles', p. 455; Marino Sañudo, 'Liber secretorum', p. 222. 65. For an outline of Saladin's extensive conquests in 1187 and 1188, see Lyons and Jackson, Saladin, pp. 267-94. 66. Matthew Paris, Chronica maiora, vol. 4, p. 301; Riley-Smith in Ibn al-Furat, 'Selections', p. 173 n. 2. 67. 'Rothelin', p. 565; 'Eracles', p. 741; Matthew Paris, Chronica maiora, vol. 4, p. 343; Ibn al-Furat, 'Selections', pp. 10-11; Abu Shama, 'Deux jardins', RHCOr, vol. 5, p. 194; Abu'1-Fida, 'Annales', p. 125. 68. 'Eracles', p. 445; 'Gestes', pp. 752-53; Marino Sañudo, 'Liber secretorum', p. 221; Abu Shama, 'Deux jardins', RHCOr, vol. 5, p. 204. 69. De constructione castri Saphet, lines 204-9, P-384; Marshall, Warfare in the Latin East, pp. 118-20. 70. The estimate of 900 knights comes from Abu Shama, 'Deux jardins', RHCOr, vol. 5, p. 204. See also 'Eracles', p. 445; 'Annales de Terre Sainte', p. 449 (text A). 71. 'Eracles', p. 441; 'Rothelin', p. 629; Marshall, Warfare in the Latin East, pp. 77-83; Marshall, 'The French Regiment', pp. 301—7. 72. Willbrand of Oldenburg, 'Itinerarium', pp. 208-10. 73. 'Eracles', pp. 403-5; Abu'1-Fida, 'Annales', p. 115. 74. Ibn al-Furat, 'Selections', p. 86. 75. Ibn al-Furat, 'Selections', p. 143; al-Makrizi, Histoire des sultans, vol. i(b), p. 84. 76. Cartulaire, vol. 4, no. 3308, pp. 291-93; Marshall, Warfare in the Latin East, pp. 117-18. 77. De constructione castri Saphet, lines 204-9, P- 3^4; Cartulaire, vol. 2, no. 2727, pp. 777-78; Marshall, Warfare in the Latin East, pp. 117-20. 78. 'Gestes', pp. 784, 786; Marino Sañudo, 'Liber secretorum', p. 228; 'Annales de Terre Sainte', p. 457 (text A). 79. 'Gestes', pp. 701, 704; Amadi, Chronique, pp. 149, 153; Bustron, Chronique, pp. 81, 85. 80. Itinerarium, pp. 62, 73; Ambroise, L'estoire de la guerre sainte, lines 3060-76, cols 82-83; 'Histoire des patriarches d'Alexandrie', extract cited in J. F. Michaud, Bibliotheque des Croisades, vol. 4 (Paris, 1829), p. 257. 81. Sicily: Itinerarium, p. 168. Acre: Richard of Devizes, Cronicón de tempore Regis Ricardi, ed. J. T. Appleby (London, 1963), p. 42. 82. See below, pp. 240-41. 83. Oliver of Paderborn, 'Historia Damiatina', p. 255.

334

NOTES TO PAGES

76-81

84. Itinerarium, p. 412. 85. Itinerarium, pp. 315-17, 428-29; Ambroise, L'estoire de la guerre sainte, lines 7778-8086, cols 208-16; Pringle, 'King Richard I and the Walls of Ascalon', pp. 136-42. 86. Ibn al-Furat, 'Selections', p. 142; al-Makrizi, Histoire des sultans, vol. i(b), p. 84. 87. A comment originally made regarding the construction of Ibelin in the 11405. See the Old French Continuation of William of Tyre in RHCOc, vol. i, pp. 697-98. 88. 'Gestes', pp. 756-57. 89. 'Gestes', pp. 732-35. 90. 'Gestes', pp. 708-9; Amadi, Chronique, p. 160; Bustron, Chronique, p. 89. 91. 'Gestes', p. 701; Bustron, Chronique, p. 81; Amadi, Chronique, p. 149. 92. 'Gestes', p. 784. 93. Oliver of Paderborn, 'Historia Damiatina', p. 176. 94. Ibn al-Furat, 'Selections', p. 141. 95. La regle du Temple, cl. 119. It has been noted recently, however, that the principal arsenal at Acre may well have been silted up during the crusader period, and that the harbour as a whole was not very user-friendly despite its economic importance. See Gertwagen, 'The Crusader Port of Acre', pp. 560-61, 569-78. Generally, see also Pryor, Geography, Technology and War, pp. 112-34.

Notes to Chapter 6, The Holy Land: The Military Orders 1. Tabulae ordinis Theutonid, nos 109-10, pp. 89-90. 2. 'Eracles', pp. 444-45; 'Gestes', p. 752; 'Annales de Terre Sainte', p. 449; Marino Sañudo, 'Liber secretorum', p. 221. 3. Tabulae ordinis Theutonid, nos 53-54, pp. 43-45, no. 63, pp. 51-53; RRH, no. 1002, p. 263. 4. D. Pringle, A. Petersen, M. Dow and C. Singer, 'Qal'at Jiddin: A Castle of the Crusader and Ottoman Periods in Galilee', Levant, 26 (1994), pp. 135-66, particularly at pp. 135-54, 159-62. For Mhalia, see also Pringle, 'Survey of Castles in the Crusader Kingdom of Jerusalem', p. 90, and see Pringle, Gazetteer, no. 152, pp. 71-72, no. 165, pp. 80-82. Amuda: see below, p. 175. 5. Tabulae ordinis Theutonid, no. 111, pp. 90-91, nos 114-15, pp. 96-98, nos 117-18, pp. 103-6. 6. Tabulae ordinis Theutonid, no. 44, pp. 35-36; Cahen, La Syrie du Nord, pp. 66768. 7. 1168: RRH, no. 447, p. 116. 1260: 'Eracles', p. 445; 'Gestes', p. 752; 'Annales de Terre Sainte', p. 449; Marino Sañudo, 'Liber secretorum', p. 221. 8. The tower of Destroit dated from the twelfth century. See Oliver of Paderborn, 'Historia Damiatina', p. 169; Johns, Guide to 'Atlit, p. 94. 9. Riley-Smith, 'The Templars and the Castle of Tortosa', pp. 284-86; Barber, The New Knighthood, pp. 79-82. 10. Cartulaire, vol. i, no. 144, pp. 116-18, no. 391, pp. 266-68 (Crac des Chevaliers),

NOTES TO PAGES 81-83

335

no. 783, pp. 491-96, no. 809, p. 505 (Margat), vol. 2, no. 2320, p. 615 (Ascalon), no. 2726, p. 777, no. 2811, pp. 815-17 (Mount Tabor), vol. 3, no. 2972, p. i, no. 2985, p. 6 (Arsuf ). 11. Cartulaire, vol. i, no. 82, pp. 76-78 (Goliath), no. 519, pp. 353-54, no. 549, pp. 371-72 (Chastel Rouge). 12. Ibn al-Furat, 'Selections', pp. 143, 147. 13. Cartulaire, vol. i, no. 783, pp. 491-96. For details of all Hospitaller properties in the area, see Riley-Smith, The Knights of St John, pp. 93-95. 14. R. P. Harper and D. Pringle, 'Belmont Castle: A Historical Notice and Preliminary Report of Excavations in 1986', Levant, 20 (1988), pp. 101-18, at p. 102. See also Pringle, Gazetteer, no. 207, p. 96.1191: Ambroise, L'estoire de la guerre sainte, line 6859, col. 183. 15. Thus, for example, Marshall, Warfare in the Latin East (published 1992), pp. 20-21, disagrees with Riley-Smith, The Knights of St John (published 1967), pp. 415-16, 436-37. See also Pringle, Gazetteer, no. 46, p. 32 (Belvoir), no. 228, p. 104 (Forbelet); Pringle, 'Survey of Castles in the Crusader Kingdom of Jerusalem', p. 90 (Forbelet). 16. Pringle, Churches, vol. i, nos 31-32, p. 95; Pringle, Gazetteer, no. 32, p. 27. 17. La Tor de l'Opital: Riley-Smith, The Knights of St John, p. 135; Pringle, Gazetteer, no. 69, p. 41. Turris Salinarum: Tibbie, Monarchy and Lordships, p. 150; Pringle, Gazetteer, no. 67, p. 41. 18. RRH, no. 949, pp. 251-52, no. 1043, p. 272. 19. Ibn al-Furat, 'Selections', pp. 115 (Latakia), 128 (Jábala). 20. Willbrand of Oldenburg, 'Itinerarium', p. 212; Ibn al-Furat, 'Selections', p. 128 n. 5; Cahen, La Syrie du Nord, p. 171. 21. Cartulaire, vol. 2, no. 2117, pp. 486-87, no. 2120, p. 489; Riley-Smith, The Knights of St John, pp. 446, 450. The Hospitallers had held Recordane since at least 1154. See Cartulaire, vol. i, no. 225, p. 173. 22. Cartulaire, vol. i, no. 879, p. 559; Pringle, The Red Tower, p. 59. 23. Cartulaire, vol. 2, no. 2141, p. 501, no. 2482, pp. 673-75. 24. Pringle, 'Templar Castles', pp. 89-109. 25. See D. Pringle, 'Templar Castles on the Road to the Jordan', in M. Barber (ed.), The Military Orders: Fighting for the Faith and Caring for the Sick (Aldershot, 1994), pp. 148-66. 26. 1187: Abu Shama, 'Deux jardins', RHCOr, vol. 4, p. 313.1192: Coni. WT, pp. 15152. 27. Matthew Paris, Chronica malora, vol. 4, pp. 288-91 (and see p. 339 for 1244); Barber, The New Knighthood, pp. 143-44. In 1252 Gaza was formally recognised as an Egyptian possession: RRH, no. 1199, p. 315. 28. RRH, no. 1318, p. 344. See also Barber, The New Knighthood, pp. 163-64. 29. ÄRH, no. 1450, p. 379. 30. Tabulae ordinis Theutonid, no. 28-29, pp. 24-25. 31. Cartulaire, vol. i, no. 938, p. 594, no. 972, pp. 616-17. 1291: 'Gestes', pp. 808, 812-13.

336

NOTES TO PAGES 83-90

32. Tabulae ordinis Theutonici, no. 44, pp. 35-36. This document refers to other sections of the walls being guarded by the Templars. 1289: 'Gestes', p. 803. 33. Cartulaire, vol. i, no. 954, p.603 (Jaffa), vol. 2, no. 2160, p.310 (Sidon). 34. Tabulae ordinis Theutonici, no. 40, pp. 32-33. Jerusalem: see above, p. 26. 35. 'Rothelin', p. 636. 36. Beaufort: Rey, Etude, pp. 127-28; Deschamps, La defense du royaume de Jerusalem, pp.198, 208. Arsuf: Ibn al-Furat, 'Selections', p. 54; Benvenisti, The Crusaders in the Holy Land, pp. 134-35. 37. De constructione castri Saphet, lines 203-4, P- 384. 38. La regle du Temple, cis 126-27. For more details on the income of the Orders, see Forey, The Military Orders, pp. 98-132. Pilgrims and crusaders: see below, p. 298. 39. Willbrand of Oldenburg, 'Itinerarium', p. 210; Deschamps, Le Crac des Chevaliers, pp. 203-4. 40. La regle du Temple, d. 228; Cartulaire, vol. 3, no. 3844, p. 453, cl. 12. 41. Ibn al-Furat, 'Selections', p. 122. 42. Abu Shama, 'Deux jardins', RHCOr, vol. 4, pp. 364-67; Ibn al-Athir, 'Kamel Altevarykh', RHCOr, vol. i, pp. 721-22; Baha'-al-Din, 'Anecdotes', pp. 111-12. For a brief description and history of Saone, see Kennedy, Crusader Castles, pp. 83-96. 43. Marshall, Warfare in the Latin East, pp. 141-42. 44. Joinville, Histoire de Saint Louis, pp. 306-8. 45. Tibbie, Monarchy and Lordships, p. 152, and see pp. 99-152, particularly pp. 120-52. 46. Dramatically confirmed by the mocking letter Baybars sent to Bohemond VI after he captured Akkar in 1271: Ibn al-Furat, 'Selections', p. 148. 47. 'Gestes', pp. 678-79. Saladin: Coni. WT, p. 98. 48. 'Eracles', pp. 432—33; Joinville, Histoire de Saint Louis, pp. 288, 290. 49. Smail, 'Crusaders' Castles of the Twelfth Century', p. 147.

Notes to Chapter 7, Cyprus: Warfare 1. See below, pp. 115-16. 2. Edbury, The Kingdom of Cyprus, pp. 1-22. G. Hill, A History of Cyprus, 4 vols (Cambridge, 1940-52), vol. 2, pp. 1-57. For more details on Isaac Komnenos prior to 1191, see W. H. Rüdt de Collenberg, 'L'empereur Isaac de Chypre et sa filie, 1155-1207', Byzantion, 28 (1968), pp. 123-77, also in W. H Rüdt de Collenberg, Families de l'Orìent latin, Xlle-XIVe siècles (London, 1983), chapter i, pp. 123-77. 3. Michael VIII's plans: L. de Mas Latrie, Histoire de l'ile de Chypre sous le regne des princes de la maison de Lusignan, 3 vols (Paris, 1852-61), vol. 3, pp. 653-55 at p. 654; Edbury, The Kingdom of Cyprus, pp. 10-11. For the problems of the Byzantine empire and its failure to regain Cyprus from Isaac, see Rüdt de Collenberg, 'L'empereur Isaac de Chypre', pp. 133-49; M. Angold,

NOTES TO PAGES 90-93

337

The Byzantine Empire, 1025-1204 (London, 1984), pp. 263-95. See also below, pp. 191-92. 4. 'Gestes', pp. 820-22; Amadi, Chronique, p. 228; Bustron, Chronique, p. 128; Pope Nicholas IV, Registres, ed. E. Langlois, 2 vols (Paris, 1886-1905), no. 6778. There were no less than five Egyptian sultans between 1291 and 1310. See Holt, The Age of the Crusades, pp. 103-13; Irwin, The Middle East in the Middle Ages, pp. 79-102. 5. For a brief history of Egypt and its problems after 1291, see Holt, The Age of the Crusades, pp. 107-206, particularly pp. 192-202. Naval weakness: Thorau, The Lion of Egypt, pp. 101, 203, 206-8, and see above, pp. 6-7. 6. Edbury, The Kingdom of Cyprus, pp. 109-11,132-33,155-56,197-211; Hill, History of Cyprus, vol. 2, pp. 379-417, vol. 3, pp. 589-91, and see below, pp. 117-22. 7. See below, pp. 108-10. 8. 'Eracles', pp. 360-62; 'Gestes', pp. 668-76; Amadi, Chronique, pp. 117-24; Edbury, The Kingdom of Cyprus, pp. 48-57; Riley-Smith, The Feudal Nobility, pp. 159-66. 9. 'Gestes', pp. 676-82; 'Eracles', pp. 367-69, 376; Amadi, Chronique, pp. 124-34; Bustron, Chronique, pp. 63-73; Edbury, The Kingdom of Cyprus, pp. 57-60. 10. 'Eracles', pp. 376-77; 'Gestes', pp. 684-94; Amadi, Chronique, pp. 136-46; Bustron, Chronique, pp. 74-79; Edbury, The Kingdom of Cyprus, pp. 60-61. 11. 'Gestes', pp. 699-724; 'Eracles', pp. 386-403; Amadi, Chronique, pp. 147-63; Bustron, Chronique, pp. 80-104; Edbury, The Kingdom of Cyprus, pp. 61-65. 12. 'Gestes', pp. 866-69, 871; Leontios Makhairas, Recital Concerning the Sweet Land of Cyprus Entitled 'Chronicle', ed. R. M. Dawkins, 2 vols (Oxford, 1932), vol. i, cc. 42-63, pp. 43-59; Amadi, Chronique, pp. 241-54, 259-62, 271-80, 298-391; Bustron, Chronique, pp. 135-41,148-62,176-243; Diomedes Strambaldi, Chroniques d'Amadi et de Strambaldi, ed. R. de Mas Latrie (Paris, 1891), pp. 1825; Amauiy of Tyre's election charter of 1306 in 'Documents chypriotes du debut du XlVe siècle', ed. C. Kohler, ROL, u (1905-8), pp. 440-52, at pp. 444-52; Edbury, The Kingdom of Cyprus, pp. 109-31; Hill, History of Cyprus, vol. 2, pp. 219-69. 13. Leontios Makhairas, Recital, cc. 90-281, pp. 81-269; Wiliam of Machaut, La prise d'Alexandre ou chronique du roí Pierre Ier de Lusignan, ed. R. de Mas Latrie (Geneva, 1877), pp. 19—265; Amadi, Chronique, pp. 408-26; Bustron, Chronique, pp. 259-76; Strambaldi, Chronique, pp. 35-114. These chroniclers believed Peter was murdered more because of his tyrannical behaviour than because of his expensive war policies. See J. Richard, 'La revolution de 1369 dans le royaume de Chypre', Btbliothèque de l'Ecole des Chartes, 90 (1952), pp. 108-23; P. Edbury, 'The Murder of King Peter I of Cyprus (1359-1369)', Journal of Medieval History, 6 (1980), pp. 219-33; P. Edbury, 'The Crusading Policy of King Peter I of Cyprus, 1359-69', in The Eastern Mediterranean Lands in the Period of the Crusades, ed. P. M. Holt (Warminster, 1977), pp. 90-105; Edbury, The Kingdom of Cyprus, pp. 175-79. 14. See above, pp. 7-9.

338

NOTES TO PAGES 93-98

15. Giorgio and Giovanni Stella, 'Annales Genuenses', ed. G. Petti Balbi, RIS n.s., vol. 17, part 2 (Bologna, 1975), p. 167. 16. 'Gestes', pp. 806-7; Marino Sañudo, 'Liber secretorum', p. 231; Edbury, The Kingdom of Cyprus, p. 99. 17. See below, pp. 108-9. 18. 'Gestes', p. 712. 19. For the history of Cyprus between 1374 and 1571, see Hill, History of Cyprus, vol. 2, pp. 404-96, vol. 3, pp. 497-1040.

Notes to Chapter 8, Cyprus: Military Architecture 1. Itinemrium pp. 181-91; Richard of Devizes, Cronicón, pp. 35-36; Ambroise, L'estoire de la guerre sainte, cols 38—42; Cont. WT, pp. 116-19; 'Grades', pp. 161-64; Roger of Howden, Chronica, ed. W. Stubbs, RS, 51, 4 vols (London, 1868-71), vol. 3, pp. 105-7; Gesta Regís Henna Secundi Benedicti Abbatis, ed. W. Stubbs, RS, 49, 2 vols (1869), vol. 2, pp. 162-64. Cyprus was renowned for its wealth. See Itinerarium, p. 204; Rüdt de Collenberg, 'L'empereur Isaac de Chypre', p. 144. 2. Itinerarium, pp. 193-94; Ambroise, L'estoire de la guerre sainte, cols 42—50; 'Eracles', pp. 165-67. 1192 rebellion: Leontios Makhairas, Recital, c. 12, p. n; 'Bracks', pp. 189-91; Amadi, Chronique, pp. 83-85; Bustron, Chronique, pp. 50-52. 3. Itinerarium, p. 199. 4. Roger of Howden, Chronica, vol. 3, p. in. 5. Originally the archaeologist A. H. S. Megaw believed this site did contain major Byzantine defences built between 654 and 1191. See A. H. S. Megaw, 'Excavations at Sandra Kolones, Paphos: Preliminary Report on the 1966-67 and 1970-71 Seasons', in Report of the Department of Antiquities, Cyprus, 1971 (Nicosia, 1971). PP-117-46; A. H. S. Megaw, 'Supplementary Excavations on a Castle Site at Paphos, Cyprus, 1970-71, Dumbarton Oaks Papers, 26 (1972), pp. 322-45. More recent evidence suggests these structures are Prankish. See A. H. S. (Peter) Megaw, 'A Castle in Cyprus Attributable to the Hospital?', in The Military Orders: Fighting for the Faith and Caring for the Sick, ed. M. Barber (Aldershot, 1994), PP- 42-516. Hill, History of Cyprus, vol. 2, p. 18. 7. C. Enlart, L'art gothique et de la Renaissance en Chypre, 2 vols (Paris, 1899), vol. 2, p. 696. 8. Richard of Devizes, Cronicón, p. 35, calls it a church, whereas Itinerarium, p. 186, calls it a small fort. See also Enlart, L'art gothique, vol. 2, pp. 678-82; A. H. S. Megaw, 'The Arts in Cyprus, B: Military Architecture', in HC, vol. 4, pp. 196-207, at p. 199. 9. Estienne de Lusignan, Description de tonte l'isle de Cypre (Paris, 1580), fol. 32. The late date of this source may well make it unreliable. 10. Megaw, 'Military Architecture', pp. 199-200.

NOTES TO PAGES 98-103

339

n. Willbrand of Oldenburg, 'Itinerarium', p. 228; 'Gestes', p. 724; Enlart, L'art gothique, vol. 2, p. 559. 12. Enlart, L'art gothique, vol. 2, pp. 571—72; Megaw, 'Military Architecture', pp. 199200. For the post-i3OO theory, see Hill, History of Cyprus, vol. 2, p. 20. 13. Bustron, Chronique, p. 24; Estienne de Lusignan, Description, fos 35v~36; Enlart, L'art gothique, vol. 2, pp. 583-95; Megaw, 'Military Architecture', p. 204; Rey, Etude, pp. 239-48. 14. Enlart, L'art gothique, vol. 2, pp. 591-95. For the use of such halls, see chapter twenty-five. 15. Leontios Makhairas, Recital, c. 552, pp. 547—49; Amadi, Chronique, pp. 477-78; Strambaldi, Chronique, pp. 232-33; Bustron, Chronique, p. 338. 16. Bustron, Chronique, pp. 23-24; Estienne de Lusignan, Description, fol. 35v; Enlart, L'art gothique, vol. 2, pp. 600-5; Megaw, 'Military Architecture', pp. 2056; Rey, Etude, pp. 249-50. 17. Bustron, Chronique, p. 23; Estienne de Lusignan, Description, fol. 35v; Enlart, L'art gothique, vol. 2, pp. 650-54; Megaw, 'Military Architecture', p. 205. 18. Itinerarium, p. 202; Ambroise, L'estoire de la guerre sainte, cols 54-55. 19. Itinerarium, pp. 197, 203. See also Edbury, The Kingdom of Cyprus, p. 9. 20. Itinerarium, p. 202. 21. Itinerarium, pp. 202-4; Richard of Devizes, Cronicón, p. 38; Roger of Howden, Chronica, vol. 3, p. in. 22. Willbrand of Oldenburg, 'Itinerarium', p. 228; Estienne de Lusignan, Description, fol. 123. 23. Enlart, L'art gothique, vol. 2, pp. 678-81. See also Megaw, 'Military Architecture', p. 199. 24. Richard of Devizes, Cronicón, p. 35; Itinerarium, p. 186. 25. Megaw, 'A Castle in Cyprus', p. 45. •2.6. 'Gestes', pp. 712-3; Amadi, Chronique, pp. 165—66; Bustron, Chronique, p. 93; Enlart, L'art gothique, vol. 2, pp. 615-18; Megaw, 'Military Architecture', p. 197. 27. Willbrand of Oldenburg, 'Itinerarium', pp. 229, 231. 28. Enlart, L'art gothique, vol. 2, pp. 673-76. 29. Megaw, 'Supplementary Excavations', pp. 323-43; Megaw, 'A Castle in Cyprus', pp. 42-51. In the former report Megaw argues that this castle is Byzantine; in the latter he proves it is actually Prankish. 1222 earthquake: Oliver of Paderborn, 'Historia Damiatina', p. 279; 'Annales de Terre Samte', p. 437 (text B). 30. Leontios Makhairas, Recital, c. 41, p. 43. 31. Estienne de Lusignan, Description, foL 24; Enlart, L'art gothique, vol. 2, pp. 61518. For Tyre, see above, p. 23. 32. Amaury also had the wooden balconies of houses facing the city streets removed, because they hampered the movement of his cavalry troops. Amadi, Chronique, p. 335; Bustron, Chronique, p. 194. 33. Enlart, L'art gothique, vol. 2, pp. 573-75, 577; Megaw, 'Military Architecture', pp. 200-3. 34. 'Manuscript recording the journey of an anonymous Englishman', trans, and

340

NOTES TO PAGES 1O3~1O6

ed. T. A. H. Mogabgab, in Supplementary Excerpts on Cyprus: Further Materials for a History of Cyprus (Nicosia, 1941), p. 58. 35. Niccolo da Martoni, 'Relation du pèlerinage a Jerusalem de Nicolas de Martoni, notaire Italien', ed. L. LeGrand, ROL, 3 (1895), pp. 566-669, at p. 628. 36. Leontios Makhairas, Recital, cc. 386-88, pp. 365-67, cc. 410-20, pp. 389-403; Amadi, Chronique, pp. 446-48, 450-54; Strambaldi, Chronique, pp. 156-60,16673; Bustron, Chronique, pp. 302-11. 37. Leontios Makhairas, Recital, c. 450, p. 435. 38. Leontios Makhairas, Recital, c. 586, p. 585. 39. Enlart, L'art gothique, vol. 2, p. 619. 40. Bustron, Chronique, p. 26. Henry II: see above, p. 102. 41. Leontios Makhairas, Recital, c. 384, p. 363. See also Strambaldi, Chronique, p. 156; Amadi, Chronique, pp. 439-40; Bustron, Chronique, p. 294. 42. Amadi, Chronique, p. 495; Bustron, Chronique, pp. 26, 352. 43. William of Machaut, La prise, pp. 258-59. See also Leontios Makhairas, Recital, c. 260, p. 241, c. 265, p. 247; Amadi, Chronique, p. 422; Bustron, Chronique, p. 271; Strambaldi, Chronique, p. 102; Enlart, L'art gothique, vol. 2, pp. 520-21. 44. Leontios Makhairas, Recital, c. 594, p. 591. 45. Leontios Makhairas, Recital, c. 597, p. 593. See also Amadi, Chronique, p. 490; Strambaldi, Chronique, pp. 250-51; Bustron, Chronique, p. 349; Estienne de Lusignan, Description, fol. 32; Enlart, L'art gothique, vol. 2, pp. 519-20. 46. Estienne de Lusignan, Description, fol. 35v; Enlart, L'art gothique, vol. 2, p. 649; Megaw, 'Military Architecture', pp. 205-6. 47. Enlart, L'art gothique, vol. 2, pp. 681-82, tentatively attributes these changes to James, although Leontios Makhairas, Recital, c. 672, p. 653, claims that the castle of Limassol was 'built' by his successor Janus (1398-1432). 48. Amadi, Chronique, p. 495; Bustron, Chronique, pp. 24, 352; Estienne de Lusignan, Description, fol. 35v; Enlart, L'art gothique, vol. 2, pp. 660-61. 49. Amadi, Chronique, p. 495; Bustron, Chronique, p. 352; Estienne de Lusignan, Description, fol. 36; Megaw, 'Military Architecture', p. 204. 50. Amadi, Chronique, p. 495; Bustron, Chronique, p. 352. 51. Leontios Makhairas, Recital, c. 377, p. 359; Amadi, Chronique, pp. 444-45; Bustron, Chronique, pp. 301—2; Strambaldi, Chronique, pp. 154-55. W3: see below, pp. 117-18. 52. Amadi, Chronique, p. 398; Bustron, Chronique, pp. 249-50. 53. Enlart, L'art gothique, vol. 2, pp. 611-15, 618-19. 54. Enlart, L'art gothique, vol. 2, pp. 523—24 (Nicosia), 676-77, 682-83 (Limassol). 55. Enlart, L'art gothique, vol. 2, pp. 567-77; Megaw, 'Military Architecture', pp. 200-3. 56. Enlart, L'art gothique, vol. 2, p. 696. 57. Estienne de Lusignan, Description, fos 35V-36. 58. Arcadia: see below, p. 210. 59. Estienne de Lusignan, Description, fol. 36. St Hilarión had existed since the eleventh century at the very least. See Megaw, 'Military Architecture', p. 204.

NOTES TO PAGES 1O6-111

341

60. Qaqun and Chastel Rouge: see above, p. 29. Kalamata, Arcadia, Corinth and Neopatras: see below, pp. 209-11. 61. Megaw, 'A Castle in Cyprus', pp. 48-51, and see below, pp. 125-27. For Belvoir, see Prawer, The Latin Kingdom of Jerusalem, pp. 300-7; Kennedy, Crusader Castles, pp. 58-61, and see above, p. 17. 62. See above, pp. 13-19, 24-26. 63. Goliath: Cartulaire, vol. i, no. 82, pp. 76-78, and see above, pp. 28-29. 64. Marshall, Warfare in the Latin East, p. 100.

Notes to Chapter 9, Cyprus: Fortifications and Internal Security 1. Gesta Regis, pp. 172-73; Roger of Howden, Chronica, vol. 3, p. 116. 2. Cont. WT, pp. 135-37; 'Eracles', pp. 189-91; Leontios Makhairas, Recital, c. 12, p. 11; Amadi, Chronique, pp. 83-85; Bustron, Chronique, pp. 50-52. 3. Cont. WT, p. 135. See also P. Edbury, 'The Templars in Cyprus', in Barber, The Military Orders, pp. 189-95, at p. 190. 4. J. Richard, 'Les révoltes chypriotes de 1191-1192 et les inféodations de Guy de Lusignan', in B. Z. Kedar, J. Riley-Smith and R. Hiestand (eds), Montjoie: Studies in Crusader History in Honour of Hans Eberhard Mayer (Aldershot, 1997)> PP-123-28; Edbury, The Kingdom of Cyprus, p. 20. 5. Edbury, The Kingdom of Cyprus, p. 15. 6. Estienne de Lusignan, Description, fol. 3iv. 7. Holy Land: see above, chapter four. Greece: see below, chapter twenty-three. 8. Philip of Mézières, The Life of St Peter Thomas, ed. J. Smet (Rome, 1954), pp. 92-93. See also Leontios Makhairas, Recital, c. 101, pp. 89-91; Amadi, Chronique, pp. 409-10; Bustron, Chronique, p. 258; Strambaldi, Chronique, p. 39. For more details on the underlying tensions between the Greek and Latin Churches, see N. Coureas, The Latin Church in Cyprus, 1195—1312 (Aldershot, 1998), especially pp. 251-317; J.Gill, 'The Tribulations of the Greek Church in Cyprus, 1191 - c. 1280', BZ, 5 (1977), PP- 73-939. Mas Latrie, Histoire, vol. 3, p. 742. 10. Leontios Makhairas, Recital, c. 445, pp. 429-31. 11. Leontios Makhairas, Recital, c. 696, p. 673, and see cc. 696-97, pp. 673-75; Amadi, Chronique, p. 513; Bustron, Chronique, p. 369; Strambaldi, Chronique, pp. 284-85. 12. 'Eracles', pp. 369, 376; 'Gestes', pp. 676-83, 684; Amadi, Chronique, pp. 131-34; Bustron, Chronique, pp. 68-73. 13. 'Gestes', p. 690; Amadi, Chronique, p. 143; Bustron, Chronique, p. 78. 14. 'Gestes', pp. 684-94; Amadi, Chronique, pp. 143-6; Bustron, Chronique, pp. 787915. 'Gestes', pp. 690-95; Amadi, Chronique, pp. 143-45; Bustron, Chronique, p. 79. 16. 'Gestes', p. 710; Amadi, Chronique, pp. 162-63; Bustron, Chronique, p. 91. 17. 'Gestes', pp. 707, 718-24; 'Eracles', pp. 399, 401-2; Amadi, Chronique, pp. 158, 173-82; Bustron, Chronique, pp. 91, 98-105.

342.

NOTES TO PAGES 111-118

18. 'Gestes', pp. 676-83; 'Eracles', pp. 367-69; Amadi, Chronique, pp. 124-32; Bustron, Chronique, pp. 63-70. 19. See above, pp. 91, m. 20. 'Gestes', p. 712, and see above, pp. 91, 94. 21. Amadi, Chronique, p. 250. See also Leontios Makhairas, Recital, c. 54, p. 53; Bustron, Chronique, p. 139. 22. Amadi, Chronique, pp. 335-36, 343-44; Bustron, Chronique, pp. 199-201, 207. 23. See above, pp. 61-63, 79> 84-86. 24. Edbury, The Kingdom of Cyprus, p. 185. Military Orders: see below, chapter twelve. 25. See above, pp. 59-61.

Notes to Chapter 10, Cyprus: Fortifications and External Secutiry 1. 2. 3. 4. 5.

See above, pp. 89-90. Megaw, 'A Castle in Cyprus', p. 45. Coni. WT, pp. 163-64; 'Eracles', pp. 205—7. 'Eracles', pp. 345-46. 'Ármales de Terre Sainte', p. 455 (text B); al-Makrizi, Histoire des sultans, vol. i(b), p. 87; Ibn al-Furat, 'Selections', pp. 152-54. For Muslim naval problems, see Pryor, Geography, Technology and War, pp. 112-34; Thorau, The Lion of Egypt, pp. 203, 206-8, and see above, pp. 6-7. 6. Amadi, Chronique, p. 238. 7. Leontios Makhairas, Recital, c. 64, p. 61. Some pirates were Turkish (see below), others were Catalan; see ibid., c. 103, p. 93. 8. Amadi, Chronique, pp. 393 (1312), 398 (1316); Bustron, Chronique, pp. 245 (1312), 249-50 (1316). See also Edbury, The Kingdom of Cyprus, p. 132. 9. Leontios Makhairas, Recital, c. 377, pp. 357-58. See also Amadi, Chronique, p. 444; Bustron, Chronique, pp. 300-1; Strambaldi, Chronique, p. 153. 10. Leontios Makhairas, Recital, c. 151, p. 133, and see cc. 139-41, pp. 121-22, cc. 150-52, pp. 131-35, for more Turkish raids. For Peter I's attacks, see below, pp. 127-29. u. Estienne de Lusignan, Description, fol. 35v. 12. See above, pp. 98-100. 13. Enlart, L'art gothique, vol. 2, pp. 615-18. 14. Leontios Makhairas, Recital, c. 221, p. 203. 15. Leontios Makhairas, Recital, c. 362, p. 341. 16. Leontios Makhairas, Recital, c. 377, pp. 357-58. 17. Leontios Makhairas, Recital, c. 377, p. 359. See also Amadi, Chronique, pp. 444— 45; Bustron, Chronique, pp. 301—2; Strambaldi, Chronique, pp. 154—55. Kyrenia: see below, pp. 120-21. 18. Amadi, Chronique, p. 495; Bustron, Chronique, p. 352. 19. Leontios Makhairas, Recital, c. 151, p. 133. 20. See above, pp. 90,102-3.

NOTES TO PAGES Il8-123

343

21. Holy Land: see above, pp. 30-35. 22. Holy Land: see above, pp. 7-9. 23. 'Gestes', pp. 806-7; Marino Sañudo, 'Liber secretorum', p. 231; Edbury, The Kingdom of Cyprus, p. 99. 24. 1244: 'Bracks', pp. 429-33; Matthew Paris, Chronica malora, vol. 4, pp. 341-43; 'Rothelin', pp. 564-65; 'Gestes', p. 740, and see above, pp. 3, 40. 25. Enlart, L'ari gothique, vol. 2, p. 522. 26. See above, pp. 91-93, 104, and below, pp. 127-28. 27. Leontios Makhairas, Recital, cc. 386-88, pp. 365-67, cc. 410-20, pp. 389-403; Amadi, Chronique, pp. 446-48; Bustron, Chronique, pp. 302-11; Strambaldi, Chronique, pp. 156-60, 166-73. 28. Leontios Makhairas, Recital, c. 384, p. 363. 29. Leontios Makhairas, Recital, c. 424, p. 405. See also Amadi, Chronique, p. 454; Bustron, Chronique, p. 312; Strambaldi, Chronique, p. 173. 30. Leontios Makhairas, Recital, cc. 433-43> pp. 4i7-^9> cc. 453-54» PP- 437-39, c. 510, pp. 499-501; Amadi, Chronique, pp. 455-57, 458, 468-69; Bustron, Chronique, pp. 312-16, 326-27; Strambaldi, Chronique, pp. 178-83, 186-87, 21112;. See also Hill, History of Cyprus, vol. 2, pp. 393-402, 406-10. 31. Leontios Makhairas, Recital, cc. 464-70, pp. 449-55; Amadi, Chronique, p. 460; Bustron, Chronique, p. 317; Strambaldi, Chronique, pp. 192-94. 32. Leontios Makhairas, Recital, c. 470, p. 453. 33. Leontios Makhairas, Recital, cc. 470-81, pp. 453-65; Amadi, Chronique, pp. 46062; Strambaldi, Chronique, pp. 194-99; Bustron, Chronique, pp. 317-19. 34. Deschamps, Le Crac des Chevaliers, pp. 177-82. 35. Leontios Makhairas, Recital, cc. 481-503, pp. 465-91; Amadi, Chronique, pp. 462-67; Bustron, Chronique, pp. 317-25; Strambaldi, Chronique, pp. 199-209, 217. 36. Amadi, Chronique, p. 454. See also Leontios Makhairas, Recital, c. 425, p. 405; Bustron, Chronique, p. 312. 37. Leontios Makhairas, Recital, cc. 470-78, pp. 453-61. 38. Hill, History of Cyprus, vol. 3, pp. 589-91. 39. See above, pp. 104-5. 40. Enlart, L'ari gothique, vol. 2, p. 658. 41. Estienne de Lusignan, Description, fol. 35v. 42. see below, pp. 240-41. 43. William of Tyre, Chronicon, vol. 2, bk. 15, c. 25, pp. 707-8; Kennedy, Crusader Castles, pp. 30-32. For a new interpretation of the role of these castles, see Ellenblum, 'Three Generations of Prankish Castle-Building', pp. 520—21. 44. Estienne de Lusignan, Description, fol. 27V. 45. Leontios Makhairas, Recital, c. 469, p. 453; Strambaldi, Chronique, p. 193. 46. Leontios Makhairas, Recital, c. 447, p. 431; Strambaldi, Chronique, p. 184. 47. Estienne de Lusignan, Description, fos 35v-36r.

344

NOTES TO PAGES 124-127

Notes to Chapter 11, Cyprus: Fortifications, Naval Warfare and the

Crusades 1. 2. 3. 4. 5. 6. 7.

Estienne de Lusignan, Description, fol. 123. Willbrand of Oldenburg, 'Itinerarium', p. 228. 'Gestes', p. 676; Bustron, Chronique, pp. 63-64. 'Eracles', p. 364. Joinville, Histoire de Saint Louis, pp. 72-74. 'Eracles', pp. 345-46, and see pp. 321-26,339-41; Jacques de Vitry, Lettres, p. 138. Thomas Wykes, 'Chronicon', ed. H. R. Luard, in Anuales Monastiä, RS, 36 (London, 1864-69), vol. 4, pp. 244—45. 8. Edbury, The Kingdom of Cyprus, pp. 74-76. 9. 'Gestes', pp. 806-7; Marino Sañudo, 'Liber secretorum', p. 231. 10. 'Eracles', pp. 345-46. u. 'Gestes', pp. 777-79; 'Annales de Terre Sainte', pp. 454-55; 'Eracles', p. 460; Marino Sañudo, 'Liber secretorum', p. 224; al-Makrizi, Histoire des sultans, vol. i(b), p. 87; Ibn al-Furat, 'Selections', pp. 152-54. 12. Honorius III, Regesta, ed. P. Pressutti, 2 vols (Rome, 1888-95, reprinted New York, 1978), nos 672-73; Mas Latrie, Histoire, vol. 2, p. 36. 13. See above, p. 102. 14. Megaw, 'A Castle in Cyprus', pp. 45-51. 15. See above, pp. 80-90. Gastria: Enlart, L'ari gothique, vol. 2, p. 656, and see below, p. 131. 16. Willbrand of Oldenburg, 'Itinerarium', p. 230. 17. 'Eracles', pp. 345-46. 18. Megaw, 'A Castle in Cyprus', pp. 48-51; K. Molin, 'Fortifications and Internal Security in the Kingdom of Cyprus, 1191-1426', in From Clermont to Jerusalem: The Crusades and Crusader Societies, 1095-1400, ed. A. V. Murray (Turnhout, i998)> PP-187-9519. For the Hospitallers and the Fifth Crusade, see Riley-Smith, The Knights of St John, pp. 129, 141-44. 20. Megaw, 'Supplementary Excavations', pp. 327, 338-39, 343; Megaw, 'A Castle in Cyprus', p. 48. 1222 earthquake: Oliver of Paderborn, 'Historia Damiatina', p. 279; 'Annales de Terre Sainte', p. 437 (text B). 21. Jacques de Vitry, Lettres, p. 140. 22. D. Jacoby, 'The Rise of a New Emporium in the Eastern Mediterranean: Famagusta in the Late Thirteenth Century', in Jacoby, Studies on the Crusader States and on Venetian Expansion, c. 8, pp. 145-79, at pp. 147-54. 23. 'Gestes', p. 811. 24. Estienne de Lusignan, Description, fol. 24. For more details on the economic importance of Limassol and Famagusta, see below, pp. 290-91. 25. 'Gestes', pp. 848-50; Marino Sañudo, 'Liber secretorum', p. 242; Amadi, Chronique, pp. 234-37; Bustron, Chronique, pp. 129-33. 26. 'Gestes', pp. 852-53; Amadi, Chronique, pp. 238-39; Bustron, Chronique, p. 133.

NOTES TO PAGES 12/-131

345

27. See above, pp. 91-93. 28. William of Machaut, La prise, pp. 68-69; Philip of Mézières, The Life of St Peter Thomas, pp. 125-35; Leontios Makhairas, Recital, cc. 159-73* PP- H3-55; Amadi, Chronique, pp. 413-15; Bustron, Chronique, pp. 262-63; Strambaldi, Chronique, pp. 63-69. 29. William of Machaut, La prise, pp. 20, 121-22; Leontios Makhairas, Recital, cc. 112-14, pp. 99-101, cc. 116-28, pp. 103-13, cc. 132-33, pp. 117-19, c. 143, p. 125, c. 150, pp. 131-33; Philip of Mézières, The Life of St Peter Thomas, pp. 96-97; Amadi, Chronique, pp. 410-15; Bustron, Chronique, pp. 259-60, 263; Strambaldi, Chronique, pp. 42-44, 46-47, 71. On the payment of tribute to Peter, see The Life of St Peter Thomas, p. 127; Leontios Makhairas, Recital, c. 124, p. 109. 30. Leontios Makhairas, Recital, c. 126, p. 111. 31. Leontios Makhairas, Recital, c. 195, p. 177, and see cc. 194-95, PP- 175-7732. Edbury, The Kingdom of Cyprus, pp. 163-64. 33. Enlart, L'ari gothique, vol. 2, pp. 615-18. 34. Leontios Makhairas, Recital, c. 64, p. 61. 35. See above, p. 72. 36. Amadi, Chronique, p. 335; Bustron, Chronique, p. 194; Estienne de Lusignan, Description, fol. 24; Enlart, L'ari gothique, vol. 2, pp. 615-18. 37. Leontios Makhairas, Recital, c. 495, p. 483; Strambaldi, Chronique, p. 205.

Notes to Chapter 12, Cyprus: The Military Orders 1. See above, pp. 59-63. 2. Boniface VIII, Registres, ed. G. Digard et al., 4 vols (Paris, 1884-1939), nos 3060-62, 3114. 3. The Templars helped finance Amaury's takeover; see Amadi, Chronique, p. 248; Barber, The New Knighthood, pp. 289-90; Riley-Smith, The Knights of St John, pp. 210-15. 4. Arrest of Templars: Leontios Makhairas, Recital, cc. 13-17, pp. 11-17; Amadi, Chronique, pp. 280-91; Bustron, Chronique, pp. 163-71; Coureas, The Latin Church in Cyprus, pp. 138-55; A. Iliéva, 'The Suppression of the Templars in Cyprus According to the Chronicle of Leontios Makhairas', in Barber, The Military Orders, pp. 212-19. Conquest of Rhodes: A. Luttrell, 'The Hospitallers of Rhodes, 1306-1421', in HC, vol. 3, pp. 278-313, at pp. 283-86; Riley-Smith, The Knights of St John, pp. 215-16. 5. For an outline of the Templars' activities on Cyprus, see Coureas, The Latin Church in Cyprus, pp. 121-38. 6. Estienne de Lusignan, Description, fol. 123. 7. Innocent III, Die Register Innocenz III, ed. O. Hagender and A. Haidacher, 2 vols. so far (Graz and Cologne, 1964- ), no. 438. Gastria: 'Eracles', pp. 315-16. 8. Hugh III held these properties until 1282. See 'Annales de Terre Sainte', p. 457 (text A); 'Gestes', p. 784; Amadi, Chronique, p. 214; Bustron, Chronique, p. 116. For more details on this dispute, see Edbury, The Kingdom of Cyprus, pp. 93-96;

346

NOTES TO PAGES 131-134

Hill, History of Cyprus, vol. 2, pp. 170—74; Barber, The New Knighthood, pp. 172-739. Bustron, Chronique, pp. 246-47. Gastria: Estienne de Lusignan, Description, fol. 36. 10. Enlart, L'art gothique, vol. 2, p. 656. 11. Bustron, Chronique, pp. 161—71; Enlart, L'art gothique, vol. 2, pp. 671-73. 12. 1279: 'Gestes', p. 784; 'Annales de Terre Sainte', p. 457 (text A); Amadi, Chronique, p. 214; Bustron, Chronique, p. 116.1232: 'Gestes', p. 710. 13. 'A Register of the Cartulary of the Cathedral of Santa Sophia of Nicosia', ed. J. L. LaMonte, Byzantion, 5 (1930), pp. 439-522, at no. 99, pp. 476-77. 14. For an outline of the Hospitallers' activities on Cyprus, see Coureas, The Latin Church in Cypurs, pp. 155-72. 15. Bustron, Chronique, pp. 170-71. The fifteenth-century tower at Khirokitia is referred to by Estienne de Lusignan, Description, fol. 35. See also Hill, History of Cyprus, vol. 2, p. 23; Enlart, L'art gothique, vol. 2, pp. 671-73. 16. 'Gestes', p. 680. 17. 'Gestes', pp. 686-89; Amadi, Chronique, pp. 139-40. 18. Cartulaire, vol. 2, no. 1354, pp. 121-22; Bustron, Chronique, p. 171; Enlart, L'art gothique, vol. 2, pp. 683-94. 19. Estienne de Lusignan, Description, fol. 18. 20. Tabulae ordinis Theutonici, no. 34, pp. 27—28, no. 71, p. 56. 21. See J. Richard, Chypre sous les Lusignans: documents chypriotes des archives du Vatican, XlVe et XVe siècles (Paris, 1962), p. 120. For a detailed discussion of the Teutonic Knights' properties in Cyprus, see W. Hubatsch, 'Der deutsche Orden und die Reichslehnschaft über Cypern', Nachrichten der Akademie der Wissenschaft im Göttingen: Philologisch-Historische Klasse (1955), pp. 245-306; Coureas, The Latin Church in Cyprus, pp. 173-78. 22. 'Gestes', pp. 718-19; Amadi, Chronique, p. 173; Bustron, Chronique, p. 98. Acre: 'Eracles', pp. 373-74, and see above, p. 59. 23. See above, pp. 125-27. 24. Armenia: Acta Aragonensia, ed. H. Pinke, 3 vols (Berlin and Leipzig, 1908-22), vol. 3, p. 146; Riley-Smith, The Knights of St John, pp. 199-200. 25. Riley-Smith, The Knights of St John, pp. 220-26. 26. Mas Latrie, Histoire, vol. 2, p. 90; Riley-Smith, The Knights of St John, p. 198. 27. Enlart, L'art gothique, vol. 2, pp. 684, 694-95. For more details on the sugar industry, see M. -L. von Wartburg, 'The Medieval Cane Sugar Industry in Cyprus: The Results of Recent Excavations', Antiquaries Journal (1983), pp. 298314. 28. Leontios Makhairas, Recital, c. 618, p. 609. For more details on the agricultural uses of fortifications, see chapter twenty-eight.

NOTES TO PAGES 137-139

347

Notes to Chapter 13, Cilician Armenia: Warfare 1. F. C. R. Robinson and P. C. Hughes: 'Lampron. Casúe of Armenian Cilicia', Anatolian Studies, 19 (1969), pp. 183-207, at pp. 188-89. 2. For general histories of Cilician Armenia during the middle ages, see T. S. R. Boase, 'The History of the Kingdom', in T. S. R. Boase (ed.), The Cilician Kingdom of Armenia (Edinburgh, 1978), pp. 1-33; S. der Nersessian, The Armenians (London, 1969); S. der Nersessian, 'The Kingdom of Cilician Armenia', in HC, vol. 2 (1962), pp. 630-59; L. M Alishan, Sissouan ou I'Armeno-Cilicie (Venice, 1899). 3. Constable Sempad, Chronique, ed. Dédéyan, pp. 63-64; Vahram of Edessa, 'Chronique rirnee des rois de la Petite Armenie', in RHCArm, vol. i, pp. 510-11. 4. Constable Sempad, Chronique, ed. Dédéyan, pp. 85-86, 92; Constable Sempad, 'Chronique du royaume de la Petite Armenie', in RHCArm, vol. i, pp. 644-45. 5. Hethoum the Historian, 'Table chronologique', p. 484; Constable Sempad, 'Chronique', RHCArm, vol. i, pp. 644-45. 6. Constable Sempad, 'Chronique', RHCArm, vol. i, p. 645. 7. For a brief outline of Turkish power at this time, see Holt, The Age of the Crusades, pp. 167-77, and particularly p. 172. See also Boase, 'The History of the Kingdom', p. 23. 8. Constable Sempad, Chronique, ed. Dédéyan, pp. 103-4. 9. Samuel of Ani, 'Extrait de la Chronographie de Samuel', in RHCArm, vol. i, p. 461; Boase, 'The History of the Kingdom', p. 25; Holt, The Age of the Crusades, p. 173 . 10. See, for example, the Turkish raids around Corycos in 1318-19 (Constable Sempad, 'Chronique', RHCArm, vol. i, p. 666) and in 1367 (Leontios Makhairas, Recital, cc. 112-13, P- 99)11. Constable Sempad, Chronique, ed. Dédéyan, pp. 117-18; Hethoum the Historian, 'Table chronologique', p. 487; Vahram of Edessa, 'Chronique rimée', pp. 521-22; Samuel of Ani, 'Chronographie', p. 461; Bar Hebraeus, Chronography, p. 446; Marino Sañudo, 'Liber secretorum', p. 223; Ibn al-Furat, 'Selections', p. 99; al-Makrizi, Histoire des sultans, vol. i(b), pp. 31, 33-34. 12. 1275: al-Makrizi, Histoire des sultans, vol. i(b), pp. 123-25; Hethoum the Historian, 'Table chronologique', p. 487; Bar Hebraeus, Chronography, pp. 452-54; Constable Sempad, 'Chronique', RHCArm, vol. i, p. 653; Marino Sañudo, 'Liber secretorum', p. 226; 'Gestes', p. 780. 1298: al-Makrizi, Histoire des sultans, vol. 2(b), pp. 60-65; Samuel of Ani, 'Chronographie', p. 463; 'Gestes', pp. 83940; Marino Sañudo, 'Liber secretorum', pp. 233-34. 1322: Samuel of Ani, 'Chronographie', p. 467. 13- 1337: Samuel of Ani, 'Chronographie', p. 468; Extract from the chronicle of Nerses Balientz in Alishan, Sissouan, p. 469. 1375: Jean Dardel, 'Chronique d'Armenie', in RHCArm, vol. 2, pp. 70-84. 14. Constable Sempad, 'Chronique', RHCArm, vol. i, pp. 646-47, 651; Constable Sempad, Chronique, ed. Dédéyan, pp. 98-100; Vahram of Edessa, 'Chronique

348

NOTES TO PAGES 139-142

rimée', p. 519; Bar Hebraeus, Chronography, p. 418; Boase, 'The History of the Kingdom', p. 25. 15. Cahen, La Syrie du Nord, p. 705. When Baybars conquered this region in 1268, Darbsak was listed as one of the castles he took from the Armenians. See Ibn al-Furat, 'Selections', pp. 126, 166; Abu'1-Fida, 'Annales', p. 152. For a brief outline of the Mongol conquests, see Holt, The Age of the Crusades-, pp. 82-89; Morgan, The Mongols, pp. 147-58. 16. Bar Hebraeus, Chronography, pp. 445-46. 17. Samuel of Ani, 'Chronographie', pp. 463, 466. 18. Samuel of Ani, 'Chronographie', p. 467. 19. For a discussion on the Seljuk army and its possible size, see A. Bombaci, 'The Army of the Saljuqs of Rum', Istituto orientale di Napoli: Annali, new series, 38 (1978), pp. 343-69. 20. Simon de Saint-Quentin, Histoire des Tartares, ed. J. Richard (Paris, 1965), p. 64. 21. Constable Sempad, Chronique, ed. Dédéyan, p. 88; Vahram of Edessa, 'Chronique rimée', p. 511. 22. Vahram of Edessa, 'Chronique rimée', p. 513. 23. Oliver of Paderborn, 'Historia Damiatina', pp. 279-80. See also Constable Sempad, 'Chronique', RHCArm, vol. i, pp. 644-46. 24. 'Gestes', p. 751; Jackson, 'The Crisis', p. 492. For a brief outline of the campaign see Morgan, The Mongols, pp. 147-58. 25. Ayalon, 'The Mamluk Army', particularly at 15 (1953), p. 222 and 16 (1954), pp. 70-71, and see above, p. 7. 26. Vahram of Edessa, 'Chronique rimée', p. 522; Constable Sempad, Chronique, ed. Dédéyan, pp. 117—18; Hethoum the Historian, 'Table chronologique', p. 487; Samuel of Ani, 'Chronographie', p. 461; Bar Hebraeus, Chronography, p. 446. 27. Bar Hebraeus, Chronography, pp. 452-53. 28. al-Makrizi, Histoire des sultans, vol. 2(b), pp. 60-61. 29. Nerses Balientz, extract in Alishan, Sissouan, p. 469. 30. al-Makrizi, Histoire des sultans, vol. 2(b), pp. 60-61. 31. Constable Sempad, 'Chronique', RHCArm, vol. i, p. 644. 32. al-Makrizi, Histoire des sultans, vol. 2(b), pp. 63-64. 33. See above, pp. 8-9. 34. Bar Hebraeus, Chronography, pp. 452-53. 35. Constable Sempad, 'Chronique', RHCArm, vol. i, p. 666. 36. Constable Sempad, 'Chronique', RHCArm, vol. i, pp. 649-50. 37. Constable Sempad, Chronique, ed. Dédéyan, p. 114. 38. Constable Sempad, 'Chronique', RHCArm, vol. i, p. 668. The precision of 'twenty-one knights', as opposed to 'hundreds' or 'thousands' of troops, may also make this one of Constable Sempad's few reliable figures. 39. M. Bartusis, The Late Byzantine Army: Arms and Society, 1204-1453 (Philadelphia, 1992), pp. 213-17. 40. Constable Sempad, Chronique, ed. Dédéyan, pp. 94-95.

NOTES TO PAGES 142-148

349

41. Constable Sempad, 'Chronique', RHCArm, vol. i, p. 661, and see p. 659. 42. Constable Sempad, Chronique, ed. Dédéyan, p. 115. 43. Constable Sempad, Chronique, ed. Dédéyan, pp. 109-10. 44. For details of the royal domain, see below, pp. 159-60, 168. 45. Hethoum the Historian, 'Table chronologique', p. 484; Constable Sempad, 'Chronique', RHCArm, vol. i, pp. 644-45. 46. Constable Sempad, Chronique, ed. Dédéyan, p. 115. 47. Constable Sempad, Chronique, ed. Dédéyan, pp. 117-18; Hethoum the Historian, 'Table chronologique', p. 487; Vahram of Edessa, 'Chronique rimée', p. 522; Samuel of Ani, 'Chronographie', p. 461; Bar Hebraeus, Chronography, p. 446; Boase, 'The History of the Kingdom', pp. 27-28. 48. Bar Hebraeus, Chronography, p. 454. 49. Constable Sempad, 'Chronique', RHCArm, vol. i, p. 661. 50. William of Tyre, Chronicon, vol. i, bk. 13, c. 10, p. 598. 51. Fedden and Thomson, Crusader Castles, pp. 55-56.

Notes to Chapter 14, Cilician Armenia: Military Architecture 1. See above, pp. 138-39. 2. 1299: 'Gestes', pp. 839-40. 1337: Nerses Balientz, extract in Alishan, Sissouan, p. 469. For a brief history of Servantikar, see also T. S. R. Boase, 'Gazetteer' (of sites in Cilician Armenia), in Boase, The Cilician Kingdom of Armenia, p. 179; P. Deschamps, 'Le chateau de Servantikar en Cilicie, le défilé de Marris et la frontière du comté d'Edesse', Syria, 18 (1937), pp. 379-88. 3. Abu-1'Fida, quoted in Deschamps, 'Le chateau de Servantikar', p. 382. 4. This was a very real danger. Servantikar itself was struck by an earthquake in 1269. See Constable Sempad, Chronique, ed. Dédéyan, p. 122. 5. Descriptions based on R. W. Edwards, The Fortifications of Armenian Cilicia, in Dumbarton Oaks Studies, 23 (Washington, DC, 1987), pp. 217-20; H. Hellenkemper, Burgen der Kreutzritterzeit in der Grafschaft Edessa und im Königreich Kleinarmenien (Bonn, 1976), pp. 111-15; Deschamps, 'Le chateau de Servantikar', pp. 381-84; Gaben, La Syrie du Nord, pp. 145-46; Sinclair, Eastern Turkey, vol. 4, pp. 318-20. 6. Robinson and Hughes, 'Lampron: Castle of Armenian Cilicia', pp. 183-207, particularly at pp. 194-95; Edwards, Fortifications, pp. 176-83. 7. J. G. Dunbar and W. W. M. Boal, 'The Castle of Vagha', Anatolian Studies, 14 (1964), pp. 175-84 (machicolation at p. 178); Edwards, Fortifications, pp. 259-65. 8. G. R. Youngs, 'Three Cilician Castles', Anatolian Studies, 15 (1965), pp. 113-34, at pp. 118-25; Edwards, Fortifications, pp. 133-35; Hellenkemper, Burgen, pp. 16568; Boase, 'Gazetteer', p. 165. 9. Youngs, 'Three Cilician Castles', pp. 113-18; Edwards, Fortifications, pp. 255-59; Hellenkemper, Burgen, pp. 188-91; Boase, 'Gazetteer', p. 184. 10. Youngs, 'Three Cilician Castles', pp. 125-33, particularly at pp. 128-30; Edwards, Fortifications, pp. 269-75; Hellenkemper, Burgen, pp. 169-87; Fedden and

35O

NOTES TO PAGES 148-150

Thomson, Crusader Castles, pp. 100-3; Müller-Wiener, Castles of the Crusaders, pp. 77-79; Boase, 'Gazetteer', p. 185. 11. Lampron: Robinson and Hughes, 'Lampron. Castle of Armenian Ciucia', p. 191. Saone: D. Pringle's introduction in Lawrence, Crusader Castles, pp. xxix-xxx. 12. For descriptions of these castles, see above, pp. 98-100. 13. J. G. Dunbar and W. W. M. Boal, 'The Castle of Azgit', in Boase, The Cuidan Kingdom of Armenia, pp. 85-91, at p. 91; Fedden and Thomson, Crusader Castles, pp. 46-47. 14. M. Gough, 'Anazarbus (Anavarza)', Anatolian Studies, 2 (1952), pp. 85-150, particularly at pp. 91,119-25; Edwards, Fortifications, pp. 65-70; Hellenkemper, Burgen, pp. 191-201; Cahen, La Syrie du Nord, p. 152; Boase, 'Gazetteer', p. 153. 15. F. Frech, 'Die armenischen Burgen', Zeitschrift der Gesellschaft für Erdkunde zu Berlin, 9 (1915), pp. 576-80, at p. 578; Hellenkemper, Burgen, pp. 137-39; Alishan, Sissouan, pp. 229-31; Boase, 'Gazetteer', p. 157; Edwards, Fortifications, p. 92. 16. Hellenkemper, Burgen, pp. 202-13; Edwards, Fortifications, pp. 233—36; Alishan, Sissouan, pp. 241-48; Fedden and Thomson, Crusader Castles, pp. 97-100; Müller-Wiener, Castles of the Crusaders, p. 77. 17. Willbrand of Oldenburg, 'Itinerarium', p. 226. 18. Edwards, Fortifications, pp. 244-50; Hellenkemper, Burgen, pp. 140-53; Cahen, La Syrie du Nord, p. 147; Müller-Wiener, Castles of the Crusaders, pp. 75-77; J. Gottwald, 'Die Burg Til im Sudöstlichen Kilikien', BZ, 40 (1940), pp. 82-103; Sinclair, Eastern Turkey, vol. 4, pp. 326-28. 19. Nerses Balientz, extract in Alishan, Sissouan, p. 469. 20. 1266: Ibn al-Furat, 'Selections', p. 99; al-Makrizi, Histoire des sultans, vol. i(b), pp. 31, 33-34; Hethoum the Historian, 'Table chronologique', p. 487; Bar Hebraeus, Chronography, p. 446; Constable Sempad, Chronique, ed. Dédéyan, pp. 117-18; Vahram of Edessa, 'Chronique rimée', pp. 521-22; Marino Sañudo, 'Liber secretorum', p. 223; Samuel of Ani, 'Chronographie', p. 261. 1298: alMakrizi, Histoire des sultans, vol. 2(b), pp. 60-65; Samuel of Ani, 'Chronographie', p. 463; 'Gestes', pp. 839-40; Marino Sañudo, 'Liber secretorum', pp. 233-34. !3O4: al-Makrizi, Histoire des sultans, vol. 2(b), p. 228. The sources for 1298 and 1304 specifically mention Toprak; those for 1266 do not. See also Boase, 'Gazetteer', pp. 183-84. 21. For a description of Montfort, see Pringle 'A Thirteenth-Century Hall', pp. 5456, and see above pp. 18-19. 22. For a description of Silifke, see Edwards, Fortifications, pp. 221-28; Hellenkemper, Burgen, pp. 249-54; Fedden and Thomson, Crusader Castles, pp. 103-5; Müller-Wiener, Castles of the Crusaders, pp. 80-81. For more details on the history and significance of this castle, see below, pp. 179—80. Crac des Chevaliers and Margat: see above, pp. 13-15. 23. Edwards, Fortifications, p. 15. For an example of slot machicolation being used in a Prankish castle, see Deschamps, Le Crac des Chevaliers, p. 182. 24. See above, pp. 143-44. 25. Prankish castles: see above, chapter two. For more details on the differences

NOTES TO PAGES 152-15J

between Armenian and crusader masonry, see Edwards, pp. 20-24.

351

Fortifications,

Notes to Chapter 15, Cuidan Armenia: Fortifications and External Security 1. Gaben, La Syrie du Nord, pp. 139-52; Edwards, Fortifications, pp. 39-40; Boase, 'The History of the Kingdom', p. 2. 2. Constable Sempad, Chronique, ed. Dédéyan, p. 115. 3. 1266 and 1275: see above, pp. 138-39. 4. Constable Sempad, Chronique, ed. Dédéyan, pp. 108-10. 5. Constable Sempad, Chronique, ed. Dédéyan, p. 88; Vahram of Edessa, 'Chronique rimée', p. 511. 6. Constable Sempad, Chronique, pp. 85-86, 92; Constable Sempad, 'Chronique', RHCArm, vol. i, pp. 644-45. 7. al-Makrizi, Histoire des sultans, vol. i(b), p. 33. 8. Constable Sempad, Chronique, ed. Dédéyan, p. 125. 9. Vahram of Edessa, 'Chronique rimée', p. 522. 10. Vahram of Edessa, 'Chronique rimée', p. 528. 11. al-Makrizi, Histoire des sultans, vol. 2(b)3 p. 64. 12. al-Makrizi, Histoire des sultans, vol. 2(b), p. 65. 13. Ibn al-Furat, 'Selections', p. 99; al-Makrizi, Histoire des sultans, vol. i(b), p. 34. 14. Willbrand of Oldenburg, 'Itinerarium', p. 224. See also Hellenkemper, Burgen, p. 264. 15. Matthew Paris, Chronica malora, vol. 4, pp. 299-311, 337-43; 'Rothelin', pp. 56165; 'Eracles', pp. 427-33, and see above, pp. 3, 40. 16. Alishan, Sissouan, pp. 172-73. 17. Boase, 'Gazetteer', pp. 163-64. 18. Jean Dardel, 'Chronique', pp. 70-84; Fedden and Thomson, Crusader Castles, pp. 99-100. 19. See above, pp. 137-39. 20. al-Makrizi, Histoire des sultans, vol. i(b), pp. 54-55; Ibn al-Furat, 'Selections', p. 166; Cahen, La Syrie du Nord, pp. 704-6; Boase, 'The History of the Kingdom', pp. 25-26. Fall of Antioch: Ibn al-Furat, 'Selections', pp. 121—26; al-Makrizi, Histoire des sultans, vol. i(b), pp. 52-54; 'Gestes', pp. 771-72; 'Annales de Terre Sainte', pp. 453-54; Marino Sañudo, 'Liber secretorum', p. 223; 'Eracles', p. 456; Bar Hebraeus, Chronography, p. 448. 21. al-Makrizi, Histoire des sultans, vol. 2(b), pp. 60-61. 22. al-Makrizi, Histoire des sultans, vol. 2(a), pp. 203-12. 23. al-Makrizi, Histoire des sultans, vol. 2(b), pp 190, 228. 24. al-Makrizi, Histoire des sultans, vol. 2(b), pp. 254-55. 25. Nerses Balientz, extract in Alishan, Sissouan, p. 469. 26. Edwards, Fortifications, pp. 244-50. 27. Jean Dardel, 'Chronique', pp. 70-84. 28. Samuel of Ani, 'Chronographie', p. 468.

352

NOTES TO PAGES 157-161

29. See above, pp. 3-4, 7-9. 30. Edwards, Fortifications, p. 38. 31. Edwards, Fortifications, p. 42. 32. Hellenkemper, Burgen, pp. 262-63; Edwards, Fortifications, pp. 41-42. 33. See above, pp. 137-38. 34. Michael the Syrian, 'Chronique', RHCArm, vol. i, p. 405. 35. Vahram of Edessa, 'Chronique rimée', p. 511. See also Hellenkemper, Burgen, p. 262. 36. Gough, 'Anazarbus', pp. 122-23; R- W. Edwards, 'The Crusader Donjon at Anavarza in Ciucia', in Abstracts of the Tenth Annual Byzantine Studies Conference (Cincinnati, 1984), pp. 53-55. Gough argues that this inscription proves that the entire keep is Armenian, but Edwards has shown that it is originally Prankish. See also Edwards, Fortifications, pp. 65-70; Hellenkemper, Burgen, pp. 191—201; Boase, 'Gazetteer', p. 153; Cahen, La Syne du Nord, p. 152. 37. Alishan, Sissouan, p. 400. 38. Youngs, 'Three Cilician Castles', p. 130. 39. Edwards, Fortifications, pp. 33-37, 41-42. 40. See above, p. 148. 41. Constable Sempad, Chronique, ed. Dédéyan, pp. 73-81; Boase, 'Gazetteer', pp. 146-48. 42. Nerses Balientz, extract in Alishan, Sissouan, p. 469. 43. See below, pp. 175-77. 44. See below, pp. 168-74. 45. Leon II: Michael the Syrian, 'Chronique', RHCArm, vol. i, p. 405; Vahram of Edessa, 'Chronique rimée', p. 511. Hethoum I: al-Makrizi, Histoire des sultans, vol. i(b), p. 33. 46. Constable Sempad, Chronique, ed. Dédéyan, pp. 73-81 and 'Chronique', RHCArm, vol. i, pp. 634-40; Boase, 'Gazetteer', pp. 146-48. 47. Gough, 'Anazarbus', pp. 122-23. See also Edwards, 'The Crusader Donjon at Anavarza', pp. 53-55; Edwards, Fortifications, pp. 65—70; Hellenkemper, Burgen, pp. 191-201, 262-63. 48. Edwards, 'The Crusader Donjon at Anavarza', pp. 53-55; Edwards, Fortifications, pp. 69-70; Kennedy, Crusader Castles, pp. 96-97. 49. P. Pattenden, 'The Byzantine Early Warning System', Byzantion, 53 (1983), pp. 258-9950. Estienne de Lusignan, Description, fol. 35V, and see above, p. 116. 51. Vahram of Edessa, 'Chronique rimée', p. 511; al-Makrizi, Histoire des sultans, vol. i(b), p. 33. 52. P. Lock, 'The Towers of Euboea: Lombard or Venetian; Agrarian or Strategic', in P. Lock and G. D. R. Sanders (eds), The Archaeology of Medieval Greece (Oxford, 1996), pp. 107-26, at pp. 107-11, and see below, pp. 251-52. 53. See above, pp. 49-50. 54. 1266: Ibn al-Furat 'Selections', p. 99. For more details on the role of Amouda, see below, pp. 175-79.

NOTES TO PAGES 1Ö1-1Ó5

353

55. Edwards, Fortifications, p. 43. 56. Edwards, Fortifications, p. 45. 57. WUlbrand of Oldenburg, 'Itinerarium', pp. 218-20. 58. Edwards, Fortifications, pp. 43-44, 65-70. 59. Edwards, Fortifications, p. 37, and see pp. 37-47. 60. See above, pp. 152-53, 155-57. 61. For descriptions of Acre and other crusader strongpoints in the Holy Land which needed massive man-made defences to make up for their exposed location on the coastal plain, see chapter two. 62. Le trésor des chartes d'Armenie, no. 18, pp. 141-43, at p. 141. 63. Constable Sempad, Chronique, ed. Dédéyan, p. 76. 64. See above, p. 28. 65. Vagha: Edwards, Fortifications, p. 260. 66. Sinclair, Eastern Turkey, vol. 4, pp. 231-51; Rey, Etude, pp. 184-204; Deschamps, La defense du comté de Tripoli, pp. 46-47, 53. 67. Gregory the Priest, 'Chronique', RHCArm, vol. i, p. 161. 68. See below, pp. 171-72. 69. Constable Sempad, Chronique, ed. Dédéyan, pp. 94-95. 70. Willbrand of Oldenburg, 'Itinerarium', p. 220. 71. Abu'1-Fida, 'Annales', pp. 40 (1170), 79 (1200/01), 83 (1202/03). 72. Constable Sempad, 'Chronique', RHCArm, vol. i, pp. 649-50. 73. It is worth noting that accounts of the First Crusade suggest that Tarsus also had complete walls around it in the 10905. For a description of the First Crusade's passage through this region, see, for example, Albert of Aix, 'Historia Hierosolymitana', in RHCOc, vol. 4, pp. 342-50. Although this evidence should be used with caution when discussing the situation more than a hundred years later, it does at least prove that the city's fortifications had not been in a permanent state of decay for many centuries. 74. 1278: al-Makrizi, Histoire des sultans, vol. i(b), p. 166. 1320: Samuel of Ani, 'Chronographie', p. 467. For full references to the raids in 1266, 1275, 1298, 1302, 1337 and 1375 see above, pp. 138-39, 156. 75. Willbrand of Oldenburg, 'Itinerarium', p. 222. 76. Bar Hebraeus, Chronography, p. 453. 77. Jean Dardel, 'Chronique', p. 73. 78. For full descriptions of the site, see Edwards, Fortifications, pp. 233-36; Hellenkemper, Burgen, pp. 202-13; Alishan, Sissouan, pp. 241-48; Fedden and Thomson, Crusader Castles, pp. 97-99; Müller-Wiener, Castles of the Crusaders, P-7779. For details on trade patterns in this area, see Cahen, La Syne du Nord, pp. 689-91. The role of the fortifications at Ayas and Corycus in protecting trade routes will be discussed in more detail below, p. 292. 80. Corycus: Edwards, Fortifications, pp. 161-66; Alishan, Sissouan, pp. 397-402; Müller-Wiener, Castles of the Crusaders, pp. 79-80. Ayas: Edwards, Fortifications, pp. 77-81; Alishan, Sissouan, pp. 432-34.

354

NOTES TO PAGES 165-170

81. Alishan, Sissouan, p. 400. 82. Bar Hebraeus, Chronography, p. 465. 83. Bar Hebraeus, Chronography, p. 453. 84. Constable Sempad, 'Chronique', RHCArm, vol. i, pp. 667-68. See also A. Luttrell, 'The Hospitallers' Interventions in Cilician Armenia, 1291—1375', in Boase, The Cilician Kingdom of Armenia, pp. 118-44, at p. 139.

85. Le trésor des chañes d'Armenie, no. 8, p. 233. 86. Richard, Chypre sous les Lusignans, p. 37. 87. Nerses Balientz, extract in Alishan, Sissouan, p. 469; Luttrell, 'The Hospitallers' Interventions', p. 139. 88. Le trésor des charles d'Armenie, Genoa: no. i, pp. 105-8, no. 10, pp. 126-28, no. 15, pp. 136-37, no. 26, pp. 154-61, no. 27, p. 162. Venice: no. 2, pp. 109-12, no. 19, pp. 143-45, no. 25, pp. 151-54, no. 31, pp. 166-68, no. 36, pp. 182-85, no. 40, pp. 193-94. Pisa: no. 16, pp. 138-39. 89. Le trésor des diartes d'Armenie^ Sicily: no. 38, pp. 186-89. Montpellier: no. 34, p. 178, no. 37, p. 185. Catalonia: no. 28, p. 163. The Bardi: no. 41, p. 195. 90. Quote based on Marco Polo, The Travels, trans. R. Latham, Penguin Classics (Harmondsworth, 1958), p. 46.

Notes to Chapter 16, Cilician Armenia: Fortifications and Internal Security 1. Constable Sempad, Chronique, ed. Dédéyan, pp. 73-81 and 'Chronique', RHCArm, vol. i, pp. 634-40; Boase, 'Gazetteer', pp. 146-48; Hellenkemper, Burgen, p. 258. 2. For a brief history of the region during the eleventh and twelfth centuries, see Boase, 'The History of the Kingdom', pp. 1-19; der Nersessian, 'The Kingdom of Cilician Armenia', pp. 630-51. 3. Robinson and Hughes, 'Lampron: Castle of Cilician Armenia', pp. 183-207; Edwards, Fortifications, pp. 176-83. 4. For a brief history of Lampron, see Robinson and Hughes, 'Lampron: Castle of Cilician Armenia', pp. 183-88. 5. Constable Sempad, 'Chronique', RHCArm, vol. i, p. 624; Vahram of Edessa, 'Chronique rimée', p. 509. 6. Leon the Great, cited in Robinson and Hughes, 'Lampron: Castle of Armenian Cilicia', p. 183. 7. Constable Sempad, Chronique, ed. Dédéyan, pp. 73-81 and 'Chronique', RHCArm, vol. i, pp. 634-40; Boase, 'Gazetteer', pp. 146-48. 8. Constable Sempad, Chronique, ed. Dédéyan, pp. 81-82. 9. Constable Sempad, 'Chronique', RHCArm, vol. i, p. 628. See also Vahram of Edessa, 'Chronique rimée', pp. 509-10. Later Leon seized Bohemond III at yet another supposedly friendly meeting and held him to ransom for a short while. See Constable Sempad, 'Chronique', RHCArm, vol. i, pp. 631-32; Boase, 'The History of the Kingdom', pp. 14-19. 10. Constable Sempad, Chronique, ed. Dédéyan, p. 73.

NOTES TO PAGES 170-1/3

355

u. Constable Sempad, Chronique, ed. Dédéyan, p. 77. 12. See above, p. 166. 13. See below, pp. 175-82. 14. Le trésor des diartes d'Armenie, no. 3, pp. 112-14; Cartulaire, vol. 2, no. 1351, p. 119. 15. Constable Sempad, Chronique, ed. Dédeyan, p. 85; Hethoum the Historian, 'Table Chronologique', p. 481; Boase, 'The History of the Kingdom', pp. 21, 23-25. 16. Le trésor des chartes d'Armenie, no. 8, pp. 122-23; Cartulaire, vol. 2, no. 1426, pp. 464-65. 17. Le trésor des chartes d'Armenie, no. 6, pp. 117-20 (Amouda), no. 18, pp. 141-43 (Harunia); Tabulae ordinis Theutonici, no. 46, pp. 37-39 (Amouda), no. 83, pp. 65-66 (Harunia). 18. Constable Sempad, Chronique, ed. Dédéyan, pp. 73-81; Boase, 'Gazetteer', p.146. 19. See, for example, Smail, 'Crusaders' Castles in the Twelfth Century', pp. 147-49, and see above, pp. 59-60, 79-86. 20. For a brief history of the region during the thirteenth and fourteenth centuries, see Boase, 'The History of the Kingdom', pp. 19-33; der Nersessian, 'The Kingdom of Cilician Armenia', pp. 645-59. 21. Constable Sernpad, Chronique, ed. Dédéyan, pp. 94-95; Bar Hebraeus, Chronography, pp. 379-80; Hethoum the Historian, 'Table chronologique', p. 485; Vahram of Edessa, 'Chronique rirnee', p. 514; Boase, 'The History of the Kingdom', pp. 22-23; Cahen, La Syrie du Nord, pp. 631-32. 22. For an outline of this conflict, see Cahen, La Syrie du Nord, pp. 596-631. It will also be discussed in more detail below. See pp. 180-81. 23. Constable Sempad, Chronique, ed. Dédéyan, pp. 93-94, 95-96; Constable Sempad, 'Chronique', RHCArm, vol. i, pp. 647-48; Hethoum the Historian, 'Table chronologique', p. 485; Samuel of Ani, 'Chronographie', p. 460; Bar Hebraeus, Chronography, pp. 375, 379-81, 389-90; Boase, 'The History of the Kingdom', pp. 23-25. 24. Bar Hebraeus, Chronography, p. 381; Cahen, La Syrie du Nord, p. 635; J. S. C. Riley-Smith, 'The Templars and the Teutonic Knights in Cilician Armenia', in Boase, The Cilician Kingdom of Armenia, pp. 92-117, at p. 113. 25. Boase, Gazetteer', p. 180. 26. Bar Hebraeus, Chronography, p. 449, and see pp. 449-50; Vahram of Edessa, 'Chronique rimée', p. 527; Constable Sempad, Chronique, ed. Dédéyan, p. 125. 27. Samuel of Ani, 'Chronographie', pp. 464-65; Hethoum the Historian, 'Table Chronologique', pp. 489-90; Boase, 'The History of the Kingdom', p. 29. 1307: Samuel of Ani, 'Chronographie', pp. 463, 466. 28. Constable Sempad, 'Chronique', RHCArm, vol. i, pp. 670-71; Jean Dardel, 'Chronique', pp. 19-21; Boase, 'The History of the Kingdom', pp. 30, 32. 29. Samuel of Ani, 'Chronographie', pp. 465-67. 30. Mardiros of Crimea, 'Liste rimée des souverains de la Petite Armenie', in RHCArm, voi. i, p. 685. For these two rulers, see also the rather unreliable

356

NOTES TO PAGES 173-1/7

and romanticised accounts in Jean Dardel, 'Chronique', pp. 20-28, 21 n. 5, and see Boase, 'The History of the Kingdom', pp. 30-31. 31. Leontios Makhairas, Recital cc. 112—13, P- 99! Boase, 'The History of the Kingdom', p. 31; P. Edbury, The Kingdom of Cyprus, pp. 163-64. As we have seen (above, pp. 127-28), during the 13605 Peter I was closely involved in the struggle against the Turks who were attacking Ciucia. 32. Jean Dardel, 'Chronique', pp. 70-84, and see pp. 16-85 for a full account of fourteenth-century Cilician Armenia. This source should be used with caution because of its colourful and romanticised style. 33. Holy Land: see above, pp. 59-63. Cyprus: see above, pp. 110-14, 130-34. 34. Bar Hebraeus, Chronography, p. 389.

Notes to Chapter 17, Cilician Armenia: The Military Orders 1. Constable Sempad, Chronique, ed. Dédéyan, p. 73; Constable Sempad, 'Chronique', RHCArm, vol. i, p. 634; Gaben, La Syrie du Nord, pp. 588-90; K. Forstreuter, Der deutsche Orden am Mittelmeer (Bonn, 1967), p. 59; RileySmith, 'The Templars and the Teutonic Knights', p. m. 2. Le trésor des chartes d'Armenie, no. 6, pp. 117-20; Tabulae ordinis Theutonici, no. 46, pp. 37-39. 3. Willbrand of Oldenburg, 'Itinerarium', pp. 222-24; Riley-Smith, 'The Tempiars and the Teutonic Knights', pp. 111-13. 4. Amouda: Edwards, Fortifications, pp. 59-61; Hellenkemper, Burgen, pp. 123-31; Forstreuter, Der deutsche Orden, p. 61. Monfort: Pringle, 'A Thirteenth-Century Hall', pp. 54-56. Judin: Pringle, Petersen, Dow and Singer, 'Qal'at Jiddin', pp. 135-66, particularly at pp. 159-62. 5. 1198: Constable Sempad, Chronique, ed. Dédéyan, p. 76. 1236: Le trésor des chartes d'Armenie, no. 18, pp. 141-43; Tabulae ordinis Theutonici, no. 83, pp. 6566. 6. Riley-Smith, 'The Templars and the Teutonic Knights', pp. 113-14. Frederick II's troops and supporters held Tyre in the kingdom of Jerusalem throughout the 12305. See above, pp. 9-10. 7. Edwards, Fortifications, pp. 143-46; Hellenkemper, Burgen, pp. 116-19; Sinclair, Eastern Turkey, vol. 4, p. 328. 8. Willbrand of Oldenburg, 'Itinerarium', p. 220; Le trésor des chartes d'Armenie, no. 6, p. 119. 9. Document trans, in Alishan, Sissouan, p. 239. 10. Edwards, Fortifications, pp. 147-49. See also Riley-Smith, 'The Templars and the Teutonic Knights', p. 114. 11. See above, pp. 171-72. 12. Le trésor des chartes d'Armenie, no. 18, p. 142. 13. For example, in 1187: Constable Sempad, Chronique, ed. Dédéyan, pp. 63-64. 14. Constable Sempad, 'Chronique', RHCArm, vol. i, pp. 648-49; Cahen, La Syrie du Nord, p. 651; der Nersessian, 'The Kingdom of Cilician Armenia', p. 652.

NOTES TO PAGES 177~l8l

357

15. 1266: Constable Sempad, Chronique, ed. Dédéyan, pp. 117-18. 16. See above, pp. 138—39, 154. 17. Forstreuter, Der deutsche Orden, pp. 62-63. For more details on the Teutonic Knights' properties in the Holy Land, see above, pp. 79-80. 18. See above, pp. 171-72. 19. Le trésor des charles d'Armenie, no. 18, p. 141. 20. Willbrand of Oldenburg, 'Itinerarium', p. 224.1266: Ibn al-Furat, 'Selections', p. 99; al-Makrizi, Histoire des sultans, vol. i(b), p. 34. 21. Le trésor des ¿hartes d'Armenie, no. 6, pp. 117-20, no. 18, pp. 141-43; Tabulae ordinis Theutonici, no. 46, pp. 37-39, no. 47, pp. 39-40, no. 83, pp. 65-66; Riley-Smith, 'The Templars and the Teutonic Knights', pp. 113-14. 22. Edwards, Fortifications, p. 59. 23. Alishan, Sissouan, p. 239; Hellenkemper, Burgen, p. 263. 24. Le trésor des diartes d'Armenie, no. 3, pp. 112-14; Cartulaire, vol. 2, no. 1351, p. 119. 25. Edwards, Fortifications, pp. 221-27; Hellenkemper, Burgen, pp. 249-54; Fedden and Thomson, Crusader Castles, pp. 103-5; Müller-Wiener, Castles of the Crusaders, pp. 80-81. 26. Le trésor des ¿hartes d'Armenie, no. 3, pp. 112-14; Cartulaire, vol. 2, no. 1351, p. 119; Boase, 'Gazetteer', pp. 158, 175. 27. See above, p. 138. 28. Constable Sempad, 'Chronique', RHCArm, vol. i, p. 645. 29. Crac des Chevaliers and Margat: see above, pp. 13-15. 30. Constable Sempad, 'Chronique', RHCArm, vol. i, p. 646. 31. Constable Sempad, Chronique, ed. Dédéyan, p. 85. 32. Le trésor des ¿hartes d'Armenie, no. 9, pp. 124-25; Cartulaire, vol. 2, no. 1427, pp. 165-66. 33. Le trésor des ¿hartes d'Armenie, no. 8, pp. 122-23; Cartulaire, vol. 2, no. 1426, pp. 464-65; Hellenkemper, Burgen, pp. 165-68; Boase, 'Gazetteer', p. 165; Youngs, 'Three Cilician Castles', pp. 118, 125. 34. For more details on this entire dispute, see Cahen, La Syrie du Nord, pp. 596631. 35. J. S. C. Riley-Smith, The Knights of St John, pp. 152-55. 36. Le trésor des ¿hartes d'Armenie, no. 11, pp. 130-31; Cartulaire, voi. 2, nos 1262-63, pp. 70-71. 37. Le trésor des ¿hartes d'Armenie, no. 12, pp. 132-33; Cartulaire, voi. 2, no. 1355, pp. 122-23. 38. 'Eracles', p. 318; 'Ármales de Terre Sainte', pp. 436-37; Hethoum the Historian, 'Table chronologique', pp. 483-84; 'Gestes', p. 664; Constable Sempad, Chronique, ed. Dédéyan, pp. 89-90; Cahen, La Syrie du Nord, pp. 621-22,630-31. For more details on the Hospitallers' involvement in these events, see RileySmith, The Knights of St John, pp. 152-61. 39. See above, pp. 69-72, 80-81, 84-85. 40. Willbrand of Oldenburg, 'Itinerarium', p. 224; Riley-Smith, The Knights of St John, p. 157.

358

NOTES TO PAGES 182-185

41. Oliver of Paderborn, 'Historia Damiatina', p. 279. See also 'Eracles', p. 347. 42. See above, pp. 171-72, and Riley-Smith, The Knights of St John, pp. 158-60. 43. Constable Sempad, 'Chronique', RHCArm, vol. i, p. 648. 44. Riley-Smith, 'The Templars and the Teutonic Knights', pp. 92-97. See also A. W. Lawrence, 'The Castle of Baghras', in Boase, The Cuidan Kingdom of Armenia, pp. 34-83, at pp. 41-43. 45. Abu Shama, 'Deux jardins', RHCOr, vol. 4, pp. 378-79; Ibn al-Athir, 'Kamel Altevarykh', RHCOr, vol. i, pp. 731-32; 'Eracles', p. 136; Constable Sempad, 'Chronique', RHCArm, vol. i, p. 636; Vahram of Edessa, 'Chronique rimée', p.512; Bar Hebraeus, Chronography, pp.336-37; Lawrence, 'Baghras', pp.44, 45; Riley-Smith, 'The Templars and the Teutonic Knights', pp. 97-98. 46. For full descriptions of the site, see Lawrence, 'Baghras', pp. 49-83; R. W. Edwards, 'Bagras and Armenian Ciucia: A Reassessment', Revue des études arméniennes, 17 (1983), pp. 415-55, at pp. 419-32. 47. Lawrence, 'Baghras', pp. 55-56, 62-63. 48. Edwards, 'Bagras and Armenian Cilicia', pp. 420-23, 426, 429. 49. Edwards, 'Bagras and Armenian Cilicia', pp. 431-32. 50. Ibn al-Athir, 'Kamel Altevarykh', vol. i, p. 732. 51. Willbrand of Oldenburg, 'Itinerarium', p. 216. 52. Innocent III, 'Opera Omnia', ed. J. P. Migne, PL, vol. 214, lib. 2, no. 259, cols 819-20. 53. For full accounts of the arguments over Baghras and the Antiochene succession dispute between the 11905 and 1216, see Cahen, La Syrie du Nord, pp. 596-623; Riley-Smith, 'The Templars and the Teutonic Knights', pp. 97-107. 54. 'Gestes', p. 663; 'Annales de Terre Sainte', p. 435 (text B); Innocent III, PL, vol. 215, lib. 8, no. 119, col. 689. See also Marino Sañudo, 'Liber secretorum', p. 203; Hethoum the Historian, 'Table chronologique', p. 480; 'Eracles', p. 257. 55. Innocent III, PL, vol. 215, lib. 7, no. 189, col. 504, lib. 8, no. 119, cols 689-90. 56. Ibn al-Athir, 'Kamel Atevarykh', RHCOr, vol. 2, pp. 317-18. 57. 'Eracles', pp. 317-18; Innocent III, PL, vol. 216, lib. 14, no. 64, cols 430-32. 58. Innocent III, PL, vol. 216, lib. 16, no. 7, cols 792-93; Riley-Smith, 'The Templars and the Teutonic Knights', p. 107. 59. Willbrand of Oldenburg, 'Itinerarium', p. 216. 60. Ibn al-Athir, 'Kamel Altevarykh', RHCOr, vol. i, p. 732. 61. Ibn al-Athir, 'Kamel Altevarykh', RHCOr, vol. 2, pp. 168-70; Constable Sempad, 'Chronique', RHCArm, vol. i, p. 648; Riley-Smith, 'The Templars and the Teutonic Knights', p. 109. 62. 1188: Abu Shama, 'Deux jardins', RHCOr, vol. 4, pp. 376-77; Ibn al-Athir, 'Kamel Altevarykh', RHCOr, vol. i, pp. 730-31; Cahen, La Syrie du Nord, pp. 144-4563. 1205: Ibn al-Athir, 'Kamel Altevarykh', RHCOr, vol. 2, pp. 317-18.1237: Matthew Paris, Chronica maiora, vol. 3, pp. 404-6; Abu'1-Fida, 'Annales', pp. 112-13.1260: Cahen, La Syrie du Nord, p. 705.

NOTES TO PAGES 185-188

359

64. Cahen, La Syrië du Nord, pp. 142-43; Sinclair, Eastern Turkey, vol. 4, pp. 314-16. 65. I have followed Cahen's argument (La Syrie du Nord, pp. 143-45) that the ruined fortress of Chilvan Kale should be identified as Roche Roussel rather than Roche Guillaume. The latter castle may be another ruined stronghold situated further east. See Boase, 'Gazetteer', pp. 159, 178. 66. Boase, 'Gazetteer', p. 177; Cahen, La Syrie du Nord, p. 141. 67. 'Eracles', pp. 405-6; Cahen, La Syrie du Nord, pp. 651-52. 68. Ihn al-Furat, 'Selections', p. 99. See also al-Makrizi, Histoire des sultans, vol. i(b), p. 34. 69. Antioch: 'Eracles', p. 318; 'Annales de Terre Sainte', pp. 436-37; Hethoum the Historian, 'Table chronologique', pp. 483-84; 'Gestes', p. 664; Constable Sempad, Chronique, ed. Dédéyan, pp. 89-90. 70. Innocent III, PL, vol. 214, cols 1005-6. 71. Riley-Smith, 'The Templars and the Teutonic Knights', p. 108, from Innocent III, PL, vol. 214, col. 1005, vol. 215, col. 504. 72. Riley-Smith, 'The Templars and the Teutonic Knights', p. no. 73. For example, in 1237 local Templars made a failed attempt to recapture Darbsak from the Muslims. See Matthew Paris, Chronica maiora, vol. 3, pp. 404-6; Abu'1-Fida, 'Annales', pp. 112-13. 74. Document trans, in Alishan, Sissouan, p. 239. 75. 'Gestes', p. 839. 76. Ihn al-Furat, 'Selections', p. 99; al-Makrizi, Histoire des sultans, vol. i(b), p. 34. 77. al-Makrizi, Histoire des sultans, vol. 2(b), p. 61; Riley-Smith, 'The Templars and the Teutonic Knights', p. 115. 78. 'Gestes', p. 839. 79. Ibn al-Furat, 'Selections', pp. 121-26; al-Makrizi, Histoire des sultans, i(b), pp. 52-54; 'Gestes', pp. 771-72; 'Annales de Terre Sainte', pp. 453-54; Marino Sañudo, 'Liber secretorum', p. 223; 'Eracles', p. 456; Bar Hebraeus, Chronography, p. 448. 80. Luttrell, 'The Hospitallers' Interventions', p. 122 n. 17. 81. J. Upton-Ward, 'The Surrender of Gaston and the Rule of the Templars', in Barber, The Military Orders, pp. 179-88. 82. 1233: Le trésor des charles d'Armenie, no. 17, p. 140. 83. Ibn al-Furat, 'Selections', p. 99. 84. Hethoum the Historian, 'Table chronologique', p. 487. For more details on the Hospitallers' involvement in this and other thirteenth-century Armenian campaigns, see Luttrell, 'The Hospitallers' Interventions', pp. 119-22; RileySmith, The Knights of St John, p. 161. 85. Acta Aragonensia, vol. 3, p. 146. 86. Leontios Makhairas, Recital cc. 116—28, pp. 103-13, cc. 132-33, pp. 117-19, c. 143, p. 125, c. 150, pp. 131-33, and see above, pp. 127-28. For more details on the Hospitallers' involvement in these and other fourteenth-century campaigns, see Luttrell, 'The Hospitallers' Interventions', pp. 122-33.

360

NOTES TO PAGES 188-197

87. For more details on these events, see Riley-Smith, The Knights of St John, pp. 198-226; Luttrell, 'The Hospitallers' Interventions', pp. 124-28, and see above, pp. 91, 112-13, 130. 88. Luttrell, 'The Hospitallers' Interventions', p. 128. 89. RHCArm, vol. i, p. xxxiv; Luttrell, 'The Hospitallers' Interventions', pp. 128-29. Notes to Chapter 18, Prankish Greece: Warfare 1. For an outline of the Fourth Crusade and its causes, see D. E. Queller and T. F. Madden, The Fourth Crusade: The Conquest of Constantinople, second edition (Philadelphia, 1997); C. M. Brand, Byzantium Confronts the West (Cambridge, Massachusetts, 1968), especially pp. 232-69, and see the introduction to this book. 2. For an outline of these events, see P. Lock, The Franks in the Aegean, 1204—1500 (London, 1995), especially pp. 35-107, 135-60. 3. For an outline of these events, see Lock, The Franks, pp. 60-134; R- L. Wolff, 'The Latin Empire of Constantinople, 1204-61', in HC, vol. 2, pp. 187-233; P. Topping, 'The Morea, 1311-1364', and 'The Morea, 1364-1460', in HC, vol. 3, pp. 104-40,141-66; K. M. Setton, 'The Catalans in Greece, 1311-1380', in HC, vol. 3, pp. 167-224. 4. For an outline of these events, see Lock, The Franks, pp. 84-86, 92-95,127-34, 155-60; Topping, 'The Morea, 1311-1364', pp. 104-40 and 'The Morea, 13641460', pp. 141-66. 5. B. Hendrickx, 'A propos du nombre des troupes de la Quatrième Croisade et de Fempereur Baudoin I', Byzantina, 3 (1971), pp. 29-41; Bartusis, The Late Byzantine Army, p. 262. 6. Geoffrey of Villehardouin, La conquête de Constantinople, ed. N. de Wailly (Paris, 1882), pp. 180, 184. 7. 1206: Villehardouin, La conquête, p. 256. 1211: Letter from Henry to the West, dated January 1212, in RHGF, vol. 18, p. 533. 8. Bartusis, The Late Byzantine Army, p, 262. 9. Henry of Valenciennes, Histoire de l'empereur Henri de Constantinople, ed. J. Longnon (Paris, 1948), pp. 46-47; letter from the emperor Henry to Pope Innocent III, September 1208, in RHGF, vol. 19, p. 514. 10. Philip Mousket, 'Chronique rimée', ed. baron de Reiffenberg, in Collection de chroniques belges inédites, 2 vols (Brussels, 1838), vol. 2, pp. 614-16; George Akropolites, 'Annales', ed. I. P. Migne, in PG, vol. 140, cols 1058-59. 11. Villehardouin, La conquête, p. 202. 12. Villehardouin, La conquête, pp. 198-230, especially at p. 208. Quote: p. 250. 13. Villehardouin, La conquête, p. 200. 14. Niketas Choniates, 'Historia', ed. J. P. Migne, in PG, vol. 139, cols 983-87, 1023-27. For the conquest of northern Greece, see also Villehardouin, La conquête, pp. 160-78; Robert of Clari, La conquête de Constantinople, ed.

NOTES TO PAGES 19/-2OO

361

P. Lauer (Paris, 1924), pp. 96-105; George Akropolites, 'Ármales', cois 9991OO2.

15. See, for example, Livre de la conqueste de la princée de l'Amorée: chronique de Marèe, 1204-1305, ed. J. Longnon (Paris, 1911), pp. 34-35; Libro de los fechos et conquistas del principado de la Marea, ed. A. Morel-Fatio (Geneva, 1885), p. 31. See also Lock, The Franks, pp. 68-74, 284. 16. Marino Sañudo, 'Istoria del regno di Romania sive regno di Morea', ed. C. Hopf, in Chroniques gréco-romanes inédites ou peu connues (Berlin, 1873), p. IDI; J. Longnon, L'empire latin de Constantinople et de la principante de Marèe (Paris, 1949), pp. 203-4. 17. To Chroniken ton Mareos. The Chronicle of Marea, ed. J. Schmitt (London, 1904). P- 446. 18. Aubrey of Trois Fontaines, 'Chronica', ed. P. Scheffer-Boichorst, in MGH SS, vol. 23 (1874), pp. 938-39; Philip Mousket, 'Chronique rimee', p. 620. 19. Marino Sañudo, 'Istoria', p. 102; Joinville, Histoire de Saint Louis, pp. 76-78. 20. Livre de la conqueste, pp. 67-68; Libro de los fechos et conquistas, pp. 48-49. 21. Bartusis, The Late Byzantine Army, p. 265. 22. Livre de la conqueste, pp. 124-31; To Chromkon tou Mareos, pp. 300-31. 23. See, for example, To Chronikon tou Mareos, p. 308. 24. Honorius III, Regesta, vol. 2, no. 4758, p. 207, no. 5464, p. 333; Gregory IX, Registres, ed. L. Auvray, 3 vols (Paris, 1890-1955), vol. 2, nos 2671, 3214; D. M. Nicol, The Despotate ofEpiros (Oxford, 1957), pp. 62-64; W. Miller, The Latins in the Levant: A History of Prankish Greece, 1204-1566 (London, 1908), pp. 84-85. 25. Livre de la conqueste, pp. 247-53; To Chronikon tou Mareos, pp. 582-89; Bartusis, The Late Byzantine Army, p. 265. 26. Bartusis, The Late Byzantine Army, p. 269. 27. Ramon, Muntaner, Crònica, ed. M. Gusta, 2 vols (Barcelona, 1979), vol. 2, p. 69. 28. 'Directorium ad faciendum passagium transmarinum', ed. C. Beazley, American Historical Review, 12 (1906-7), pp. 810-57 and 13 (1907-8), pp. 66-115, at 13 (1907-8), p. 80. 29. Muntaner, Crònica, vol. 2, pp. 122-24. 30. Muntaner, Crònica, vol. 2, p. 103. 31. Nouvelles recherches historiques sur la principante francaise de Marèe et ses hautes baronnies a la suite de la Quatrìème Croisade, ed. J. A. C. Buchón, 2 vols (Paris, 1843), vol. 2, no. 25, pp. 143-53. See also Pryor, Geography, Technology and War, pp. 165-92, especially pp. 165-76; A Luttrell, 'The Latins and Life on the Smaller Aegean Islands, 1204-1453', in B. Arbel, B. Hamilton and D. Jacoby (eds), Latins and Greeks in the Eastern Mediterranean after 1204 (London, 1989), pp. 146-57, at p. 147 32. Marino Sañudo, 'Istoria', pp. 113-14. 33. Pryor, Geography, Technology and War, pp. 153-61. 34. Andrea Dandolo, 'Chronicon Venetum', ed. E. Pastorello, in RIS n.s., vol. 12,

302

NOTES TO PAGES 2OO-2O4

part i (1938), pp. 280 (Crete), 283, 367 (Corfu); Marino Sañudo, 'Vite de' duchi di Venezia', ed. L. A. Muratori, in RIS, vol. 22 (1733), cols 533, 536 (Crete); Andrea Navagerio, 'Storia Venezia', ed. L. A. Muratori, in RIS, vol. 23 (1733), cols 984-89 (Crete and Corfu); Martin da Canal, 'La chronique des Veneciens', ed. G. Galvani, in Archivio storico italiano, vol. 8 (1845), p. 351 (Crete). See also J. K. Fotheringham, Marco Sañudo, Conqueror of the Archipelago (Oxford, 1915), pp. 81-87. After the Fourth Crusade, Crete was originally claimed by Boniface of Montferrat, but then sold by him to the Venetians. See Urkunden zur älteren Handels- und Staatsgeschichte der Republik Venedig mit besonderer Beziehung auf Byzanz und die Levante, ed. G. Tafel and G. Thomas, 3 vols (Vienna, 1856-57), vol. i, pp. 512-15. 35. Livre de la conqueste, pp. 81-87; Libro de los fechos et conquistas, pp. 49-52; Marino Sañudo, 'Istoria', pp. 103-6; Andrea Dandolo, 'Chronicum Venetum', p. 306. The final peace treaty ending this war was not signed until 1262; see Tafel and Thomas, Urkunden, vol. 3, pp. 46-55. See also Longon, L'empire latin, pp. 220-23; Lock, The Franks, pp. 88-92. 36. See below, pp. 253-57. 37. Luttrell, 'The Latins and Life on Smaller Aegean Islands', p. 146. Marino Sañudo ('Istoria', p. 143) noted that Latins on other Aegean islands were also heavily outnumbered. 38. Villehardouin, La conquete, pp. 92, 142. 39. 1235: Philip Mousket, 'Chronique rimée', pp. 614—16; Martin da Canal, 'La chronique des Veneciens', pp. 363-67; Andrea Dandolo, 'Chronicon Venetum', p. 295; Andrea Navagerio, 'Storia Venezia', col. 992; George Akropolites, 'Annales', cols 1058—59. 1236: Aubrey of Trois Fontaines, 'Chronica', pp. 938-39; Philip Mousket, 'Chronique rimée', p. 620. 40. Muntaner, Crònica, vol. 2, pp. 85-87. For further examples of the great skill and ferocity of the Catalans and Aragonese in naval warfare, see J. H. Pryor, 'The Naval Battles of Roger of Lauria', in J. H. Pryor, Commerce, Shipping and Naval Warfare in the Medieval Mediterranean (London, 1987), c. 6, pp. 179-216. 41. Bartusis, The Late Byzantine Army, p. 68. 42. Marino Sañudo, 'Istoria', p. 132. 43. Muntaner, Crònica, vol. 2, p. 74, and see Pryor, Geography, Technology and War, pp. 165-73.

Notes to Chapter 19, Prankish Greece: Military Architecture 1. 2. 3. 4.

Bon, La Marèe franque, p. 267. Bon, La Marèe franque, pp. 637-39. Bon, La Marèe franque, pp. 638-39; Libro de los fechos et conquistas, p. 49. Bon, La Marèe franque, pp. 637, 638; C. Foss and D. Winfield, Byzantine Fortifications: An Introduction (Pretoria, 1986), pp. 30-31, 162-64. 5. Bon, La Marèe franque, p. 645. For the later history of Androusa, see ibid., pp. 411-12; Miller, The Latins in the Levant, pp. 391-92, 448-49.

NOTES TO PAGES 2O4-2OÓ

363

6. Bon, La Marèe franque, p. 639. The following scholars all agree with Bon: K.Andrews, Castles of the Marea (Princeton, 1953), pp. 161, 173; S. Runciman, Mistra (London, 1980), pp. 29-30, 95; Müller-Wiener, Castles of the Crusaders, pp. 84-85. For a description and history of Mistra see Andrews, Castles, pp. 161-82, and, in far greater detail, Runciman, Mistra, pp. 9-146. See also To Chronikon tou Mareos, p. 200; Libro de los fechos et conquistas, p. 49. 7. Foss and Winfield, Byzantine Fortifications, pp. 30, 34. 8. For more details, see Andrews, Castles, pp. 135,137-38,143-45 (Corinth); pp. 90, 91-92, 94-105 (Nauplia); pp. 192—96, 198-99, 209-10 (Monemvasia); pp. 116, 117-19, 129 (Patras). 9. A. Bon, 'Fortéresses medievales de la Grece centrale', Bulletin de correspondance hellénique, 61 (1937), pp. 136-208, at p. 163. 10. Andrews, Castles, pp. 74-78, 81. 11. Villehardouin, La conquête, p. 190. 12. Foss and Winfield, Byzantine Fortifications, pp. 154-55. 1225: Nikephoros Gregoras, 'Nicephori Gregorae Byzantina historia', ed. L. Schopen, in CSHB, 3 vols (Bonn, 1829-55), vol. i, p. 25; Aubrey of Trois Fontaines, 'Chronica', p. 911; George Akropolites, 'Annales', cols 1038-39. 13. Foss and Winfield, Byzantine Fortifications, pp. 41-70; A. van Millingen, Byzantine Constantinople: The Walls of the City and Adjoining Sites (London, 1899), pp. 51-58, 122-27, 164-7414. Villehardouin, La conquête, p. 134, and see ibid., pp. 90-102; Robert of Clari, La conquête, pp. 44-51; Niketas Choniates, 'Historia', cois 923-27. 15. Niketas Choniates, 'Historia', col. 926; Villehardouin, La conquête, pp. 96-100; Robert of Clari, La conquête, pp. 45, 47, 51. 16. Quote: Villehardouin, La conquête, p. 134. See also ibid., pp. 138-42; Robert of Clari, La conquête, pp. 62, 69-77; Niketas Choniates, 'Historia', cois 947, 951-54. For a description of Constantinople's sea walls, see van Millingen, Byzantine Constantinople, pp. 178-267; Foss and Winfield, Byzantine Fortifications, PP- 70-7317. Foss and Winfield, Byzantine Fortifications, p. 42. 18. George Pachymeres, 'Georgii Pachymeris de Michaela et Andronico Palaeologis', ed. I. Bekker, in CSHB, 2 vols (Bonn, 1835), vol. I, pp. 186, 187. The Blachernae palace also had to be restored at this time, because it had been left in such a bad state by the last Latin emperor, Baldwin II. See ibid., vol. i, pp. 144, 161. 19. Villehardouin, La conquête, p. 166. 20. Villehardouin, La conquête, pp. 200, 204-6 (Arcadiopolis, Stenimaka, Philipoppolis); pp. 240-44 (Rousion); pp. 246-48 (Rodosto); pp. 206-16, 234-36 (Adrianople). For Adrianople, see also Niketas Choniates, 'Historia', col. 1002; Robert of Clari, La conquête, pp. 105-6. 21. W. Miller, Essays on the Latin Orient (Cambridge, 1921), p. 279; Foss and Winfield, Byzantine Fortifications, pp. 5, 8; M. Vickers, 'The Byzantine Sea Walls

364

NOTES TO PAGES 200-209

of Thessaloniki', Balkan Studies, 2 (1970), pp. 261-78. Thessaloniki's defences are very similar to those of Antioch, described above, pp. 24-26. 22. 1210: Lock, The Franks, p. 60. 1224: Nikephoros Gregoras, 'Byzantina historia', vol. i, pp. 25-28; George Akropolites, 'Annales', cols 1035-38; B. Sinogowitz, 'Zur Eroberung Thessalonikes im Herbst 1224', BZ, 45 (1952), p. 28. 1209: the Lombard rebellion will be discussed in detail below, pp. 258-59. 23. Niketas Choniates, 'Historia', cols 651-54; Miller, Essays on the Latin Orient, pp. 275-76; Vickers, 'The Byzantine Sea Walls of Thessaloniki', pp. 272-74. 24. Henry of Valenciennes, Histoire de l'empereur, pp. 97-104. For the initial invasion of northern Greece, see Vulehardouin, La conquête, pp. 160-78; Robert of Clari, La conquête, pp. 96-105; Niketas Choniates, 'Historia', cois 983-87, 1023-27; George Akropolites, 'Annales', cols 999-1002. 25. 1206-7: Villehardouin, La conquête, pp. 232-34; Niketas Choniates, 'Historia', col. 1006. 26. 1249: Livre de la conqueste, pp. 67-68, 72-74; Libro de los fechos et conquistas, pp. 48-49. 1259-62: George Akropolites, 'Annales', cols 1195-99; George Pachymeres, 'De Michaele et Andronico', vol. i, pp. 85-89; Nikephoros Gregoras, 'Byzantina historia', vol. i, pp. 74-75, 79-80; Livre de la conqueste, pp. 92-123; Libro de los fechos et conquistas, pp. 55-69; Marino Sañudo, 'Istoria', pp. 107-8. For the background to the battle of Pelagonia, see also D. J. Geanakoplos, 'Graeco-Latin Relations on the Eve of the Byzantine Restoration: The Battle of Pelagonia, 1259', Dumbarton Oaks Papers, 7 (1953), pp. 101-41. 27. Andrews, Castles, pp. 206-7,209-10. For a description and history of the whole site, see ibid., pp. 192-210. 28. Villehardouin, La conquête, pp. 272, 288. See also George Akropolites, 'Annales', col. 995. 29. Villehardouin, La conquête, p. 272. The Catalans reoccupied the same site over a century later, when Andronicus II employed them to fight the Turks in Asia Minor. See Muntaner, Crònica, vol. 2, pp. 71-74; George Pachymeres, 'De Michaele et Andronico', vol. 2, pp. 398-99. 30. Vulehardouin, La conquête, pp. 188, 274-76. 31. Henry of Valenciennes, Histoire de l'empereur, pp. 50-52. 32. Innocent III, PI, vol. 216, no. 164, col. 340; Miller, The Latins in the Levant, p. 64; E. Gerland, Neue Quellen zur Geschichte des lateinischen Erzbistmus Potras (Leipzig 1903), p. 14. Rivalry between the barons and archbishops of Patras came to an end when William Aleman's successor sold the lordship to the church and returned to western Europe; Libro de los fechos et conquistas, p. 88. During the fourteenth century the city was allied to, and eventually came under the protection of, Venice; F. Thiriet, Régestes des deliberations du sénat de Venise concernant la Romanie, 3 vols (Paris, 1958-61), vol. i, no. 520, p. 130. During the early fifteenth century Venice rented the entire city from the church for a while, until the Greeks of Mistra captured it in 1430; Marino Sañudo, 'Vite de' duchi', cols 839, 917. See also Miller, The Latins in the Levant, pp. 363-64; Gerland, Neue Quellen, pp. 55-67,149-73; Andrews, Castles,

NOTES TO PAGES 2O9-211

365

pp. 116-19; Traquair, 'Mediaeval Fortresses', 279-80. Patras's close links with Venice will be discussed again below, p. 230. 33. Livre de la conqueste, pp. 37—38. See also Libro de los fechos et conquistas, p. 28. 34. Andrews, Castles, p. 34, and see pp. 30-35; Bon, La Marèe franque, pp. 606-68; Traquair, 'Mediaeval Fortresses', pp. 271-72. 35. Bon, La Moree franque, pp. 669-70; Andrews, Castles, pp. 85-89. As was noted above (p. 100), Arcadia can be compared with the Greek mountain fortresses occupied by the Latins on Cyprus. 36. Livre de la conqueste, p. 39. See also Libro de los fechos et conquistas, p. 28. 37. Andrews, Castles, p. 89; T. S. R. Boase, 'The Arts in Prankish Greece and Rhodes', in HC, vol. 4, pp. 208-28, at p. 219. 38. Coron: Villehardouin, La conquête, p. 196; Livre de la conqueste, p. 37. Patras: ibid., p. 30. Athens and Thebes: Niketas Choniates, 'Historia', col. 995. 39. The exact date of the fall of Corinth, Argos and Nauplia remains unclear. They were initially besieged by Boniface of Montferrat in late 1204: Villehardouin, La conquête, p. 196; Niketas Choniates, 'Historia', cois 991, 998. According to the Chronicle of Morea, Corinth and Nauplia were not captured until the 12405: Livre de la conqueste, pp. 68-71; Libro de los fechos et conquistas, p. 48; To Chromkon tou Mareos, pp. 188-97. However, papal documents from the reign of Innocent III alluding to the fall of Corinth (PL, vol. 216, no. 6, cols 201-2) and Argos (PL, vol. 216, no. 77, col. 598) suggest that the much earlier dates given above are correct. For Argos, see also To Chronikon tou Mareos, p. 104; Libro de los fechos et conquistas, p. 24. Libro de los fechos et conquistas correctly attributes the capture of Corinth to Geoffrey I of Villehardouin (p. 43), but subsequently appears to make the same mistake as the Livre de la conqueste by implying that Corinth fell much later in the thirteenth century (p. 48). See also J. Longnon, 'Problèmes de l'histoire de la principauté de Moree', Journal des savants (1946), pp. 77-93,147-61, at pp. 156-57; Bon, La Moree franque, p. 68. 40. Neopatras and surrounding territories had been granted to Boniface of Montferrat's followers in the partition treaty of 1204; Tafel and Thomas, Urkunden, vol. i, pp. 486-88. It presumably fell to the crusaders during Boniface's campaign in the area; Villehardouin, La conquête, p. 178; Niketas Choniates, 'Historia', cois 986, 990-91; Nicol, The Despotate, pp. 35-36, 57-58. In 1319 the Catalans captured Neopatras, and it became an important duchy under their rule until the Turkish invasion of 1390. See Marino Sañudo, 'Epistulae', ed. J. Bongars, in Gesta Dei per Francos, vol. 2, ep. 3 (1325), p. 291; A. Rubio y Lluch, 'Els Castells Catalans de la Grècia continental', Annuari de l'institut d'estudis Catalans, z (1908), pp. 364-425, at pp. 399-413. 41. For general descriptions of these sites, see, Corinth: Andrews, Castles, pp. 13845. Argos: Bon, La Moree franque, pp. 674-76; Andrews, Castles, pp. 107-15. Neopatras: Rubio y Lluch, 'Els Castells', pp. 399-400. Thebes: Bon, 'Forteresses medievales', pp. 189-91. 42. For general descriptions of these sites, see, Patras: Bon, La Moree franque,

306

NOTES TO PAGES 211-214

pp. 670-73; Andrews, Castles, pp. 119-29. Nauplia: ibid., pp. 92-105. Coron: ibid., pp. 15-23. 43. Good examples of such masonry can still be seen at Argos and Patras. See Andrews, Castles, p. 113, and figs 125-27 (Argos), p. 126 and fig. 144 (Patras); Bon, La Marèe franque, p. 673 (Patras). 44. Kalamata: Andrews, Castles, pp. 34-35. Arcadia: ibid., p. 89; Boase, 'The Arts in Prankish Greece', p. 219. Neopatras: Rubio y Lluch, 'Els Castells', p. 400. Corinth: Bon, La Marèe franque, p. 674; Andrews, Castles, p. 140. 45. Libro de los fechos et conquistas, p. 49. 46. Miller, The Latins in the Levant, pp. 401-2; P. Lock, 'The Prankish Tower on the Acropolis, Athens: The Photographs of William J. Stillman', ABSA, 82 (1987), pp. 131-33; P. Lock, 'The Prankish Towers of Central Greece', ABSA, 81 (1986), pp. 101-23, at pp. 107, 111-12. 47. Niketas Choniates, 'Historia', col. 991; Miller, The Latins in the Levant, pp. 31-32. On the great strength of the acropolis in general, see K. M. Setton, Catalan Domination of Athens, 1311—1388 (Cambridge, Massachusetts, 1948), pp. 188-89. 48. Corinth: Andrews, Castles, pp. 140-41. 49. Marino Sañudo, 'Istoria', p. 106. 50. Livre de la conqueste, p. 71. 51. Nauplia: Bon, La Marèe franque, p. 677; Andrews, Castles, pp. 93-94. Boudonitza: Bon, 'Fortéresses medievales', p. 161. Salona: ibid., p. 177. 52. This was destroyed by the Catalans to prevent its recapture by the Franks. However, the Catalans probably retained some urban defences at Thebes. See Livre de la conqueste, pp. 220-21; To Chronikon tou Mareos, p. 524; Bon, 'Fortéresses medievales', pp. 187—91. 53. Henry of Valenciennes, Histoire de l'empereur, pp. 111-13. 54. Cyzicus: Villehardouin, La conquête, p. 272. 55. Bon, 'Fortéresses medievales', pp. 148-63; Boase, 'The Arts in Prankish Greece', pp. 214-15. 56. For general descriptions of these sites, see, Modon: Andrews, Castles, pp. 61-83. Salona (Amphissa): Bon, 'Fortéresses medievales', pp. 164-86; Rubio y Lluch, 'Els Castells', pp. 413—25; Boase, 'The Arts in Prankish Greece', p. 215. Akova: Bon, La Marèe franque, pp. 634-35. Zeitoun (Lamia): Rubio y Lluch, 'Els Castells', pp. 393-98; Boase, 'The Arts in Prankish Greece', p. 216. 57. 1205: Villehardouin, La conquête, p. 194. 1209: Martin da Canal, 'La chronique des Veneciens', p. 349. Modon had originally been taken from the Greeks by Prankish crusaders, but was recognised as Venetian in 1209 in return for Venice's military support; see Andrea Dandolo, 'Chronicum Venetum', pp. 283-84; Tafel and Thomas, Urkunden, vol. 2, pp. 97-100. 58. Bon, La Marèe franque, pp. 634-35. 59. Bon, 'Fortéresses medievales', pp. 152-61, particularly at p. 161. 60. Bon, 'Fortéresses medievales', pp. 179-83; Rubio y Lluch, 'Els Castells', pp. 415i?. 61. To Chronikon tou Mareos, pp. 176-82; Libro de los fechos et conquistas, p. 49,

NOTES TO PAGES 214-218

367

which incorrectly dates these events to c. 1256 and the reign of William II; Honorius III, Regesta, vol. 2, nos 3162, 3163, pp. 516-17, no. 4480, p. 159. (Full text of no. 4480 also in Innocent III, PL, vol. 216, cols 968-72.) See also Bon, La Marèe franque, pp. 94-97; R. L. Wolff, 'Politics in the Latin Patriarchate of Constantinople', Dumbarton Oaks Papers, 8 (1954), pp. 225—303, at p. 274. 62. Bon, La Marèe franque, pp. 608—29 (fireplaces, pp. 621-22); Andrews, Castles, pp. 149-58; Traquair, 'Mediaeval Fortresses', pp. 272-79. Traquair should be used with caution, as he incorrectly states that Chlemoutsi was built in the 14308. He is also used by Boase, 'The Arts in Prankish Greece', pp. 217-18. 63. Bon, La Marèe franque, pp. 602—7. See also Traquair, 'Mediaeval Fortresses', pp. 272, 275-79, although this description is less reliable, and the attached history of the site appears to confuse it with Chlemoutsi. 64. Bon, La Marèe franque, pp. 324, 607. 65. Bon, 'Fortéresses medievales', pp. 194-206; Rubio y Lluch, 'Els Castells', pp. 38387. 66. To Chronikon tou Mareos, pp. 130, 131, 210; Libro de los fechos et conquistas, p. 29; Livre de la conqueste, p. 79. 67. Bon, La Marèe franque, pp. 629-30, 631. 68. Bon, La Marèe franque, pp. 630—33. 69. Androusa: Libro de los fechos et conquistas, p. 49; Bon, La Marèe franque, pp. 637-39. Old Navarino (Port de Jone): Livre de la conqueste, p. 221; Libro de los fechos et conquistas, p. 103; Bon, La Marèe franque, pp. 668-69. For a detailed description, see also Andrews, Castles, pp. 42-48. Mistra: Libro de los fechos et conquistas, p. 49; To Chronikon tou Moveos, p. 200; Bon, La Marèe franque, pp. 639-42, and see also Andrews, Castles, pp. 168-74; Runciman, Mistra, pp. 29-30. Beaufort (Leutron): Libro de los fechos et conquistas, p. 49; To Chronikon tou Mareos, p. 202; Bon, La Marèe franque, p. 504. Old Mania: Livre de la conqueste, pp. 74-75; Libro de los fechos et conquistas, p. 49; Bon, La Marèe franque, pp. 503-4; R. Traquair, 'Laconia: I, Mediaeval Fortresses', ABSA, 12 (1905-6), pp. 259-76, at pp. 275-76. Gerald: To Chronikon tou Mareos, pp. 130, 210; Bon, pp. 642-45. 70. P. Burridge, 'The Castle of Vardounia and Defence in the Southern Taygetos', in Lock and Sanders, The Archaeology of Medieval Greece, pp. 19-28, at p. 22. 71. Bon, La Marèe franque, p. 639; Foss and Winfield, Byzantine Fortifications, p. 30, and see above, pp. 204-5. 72. Bon, La Marèe franque, p. 632. 73. To Chronikon tou Mareos, pp. 130, 131; Livre de la conqueste, p. 44; Bon, La Marèe franque, pp. 633-34, 636. 74. This book relies on the Greek ( To Chronikon tou Mareos), the Aragonese (Libro de los fechos et conquistas) and the French (Livre de la conqueste) versions. The Italian version is simply a shortened sixteenth-century translation of the Greek version. For more details on these texts, see D. Jacoby, 'Quelques considerations sur les versions de la chronique de Morée', Journal des Savants (1968), pp. 13389; Lock, The Franks, pp. 21-24.

368

NOTES TO PAGES 218-222

75. See above, p. 210 n. 39. 76. Around 1287: Livre de la conqueste, pp. 220-21; To Chronikon tou Moreos, p. 524. As noted previously, the Byzantine castle at Thebes had already withstood a major siege in 1209; see Henry of Valenciennes, Histoire de l'empereur, pp. 111-13. 77. To Chronikon tou Moreos, pp. 130, 131, 210. 78. See above, p. 102. 79. Lock, The Franks, p. 218. 80. Pringle's introduction to Lawrence, Crusader Castles, p.xxx. 81. Kennedy, Crusader Castles, p. 33. 82. Holy Land: see above, pp. 29, 79-80. Cyprus: see above, pp. 101-2, 106 and Molin, 'Fortifications and Internal Security in the Kingdom of Cyprus, 11911426', pp. 187-99. 83. Edwards, 'The Crusader Donjon at Anavarza in Ciucia', pp. 53-55, and see above, p. 160. 84. Lock, 'The Prankish Towers of Central Greece', p. 102; P. Lock, 'The Medieval Towers in Greece: A Problem in Chronolgy and Function', in Arbel, Hamilton and Jacoby, Latins and Greeks in the Eastern Mediterranean after 1204, pp. 12945, at pp. 132-33. 85. Lock, 'The Medieval Towers of Greece', p. 133; Bon, 'Fortéresses medievales', pp. 146-48; Lock, 'The Prankish Towers of Central Greece', pp. 113,114-15,121, 122. 86. Lock, 'The Prankish Towers of Central Greece', pp. 113, 114; Bon, 'Fortéresses medievales', p. 146. For Haliartos, see also J. Bintliff, 'The Prankish Countryside in Central Greece: The Evidence from Archaeological Field Survey', in Lock and Sanders, The Archaeology of Medieval Greece, pp. 1—18, at p. 6. 87. Bon, 'Fortéresses medievales', pp. 147-48; Lock, 'The Medieval Towers of Greece', pp. 132-33; Lock, 'The Prankish Towers of Central Greece', pp. 113,121,122. 88. Bon, La Marèe franque, p. 650. 89. Lock, 'The Prankish Tower on the Acropolis, Athens', p. 133; Lock, 'The Medieval Towers of Greece', pp. 132-33; Lock, 'The Prankish Towers of Central Greece', pp. 107, 111-12. 90. Lock, 'The Medieval Towers of Greece', pp. 132-33. 91. Bon, 'Fortéresses medievales', pp. 147-48, 161. 92. Lock, 'The Prankish Tower on the Acropolis, Athens', p. 131; Bon, 'Fortéresses medievales', p. 161. 93. Bon, La Marèe franque, p. 677; Andrews, Castles, pp. 93-94. 94. Lock, 'The Prankish Tower on the Acropolis, Athens', p. 133; Longnon, 'Problèmes de l'histoire', pp. 89-90; Lock, 'The Prankish Towers of Central Greece', p. 112. 95. Lock, 'The Medieval Towers of Greece', p. 136; Lock, 'The Prankish Towers of Central Greece', pp. 106-8. 96. Lock, 'The Towers of Euboea', p. 113, no. 13, and p. 124, plate 5. 97. Quote: Lock, 'The Towers of Euboea', p. 108. See also Lock, 'The Prankish Towers of Central Greece', pp. 102,104; Lock, 'The Medieval Towers of Greece',

NOTES TO PAGES 222-223

3^9

pp. 137, 139-41; M. Balard, La Romanie génoise, Xlle-début du XVe siede), 2 vols (Rome and Genoa, 1978), vol. i, pp. 445-46. 98. Marino Sañudo, 'Vite de' duchi', cols 543-45. See also Andrea Navagerio, 'Storia Venezia', col. 987; Niketas Choniates, 'Historia', col. 1030; Andrea Dandolo, 'Chronicum Venetum', pp. 283-84; Fotheringham, Marco Sañudo, pp. 51, 82-83. 99. Daniele Barbaro, 'Cronica del Trivisano della città di Venezia', extract cited in Fotheringham, Marco Sañudo, p. 106; 'Cronica antica di Venetia', extract cited in Fotheringham, Marco Sañudo, p. rio; Sauger, 'Histoire nouvelle', extract cited in Fotheringham, Marco Sañudo, p. 115; Boase, 'The Arts in Prankish Greece', p. 222; Fotheringham, Marco Sañudo, pp. 41-44. 100. Sauger, 'Histoire nouvelle', in Fotheringham, Marco Sañudo, p. 115; Boase, 'The Arts in Prankish Greece', p. 222; Fotheringham, Marco Sañudo, pp. 70-71. 101. Livre de la conqueste, pp. 43-44; Libro de los fechos et conquistas, p. 28. 102. The lords of Patras, Karytaina and Kalavryta could exercise high justice, or cases involving crimes punishable by loss of life or limb. See Les Assises de Romaniee, ed. G. Recoura (Paris, 1930) (henceforth Assises), articles 43, 94. 103. The history of Karytaina can be used to illustrate these problems. In 1275 Geoffrey de Briel, only the third lord of the dynasty, died childless, and his lordship was divided between the prince of Achaia and Geoffrey's widow. Karytaina was subsequently granted to a succession of different vassals before being lost to the Greeks of Mistra early in the fourteenth century. See Livre de la conqueste, pp. 194-95, and p. 405; Libro de los fechos et conquistas, p. 141; Bon, La Marèe franque, pp. 366-69. Emigration back to Europe, the harsh living conditions, and in particular military setbacks such as the battle of Halmyros against the Catalans (1311) accounted for the disappearance of many Prankish families early in the fourteenth century. See Bon, La Marèe franque, pp. 195-97; Miller, The Latins in the Levant, pp. 146-48; Topping, 'The Morea, 1311-1364', pp. 120—21; Longnon, L'empire latin, pp. 314-16. 104. See above, p. 209 n. 32. 105. Bon, La Marèe franque, pp. 629-33. Karytaina owed twenty-two knights' fees, and was therefore only smaller than Patras and Akova (twenty-four fees each). See Livre de la conqueste, pp. 43-44; Libro de los fechos et conquistas, pp. 28-29. 106. Honorius III appealed for help in defending Boudonitza and neighbouring Salona against the Greeks in the i22os. See Honorius III, Regesta, vol. 2, no. 4758, p. 207, no. 5464, p. 333; Miller, The Latins in the Levant, pp. 84-85; Nicol, The Despotate, pp. 62-64. 107. The Pallavicini lord of Boudonitza was killed at the battle of Halmyros (1311), but his widow married a Venetian called Andrea Cornaro. Thereafter Boudonitza increasingly came under the control of Venice, whose military and political power safeguarded its virtual independence from the Catalans. See Marino Sañudo, 'Istoria', p. 125; Libro de los fechos et conquistas, p. 120; Setton, Catalan Domination, pp. 33, 105-6; Miller, The Latins in the Levant, p. 248. 108. Miller, The Latins in the Levant, pp. 373-75. For a general history of the site,

37O

109. 110. 111.

112.

113. 114. 115. 116. 117. 118. 119. 120. 121.

NOTES TO PAGES 223~228

see Bon, 'Fortéresses medievales', pp. 148-51. For variations in masonry, see ibid., pp. 160-61. Bon, 'Fortéresses medievales', pp. 164-86; Rubio y Lluch, 'Els Castells', pp. 41325. Bon, La Marèe franque, pp. 637-39. Othon de la Roche's ownership of Livadia is confirmed by documents dating from 1214, issued during his dispute with the papacy. See L. A. Muratori, ed., Antiquitates italicae medii aevi, 6 vols (Rome, 1738-42), vol. 5, cols 83336; Miller, The Latins in the Levant, pp. 69-70. Othon's arguments with the church were similar to, and linked with, those of Geoffrey of Villehardouin; see above, pp. 213-14. For the subsequent history of Livadia, see Bon, 'Fortéresses medievales', pp. 191-206; Rubio y Lluch, 'Els Castells', PP- 374-87. Foss and Winfield, Byzantine Fortifications, p. 42. Urgent repairs were carried out after Michael VIII Palaiologos regained the city in 1261; see George Pachymeres, 'De Michaele et Andronico', vol. i, pp. 186, 187. Thebes: see above, p. 212. Chlemoutsi: see above, pp. 213-14. Libro de los fechos et conquistas, p. 28; Livre de la conqueste, pp. 43-44; Gerland, Neue Quellen, p. 14; Bon, La Marèe franque, p. 106 n. 3. See above, Chapter 2. Bon, La Marèe franque, pp. 629-33. Cyprus: see above, pp. 98-100. Cilician Armenia: see above, pp. 145-48. Bon, La Morée franque, pp. 602-7. Holy Land: see above, pp. 7-9, 24-26, 35. Greece: see above, pp. 194-202. See below, pp. 255-56. For example, in the 13205 the Greeks obtained Karytaina and other neighbouring castles by bribing their Latin garrisons; see Livre de la conqueste, pp. 404—5; Libro de los fechos et conquistas, pp. 140-43.

Notes to Chapter 20, Prankish Greece: Fortifications, Naval Warfare and the Crusades 1. 2. 3. 4. 5.

Villehardouin, La conquete, pp. 100, 222. Villehardouin, La conquete, p. 194. Villehardouin, La conquete, p. 272. Villehardouin, La conquete, p. 236. Henry of Valenciennes, Histoire de l'empereur, pp. 59-61,104-6. This rebellion will be discussed in more detail below, pp. 258-59. 6. Actes relatifs a la prinàpauté de Morée, 1289-1300, ed. C. Perrat and J. Longnon (Paris, 1967), pp. 207-11; Livre de la conqueste, pp. 160, 170-76; Libro de los fechos et conquistas, pp. 91, 92; Marino Sañudo, 'Istoria', pp. 118-19; Longnon, L'empire latin, pp. 230-40; Lock, The Franks, pp. 84-86, and see above, p. 193. 7. There appear to be no remains on the site which predate 1204; see Bon, La Morée franque, pp. 324, 607.

NOTES TO PAGES 228-231

371

8. Buchón, Nouvelles recherches, vol. 2, no. 19, pp. 326-27, and see no. 24, pp. 33031, no. 25, p. 331, no. 27, pp. 332-33. See also Livre de la conqueste, pp. 178, 192-93. 9. Livre de la conqueste, p. 221; Libro de los fechos et conquistas, p. 103. For a description, see Bon, La Marèe franque, pp. 668-69; Andrews, Castles, pp. 404910. Livre de la conqueste, pp. 220-21; To Chronikon tou Mareos, p. 524; Lock, The Franks, pp. 78-79. 11. Naxos: Sauger, 'Histoire nouvelle', in Fotheringham, Marco Sañudo, p. 115, and see pp. 70-71; Boase, 'The Arts in Prankish Greece', p. 222. Monemvasia: Andrews, Castles, pp. 192-210. 12. See above, pp. 196, 197. 13. VUlehardouin, La conquête, pp. 284-88; George Akropolites, 'Annales', col. 1042. 14. Modon and Coron did not fall until 1500; see Miller, The Latins in the Levant, pp. 495-98; Andrews, Castles, pp. 15, 60. 15. George Akropolites, 'Annales', cols 1207-11; Nikephoros Gregoras, 'Byzantina historia', vol. i, pp. 85-86; Andrea Dandolo, 'Chronicum Venetum', pp. 311, 369; George Pachymeres, 'De Michaele et Andronico', vol. i, pp. 140-48; Andrea Navagerio, 'Storia Venezia', col. 199; Marino Sañudo, 'Istoria', p. 115; Libro de los fechos et conquistas, pp. 21, 51; Livre de la conqueste, pp. 26-27. Acre etc.: see above, pp. 45-46, 50-51. 16. VUlehardouin, La conquête, pp. 278-82. 17. 1209: Andrea Dandolo, 'Chronicum Venetum', pp. 283-84; Tafel and Thomas, Urkunden, vol. 2, pp. 97-100. 18. Livre de la conqueste, p. 68. See also Libro de los fechos et conquistas, p. 48. It should be noted that the chronicle incorrectly dates the fan of Nauplia to this period; it actually fell in 1212. 19. lacobo Zeno, 'Vita Caroli Zeni', ed. G. Zonta, RIS n.s., vol. 19, part 6 (Bologna, 1931-41), pp. 10-11. 20. In 1373 Venice sent two galleys to Patras. See Thiriet, Régestes, vol. i, nos 520, 522, p. 130. For more details on the background to these events, see Bon, La Morée franque, pp. 247-51; Miller, The Latins in the Levant, pp. 287-90; Gerland, Neue Quellen, pp. 39-42. 21. Libro de los fechos et conquistas, pp. 121-24,127-37; Muntaner, Crònica, vol. 2, pp. 159-60, 164-65. For more details on the background to this dispute, see Miller, The Latins in the Levant, pp. 252-57. 22. Livre de la conqueste, pp. 35—36. 23. Pryor, Geography, Technology and War, pp. 153-61, 165-75, and see above, pp. 199-200. 24. See, for example, Marino Sañudo, 'Istoria', pp. 132, 146. 25. See above, p. 222. 26. Niccolo da Martorii, 'Relation du pèlerinage', pp. 646-48; Luttrell, 'The Latins and Life on the Smaller Aegean Islands', p. 151.

372

NOTES TO PAGES 231-234

27. Fotheringham, Marco Sañudo, p. 78. 28. See above, pp. 221—22. 29. Sauger, 'Histoire nouvelle', extract cited in Fotheringham, Marco Sañudo, p. 115; Fotheringham, Marco Sañudo, p. 71. 30. Marino Sañudo, 'Istoria', pp. 113—14. 31. Lock, 'The Towers of Euboea', pp. 112 (Rovies), 117 (Aliberi-Matsoukela). 125558: Marino Sañudo, 'Istoria', pp. 103-6, and see Livre de la conqueste, pp. 81-87; Libro de los fechos et conquistas, pp. 49-52; Miller, The Latins in the Levant, pp. 1OO-6; Longnon, L'empire latin, pp. 220-23; Lock, The Franks, pp. 90-91. 32. Livre de la conqueste, pp. 301-2. See also Muntaner, Crònica, vol. i, pp. 186-87. The town of Monemvasia was eventually fortified, either by the Venetians or the Turks, as further protection against such raids. See Bon, La Moree franque, p. 492; Andrews, Castles, pp. 202-3. 33. Buchón, Nouvelles recherches, vol. 2, no. 25, pp. 143-55 (St George, p. 146), no. 26, pp. 153-55, no. 27, pp. 155-56, no. 28, pp. 157-58, no. 33, pp. 204-7 (particularly at p. 204). 34. Buchón, Nouvelles recherches, vol. 2, no. 15, pp. 109-14, and see no. 29, pp. 158-60. 35. For more details on the Turkish naval threat and the creation of the Holy League, see J. Gay, Le Pape Clement VI et les affaires d'Orient, 1342-1352 (Paris, 1904); Pryor, Geography, Technology and War, pp. 165-73; Topping, 'The Morea, 1311-1364', p. 133. 36. Nikephoros Gregoras, 'Byzantina historia', vol. 2, p. 689; Clement VI, Lettres closes, patentes et curiales, 3 vols (Paris, 1958-59), vol. i, no. 1350, p. 335; Marino Sañudo, 'Vite de' duchi', col. 610; Johannis Villani, 'Fiorentini historia universalis', ed. L. A. Muratori, RIS, vol. 13 (1728), col. 917; Gay, Le Pape Clement, pp. 32-43. 37. G. Wheeler, Journey into Greece (London, 1682), cited in Andrews, Castles, p. 187. See also ibid., pp. 187-91. 38. For descriptions of Famagusta, see above, pp. 102-4,128- For more details on the Hospitallers' activities and fortifications on Rhodes, see below, pp. 265—66. 39. For Corycus and other Cypriot bases in Asia Minor, see above, pp. 127-28. 40. Livre de la conqueste, p. 68. 41. Muntaner, Crònica, vol. 2, pp. 79-94. See also pp. 94-104. 42. Muntaner, Crònica, vol. 2, p. 103. 43. Livre de la conqueste, pp. 301-2; Muntaner, Crònica, vol. i, p. 186. 44. Andrea Dandolo, 'Chronicum Venetum', pp. 283, 367; Andrea Navagerio, 'Storia Venezia', cols 986-87. 45. Nikephoros Gregoras, 'Byzantina historia', vol. 3, pp. 46-51. 46. Marino Sañudo, 'Vite de' duchi', cols 536, 543-45, and see Fotheringham, Marco Sañudo, pp. 81-87.

NOTES TO PAGES 236-240

3/3

Notes to Chapter 21, Prankish Greece: The Offensive Functions of Inland Fortifications 1. Villehardouin, La conquête, pp. 154-66; Robert of Clari, La conquête, pp. 96-105; Aubrey of Trois Fontaines, 'Chronica', p. 885; Niketas Choniates, 'Historia', cols 983-87; Nikephoros Gregoras, 'Byzantina historia', vol. i, p. 14. 2. Villehardouin, La conquête, pp. 198-200, 202-22. 3. Livre de la conqueste, pp. 176-77, and see pp. 176-81. 4. Livre de la conqueste, pp. 404-5; Libro de los fechos et conquistas, pp. 140—43. By the 13505, when Niccolo Acciajuoli was castellan of Corinth and was responsible for the defence of this area against the Greeks, St George had returned to Latin control. See Buchón, Nouvelles recherches, vol. 2, no. 25, p. 146. 5. Muntaner, Crònica, vol. 2, pp. 94-95. 6. Villehardouin, La conquête, pp. 240-42; Niketas Choniates, 'Historia', col. 1018; Letter from the emperor Henry to the West, September 1206, in RHGF, vol. 18 (1879), pp. 527-28. 7. Letter from the emperor Henry to the West, RHGF, vol. 18, pp. 528-29; Villehardouin, La conquête, p. 270; Niketas Choniates, 'Historia', col. 1031. 8. Livre de la conqueste, pp. 178-80, and see also pp. 180-82; Libro de los fechos et conquistas, p. 84. 9. Livre de la conqueste, p. 99. It seems William II of Villehardouin was particularly keen to recapture Salonika for the Franks. See Libro de los fechos et conquistas, pp. 55-56. Saladin: see above, pp. 73-74. 10. The Franks certainly lost many men, but the figure of 700 knights given by Muntaner (Crònica, vol. 2, pp. 122-24) is surely an exaggeration. For Livadia, see, Diplomatari de l'Oríent cátala, ed. A. Rubio y Lluch (Barcelona, 1947), no. 186, pp. 227-28, no. 268, pp. 352-53; Rubio y Lluch, 'Els Castells', p. 375. 11. To Chronikon tou Mareos, pp. 200-2 (quote p. 202); Libro de los fechos et conquistas, p. 49. 12. Livre de la conqueste, pp. 74-75. See also Libro de los fechos et conquistas, p. 49; To Chronikon tou Mareos, p. 202. 13. Villehardouin La conquête, p. 250. 14. Henry of Valenciennes, Histoire de l'empereur, pp. 44-48, 50, 52. 15. Marino Sañudo, 'Istoria', p. 116. 16. Livre de la conqueste, pp. 220-21; To Chronikon tou Mareos, p. 524. 17. Villehardouin, La conquête, p. 292. 18. Villehardouin, La conquête, pp. 246-50; Niketas Choniates, 'Historia', cois 1181919. See above, pp. 40-41, 76-77. 20. Livre de la conqueste, pp. 68-69; Libro de los fechos et conquistas, pp. 25-26; Niketas Choniates, 'Historia', col. 998; Andrews, Castles, pp. 136, 140, 224; Boase, 'The Arts in Prankish Greece', p. 217. 21. Livre de la conqueste, p. 30. 22. Villehardouin, La conquête, p. 92. During the First Crusade Bohemond had

374

NOTES TO PAGES 240-245

been one of the key leaders involved in the siege of Antioch (1097-98), which was also conducted from small castles built by the crusaders to blockade the city's gates. See Gesta Francorum, ed. and trans. R. Hill (Oxford, 1962), pp. 30, 39-40, 42, 4323. Livre de la conqueste, p. 33; Libro de los fechos et conquistas, p. 24. 24. Villehardouin, La conquête, pp. 92-94. 25. Livre de la conqueste, p. 30. 26. Livre de la conqueste, p. 69, and see Libro de los fechos et conquistas, p. 26. 27. Villehardouin, La conquête, p. 92. 28. These tactics can be compared with contemporary events in the Holy Land, where we have seen that a number of sieges, most famously that of Acre during the Third Crusade, were conducted from local makeshift fortifications. See above, p. 76. Fall of Corinth: see above, p. 210 n. 39. 29. Villehardouin, La conquête, pp. 212-22; letter from the emperor Henry to Innocent III, June 1205, RHGF, vol. 18, p. 527. 30. Henry of Valenciennes, Histoire de l'empereur, pp. 50-52. 31. 1271: Livre de la conqueste, pp. 178-80, and see also pp. 180-82; Libro de los fechos, p. 84. 1302: Livre de la conqueste, p. 365. 32. Henry of Valenciennes, Histoire de l'empereur, pp. 37—38. 33. Serres: Philip Mousket, 'Chronique rimée', voi. 2, pp. 408-9. Poimanenon: Nikephoros Gregoras, 'Byzantina historia', vol. i, p. 25; Aubrey of Trois Fontaines, 'Chronica', p. 911; George Akropolites, 'Annales', cols 1038-39.

Notes to Chapter 22, Prankish Greece: The Defensive Functions of Inland Fortifiactions 1. Philip Mousket, 'Chronique rimée', vol. 2, pp. 614-16; Andrea Dandolo, 'Chronicum Venetum', p. 295; Andrea Navagerio, 'Storia Venezia', col. 992; George Akropolites, 'Annales', cols 1058-59. 2. Aubrey of Trois Fontaines, 'Chronica', pp. 938-39; Philip Mousket, 'Chronique rimée', vol. 2, p. 620. 3. Villehardouin, La conquête, pp. 240, 244-46. 4. Villehardouin, La conquête, pp. 240-62; Niketas Choniates, 'Historia', cois 101523, 1031-34; letter from the emperor Henry to the West, September 1206, RHGF, vol. 18, pp. 528-29. 5. Livre de la conqueste, p. 385. The tactic of demolishing buildings and outer defences which could be used by the enemy was also adopted in the Holy Land. See above, pp. 34-35. 6. Livre de la conqueste, pp. 380-89 (quote, p. 387). For the background to this dispute, see Longnon, L'empire latin, pp. 285-86. 7. Johannis Villani, 'Florentini historia universalis', c. 190, col. 717. The pope even treated this expedition as a crusade, granting indulgences to those who went. See John XXII, Lettres communes, ed. G. Mollat, 16 vols (Paris 1904-47),

NOTES TO PAGES 245-250

375

vol. 9, no. 49924, p. 345. See also Buchón, Nouvelles recherches, vol. i, pp. 30-33; Setton, Catalan Domination, pp. 38-41. 8. Holy Land: see above, pp. 3, 40. 9. Livre de la conqueste, pp. 32-34; Libro de los fechos et conquistas, pp. 23-26; Niketas Choniates, 'Historia', cols 991, 998. For the history of Corinth in the fifteenth century, see Andrews, Castles, p. 137. 10. 1205: Niketas Choniates, 'Historia', col. 991. 12205: Honorius III, Regesta, vol. 2, no. 4758, p. 207, no. 5464, p. 333. 11. Assises, article 19; Livre de la conqueste, pp. 120-23 (quote: p. 121); Libro de los fechos et conquistas, pp. 65-66, 67-68. 12. Bartusis, The Late Byzantine Army, pp. 115-17. 13. Nikephoros Gregoras, 'Byzantina historia', vol. i, p. 246. 14. Livre de la conqueste, pp. 301-2, and see Muntaner, Crònica, vol. i, p. 186. 15. To Chronikon tou Moreos, p. 308. 16. Livre de la conqueste, pp. 328-29. 17. Buchón, Nouvelles recherches, vol. 2, no. 15, pp. 109-14, and see no. 29, pp. 15860. 18. Quotes: Bintiiff, 'The Prankish Countryside in Central Greece', p. 6. 19. Bintiiff, 'The Prankish Countryside in Central Greece', pp. 6-7. 20. Lock, 'The Towers of Euboea', no. 28, pp. 115-16. 21. Ellenblum, 'Three Generations of Prankish Castle-Building', pp. 520-23. 22. See, for example, Lock, The Franks, p. 293. 23. Holy Land: Ellenblum, Prankish Rural Settlement, especially pp. 3-6, 9, 12-19, and see above, pp. 55-57. 24. Lock, 'The Medieval Towers of Greece', pp. 136-37; account of the Serbian chronicler Daniel, in A. Soloviev, 'Histoire du monastère russe au Mont-Athos', Byzantion, 8 (1933), pp. 213-38, at pp. 223-24. 25. Bartusis, The Late Byzantine Army, pp. 316—17. 26. Innocent III, PL, vol. 216, no. 168, cols 956-57. 27. Fourteenth-century survey of royal rights and properties in the Morea, reproduced in Bon, La Marèe franque, pp. 691-92, at p. 692. See also ibid., pp. 279, 646-48. 28. Bon, 'Forteresses medievales', pp. 162, 184. 29. Livre de la conqueste, p. 71. 30. Sauger, 'Histoire nouvelle', extract in Fotheringham, Marco Sañudo, p. 115; Fotheringham, Marco Sañudo, p. 71. Generally, the inhabitants of Chios were segregated in the same way, with the Genoese living in the actual citadel and the Greeks inhabiting the surrounding bourg. See Balard, La Romanie génoise, vol. i, p. 226. 31. Villehardouin, La conquéte, p. 232. 32. Villehardouin, La conquête, p. 250; Niketas Choniates, 'Historia', col. 1006. 33. Livre de la conqueste, p. 48. 34. See above, pp. 47-51, 54-59-

3/6

NOTES TO PAGES 250-255

35. Livre de la conqueste, pp. 243-52 (quote: p. 251); Libro de los fechos et conquistas, pp. loo-ioi. 36. Assises, article 26. 37. See above, pp. 3-4, 7-11, 192, 194-96. 38. Corinth: Bon, La Marèe franque, p. 473; Andrews, Castles, pp. 136-37. Monemvasia: ibid., p. 207. Thebes and Salona: Bon, 'Fortéresses medievales', pp. 164, 187. 39. Ludolph of Sudheim, 'De itinere Terrae Sanctae', p. 28. 40. Lock, 'The Prankish Tower on the Acropolis, Athens', p. 133; Lock, 'The Prankish Towers of Central Greece', p. 112. Buffavento: see above, p. 116. 41. Boeotia: Bon, 'Fortéresses medievales', p. 148. Euboea: Lock, 'The Prankish Towers of Central Greece', p. 102. Lock has since changed his views; see next footnote. 42. Euboea: Lock, 'The Towers of Euboea', pp. 107-11, 118. See also Lock, 'The Medieval Towers of Greece', pp. 139-40. Haliartos: Bintuff, 'The Prankish Countryside in Central Greece', p. 6. 43. This was particularly the case in Cuidan Armenia; see above, pp. 157-61. Notes to Chapter 23, Prankish Greece: Fortifications and Internal Security 1. Assises, article 70; Livre de la conqueste, p. 47; Libro de los fechos et conquistas, p. 32. 2. Livre de la conqueste, pp. 34-35. See also Libro de los fechos et conquistas, p. 31. 3. Lock, The Franks, pp. 284-90. That Greeks in Achaia performed military service is shown in Assises, article 71. The Greeks of Naxos are said to have helped their Venetian conquerors combat pirates. See Andrea Dandolo, 'Chronicum Venetum', p. 282; Fotheringham, Marco Sañudo, p. 57. 4. See above, pp. 211, 213, 218-22, 224. 5. Villehardouin, La conquête, p. 196. 6. Livre de la conqueste, p. 79. 7. Assises, article 94. 8. Marino Sañudo, 'Istoria', p. 106. 9. Andrea Dandolo, 'Chronicum Venetum', pp. 284-85, 288, 292 (see also p. 304); Andrea Navagerio, 'Storia Venezia', cols 987-92; Marino Sañudo, 'Vite de' duchi', cols 545, 547, 549 (see also col. 557); Fotheringham, Marco Sañudo, pp. 87-103. 10. Andrea Navagerio, 'Storia Venezia', col. 991; Marino Sañudo, 'Vite de' duchi', col. 549; Fotheringham, Marco Sañudo, p. loi. 11. Livre de la conqueste, pp. 318-27. 12. This is made clear by the fact that a Latin garrison defended St George during another Greek rebellion in 1302 which will be discussed shortly. See Livre de la conqueste, p. 367. 13. Livre de la conqueste, pp. 362-65, 373. 14. Fotheringham, Marco Sañudo, pp. 89-92.

NOTES TO PAGES 255-258

377

15. John VI Kantakouzenos, 'loannis Cantacuzeni eximperatoris historiarum libri IV', ed. L. Schopen and B. Niehbuhr, in CHSB, 3 vols (Bonn, 1828-32), voi. i, pp. 370-79; M. Balard, 'The Genoese in the Aegean, 1204-1566', in Arbel, Hamilton and Jacoby, Latins and Greeks in the Eastern Mediterranean after 1204, pp. 158-74, at pp. 162-63; Balard, La Romanie génoise, vol. i, pp. 121-23. 16. Hattin: see above, p. 58. 17. Villehardouin, La conquête, pp. 288-92. 18. Halmyros: after the battle there were not enough Latins left in Livadia to prevent Greek rebels from delivering the city to the Catalans. See Rubio y Lluch, Diplomatari de l'Orient cátala, no. 186, pp. 227—28, no. 268, pp. 352-53; Rubio y Lluch, 'Els Castells', p. 375. Pelagonia: Marino Sañudo, 'Istoria', p. 107. 19. Bon, La Marèe franque, pp. 386-89, 504, 650-52. 20. Livre de la conquiste, pp. 365-74. 21. Villehardouin, La conquête, pp. 204, 236-38, 260-62; letter from the emperor Henry to the West, September 1206, RHGF, vol. 18, pp. 528-29. See also Niketas Choniates, 'Historia', cois 1015, 1031. 22. John VI Kantakouzenos, 'loannis Cantacuzeni eximperatoris historiarum libri IV', vol. 3, pp. 83-85; Balard, 'The Genoese in the Aegean', pp. 164-65; Balard, La Romanie génoise, vol i, pp. 124-25. 23. Setton, Catalan Domination, pp. 83-85; Rubio y Lluch, 'Els Castells', pp. 381-82, 389-90; Setton, 'The Catalans in Greece, 1311-1380', pp. 206-7. 24. Marino Sañudo, 'Istoria', p. 121; George Pachymeres, 'De Michaele et Andronico', vol. i, pp. 342-48; Rubio y Lluch, 'Els Castells', p. 399; Miller, The Latins in the Levant, pp. 131-35. 25. Geoffrey I kept Corinth after it finally surrendered in 1210: Livre de la conqueste, p. 69. Kalamata was granted to Geoffrey I by William Champlitte in c. 1205-9: Livre de la conqueste, p. 42; Libro de los fechos et conquistas, pp. 31-32. Androusa was built, or, as has been suggested, reoccupied and repaired by William II: Libro de los fechos et conquistas, p. 49. Chlemoutsi was built from scratch by Geoffrey I: Bon, La Marèe franque, pp. 608-29. Glarentza's walls seem to have been built from scratch by the Achaian rulers in the thirteenth century: Libro de los fechos et conquistas, p. 49; Bon, La Morée franque, pp. 602-7. Monemvasia was garrisoned by William U's troops after it surrendered: To Chronikon tou Mareos, p. 198. Mistra was built or, as has been suggested, repaired by William II: Libro de los fechos et conquistas, p. 49. 26. Salonika: This city was occupied by Boniface of Montferrat after he had come to terms with Baldwin of Flanders. See Villehardouin La conquête, pp. 176-80; Robert of Clari, La conquête, pp. 104-5. Constantinople: Andrea Dandolo, 'Chronicum Venetum', p. 281; Andrea Navagerio, 'Storia Venezia', cols 984-86; Robert of Clari, La conquête, p. 102; Villehardouin, La conquête, p. 136; Lock, The Franks, pp. 142-46. 27. Villehardouin, La conquête, pp. 232-34, 272. 28. Livre de la conqueste, p. 338. 29. Livre de la conqueste, p. 321.

378

NOTES TO PAGES 258-263

30. Livre de la conqueste, pp. 404-5; Libro de los fechos et conquistas, pp. 140-43. 31. Villehardouin, La conquête, pp. 164-66; Robert of Clari, La conquête, PP- 97-98. 32. Villehardouin, La conquête, pp. 166-78; Robert of Clari, La conquête, pp. 97-100, 104-5; Longnon, L'empire latín, pp. 55-61. 33. Henry of Valenciennes, Histoire de l'empereur, pp. 55-118. See also Longnon, L'empire latin, pp. 106-11; Miller, The Latins in the Levant, pp. 72-75; Lock, The Franks, pp. 57-60. 34. Henry of Valenciennes, Histoire de l'empereur, pp. 59-61, 104-6. 35. Henry of Valenciennes, Histoire de l'empereur, pp. 79-118. 36. Frederick III had recognised Roger as vicar-general by August 1366. See Rubio y Lluch, Diplomatari de I'Orient cátala, no. 271, p. 355. The pope also complained about Roger's actions, and his use of Turkish mercenaries. See Lettres secretes et curíales se rapportant a la France, ed. G. Mollai (Paris 1955), nos 1047, 1050, p. 163; Setton, 'The Catalans in Greece, 1311-1380', pp. 198-99, 202-4. 37. Livre de la conqueste, p. 85; Libro de los fechos et conquistas, pp. 50-51; Marino Sañudo, 'Istoria', p. 105. 38. Libro de los fechos et conquistas, pp. 152-55, and see above, p. 230. 39. Livre de la conqueste, p. 30. See also Bon, La Morée franque, pp. 318-20. 40. Livre de la conqueste, p. 137. 41. See above, pp. 104-5. 42. Innocent III, PL, vol. 216, no. 27, col. 564; Lock, The Franks, p. 212. 43. Holy Land: see above, pp. 54-59.

Notes to Chapter 24, Prankish Greece: The Military Orders 1. Libro de los fechos et conquistas, pp. 30-31; De constructione castri Saphet, lines 204-10, p. 384. 2. Gregory IX: letter of igth May, 1241, reproduced in Forstreuter, Der deutsche Orden, pp. 237-38, at p. 237; Assises, article 48. 3. Forstreuter, Der deutsche Orden, pp. 73-74; Bon, La Morée franque, pp. 343-44, 429. Monfort: Pringle, 'A Thirteenth-Century Hall', pp. 54-56. Judin: Pringle, Petersen, Dow and Singer, 'Qal'at Jiddin', pp. 135-62. Also, see above, pp. 18-19, 79-80. 4. Tabulae ordinis Theutonici, no. 133, p. 134.

5. Libro de los fechos et conquistas, p. 103. 6. Livre de la conqueste, pp. 404-5; Libro de los fechos et conquistas, pp. 140-43. 7. Ludolph of Sudheim, 'De itinere Terrae Sanctae', p. 23. 8. Forstreuter, Der deutsche Orden, pp. 80-81. 9. Forstreuter, Der deutsche Orden, pp. 71-72. io. Innocent III, PL, vol. 216, no. 149, col. 329. u. Libro de los fechos et conquistas, p. 129; Bon, La Morée franque, p. 338 n. 7. 12. Fourteenth-century survey of royal properties in the Morea, reproduced in Bon, La Marèe franque, p. 690.

NOTES TO PAGES 263-267

379

13. Survey of royal properties in Morea, 1371, in Bon, La Marèe franque, p. 690; Innocent III, PL, vol. 216, no. 150, col. 330. 14. Innocent III, PL, vol. 216, no. 136, cols 323-24; Rubio y Lluch, 'Els Castells', PP- 393-9815. Henry of Valenciennes, Histoire de l'empereur p. no, no n. 4. See also P. Lock, 'The Military Orders in Mainland Greece', in Barber, The Military Orders, PP- 333-339, at pp. 334-3516. Innocent III, PL, vol. 216, no. 144, col, 328; Lock, 'The Military Orders', p. 335. 17. Rubio y Lluch, Diplomatari de l'Orient cátala, no. 63, pp. 78-79. 18. Niccolo da Martoni, 'Relation du pèlerinage', pp. 655-56. 19. Innocent III, PL, vol. 216, lib. 13, no. 120, cols 307—8, lib. 15, no. 69, cols 591-94. 20. Marino Sañudo, 'Istoria', p. 130; Livre de la conqueste, pp. 216-17; Miller, The Latins in the Levant, pp. 133-35. 21. Marino Sañudo, 'Epistulae', ep. 3, p. 293. 22. To Chronikon ton Mareos, p. 214. 23. Riley Smith, The Knights of St John, pp. 215-16; A. Luttrell, 'The Hospitallers at Rhodes, 1306-1421', in HC, vol. 3, pp. 278-313, at pp. 283-88. 24. Boase, 'The Arts in Prankish Greece', pp. 240-43; A. Luttrell, 'Lindos and the Defence of Rhodes, 1306-1522', in A. Luttrell, The Hospitallers of Rhodes and their Mediterranean World (Aldershot, 1992), c. 7, pp. 317-32. 25. Ludolph of Sudheim, 'De itinere Terrae Sanctae', pp. 27-28; Boase, 'The Arts in Prankish Greece', pp. 231-40; A. Luttrell, 'The Military and Naval Organization of the Hospitallers at Rhodes, 1310-1444', in Lutrell, The Hospitallers of Rhodes and their Mediterranean World, c. 19, pp. 133-53, especially at pp. 136-44. 26. Luttrell, 'The Hospitallers at Rhodes', pp. 286-96; Luttrell, 'The Military and Naval Organization of the Hospitallers', pp. 144-47; A. Luttrell, 'The Hospitallers of Rhodes Confront the Turks, 1306-1421', in Luttrell, The Hospitallers of Rhodes and their Mediterranean World, c. 2, pp. 80—116, and see above, pp. 232-33. 27. Innocent VI, Lettres secretes et curiales, ed. P. Gasnault, M. H. Laurent and N. Gotteri, 4 vols (Paris and Rome, 1959-76), vol. 4, no. 2133, pp. 75-76, no. 2134, p. 76; Luttrell, 'The Hospitallers at Rhodes', pp. 296-97. 28. Libro de los fechos et conquistas, pp. 159-60. 29. Luttrell, 'The Hospitallers at Rhodes', pp. 299-313; Lutrell, 'The Hospitallers of Rhodes Confront the Turks', pp. 90-97; A. Luttrell, 'Intrigue, Schism and Violence among the Hospitallers of Rhodes, 1377-1384', in A. Luttrell, The Hospitallers in Cyprus, Rhodes, Greece and the West, 1291-1440 (Aldershot, 1979), c. 23, pp. 30-48. 30. Innocent III, PL, vol. 216, no. 136, col. 323. 31. See above, pp. 192, 198. 32. 'Régestes des empereurs latins de Constantinople, 1204-1261/72', ed. B. Hendrickx, Byzantina, 14 (1988), no. 41, p. 37, and see Lock, 'The Military Orders', PP- 335-36. 33. Port: Niccolo da Martoni, 'Relation du pèlerinage', p. 655. See also A. Luttrell,

380

NOTES TO PAGES 267-2/4

'La Corona de Aragon y la Grecia catalana, 1379-1394', in A. Luttrell, Latin Greece, the Hospitallers and the Crusades, 1291—1440 (London, 1982), c. 11, pp. 219-52, at pp. 241, 247-48. 34. Tabulae ordinis Theutonid, no. 133, p. 134; Forstreuter, Der deutsche Orden, PP- 75> 77-78, 78 n. 16; Bon, La Marèe franque, p. 429. 35. For Kolossi and similar properties belonging to the Orders on Cyprus, see above, pp. 130-34. 36. See above, pp. 59-63, 79-86. Notes to Chapter 25, The Non-Miltary Functions of Fortifications: Residences 1. Ellenblum, 'Three Generations of Prankish Castle-Building', pp. 520-23. 2. A. Bon, La Marèe franque, pp. 608-29, particularly at pp. 608-22; Andrews, Castles, pp. 149-58, particularly at pp. 154-58. 3. Niccolo da Martoni, 'Relation du pèlerinage', p. 651, and see pp. 647—53, 656. See also Setton, Catalan Domination, pp. 227—32. 4. Megaw, 'Military Architecture', pp. 203, 204-5; Enlart, L'art gothique, vol. 2, pp. 525-38, 575-77» 590-955. Enlart, L'art gothique, vol. 2, p. 589. 6. Montfort: E. W. G. Masterman, 'A Crusaders' Fortress in Palestine', PEQ, 60 (1928), pp. 91-97, at p. 96. A more recent survey of Montfort makes no reference to a garden; see Pringle, 'A Thirteenth-Century Hall', pp. 52-81. Athens: Lock, 'The Prankish Tower on the Acropolis, Athens', p. 133. 7. Megaw, 'Supplementary Excavations' pp. 324-25; Benvenisti, The Crusaders in the Holy Land, pp. 374-76. 8. Willbrand of Oldenburg, 'Itinerarium', p. 206. 9. Willbrand of Oldenburg, 'Itinerarium', p. 206. 10. Livre de la conqueste, pp. 220-21; To Chronikon tou Mareos, p. 524. 11. J. Folda, 'Crusader Frescoes at Crac des Chevaliers and Marqab Castle', Dumbarton Oaks Papers, 36 (1982), pp. 177-210. Kerak: Deschamps, La defense du royanme de Jerusalem, pp. 87-88. 12. Matthew Paris, Chronica maiora, vol. 4, p. 256. 13. Deschamps, Le Crac des Chevaliers, pp. 209-10. 14. Oliver of Paderborn, 'Historia Damiatina', p. 171. 15. Joinville, Histoire de Saint Louis, p. 282; 'Gestes', p. 741. 16. Joinville, Histoire de Saint Louis, pp. 282-84 (Jaffa), 318-20, 324-26 (Sidon). 17. Cartulaire, vol. 2, nos 1602-3, PP- 238—40; Deschamps, Le Crac des Chevaliers, pp. 126-27. 18. Livre de la conqueste, pp. 329-34. 19. Pringle, 'A Thirteenth-Century Hall', pp. 60-75. 20. Kyrenia and St Hilarión: Enlart, L'art gothique, vol. 2, pp. 576,591-94. Karytaina and Chlemoutsi: Bon, La Marèe franque, pp. 614, 632. Beaufort: Deschamps, La defense du royanme de Jerusalem, pp. 206-8; Pringle, Gazetteer, no. 44, p. 31.

NOTES TO PAGES 274-277

381

Sidon: Kalayan, 'The Sea Castle of Sidon', pp. 81-89; Deschampa, La defense du royanme de Jerusalem, p. 232. Crac des Chevaliers: Deschamps, Le Crac des Chevaliers, pp. 213-24. 21. 'Gestes', p. 793. 22. Livre de la conqueste, pp. 397-99. 23. Livre de la conqueste, p. 344. 24. Sañudo, 'Istoria', p. loi. Chlemoutsi: Bon, La Marèe franque, pp. 608-22; Andrews, Castles, pp. 154-58. 25. Joinville, Histoire de Saint Louis, pp. 282-84. 26. Livre de la conqueste, pp. 120-23; Libro de los fechos et conquistas, pp. 65-66, 67-68. 27. Henry of Valenciennes, Histoire de l'empereur, p. 62. 28. Pringle, The Red Tower, pp. 15, 105, 107, 124. 29. Documents sur le régime des tenes dans la printipauté de Morée au XlVe siècle, ed. J. Longnon and P. Topping (Paris, 1969), pp. 70-71; P. Lock, 'The Prankish Towers of Central Greece', p. no. Kolossi: Enlart, L'art gothique, vol. 2, pp. 6839430. Bintliff, 'The Prankish Countryside in Central Greece', p. 6, and see pp. 1-18. 31. Lock, 'The Prankish Towers of Central Greece', p. 111; Lock, 'The Medieval Towers of Prankish Greece' p. 138. 32. Lock, 'The Towers of Euboea', pp. 107-26. 33. Lock, 'The Towers of Euboea', p. 113, no. 13. 34. 'Gestes', pp. 814-15; Jacoby, 'Crusader Acre in the Thirteenth Century', pp. 19-36; Rey, 'Etude sur la topographic', pp. 137-40; Pringle, 'Towers in Crusader Palestine', pp. 337-38. 35. 'Gestes', p. 747; 'Annales de Terre Sainte', p. 448 (text B). 36. Athens: Lock, 'The Prankish Tower on the Acropolis, Athens', pp. 131-33; Lock, 'The Medieval Towers of Prankish Greece', p. 132. Negroponte: Lock, 'The Towers of Euboea', p. 116, no. 31. 37. Lock, 'The Towers of Euboea', p. 113, no. 13. 38. Enlart, L'art gothique, vol. 2, p. 576. 39. D. Renn, 'Burhgeat and Gonfanon: Two Sidelights from the Bayeux Tapestry', Anglo-Norman Studies, 16 (1994), pp. 177-98, at pp. 182-83, cited from C. Coulson, 'Cultural Realities and Reappraisals in English Castle-Study', Journal of Medieval History, 22 (1996), pp. 171-207, at p. 175. I am very grateful to Dr Coulson for providing me with this information.

Notes to Chapter 26, The Non-Miltary Functions of Fortifications: Prisons 1. 2. 3. 4.

La regle du Temple, els 554, 573, 592, 593, 603. 'Gestes', p. 805. Willbrand of Oldenburg, 'Itinerarium', p. 204. The ditches at Saphet, for example, were more than fifteen metres deep. See De construction castri Saphet', p. 383, line 171.

382

NOTES TO PAGES 277-281

5. Deschamps, La defense du comté de Trìpoli, p. 231; G. Saadé, 'Histoire du chateau de Saladin', Studi Medievali, series 3,9 (1968), pp. 980-1016, at p. 1004; Kennedy, Crusader Castles, p. 96. 6. 'Gestes', p. 788; Deschamps, La defense du comté de Trìpoli, p. 207. 7. Amadi, Chronique, pp. 386, 388, 390; Bustron, Chronique, pp. 143-45. 8. Leontios Makhairas, Recital, e. 85, p. 77, e. 257, pp. 237-39; Amadi, Chronique, p. 408; Bustron, Chronique, p. 257; Strambaldi, Chronique, pp. 34-35, 101. 9. Amadi, Chronique, p. 393; Bustron, Chronique, p. 245. 10. Makhairas, Recital, cc. 610-11, pp. 601-3; Strambaldi, Chronique, pp. 255-56. 11. William of Machaut, La prise, pp. 258-59, 265. See also Makhairas, Recital, c. 260, p. 241, c. 265, p. 247; Amadi, Chronique, p. 422; Bustron, Chronique, p. 271; Strambaldi, Chronique, p. 102; Enlart, Van gothique, vol. 2, pp. 520-21. 12. Makhairas, Recital, c. So, p. 75. 13. Samuel of Ani, 'Chronographie', pp. 466-67. 14. Livre de la conqueste, p. 149. 15. The Crusade and Death of Richard I, ed. R. C. Johnston (Oxford, 1961), p. 32. 16. Ibn al-Furat, 'Selections', p. 160. 17. Ibn al-Furat, 'Selections', pp. 88-89; De construction castri Saphet, pp. 381-82, lines 115-24. 18. Makhairas, Recital, cc. 594-97, pp. 591-93. 19. Arsuf: Ibn al-Furat, 'Selections', p. 78; al-Makrizi, Histoire des sultans, vol. i(b), p. 10. Damascus: ibid., vol, i(a), p. 168. 20. Matthew Paris, Chronica maiora, vol. 2, p. 371; Arnold of Lübeck, 'Chronica Slavorum', p. 178; Deschamps, La defense du comté de Trìpoli, pp. 279-80. 21. George Pachymeres, 'De Michaele et Andronico', vol. i, p. 209.

22. Livre de la conqueste, pp. 245-47, 260. Chlemoutsi: Bon, La Marèe franque, pp. 608-22. Notes to Chapter 27, The Non-Miltary Functions of Fortifications: Taxation, Justice and Administration 1. Assises, articles 43, 177, and see article 9. 2. Assises, article 94. In 1275, for example, the high court gathered at Andravida to hear an important land dispute. See Livre de la conqueste, pp. 197-211. 3. 'Nouvelles preuves de l'histoire de Chypre sous le regne des princes de la maíson de Lusignan', ed. L. de Mas Latrie, Bibliothèque de l'Ecole des Chañes, 35 (Paris, 1874), pp. 120-21; Edbury, The Kingdom of Cyprus, p. 194. 4. 'Livre de Jean d'Ibelin', RHCLois, vol. i, p. 420. 5. Livre de la conqueste, pp. 328-29. See also Bon, La Morée franque, pp. 657-58. 6. De constructione castri Saphet, p. 386, lines 257-62. 7. Cartulaire, vol. 3, no. 3051, p. 64. See also Kedar, 'The Subjected Muslims', pp. 160-74. 8. For this and similar twelfth-century examples, see Smail, Crusading Warfare, p. 61.

NOTES TO PAGES 281-284

383

9. 'Gestes', p. 815; Barber, The New Knighthood, pp. 267-68, 311. 10. Joinville, Histoire de Saint Louis, pp. 222-26. 11. La regle du Temple, cl. 89. 12. La regle du Temple, cl. 129. 13. Tabulae ordinis Theutonici, no. 100, p. 78-81. For the treasuries of the Hospitallers and Teutonic Knights, see also Riley-Smith, The Knights of St John, pp. 310-12; Sterns, 'The Teutonic Knights in the Crusader States', p. 329. 14. De constructione castri Saphet, p. 384, line 182. 15. Pringle, The Red Tower, p. 13, and see p. 60. Tibbie, Monarchy and Lordships, p. 142, disagrees with Pringle, and believes Qaqun was held by the Templars rather than the lords of Caesarea. 16. Prawer, The Latin Kingdom of Jerusalem, pp. 289-94; Pringle, 'Towers in Crusader Palestine', p. 338 17. Bon, La Marèe franque, p. 612. 18. Leontios Makhairas, Recital, cc. 594-97, pp. 591-93; Amadi, Chronique, p. 490; Bustron, Chronique, p. 349; Strambaldi, Chronique, pp. 250-51; Edbury, The Kingdom of Cyprus, pp. 191-92; J. Richard, 'The Institutions of the Kingdom of Cyprus', in HC, vol. 6, pp. 150-74, at pp. 162—63; Enlart, L'ari gothique, vol. 2, pp. 519-20. 19. Cartulaire, vol. 2, no. 1193, pp. 31—40; Deschamps, La defense du comté de Tripoli, pp. 279-84. Notes to Chapter 28, The Non-Miltary Functions of Fortifications: Agriculture and Daily Life 1. De constructione castri Saphet, p. 386, lines 257-58. 2. Matthew Paris, Chronica malora, vol. 3, p. 14; Oliver of Paderborn, 'Historia Damiatina', pp. 171-72. 3. Cartulaire, vol. 2, no. 2726, p. 777, no. 2811, pp. 815-17; Riley-Smith, The Knights of St John, pp. 413-17. 4. Willbrand of Oldenburg, 'Itinerarium', pp. 202, 206, 208, 210; Burchard of Mount Sion, 'Descriptio', pp. 23, 29, 33-34. 5. Saphet: De constructione castri Saphet, pp. 384-85, lines 215-33. 'Atlit: Oliver of Paderborn, 'Historia Damiatina', p. 171; Matthew Paris, Chronica malora, vol. 3. P- H6. William of Tyre, Chronicon, vol. i, bk 11, c. 26, p. 535; J. Prawer, Crusader Institutions (Oxford, 1980), pp. 109-10. 7. Mhalia: Tabulae ordinis Theutonici, no. 112, pp. 91-94; Kolossi: Enlart, L'art gothique, vol. 2, p. 694. 8. Oliver of Paderborn, 'Historia Damiatina', p. 171; Matthew Paris, Chronica malora, vol. 3, p. 14; Johns, Guide to 'Atlit, p. 72. 9. Corfu: Thiriet, Régestes, vol. i, no. 850, p. 202. Castellorizzo: Le saint voyage de Jhérusalem du seigneur d'Anglure, ed. F. Bonnardot and A. Longnon (Paris, 1878), pp. 89-91.

384

NOTES TO PAGES 284-286

10. Benvenisti, The Crusaders in the Holy Land, p. 251; Deschamps, La defense du royanme de Jerusalem, p. 124; Pringle, 'Survey of Castles in the Crusader Kingdom of Jerusalem', p. 89; Pringle, Gazetteer, no. 133, pp. 62-64. 11. Livre de la conqueste, p. 273. 12. Villehardouin, La conquête, p. 250. 13. Makhairas, Recital, c. 470, p. 453; Strambaldi, Chronique, p. 194. 14. Margat: Willbrand of Oldenburg, 'Itinerarium', p. 210. Saphet: De constructione castri Saphet, p. 385, lines 229-33. 15. Ibn al-Furat, 'Selections', p. 85. 16. Margat: Willbrand of Oldenburg, 'Itinerarium', p. 210. Cave de Tyron: Deschamps, La defense du royanme de Jerusalem, pp. 219-20. 17. 'Gestes', p. 686. 18. Marino Sañudo, 'Líber secretorum', p. 229. 19. La regle du Temple, cl. 126, deals with supplies of this type being provided for castles. Stables situated within the outer defences at 'Atlit could accommodate a large number of cattle and horses. See C. N. Johns, 'Excavations at Pilgrims' Castle ('Atlit): Stables at the South-West of the Suburb', QDAP, 5 (1935-36), pp. 31-60. 20. Assises, article 159. 21. Deschamps, Le Crac des Chevaliers, p. 155; Lawrence, 'The Castle of Baghras', pp. 58-59. See also Kennedy, Crusader Castles, p. 100. 22. Johns, 'The Faubourg and its Defences', p. 120. 23. Johns, Guide to 'Atlit, pp. 96-97. 24. William of Tyre, Chronicon, vol. 2, bk 22, c. 31, p. 1059. 25. Deschamps, Le Crac des Chevaliers, p. 189. 26. Beaufort: Deschamps, La defense du royanme de Jerusalem, p. 199. See also Kennedy, Crusader Castles, pp. 99-100; Deschamps, La defense du comté de Tripoli, p. 153. 27. Deschamps, Le Crac des Chevaliers, p. 152. 28. 'Gestes', p. 817; Bustron, Chronique, p. 127. 29. Paphos: Megaw, 'Supplementary Excavations', p. 324. Montfort: Pringle, 'A Thirteenth-Century Hall', pp. 57-60, 68-71. 30. Deschamps, Le Crac des Chevaliers, pp. 203-4; Deschamps, La defense du comté de Trìpoli, pp. 153-54. 31. Crac des Chevaliers and Margat: Willbrand of Oldenburg, 'Itinerarium', pp. 208-10. Saphet: De constructione castri Saphet, p. 384, lines 204-6. 32. Benvenisti, The Crusaders in the Holy Land, p. 299, and see also pp. 374—75 for more details on crusader kitchens. 33. De constructione castri Saphet, p. 384, lines 210-15. 34. La regle du Temple, cl. 196. 35. La regle du Temple, cl. 662. 36. Megaw, 'Supplementary Excavations', pp. 333-34, 335, 337. 37. La regle du Temple, els 325 (masons), 604 (cobblers).

NOTES TO PAGES 286-290

385

38. Saphet: De constmctione castri Saphet, p. 384, lines 185-90. Paphos: Megaw, 'Supplementary Excavations', p. 332. 39. Tabulae ordinis Theutonid, nos 27-29, pp. 23-25; Rey, 'Etude sur la topographic', p. 125. 40. See, for example, Baybars' preparations for his campaign against Saphet in 1266: al-Makrizi, Histoire des sultans, vol. i(b), p. 28. 41. De constmctione castri Saphet, p. 384, lines 204-10. 42. De constmctione castri Saphet, pp. 381-82, lines 115-24. 43. al-Makrizi, Histoire des sultans, vol. i(b), pp. 27-31; Ibn al-Furat, 'Selections' pp. 88-96; 'Gestes', pp. 764-66; 'Eracles', pp. 454-55; Marino Sañudo, 'Liber secretorum', pp. 222-23. 44. For more details on turcopoles, see Marshall, Warfare in the Latin East, pp. 58-60. 45. 'Gestes', p. 756. See also Kedar, 'The Subjected Muslims', pp. 152-54. 46. Marshall, Warfare in the Latin East, p. 119. 47. Amadi, Chronique, p. 238. See also 'Gestes', pp. 849-50; Bustron, Chronique, P-13348. Livre de la conqueste, p. 71. 49. Assises, articles 70, 71. 50. Nikephoros Gregoras, 'Historia Byzantina', vol. i, pp. 25-28; George Akropolites, 'Annales', cols 1035-38; Sinogowitz, 'Zur Eroberung Thessalonikes im Herbst 1224', p. 28. 51. Ibn al-Furat, 'Selections', p. 139. 52. Cartulaire, vol. 4, no. 3308, pp. 291-93. 53. RRH, no. 1329, p. 347. 54. Ibn al-Furat, 'Selections', pp. i43~49;'Gestes' pp. 768, 777; al-Makrizi, Histoire des sultans, vol. i(b), pp. 84-85. 55. Ibn al-Furat, 'Selections', p. 130; al-Makrizi, Histoire des sultans, vol. i(b), p. 56; Cartulaire, vol. 3, no. 3400, p 231. For the capture of Montfort in 1271, see Ibn al-Furat, 'Selections', pp. 106-12; al-Makrizi, Histoire des sultans, i(b), p. 87; 'Gestes' p. 778. 56. Marino Sañudo, 'Istoria', pp. 118-19; Livre de la conqueste, pp. 160,170-76; Libro de los fechos et conquistas, pp. 91, 92, and see above, p. 193. 57. Buchón, Nouvelles recherches, vol. 2, nos 24, 25, 27, pp. 330-31, 332-33. 58. Buchón, Nouvelles recherches, vol. 2, nos 25-29, pp. 143-60.

Notes to Chapter 29, The Non-Miltary Functions of Fortifications: Trade 1. Estienne de Lusignan, Description, fol. 24. 2. Jacoby, 'The Rise of a New Emporium in the Eastern Mediterranean', pp. 14950. 3. Frederick II: 'Gestes', p. 676; Bustron, Chronique, pp. 63-64;. Louis IX: Joinville, Histoire de Saint Louis, pp. 72-74.

386

NOTES TO PAGES 290-294

4. Jacoby, 'The Rise of a New Emporium in the Eastern Mediterranean', pp. 14754; Enlart, L'art gothique, vol. 2, pp. 673-83. 5. The defences of Acre and Tyre protected the people and economic prosperity of these cities on many occasions. See, for example, al-Makrizi, Histoire des sultans, vol. i(a), pp. 199-200; Ibn al-Furat, 'Selections', pp. 57-59 (Mamluk raid on Acre in 1263), and al-Makrizi, Histoire des sultans, vol. i(b), pp. 27-28; Ibn al-Furat, 'Selections', p. 87 (Mamluk raid on Acre and Tyre in 1266). See also above, p. 39. 6. Buchón, Nouvettes recherches, vol. 2, no. 9, pp. 98-103; Bon, La Marèe franane, pp. 320-22. 7. Lock, The Franks, pp. 154-55, 252-53; Miller, The Latins in the Levant, pp. 495, 498; Andrews, Castles, pp. 14» 598. Patras: Gerland, Neue Quellen, pp. 89-92. Monemvasia: George Sphrantzes, 'Chronicon minus', ed. I. Bekker, CSHB (Bonn, 1838), pp. 399-404. 9. For detailed descriptions of these, see Rey, 'Etude sur la topographic', pp. 137 (Venice), 137-38 (Genoa), 138-39 (Pisa); Benvenisti, The Crusaders in the Holy Land, pp. 98-100 (Pisa), 100-2 (Genoa), 102-4 (Venice); Jacoby, 'Crusader Acre in the Thirteenth Century', pp. 19-26 (Pisa), 26-30 (Genoa), 30-36 (Venice); Kesten, The Old City of Acre, pp. 29-39 (Genoa), 40-45 (Venice), 59-67 (Pisa). 10. 'Gestes', pp. 742-48. For more details on this dispute, see Riley-Smith, The Feudal Nobility, pp. 215-17. 11. Makhairas, Recital, c. 372, p. 353. 12. 'Gestes', p. 829. 13. See, for example, Thiriet, Régestes, vol. i, no. 371, p. 97, and see Lock, 'The Medieval Towers of Prankish Greece', p. 139; Lock, 'The Frankish Towers of Central Greece', pp. 108-9; Lock, 'The Towers of Euboea', p. ill, nos 3, 4, p. 116, no. 31, p. 117, no. 35. As we have seen, towers of this kind also often acted as residences and status symbols. See chapter twenty-five. 14. Lock, 'The Towers of Euboea', p. 117, no. 35. 15. Le trésor des ¿hartes d'Armenie, no. 23, pp. 170-75, and see pp. 35-38, 49-50. For trade patterns in this area, see also Cahen, La Syrie du Nord, pp. 689-91; Barber, The New Knighthood, p. 240. 16. 'Gestes', pp. 768-69; Jacoby, 'Crusader Acre in the Thirteenth Century', pp. 810; Gertwagen, 'The Crusader Port of Acre', pp. 559-60. 17. Andrews, Castles, pp. 187-91, particularly p. 187. For Venetian defence measures regarding this bridge, see also Lock, 'The Towers of Euboea', p. 118. 18. RRH, no. 977, p. 257. 19. Jacoby, 'Crusader Acre in the Thirteenth Century', pp. 15-19. For a recent reinterpretation of the layout of Acre's harbour, see Gertwagen, 'The Crusader Port of Acre', pp. 554-68. 20. Oliver of Paderborn, 'Historia Damiatina', p. 169; Johns, Guide to 'Atlit, pp. 9498. 21. RRH, no. 1043, p. 272. The Templars were exempt from paying when using

NOTES TO PAGES 294-297

387

this gate. See also Cartulaire, vol. 2, no. 2,058, pp. 455-57; Deschampa, La defense du comté de Tripoli, pp. 284—85. 22. Document translated and reproduced in Alishan, Sissouan, p. 239. See also Edwards, Fortifications, pp. 147-49; Forstreuter, Der deutsche Orden, p. 65. 23. 'Gestes', p. 839.

Notes to Chapter 30, The Non-Miltary Functions of Fortifications: The Church, Pilgrims and Crusaders 1. Oliver of Paderborn, 'Historia Damiatina', p. 169. 2. 'Les chemins et pèlerinages de la Terre Sainte', ed. H. Michelant and G. Raynaud, in Itinéraires a Jerusalem (Geneva, 1882), p. 180; Johns, Guide to 'Atlit, P. 553. Joinville, Histoire de Saint Louis, pp. 328-30; Barber, The New Knighthood, p. 81. 4. De construction castri Saphet, pp. 386-87, lines 268-90. 5. La regle du Temple, cl. 188, mentions paupers being fed at castles. For more details on infirmaries, see Sterns, 'The Teutonic Knights in the Crusader States', pp. 341-48; Barber, The New Knighthood, pp. 217-18; A. Luttrell, 'The Hospitallers' Medical Tradition, 1291-1530', in Barber, The Military Orders, pp. 64-81. 6. Willbrand of Oldenburg, 'Itinerarium', p. 212; Burchard of Mount Sion, 'Descriptio', pp. 30-31; B. Hamilton, The Latin Church in the Crusader States: The Secular Church (London, 1980), p. 215. 7. Folda, 'Crusader Frescoes', pp. 187-196, particularly at pp. 193-96. 8. Ibn al-Furat, 'Selections', pp. 126, 161-62, 165; al-Makrizi, Histoire des sultans, vol. i(b), p. 127; Cahen, La Syne du Nord, pp. 697-98, 717. 9. Bar Hebraeus, Chronography, p. 453; Alishan, Sissouan, pp. 172-73. 10. Robinson and Hughes, 'Lampron: Castle of Armenian Cilicia', pp. 188—89. 11. Ibn al-Furat, 'Selections', pp. 56-57. 12. Benvenisti, The Crusaders in the Holy Land, pp. 358-59; Pringle, Churches, vol. 2, pp. 66-68. 13. De constructione castri Saphet, p. 382, lines 124-28. 14. Tabulae ordinis Theutonici, no. 72, pp. 56-57. 15. Oliver of Paderborn, 'Historia Damiatina', p. 171. 16. Deschamps, Le Crac des Chevaliers, pp. 197-201; Deschamps, La defense du comté de Trìpoli, pp. 254, 277-78; Rey, Etude, pp. 26-28, 48-49, 88-89; Barber, The New Knighthood, p. 194. 17. Johns, Guide to 'Atlit, pp. 52-58; Pringle, 'Reconstructing the Castle of Safad', pp. 147-48; E. Lambert, L''architecture des Templiers (Paris, 1978), pp. 5—19,30-31, 92-93; Barber, The New Knighthood, pp. 194-95; Pringle Churches, vol. i, no. 27, PP- 75-So. 18. Deschamps, Le Crac des Chevaliers, pp. 216-24; Harper and Pringle, 'Belmont Castle', pp. 104, 116. 19. Johns, Guide to 'Atlit, p. 24.

388

NOTES TO PAGES 297-307

20. Beaufort: Deschamps, La defense du royanme de Jerusalem, pp. 206-8; Pringle, Gazetteer, no. 44, p. 31. Sidon: Kalayan, 'The Sea Castle of Sidon', pp. 81-89; Deschamps, La defense du royanme de Jerusalem, p. 232. 21. For more details on these issues, see Riley-Smith, The Knights of St John, PP- 375-420; Barber, The New Knighthood, pp. 55-63, 195-98. 22. Riley-Smith, The Knights of St John, pp. 410-13. 23. Riley-Smith, 'The Templars and the Castle of Tortosa', pp. 278-88; Barber, The New Knighthood, pp. 81-83. 24. Sidon: Ernoul, Chronique d'Ernoul et de Bernard le Trésoríer, p. 459; 'Eracles', p. 365; 'Gestes', p. 676; Kennedy, Crusader Castles, pp. 122—24. 'Atlit: Oliver of Paderborn, 'Historia Damiatina', pp. 168, 207. 25. Joinville, Histoire de Saint Louis, p. 284. Richard I: Itinerarium, p. 317.

Notes to Chapter 31, Conclusion 1. Pryor, Geography, Technology and War, p. 112. 2. Ibn al-Furat, 'Selections', p. 54.

Bibliography PRIMARY SOURCES

Western Sources Acta Aragonensia, ed. H. Pinke, 3 vols (Berlin and Leipzig, 1908—22). Actes relatifs a la principante de Marèe, 1289-1300, ed. C. Perrat and J. Longnon (Paris, 1967). Amadi, Francesco, Chroniques d'Amadi et de Stmtnbaldi, ed. R. de Mas Latrie (Paris, 1891). Ambroise, L'estoire de la guerre sainte, ed. G. Paris (Paris, 1897). 'Annales Genuenses', ed. G. Petti Balbi, RIS n.s., vol. 17, part 2 (Bologna, 1975). 'Annales de Terre Sainte', ed. R. Röhricht and G. Raynaud, in Archives de l'Orient latin, 2 (1884), pp. 429-61. Antiquitates italicae medii aevi, ed. L. A. Muratori, 6 vols (Rome, 1738-42), vol. 5. Arnold of Lübeck, 'Chronica Slavorum', ed. J. Lappenberg, in MGH SS, vol. 21 (1868). Les Assises de Romanice, ed. G. Recoura (Paris, 1930). Aubrey of Trois Fontaines, 'Chronica', ed. P. Scheffer-Boichorst, in MGH SS, vol. 23 (1874). 'Bans et ordonnances des rois de Chypre', in RHCLois, vol. 2. Barbaro, Daniele, 'Cronica del Trivisano della città di Venezia', short extract in J. K. Fotheringham, Marco Sañudo: Conqueror of the Archipelago (Oxford, 1915), p. 106. Benvenuto di San Giorgio, 'Historia Montisferrati', ed. L. A. Muratori, in RIS, vol. 23 (i733)Boniface VIII, Registres, ed. G. Digard et al., 4 vols (Paris, 1884-1939). Burchard of Mount Sion, 'Descriptio Terrae Sanctae', ed. J. Laurent, in Peregrinatores medii aevi quatuor (Leipzig, 1873). Bustron, Florio, Chronique de l'ile de Chypre, ed. R. de Mas Latrie (Paris, 1886). Cartulaire generai de l'orare des Hospitaliers de St Jean de Jerusalem, 1100-1310, ed. J. Delaville Ie Roulx, 4 vols (Paris, 1894—1906). Cartulaire general de Vordre du Temple, niyï-iiso, ed. G. Albon (marquis d') (Paris, 1913)'The Cartulaire de Manosque: A Grant to the Templars in Latin Syria and a Charter of King Hugh I of Cyprus', ed. P. W. Edbury, Bulletin of the Institute of Historical Research, 51 (1978), pp. 174-81.

390

UNKNOWN CRUSADER CASTLES

Chañes de Terre Sainte provenant de l'abbaye de Josaphat, ed. H. F. Delaborde (Paris, 1880). Chronica de Mauros, ed. J. Stevenson (Edinburgh, 1835). Chronique d'Ernoul et de Bernard le Trésorier, ed. L. de Mas Latrie (Paris, 1871). Clement VI, Lettres closes, patentes et curíales, ed. E. Deprez, 3 vols (Paris, 1958-59). La continuation de Guillaume de Tyr (1184-1197), ed. M. R. Morgan (Paris, 1982). 'Continuation de Guillaume de Tyr de 1229 a 1261, dite du manuscrit de Rothelin', in RHCOc, voi. 2. 'Cronica antica di Venetia', short extract in J. K. Fotheringham, Marco Sañudo: Conqueror of the Archipelago (Oxford, 1915), p. no. Daniel, Russian abbot; account of journey to Mount Athos summarised in A. Soloviev, 'Histoire du monastère russe au Mont-Athos', Byzantion, 8 (1933), pp. 213-38. De constructione castri Saphet, in R. B. C. Huygens (ed.), 'Un nouveau texte du traite De constructione castri Saphet', Studi Medievali, 3rd series, 6, part i (1965), pp. 35587. Diplomatari de l'Orìent cátala, ed. A. Rubio y Lluch (Barcelona, 1947). 'Directorium ad faciendum passagium transmarinum', ed. C. Beadey, American Historical Review, 12 (1906-7), pp. 810-57; 13 (1907-8), pp. 66-115. Documents chypriotes des archives du Vatican, XlVe et XV siècles, ed. J. Richard (Paris, 1962). 'Documents chypriotes du début du XlVe siècle', ed. C. Kohler, J?OL, il (1905-8), pp. 440-52'Documents nouveaux servant de preuves a l'histoire de l'ile de Chypre sous le regne des princes de la maison de Lusignan', ed. L. de Mas Latrie, in Collection des documents inédits: mélanges historiques, vol. 4 (1882). Documents sur le regime des terres dans la principauté de Marèe au XIV siede, ed. J. Longnon and P. Topping (Paris, 1969). Estienne de Lusignan, Description de toute l'isle de Cypre (Paris, 1580). 'L'estoire de Eracles empereur et la conqueste de la terre d'Outremer', in RHCOc, Vols 1—2.

Fulcher of Chartres, Historia Hierosolymitana, ed. H. Hagenmeyer (Heidelberg, 1913). Geoffrey of Villehardouin, La conquête de Constantinople, ed. N. de Wailly (Paris, 1882). Gesta Regis Henna Secundi Benedicti Abbatìs, ed. W. Stubbs, RS, 49, 2 vols (London, 1869). 'Les Gestes des Chiprois', in RHCArm, vol. 2. Gregory IX, Registres, ed. L. Auvray, 3 vols (Paris, 1890-1955). Henry, Latin emperor (1206-16); various letters sent to Innocent III and the West, in RHGF, vol. 18 (1879), pp. 525-33; vol. 19 (1880), p. 514. Henry of Valenciennes, Histoire de l'empereur Henri de Constantinople, ed. J. Longnon (Paris, 1948). Honorius III, Regesta, ed. P. Pressutti, 2 vols (Rome, 1888-95, reprinted New York, 1978).

BIBLIOGRAPHY

3pl

Innocent III, 'Opera Omnia', ed. J. P. Migne, in PL, vols 214-17. Innocent III, Die Register Innocenz III, ed. O. Hagender and A. Haidacher, 2 vols so far (Graz and Cologne, 1964- ). Itinerarium peregrinarían et gesta Regis Ricardi, ed. W. Stubbs, RS, 38,2 vols (London, 1864-65), vol. i. Jacques de Vitry, Lettres de Jacques de Vitry, ed. R. B. C. Huygens (Leyden, 1960). John XXII, Lettres communes, ed. G. Mollat, 16 vols (Paris, 1904-47). John of Joinville, Histoire de Saint Louis, ed. N. de Wailly (Paris, 1874). Libro de los fechos et conquistas del principado de la Marea, ed. A. Morel-Fatio (Geneva, 1885). 'Livre de Jean d'Ibelin', in RHCLois, voi. i. Livre de la conqueste de la princée de l'Amorée: Chronique de Morée, 1204—1305, ed. J. Longnon (Paris, 1911). Ludolph of Sudheim, 'De itinere Terrae Sanctae', ed. F. Deycks, in Bibliothek des litterarischen Vereins in Stuttgart, vol. 25 (Stuttgart, 1851). Martin da Canal, 'La chronique des Veneciens', ed. G. Galvani, in Archivio storico italiano, vol. 8 (1845). Matthew Paris, Chronica maiora, ed. H. R. Luard, RS, 57, 7 vols (London, 1872-83). Muntaner, Ramon, Crònica, ed. M. Gusta, 2 vols (Barcelona, 1979). Muntaner, Ramon, The Chronicle of Muntaner, trans. A Goodenough, 2 vols (London, 1920-21). Navagerio, Andrea, 'Storia Venezia', ed. L. A. Muratori, in JRJS, voi. 23 (1733). Neue Quellen zur Geschichte des lateinischen Erzbistmus Patras, ed. E. Gerland (Leipzig, 1903). Niccolo da Martoni, 'Relation du pèlerinage a Jerusalem de Nicolas de Martoni, notaire Italien, 1394-1395', ed. L. LeGrand, ROL, 3 (1895), pp. 566-669. Nicholas IV, Registres, ed. E. Langlois, 2 vols (Paris, 1886-1905). 'Nouvelles preuves de l'histoire de Chypre sous le règne des princes de la maison de Lusignan', ed. L. de Mas Latrie, in Bibliothèque de l'Ecole des Charles, voi. 32 (1871), pp. 341-78; vol. 34 (1873), pp. 47-87; vol. 35 (1874), pp- 99-158. Nouvelles recherches historiques sur la principante francaise de Morée et ses hautes baronnies a la suite de la Quatrième Croisade, ed. J. A. C. Buchón, 2 vols (Paris, 1843)Oliver of Paderborn, 'Historia Damiatina', ed. H. Hoogeweg, in Bibliothek des litterarischen Vereins in Stuttgart, vol. 202 (Tübingen, 1894). Philip Mousket, 'Chronique rimée', ed. baron de Reiffenberg in Collection de chroniques belges inédites, z vols (Brüssels, 1838), vol. 2. Philip of Mézières, 'Le songe du vieil pelerin', extracts in L. de Mas Latrie, Histoire de l'ile de Chypre sous le règne des princes de la maison de Lusignan, 3 vols (Paris, 1852-61), vol. 2, pp. 332-33Philip of Mézières, The Life of St Peter Thomas, ed. J. Smet (Rome, 1954). Philip of Novara, Mémoires, ed. C. Kohler (Paris, 1913). Philip of Novara, The Wars of Frederick II against the Ibelins in Syria and Cyprus, ed. and trans. J. L. LaMonte (New York, 1936).

392

UNKNOWN

CRUSADER CASTLES

Regesta Regni Hierosolymitani (1097-1291), ed. R. Röhricht (Innsbruck, 1893; Additamentum, 1904). Régestes des deliberations du sénat de Venise concernant la Romanie, ed. F. Thiriet, 3 vols (Paris, 1958-61). 'Régestes des empereurs latins de Constantinople, 1204-1261/72', ed. B. Hendrickx, Byzantina, 14 (1988). 'A Register of the Cartulary of the Cathedral of Santa Sophia of Nicosia', ed. J. L. LaMonte, Byzantion, 5 (1930), pp. 439-522. La regle du Temple, ed. H. de Curzon (Paris, 1886). Richard of Devizes, Chronicon de tempore Regis Ricardi Primi, ed. J. T. Appleby (London, 1963). Robert of Clari, La conquête de Constantinople, ed. P. Lauer (Paris, 1924). Roger of Howden, Chronica, ed. W. Stubbs, RS, 51, 4 vols (London, 1868-71). Le saint voyage de Jhérusalem du seigneur d'Anglure, ed. F. Bonnardot and A. Longnon (Paris, 1878). Sañudo, Marino (Torsello, or the Elder), 'Liber secretorum fidelium crucis', ed. J. Bongars, in Gesta Dei per Francos, 2 vols (Hannau, 1611), voi. 2. Sañudo, Marino (Torsello, or the Elder), 'Epistulae', ed. J. Bongars, in Gesta Dei per Francos, 2 vols (Hannau, 1611), voi. 2. Sañudo, Marino (Torsello, or the Elder), 'Istoria del regno di Romania sive regno di Morea', ed. C. Hopf, in Chroniques gréco-romanes inédites ou peu connues (Berlin, 1873). Sañudo, Marino (the Younger), 'Vite de' duchi di Venezia', ed. L. A. Muratori, RIS, vol. 22 (1733). Sauger, Robert, Histoire nouvelle des anciens dues de l'Archipel (Paris, 1699), extract in J. K. Fotheringham, Marco Sañudo: Conqueror of the Archipelago (Oxford, 1915), pp. 113-22. Simon de Saint-Quentin, Histoire des Tartares, ed. J. Richard (Paris, 1965). Strambaldi, Diomedes, Chroniques d'Amadi et de Strambaldi, ed. R. de Mas Latrie (Paris, 1891). Supplementary Excerpts on Cyprus: Further Materials for a History of Cyprus, trans. and ed. T. A. H. Mogabgab (Nicosia, 1941). Surveys of royal properties in the Morea undertaken in the fourteenth century, reproduced in A. Bon, La Marèe franque: recherches historiques, topographiques et archéologiques sur la principauté d'Acaïe, 1205-1430 (Paris, 1969), pp. 689-92. Tabulae ordinis Theutonici, ed. E. Strehlke (Berlin, 1869). 'The Templars and the Castle of Tortosa in Syria: An Unknown Document Concerning the Acquisition of the Fortress', ed. J. S. C. Riley-Smith, EHR, 84 (1969), pp. 278-88. Thomas Wykes, 'Chronicon', ed. H. R. Luard, in Annales Monastici, 5 vols, RS, 36 (London, 1864-69), vol. 4. Urban V, Lettres secretes et curiales du pape Urbain V se rapportant a la France, ed. G. Mollat (Paris, 1955). Urkunden zur älteren Handels- und Staatsgeschichte der Republik Venedig mit

BIBLIOGRAPHY

393

besonderer Beziehung au/Byzanz und die Levante, ed. G. Tafel and G. Thomas, 2 vols (Vienna, 1856). Villani, Johannis, 'Florentini historia universalis', ed. L. A. Muratori, RIS, vol. 13 (1728). Willbrand of Oldenburg, 'Itinerarium Terrae Sanctae', ed. S. de Sandoli, in lunera Hierosolymitana Crucesignatorum, 4 vols ( Jerusalem, 1978-84), vol. 3. William of Machaut, La prise d'Alexandre, ou chronique du roi Pierre Ier de Lusignan, ed. L. de Mas Latrie (Geneva, 1877). William of Nangis, 'Gesta Ludovici IX', in RHGF, vol. 20 (1881). William of Tyre, Chronicon, ed. R. B. C. Huygens, 2 vols (Turnhout, 1986). Zeno, lacobo, 'Vita Caroli Zeni', ed. G. Zonta, RIS n.s., vol. 19, part 6 (1931-41).

Greek Sources Akropolites, George, 'Annales', ed. J. P. Migne, in PG, vol. 140 (1887). Choniates, Niketas, 'Historia', ed. J. P. Migne, in PG, vol. 139 (1894). To Chroniken tou Mareos: The Chronicle of Marea, ed. J. Schmitt (London, 1904). Gregoras, Nikephoros, 'Nicephori Gregorae Byzantina historia', ed. L. Schopen, in CSHB, 3 vols (Bonn, 1829-55). John VI Kantkouzenos, 'loannis Cantacuzeni eximperatoris historiarum libri IV', ed. L. Schopen and B. Niehbuhr, in CHSB, 3 vols (Bonn, 1828-32). Makhairas, Leontios, Recital Concerning the Sweet Land of Cyprus, Entitled 'Chronicle', ed. and trans. R. M. Dawkins, 2 vols (Oxford, 1932). Pachymeres, George, 'Georgii Pachymeris de Michaele et Andronico Palaeologis', ed. I. Bekker, in CHSB, 2 vols (Bonn, 1835).

Armenian and Syriac Sources Bar Hebraeus, The Chronography of Gregory Abu'l Faraj, the Son of Aaron, the Hebrew Physician Commonly Known as Bar Hebraeus, trans. E. A. Wallis Budge (Oxford, 1932). Constable Sempad, 'Chronique du royaume de la Petite Armenie', in RHCArm, vol. i. Constable Sempad, La chronique attribuée au connétabk Smbat, ed. and trans. G. Dédéyan (Paris, 1980). Hethoum the Historian, 'Table chronologique de Héthoum, corate de Gorigos', in RHCArm, voi. i. Jean Dardel, 'Chronique d'Armenie', in RHCArm, voi. 2. Lettres pontificales concernant l'histoire de la Petite Armenie au XTVe siècle, ed. C. Kohler (Paris, 1909). Mardiros of Crimea, 'Liste rimée des souverains de la Petite Armenie', in RHCArm, voi. i. Michael the Syrian, 'Chronique', in RHCArm, voi. i. Nerses Balientz, short extract in L. M. Alishan, Sissouan ou l'Arméno-CiUcie (Venice, 1899). P- 469-

394

UNKNOWN CRUSADER CASTLES

Samuel of Ani, 'Extrait de la Chronographie de Samuel', in RHCArm, vol. i. Le trésor des chañes d'Armenie, ou cartulaire de la chancellerie royale des Roupéniens, ed. V. Langlois (Venice, 1863). Vahram of Edessa, 'Chronique rimée des rois de la Petite Armenie', in RCHArm, vol. i.

Muslim Sources Abu'1-Fida, 'Annales', in RHCOr, vol. i. Abu Shama, 'Le livre des deux jardins', in RHCOr, vols 4-5. al-'Ayni, 'Le collier de perles', in RHCOr, vol. 2. al-Makin Ibn al'Amid, Chronique des Ayyoubides (602-658/1205-6 -1259-60), trans. A.-M. Eddé and F. Micheau (Paris, 1994). al-Makrizi, 'Histoire d'Egypte', trans. E. Blochet, ROL, 9 (1902). al-Makrizi, Histoire des sultans mamlouks de I'Egypte, trans. M. E. Quatremère, 2 vols in 4 parts (Paris, 1845). Baha'-al-Din, 'Anecdotes et beaux traits de la vie de sultan Youssof', in RHCOr, vol. 3. 'Histoire des patriarches d'Alexandrie', extracts in J. F. Michaud, Bibliothèque des Croisades, vol. 4 (Paris, 1829). Ibn 'Abd al-Rahim, 'Vie de Kalavun', extracts in J. F. Michaud, Bibliothèque des Croisades, vol. 4 (Paris, 1829). Ibn al-Athir, 'Kamel Altevarykh', in RHCOr, vols 1-2. Ibn al-Furat, 'Selections from the Tarikh al-Duwal wa'1-Muluk', in Ayyubids, Mameluks and Crusaders, ed. and trans. U. and M. C. Lyons, with an introduction by J. S. C. Riley-Smith, 2 vols (Cambridge, 1971). Ibn Jubair, 'Extrait du voyage d'Ibn Djobeir', in RHCOr, vol. 3. Imad ad-Din al-Isfahani, Conquête de la Syne et de la Palestine par Saladin, trans. H. Masse (Paris, 1972). 'Izz al-Din Ibn Saddad, Description de la Syrie du Nord, trans. A-M Eddé-Terrasse (Damascus, 1984). Kamal-ad-Din, 'L'histoire d'Alep', trans. E. Blochet, ROL, 5 (1897).

SECONDARY SOURCES

Alishan, L. M., Sissouan ou l'Arméno-CiUcie (Venice, 1899). Andrews, K., Castles of the Morea (Princeton, 1953). Angold, M., A Byzantine Government in Exile: Government and Society under the Laskarids ofNicaea, 1204-1261 (London, 1975). Angold, M., The Byzantine Empire, 1025-1204 (London, 1984). Arbel, B., B.Hamilton and D.lacoby (eds), Latins and Greeks in the Easstern Mediterranean after 1204 (London, 1989). Ayalon, D., 'Studies on the Structure of the Mamluk Army', Bulletin of Oriental and African Studies, 15 (1950), pp. 203-28, 448-76; 16 (1954), pp. 57-90.

BIBLIOGRAPHY

395

Balard, M., 'L'activité commerciale en Chypre dans les années 1300', in P. Edbury (ed.), Crusade and Settlement (Cardiff, 1985), pp. 251-67. Balard, M., La Romanie génoise, Xlle- début du Xve siècle, 2 vols (Rome and Genoa, 1978). Balard, M., 'The Genoese in the Aegean, 1204-1566', in B. Arbel, B. Hamilton and D. Jacoby (eds), Latins and Greeks in the Eastern Mediterranean after 1204 (London, 1989), PP-158-74Barber, M., 'Supplying the Crusader States: The Role of the Templars', in B. Z. Kedar (ed.), The Horns ofHattin (London and Jerusalem, 1992), pp. 314-26. Barber, M., The New Knighthood: A History of the Order of the Temple. (Cambridge, 1994)Bartusis, M. C., The Late Byzantine Army: Arms and Society, 1204-1453 (Philadelphia, 1992). Benvenisti, M., The Crusaders in the Holy Land (Jerusalem, 1970). Bintìiff, J., 'The Prankish Countryside in Central Greece: The Evidence from Archaeological Field Survey', in P. Lock and G. D. R. Sanders (eds), The Archaeology of Medieval Greece (Oxford, 1996), pp. 1-18. Boas, A., Crusader Archaeology: The Material Culture of the Latin East (London, 1999). Boase, T. S. R., Kingdoms and Strongholds of the Crusaders (London, 1971). Boase, T. S. R., 'Military Architecture in the Crusader States', in HC, vol. 4 (1977), pp. 140-64. Boase, T. S. R., 'The Arts in Prankish Greece and Rhodes', in HC, vol. 4 (1977), pp. 208-50. Boase, T. S. R., 'The History of the Kingdom', in idem (ed.), The Cilician Kingdom of Armenia (Edinburgh, 1978), pp. 1-33. Bombaci, A., 'The Army of the Saljuqs of Rum', Istituto orientale di Napoli: annali, new series, 38 (1978), pp. 343-69Bon, A., 'Fortéresses medievales de la Grece centrale', Bulletin de correspondence hellénique, 61 (1937), pp. 136-208. Bon, A., La Marèe franque: recherches historiques, topographiques et archéologiques sur la principauté d'Achaie, 1204-1430, with a volume of plates and plans (Paris, 1969). Brand, C. M., Byzantium Confronts the West (Cambridge, Massachusetts, 1968). Braune, M., 'Die mittelalterlichen Befestigungen der Stadt Tortosa/Tartus', Damaszener Mitteilungen, 2 (1985), pp. 44-54. Burridge, P., 'The Castle of Vardounia and Defence in the Southern Taygetos', in P. Lock and G. D. R. Sanders (eds), The Archaeology of Medieval Greece (Oxford, 1996), pp. 19-^8. Gaben, C., La Syrie du Nord a l'époque des croisades et la principauté franque d'Antioche (Paris, 1940). Canard, M., 'Le royaume d'Arménie-Cilicie et les mamelouks jusqu'au traite de 1285', Revue des études arméniennes, 4 (1967), pp. 217-59. Chevedden, P., 'Crusader Warfare Revisited: A Revisionist Look at R. C. Smail's Thesis: Fortifications and Development of Defensive Planning', Journal of the Rocky Mountain Medieval and Renaissance Association, 14 (1993), pp. 13-31.

396

UNKNOWN

CRUSADER

CASTLES

Chevedden, P., L. Eigenbrod, V. Foley and W. Soedel, 'The Trebuchet', Scientific American, 273, no. i (July, 1995), pp.58-63. Contamine, P., War in the Middle Ages, trans. M.Jones (Oxford, 1984). Coulson, C., 'Cultural Realities and Reappraisals in English Castle-Study', Journal of Medieval History, 22 (1996), pp. 171-207. Coureas, N., The Latin Church in Cyprus, 1195-1312 (Aldershot, 1997). Der Nersessian, S., The Armenians (London, 1969). Der Nersessian, S., 'The Kingdom of Cilician Armenia', in HC, vol. 2 (1962), pp. 630-59. Deschamps, P., 'Les entrees des chateaux des croisés en Syrië et leurs defenses', Syria, 13 (1932), pp. 369-87Deschamps, P., 'Le chateau de Servantikar en Glide, le défilé de Marris et la frontière du comté d'Edesse', Syria, 18 (1937), pp. 379-88. Deschamps, P., Les chateaux des croisés en Terre Sainte, vol i, Le Crac des Chevaliers (Paris, 1934); vol. 2, La defense du royaume de Jerusalem (Paris, 1939); vol. 3, La defense du comté de Tripoli et de la principauté d'Antioche (Paris, 1973). Dunbar, J. G. and W. W. M. Boal, 'The Castle of Azgit', in T. S. R. Boase (ed.), The Cilician Kingdom of Armenia (Edinburgh, 1978), pp. 85-91. Dunbar, J. G. and W. W. M. Boal, 'The Castle of Vagha', Anatolian Studies, 14 (1964), pp. 175-84Edbury, P. W. (ed.), Crusade and Settlement (Cardiff, 1985). Edbury, P. W., 'The Crusading Policy of King Peter I of Cyprus, 1359-69', in P. Holt (ed.), The Eastern Mediterranean Lands in the Period of the Crusades (Warminster, i977)> PP- 90-105. Edbury, P. W., The Kingdom of Cyprus and the Crusades, 1191-1374 (Cambridge, 1991)Edbury, P. W., 'The Murder of King Peter I of Cyprus (1359-1369)', Journal of Medieval History, 6 (1980), pp. 219-33. Edwards, R. W., 'Bagras and Armenian Ciucia: A Reassessment', Revue des études arméniennes, 17 (1983), pp. 415-55. Edwards, R. W., 'The Crusader Donjon at Anavarza in Cilicia', Abstracts of the Tenth Annual Byzantine Studies Conference (Cincinnati, 1984), pp. 53-55. Edwards, R. W., The Fortifications of Cilican Armenia, Dumbarton Oaks Studies, 23 (Washington, DC, 1987). Ehrenkreutz, A. S., 'The Place of Saladin in the Naval History of the Mediterranean Sea in the Middle Ages', Journal of the American Oriental Society, 75 (1955), pp. 100-16. Ellenblum, R., Prankish Rural Settlement in the Latin Kingdom of Jerusalem (Cambridge, 1998). Ellenblum, R., 'Three Generations of Prankish Castle-Building in the Latin Kingdom of Jerusalem', in M. Baiard (ed.), Autour de la Première Croisade (Paris, 1996), pp. 517-71Enlart, C., L'ari gothiaue et de la Renaissance en Chypre, 2 vols (Paris, 1899). Fedden, R. and J. Thomson,' Crusader Castles (London, 1957).

BIBLIOGRAPHY

397

Folda, }., 'The Crusader Frescoes at Crac des Chevaliers and Marqab Castle', Dumbarton Oaks Papers, 36 (1982), pp. 177-210. Forey, A. J., 'The Military Order of St Thomas of Acre', EHR, 92 (1977), pp. 481-503. Forey, A. J., The Military Orders from the Twelfth to the Early Fourteenth Centuries (London, 1992). Forstreuter, K., Der deutsche Orden am Mittelmeer (Bonn, 1967). Foss, C., and D. Winfield, Byzantine Fortifications: An Introduction (Pretoria, 1986). Fotheringham, J. K., Marco Sañudo: Conqueror of the Archipelago (Oxford, 1915). France, J., Western Warfare in the Age of the Crusades, 1000—1300 (London, 1999). Frankel, R., 'The North-Western Corner of Crusader Acre', Israeli Exploration Journal, 37, part 4 (1987), pp. 256-61. Frech, F., 'Die armenischen Burgen', Zeitschrift der Gesellschaft für Erdkunde zu Berlin, 9 (1915), pp. 576-80. Gay, J., Le Pape Clement VI et les affaires d'Orient, 1342-1352 (Paris, 1904). Geanakoplos, D. J., 'Greco-Latin Relations on the Eve of the Byzantine Restoration: The Battle of Pelagonia, 1259', Dumbarton Oaks Papers, 7 (1953), pp. 101-41. Geanakoplos, D. J., Emperor Michael VIII Palaeologus and the West, 1258-82: A Study in Byzantine-Latin Relations (Cambridge, Massachusetts, 1959). Gertwagen, R., 'The Crusader Port of Acre: Layout and Problems of Maintenance', in M. Baiard (ed.), Autour de la Première Croisade (Paris, 1996), pp. 553-82. Gill, J., 'The Tribulations of the Greek Church in Cyprus, 1196-1280', Byzantinische Forschungen, 5 (1977). PP- 73~93Godfrey, J., 1204: The Unholy Crusade (Oxford, 1980). Gottwald, J., 'Burgen und Kirchen im mittleren Kilikien', BZ, 41 (1941), pp. 82-103. Gottwald, J., 'Die Burg Til im sudöstlichen Kilikien', BZ, 40 (1940), pp. 89—104. Gottwald, J., 'Die Kirche und das Schloss Paperon in Kilikisch-Armenien', BZ, 36 (1936), pp. 86-100. Gough, M., 'Anazarbus (Anavarza)', Anatolian Studies, 2 (1952), pp. 85-150. Grabois, A., 'La cité de Baniyas et le chateau de Subeibeh pendant les croisades', Cahiers de civilisation medievale, 13 (1970), pp. 43-62. Hamilton, B., The Latin Church in the Crusader States: The Secular Church (London, 1980). Harper, R. and D. Pringle, 'Belmont Castle: A Historical Notice and Preliminary Report of Excavation in 1986', Levant, 20 (1988), pp. 101-18. Heffening, W., 'Eine Burgruine im Taurus', Reportium für Kunstwissenschaft, 45 (1925), pp. 179-89Hellenkemper, H., Burgen der Kreuzritterzeit in der Grafschaft Edessa und im Königreich Kleinarmenien (Bonn, 1976). Hendrickx, B., 'A propos du nombre des troupes de la Quatrième Croisade et de l'empereur Baudoin I', Byzantina, 3 (1971), pp. 29-41. Hill, G., A History of Cyprus, 4 vols (Cambridge, 1940-52). Hillenbrand, C., The Crusades: Islamic Perspectives (Edinburgh, 1999). Holt, P., 'Baybars's Treaty with the Lady of Beirut in 667\i269', in P. Edbury (ed.), Crusade and Settlement (Cardiff, 1985), pp. 242-48.

398

U N K N O W N CRUSADER CASTLES

Holt, P., The Age of the Crusades: The Near East from the Eleventh Century to 1517 (London, 1986). Holt, P. (ed.), The Eastern Mediterranean Lands in the Period of the Crusades (Warminster, 1977). Holum, K. G., R. L. Hohlfelder, R. J. Bull and A. Raban, King Herod's Dream: Caesarea on the Sea (New Haven and London, 1988). Hubatsch, W., 'Der deutsche Orden und die Reichslehnschaft über Cypern', Nachrichten der Akademie der Wissenschaftim Göttingen. Philologisch-Historische Klasse (1955), PP- 245-306. Humphreys, R. S., From Saladin to the Mongols: The Ayyubids of Damascus, 1193-1260 (New York, 1973). Humphreys, R. S., 'The Emergence of the Mamluk Army', Studia Islamica, 45 (1977), pp. 67-99. I'Anson, E., 'Medieval and Other Buildings in the Island of Cyprus', Transactions of the Royal Institute of British Architects, session 1882-83 (London, 1883), pp. 13-33. Ilieva, A., 'The Suppression of the Templars in Cyprus According to the Chronicle of Leontios Makhairas', in M. Barber (ed.), The Military Orders: Fighting for the Faith and Caring for the Sick (Aldershot, 1994), p. 212-19. Imber, C., The Ottoman Empire, 1300-1481 (Istanbul, 1990). Inalcik, H., The Ottoman Empire: The Classical Age, 1300-1600, trans. N. Itzkowitz and C. Imber (London, 1973). Inalcik, H., 'The Ottoman Turks and the Crusades, 1329-1451' and 'The Ottoman Turks and the Crusades, 1451-1522', both in HC, vol. 6 (1989), pp. 222-75, 3U-53Irwin, R., 'The Mamluk Conquest of the County of Tripoli', in P. Edbury (ed.), Crusade and Settlement (Cardiff, 1985), pp. 246-49. Irwin, R., The Middle East in the Middle Ages: The Early Years of the Mamluk Sultanate, 1250-1382 (Beckenham, 1986). Jacoby, D., Studies on the Crusader States and on Venetian Expansion (Northampton, 1989), containing 'Crusader Acre in the Thirteenth Century: Urban Layout and Topography', ch. 5, pp. 1-45; 'Montmusard, Suburb of Crusader Acre: The First Stage of its Development', ch. 6, pp. 205-17; 'The Rise of a New Emporium in the Eastern Mediterranean: Famagusta in the Late Thirteenth Century', ch. 8, pp. 145-79Jacoby, D., 'The Encounter of Two Societies: Western Conquerors and Byzantines in the Peloponnese after the Fourth Crusade', American Historical Review, 78 (i973), PP- 873-906. Jackson, P., 'The Crisis in the Holy Land in 1260', EHR, 95 (1980), pp. 481-513. Jackson, P., 'The Crusades of 1239-41 and their Aftermath', Bulletin of the School of Oriental and African Studies, 50 (1987), pp. 32-60. Jackson, P., 'The End of Hohenstaufen Rule in Syria', Bulletin of the Institute of Historical Research, 59 (1986), pp. 20-36. Johns, C.N., A Guide to 'Atlit (Jerusalem, 1947). Johns, C. N., 'Excavations at Pilgrims' Castle ('Atlit): The Faubourg and its Defences', QDAP, i (1932), pp. 111-29; 'The South-Eastern Cemetary', QDAP, 2 (1932-33),

BIBLIOGRAPHY

399

pp. 41-104; 'The Ancient Tell and the Outer Defences of the Castle', QDAP, 3 (1933-34), PP-145-64; 'An Unfinished Church in the Suburbs', QDAP, 4 (1934-35), pp. 122-37; 'Stables at the South West of the Suburbs', QDAP, 5 (1935-36), pp. 3160. C. N. Johns, Pilgrims' Castle CAtlit), David's Tower (Jerusalem) and Qal'at ar-Rabad ('Ajlun), ed. D. Pringle with a foreword by A. Johns (Aldershot, 1997). Collected studies containing all other listed works by C. N. Johns. Johns, C. N., 'The Citadel, Jerusalem: A Summary of Work since 1934', QDAP, 14 (1950), pp. 121-90. Kalayan, H., 'The Sea Castle of Sidon', Bulletin du Musée de Beyrouth, 26 (1973), pp. 81-89. Kedar, B. Z., 'The Subjected Muslims of the Prankish Levant', in J. M. Powell (ed.), Muslims under Latin Rule, 1100-1300 (Princeton, 1990). Kennedy, H., Crusader Castles (Cambridge, 1994). Kesten, A., The Old City of Acre: Re-Examination Report, 1993 (Acre, 1993). Lawrence, A. W., 'The Castle of Baghras', in T. S. R. Boase (ed.), The Cuidan Kingdom of Armenia (Edinburgh, 1978), pp. 34-84. Lawrence, T. E., Crusader Castles, ed. D. Pringle (Oxford, 1988). Lock, P., 'Castles and Seigneurial Influence in Latin Greece', in A. V. Murray (ed.), From Clermont to Jerusalem: The Crusades and Crusader Societies, 1095-1500 (Turnhout, 1998), pp. 173-85. Lock, P., 'The Prankish Tower on the Acropolis, Athens: The Photographs of William J. Stillman', ABSA, 82 (1987), pp. 131-33. Lock, P., 'The Prankish Towers of Central Greece', ABSA, 81 (1986), pp. 101-23. Lock, P., The Franks in the Aegean, 1204-1500 (London, 1995). Lock, P., 'The Medieval Towers of Prankish Greece: A Problem in Chronology and Function', in B. Arbel, B. Hamilton and D. Jacoby (eds), Latins and Greeks in the Eastern Mediterranean after 1204 (London, 1989), pp. 129-45. Lock, P., 'The Towers of Euboea: Lombard or Venetian; Agrarian or Strategic', in P. Lock and G. D. R. Sanders (eds), The Archaeology of Medieval Greece (Oxford, 1996). PP-107-26. Longnon, J., L'empire latin de Constantinople et la principauté de Marèe (Paris, 1949). Longnon, J., Les compagnons de Villehardouin: recherches sur les croisés de la Quatrième Croisade (Geneva and Paris, 1978). Longnon, J., 'Problèmes de Fhistoire de la principauté de Morée', Journal des Savants (1946), pp. 77-93, 147-61. Longnon, J., 'The Prankish States in Greece, 1204-1311', in HC, vol. 2 (1962), pp. 235-74. Luttrell, A., Latin Greece, the Hospitallers and the Crusades, 1291-1440 (Aldershot, 1982), containing 'La corona de Aragon y la Grecia catalana, 1379-1394', eh. 11, pp. 219-52; 'The Crusades in the Fourteenth Century", eh. 16, pp. 122-54; 'The Hospitallers of Rhodes: Perspectives, Problems and Possibilities', eh. i, pp. 243-66. Luttrell, A., The Hospitallers in Cyprus, Rhodes, Greece and the West, 1291-1440

40O

UNKNOWN CRUSADER CASTLES

(Aldershot, 1979), containing 'The Hospitallers at Rhodes, 1306-1421', eh. i, pp. 278-313; 'The Hospitallers in Cyprus after 1291', ch. 2, pp. 161-71; 'Venice and the Knights Hospitallers of Rhodes in the Fourteenth Century', ch. 5, pp. 195-212; 'The Principality of Achaea in 1377', ch. 22, pp. 340-45; 'Intrigue, Schism and Violence among Hospitallers of Rhodes, 1377-1384', ch. 23, pp. 30-48. Luttrell, A., 'The Hospitallers' Interventions in Cilician Armenia, 1291—1375', in T. S. R. Boase (ed.), The Cilician Kingdom of Armenia (Edinburgh, 1978), pp. 118-44. Luttrell, A., 'The Hospitallers at Rhodes, 1306-1421', in HC, vol. 3 (1975), pp. 278-313. Luttrell, A., The Hospitallers of Rhodes and their Mediterranean World (Aldershot, 1992), containing 'The Hospitallers of Rhodes Confront the Turks, 1306-1421', ch. 2, pp. 80-116; 'Settlement on Rhodes, 1306-1366', ch. 5, pp. 273-81; 'Lindos and the Defence of Rhodes, 1306-1522', ch. 7, pp. 317-32; 'The Hospitallers in Cyprus, 1310-1378', ch. 9, pp. 155-84; 'The Military and Naval Organization of the Hospitallers at Rhodes, 1310-1444', ch. 19, pp. 133-53. Luttrell, A., 'The Latins and Life on the Smaller Aegean Islands, 1204—1453', in B. Arbel, B. Hamilton and D. Jacoby (eds), Latins and Greeks in the Eastern Mediterranean after 1204 (London, 1989), pp. 146-57. Lyons, M. C. and D. E. P. Jackson, Saladin: The Politics of Holy War (Cambridge, 1982). Markowski, R., 'Richard Lionheart: Bad King, Bad Crusader?', Journal of Medieval History, 23 (1997), pp. 351-65. Marshall, C., 'The French Regiment in the Latin East, 1254-1291', Journal of Medieval History, 15 (1989), pp. 301-7Marshall, C., Warfare in the Latin East, 1192-1291 (Cambridge, 1992). Mas Latrie, L. de, Histoire de l'ile de Chypre sous le regne des princes de la maison de Lusignan, 3 vols (Paris, 1852-61). Masterman, E. W. G., 'A Crusaders' Fortress in Palestine', PEQ (1928), pp. 91-97. Mayer, H. E., 'Latins, Muslims and Greeks in the Latin Kingdom of Jerusalem', History, 63 (1978), pp. 175-92. Megaw, A. H. S., 'A Castle in Cyprus Attributable to the Hospital?', in M. Barber (ed.), The Military Orders: Fighting for the Faith and Caring for the Sick (Aldershot, 1994), PP- 42-51Megaw, A. H. S., 'Excavations at Sandra Kolones, Paphos: Preliminary Report on the 1966-67 and 1970-71 Seasons', in Report of the Department of Antiquities, Cyprus, 1971 (Nicosia, 1971), pp. 117-46. Megaw, A. H. S., 'Supplementary Excavations on a Castle Site at Paphos, Cyprus, 1970-71', Dumbarton Oaks Papers, 26 (1972), pp. 322-45. Megaw, A. H. S., 'The Arts in Cyprus. B: Military Architecture', in HC, vol. 4 (1977), pp. 196-207. Mesnil du Buisson, R. du, 'Les anciennes defenses de Beyrouth', Syria, 2 (1921), PP- 235-57» 317-27Miller, W., Essays on the Latin Orient (Cambridge, 1921). Miller, W., The Latins in the Levant: A History of Prankish Greece, 1204-1566 (London, 1908).

BIBLIOGRAPHY

401

Molin, K., 'Fortifications and Internal Security in the Kingdom of Cyprus, 1191—1426', in A. V. Murray (ed.), From Clermont to Jerusalem: The Crusades and Crusader Societies, 1095-1500 (Turnhout, 1998), pp. 187-99. Molin, K., 'The Non-Military Functions of Crusader Fortifications, iiSj-c. 1380', Journal of Medieval History, vol. 23, no. 4 (1997), pp. 367-88. Morgan, D., The Mongols (Oxford, 1986). Müller-Wiener, W., Castks of the Crusaders, trans. ]. Brownjohn (London, 1966). Murray, A. V. (ed.), From Clermont to Jerusalem. The Crusades and Crusader Societies, 1095-1500 (Turnhout, 1998). Nicol, D.M., The Despotate of Epirus (Oxford, 1957). Nicol, D.M., The Last Centuries of Byzantium, 1261—1453 (London, 1972). Pattenden, P., "The Byzantine Early Warning System', Byzantion, 53 (1983), pp. 258-99. Powell, }. M., Anatomy of a Crusade, 1213-1221 (Philadelphia, 1986). Prawer, }., Crusader Institutions (Oxford, 1980). Prawer, }., The Latin Kingdom of Jerusalem (London, 1972). Pringle, D., 'A Castle in the Sand: Mottes in the Crusader East', Cháteau-Gaillard, 18 (Caen, 1998), pp. 187-91. Pringle, D., 'Aqua Bella: The Interpretation of a Crusader Courtyard Building', in B. Z. Kedar (ed.), The Horns of Hattin (London and Jerusalem, 1992), pp. 146-67. Pringle, D., 'A Thirteenth-Century Hall at Montfort Castle in Western Galilee', Antiquaries Journal, 66 (1986), pp. 52-81. Pringle, D., 'Crusader Castles: The First Generation', Fortress, i (1989), pp. 14-25. Pringle, D., 'King Richard I and the Walls of Ascalon', PEQ, 116 (1984), pp. 133-47. Pringle, D., 'Magna Mahumeria (al-Bira): The Archaeology of a Prankish New Town in Palestine', in P. Edbury (ed.), Crusade and Settlement (Cardiff, 1985), pp. 147-68. Pringle, D., A. Petersen, M. Dow and C. Singer, 'Qal'at Jiddin: A Castle of the Crusader and Ottoman Periods in Galilee', Levant, 26 (1994), pp. 135-66. Pringle, D., 'Review Article: Reconstructing the Castle of Safad', PEQ, 117 (1985), pp. 139-49Pringle, D., Secular Buildings in the Crusader Kingdom of Jerusalem: An Archaeological Gazetteer (Cambridge, 1997). Pringle, D., 'Survey of Castles in the Crusader Kingdom of Jerusalem, 1989: Preliminary Report', Levant, 23 (1991), pp. 87-91. Pringle, D., 'Templar Castles between Jaffa and Jerusalem', in H. Nicholson (ed.), The Military Orders 2: Welfare and Warfare (Aldershot, 1998), pp. 89-109. Pringle, D., 'Templar Castles on the Road to the Jordan', in M. Barber (ed.), The Military Orders: Fighting for the Faith and Caring for the Sick (Aldershot, 1994), pp. 148-66. Pringle, D., The Churches of the Crusader Kingdom of Jerusalem: A Corpus, vol i, A-K (Excluding Acre and Jerusalem) (Cambridge, 1993); vol. 2, L-Z (Excluding Tyre) (Cambridge, 1998). Pringle, D., The Red Tower (London, 1986).

4O2

UNKNOWN CRUSADER CASTLES

Pringle, D., 'Towers in Crusader Palestine', Chàteau-Gaillard, 16 (Caen, 1994), PP- 335-50. Pringle, D., 'Town Defences in the Crusader Kingdom of Jerusalem', in I. Corns and M. Wolfe (eds), The Medieval City under Siege (Woodbridge, 1995), pp. 69-121. Pryor, J. H., Geography, Technology and War: Studies in the Maritime History of the Mediterranean, 649-1571 (Cambridge, 1988). Pryor, J. H., 'The Naval Battles of Roger of Lauria', in J. H. Pryor, Commerce, Shipping and Naval Warfare in the Medieval Mediterranean (London, 1987), ch. 6, pp. 179-216. Queller, D. E., and T. F. Madden, The Fourth Crusade: The Conquest of Constantinople, second edition (Philadelphia, 1997). Razi, Z. and E. Braun, 'The Lost Crusader Castle of Tiberias', in B. Z. Kedar (ed.), The Horns ofHattin (London and Jerusalem, 1992), pp. 216-27. Rey, E. G., 'Etude sur la topographic de la ville d'Acre au XlIIe siècle', Mémoires de la société nationale des antiquaries de France, 39 (1878), pp. 115—45. Rey, E. G., Etude sur les monuments de l'architecture militaire des croisés en Syrië et dans l'ile de Chypre (Paris, 1871). Richard, J., 'La revolution de 1369 dans le royaume de Chypre', Bibliothèque de l'Ecole des Chartes, 110 (1952), pp. 108-23. Richard, J., 'Les révoltes chypriotes de 1191-1192 et les inféodations de Guy de Lusignan', in B. Z. Kedar, J. Riley-Smith and R. Hiestand (eds), Montjoie: Studies in Crusader History in Honour of Hans Eberhard Mayer (Aldershot, 1997), pp. 123-28. Richard, J., Saint Louis: Crusader King of France, ed. S. Lloyd and trans. J. Birrel (Cambridge, 1992). Richard, J., 'The Institutions of the Kingdom of Cyprus', in HC, vol. 6 (1989), pp. 150-74. Riley-Smith, J. S. C., The Feudal Nobility and the Kingdom of Jerusalem, 1174-1277 (London, 1973). Riley-Smith, J. S. C., The Knights of St John in Jerusalem and Cyprus, c. 1050-1310 (London, 1967). Riley-Smith, J. S. C., 'The Templars and the Teutonic Knights in Cilician Armenia', in T. S. R. Boase (ed.), The Cilician Kingdom of Armenia (Edinburgh, 1978), pp. 92-117. Robinson, F. C. R. and P. C. Hughes, 'Lampron: Castle of Armenian Ciucia', Anatolian Studies, 19 (1969), pp. 183-207. Roll, I., 'Medieval Apollonia-Arsuf: A Fortified Coastal Town in the Levant of the Early Muslim and Crusader Periods', in M. Baiard (ed.), Autour de la Première Croisade (Paris, 1996), pp. 595-606. Rubio y Lluch, A., 'Els Castells Catalans de la Grècia continental', Annuari de ITnstitut d'estudis Catalans, 2 (1908), pp. 364-425. Rüdt de Collenberg, W., Families de l'Orient latin, Xlle-XIVe siecles (London, 1983), containing 'L'empereur Isaac de Chypre et sa filié, 1155-1207', ch. i, pp. 123-77. Runciman, S., Mistra (London, 1980). Saadé, G., 'Histoire du chateau de Saladin', Studi Medievali, series 3, 9 (1968), pp. 980-1016.

BIBLIOGRAPHY

403

Setton, K.M. (general ed.), A History of the Crusades, 6 vols (Madison, 1955-89). Setton, K.M., Catalan Domination of Athens, 1311-1388 (Cambridge, Massachusetts, 1948). Setton, K. M., 'The Catalans in Greece, 1311-1380', and 'The Catalans and Florentines in Greece, 1380-1462', both in HC, vol. 3 (1975), pp. 167-224, 225-77. Setton, K. M., The Papacy and the Levant, 1204-1571, vol. i, The Thirteenth and Fourteenth Centuries (Philadelphia, 1976). Sigonowitz, B., 'Zur Eroberung Thessalonikes im Herbst 1224', BZ, 45 (1952), p. 28. Sinclair, T. A., Eastern Turkey: An Architectural and Archaeological Survey, 4 vols (London, 1990). Smail, R. C., 'Crusaders' Castles of the Twelfth Century', Cambridge Historical Journal, 10 (1951), pp. 133-49Smail, R. C., Crusading Warfare, 1097-1193, second edition with an introduction by C. Marshall (Cambridge, 1995). Sterns, I., 'The Teutonic Knights in the Crusader States', in HC, vol. 5 (1985), pp. 315-78. Thomson, J., 'Castles in Cilicia', Geographical Magazine, 23 (April, 1951), pp. 569-77. Thorau, P., The Lion of Egypt: Sultan Baybars I and the Near East in the Thirteenth Century, trans. P. M. Holt (London, 1992). Tibbie, S., Monarchy and Lordships in the Latin Kingdom of Jerusalem, 1099-1291 (Oxford, 1989). Topping, P., 'The Morea, 1311-1364' and 'The Morea, 1364-1460', both in HC, vol. 3 (i975). PP-104-40, 141-66. Traquair, R., 'Prankish Architecture in Greece', Journal of the Royal Institute of British Architects, 31 (1923-24), pp. 33-48, 73-86. Traquair, R., 'Laconia, I: Mediaeval Fortresses', ABSA, 12 (1905-6), pp. 259-76. Traquair, R., 'Mediaeval Fortresses of the North-Western Peloponnesus', ABSA, 13 (1906—7), pp. 268-81. Upton-Ward, J., 'The Surrender of Gaston and the Rule of the Templars', in M. Barber (ed.), The Military Orders: Fighting for the Faith and Caring for the Sick (Aldershot, 1994), pp. 179-88. Van Milh'ngen, A., Byzantine Constantinople: The Walls of the City and Adjoining Sites (London, 1899). Vickers, M., 'The Byzantine Sea Walls of Thessaloniki', Balkan Studies, 2 (1970), pp. 261-78. Wartburg, M-L von, 'The Medieval Cane Sugar Industry in Cyprus: Results of Recent Excavations', Antiquaries Journal (1983), pp. 298-314. Wolff, R. L., 'Mortgage and Redemption of an Emperor's Son: Castile and the Latin Empire of Constantinople', Speculum, 29 (1954), pp. 45-84. Wolff, R. L, 'Politics in the Latin Patriarchate of Constantinople, 1204-1261', Dumbarton Oaks Papers, 8 (1954), pp. 225-303. Wolff, R. L., 'The Latin Empire of Constantinople, 1204-1261', in HC, vol. 2 (1962), pp. 187-233. Youngs, G. R., 'Three Cilician Castles', Anatolian Studies, 15 (1965), pp. 113—34.

This page intentionally left blank

Index

Acciajuoli lords of Corinth and Athens 203, 211, 221, 246. See also Niccolo Acciajuoli Accursed Tower (Acre) 83 Achaia, principality of, 192, 193-94, 197-99, 201, 202, 203, 208, 212, 214, 215, 217, 223, 224, 227, 229, 230, 232, 236, 244, 246, 247, 250, 253, 254, 255, 257-58, 259-60, 262-63, 266, 267, 268, 271, 273, 275, 278, 280-81, 282, 284, 289, 304, 305, 306 Acre xiv, xvi, xxi, 3, 4, 6, 7, 9, 10, 11, 23, 24, 26, 30, 32, 34, 35, 39, 42, 43, 44, 45, 46, 47, 49, 51, 52, 54, 55, 57, 58, 59, 60, 61, 62, 64, 65, 66, 67, 69, 70, 73, 74, 77, 78, 79, 83, 84, 86, 89, 90, 107, 108, 112, 115, 118, 119, 124, 125, 127, 128, 130, 133, 150, 157, 162, 165, 178, 188, 225, 229, 260, 265, 273, 274, 276, 277, 278, 281, 282, 283, 284, 286, 290, 291-92, 293, 296, 299, 301, 305, 306, 313 n.47, 324 n. 23; description 23, 24-25, 32, 61, 319 n. 74, 324 n. 23, 326 n. 59, 334 n.95; 1189-91 siege 6, 66, 67, 76, 278; 1291 siege 8, 10, 23, 32, 34-35, 51, 73, 94, 125, 312 n.40 acropolis (Athens) 219, 251-52, 276; description 211, 221, 272 Adana 160, 162, 163, 168; description 162, 163 al-Adil, sultan of Egypt, 4, 18, 54, 64 Adrianople 206, 207, 233, 236, 237, 241, 242, 244, 258

Aetos, description 220-21 Agridi, battle (1232) 91, 94, 111-12, 133 Aimery Barlais, Cypriot baiili of Frederick II, 111, 112 Aimery of Lusignan, king of Cyprus, 102, 115, 132, 175 Ak Kale 158 Akkar (Gibelcar), 24, 36, 48, 49-50, 54, 60, 64, 72, 76, 81, 86, 161, 336 n. 46; description 19 Akova (Mathegriffon) 212, 219, 223, 236, 237; description 213 Alanya 138 Aleppo xiv, xvi, xvii, 4, 38, 68, 75, 139, 140, 141, 156, 183-84, 185 Alexander IV, pope, 18, 81 Alexandretta 152, 180, 186 Alexandria xiv, 77, 127; crusade against (1365) 120, 127-28 Alexis, Cypriot rebel, 110 Aliberi-Matsoukela (Euboea) 232, 292 Alice, niece of Leon II, 180 Almogavers 199 Amanus Gates 137, 138, 141, 143, 145, 152-53, 160, 161, 175, 176, 177, 178 Amaury of Tyre, 'governor' of Cyprus, 91, 103, 112-13, 114, 116, 128, 130, 188, 277, 339 n. 32 Amouda (Amuda) 80, 154, 155, 156, 158, 159, 161, 170-71, 177-78, 186, 187, 307; description 161, 175-76, 177 Amphikleia: see Dadi Amphissa: see Salona

406

UNKNOWN CRUSADER

Anamur 128, 138, 188 Anavarza (Anazarbus) 140, 158-59, 160, 168, 172, 219; description 148, 160, 162, 352 n. 36 Anazarbus: see Anavarza Anceau de Brie, Ibelin supporter, 111, 114 Andravida (Andreville) 247, 260, 280, 282 Andreville: see Andravida Andrew II, king of Hungary 10, 273 Andronikos Komnenos, Byzantine commander, 163 Andronikos II Palaiologos, Byzantine emperor, 192, 201, 246-47, 255, 291 Androusa 210, 217, 219, 224, 257, 280; description 203-4 Anna Palaiologina, ruler of Epiros, 244 Antalya: see Satalia. Antioch xiv, xvi, 3, 13, 36, 39, 42, 46, 60, 68, 80, 85, 137, 143, 152, 155, 163, 169, 172, 180, 181, 182, 186, 187, 260, 281, 294, 296, 323 n. 5, 374 n. 22; description 24-25, 163; 1268 siege 24, 42, 43, 47, 50, 84; succession dispute (1201-1220s) 60, 68, 180-81, 182-85, 186 Antioch on the Meander, battle of (1211), 194 Antiochetta 188 Apalire (Naxos) 231; description 222 Apano-Castro (Naxos) 231 Appalonia 205 Aragonese crusade (1269) 11, 69 Aragonese dukes of Athens 193, 224, 257, 259 Arcadia 106, 211, 212, 216, 217, 219, 280; description 210 Arcadiopolis 206, 243-44 Areas 284 Archipelago, duchy of: see Naxos archons 197, 253 Argos 217, 218, 365 n. 39; description 210-11

CASTLES

al-Arimah 38-39, 49-50, 80; description 14-16 armies; size, strength and composition: Armenians 141-44; Ayyubids 4-6; Bulgars and Cumans 194-96; Catalans 199; Cypriot kings 94; Genoese invaders of Cyprus 93-94; Greek states after Fourth Crusade 194-99; Khwarizmians 3; Latins in Greece 194-99; Latins in the Holy Land 10-11; Mamluks 7-9, 41-42, 93, 141; Mongols 41, 140; Seljuk Turks 140, 141 Arsuf 19, 26, 45, 67, 81, 84, 279, 305; description 20-22 Arsuz: see Port Bonnel Arta 191, 244, 245 Ascalon 3, 19, 26, 27, 31, 40, 43, 45, 46, 53, 65, 66, 67, 73, 74, 77, 81, 82, 122, 305; description 20-22, 316 n.43 al-Ashraf Khalil, sultan of Egypt, 90, 125 Assassins 71, 75 Assizes of Romania 246, 253, 254, 280, 285, 288 Athens, city and duchy of, 191, 192-93, 196, 198-99, 201, 208, 210, 219, 221, 223, 224, 232, 233, 239, 245, 246, 247-48, 252, 257, 259, 261, 263, 267, 272, 273, 302, 304. See also acropolis (Athens) 'Atlit (Chastel Pelerin/Pilgrims' Castle) 4, 13, 19, 22, 28, 30, 31, 32, 33, 35, 44, 45^6, 47, 52-53, 54, 55, 56, 57, 59-60, 66, 70, 72, 74, 77, 78, 79, 80, 83, 86, 107, 130, 155, 163, 225, 272, 273, 277, 280, 283, 284, 285, 293, 295, 296-97, 298, 299, 300, 304, 306, 307; description 16, 28, 273, 297; 1220 siege 33, 34, 39, 40, 44, 46, 55, 76 auberge (Acre) 274

INDEX

Ayas 156, 157, 158, 160, 165-66, 168, 169, 170, 292; description 165 Ayyubids xvii, 4, 7, 26, 30, 35, 137 Azgit 148 Baghras (Gaston) 68, 152, 186, 187, 285, 306, 307; description 182-85 baillis of Frederick II on Cyprus, 91, 111-12, 132 Baldwin II, king of Jerusalem, 281 Baldwin I, Latin emperor, 194, 196, 206, 207, 224, 241, 242, 258 Baldwin II, Latin emperor, 229, 363 n. 18 Baldwin, son of Henry of Silifke, 170, 180 Balian of Ibelin, lord of Arsuf, 81 Banyas 49, 65, 69, 72 Banyas (Valania) near Margat, 81, 295-96, 297 Bardi, Florentine bankers, 166 Bar Hebraeus, chronicler, 165 Bartholomew de Ravendel, lord of Maraclea, 22, 65, 70-71, 86 Bartolomeo Ghisi, lord of Tinos and Mikonos, 200, 231 Baybars, sultan of Egypt, 4, 6, 7, 8-9, 11, 17, 18, 21, 22, 34, 36, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 52, 53, 54, 55, 57, 58, 64, 70, 71, 72, 77, 81, 82, 84, 93, 116, 125, 128, 138, 141, 152, 153, 155-56, 157, 177, 187, 240, 271, 278, 279, 281, 287, 288-89, 296, 307, 324 n. 23, 327 n. 79-80, 348 n. 15 Beaufort (Leutron) (Achaia) 217, 219, 239 Beaufort (Holy Land) 3, 28, 39, 40, 48, 49, 50, 65, 69, 74, 79, 80, 84, 273, 285, 297, 305; description 18, 297; 1268 siege 9, 34, 48 Beaufort (smaller Achaian fortress near Karytaina) 255-56, 258 Beirut 3, 23, 26, 31, 46, 49, 52, 74,

407

79, 85, 86, 272, 277, 281, 315 n. 34; description 19-20, 28, 272; imperialist siege (1231-32) 10, 28, 44-45, 46, 76, 77, 78, 91, 102 Belén Pass (Syrian Gates) 143, 152-53, 160, 182, 185 Belmont 81, 297 Belvoir 3, 19, 20, 36, 38, 49, 53, 81, 106, 126, 272, 286, 297, 300, 303; description 17, 297 Bertrand II, Embriaco lord of Gibelet, 61 Bethgibelin 82, 122 Bethlehem 3 Bikisrail 68, 181 al-Bira 142 Bizöe 243-^4 Blachernae palace (Constantinople) 206, 363 n. 18 Black Tower (Hasanbeyli(?)), 176, 178, 294, 307 Blanchegarde 122 Bodrum (Cilician Armenia) 148, 158 Bodrum (Turkey) 265, 266 Boas, A, archaeologist, xviii Bohemond I, ruler of Antioch, 373-74 n. 22 Bohemond III, ruler of Antioch/Tripoli 60, 68, 169, 180, 354 n. 9 Bohemond IV, ruler of Antioch/Tripoli, 60, 62, 68, 80, 172, 180-81, 182, 183-85, 187 Bohemond V, ruler of Antioch/Tripoli, 71 Bohemond VI, ruler of Antioch/Tripoli 36, 65, 68-69, 82, 86, 155, 331-32 n. 31, 336 n. 46 Bohemond VII, ruler of Antioch/Tripoli, 61, 62, 77, 113, 277, 285 Bohemond, the castle of (Constantinople), 240 Bokebais (Abu Qubais) 71

408

UNKNOWN

CRUSADER

Bon, A., archaeologist, 203-4, 215-17, 218, 219, 222 Boniface VIII, pope, 130 Boniface of Montferrat, king of Thessaloniki, 197, 207, 234, 236, 245, 246, 257, 258, 263 Boniface, traitor at St George (Achaia), 258 Boudonitza, 198, 204, 219, 221, 223, 246, 249, 369 n. 107; description 212, 213 bourg, see military architecture Brindisi 228, 245, 289, 291, 305 Buffavento 95, 98, 106, 114, 116, 119, 122, 130, 148, 161, 251, 277-78; description 100-1, 105; 1232 siege 111-12 Bulgars 89, 98, 120, 191, 194-96, 197, 201, 206, 227, 228, 236, 237, 239, 240, 241, 242, 243-44, 250, 251, 257, 284, 301, 302 Burchard of Mount Sion, pilgrim, 283 Byzantine empire xiii, 89-90, 137, 160, 168, 191, 192-93, 194, 233-34, 253, 300 Caco: see Qaqun Caesarea (Holy Land) 26, 29, 31, 34, 39, 45, 46, 54, 65, 67, 82, 83, 85, 225, 275, 282; description 20-22, 27-28, 326 n. 57 Caesarea (Kayseri) (Asia Minor) 138, 140, 152, 153 Cairo xiv, xvi, xvii, 3, 34, 90, 144 Camardias: see Goumardias. Canamella (al-Tina/al-Tinat) 180, 186, 188 Cannaqui, Greek pirate, 115 Carpamos 265 carrier pigeons 49 Casal des Plains 66, 78, 83 Casal Imbert 77 Castellorizzo 251, 265, 284

CASTLES

Castellum Novum (Cilician Armenia): see Norpert Castiello Nuevo 262, 268 Castle of Roger the Lombard, 54; description 28-29 Castle of the Snakes: see Yuan castrum: see military architecture Catalans (and the Catalan company) 41, 192-93, 194, 199, 201, 212, 213, 215, 221, 223, 224, 231, 232, 233-34, 237, 238, 239, 245, 246-47, 249, 251, 256, 257, 259, 263, 264, 289, 302, 305 Catalonia 166 catapults: see siege weapons Catholicos (head of Armenian church), 155, 170, 278, 296 Cave de Tyron 3, 49, 79, 285; description 19 Caymont 83 Cephalonia 191, 244, 245, 250 Cephissus, battle of (1311): see Halmyros Ceyhan, river, 156-57, 161, 178 Charax 208 Charles I of Anjou, king of Sicily, suzerain (later prince) of Achaia and claimant to the kingdom of Jerusalem, 9, 60, 131, 193, 197, 227, 231, 237, 289 Charles II of Anjou, king of Naples and prince of Achaia, 244-45 Chastel Blanc, 29, 38-40, 49-50, 52, 53, 64, 69, 75, 80, 81, 82, 84-85, 150, 297, 298; description 14-16, 297; 1271 siege 16, 48, 75 Chastelez ( Jacob's Ford) 38 Chastel Pelerin: see 'Atlit Chastel Rouge, 49, 81, 106; description 29, 30 Chäteauneuf (Achaia) 280-81 Chàteauneuf (Hunin) (Holy Land) 3, 38, 49 Chilvan Kale: see Roche Roussel

INDEX

409

279, 295, 300, 302, 305; description 205-6; 1203 and 1204 sieges 206, 240-41; 1235 and 1236 sieges 197, 200-1, 228, 229, 242; Greek recapture (1261) 229 Corfu 191, 200, 234, 284 Corinth 106, 122, 194, 197, 199, 204, 212, 213, 217, 218, 219, 232, 240-41, 245-46, 246, 251, 255, 257, 274, 289, 302, 365 n. 39; description 210-11 Coron 191, 229, 233, 249, 254, 262, 291; description 210-11 Corycus 128, 129, 142, 158, 159, 165, 171, 172, 173, 233, 292, 347 n. 10; description 165 Cos 251 cost of warfare and fortifications 10-11, 35, 79-80, 84, 85-86, 107, 223-25 Crac des Chevaliers xviii, 19, 24, 28, 30, 31, 32, 34, 35, 38-39, 40, 49-50, 52, 53, 59, 64, 68, 69, 71, 72, 75-76, 78, 79, 80, 81, 84-85, 107, 121, 123, 128, 130, 150, 179, 181, 182, 204, 225, 237, 262, 272, 273, 285, 286, 287, 288-89, 290, 296, 297, 298, 299, 300, 303, 306, 307; description 13-16, 273, 286; 1271 siege 9, 16, 36, 289 Crete 109, 191, 200, 222, 234, 251, 254, 255, 361-62 n. 34 Constantine, Armenian Catholicos, 278 Crevecoeur 255 Crucemont 239 Constantine, Hethoumid leader and father of Hethoum I, 163-64, Cumans 191, 194-96, 201, 236, 239, 242, 244, 250, 284, 301 171-72, 173, 182 Cumbethefort 176 Constantine, lord of Gabán, 172 Constantine, son of Henry of Silifke, Cursat 296 170, 180 Cyzicus 212, 227, 229, 230, 233, 239; description 208 Constantinople xiii, xvii, 36, 37, 90, 137, 152, 160, 168, 191, 192, 193, 194, 196, 197, 200, 207, 224, 227, Dadi (Amphikleia) 220, 221 230, 231, 233, 236, 239, 240, 242, Damascus xiv, xvi, xvii, 3, 4, 6, 9, 20, 26, 244, 246, 253, 256, 257, 258, 260, 33, 38, 55, 70, 71, 80, 140, 279, 281

Chios 191, 200, 201, 221, 234, 255, 257, 375 n. 30 Chlemoutsi 35, 222, 224, 257, 260, 262, 268, 271-72, 273, 274, 279, 282; description 213, 214 Christopoli (Kavala) 206, 227, 233, 236, 247, 259 Chronicle of Morea, 197, 198, 199, 203-4, 211, 214, 215-19, 222, 228, 236, 238, 249, 255, 262, 279, 280, 367 n. 74 Cilician Gates 137, 138, 147, 152, 160 Civetot 208 Clarence: see Glarentza Clement V, pope, 133 Clermont: see Chlemoutsi Goliath 30, 52, 54, 55, 56, 59, 64, 81, 107, 303, 307; description 28, 29 Conrad, son of Frederick II and king of Jerusalem, 9 Conrad of Hildesheim, imperial chancellor, 175 Conrad of Mainz, papal legate, 175 Conrad of Montferrat, king elect of Jerusalem, 89 Conradin, king of Jerusalem, 9 Constance of Aragon, queen of Cyprus, 173 Constantine IV (Guy of Lusignan), king of Cilician Armenia, 173 Constantine VI, king of Cilician Armenia, 173

41O

UNKNOWN CRUSADER

Damietta 124, 125, 127 Darbsak (Trapesac) 68, 73, 139, 182, 183, 185, 348 n. 15, 359 n. 73 Darkoush 68 Darum 38 De constructione castri Saphet 16-17, 41, 48, 286, 295 Demetrius, king of Thessaloniki, 258-59 Demotika 206, 227, 236, 237 Deschamps, P., archaeologist, 50, 51 Destroit 34, 285, 293, 295 Doc 54-55, 64, 82 earthquakes 13-14, 126, 147, 164, 321 n. 108, 349 n. 4 Edessa, xiv, xvi, 168 Edmund, brother of Edward I of England, 125 Edward I, king of England, crusade of (1271-72), 10, 67, 69, 77, 124, 125 Edwards, R., archaeologist, xix-xx, 157-58 Eleanor of Aragon, queen of Cyprus, 93, 98, 120, 123 Ellenblum, R, archaeologist, xviii-xx, 53, 56, 58 Embriaco lords of Gibelet 3, 23, 60-62, 277, 318 n. 61 Enlart, C, archaeologist and architectural historian, 101, 102, 103, 131 Epiros, despotate of, 191, 192, 198-99, 201, 207, 223, 242, 244-45, 246, 250, 264, 267, 279, 288, 301, 306 Episkopi 132 Eschive de Montbéliard, daughterin-law of John of Ibelin, 111 Estienne of Lusignan, historian, 132, 290 Euboea 161, 191, 192, 200, 221, 222, 231, 232, 233, 234, 248, 252, 259, 264, 276, 292, 303

CASTLES

Famagusta xviii, 23, 61, 90, 93, 95, 98, 106, 107, 109, 113, 114, 115, 116, 117, 118-19, 121, 122, 123, 124, 126, 127, 128, 129, 130, 131, 133, 233, 260, 278, 290, 291, 299, 302, 305; description 95-96, 101-102, 102-104, 105, 128; 1373 siege 90, 103, 120 Fedden, R., historian, 50 Ferdinand of Majorca, claimant to the principality of Achaia, 230 Fermia: see Thermia First Crusade (1095-99) xiii-xiv, xviii, xxi, 19, 28, 37, 56, 58, 160, 219, 353 n. 73, 373-74 n. 22 Fifth Crusade (1217-21) xvii, 10, 16, 20, 39, 41, 57, 67, 69, 77, 115-16, 124, 125-27, 273, 296, 298 Florence 62 Florent of Hainault, prince of Achaia, 199, 250, 255 Forbelet 81 fortifications, defensive functions: fortifications to defend territories against major invasions 38—53, 118-23, 152-57, 243^7; to shelter people against raiders and invaders 39, 47-51, 55, 154-55, 161-67, 247—50; fortifications receiving seaborne reinforcements 44-45, 228-29; sheltering Christians evacuated by sea 45^6, 51, 228-29; protecting sea routes, anchorages and Christian outposts 124-29, 230-32; strategic role of fortifications 51-53, 157-67; castles as look out posts and intervisibuity between fortifications 49-50, 116, 157-61, 251-52; fortifications as shelter against pirates and seaborne raiders 115-18, 230-32; to shelter people and property against minor attackers and criminals 11, 54-57, 247-50

INPEX

—, Internal security: fortifications and internal warfare amongst fellow Christians 9-10, 59-63, 86, 91-93, 110-14, 130, 168-74, 180-85, 258-61; fortifications as shelter against rebellions by local peoples 11, 57-59, 90, 108-10, 253-58 —, Offensive functions: fortifications as a means of gaining and consolidating control over territories 64-67, 227-30, 236-37, 238-40; as starting and supply points for naval campaigns or coastal landings 77, 124-29, 227-30, 232-35; to shelter Christian field armies 65-66, 241-42; as starting and supply points for sieges, raids or sorties 67-73, 237-38; as a means of controlling local areas and peoples 57-59, 71-72, 108-10, 237, 253-54; fortification garrisons as a supply of troops for field armies 73-76, 238; minor fortifications to blockade larger strongpoints 76, 240-42 —, Non-military functions: use of fortifications as arsenals 286-87; residences 272-76; status symbols, 275-76; prisons 277-79 venues for banquets, meetings, weddings, 273-76; venues for courts 280; to store taxes, revenues, important documents, 280-82; to protect and exploit agricultural activities 284-85, 288-89; to protect and regulate trade 291-94; to protect the church 295-96; religious significance of and activities within fortifications 296-98 Fourth Crusade (1202-4) xvii, 35, 89, 93, 122, 126, 137, 168, 191, 197, 200, 201, 205, 206-7, 210, 212, 218, 227, 229, 231, 234, 240, 267, 268, 295

411

Frederick II, 9, 10, 76, 81, 83, 132, 176; crusade of (1228-29) 3, 9-10, 20, 22, 26, 48, 59-60, 61, 62, 65, 80, 86, 91, 110-12, 113, 122, 124, 125, 132, 133, 174, 290, 298, 304, 305, 329 n. 32 Frederick III, king of Sicily, 259 French regiment at Acre 10-11, 60, 74 Gabán 138, 141, 143, 153, 155, 159, 172 Gallipoli 201, 233-34, 237, 246, 305 Gardiki 247, 264, 265, 267, 268, 306 garrisons, size and composition of, 41-42, 43-44, 75-76, 84-85; non-combattants and servants 287-88 Gastria, 126, 131, 132, 133, 134, 306, 307; description 131 Gaston: see Baghras Gautier I, duke of Athens 192-93, 199 Gautier II of Brienne, titular duke of Athens, 239, 245, 256, 302 Gaza 53, 83, 122 Genoa xvi, xxi, 61, 90, 121, 230, 234, 255, 276 Genoese xvii, xviii, xxi; in Holy Land 45, 46, 61-62, 77, 89, 276, 291-92, 305; on Cyprus 98, 105, 107, 109, 111, 115, 116, 117-18, 278, 292; invasion of Cyprus (1373-74) 90 93-94, 103-4, 105, 117-18, 120-23, 130, 260, 284, 302, 303; in Cuidan Armenia 166; in Prankish Greece and the Aegean 191, 200, 221-22, 233, 234, 255, 257, 265 Geoffrey de Bruyere/Briel, lord of Karytaina, 218, 369 n. 103 Geoffrey I of Villehardouin, prince of Achaia, 214, 245, Geoffrey II of Villehardouin, prince of Achaia, 197, 274-75

412

UNKNOWN CRUSADER CASTLES

Geoffrey of Villehardouin, member of Fourth Crusade and chronicler, 196, 200, 206, 227, 241, 254 Gerald 217, 218, 219, 223 Gibelcar: see Akkar Gibelet 3, 26, 44, 60-62, 77, 85, 277, 318 n. 61; description 23 Giovanni de lo Cavo, pirate and Byzantine admiral, 201 Glarentza (Clarence) 217, 222, 225, 228, 230, 234, 237-38, 242, 248, 249, 257, 280, 282, 291, 305; description 214-15 Gökvelioglou 148, 158, 180 Goumardias (Camardias) 170, 179, 180 Grande Maigne: see Old Mania Gravia 261 Gregory IX, pope, 9, 47, 198, 262, 296 Guérin, master of the Hospitallers, 182 Guy II Embriaco, lord of Gibelet, 277 Guy I de la Roche, duke of Athens, 200, 232, 259, 265 Guy II de la Roche, duke of Athens, 273 Guy of Lusignan, king of Jerusalem and ruler of Cyprus, 89, 100-1, 102, 108-109, 115, 124, 125, 131 Guy of Lusignan, king of Cilician Armenia: see Constantine IV Hadjar Shoglan: see Roche Roussel Haifa 26, 28, 45, 65, 225 Haliartos (Moulki) 221, 222, 247, 252; description 220 Halmyros, battle of (1311) 193, 238, 245, 255, 369 n. 103, 369 n. 107 halqa 7 Kama 50, 69, 71, 75, 332 n. 41 Harunia (Haruniye/Haroun) 158, 159, 162, 170-71, 177-78, 181, 187, 306; description 175-76 Hasanbeyli: see Black Tower. Hattin, battle of (1187) xvi, xxi, 3, 6,

39, 45, 58, 65, 66, 73, 81, 83, 85, 109, 182, 255, 304 Hellenkemper, H., archaeologist, xix-xx, 158 Henry I, king of Cyprus, 91, 110, 114, 126 Henry II, king of Cyprus and Jerusalem, 9, 10, 60, 91, 94, 102-3, 104, 105, 106- 107, 112-13, 114, 116, 117, 118, 125, 127, 128, 130, 173, 188, 265, 274, 277, 290 Henry VI, German emperor, 132, 175; crusade of (1197-98) 67, 68 Henry, Latin emperor, 194, 196, 207, 208, 212, 227, 229, 236, 237, 239, 242, 243-44, 258-59, 264, 267, 275 Henry, brother of Bohemond V, 71 Henry, lord of Silifke, 170, 180 Henry of Valenciennes, chronicler, 208 Hermann of Salza, master of the Teutonic Knights, 175 Hethoum I, king of Cilician Armenia, 138, 141, 142, 143, 152, 153, 156, 159, 161, 162, 167, 171, 172, 173, 176, 177-78, 179, 182, 186 Hethoum II, king of Cilician Armenia, 142, 143, 172-73 Hethoum, lord of Lampron, 169—70 Hethoumids 137, 163-64, 168-70, 171-72, 173, 187 Hexamilion Wall 246 Hisn al-Akrad: see Crac des Chevaliers Holy League 232-33 Holy Sepulchre ( Jerusalem), 297 Horns 41, 50, 68, 69, 75, 188, 332 n.41 Honorius III, pope, 125, 126, 198, 214, 246 Hospitallers xvi, 3, 13-14, 17-18, 20-21, 28, 31, 41, 42, 43-44, 45, 49, 52, 54-55, 60, 61, 68, 69-70, 71-73, 74-76, 79-86, 106, 107, 113, 126, 127, 128, 130-34, 150, 170, 171, 172, 174, 175, 178, 191, 194,

INDEX

237, 251, 272, 273, 274, 275, 277, 281, 282, 283, 284, 285, 287, 288-89, 293-94, 295-98, 299, 303, 305-7, 332 n. 41; strongholds and properties in the Holy Land 61, 81-82, 83-84, 150, 274; on Cyprus 132; in Cuidan Armenia 150, 170, 179-82, 187-88; in Prankish Greece and Rhodes 232-33, 262, 263, 264-68 Hugh I, king of Cyprus, 91, 102, 126, 132 Hugh III, king of Cyprus and Jerusalem, 9, 60, 125, 131, Hugh IV, king of Cyprus, 116, 278 Hugh of Lusignan, claimant to principality of Achaia, 230, 259-60 Hugues de Bruyere/Briel, lord of Karytaina, 218 Hunin: see Cháteauneuf (Holy Land) Ibelin, castle, 122, 334 n. 87 Ibelin war with Frederick II and his forces (1228-42), 9-10, 28, 44-45, 59-60, 76, 77, 90-91, 94, 98, 102, 110-12, 114, 122, 132, 133, 285, 304, 326 n. 59. See also John of Ibelin. Ibn al-Athir, chronicler, 183, 185 Ibn al-Furat, chronicler, 72, 82, 154, 186, 188, 278 flan Kale: see Yuan Innocent III, pope, 126, 131, 183, 184, 218, 249, 264 Innocent VI, pope, 266 loannina 199, 250 Isaac Komnenos, ruler of Cyprus, 89, 95-96, 100-1, 107, 279 Isabella II, queen of Jerusalem (1212-28) 9 Isabelle, queen of Cuidan Armenia, 163, 171-72, 173, 176, 178, 182 Isabelle of Villehardouin, princess of Achaia, 247, 257, 273

413

Jábala 52, 68, 69, 82, 181 Jacques de Vitry, preacher, 49, 57, 127 Jaffa xvi, 3, 19, 26, 33, 40, 44, 59, 65, 66, 67, 73, 74, 76, 83, 85, 225, 273, 275, 298, 330 n. 10; description 20-22, 28 James I, king of Cyprus, 104, 105, 107, 118, 121-22, 277, 282, 299 James of Lusignan, brother of King Peter I of Cyprus, 93 Janus, king of Cyprus, 105, 110, 121 Jean Dardel, chronicler, 173 Jericho 83 Jerusalem xiii, xiv, xvi, 3, 32, 40-43, 44, 46, 47, 53, 57, 58-59, 65-67, 70, 77, 81, 83, 84, 121, 155, 232, 253, 295, 320 n. 86, 324 n. 20; description 23-24, 26 Jiddin: see Judin Joanna, queen of Naples and princess of Achaia, 266 John XXII, pope, 166, 188 John, Armenian Catholicos, 170 John Asen II, ruler of Bulgaria, 200-1, 243 John of Brienne, king of Jerusalem and Latin co-emperor, 23, 103, 196 John I of Ibelin, lord of Beirut 21, 44-45, 86, 91, 110-112, 114, 132, 133 John II of Ibelin, lord of Beirut, 293 John of Joinville, chronicler and crusader 21, 57, 65, 85, 275, 281, 295 John of Lusignan (and Antioch), brother of King Peter I and Constable of Cyprus, 93, 98, 110, 121 John III Vatatzes, Nicaean emperor, 200-1, 243 Jordan, river 49, 70, 72, 83, 126, 283 Joscelin, son of Henry of Silifke, 170, 180

414

UNKNOWN CRUSADER CASTLES

Judin (Jiddin) 175, 262; description 79-80 Julian, lord of Sidon 48, 69, 79, 80 Kafr Dubbin, 68 Kalamata 106, 211, 213, 217, 219, 232, 247, 254, 257; description 209-10 Kalavryta 217, 223 Kalavun, sultan of Egypt, 7, 8, 11, 22, 55, 71, 156 Kalojan, ruler of Bulgaria, 196, 206, 207, 227, 240, 241, 243-44, 250, 251, 255, 256, 257 al-Kamil, sultan of Egypt, 4, 58 Kantara 91, 95, 98, 106, 110, 112, 114, 116, 119, 122, 130, 148; description 100-101, 105; 1229-30 siege 110-11 Karamania, 173 Karytaina 217, 218, 219, 222, 223, 236, 237-38, 242, 247, 251, 255, 258, 263, 273, 369 n. 103, 370 n. 121; description 215-16, 225 Kavala: see Christopoli Kay-Kavus, sultan of Rum, 138 Kay-Khushraw I, sultan of Rum, 138 Kay-Qubad I, sultan of Rum, 138 Kennedy, H., historian, xix Kerak, 4, 38, 42, 272, 285 Khirokitia, 132, 133-34; description 131 Khwarizmians and Khwarizmian invasion (1244) 3, 6, 10, 26, 40, 43, 47, 54, 55, 67, 74, 119, 155, 157, 245, 320 n. 86 Kibotos 229 Kolossi, 133-34, 268, 275, 282, 284, 307; description 132 Kyrenia 90, 91, 95, 101, 104, 106, 107, 110, 113, 114, 116, 117, 118-19, 120, 123, 130, 272, 273, 276, 277, 278, 280, 299; description

96-98, 102-103, 105; 1232-33 siege 91, 98, 111-12; 1374 siege 90, 98, 120-21, 122-23, 128-29, 284 La Cava, 118, 122; description, 105 La Fève 38 La Forbie, battle of (1244) 3, 6, 10, 40, 55, 73, 74, 83 Laffustan (Phostena) 263, 307 Lajazzo: see Ayas Lamia: see Zeitoun Lampron 138, 157, 159, 162, 168-70, 174, 188, 296, 304; description 147-48, 168-69 Larissa 207, 259 Larnaca 105 Latakia 65, 68, 71, 82, 85, 86, 285 Latin empire, 191, 192, 194-96, 200-1, 227, 228, 229, 243-44, 251, 256, 258, 263, 301, 302 Latrun (Toron des Chevaliers) 66-67, 83 Lausada 138 Lefkosia: see Nicosia Leon II, king of Cuidan Armenia, 60, 80, 137-38, 140, 143, 144, 150, 153, 155, 158, 159, 161, 162, 163, 165, 167, 168-71, 173, 174, 175, 176, 177-78, 179-82, 183-85, 186, 187, 307, 354 n. 9 Leon III, king of Cilician Armenia 143, 153, 154, 156, 172 Leon W, king of Cilician Armenia, 140, 173 Leon V, king of Cilician Armenia, 145, 173 Leon VI, king of Cilician Armenia, 164, 173 Leon Sgouros, lord of Corinth, 211, 246 Leontios Makhairas, chronicler, 278 Leutron: see Beaufort (Achaia) Licario of Karystos, Byzantine mercenary, 192

INDEX

Limassol 6-7, 89, 95, 98, 106, 109, 113, 114, 115-16, 117, 118, 122, 126, 127, 128, 130, 132, 133, 134, 290-91, 292; description 95-96, 101-102, 105, 124, 125 Lindos (Rhodes) 265 Livadia 217, 223, 224, 238, 257; description 215 Lombard rebellion in Thessaly (1207-9) 207, 211, 227, 258-59, 264, 275 Lombard troops of Frederick II: see Ibelin war with Frederick II Louis IX, king of France, 9, 10-11, 35, 74, 193; first crusade (1248-54): xvii, 3-4, 20, 23, 24, 27-28, 47, 48, 57, 65, 67, 69, 72, 85, 124, 125, 127, 197, 201, 225, 273, 275, 281, 290, 298, 305; second crusade (1270) 77 Loulon 138, 160 Ludolph of Sudheim, pilgrim, 251, 263, 266 Lydda 3 Maditos 237 Maina: see Old Mania al-Makrizi, chronicler, 153, 154, 156 Maldoim 38 Malik an-Nasir Dawud, lord of Kerak, 42-43 Mamluks and Mamluk forces 4, 7-8, 30, 35, 93-94, 115, 119, 138-39, 140-41, 155, 225, 301 Manuel I Komnenos, Byzantine emperor, 205 Manzikert, battle of (1071), 168 Mara Brankowicz 221 Maraclea 31, 65, 70-71, 78, 86, 127, 318 n. 57-58; description 22 Marash 138, 152, 157, 294 Marco Polo, merchant, 166 Marco I Sañudo, duke of the Archipelago (Naxos), 222, 228, 231

415

Marco II Sañudo, duke of the Archipelago (Naxos), 231 Margarita Tower (Nicosia), 104-105, 119, 278 Margat 8, 19, 22, 23, 28, 31, 38-39, 40, 52, 53, 68, 69, 71, 72, 73, 75-76, 78, 81, 84- 85, 86, 128, 150, 179, 181, 272, 273, 279, 282, 284, 285, 286, 287, 296, 297, 299, 306; description 13-14, 16; 1281 siege: 40, 42, 73; 1285 siege 8, 16; toll station 293-94 Maria of Antioch, claimant to the kingdom of Jerusalem, 60 Markopoulo 221, 222; description 219-20 Marqab: see Margat Marshall, C., historian, xviii, xix, xx Mathegriffon: see Akova Matthew Paris, chronicler, 53 Mayer, H., historian, 57, 58 Melings 238-39, 254 Mhalia (Castrum Regis) 79, 283-84 Michael Komnenos Doukas, despot of Epiros, 207 Michael VIII Palaiologös, Byzantine emperor, 89-90, 192, 193, 198, 201, 206, 228, 236, 237-38, 239, 255 Michael the Syrian, chronicler, 158 Mikonos 200, 231 military architecture: major castles 13-19, 96-101, 102-5, 145-51, 203-19; minor castles 28-30, 95-96, 101-2, 105, 219-22; urban defences 20-21, 24-28, 42-44, 51, 98, 103-5, 151, 161-67, 205-7, 214-15; keeps and donjons 14, 16, 18, 38, 80, 175, 211, 213, 218-19, 224; citadels 19-24, 42-44, 47-48, 51; bourg defences 28, 47, 54, 162-63, 249-50; castrum castles 28, 29, 37, 54, 105, 107, 122, 299; walls and ramparts 22, 31, 145-47, 203—4, 213; techniques used to

4i6

U N K N O W N CRUSADER CASTLES

strengthen walls and ramparts 22, 31, 145-47; flanking towers 31-32, 33-34, 145-47, 203-4, 213; concentric fortifications 34, 36-37; gateways, posterns barbicans and outworks 26, 27, 32-33, 34-35, 145-47, 214-15; moats and ditches 32; isolated towers and watch towers 29, 56-57, 81, 82, 101-2, 131-33, 176, 219-22, 247-49, 262-63; architectural developments, influences and construction of fortifications 29-34, 76-77, 106-7, 147-51, 203-4, 215-19, 222-26; workshops, kitchens, baths, mills, stables etc. 285-87; chapels and monastic structures 297 mines: see siege weapons Misis 158, 160, 162, 163, 168, 176, 180; description 162, 163 Mistra 192, 193, 194, 198, 201, 216, 217, 219, 225, 228, 232, 236, 237, 238-39, 242, 246, 247, 255, 256, 257, 260, 262-63, 267, 280-81, 284, 289, 304; description 204, 205, 217 Modon 191, 204, 227, 228, 229, 233, 249, 267, 291, 366 n. 57; description 212 Monemvasia 192, 193, 198, 204, 218, 228, 229, 232, 233, 234, 236, 237, 242, 246, 247, 251, 255, 256, 257, 260, 291; description 208, 372 n. 32 Mongols 4, 7, 41, 46, 48, 68-69, 79, 84, 127, 138, 139-40, 141, 142, 143, 154, 155, 156, 164, 165, 173, 185, 188, 292, 301 Mons Glavianus 281 Montferrand 69, 71, 75 Montfort 11, 13, 23, 24, 36, 40, 47, 69, 74, 79-80, 150, 175, 219, 262, 272, 273, 281, 286, 289, 290, 296, 299, 301, 323 n. 142; description 18-19, 80, 273; 1271 siege 36, 289

Montmusard (Acre) 24, 44, 274, 319 n.75 Montreal 36, 38-39, 283 Morea: see Achaia, Chronicle of Morea. Morocco 117 Mostenitsa 262, 263, 267-68 Moulki: see Haliartos Mount Athos 249 Mount Silpius 24, 25, 43 Mount Tabor, 26, 67, 70, 72, 81, 283, 287, 296; description 18 al-Muazzam 26, 33, 39, 40, 44, 55 Naples, rulers of, 193, 228, 289 Nauplia (Nauplion) 204, 218, 221, 229, 249, 288, 365 n. 39; description 210-11, 211- 12 Navarrese company, 194 navies and naval strengths: Byzantine 201; Catalan 201; Latin and Italian 6-7, 45-47, 200-1; Muslim 6-7, 47, 90; Turkish 201 Naxos 191, 222, 231, 233; castle 222, 228, 231, 250 Nazareth 3, 296 Negroponte 233, 276, 292 Neopatras 106, 213, 219, 224, 257, 264, 365 n. 40; description 210-11 Nephin 23, 32, 60-61, 62, 86, 277, 318 n. 57; description 22 Nerses Balientz, chronicler, 148-49 Newark 276 Nicaea, Greek empire of, 191, 192, 194, 196, 197, 198, 201, 205, 208, 229, 238, 242, 267, 289, 301 Niccolo Acciajuoli, lord of Corinth, 193, 199, 232, 247, 289 Niccolo da Martoni, pilgrim, 231, 264, 267 Nicholas IV, pope, 90 Nicholas II of Saint Omer, co-ruler of Thebes, 212, 218, 228 Nicholas III of Saint Omer, baili of Achaia, 262

INDEX

Nicomedia 208, 227, 229, 239, 255 Nicosia 90, 91, 95, 98, 100, 106, 107, 108, 109, 110, 111, 113, 114, 115, 116, 118, 119, 120, 121, 122, 123, 124, 130, 131, 132, 133, 260, 272, 278, 282, 285, 299; description 95-96, 101-102, 102-103, 104-105 Nikli 275 Norpert 170, 179, 180 Nur al-Din, ruler of Aleppo, xv, 38, 52, 80 Odo of Chäteauroux, papal legate accompanying Louis IX, 85 Odo of Montbéliard, constable of Jerusalem, 22, 86 Old Mania (Maina/Grande Maigne) 192, 193, 198, 217, 238-39, 246, 255, 260 Old Navarino 217, 219, 228, 259, 280 Oliver of Paderborn, crusader and chronicler, 39, 66, 70, 273 Oreus (Euboea) 234 Oschin, king of Cilician Armenia, 173, 188, 278 Oschin, lord of Corycus, 142, 173 Oschin, lord of Lampron, 169 Otto IV, German emperor, 175 Ottoman and other Turks in Asia Minor xiii, xvii, xviii, xxi, 90, 94, 105, 118, 138, 194, 199-200, 201, 204, 213, 215, 223, 231, 232-33, 246, 251, 265-66, 284, 289, 291, 301, 307 Palaiocastro (Crete) 234 Palaiopolis 263, 267, 307 Pamphilon 239, 241; description 208 Panidos 233 Panormos 208 Paphos, 90, 95, 98, 106, 109, 113, 114, 115, 116, 117, 117-18, 122, 124, 125-27, 130, 132, 218, 272, 286, 306; description 96, 101-102, 105, 338 n. 5

417

Patras 204, 223, 224, 230, 240, 254, 259, 291, 364-65 n. 32; description 209, 210-11 Pelagonia, battle of (1259), 192, 208, 238, 246, 255, 275 Pendeskouphi (Corinth) 240 Pergamum 267 Peter I, king of Cyprus, 91-93, 98, 104, 109, 116, 119-20, 123, 127-28, 129, 173, 188, 278, 337 n. 13 Peter II, king of Cyprus, 93, 103, 104, 122 Peter of Bracieux, lord of Cyzicus, 208 Peter Campofregoso, Genoese admiral, 104 Peter Thomas, papal legate, 109 Philip, king of Cilician Armenia, 140, 172, 173 Philip of Novara, chronicler and Ibelin supporter, 132, 133 Philip of Savoy, prince of Achaia, 255-56, 257-58, 274 Philip II of Taranto, prince of Achaia, 230 Philippa, niece of Leon II, 169 Philippopolis 206, 208, 239, 256-57 Phostena: see Laffustan PhyUa (Euboea) 248 Pigae: see Spiga Pilgrims' Castle: see 'Atlit piracy and seaborne raiding 89, 90, 199-201, 230-35 Pisa xvi, 230 Pisans, in Holy Land 61, 276, 291; in Cilician Armenia 166 Poimanenon, battle of (1224), 242 Politika (Euboea) 276; description 221, 276 Port Bonnel (Arsuz) 182, 186, 187 Port de Jone: see Old Navarino Pringle, D, archaeologist, xviii, xix Qaqun, 54, 68, 69, 70, 106, 282; description 29

4i8

UNKNOWN CRUSADER CASTLES

Rafania 289 Ramlah 3 Ramon Muntaner, chronicler, 199, 233-34 Ravennika 263-64, 267, 268, 306 Raymond III, count of Tripoli, 45 Raymond, son of Bohemond III, 180 Raymond Roupen, claimant to Antioch and Cilician Armenia, 60, 68, 82, 163, 171, 172, 180-81, 183, 186 Recordane, 54-55, 64, 82, 284, 328 n. 11 Red Tower, 82, 275 Renaud III, lord of Nephin, 60-61 Renier of Trit, lord of Philippopolis 256-57 Rey, E. G., archaeologist, 50, 51, 52 Rhodes 116, 127, 128, 130, 133, 188, 191, 232-33, 251, 267, 302; description 265-66 Richard I, king of England, xvi, xvii, xviii, xxi, 3, 6, 20, 40-41, 44, 53, 66, 67, 76-77, 82, 83, 89, 90, 95-96, 100-1, 106, 108, 110, 131, 219, 298, 324 n. 20 Richard, count of Cephalonia, 250 Richard of Cornwall's crusade (1240^1) 3, 20, 22, 45, 53 Richard Filangieri, imperial marshal of Frederick II, 10, 76, 91, 102, 111-12 Robert of Taranto, prince of Achaia, 230, 259-60 Roche Guillaume 182, 183, 185, 186, 187, 359 n. 65 Roche Roussel (Hadjar Shoglan/Chilvan Kale) 68, 73, 182, 183, 186, 187, 359 n. 65; description 185 Rodosto 206, 227, 233, 241 Roger de Lluria, vicar-general of the duchy of Athens, 259 Rouad, island of, 127, 128, 129, 133, 288

Roupen III, ruler of Cilician Armenia, 169 Roupenids 137, 168-70 Rousion 206, 237, 243 Rovies (Euboea) 232 Safad: see Saphet Safitha: see Chastel Blanc Ste Hélène 255 St George (Achaia) 232, 236, 255, 256, 258, 263, 267, 373 n. 4 St George (Cyprus) 133 St George (Holy Land) 69 St Hilarion 91, 95, 106, 110, 114, 116, 119, 120, 122, 123, 130, 148, 272, 273; description 98-101, 105, 272, 340 n.59; 1229-30 siege 110-12; 1232 siege 111-12 St Lazarus (Order of) 44, 293 St Mary of the Latins 82 St Nicholas' Gate (Acre) 83 Saint Omer 249 St Sabas, War of, 61 St Stephen's Gate (Jerusalem) 26 Saladin, sultan of Egypt, xvi, xvii, xviii, xxi, 4, 6, 7, 11, 17, 18, 20, 22, 23, 33, 38-39, 40-41, 44, 45, 47, 49, 52, 53, 54, 55, 57, 59, 64, 66, 77, 81, 83, 86, 109, 123, 137, 168, 182, 184, 238, 240, 283, 285, 295, 302, 315 n. 34, 316 n. 37; 1187 campaign xiv, 3 6, 38-39, 53, 67, 73-74, 238; 1188 campaign xvi, 3, 6, 38-39, 40, 45, 53, 68, 85, 182, 323 n. 5 al-Salih, sultan of Egypt, 42 Salona 198, 212, 219, 223, 249, 251; description 213 Samaria 249 Sañudo dukes of Naxos, 191, 249-50 Saone 85, 148, 277 Saphet (Safad) 3, 19, 24, 28, 31, 32, 33, 35, 36, 38, 39, 40, 47, 48-49, 53, 54, 65, 70, 72, 74, 75, 78, 79, 80, 84, 119, 155, 237, 262, 278,

INDEX

419

281-82, 283, 284, 286-87, 288, 295, Smail, R. C., historian, xviii, xx, 296, 297, 301, 303, 305, 313 n. 47; 51-52, 86 description 16-17, 314 n. 19, 321 Smyrna 188, 233, 234, 266, 305 n. Ill; 1266 siege 7, 9, 17, 36, 41, 48 Spiga (Pigae) 208; description 205 Satalia (Antalya) 128, 129, 138, 188, 267 Stenimaka 206, 256 Second Crusade (1147-49) xvi Subeibe 49 Seleucia: see Silifke Suda (Crete) 254 Seljuk Turks and Seljuk forces xiii, Sykaminon 264, 267 137, 138, 140, 141, 153, 155, 168, Syrian Gates: see Belén Pass 170, 174, 179- 80, 194, 301, 306 Selymbria 243-44 Tagliacozzo, battle of (1268), 197 Sempad, king of Cilician Armenia, 172 Tarsus 142, 143, 160, 162, 168, 169, Sempad, Armenian Constable and 171, 176, 353 n. 73; description chronicler, 142, 143, 163-64 162, 163-64 Serres 207, 242, 250, 251, 257, 259 Tchorlu 196, 236 Servantikar 148, 150, 151, 152, 153, Telos 265 157, 159, 162, 166, 176, 179, 187, Templars xvi, 3, 9, 14-17, 19, 32, 294, 299, 349 n. 4; description 33-34, 35, 38-39, 41, 43-44, 49, 145-47 51, 52, 54-55, 57, 59-61 62, 66-67, Sherborne 276 68, 69, 70, 71, 72, 73, 74-76, 77, Siderokastron 257 79-86, 89, 108, 113, 124, 126, 127, Sidon 3, 18, 19, 26, 31, 45, 46, 47, 48, 130-34, 163, 171, 174, 175, 180, 51, 60, 65, 67, 69, 72, 77, 79, 80, 237, 278, 280, 281-82, 283, 285, 83, 85, 225, 229, 273, 285, 286, 286-88, 293, 295, 296-98, 304, 297, 298, 305, 330 n. 10; 305-7, 332 n.41; strongholds and description 22-23, 28, 297 properties in the Holy Land 61, siege weapons and engineers 8-9, 141, 80-81, 82-84; on Cyprus 131-32, 143-44, 147, 225, 286-87, 312 133; in Cilician Armenia 182-87; in n. 35; catapults and trebuchets 8-9, Prankish Greece 262, 263-65, 267, 268. 30-31, 32-34, 35, 37, 61, 93, 119, 141, 147, 256; mines 8, 31, 34, 141 Templar headquarters (Acre), 51, 59, Sigouri, 107, 118, 122, 299; 61, 133, 281; (Paris) 281 description 105 Teutonic Knights xvi, 16, 18-19, 26, Silifke (Seleucia) 14, 137, 138, 143, 36, 43-44, 61, 73, 74, 84, 132-33, 150, 152, 153, 155, 159, 170, 172, 161, 162, 170, 170-71, 172, 174, 178, 181-82, 187, 188, 299, 306; 182, 272, 273, 281, 283-84, 295, description 179-80 296, 297-98, 299, 304, 305-7; Sion Gate (Jerusalem) 26 strongholds and properties in the Siquinum 188 Holy Land 61, 79-80, 83-84, 273, Siros 231 286- 87; on Cyprus 132-33; in Sis 138, 140, 141, 147, 148, 154, 155, Cilician Armenia 161, 175-79, 181, 156, 157, 160, 161, 165, 168, 169, 187, 294; in Prankish Greece 173, 178, 278; description 164-65; 262-63, 267-68 1375 siege 164-65, 173 Thebes 198, 218, 224, 239, 251, 257,

42O

UNKNOWN CRUSADER CASTLES

259, 272; description 210-11, 212, 272, 366 n. 52 Theobald of Champagne's crusade (1239-40) 3, 20, 22, 53 Theodore, despot of Epiros, 207, 210, 246 Theodore I Laskaris, Nicaean emperor, 227, 229, 239^10 Theodosius II, Byzantine emperor, 205 Thermia (Fermia) 231 Thessaloniki, city and kingdom of, 191, 192, 194-96, 197, 198, 227, 233, 236, 246, 257, 258, 260, 275; description of city 206-7; 1224 siege 207, 288 Third Crusade (1189-92) xvi, xvii, xviii, 3, 6, 20, 40-41, 44, 66, 76-77, 82, 83, 89, 271 Thomas, despot of Epiros, 250, 279 Thomson, J., historian, 50 Thoros, brother of Hethoum II, 172 Thoros, ruler of Cilician Armenia, 163 Thurion 220, 221 Tiberias 3, 26-27, 31, 43, 45, 58, 65, 73, 86; description 22 Tibnin: see Toron Til Hamdoun: see Toprak Tinos 200, 231 Toprak (Tu Hamdoun) 154, 156, 157, 158, 159, 160; description 148-49, 157 Tor de l'Opital 54, 82 Toron (Tibnin) 3, 49, 68, 69, 80 Toron des Chevaliers: see Latrun Tortosa 19, 22, 30, 31, 32, 33, 38-39, 40, 44, 49, 52, 53, 75, 80, 82, 84-85, 86, 119, 127, 150, 183, 288, 295, 298, 299, 301, 306; description 14-16 Tower of David (Jerusalem), 23, 41, 42-43, 59, 304 Tower of the Flies (Acre) 46, 292 Trapesac: see Darbsak

Trebizond 191 trebuchet: see siege weapons Tripoli xiv, xvi, 3, 28, 39, 43, 44, 45, 49, 54, 60-62, 64, 67, 68, 77, 78, 80, 83, 260, 277, 281, 284, 285, 290, 323 n. 5; description 24, 25-26; 1289 siege 7, 24, 51, 83 Tukhlah 29, 50, 54, 82, 219 Tumlu 148, 158, 160, 168 Turris Salinarum 82 Tyre xiv, xvi, 3, 6, 24, 35, 39, 43, 45, 46, 47, 49, 54, 57, 62, 77, 78, 82, 84, 90, 91, 103, 107, 119, 144, 151, 291, 299, 305; description 23, 25-26; 1187 siege: 67, 302, 323 n. 5; Ibelin attack on (1242) 10, 77, 326 n.59 Vagha 148, 150, 153, 155, 160, 163, 166, 168, 296; description 147-48 Vahram of Edessa, chronicler, 154, 158 Valania: see Banyas Vardounia 217 Vasil, citizen of Tarsus, 163-64 Vasil, lord of Vaner, 170, 180 Veligosti 247 Venice xvi, 223, 228, 230, 234 Venetians xvii, xviii; in Holy Land 46, 61-62, 77, 276, 291-92, 305; on Cyprus 90, 94, 95, 98, 104, 105, 109, 123; in Cilician Armenia 166, 292; during the Fourth Crusade and in Prankish Greece, 191, 194, 200, 201, 204, 206, 211, 212, 22122, 228, 229-30, 231, 232-33, 234, 240, 241, 243, 250, 252, 254, 257, 259, 276, 284, 291, 292 Villehardouin: see Geoffrey, Isabelle, William of Villehardouin Vostitza 230, 274 William Alemán, lord of Patras, 209, 211, 214, 224, 254, 364 n. 32

INDEX

William I of Champlitte, prince of Achaia, 209, 230, 240, 245, 263 William de la Roche, duke of Athens, 264 William Sañudo, lord of Siros, 231 William II of Villehardouin, prince of Achaia, 192, 193, 197, 200, 201, 203-4, 208, 211, 217, 227, 229, 230, 232, 237-38, 238-39, 241, 246, 247, 254, 259, 265, 275, 278, 279, 289 William II, king of Sicily, 67 Willbrand of Oldenburg, German pilgrim and imperial representative, 19-20, 54, 75, 82, 84, 98, 101, 102, 124, 126, 148-49, 154, 162, 163-65, 175, 176, 183, 185, 272, 277, 283

421

William of Tyre, chronicler, xviii, 36, 285 Winfield, D., archaeologist, 204, 217, 218, 219 Xoz Jor 153

Yermasoia 131, 132, 133-34 Yuan (Ilan Kale/The Castle of the Snakes), 150, 154, 157, 158, 159, 160, 162, 168, 179, 299; description 148, 149 Zabel, queen of Cilician Armenia: see Isabelle Zeitoun 212, 263-64, 265, 267, 268, 306 Zengids 6