The World of the Skandapuran.a

Supplement

to

Groningen Oriental Studies

Published under the auspices of the J. Gonda FoundationRoyal Netherlands Academy of Arts and Sciences

Editor

H.T. Bakker, Groningen

Editorial Board

P.C. Bisschop • D.D.S. Goodall

H. Isaacson • G.J. Meulenbeld

Advisory Board

R.F. Gombrich, Oxford • J. Heesterman †, Leiden

D. Shulman, Jerusalem • J. Williams, Berkeley

for

Aafke

The Skandapuran. a research team on Mount Kalanjara

from left to right: Mudassir Ansari, Elizabeth Cecil, Peter Bisschop, Nina Mirnig,Hans Bakker, Anna Ślączka, Judit Torzsok, Yuko Yokochi, Bijendra Singh

Hans T. Bakker

The World of the Skandapuran. a

Northern India in the Sixth and Seventh Centuries

Leiden | Boston

Typesetting and layout: H.T. Bakker

Research and production of this book have been made possible byfinancial support from:

the Netherlands Organisation for Scientific Research

the European Research Council

ISBN . . .

...........

Copyright c© 2014 by Koninklijke Brill NV, Leiden, The Netherlands

............

Preface

In this book our research on the original Skandapuran. a, pursued over along period of time, comes to a temporary conclusion. And yet, the resultsof our studies cannot be qualified as anything but provisional. Though wehave been working on this Puran. a as a team since the early 1990s, thesheer size of the text has limited the degree to which we are able tocomprehend its full significance. At the moment of writing the adhyayas1 to 69 and 167 have been critically edited and studied in detail, whichis about one third of the text. However, despite this limited progress, ithas become abundantly clear that the Skandapuran. a is a major exampleof Sanskrit literature, an invaluable source for the cultural and religiousworld of early medieval northern India. To explore this world is the aimof this monograph.

The book is divided into two parts.

The first part is a survey of the political landscape, from the fall of theGupta Empire to the chaotic days of Hars.avardhana’s successors. This isa turbulent period in Indian history, seeing the fall of one empire and therise of another in barely more than one century. We have to situate thecomposition of our text in this landscape. It is my sincere conviction thatan appraisal of its religious content cannot be done properly, unless thematrix of the political history of North India in the sixth and seventhcenturies is first carefully worked out.

The present survey is not an economic and social history of early me-dieval India, nor are questions regarding the changing nature and organ-isation of the state addressed. This I have to leave to historians. Thepresent historical survey is informed by the question of how the cultureand religion embedded in the Skandapuran. a could have developed withinthe political ramifications of the investigated period. Due to the scarcityof our historical sources as well as the nature thereof, such historiographynecessarily assumes the form of a description of royal dynasties, their com-ing to power, their wars and decline. I am aware that many Indologistsfind this a rather tedious branch of the discipline. If it does not put themoff altogether, an option could be to pass over Part I and to begin with

vii

viii Preface

Part II, in which, through cross-references, readers are directed towardsthe relevant historical circumstances elaborated in Part I.

In order to lay bare the historical ramifications, my methodologicalstrategy has been to give priority to the Sanskrit sources, which for thegreater part are inscriptions. I have presented these sources by renderingthem into English, while the Sanskrit texts are given in the footnotes,in orthographic spelling and, when needed, in slightly emended form, if Iwas confident of the intended reading. When the significance hinges on aproblematic or dubious reading this has been made explicit. For consultingthe original, diplomatic edition, scholars are referred to the publicationsof the inscriptions in Epigraphia Indica etc.

The preference given to the primary sources has also been motivatedby the state of Indian historiography. I find the extent to which old ideasare being recycled without being checked against the actual sources oftenunacceptable. Instances thereof include, for example, the untenable theorythat the Later Guptas moved to Malwa after their defeat by the Maukharis(discussed in n. 247 on p. 86), and the supposedly epoch-making battlebetween Hars.avardhana of Kanauj and Pulakeśin II of Badami, which isgenerally believed to have been ‘a turning point in the history of India,’discussed on p. 109.

Part II of this study is devoted to the exploration of the Skandapura-n. a’s connection with the major religious movement of its time, PaśupataSaivism, and of historical sites where archaeological remnants still testifyto this connection. Here we had to be pragmatic. It was impossible to visitand explore all the places mentioned in the text for obvious reasons. Weselected six sites or areas which play an important part in the mythologyrelated by our text and which have preserved enough sixth and seventh-century remains to allow an assessment of their links with the Skandapura-n. a. The findings of our fieldwork sometimes reached beyond the world ofthe Puran. a. This, for instance, was the case on Kalinjar Hill (Kalanjara).The impressiveness of the remains there and their relative obscurity haveled me to indulge in them a little more than can strictly be justified bythe Skandapuran. a.

The fieldwork that underlies Part II has partly been a joint undertakingby the team of scholars that has worked with me on the Skandapuran. aover the last five years. A photograph of this team on Mount Kalanjara ispresented in the frontispiece. The sixth site which I have chosen to discussin this book, Kot.ivars.a or Bangarh in West Bengal, I did not visit myself.It was investigated, however, by my friend and teammate Yuko Yokochi(Kyoto), and I am grateful that she has been willing to share her findingsand to write this chapter together with me. Professor Yokochi has also

Preface ix

prepared an edition of the Kot. ıvars.amahatmya in Skandapuran. a 171, andwe present this text in an Appendix to this chapter.

The purpose of the six studies contained in Part II is to give the Puran. aa Sitz im Leben. The Mahatmyas related in the text were not value-free.They served the interests of the Maheśvara or Paśupata communities byproviding them with an authoritative basis for the claims to holiness ofthe sites and institutions in their charge. As such the text supported thesefellow believers in their rivalry with other religious organisations, andthis may have been one of the considerations that inspired some learneddevotees to compose a huge work like the Skandapuran. a. Though thepolitical history in Part I is meant to supply a coherent framework, thesix studies still remain fragmented to some extent. In the present stateof our knowledge this seems unavoidable. However, if the mosaic work ofPart II can evoke a sense of the historical actuality of the world of theSkandapuran. a, my aim has been achieved.

As I have noted above, the present book is very much the product ofcollaborative work.

A team was formed in September 2008, when we were awarded a grantfor the project A Historical Enquiry Concerning the Composition andSpread of the Skandapuran. a by the Netherlands Organisation for Scien-tific Research (NWO) to continue our efforts to disclose the world of theSkandapuran. a. In addition to the old team members, Harunaga Isaac-son (Hamburg), Yuko Yokochi (Kyoto) and Peter Bisschop (Leiden), MsNatasja Bosma and Ms Martine Kropman joined our group and were in-stalled as PhD researchers at the University of Groningen. We regret thatMs Kropman was obliged to resign for reasons of health, but in the yearthat she worked with us she contributed important new ideas, which havebeen incorporated into this book. Ms Kropman was succeeded by Dr NinaMirnig, who as a postdoc researcher explored the history of early Saivismin Nepal and its connection with northern India. In writing this book Ihave greatly benefited from the results of her studies, and from those ob-tained by Ms Natasja Bosma, who investigated forms of early Saivism inDaks.in. a Kosala (Chhattisgarh). Drafts of this study in various stages havebeen discussed during the meetings of the research team, which has beenreinforced by Dr Judit Torzsok (Lille) since 2010. Unforgettable was ourjoint fieldwork trip in January 2012, which has contributed so much tothe present study. I owe a large debt of gratitude to all members of thisresearch group.

In the course of my research within the framework of the Skandapuran. ascheme, I became increasingly aware of the important role of temples andmonastic institutions with their landed estates in the formation of Gupta

x Preface

and post-Gupta culture and society. The Paśupata movement played a keyrole in this development and the Skandapuran. a is just one of its tangibleand remaining results. My thoughts on this subject were brought intofocus by intensive discussions and joint fieldwork with Michael Willis,curator South Asian Collections at the British Museum. Gradually theplan evolved to couch our ideas into a concrete research scheme. The firstoutcome of this joint effort was the project Politics, Ritual and Religion:Cultural Formation in Early Medieval India, which was awarded a grantby the Leverhulme Trust (2011). Our work at this project inspired a morecomprehensive, cooperative ambition, resulting in the research proposalBeyond Boundaries: Religion, Region, Language and the State, which wasawarded a ‘Synergy Grant’ by the European Research Council (2013).

The present book is very much the consequence of the forces com-bined in these three programmes, the ‘Skandapuran. a Project’ (NWO),the ‘Politics, Ritual and Religion Project’ (Leverhulme) and the ‘BeyondBoundaries’ programme (ERC). I am deeply grateful to these three bodiesfor the trust placed in our research and the financial facilities put at ourdisposal, and in particular my thanks are due to my friend and colleagueMichael Willis who took the lead in applying to the latter two of thesethree institutions.

In addition to the people and institutions mentioned above, I havebeen helped at various stages by friends who read parts of this studyand from whose criticism and suggestions I have derived great benefit. Iwould particularly like to mention Alexis Sanderson (Oxford), Arlo Grif-fiths (EFEO), and Phyllis Granoff (Yale). A special word of thanks is dueto Yuko Yokochi, who not only joined me in my research during everystage of this study, contributing to it and providing many of its ideas, butwho also went meticulously through the whole manuscript, keeping mefrom many inaccuracies or worse by her Sanskrit learning.

I am also indebted to Julia Harvey who tirelessly ploughed her waythrough the English, and to Egbert Forsten, my life-long friend and pub-lisher, who continued to give valuable advice on the graphic design of thisbook, even after the Groningen Oriental Studies was taken over by BrillLeiden. I am grateful to my new publisher Brill for continuing this seriesin Egbert’s spirit.

However, the greatest debt of gratitude I owe to my wife Aafke, whoselove has never failed and thanks to whose patient support this work hasseen the light of day. This book is dedicated to her.

Hans Bakker

Table of Contents

Introduction

A tour d’horizon . . . . . . . . . . . . . . . . . . . . . . 1

Preamble . . . . . . . . . . . . . . . . . . . . . . . . . . 1The geographical and historical perimeter . . . . . . . . . . 1The cultural framework . . . . . . . . . . . . . . . . . . . 4

The Vais.n. ava world . . . . . . . . . . . . . . . . . . . 4The sphere of Goddess worship . . . . . . . . . . . . . . 5The outside world . . . . . . . . . . . . . . . . . . . . 7The heterodox religions and the Sam. khya . . . . . . . . . 8The Śaiva world . . . . . . . . . . . . . . . . . . . . . 9

The centre . . . . . . . . . . . . . . . . . . . . . . . . 10

Composition and spread of the Skandapuran.a:A possible scenario . . . . . . . . . . . . . . . . . . . . 12

I The World of PowerThe Political Landscape

from Skandagupta to Hars.avardhana’s successors

The Twilight of the Gupta Empire . . . . . . . . . . . 25

Hun. a Attacks and Inner Fragmentation . . . . . . . . . . 25Skandagupta . . . . . . . . . . . . . . . . . . . . . . 25Budhagupta . . . . . . . . . . . . . . . . . . . . . . 28Harivarman . . . . . . . . . . . . . . . . . . . . . . . 29Toraman. a . . . . . . . . . . . . . . . . . . . . . . . 30The Aulikara rulers of Daśapura, Mandasor . . . . . . . 32Yaśodharman . . . . . . . . . . . . . . . . . . . . . . 35Mihirakula . . . . . . . . . . . . . . . . . . . . . . . 37

xi

xii Contents

The Maukharis . . . . . . . . . . . . . . . . . . . . . . . 41

The Origin of the Maukhari Dynasty . . . . . . . . . . . . 41Harivarman . . . . . . . . . . . . . . . . . . . . . . . 41The Maukharis of the Gaya Region . . . . . . . . . . . 42

The Maukharis of Kanauj . . . . . . . . . . . . . . . . . 47Iśvaravarman . . . . . . . . . . . . . . . . . . . . . . 51Iśanavarman . . . . . . . . . . . . . . . . . . . . . . 53

The Sulikas . . . . . . . . . . . . . . . . . . . . . 53The Andhras . . . . . . . . . . . . . . . . . . . . . 55The Gaud. as . . . . . . . . . . . . . . . . . . . . . 57

A Period of Stability and Cultural Bloom . . . . . . . . . 62Śarvavarman and Avantivarman . . . . . . . . . . . . . 62The Kaumudımahotsava . . . . . . . . . . . . . . . . 71Viśakhadatta . . . . . . . . . . . . . . . . . . . . . . 72

A Reshuffle of Power in Madhyadeśa . . . . . . . . . . 77

The Thanesar–Kanauj Axis . . . . . . . . . . . . . . . . . 77Preamble . . . . . . . . . . . . . . . . . . . . . . . . 77Sthaneśvara . . . . . . . . . . . . . . . . . . . . . . . 77Malava . . . . . . . . . . . . . . . . . . . . . . . . . 80The Malava–Gaud. a conspiracy . . . . . . . . . . . . . 84Hars.avardhana’s consecration . . . . . . . . . . . . . . 86Śaśanka . . . . . . . . . . . . . . . . . . . . . . . . 88

The Reign of Hars.avardhana (ad 612–647) . . . . . . . 95

A Literary and a Buddhist Record . . . . . . . . . . . . . 95Ban. a’s introduction to the court . . . . . . . . . . . . . 95Hars.a’s campaigns against Gaud. a . . . . . . . . . . . . 97

The Establishment of an Empire . . . . . . . . . . . . . 102The East: Bengal and Orissa . . . . . . . . . . . . . . 102The South: Hars.a’s struggle with the Calukyas . . . . . 104A looming threat from the West . . . . . . . . . . . . 113Foreign Affairs: The empire in the international arena . . 119The ‘Arena of Charitable Offerings’ . . . . . . . . . . 123The end of an epoch . . . . . . . . . . . . . . . . . . 128The aftermath: the rise of the Gupta kings of Magadha . 130The Skandapuran. a reaches Nepal . . . . . . . . . . . . 133

Contents xiii

II The World of CultureThe Skandapuran. a

its composition and positionwithin the literary and religious landscape

The Paśupata Movement and the Skandapuran.ad Benares d . . . . . . . . . . . . . . . . . . . . . . . . . 137

The Composition and Transmission of the Skandapuran. a . 137The historical time frame . . . . . . . . . . . . . . . 137Varan. ası . . . . . . . . . . . . . . . . . . . . . . . . 138The Paśupata history as told in SPS 167 . . . . . . . . 140The Junvanı Copperplate Inscription of Mahaśivagupta . 143The Lagud. i tradition . . . . . . . . . . . . . . . . . . 145Diversity and coherence of the Paśupata movement . . . 147

The Kapalikas of Magadha . . . . . . . . . . . . . . . . 151The Mahavrata . . . . . . . . . . . . . . . . . . . . 153

The Skandapuran.a and Ban.a’s Hars.acaritad The Sarasvatı, Kuruks.etra and Thanesar d . . . . . . . . . 155

Preamble . . . . . . . . . . . . . . . . . . . . . . . . . 155

The Puran. a in Performance . . . . . . . . . . . . . . . . 156

Ban. a’s Family Tree and the Foundation of Sthaneśvara . . . 158The myth of Dadhıca . . . . . . . . . . . . . . . . . 158

The Abhis.eka of Skanda and Hars.avardhana . . . . . . . . 161Nılalohita and Kuruks.etra . . . . . . . . . . . . . . . 165Kuruks.etra . . . . . . . . . . . . . . . . . . . . . . 169

Conclusion . . . . . . . . . . . . . . . . . . . . . . . . 172

The Myth of Daks.a’s Sacrificed Hardwar and Rishikesh d . . . . . . . . . . . . . . . . . . 173

Preamble . . . . . . . . . . . . . . . . . . . . . . . . . 173

Bhadreśvara . . . . . . . . . . . . . . . . . . . . . . . . 174The initiation of the gods in the Paśupata vrata . . . . 174Bhadreśvara: the site . . . . . . . . . . . . . . . . . . 178Kubjamraka . . . . . . . . . . . . . . . . . . . . . . 181Brahmavarta . . . . . . . . . . . . . . . . . . . . . 183

Kanakhala . . . . . . . . . . . . . . . . . . . . . . . . 185

Conclusion . . . . . . . . . . . . . . . . . . . . . . . . 186

xiv Contents

The Seven Brahmins, King Udayana,and Śveta’s Struggle with Deathd Kalanjara Hill d . . . . . . . . . . . . . . . . . . . . . . 189

The Legend of the Seven Brahmins . . . . . . . . . . . . 189Historical evidence . . . . . . . . . . . . . . . . . . . 189Literary evidence . . . . . . . . . . . . . . . . . . . 192

The seven deer of Kalanjara . . . . . . . . . . . . . . 193Mr.gadhara . . . . . . . . . . . . . . . . . . . . . . 194

King Udayana . . . . . . . . . . . . . . . . . . . . . . 200The Kalacuris of Ujjain . . . . . . . . . . . . . . . . 203Udayana, Vasanta, and the carrier . . . . . . . . . . . 205

The Story of Śveta . . . . . . . . . . . . . . . . . . . . 209Śveta in the Skandapuran. a . . . . . . . . . . . . . . . 210The Kalacuris of Kalanjara . . . . . . . . . . . . . . . 213The twenty-eight incarnations . . . . . . . . . . . . . 214

The Paśupatas and Śaiva of the Nılakan. t.ha Temple Complex 215

Summary . . . . . . . . . . . . . . . . . . . . . . . . 218

The Myth of Tilottama and Man.d.aleśvarad Mun.d. eśvarı Hill d . . . . . . . . . . . . . . . . . . . . . . 221

Man. d. aleśvarasvamin and the Skandapuran. a . . . . . . . . 221The myth of Tilottama . . . . . . . . . . . . . . . . 221Mun.d. eśvarı Hill . . . . . . . . . . . . . . . . . . . . 222The inscription of Udayasena, Year 30 . . . . . . . . . 225

The Sites of Mun. d. eśvarı Hill . . . . . . . . . . . . . . . 228The Temple of Man. d. aleśvara . . . . . . . . . . . . . . 228Vinıteśvara . . . . . . . . . . . . . . . . . . . . . . 231The Tapovana . . . . . . . . . . . . . . . . . . . . . 234Narayan. adevakula . . . . . . . . . . . . . . . . . . . 235

Conclusion . . . . . . . . . . . . . . . . . . . . . . . . 236

Goddesses from the Eastd Bangarh or Kot.ivars.a d

In cooperation with Yuko Yokochi . . . . . . . . . . . . . . . 241

Introduction . . . . . . . . . . . . . . . . . . . . . . . 241

Bangarh or Kot.ivars.a: the Material Evidence . . . . . . . . 241The Damodarpur copper-plate inscriptions . . . . . . . 241The Bangarh excavations . . . . . . . . . . . . . . . 247

Contents xv

Kot.ivars.a in Sanskrit Literature . . . . . . . . . . . . . . 251Vais.n. ava sources of the Gupta age . . . . . . . . . . . 251Kot.ivars.a and the Skandapuran. a . . . . . . . . . . . . 252The Matr.tantras or the seven Yamalas . . . . . . . . . 257

Later Skandapuran. a Transmission and the Pala Kingdom . 260

Appendixd The Kot.ıvars.amahatmya of the Skandapuran. a d

Edited by Yuko Yokochi (Kyoto) . . . . . . . . . . . . . . . 263

Introduction . . . . . . . . . . . . . . . . . . . . . . . 263

Part I (S recension): KoM(S) 1–23 . . . . . . . . . . . . . 263

Part I (RA recension): KoM(RA) 1–56 . . . . . . . . . . . 265

Part II (S & RA recensions): KoM 24–44 . . . . . . . . . . 268

ReferencesFigures . . . . . . . . . . . . . . . . . . . . . . . . . . . 273

Plates . . . . . . . . . . . . . . . . . . . . . . . . . . . 273

Bibliography . . . . . . . . . . . . . . . . . . . . . . . 277

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301

Introduction

A tour d’horizon

Preamble

Fifteen years ago, in 1998, when we were working on the Prolegomenaof the first volume of our critical edition of the Skandapuran. a, we wrotethat,

In the present state of our knowledge, with so much of the text notyet critically established and carefully examined, we cannot assign avery precise date to it; we postpone a full discussion of this point toa later occasion. It must suffice to say that we are inclined to situateit between the sixth and eighth century ad. A firm terminus antequem is of course given by the date of the earliest manuscript, viz.ad 810. (SP I, 5)

Now that our edition has covered about one third of the text, the situationhas improved, though still precious little can be said with great confidence.Yet it seemed to us no longer warranted once again to ‘postpone a fulldiscussion.’ The ‘later occasion’ has arrived and the present book is meantto fulfill this long-standing desideratum, for better or for worse. Latergenerations of scholars, we have little doubt, will find much to reject andto improve, but we welcome that as the natural way of making progress inthe domain of historical knowledge. Our present endeavour is meant to bea modest contribution to this continuing process of learning, a true workof the longue duree.

The geographical and historical perimeter

The Skandapuran. a in its first redaction pertains to Northern India of thesixth and seventh centuries.

The geographical heart of this text is located between Kuruks.etra inthe west, Varan. ası in the east, the Vindhya Range in the south and Uttara-khand in the Himalaya in the north.1 This is not to say that its composers

1 Cf. Bisschop’s observation regarding the distribution of places listed in SP 167: ‘A

1

2 Hans Bakker

had no knowledge of the cultural and geographical identity of the Subcon-tinent as a whole. They were very aware that the religious movement,among whose proponents they counted themselves, had its origin in southGujarat (Karohan. a). The legendary history of the Pancartha doctrine aspropounded by Lagud. i, told in adhyaya 167, is there to prove it. Theyalso had enough information about outlying Śaiva or Paśupata centresto mention them and to include their mythology in the text. This is at-tested by the Mahatmyas of Śankukarn. eśvara at the mouth of the RiverIndus,2 Puriścandra in the Western Ghat.s,3 two Gokarn. as, a northern anda southern one,4 Śrıparvata in Andhra Pradesh,5 and Kot.ivars.a in WestBengal.6

What these latter Mahatmyas have in common, with the exception ofthe one of Kot.ıvars.a, is that they hardly, if at all, betray first-hand infor-mation about the topography. This differentiates them from, for instance,the Mahatmyas of Kuruks.etra, Bhadreśvara, or Varan. ası. The importanceof these places is acknowledged and guaranteed by connecting them withŚaiva mythology, but the text does not bear witness to a direct acquain-

few things strike me as significant. First, the absence of any place south of the Dec-can. The north-east is likewise conspicuously absent. The Himalaya, by contrast,takes up a prominent position. Śiva’s longstanding connection with mountains ingeneral, and the Himalaya in particular, is well-known. A similar picture arisesfrom the mythology of Śiva in the entire Skandapuran. a. It is especially the presentGarhwal region down to Kuruks.etra which is richly endowed with sanctuaries: pos-sibly all sanctuaries listed from Mahalaya up to Taks.akeśvara are to be located inthis area. If so, more than half of the mundane sanctuaries mentioned (thirteenout of twenty-four) belong to this area, which indicates its prominent position forearly Śaivism. Ancient Kashmir, on the other hand, is another area conspicuousby its absence’ (Bisschop 2006, 14).

2 SPBh 73.8–106. Granoff 2004, 114f. argues that, ‘a story like the one in SPBh 73represents an effort to remake a holy site, once sacred to Śankukarn. a, into a Śaivasite. [. . . ] The existence of a holy site bearing the name of Śankukarn. a furthersuggests that the Gan. a Śankukarn. a may well have been a figure, like Ghan. t.akarn. aor Naigames.in, to whom an independent cult was dedicated.’ See also below, n. 471on p. 151 and n. 534 on p. 173; Bisschop 2006, 220.

3 SPBh 122.73–88. The place is situated near the Hariścandra hilltop, modern Hariś-candragarh, 80 km north of Pune. Bisschop 2006, 203.

4 The southern one is the well-known Gokarn. a on the west coast south of Goa. Thenorthern one may refer to the Gokarn. eśvara in the Kathmandu Valley. Yokochi2013b (SP III, 105) thinks that the Mahatmya told in SP 60.22–71 may refer tothe northern Gokarn. a, but whether it is to be located in Nepal remains an openquestion.

5 SPBh 70.42–59. This hill is to be identified with Śrıśaila in the Kurnool District ofAndhra Pradesh. See Bakker 1997, 46; SP I, 107; Bisschop 2006, 201. It became amajor centre under the dynasty of the Vis.n. ukun.d. ins (below, p. 3).

6 SPBh 171.78–137. For this Mahatmya see below, p. 263 ff. The place can be iden-tified with Bangarh near the village of Gangarampur in the Dakshin DinajpurDistrict of West Bengal (India). For further discussion see below, p. 252.

Introduction 3

tanceship of the composers with the local situation.Regarding the time-frame, it can be observed that the composers were

very well acquainted with the Mahabharata,7 Harivam. śa,8 and the Pura-n. a,9 from which texts verses, whole passages, or myths are occasionallyborrowed. This makes the beginning of the sixth century a plausible ter-minus post quem.

On the other hand, the Skandapuran. a specifies a holy place in Andhra,Citraratha (SPS 167.92–94), which we know from epigraphical and archae-ological sources pertaining to the fourth and fifth centuries to have been animportant sanctuary of the Śalankayana dynasty, whose kings are said tohave been devotees of Citrarathasvamin.10 The site, identified with Vengı-pura, modern Peddavegi near Ellore, obviously went into a decline whenthe Śalankayanas were being eclipsed by the Vis.n. ukun. d. ins under KingMadhavavarman II in the second half of the fifth century, a dynasty thatdirected its devotion to the Lord of the Śrıparvata.11 The Skandapuran. ais the last text known to us that mentions Citraratha. Though this evi-dence does not contradict our proposed terminus post quem, it may pointnevertheless to a date for our text that cannot be much later than thesixth or seventh century.12

This tallies with the fact that the Skandapuran. a seems to be barelyacquainted with Tantric forms of Saivism. Here we should tread cautiously.In SP I, 4 we observed:

The background of the text must presumably be sought in circlesof lay Śaiva devotees, associated perhaps rather with Paśupatas ofsome kind than with Tantric Śaivas proper. The SP does howevershow awareness of Tantric Śaivism of some sort. See especially the

7 E.g. SPBh 164.1–3 ≈ MBh 9.43.50–52 (Skanda’s abhis.eka). See below, p. 162 andSP I, 25f.

8 E.g. SP 57.39 = HV 19.18 (the story of the seven brahmins; see below, p. 193). Cf.also the Mahatmya of Kot.ıvars.a (p. 252).

9 The reference in SP 1.9 to ‘the Puran. a’ may refer to a text basically consistingof the Puran. apancalaks.an. a (see SP I, 20–22). Our text seems to borrow directlyfrom (an early) version of the Vayupuran. a/Brahman. d. apuran. a in the Nılakan. t.hamyth—SPBh 114 ≈ VaPVenk 1.54 (Bd. P 1.2.25)—and in its concluding descriptionof the Śivapura (SPBh 183.1–40 ≈ VaPVenk 2.39.209–255, Bd. P 3.4.2.211–238). SeeBisschop 2007.

10 E.g. the Plates of Vijayadevavarman in EI IX (1907–08), 58: bhagavato cittaratha-sami-padanujjhatassa bappabhat.t.arakapadabhattassa paramamahessarassa salanka-yanassa assamedhayajino maharajasirıvijayadevavammassa vayan. ena elure mul.u-d. a-pamukho gamo bhan. itavvo | Cf. HCI III, 204–06; Bisschop 2006, 200.

11 E.g. the Ipur Plates (set I) of Madhavavarman, Year 37 (Sankaranarayanan 1977,159): bhagavacchrıparvatasvamipadanudhyatasya vis.n. ukun. d. ınam [. . . ] śrıgovinda-varmanah. putrah. [. . . ] maharajaśrımadhavavarma vijayaskandhavarat [. . . ]. Cf.Sankaranarayanan 1977, 19, 54, 141.

12 See Bisschop 2006, 14, 200.

4 Hans Bakker

intriguing reference to and list of what are called Matr.tantras (SPBh

171.127–131).

This ‘intriguing reference’ is found in the context of the Kot. ıvars.amaha-tmya. Again, Kot.ivars.a seems to be the exception. We will deal with thispassage in greater detail below (p. 257). In the present context it willsuffice to note that this reference appears to be isolated and inconclusive.

Apart from this instance, we have not been able to trace any knowledgeof texts, rituals, or ideas typical of the Mantramarga in the Skandapuran. aas represented in the Nepalese manuscripts. The Mahatmya of Kot.ıvars.a isthe last one presented in the text, contained in adhyaya 171.13 Obviouslythis Mahatmya was included in a phase in which the first redaction of thetext was about to fix the final composition. Here the geographical andreligious scope of the composers reached its maximum span.

The above observations lead us to consider the end of the seventhcentury as a plausible terminus ante quem.

The cultural framework

These two termini serve as points of departure for our reconstruction ofthe cultural framework in which the composition of the Skandapuran. a wasembedded. This is the book’s principal aim.

The Vais.n. ava world

The composers of the Skandapuran. a contributed to the rising tide of earlyŚaiva or Paśupata religion. Whether they were ascetics, acaryas, laymendevotees (laukika), or a mix, they belonged to a milieu of learned Maheśva-ras. As such they were well aware that the Vais.n. avas had been and werestill serious competitors in the pursuit of popular support and patronageof the rich and powerful.

However it seems that Vaisnavism was compromised by having beenthe offical religion of the empire and, since the fall of the Guptas, Vais.n. a-vas were in bad grace. All the rulers and dynasties that play a role in thisbook had turned to Saivism: Maitrakas, Kalacuris, Aulikaras, Maukharis,Hun. as, Vardhanas, Pan.d. uvam. śins, Śaśanka, and the Licchavis. Only the

13 SPBh 171.78–137. After this Mahatmya the text continues, in SPBh 172, with themyth of Indra clipping the wings of the mountains, Prahlada’s fight with Vis.n. u(see Granoff 2004, 128–30), after which he becomes the Sam. khya teacher Panca-śikha (SPBh 172.63; see below, p. 8); in SPBh 173 with a eulogy on the greatnessof brahmins; in SPBh 174–182 with an exposition of Paśupata yoga; and in SPBh

183, the last chapter of the text, with a description of the Śivapura and Vya-sa’s practising of Paśupata yoga, after which he sets out on a tırthayatra (SPBh

183.61–62).

Introduction 5

Later Gupta king Adityasena is a paramabhagavata; his successors, Deva-gupta and Vis.n. ugupta, were again paramamaheśvaras. The preponderanceof Saivism over Vaisnavism owed just as much or more, in my view, topolitical considerations, than to an alleged greater expediency in providingthe religious rituals that were in demand by kings and their ministers.

The intricate relationship between the Vais.n. avas and Śaivas in thisformative stage of ‘the Śaiva Age,’ was the subject of a penetrating analysisby Phyllis Granoff, who observed:

While there is no denying the element of sectarian propaganda in the earlySkandapuran. a stories [. . . ], the combined weight of the evidence from othertexts suggests that sectarian concerns are not the sole factor responsiblefor the form these stories took. [. . . ]

My hypothesis is that [the tradition] remakes itself in conformity withsomething that is happening around it in Vaisnavism. This is the restruc-turing of earlier, in some cases, cosmogonic stories, and the creation of newstories with a single pattern: the appearance of a god in the world of mor-tals and demons in order to destroy evil. It is perhaps not incidental thatone of the earliest such stories in the Śaiva corpus, the story of the defeat ofAndhaka,14 is explicitly linked to the Vais.n. ava corpus in that Andhaka isthe child of the demon Hiran. yaks.a,15 who is defeated by Vis.n. u in the boarincarnation,16 and whose brother Hiran. yakaśipu is defeated by Vis.n. u asthe Man-lion.17 I believe that the stories in the early Skandapuran. a makeclear some of the problems that both the newly emerging Vaisnavism andSaivism, which owed so much to Vais.n. ava mythological models, faced.18

The Vais.n. ava myths accommodated to the Saivism of the Skandapura-n. a occupy a comparatively large proportion of the text.19 This seems tosupport the view that Vaisnavism was a major concern in early medievalSaivism. But the attitude of the composers could not be better illustratedthan by the fact that within the entire Mahatmya of Varan. ası (SP 26–31.14), with all its detailed descriptions of bathing places and sacred spotsin the holy town, not a single Vais.n. ava tırtha is mentioned.

The sphere of Goddess worshipThe text’s attitude is different when it comes to accommodating forms ofgoddess worship to the world of the Skandapuran. a: viz. embracing ratherthan opposing.

14 The Andhaka Cycle (including many subplots) in the Skandapuran. a roughly coversSPBh 73–157. It comes to an end when Devı accepts Andhaka as her son (SPBh

157). See below, p. 69.15 SPBh 73.89–94.16 SPBh 99–108.17 SPBh 71.1–47.18 Granoff 2004, 130, 133f.19 SPBh 70–72, 75–110, 115–122.

6 Hans Bakker

Taking their starting point from Śiva’s spouse, Parvatı, the embodi-ment of the brahminical ideal of womanhood, a chaste and loyal wife, whois an affectionate mother to boot, the composers of our text envisage twomore classes of goddesses which are subsumed under, and at the same timeincluded by Parvatı, ‘Mother of the World.’ In the Vindhyavasinı Cycle,which is the subject of SP III, Parvatı sloughs off her dark skin/side, whichbecomes embodied in the goddess Kauśikı, alias Vindhyavasinı, since sheis located in the Vindhya mountains. She is the personification of theformidable, dark-skinned, virgin warrior and demon-slayer. The figure ofVindhyavasinı contains a type of deity that might have emerged from god-desses worshipped in the Vindhyas. This warrior goddess, again, is con-sidered to be the source of hosts of terrifying goddesses called ‘Mothers,’often with animal or bird’s heads, who represent a plethora of female folkdeities (SP 64.19–29).20 That we are here concerned with local goddessesfollows from SP 68.1–9, in which these goddesses are assigned by Kau-śikı to various actual countries, cities and villages, such as, for instance,Revatı to Gokarn. a, Prabha to Vatsagulma, Laks.mı to Mt Kola, S. as.t.hı toKaśmıra, Bahumam. sa to Kot.ıvars.a, and Citraghan. t.a to Varan. ası.21

Lifting up the tail of Indra’s en-emy (i.e. the buffalo demon),putting her foot on his head withforce, and taking her trident,she pierced his back and quicklyrobbed the demon of his life.22

Skandapuran. a 68.22

Plate 0Mahis. asuramardinı

Nachna, Ramvan Museum

20 See Yokochi (2013b) in SP III, 25–32.21 In SPBh 164.142–178 Kauśikı gives these Mothers (Matr.s) to Skanda on the occa-

sion of his abhis.eka. In this context more than two hundred names of goddessesare mentioned. See Yokochi 2013b (SP III, 135f.).

22 SP 68.22: udgr.hya sa valadhim indraśatroh. kr. tva ca padam. śirasi prasahya | triśu-lam adaya bibheda pr.s. t.he vyayojayac casubhir aśu daityam ‖

Introduction 7

The Skandapuran. a thus represents one of the first Sanskrit texts thatgives prominence to this type of goddess devotion, and in its systematicattempt to accommodate these female deities to the highbrow world ofŚaiva mythology it contributed significantly to the emancipation of folkreligion. This development was well underway in the sixth century. Brah-manical sites of goddess worship as attested by the artistic production oficons, especially of Mahis.asuramardinı, came to the fore in northern Indiaduring Gupta rule, and this process continued to grow in the followingcenturies.

SP 68.22 describes how Kauśikı-Vindhyavasinı slays the buffalo de-mon, and this description agrees exactly with an iconographic type of theMahis.asuramardinı icon that made its appearance by the end of the fifthcentury, a type that Yokochi has classified as ‘Vindhya subtype.’23 An im-age found in Nachna, now in the Ramvan Museum, represents the earliestknown specimen of this type (Plate 0).24 Styllistically it may be dated toaround ad 500. If we follow Yokochi 2013b, and allow about half a century‘for the dissemination of a specific icon, ad 550 would be plausible as asafe upper limit for this description’ in the Skandapuran. a.25

The emancipation of female deities initiated by the Skandapuran. a fi-nally led to a dominance of the Goddess in some forms of Tantric Saivism,in which she was conceived of as a supreme, feminine power, śakti, some-times even transcending Śiva himself. This concept of śakti, however, isstill unknown to the Skandapuran. a,26 though the start of this developmentis maybe visible in the Skandapuran. a’s reference to the Matr.tantras.27

The outside worldThe Skandapuran. a professes to be a staunchly orthodox work, which isshown, inter alia, by the nearly total neglect of anything outside the worldof Brahmanism. It does not seem to know the Hun. as, mentions the coun-tries of the Persians (Paraśıka) and Yavanas only once, viz. as places towhich the goddesses Upaka and Vayası are assigned by Kauśikı,28 and theword mleccha occurs in the text only twice, once when the demons withangry-faces are compared to Mlecchas, at the moment that they struck

23 See Yokochi 1999 and 2004b, 137–152.24 Another, somewhat later, example of the ‘Northern subtype,’ is illustrated in Plate

3 (below, p. 44).25 SP III, 57.26 In the extraordinary hymn to the Goddess in Dan.d. aka metre found in SP 32.113–

117, she is once called ‘power (śakti) of Maheśa.’ This is, as far as we have beenable to check, the only occurrence in the Skandapuran. a in which the Goddess iscalled the Śakti of Śiva.

27 Above, p. 4 and below, p. 257.28 SP 68.4: vatsagulme prabham. devım. laks.mım. kolagirav api | upakam. paraśıkes.u

vayasım. yavanes.u ca ‖ 4 ‖

8 Hans Bakker

down with their axes, lances, swords and hatchets the elder brother ofSkanda (i.e. Vinayaka),29 and a second time, when it mentions a rebirth,after a stay in hell, among the Mlecchas where one lives like an immortal,a rebirth that is the result of a donation of food in a preceding life, evenif that food was impure and not consecrated.30

The heterodox religions and the Sam. khya

The heterodox religions meet with similar neglect. Only once, apparently,when the text sums up the people who will go to the Mahapadma Hell,does it refer to heretic mendicants, with whom probably the Buddhists aremeant, Nirgranthas and Ajıvakas and the brahmins who support them.31

In a similar context, the followers of the false doctrine of Kapila, i.e. theSam. khyas, are condemned, but here the situation is more complex.32

Although Paśupata Saivism considers the Sam. khya philosophy as suchinferior, it does not usually condemn it since its doctrine is acknowledgedas the basis of its own system. This is explicitly the case in the Skanda-puran. a chapters that deal with the Paśupata yoga (SPBh 174–181). More-over, SPBh 172.59–66 tells us that the Daitya Prahlada turns to Asuri,the pupil of Kapila, and becomes the Sam. khya acarya Pancaśikha, who,thanks to the Sam. khyasiddhanta, transcends the threefold suffering andreaches brahman.33 The declaration of the Sam. khya as a false doctrine

29 SPBh 136.19: jambhe ’vasanne ’ribale prasanne, s.ad. danavah. s.an. mukhapurvajam.tam | kut.haraśulasiparaśvadhais tu, mleccha yatha kruddhamukhas tataks.uh. ‖

30 SPBh 111.16cd–19cd:avadhutam avajnatam asatkr. tam athapi ca |annam. pradadyad yo devyo yadr.śam. tadr.śam. śubhah. |tasyapi narake ghore yatyamanasya raks.asaih. ‖upatis. t.heta bahudha tr.ptir yenasya jayate |yadi manus.yam ayati kadacit sa narah. punah. |mlecches.u bhogı bhavati ramate ca yathamarah. |annam evam. * viśis. t.am. hi tasmad na paramam. bhuvi ‖

* Bhat.t.araı’s edition reads annemevam. .31 SP 45.11: dus.t.aśravan. anirgranthas tathaivajıvakaś ca ye | bhaktas tebhyo dvijo yaś

ca narakam. samprapadyate ‖32 SP 48.10: śastran. i harate yaś ca goghno yaś ca narah. smr.tah. | kapilaś ca vr. thadr.s. t.ir

vedadharmarthadus.akah. | nastikaś ca duratmana ity ete tam. prayanti vai ‖33 SPBh 172.59–62:

param. tattvam athanves.t.um. gun. atıtam ajam. śivam |saks. ad bhagavatah. śis.yam. kapilasyasurim. munim ‖ 59 ‖śis.yatvenopasam. gamya moks.avidyam avaptavan |prakr. tınam. param. tattvam. vikaran. am. param. tatha ‖ 60 ‖ks.etrajnam. purus.am. jnatva bhoktaram. bhoktr.varjitam |moks.avidyapararthajnah. sam. khyasiddhantaparagah. ‖ 61 ‖siddhah. pancaśikho namna so ’bhavad munisattama |sa yogam. paramam. labdhva sam. khyacaryo mahamunih. |duh. khatrayavinirmuktah. praptavan padam avyayam ‖ 62 ‖

Introduction 9

seems, therefore, to be out of place, and may be seen as an indicationthat the chapters on the thirteen hells (SP 37–51) have a non-Paśupatabackground.34

The Śaiva worldFascination with hell was to become a regular feature of texts belong-ing to the Śaiva Siddhanta corpus, in which the concept of 32 basic hellsprevails.35 The Skandapuran. a’s representation is closer to the hell descrip-tions that are found in the Śivadharma corpus, a collection of texts thatwas intended for the Maheśvaras at large, and of which the nucleus isgenerally conceived of as predating the Śaiva Siddhanta.36

In addition to the number of hells, viz. thirteen, the hell conception ofthe Skandapuran. a distinguishes itself from all similar descriptions in oneother important aspect. This is the concept of ‘elevations’ (ucchraya). Thisconcept, explained in SP 37.9, is not known from Śaiva literature otherthan the Skandapuran. a, but it is strikingly similar to that of the utsada

Cf. Sam. khyakarika 70: etat pavitram agryam. munir [i.e. Kapila] asuraye ’nukampa-ya pradadau | asurir api pancaśikhaya tena ca bahudha kr. tam. tantram ‖ 70 ‖Larson 1987, 107f. formulates a scholarly consensus to the effect that ‘by the sixth-century of the Common Era (the approximate date of the Chinese translation) andthereafter, the writers of Sam. khya texts (a) identified Kapila as the founder of thesystem, (b) recognized Asuri as someone who inherited the teaching, (c) consideredPancaśikha as someone who further formulated the system (tantra) and widelydisseminated it, and (d) described Iśvarakr.s.n. a as someone who summarized andsimplified the old system after an interval of some centuries (Paramartha’s Chinesetranslation) or more (Yuktidıpika).’

34 See the Introduction to SP II B, pp. 9–10.35 Sanderson 2003–04, 422; Goodall 2004, 282f. n. 490.36 Sanderson forthcoming, 8f: ‘The Śivadharma and the Śivadharmottara were pro-

duced when initiatory Śaivism was restricted to ascetics or at least in the context ofthat form of Śaivism. That the Śivadharmottara was written in the context of theearlier ascetic Śaivism is evident from its instruction (2.147–50) that there shouldbe a shrine next to a monastery’s hall for public instruction in which one shouldinstall an image of Śiva as Lakulıśa. [. . . ] But later texts of this corpus, such asthe Śivadharmasam. graha and the Vr.s.asarasam. graha, show awareness of the laterform of the religion, in which Śaivism had diversified through its expansion amonghouseholders. Thus the Śivadharmasam. graha shows evidence of dependence onthe Niśvasamukha and the Niśvasaguhya.’ The latter two texts, belonging to thecorpus of the Niśvasatattvasam. hita, are generally seen as early representatives ofthe Mantramarga, initiating a development that would lead to the mature ŚaivaSiddhanta. Sanderson 2013, 234f., following Goodall and Isaacson (2007, 6), datesthis corpus very early indeed, assigning a date of c. 450–550 to the Niśvasamula,the oldest text in the corpus. Since this dating is mainly based on palaeography,and developments therein are difficult to pin-point within an interval of a hundredyears (cf. Sanderson 2013, 221: ‘notoriously unreliable’), we have serious doubtsabout this early date and find it more likely that the corpus began to evolve onecentury later, that is more or less simultaneously with the Skandapuran. a (cf. below,n. 469 on p. 150).

10 Hans Bakker

(Pali: ussada) found in (later) Buddhist literature on hells. It signifies asort of supplementary hell belonging to a major hell (mahanaraka), inwhich the sinner finds some temporary relief.37 Within the Śaiva contextit makes an archaic impression, which, on the one hand, may point toa certain indirect Buddhist influence,38 and, on the other, may be anindication that the description of the thirteen hells in the Skandapuran. ais earlier than related descriptions in the Śivadharma corpus.

Research on this corpus is still in its infancy. Although, admittedly, theŚivadharma seems to have originated in an environment that may havebeen close to the milieu that produced the Skandapuran. a and would merita thorough investigation, we have to limit ourselves in the present contextto a brief mention of one more issue, viz. the arrangement of forty Śaivaholy places into five sets of eight ayatanas, the Pancas.t.aka. This idea hasbeen dealt with by Sanderson, Goodall and Bisschop. Sanderson believesthat this classification ‘was already current when the first scriptures ofthe Mantramarga came into existence, which is not likely to be later thanthe sixth century,’ and Goodall observed that these five sets of eight form‘an earlier, not exclusively Tantric structure.’39 As pointed out by Bis-schop, it ‘is especially significant that the Śivadharma corpus also liststhese five ogdoads (as.t.akas) and mentions no other,’ as is common inlater Śaiva scriptures. He further observed that the Skandapuran. a doesnot yet seem to know this classification of forty places, but that ‘aboutthirty of them are mentioned in the ayatana list of the Skandapuran. a.’40

Like its description of the thirteen hells, the list of ayatanas in SPS 167thus suggests that our text is either more conservative, which would agreewith the Puranic genre, or is slightly earlier than the Śivadharma corpus.In both cases a sixth century date would match the evidence.

The centreA central aim of the Skandapuran. a was to sanctify the landscape of North-ern India. Many of the myths, regardless of over how many chapters theyextend, are geared towards a final section, in which what has been told isrelated to a specific locality. The Andhaka myth cycle appears to be anexception to this rule. Nevertheless, with some exaggeration, it could bemaintained that the Puran. a is to a large extent a sequence of Mahatmyas,whose purpose is to legitimize and lend authority to local communities ofthe Śaiva movement in the sixth and seventh centuries, thus supporting

37 SP 37.9. For the Buddhist utsada see Edgerton’s BHSD s.v.; Kirfel 1967, 201f.;Van Put 2007, 205ff.

38 See the Introduction to SP II B, p. 9.39 Sanderson 2003–04, 405f.; Goodall 2004, 315.40 Bisschop 2006, 27. For a thorough investigation of the Pancas.t.aka in the Śaiva

scriptures see Bisschop 2006, 27–34.

Introduction 11

them in their rivalry with other religious groups. It does so by declaringthe transcendent quality of a particular site, of which the intrinsically di-vine and soteriological value is imparted by representing it as the scene ofprimordial events.

Having stated this much, the reader may ask, why Skandapuran. a, whynot just a Maheśvara or Mahadeva Puran. a? The answer is not easily given.Actually, Skanda does not play a major role in this Puran. a, nor is heconceived of as an object of a special cult besides as a son of Śiva.41 Hisbirth and deeds are only reached at the very end of the text, contained inchapters SPBh 163–65, and 171.42 Apart from occasional references to hisassociation with the Matr.s, Skanda in this Puran. a is first and foremost aSenapati, the god of war. As such he outclasses Indra, who is cut down tosize in SPBh 171.1–77.

The Skandapuran. a may thus be seen as articulating a change of per-spective in a time that was riven by war. It mirrors these conflicts in themany chapters in which it describes in detail the recurrent battles be-tween Devas and Asuras. In addition to the significant role given in themto the warrior goddess Kauśikı Vindhyavasinı (SP 61–66), who as suchmakes her first appearance in the Sanskrit literature, it is the young armygeneral Skanda, Śiva’s son, rather than the old king of the gods, Indra,who should be held in high regard. Or in the blustery words of Skandahimself: ‘Brahma and the rest, Siddhas, and those who know the Vedas,they know who I am; it is immaterial whether you know (who I am, ornot), and for what reasons should I pay my respects to you, O Śakra, youwhose kingship is obtained by tapas and who will soon go down?’43

The centre in which and from which the world is viewed in this way isVaran. ası. The Mahatmya of this holy town is by far the longest and mostdetailed in the Puran. a. Varan. ası is the town where Śiva created by hismind his own divine palace (vimana) of matchless beauty, as well as apleasure grove, more beautiful than any other divine garden (SP 26.64–

41 Yokochi SP III, 24f. n. 45: ‘For a historical study of various characteristics of Ska-nda, see Mann 2001, 2007 and 2012. There are, however, significant differences.According to Mann [2012], the rise of Skanda in mainstream Hinduism as the godof war and Śiva’s son caused the decline of Skanda’s cult. In the case of Vindhya-vasinı, on the other hand, her evolution into the Warrior Goddess was the keyfactor in her establishment in mainstream Hinduism. The decline of Skanda’s cultin this respect may have been caused by the rise of the Warrior Goddess, by whomSkanda was replaced in the function of war god.’

42 The narrative structure required that the birth of Skanda is anticipated in SPBh

72. For a summary of these chapters see Mann 2012, 187–95.43 SPBh 171.7cd–8cd: brahmadya mam. vijananti siddha vedavidaś ca ye | tapasa la-

bdharajyena na cirad astam es.yata | kim. tvaya janata śakra karyam. kim. pujitenava ‖ 8 ‖

12 Hans Bakker

67). In the words of the myth-maker:

This Avimuktaks.etra is holier than other holy places where Devaresides, such as Mahalaya, Kedara, and Madhyama; this is becausethere one only attains gan. a-hood, whereas here one attains finalrelease, which is superior to the state of being a Gan. apati. Eversince the worlds were created, Deva has not left (mukta) it, hence itsname ‘Avimukta.’ When one sees the Avimukteśvara linga, one willbe freed at once from all the fetters of bound souls (paśupaśa).44

The history of Varan. ası has been investigated in the Introduction to Vol-ume II A of our edition of the Skandapuran. a. There it was tentatively con-cluded, after collating all available evidence, that the composition of theMahatmya took place in ‘the 6th or, maybe, first half of the 7th century,’and that the first redaction of the text as a whole was finished ‘under therule of either the Maukharis or Hars.avardhana of Kanauj.’45 This conclu-sion has been roughly confirmed in the following years of research, whichhave resulted in the present study, although at this stage, more than inthe past, we take into consideration that the process of composition andredaction of the text may have extended over a considerable period—theduration of a man’s working life, say about thirty to fifty years.

This is how far historical research can bring us to date. As remarkedabove, precious little can be said with great confidence. The hypothesisadvanced in this book is merely an informed guess. It leaves various sce-narios open. However, in order to satisfy our desire for a palpable resultand a real picture, I have worked out one scenario in more detail andcouched it in an apodictic style. I hope that the reader can appreciate itfor what it is and will not lose sight of its highly speculative nature.

Composition and spread of the Skandapuran.a:A possible scenario

When Avantivarman ascended the Maukhari throne in Kanyakubja in thelast quarter of the sixth century, it may have appeared as if the old daysof stability and prosperity had returned to Madhyadeśa. Thanks to theincessant war efforts of his grandfather, Iśanavarman, the cruel intruderscalled Hun. as had been driven back to the foothills of the western Hima-layas after a long and devastating period of war. A close friendship haddeveloped between the rulers in Kanyakubja and Sthaneśvara, where the

44 Synoptic paraphrase of SP 29.53–57 (SP II A, 129f., 214).45 SP II A, 52.

Introduction 13

dynasty of the Vardhanas guarded the western part of the kingdom. Theeastern enemies, the Gaud. as and their allies the Guptas, had been forcedto take refuge at the borders of the ocean, where they were being kept incheck by Kanyakubja’s powerful southern allies in Daks.in. a Kosala, whotraced their respectable pedigree straight back to Pan.d.u and his mightyson Arjuna.

The ancient land of the Buddha and the cradle of empire was firmly un-der control. Avantivarman proudly bore the title ‘sovereign of Magadha.’ ABuddhist settlement there was developing into a place of learning of highinternational repute. The university of Nalanda attracted students andscholars from all over India and abroad, and the Maukhari king, thoughnot a Buddhist, prided himself on being its chancellor.

The monarch watched over the Bull of the Dharma, which was shep-herded by his countrymen. The Bull, shown on the royal seal, had inrecent years become a forceful emblem, a symbol appropriated by anotherreligion, one to which the Maukharis had confessed ever since they hadthrown off the yoke of the Imperial Guptas with their state deity Vis.n. u.46

Worship of Śiva had opened up new avenues for the imagination and en-shrined royal authority in burgeoning forms of early Tantric Hinduism.

Though familiar with all sorts of asceticism, Northern India in thesixth century saw a new type of strange sadhus travelling around, whosmeared themselves with ashes and imitated the god of their devotion,Śiva Paśupati. A lineage of gurus pertaining to this movement had settledin Kanyakubja, an establishment founded in the capital by a saint fromKuruks.etra, the ancient battlefield, now firmly under the control of thefriendly Vardhanas or Pus.yabhutis, who themselves had become staunchfollowers of this type of religion. Avantivarman, too, was well disposedtowards them and invited some of them to his court.

The Paśupatas, as these Śiva worshippers were called, made good useof the patronage that fell to their lot. They set up religious centres (stha-na), temples (ayatana) and monasteries (mat.ha) in the country’s mosthallowed places, such as the Kapalasthana in Kuruks.etra, Bhadreśvaranear Gangadvara, the great Deva temple, ayatana, in Prayaga, and thesiddhasthana, ‘home of the saints,’ called Madhyameśvara, circa one kilo-meter north of the renowned cremation grounds of Avimukta or Varan. ası.

A network of itinerant sadhus connected these centres, which becamewell integrated with the local religious infrastructure and developed into

46 Cf. Dundas 2006, 390: ‘It might well be argued that a standard feature of consol-idation of imperial state formations of any variety at any place and time is thefostering of quasi-imperial religions whose totalizing claims and concomitant sym-bolic discourse connected to kingly authority reciprocally bolster the ideology ofsecular power with its ambitions of control and expansion.’

14 Hans Bakker

junctions within a fabric of yogins and religious teachers. The Paśupa-tas had had a good look at their Buddhist counterparts and had copiedtheir formula for success, namely a standing organisation of professionalreligious specialists—yogins, ascetics, and acaryas—supported by a fol-lowing of ordinary devotees, the Maheśvara community at large, to whosespiritual needs it catered.47 One of the peculiar facilities offered to thecommunity of laukikas, by at least some of these Paśupata ascetics, wasto extend services in and around the cremation grounds. Living in thecremation ground was a highly acclaimed strategy within Paśupata as-ceticism. Mahakala in Ujjain, Avimukteśvara in Benares, Paśupatinathain Nepal, to mention just the best known, were run by Paśupatas andbecame key to their success.

Avantivarman, therefore, acted in tune with the spirit of his time whenhe supported the movement. Earlier his uncle Suryavarman had spentlarge sums on the rebuilding of a dilapidated temple of the ‘Foe of Andha-ka,’ whose images were beginning to appear around this time. The princehad hired a poet to sing the praises of the god as well as of himself,chiselled into stone, for everyone to read:

May that figure of Andhaka’s Foe, on whose body snakes glimmer,offer you a stable abode—a figure who wears a lion skin that isslightly crimsoned by the light of the jewel in the hood of the serpent[that is his sacred thread], and who reddens the white line of skullsthat is the chaplet by the radiance from his third eye, and who bearson his crest the slender, darkness dispelling digit of the moon.

He (the prince) had youth that was beautiful like the waxing moonand dear to all the world; he was at peace and his mind was devotedto reflection on the branches of learning; he had mastered fully (all)the arts; it was as if Laks.mı (fortune), Kırti (fame) and Sarasvatı(learning), among others, vied with one another for his patronage:in the world, women in love experience the feeling (of love) all themore, if their lover is beloved.48

Poets were held in high esteem and Avantivarman invited them to hiscourt. Imagine the glamorous world in which plays like the Mudraraks.asawere staged, attended by the playwright Viśakhadatta in person, or theKaumudımahotsava, to mention another play, in which the entrance ofthe ruler himself is announced:

A son of the House of Magadha has arrived, thronged by hundredsof eminent ministers, like the moon enhanced by an aureole of stars,

47 Cf. Brancaccio 2011, 154f.48 For the Sanskrit text of these two verses see n. 202 on p. 70 and n. 203 on p. 71.

Introduction 15

that prince, who is a feast for the eyes of his delighted subjects.49

This is the world in which Sanskrit flourished, the world in which the kaviBhatsu, Ban. a’s respected teacher, was honoured by crowned heads. Thiscourt was sustained by the inhabitants of Kanyakubja who, in the words ofthe famous Chinese pilgrim Xuanzang, were ‘honest and sincere, noble andgracious in appearance, clothed in ornamented and bright-shining fabrics,’inhabitants who ‘applied themselves much to learning, and in their travelswere very much given to discussion on religious subjects, whereas the fameof their pure language was far spread.’50 To this court the leading figuresof the community of Maheśvaras were also welcomed. Sanskrit was theirlanguage and in Sanskrit they composed their learned treatises and wittymythology.

For learned treatises and religious expositions the educated classes ofnorthern India looked to Varan. ası. This trading town on the River Gangeshad emerged under the Guptas in the preceding century as a centre of tra-ditional Hindu learning. The arrival of the Paśupata movement added toits reputation for holiness, whereas the collective Sanskrit learning of thetown added to the literary achievements of the Maheśvaras.

The composition of the two classic Sanskrit epics was closed by thefourth century. Some of the new religious ideas concerning the god Śivahad still made it into the latest layers of the Mahabharata. After theRamayan. a had been completed, mythology related to the tutelary deityof the Gupta Empire, Vis.n. u, and his popular manifestation of Kr.s.n. a inparticular, had found expression in an Appendix to the great epic, theHarivam. śa, as well as in a new type of Sanskrit text styled ‘Ancient Lore,’i.e. Puran. a.

The Puran. a as a literary genre in its first stages of development dealtwith the creation of the universe, the origin of the world and its royal dy-nasties. However, as for instance the Vis.n. upuran. a had shown, the genrealso lent itself perfectly to the circulation of popular, religious and mytho-logical material. After the civilized world had recovered from a period ofdevastating wars and invasions, and now that Vis.n. u had ceded his placeof prominence to Śiva, the sixth century embraced a new form of devo-tion. The time had come to collect the mythology of the Great God. Inthe words of the Skandapuran. a: ‘having heard the story of Bharata aswell as the Ancient Lore, we wish to hear about the birth of Śiva’s son,Karttikeya.’51

49 For the Sanskrit text of this verse see n. 210 on p. 72.50 Si-yu-ki I, 206f.51 See below, p. 124.

16 Hans Bakker

A senior brahmin member of the Maheśvara community in Varan. ası,well-versed in Sanskrit literature, an expert on the epic tradition, initiatedin the Paśupata sacred texts, in short, a brahmin with great prestigeamong his fellow believers, charismatic and dynamic, that man, let us callhim the Suta, took the initiative to fulfill this wish and to compose aPuran. a text that would do justice to the rich mythology of Śiva and hisfamily, that would be accessible to the whole community, and, last but notleast, that would validate local claims of the sanctity of holy places andreligious establishments by telling their Mahatmyas. In order to possessthis authority, the text should be in the anonymous, pseudohistoric style ofthe Puran. a, reportedly spoken by a sage of yore with intimate knowledgeof the Great God’s own thoughts and deeds.

It happened in the days of Avantivarman’s reign that a group of kin-dred spirits and literary talents convened in an institution of the commu-nity in Benares. They discussed the plan and pledged their commitment.The Suta, the editor-in-chief, began his composition in Śloka verses, whilean inventory was being agreed on of the myths, stories, topics and placesthat had to be treated in the course of the work, a narrative that wasdesigned to lead to the birth, consecration and heroic deeds of Karttike-ya, but should not reach that point before an extensive cycle of Andhakamyths had been told first—Andhaka who, like the Mleccha foes of theMaukharis, could not be defeated until after an endless series of battles.52

The materials were arranged in a preliminary order in versified form.This inventory or blueprint, Anukraman. ika as it was called, has survivedand come down to us in the second adhyaya of the Skandapuran. a. Theeditor-in-chief was assisted by some editors who were assigned specific por-tions of the composition. The Paśupata network was called in to assembleinformation about places sacred to the Maheśvara community. Sometimesthis resulted in new collaborators entering the group, bringing in localknowledge couched in Mahatmya-style texts, sometimes the editor him-self used the information to compose the story. Rarely were ready-madetexts taken from existing literature. The Suta guaranteed the unity ofliterary style and the quality of the Sanskrit, but this could not preventminor differences remaining. He also took great care that the arrangementof stories, the complex narrative structure of the text, remained consis-tent and logical. However, soon it appeared that the original blueprintcould not be implemented except in broad outline; the myths and storiescomposed had too powerful a dynamic of their own to link up with eachother perfectly. Here the genius of the Suta was most needed and he did

52 SPBh 130–56. Andhaka becomes Śiva’s devotee—his gan. a and adopted son, a myththat was probably narrated for the first time in our text.

Introduction 17

a brilliant job.53

The Paśupata network was strongest along the east-west axis, Vara-n. ası–Kanyakubja–Kuruks.etra. It had been decided to begin in the west,since it was one of the underlying aims of the work to cover, or rather torecover the entire landscape of northern India, transforming it into sacredspace, a landscape on which the deeds of the Great God and his entouragehad bestowed holiness at the beginning of time. The work was well underway—the myths relating to Kuruks.etra and the Sarasvatı, Sthaneśvara,Bhadreśvara and Kanakhala, and Varan. ası itself had been composed, andthe Vindhyavasinı Cycle was drafted—when political reality threatened todisrupt the literary activity. A joint attack from the east and the south-west brought to an end the rule of the Maukharis, just when Grahavarmanhad succeeded his father Avantivarman, while that of its allies in Thane-sar was shaking on its foundations. For a while Benares came under thecontrol of the easterners, the Gaud. as.

A young prince, a kumara from Sthaneśvara, installed as chief of the armyon the banks of the Sarasvatı, as it were the embodiment of Skanda himself,came to the rescue of the kingdom of Kanyakubja. In a war that lastedseveral years, Hars.avardhana succeeded in pushing the Gaud. as under theirking ‘Moon,’ Śaśanka, back across the rivers Son. a and Gan.d. akı.

In about ad 606, the political situation had stabilized enough toorganize a magnificent royal coronation ceremony. Hars.avardhana wasenthroned in Kanyakubja. It would take Hars.a six more years, however,to consolidate his sovereignty over the combined hereditary lands of theVardhanas and Maukharis, including Magadha, and before finally, toparaphrase the closing metaphor of Hars.a’s Deeds sung by the greatestwriter of the time, Ban. a—after a day of bloody contest, at the fall ofnight, while the sinking sun crimsoned the sky and waters of the ocean,the Fame of his House, the Glory of his Rule, and the Force of his Destinyunited to hand over to him a pale-looking Moon.54

Varan. ası was back in the kingdom of Kanyakubja, but the new politicalsituation had an effect on the perspective and scope of the composition inprogress.

To begin with, the historical consecration of a young prince (kumara)on the banks of the Sarasvatı to lead an army against the Gaud. a king‘Moon’ (śaśanka), reflected the mythology of Skanda, the main subject ofthe Puran. a—Skanda, the god of war, who, after his consecration as sena-

53 An illustration of this intricate process is the inclusion of the legend of the sevenbrahmins into the Vindhyavasinı Cycle, for which see Yokochi’s Introduction SPIII, 15–22.

54 For the Sanskrit text see below, n. 270 on p. 93.

18 Hans Bakker

pati on the banks of the same river, led the Devas against the Asuras inorder to destroy the demon king ‘Star’ (taraka).55 The composers decidedto bring their work to the attention of King Hars.avardhana, solicitinghis blessing. After all, Hars.a himself confessed to be a paramamaheśvara,and his court offered a venue to the most promising literary men of thecountry, among whom was the king himself.

Secondly, the king’s military successes against Gaud. a called atten-tion to the east, bringing a Śaiva settlement in western Gaud. a within thepurview of the composers. The Suta, or his successor, made the decisionto conlude the sanctification of the sacred landscape of northern India inKot.ivars.a, an important commercial and religious centre in the provinceof Pun.d. ra, which was situated 80 km north-east of the army camp of KingHars.a on the Lower Ganges, the camp where the king would eventuallymeet the Chinese pilgrim. The concluding chapters of the Puran. a werereserved for philosophy and an exposition of Paśupata yoga, which, alongwith devotion and pilgrimage, would bring the Maheśvara, yogin and lay-man alike, to paradise, the City of Śiva at the top of the universe (SPBh

183), while the accomplished yogin may unite with Śiva (SPBh 182).

The day arrived when the composition of the Puran. a was concluded andthe text could be copied into a carefully prepared book, a pustaka, thatcould be offered to the Great God and donated to the king and the com-munity of the Maheśvaras. As usual when a work of such magnitude wascompleted, a solemn occasion had to be found when parts of the work couldbe recited and the book could be consecrated and ritually entrusted to atemple. Such an occasion was King Hars.a’s ‘arena of charitable offerings,’a spectacular event that was staged every five years at the confluence ofthe Ganga and Yamuna. The great ayatana or temple of Deva there wouldbe an excellent repository.

The permission was obtained. In the middle of Hars.a’s reign, when hewas at the pinnacle of power, a great assembly of feudatories, Śraman. as,and Brahmins, convened in Kanyakubja around the beginning of the NewYear, in preparation of the quinquennial event. The procession to Prayagaand the festivities there were part of the Festival of Spring in the month ofCaitra. The king rode on his magnificent elephant Darpaśata towards Pra-yaga, scattering pearls and other riches, while dressed as Indra. His mobilecourt offered splendid opportunities for staging theatrical productions,first and foremost, of course, those of his own. The Sutradhara in theRatnavalı and Priyadarśika introduces Hars.a’s plays:

Today, on the occasion of the Spring Festival, I have been respect-fully called by the assembly of kings, which has convened from all

55 SPBh 163–65. Below, p. 161f.

Introduction 19

quarters of the world, and which is subservient to the lotus-feetof King Śrı-Hars.adeva. I have been addressed as follows: ‘We haveheard by hearsay that a play entitled Ratnavalı, which is embellishedby an unprecedented arrangement of the material, was composed byour lord Śrı-Hars.a, but we have not yet seen it performed.’56

The play turned out to be a great success, and would stand the test oftime. But Hars.avardhana was too great a king to hear only his own voice.A date for the first recitation of the Skandapuran. a was agreed on. TheSuta and his team were offered their platform. In order to sustain theillusion of its being a work from time immemorial, an essential featureof the genre of Ancient Lore, a professional reader, a pustakavacaka, wasasked to recite it. The first presentation of the work went ahead before anaudience including the king, courtiers, sadhus, monks, literati of all sorts,pandits and a selection of educated Maheśvaras. It was a great tamaśa,going by the consolidated words of the Suta and his fellow kavi, Ban. a:

The sages, assembling in Prayaga to bathe in the confluence of the Gangaand Yamuna on the day of full moon, see the Singer of Ancient Lore comingtowards them to pay his respects.

Dressed in white silk made in Pun.d. ra, his forehead marked by a tilaka con-sisting of lines of orpiment on a white clay coating, his topknot ornamentedby a small bunch of flowers, his lips reddened by betel, and his eyes beau-tified by lines of collyrium, he takes his seat and begins his performance.

He pauses for a moment before he places, on a desk made of reed stalksthat is put in front of him, a pustaka, which, although its wrapping hasbeen removed by that time, is still wrapped, as it were, in the halo of hisnails, which shine softly like the fibres of a lotus.

They ask him about the birth of Karttikeya, a story that equals the Maha-bharata and surpasses the Puran. a, both of which he had recited in theNaimis.a forest on the occasion of a brahmasattra.

Then, while he assigns two places behind him to two flautists, Madhukara,‘the bee,’ and Paravata, ‘the turtle-dove,’ his close associates, he turns overthe frontispiece, takes a small bundle of folios, and announces the story ofthe birth of Skanda, of his friendliness towards brahmins, his glory and hisheroism, greater than that of the gods.

By his chanting he enchants the hearts of his audience with sweet intona-tions, evoking as it were, the tinkling of the anklets of Sarasvatı, as shepresents herself in his mouth, while it seems as if, by the sparkling of histeeth, he whitewashes the ink-stained syllables and worships the book withshowers of white flowers.57

56 For the Sanskrit text see below, n. 393 on p. 126.57 This is a coalescence of two passages, SP 1.4–13 and HC 3 p. 137f. For the separate

texts see p. 124 and p. 156.

20 Hans Bakker

The performance received favourable reactions. After having done theirritual duties, attended the great potlatch ceremony at the confluence, andpaid obeisance to the Great God in his temple and the king in his court,the Suta and his entourage returned to Varan. ası. More copies of the bookwere produced. Small emendations were made and the first transcriber’sfaults slipped in. The different versions of the text were born.

The subsequent transmission and distribution of the Puran. a over var-ious centres of the Maheśvaras added more flaws. The copying took placein focal points of Sanskrit learning, to the west and the east of Varan. ası.In Magadha, some Paśupata acaryas were not entirely satisfied with thetext. They missed in particular an account of the Lakulıśa tradition intheir own country, and, in general, they felt that the holy places in theeast and in the north, in Magadha, Orissa and Nepal, had not been donejustice. They decided to amend this shortcoming by inserting an addi-tional list of tırthas in an adhyaya that appeared to be the right place forit.58

While these processes were underway, the political situation in Indiachanged dramatically. What a few years earlier had still seemed far awayor downright impossible, happened. Hars.a’s empire collapsed. Chaosprevailed all over northern India, whereas the north-east was confrontedby a military invasion from Nepal and Tibet.

Magadha was the first country in which order was restored under theauthority of the dynasty of the Later Guptas. The daughter of Adityaguptamarried the Maukhari prince Bhogavarman, a wise move, contributingsignificantly to political stability. And while the kingdoms of Kanyakubja,Pun.d. ra and Kamarupa were still in disarray, the Gupta House of Maga-dha consolidated its power further by re-establishing good relations withits northern neighbour, the Licchavi kingdom of Nepal. A daughter bornof the marriage with the Maukhari prince, Vatsadevı, was married off tothe Licchavi king Śivadeva.

During the last two decades of the seventh century, relations with Ne-pal became close and cultural exchange between the two countries intensi-fied. Paśupata yogins and acaryas wandered from Magadha into Nepal tovisit the great shrine of Paśupatinatha, which had developed into a statesanctuary and received substantial financial support from Vatsadevı andher Nepalese husband. The priesthood of this temple was firmly in thehands of a local branch of Paśupatas. They were happy with the growingreputation of their temple. It brought them pilgrims from afar and theircoffers filled accordingly. At the same time the intensive traffic kept them

58 SPS 167.163–87. Below, p. 137.

Introduction 21

up-to-date with new religious developments and informed about the latestliterary productions.

Thus the reputation of the Skandapuran. a spread to Nepal, and friendsin Magadha were asked for a copy. They brought one, naturally a manu-script that contained the insertion mentioning Paśupatinatha in Nepala.The new acquisition was treasured. In order to preserve the text, themanuscript was copied in the century that followed. And so it happenedthat on the twelfth day of the bright half of the month of Caitra in theyear 234 (= ad 810/11) a scribe in Nepal could complete his work onthe Skandapuran. a, a labour that he had undertaken for the sake of theperfection of all beings. It would become our manuscript S1. And, if it hasnot contributed to our perfection, we are certainly the only ones to blame.

I

The World of Power

The Political Landscapefrom Skandagupta to Hars.avardhana’s successors

The Twilight of the Gupta Empire

Hun.a Attacks and Inner Fragmentation

Skandagupta

The heyday of the Gupta Empire was over when, in ad 456, Skanda-gupta was able to proclaim himself emperor after a protracted war ofsuccession, a struggle which may have lasted for about ten years. As usual,the triumph is expressed through religious metaphor, in which the emperoris compared to the king of the gods, Indra, who is confirmed in his royaloffice by the dynasty’s supreme, tutelary deity Vis.n. u.59

The inscription testifies to the fact that Skandagupta’s authority stillstretched from coast to coast, and that even in the far-off western provinceof Saurashtra he was able to appoint a governor, called Parn. adatta, whowas still willing to accept the emperor’s overlordship.60 However, the sameinscription also obliquely refers to the dangers that had nearly preventedhis accession to power: the Nagas and the Mlecchas.61 The Nagas may have

59 Skandagupta’s Junagar.h Rock Inscription (CII III (1888), 58f. v. 1; Sircar SI I,308):

śriyam abhimatabhogyam. naikakalapanıtam. ,tridaśapatisukhartham. yo baler ajahara |kamalanilayanayah. śaśvatam. dhama laks.myah. ,sa jayati vijitartir vis.n. ur atyantajis.n. uh. ‖‘Vis.n. u, whose triumph is supreme, who has triumphed over pain, the perpet-ual abode of Laks.mı, who is alighting on (her) lotus throne—He is (truly) tri-umphant, when He takes Śrı, i.e. Royal Glory, away from (the demon king) Bali,in order to make the king of the gods (Indra) happy, (Royal Glory) that is his(Indra’s) dear object of enjoyment but that had been denied to him more thanonce.’

Cf. the Kahaum Stone Pillar Inscription of Skandagupta (CII III (1888), 67):śakropamasya ks. itipaśatapateh. skandaguptasya. For the Gupta succession wars seeBakker 2006.

60 Virkus 2004, 60, 72f.61 Junagar.h Rock Inscription of Skandagupta vv. 2–4 (CII III (1888), 59; Sircar SI I,

308f.):tadanu jayati śaśvat śrıpariks. iptavaks. ah. ,svabhujajanitavıryo rajarajadhirajah. |

25

26 Hans Bakker

sided with the Gupta viceroy of Vidiśa, Ghat.otkaca, in his bid for powerin the final years of Kumaragupta’s reign, an alliance to which the king ofthe Eastern Vakat.akas, Pravarasena II, the brother-in-law of Ghat.otkaca,had also belonged. The Kevala Narasim. ha Temple Inscription tells thedramatic story of how this party finally lost the war, and Ghat.otkaca’swidow was brought home to Pravarapura by her brother, the Vakat.a-ka king.62 On the other hand, the Mlecchas, among whom we certainlyhave to count the Hun. as, may have been temporarily thwarted, but theyremained a constant threat to the empire and would soon reappear in thehistory of South Asia. Skandagupta himself refers explicitly to the Hun. asin his Bhitarı Stone Pillar Inscription.63

Despite these proud declarations of supremacy, the Gupta empire wasin fact considerably weakened after it lost its main ally in the south, theVakat.akas. Without their stabilizing influence, the control by the centreover the western part of the empire (Malwa), was undermined. After atumultuous but successful start, a variety of indications point to the in-creasing difficulty Skandagupta had in preserving his grip on the extensivekingdom.64 His last coins date from ad 468 and he may not have reignedmuch longer. How the emperor met his end is unknown, but he left be-

narapatibhujaganam. manadarpotphan. anam. ,pratikr. tigarud. a[jnam. ] nirvis. ı[m. ] cavakartta ‖ 2 ‖nr.patigun. aniketah. skandaguptah. pr. thuśrıh. ,caturu[dadhijal]antam. sphıtaparyantadeśam |avanim avanatarir yah. cakaratmasam. stham. ,pitari surasakhitvam. praptavaty atmaśaktya ‖ 3 ‖api ca jita[m e]va tena, prathayanti yaśam. si yasya ripavo ’pi |amulabhagnadarpa, ni[rvacana] mlecchadeśes.u ‖ 4 ‖‘Following Him, the supreme king of kings is forever triumphant, whose breastis embraced by Śrı (Glory), whose heroism is brought about by his own arms,the one who picks the antidotal herb (nirvis. ı), namely the Garud. a/imperial com-mand, a remedy against the Serpent kings whose hoods were uplifted in prideand arrogance. When his father had won the company of the gods, he, Skanda-gupta of great glory and possessed of kingly virtues, forced his enemies to theirknees and conquered the earth by his own power, up to the four oceans, (mak-ing) the bordering countries thrive. Even his enemies spread his fame in Barbariccountries that he truly had obtained a victory, without words, (but) through theeradication of their pride.’

In note 3, placed twice ad 4d, Sircar remarks (ibid.): ‘This is Fleet’s conjecturalreading. Mlechchha may indicate the Hun. as who may have advanced against Cen-tral India about the end of Kumaragupta’s reign.’ Fleet’s only conjecture was toread nirvacana instead of niva. . . , and his edition suggests that he could read mle-cchadeśes.u without problems (CII III (1888), 59), and so do Bhandarkar et al. (CIIIII (1981), 299). The latter edition, however, conjectures niva[rtita*]. Somethingseems to have gone wrong with the notation in Sircar’s edition.

62 Bakker 2006, 172–181.63 CII III (1888), 55. See Bakker 2005.64 Gupta 1974–79 I, 336–38.

The Twilight of the Gupta Empire 27

hind an empire that was riven by internal strife and on the verge of fallingapart.

The Imperial Guptas

Kumaragupta I

Skandagupta (455–69)

x

Purugupta x Candradevı

Narasim. haguptax Mitradevı

Kumaragupta II (473)

Vis.n. ugupta (543?)

Vainyagupta

Budhagupta (476–95)

Anantadevı

Figure 1

Genealogy of the Imperial Guptas (after Kumaragupta I)

Between the years 468 and 476 we have to situate at least two Gupta kingsruling the centre of the former empire: Kumaragupta II, whose SarnathStone Inscription dates from ad 473 and whose seals have been found inNalanda and Bhitarı,65 testifying to his rule in Eastern Uttar Pradeshand Bihar, and his father Narasim. hagupta, two of whose seals have alsobeen found in Nalanda.66 Both these kings belonged to the branch of theImperial Gupta family that had been pushed out by Skandagupta, whenhe prevented Narasim. ha’s father, Purugupta, apparently the legal heir toKumaragupta I, from ascending the throne (see Figure 1).

65 CII III (1981), 321f., 355–360. Bhandakar et al. ascribed these seals to a Kumara-gupta III. For the identities of Kumaragupta II and III see Willis 2005.

66 CII III (1981), 354f.

28 Hans Bakker

Budhagupta

It is unknown who was ruling the western part of the empire from thecapital Vidiśa during the decade following Skandagupta. It would seemthat the Imperial Guptas had little or no control here, and that regionalpowers were beginning to gain the upper hand. The next Gupta king ofwhom we have dated evidence, Budhagupta, another son of Purugupta,began his career as a ruler over the central and eastern part of the empire,like his elder brother and nephew before him during their interregnum. Allinscriptions that testify to Budhagupta’s rule, from ad 476/77 (SarnathStone Inscription) to the Damodarpur Copperplate Inscription, Year 163of ad 482, have been found in the east, Uttar Pradesh, Bihar and WestBengal.67 Silver coins of Budhagupta of unknown provenance, which im-itate those of Skandagupta, attest that, in the Gupta Year 175, i.e. ad494, this ‘Lord of the Earth, the illustrious Budhagupta, after conqueringthe earth, is conquering heaven!’68 The latter endeavour might have metwith success.

Budhagupta thus had a long reign of about eighteen years, but theonly indication that he succeeded in restoring some sort of authority inthe western part of the empire comes from the famous Eran Stone PillarInscription of ad 484.69 This records the good fortunes of the ‘ShiningMoon,’ Suraśmicandra, who is said to protect the earth from the Gangesto the Narmada, while he qualifies himself as a Maharaja, previously thetitle of the viceroys of Vidiśa. The dating of the record (in the Gupta year165) with the words: ‘when Budhagupta was king,’70 shows that Suraśmi-candra and his feudatories were still aware of Gupta rule, but the total lackof honorifics strongly suggests that Budhagupta actually wielded little ifany authority in Vidiśa and Eran.

The precarious condition of the empire may best be illustrated bythe situation in Bundelkhand, immediately to the south of the Guptaheartland. Here, in Nachna (the old capital of the Vakat.akas accordingto A.M. Shastri) and nearby Ganj, two inscriptions have been found thatbelong to a king Vyaghradeva, who had switched his allegiance away fromthe Guptas and acknowledged the overlordship of the Vakat.aka king Pr.thi-vıs.en. a II.71 This Vyaghradeva probably ruled during the reign of Budha-gupta, since two inscriptions of Jayanatha, his son by his wife Ajjhitadevı,

67 CII III (1981), 332–339, 342–351. For the Damodarpur Copperplates see EI XV(1919–20), 113–145, and below, p. 242ff.

68 Allan 1914, 153 (emended): vijitavanir avanipati(h. ), śrıbudhagupto div(am. ) jayati(metre: first hemistich of an Upagıti (12, 15)).

69 CII III (1888), 88–90. SI I, 334–36.70 Sircar SI I, 335: bhupatau ca budhagupte.71 See Bakker 1997, 47ff. CII V, 91f.: vakat.akanam. maharajaśrıpr. thivıs.en. apada-

nudhyato vyaghradevo.

The Twilight of the Gupta Empire 29

are dated in the Gupta years 174 and 177 (i.e. ad 493, 496).72

At the same time somewhat to the east, in Baghelkhand, another pettyking, Harivarman, still acknowledged the suzerainty of Budhagupta (para-madevabudhagupte rajani), in an inscription dated in the Gupta Year 168(ad 487/88).73 This inscription was found in Shankarpur in the Sidhi Dis-trict between the Son and its tributary the Gopad River. The distancebetween the findspots of the inscriptions of Vyaghradeva and Harivarman(Ganj and Shankarpur) is only 160 km as the crow flies, but the respectiveterritories of the two kings were separated by difficult terrain, criss-crossedby several rivers, among which the Son, and the Kaimur Range. Despitehis allegiance to Budhagupta, the successors of Harivarman, as we willsee, would play a greater part in the fall of the Gupta empire than theirdisloyal western neighbours of the Uccakalpa dynasty. So who was thisHarivarman?

HarivarmanA Maharaja Harivarman is mentioned in the sixth-century inscriptionsof the Maukhari rulers of Kanauj as their ancestor. The only inscriptionthat we have of a king of that name, however, is the copperplate foundin Shankarpur. Ever since its discovery in 1977, the Harivarman of thisinscription has been identified with the ancestor of the Maukharis, butA.M. Shastri expressed scepticism when he called it ‘a brilliant conjecturein the absence of any dependable evidence on the point.’74 The plate hasbeen published by Balchandra Jain, and again by Thaplyal,75 but a closerexamination of the inscription is still warranted.

On the Shankarpur Copperplate, Harivarman mentions Śarvasvami-nı, Vijayavarman and Gıtavarman as mother, father and grandfather, allthree unknown from any other inscription or seal. This Harivarman be-longs to a kula, but the specification of this family is unclear; it definitelydoes not seem to be Mukhara/Maukhara or Aśvapati/Madra/Malava (seebelow). I read the passage at issue as: [2] śrımaharaja x x x 〈˘ ˘ 〉[3] kulo-dbhutena. Obviously there are several ways to restore the lost text,76 but

72 CII III (1888), 118 ll. 4–5, pp. 122, 131, 136. We accept Mirashi’s view as the mostplausible, viz. that ‘Vyaghradeva probably belonged to the Uchchakalpa dynasty’(CII V, xxviii). This dynasty is referred to as that of Uccakalpa on account ofthe place of issue of their later charters. Most of these inscriptions were found inKhoh, in the Nagod District, and the earliest one in Karı-talaı, 35 km north-eastof Mud. wara (Murwara).

73 JESI IV, 62–66; Thaplyal 1985, 156f.; Bakker 1997, 48; Willis 2005, 148f.74 Shastri 1997, 66.75 JESI IV (1977), 62–66; Thaplyal 1985, 156–57.76 The reading of the three syllables x are uncertain, but the first and third seem to be

sa and na (possibly va), whereas the second is variously, but unconvincingly, read ast.a (Jain 1997, 64), la (Ramesh/Thaplyal) or rya (K.V. Ramesh in Jain 1977, 66 n.

30 Hans Bakker

Maharaja Sadanavarman(?) Kula

Maharaja Gıtavarman

Maharaja Vijayavarman x Mahadevı Śarvasvaminı

Maharaja Harivarman

Figure 2

The pedigree of Harivarman given in the Shankarpur Copperplate Inscription

a tentative, translation (based on an emended reading) runs as follows:

On the above specified day, the illustrious Maharaja Harivarman,[. . . ] grandson of the illustrious Maharaja Gıtavarman, who belongsto the lineage of the illustrious Maharaja Sadanavarman, has grantedan agrahara settlement [. . . ].77

The rest of Harivarman’s pedigree as given in the Shankarpur Copperplateis clear and is represented in Figure 2.

Despite these uncertainties and apparent discrepancies, there are goodreasons for identifying this Harivarman with the ancestor of the Maukha-ris. We will come back to this later, but first it is necessary to considerthe other axe at the root of the Gupta tree.

Toraman. a

The Eran Inscription of ad 484, which broadcasted the good fortunesof Maharaja Suraśmicandra, is actually mainly concerned with the good

17). I read the second syllable as either a ‘round da with its mouth open to the right’(Dani 1986, 283; cf. da in paradattam in line 6 of the inscription), or dra, resemblingdra in rudra and bapidra in line 9. At least one and probably two syllables aremissing after na at the end of line 2, a word or part of a word that should beread in composito together with the following kula in line 3: ◦sadanavarmma◦ (?).The reading kulodbhutena, first proposed by Ramesh (in Jain 1977, 64 n. 3), seemsto me to be virtually certain, notwithstanding the fact that the first editor of theinscription, Balchandra Jain, proposed reading: [2]Sa[Śa]- [3] ka-[dauhi ]te(tre)na.What raises suspicion is that the dynastic (kula) name appears to be preceded byśrımaharaja.

77 Based on a photograph by Michael Willis, http://commons.wikimedia.org/wiki/File:Budhagupta.jpg (cf. Jain in JESI IV (1977)): asyam. divasapurvayam. ,śrımaharaja(sadana)〈varma〉kulodbhutena śrımaharajagıtavarmapautren. a [. . . ] śrı-maharajaharivarman. a [. . . ] agraharo ’tisr.s. t.ah. [. . . ].

The Twilight of the Gupta Empire 31

deeds of what must have been his feudatory, the (local) Maharaja Matr.-vis.n. u, who, together with his younger brother Dhanyavis.n. u, erected theGarud. a pillar (dhvajastambha) on which the inscription is found, in orderto honour Janardana, i.e. Vis.n. u, a monument that still adorns the sacredgrounds of Eran. Though rather isolated today, in classical times Eran(Airikin. a) guarded the main passage from Mathura, via Gwalior, throughthe valley of the Betwa to the rich province of Daśarn. a (Eastern Mal-wa) and its capital Vidiśa. And so it became a battlefield where invadersand local kings regularly clashed,78 the field where the fate of the westernGupta empire was sealed. Matr.vis.n. u had to face a formidable enemy fromabroad, previously encountered by Skandagupta: the Hun. as, also knownas the ‘White Huns,’ or Hephtalites.

A few years after the proud erection of the pillar, Matr.vis.n. u foughtToraman. a; he not only lost the battle but also his life. His younger brotherDhanyavis.n. u survived and submitted to the Hun. a chief.79 The latter obvi-ously allowed him to stay on, and even granted him permission to executethe last will of his brother, the accomplishment of their joint project inhonour of their Lord, who was believed to be ‘the pillar of the great cosmichouse.’80 As far as we can tell, this project had begun with the erectionof the inscribed Garud. a pillar and was now completed by the survivingbrother by erecting a temple of Lord Narayan. a in his manifestation as thecosmic boar, the refuge of the world.81

The sacred compound at Eran already possessed an anthropomorphicimage of Varaha (at present in Sagar University), made after the famousUdayagiri example, but now it was to receive a still larger Varaha templeand image (still in situ) in his theriomorphic form, after the Ramagiriexample.82 The royal symbolism of the Varaha figure had proven to bepowerful,83 one that could even be shared by former enemies, the Eranmonument seems to imply, since ‘it was made for the benefit of all livingbeings.’84 The monument epitomizes the emotion that must have been feltby the brothers Matr.vis.n. u and Dhanyavis.n. u when they saw the CentralAsian horsemen shaking the foundations of the great Gupta Empire, andas such it bears testimony to their trust in God, who, at times when theworld was sinking,

78 Casile 2009 I, 76–83.79 Eran Stone Boar Inscription of Toraman. a, CII III (1888), 158–161.80 CII III (1888), 159: trailokyamahagr.hastambhah. .81 CII III (1888), 160: bhagavato varahamurter jagatparayan. asya narayan. asya.82 Bakker 1997, Plate xxxii.83 Willis 2009, 41ff. For the Varaha myth see e.g. Gonda 1969, 129–45, and most

recently Kintaert 2011–12, 91f., Fig. 1.84 CII III (1888), 160: svasty astu gobrahman. apurogabhyah. sarvaprajabhya(h. ).

32 Hans Bakker

Turned into a boar (Varaha) of great lustre and possessed of onetusk, appearing fearsome due to his bloodshot eyes, illuminating the(dark) smoke by his radiance, he expanded in that (nether) region.

Having taken the earth upon his single shining tusk, he, the imper-ishable one, lifted (her) up, O hero, a thousand yojanas.

And while she was being lifted, a trembling originated, and all thegods were shaken as well as the seers rich in tapas.

The whole cosmos burst into lamentation, heaven, atmosphere andearth. Not a single god or human being stood firm.85

The image is the first of its kind that shows these trembling beings clingingto the sides of the boar in the hope of being brought to safety. As ithappened, that safety was being provided in the political realm by Hu-n. a rule, which, as the Eran Varaha seems to prove, was mild and liberalenough to allow Dhanyavis.n. u to celebrate his faith, ‘in the first year thatthe Maharajadhiraja, the illustrious Toraman. a of great fame and greatlustre, was ruling the earth.’86

The Aulikara rulers of Daśapura, Mandasor

Even with the disappearance of Suraśmicandra from the scene, the deathof Matr.vis.n. u, and Dhanyavis.n. u’s switching of sides, not all the Guptaresistence in the western part of the empire was yet broken. In anotherEran inscription of ad 510/11 we learn about a śrı Bhanugupta, whose re-lationship with the main Gupta line remains obscure, but who might havebelonged to the Gupta nobility of Daśarn. a. This inscription reports thata pitched battle (yuddham. sumahatprakaśam) took place at Eran betweenPrince Bhanugupta, ‘the bravest man on earth,’87 and an unspecified en-emy, who, however, may be identified with Toraman. a with a reasonabledegree of certainty. Bhanugupta’s general Goparaja fell in this battle, anda funerary monument was erected for him and his wife, who committedsatı. The story is recorded on the remains of this monument, which hadbeen transformed into a linga that was found on the banks of the BınaRiver 800 m SE of Eran near the village of Pehelejpur.88

The inscription is inconclusive as to who won the battle, but in thelight of what follows it may be suggested that it records the first of aseries of setbacks for the Hun. a conqueror. In his fight against the invader,Bhanugupta may have had the support of the ruler of Mandasor, King

85 MBh 3 App. I No.16, ll. 91–99.86 CII III (1888), 159: vars.e prathame pr. thivım pr. thukırtau pr. thudyutau maharajadhi-

rajaśrıtoraman. e praśasati |87 CII III (1888), 92: jagati pravıro raja mahan parthasamo ’tiśurah. .88 CII III (1888), 91–93.

The Twilight of the Gupta Empire 33

Prakaśadharman, since the former’s sister(?) Bhanugupta was marriedto Ravikırti, the prime minister (amatya) under Prakaśadharman’s fatherRajyavardhana; Ravikırti’s son Bhagavaddos.a, nephew(?) of Bhanugupta,served as prime minister under Prakaśadharman and may have fought onthe side of his maternal uncle (see Figure 3).89 As we will see, this emergingdynasty of Mandasor kings, referred to as the ‘Later Aulikaras,’ was onits way to turning the tables on their former Gupta overlords.

Generals Early Aulikaras Guptas Later Aulikaras

Vayuraks.ita GovindaguptaViśvavarman

KumaraguptaBandhuvarman

Dattabhat.a [Prabhakara] Skandagupta

Ravikırti Rajyavardhanax Bhanugupta Bhanugupta

Bhagavaddos.a Prakaśadharman

Yaśodharman/Vis.n. uvardhana

Kumaravarman(?)

Figure 3

The Guptas, the Early and Later Aulikaras, and their ministers/generals inMandasor

The fall of the Gupta empire offered challenges and opportunities to formerfeudatories and their subordinates. Along the ancient caravan route thatled from the seaport of Bhr.gukaccha (Bharukaccha/Broach), at the Gulfof Cambay, via Ujjayanı, capital of Western Malwa, to Mathura and thegreat Ganges-Yamuna Plain, a series of important centres had becomeprominent: Daśapura (Mandasor), Madhyamika (the present village ofNagarı) near the citadel of Chitorgar.h (Chitor, Rajasthan), and Jambu-marga, which may be identified with Xuanzang’s Mo-hi-shi-fa-lo-pu-lo,

89 Salomon 1989, 26. Mandasor Stone Inscription of Yaśodharman Year 589 (CII III(1888), 153; SI I 414f. vss. 10–16).

34 Hans Bakker

Maheśvarapura. Along this route, as we will discuss later, the Paśupatamovement, which originated in south Gujarat in the second century ad,spread to the north.

This route was, at least partly, under the control of Daśapura, whoserulers had acknowledged Gupta suzerainty during the greater part of thefifth century. They had played a role in the succession after Candragupta’sdeath, when a faction led by General Vayuraks.ita had sided with Go-vindagupta, viceroy in Vidiśa. After Kumaragupta had eliminated his el-der brother, he restored the so-called ‘Early Aulikaras,’ Viśvavarman andBandhuvarman, to power in Mandasor (see Figure 3).90 The son of Vayu-raks.ita, General Dattabhat.a, entered the service of the usurper Prabha-kara, bhumipati of Mandasor, who had pledged allegiance to Skandagupta:the latter may have deployed the general against the Hun. as.91

When the Gupta empire crumbled, the rulers of Mandasor becamemore directly exposed to the foreign invaders. As the three copperplatecharters found in Sanjeli indicate, Toraman. a had penetrated along thistrade route at least as far as northern Gujarat, and possibly the tradingport of Bharukaccha,92 after his first victory in Eran over Suraśmicandraand Budhagupta.93 The second battle of Eran, however, may have beena turning point. At that time the Later Aulikaras were still fighting onthe side of Bhanugupta, as we noted above, but four years later, Pra-kaśadharman stood alone against the mighty Toraman. a. If we accept theevidence of the Rishtal Inscription, he gained a victory. This inscriptionrecords the events of ad 515 and gives us some insight into the practiceof war in those days.

He [i.e. Prakaśadharman], who by battle rendered the title ‘Lord’ ofthe Hun. a king false, (though it) had been firmly established on earthup to Toraman. a, whose footstool had glittered with the sparklingjewels in the crowns of kings (that had bowed at his feet);

By whom auspicious seats were offered to the ascetics, splendid ones,made of the long tusks of that same (king’s) elephants, whose rut

90 See Bakker 2006, 171 n. 17.91 Sircar SI I, 406–09; Bakker 2006, 171.92 Chakravarti 2008, 396.93 These copperplates were found buried in the fields near the village of Sanjeli,

which lies 144 km east of Ahmedabad and 108 km north-east of Baroda. The firstof the three inscriptions refers to the third year of the reign of Toraman. a: [. . . ]paramabhat.t.arakamaharajadhirajaśrıtoraman. e pr. thivım anuśasati [. . . ]. It recordsthe pious gift of a group of merchants, residential ones and from abroad (caturdiśa-bhyagatakavaideśyavan. ijakah. ) in Vadrapali in the district of Śivabhagapura, where(Maharaja) Bhuta was vis.ayapati by the grace (prasada) of Toraman. a (Mehta &Thakkar 1978, 14). Cf. Ramesh 1986; Salomon 1989, 29; Virkus 2004, 133–37.

The Twilight of the Gupta Empire 35

was dripping (from their temples) while they were being shot downby (his) arrows at the battle front;

And by whom were carried off a choice of ladies of the harem ofthat same (king), whom he had defeated by his vigour in the thickof battle, after which he offered them to Lord Vr.s.abhadhvaja (i.e.Śiva) to mark the strength of the arms of the ‘Light of the World’(Lokaprakaśa, i.e. Prakaśadharman).94

For the time being the Huns had been warded off. The Gupta empire hadpaid the price, but a new generation of warriors had entered the ring.In Mandasor Prakaśadharman was succeeded by his son Yaśodharman. Inthe Panjab the Hun. as rallied around Toraman. a’s son Mihirakula. Anotherround of devastating wars began.

Yaśodharman

The final collapse of the Gupta empire in the first two decades of the sixthcentury resulted in several important changes in the religious landscapeof Northern India. The Gupta emperors had actively promoted the Bha-gavata religion. Thanks to their support great Vais.n. ava monuments hadarisen in Udayagiri, Bhitargaon,95 Bhitarı, Deogarh and Eran, to mentionjust the best known. In the sixth century this type of devotion came moreand more to be seen as a religion of the past.

The Mandasor rulers were staunch Śaivas and so were the Maukharis,and, apparently also Mihirakula (below, p. 38). Under Yaśodharman thekingdom of Daśapura evolved in the third and fourth decades of the sixth

94 Rishtal Inscription of Aulikara Prakaśadharman, Year 570 (515 ad), vss. 16–18 inSalomon 1989, 4:

a toraman. anr.pater nr.pamauliratna-jyotsnapratanaśabalıkr. tapadapıt.hat |hun. adhipasya bhuvi yena gatah. pratis. t.ham. ,nıto yudha vitathatam adhirajaśabdah. ‖16 ‖sam. gramamurdhani vipat.hanipatitanam. ,tasyaiva yena madavarimucam. gajanam |ay(am)idantaghat.itani taponidhi(bh)y(o),bhadrasanani rucimanti niveditani ‖17 ‖tasyaiva c(a)havamukhe tarasa jitasya,yenavarodhanavarapramadah. pramathya |lokaprakaśabhujavikramacihnahetor,viśran. ita bhagavate vr.s.abhadhvajaya ‖18 ‖

95 The original idol of the Bhitargaon brick temple (before ad 460) has not been pre-served, but the Śaiva-Vais.n. ava mixture of terracotta sculpture ‘suggests a templededicated to Hari-Hara’ (Williams 1982, 83).

36 Hans Bakker

century into an important cultural centre.96 Mandasor and the surround-ing sites are replete with sixth-century archaeological material, which sup-ports our idea that this area had played an important role in transmittingthe Paśupata movement to northern India.

On the route to the north, halfway between Daśapura and Jambu-marga, lies ancient Madhyamika, the present village of Nagarı. LikeMandasor, Nagarı contains sixth-century remains that testify to theexistence of a Śaiva temple. A large lintel of the original toran. a of thistemple has been preserved. The lintel contains sculptured panels at bothsides, which illustrate scenes from Śaiva (i.e. Paśupata) mythology. Thepanels on the obverse side seem to contain a true daks. in. amurti, in whichPaśupati initiates Daks.a in the Paśupata vrata, discussed in Bakker2004c. The panels on the reverse side of the lintel also depict a Paśupatatheme, viz. Arjuna obtaining the Paśupata weapon. They tell the storyof the Kiratarjunıya. This myth is best known from Bharavi’s epic poem,and there are good reasons to believe that this poet, who himself mighthave been a Paśupata in view of the first element of his name, bha,97 hadclose ties with the court of Yaśodharman.

Dan. d. in (mid 7th century) mentions in the autobiographical introduc-tion to his Avantisundarıkatha that his great-grandfather Damodara (firsthalf of the sixth century) was introduced to a King Vis.n. uvardhana throughhis friendship with Bharavi.98 Only one king named ‘Vis.n. uvardhana’ dur-ing the first half of the sixth century is known: the Aulikara king Yaśo-dharman, who in his Mandasor Stone Inscription, Year 589 (ad 532) isalso referred to by his name Vis.n. uvardhana.99 Combining these facts, wecan infer that the poet Bharavi may have lived at the court of this Yaśo-dharman in Daśapura.

In his Kiratarjunıya (KA), by representing Arjuna’s thoughts, Bharaviformulates a political ideology that may have been shared by the poet’spatron. Warder paraphrases Arjuna’s resolve in Canto 11.59–64 as follows:

People lacking in self-respect, who are without firmness and thereforeallow themselves to be oppressed through this defect of power, arecontemptible. Their condition is equal to that of grass (they will be

96 An epitome of the state of the art is found in the works of Varahamihira, whichdemonstrate the high level of cultural and scientific scholarship at this time.Varahamihira is believed to have worked at the Aulikara court of Yaśodharman(Pingree 1981, 32).

97 For Paśupata names beginning with bha see also n. 666 on p. 226.98 Warder 1989–1992 III, 198f. Even if the Avantisundarıkatha was not by the same

author who wrote the Daśakumaracarita (see Lienhard 1984, 234f.), the traditionit reports may be authentic.

99 For the identity of Yaśodharman and Vis.n. uvardhana see Sircar SI I, 411f. n. 1;Salomon 1989, 13–17.

The Twilight of the Gupta Empire 37

‘Give way to nobody,’ thismaxim the great seer hastaught me; the beast thatwas after my death hasbeen killed by me, for tostand by one’s maxims isthe ornament of the good.

Kiratarjunıya 14.14

Plate 1Third panel on the left of the reverse side of the Nagarı lintel

trampled on). Fortune and honour resort to a man only as long ashe does not lose his self-respect: only then is he a real man.100

Bharavi may have played on his patron’s self-conceit by casting his heroArjuna in the mould of Yaśodharman’s sentiments (see KA 14.14 next toPlate 1). The Mandasor Stone Pillar Inscription of Yaśodharman opens ina similar vein.

May that flying banner of Śulapan. i (i.e. Śiva) destroy the forcesof your enemy, the banner that bears the Bull, who is marked bythe prints of the five fingers of the daughter of the mountain (i.e.Parvatı), due to whose terrific bellowing the quarters vibrate andthe demons are gripped with fear, and who cracks the rocks of thecliffs of Mount Sumeru by the pounding of his horns.101

This is the language of war. The Śaiva Age had begun.102

MihirakulaThe enemy lurked indeed in the foothills of the Himalaya mountains,in the district around Śakala (Sialkot),103 and had recently again made

100 Warder op. cit., 205.101 CII III (1888), 146 vs. 1:

vepante yasya bhımastanitabhayasamudbhrantadaitya digantah. ,śr. ngaghataih. sumeror vighat.itadr.s.adah. kandara yah. karoti |uks. an. am. tam. dadhanah. ks. itidharatanayadatta(pancangula)nkam. ,draghis. t.hah. śulapan. eh. ks.apayatu bhavatam. śatrutejam. si ketuh. ‖

102 Sanderson 2009, 258–60. Cf. Mirnig 2013, 336.103 See below, p. 113.

38 Hans Bakker

formidable inroads into central India. By the 15th year of his reign (c.ad 530), Mihirakula had succeeded in recovering much of the terrainthat his father had apparently lost, as is proven by his Gwalior StoneInscription.104 Although this inscription is mainly concerned with the es-tablishment of a Sun temple on top of the Gwalior hill (fort) by Matr.ceta,it provides some information about Mihirakula, who is said to rule theworld.

He (i.e. Toraman. a), who had raised his family to fame, had a son ofunequaled prowess, a lord of the earth whose name was Mihirakula,and who, (though) unbent, [was bending to] Paśupati. 105

Plate 2Gwalior Stone Inscription of Mihirakula

A war between the ruler of Daśapura and the Hun. a king was unavoid-able. A decisive battle took place between ad 530 and 532 in the valleyof the Betwa (see below, p. 52), and Yaśodharman claimed victory: heerected a pair of ran. astambhas in Mandasor, ‘columns of victory in war,’

104 CII III (1888), 162. SI I, 424–26.105 CII III (1888), 162 (metre Arya):

tasyoditakulakırteh. putro ’tulavikramah. patih. pr. thvyah. |mihirakuletikhyato ’bhango yah. paśupatim a[vanatah. ] ‖ 4 ‖

In the fourth pada of this verse Paśupati is mentioned in the accusative, but theverb on which it depends is illegible. Fleet (loc. cit. n. 6) rightly thinks that ‘werequire here a verb, supplying the antithesis to (a)bhango, “without breaking;unbroken.”’ ◦tima◦ seems to be followed by three aks.aras, which could possiblygive: ◦timavanatah. . Sircar SI I, 425 n. 6 conjectures that we may read arcayati( ˘ , ˘ ˘ ), ‘worships.’ This would stretch the metre in which the last syllable isusually prosodically long, also when it ends in a short vowel.

The Twilight of the Gupta Empire 39

on which the inscription at issue is found in duplicate.106 Early in thefourth decade of the sixth century, Mihirakula was finally brought undercontrol, it seems. This is Yaśodharman’s rhetorical reply to Mihirakula’sboast:

By that King Mihirakula, whose head had never been forced to bowin humility by anyone save Sthan.u (i.e. Paśupati), and the embraceof whose arms gave the Himalaya Mountain the illusion of being‘impregnable,’ even by him (i.e. Mihirakula) the feet of this (Yaśo-dharman) were humbly worshipped with an offering of flowers (fallenfrom) his crest, when the strength of his arms bent that (Hun. a)monarch’s head painfully down into obeisance.107

The Mandasor Stone Inscription of Yaśodharman alias Vis.n. uvardhana,Year 589 reveals how Yaśodharman achieved this remarkable result.108

He led what seems to have been a coalition of Indian kings against theircommon northern enemy. Thus I interpret the following verse.

After having caused to submit by peaceful means and by force theeastern kings and the many, very powerful northern ones (respec-tively), he bore as his second name ‘Rajadhiraja Parameśvara,’ that(title) desirable, (but) difficult to obtain in the world.109

The ‘very powerful northern kings’ must have included the Hun. as, whosesubmission by force is recorded in the undated Mandasor Stone Pillar In-scription (above, p. 39). The ‘eastern ones,’ who were won over by ‘peacefulmeans,’ obviously had been part of a ‘coalition of the willing’ against the

106 Fleet in CII III (1888), 143-45.107 Mandasor Stone Pillar Inscription of Yaśodharman, CII III (1888), 146f.; SI I, 419

n. 4 vs. 6:

sthan. or anyatra yena pran. atikr.pan. atam. prapitam. nottamangam. ,yasyaślis. t.o bhujabhyam. vahati himagirir durgaśabdabhimanam |nıcais tenapi yasya pran. atibhujabalavarjanaklis. t.amurdhna,cud. apus.popaharair mihirakulanr.pen. arcitam. padayugmam ‖ 6 ‖

This verse rebukes the verse quoted above from Mihirakula’s Gwalior inscription(p. 38). It is a clear case of intertextuality and it seems to prove that kings wereactually familiar with the contents of the inscriptions of their adversaries.

108 This inscription was engraved by Govinda, the same person who engraved theundated Mandasor Stone Pillar Inscription of Yaśodharman. Consequently, thedate of the latter inscription ‘must fall within a few years on either side of thatdate,’ viz. ad 532 (Fleet in CII III (1888), 146). The rear side of this stone slabcontains a depiction which might also be significant (see below, p. 54).

109 Mandasor Stone Inscription of Yaśodharman/Vis.n. uvardhana (Malava Year 589,ad 532), CII III (1888), 153; SI I, 413 vs. 7: praco nr.pan subr.hataś ca bahunudıcan, samna yudha ca vaśagan pravidhaya yena | namaparam. jagati kantam adodurapam. , rajadhirajaparameśvara ity udud. ham ‖ 7 ‖

40 Hans Bakker

Hun. as and their feudatories. After this victory Yaśodharman made ex-travagant claims, such as that kings from the Lauhitya (Brahmaputra)River in the north-east to the western ocean and from the Himalaya inthe north-west to the Mahendra mountains in the south-east, were bowingto his feet.110

But who were Yaśodharman’s eastern allies? For the answer we willhave to turn our attention again to the centre, where we see the power ofthe Maukharis rising out of the ruins of the Gupta empire.

110 CII III (1888), 146.

The Maukharis

The Origin of the Maukhari Dynasty

Harivarman

The Maukhari dynasty of Kanauj has left us two inscriptions (the Jaunpurand Haraha Inscriptions), seven seals and a number of coins. To thesedocuments may be added the earlier inscription of the Gupta feudatorynamed Harivarman, if this is the same Harivarman to whom the Maukharikings trace back their pedigree (see Figure 5).111

According to the Haraha Inscription, on the other hand, the Mukharakings, headed by Harivarman, did not descend from the ancestors men-tioned in the Shankarpur Copperplate Inscription, but from the hundred

111 EI XXIII (1935–36), 42–52 reports the existence of Three Maukhari Inscriptions onYupas: Krita Year 295, which seem to belong to the third century, if the equationof Kr.ta wih the Vikrama Era is correct. These three short inscriptions mentionthe three sons of a certain Bala, Balavardhana, Somadeva and Balasim. ha, who arecalled ‘Governors’ or ‘High Commanders’ (mahasenapati). These commanders areapparently also qualified by the term mokhari (= maukhari), but the readings arefar from certain. If correct, however, this could mean that they were descendantsof Mukhara, which means ‘loudmouth’ or ‘leader,’ but one would have expectedthat in that case this would have been said of Bala, not of his sons (cf. Editor’snote ibid. p. 52 n. 8). Alternatively, mokhari is a variant of mukhara and only aqualification of the three sons, as mahasenapati is. The three yupas, which com-memorate three Triratra sacrifices, are found in Barwa, Kota District (Rajasthan).There appears to be nothing to link these three mokharis to the Maukhari kingsof Kanauj.However, a possible reference to these mokharis may be found in the Chandravalli

Inscription of the Kadamba king Mayuraśarman. The date of this inscription isuncertain. Sircar dates this king to c. ad 330–60 (SI I, 473; cf. HCI III, 272),but Pires 1934, 34 dates the Chandravalli Inscription ‘to c. 283 A.D.’ The inscrip-tion credits King Mayuraśarman with victories over the Traikut.a, Abhıra, Pallava,Pariyatrika, Śakasthana, Sendraka, Pun. yat.a, and Maukhari. According to Sircar(SI I, 473f.) the Prakrit names in the inscription read: Tekud. a, Abhira, Pallava,Puriyotika, Sakastha[na], Sayinthaka, Pun. ad. a, and Mokari. Sircar, following Tri-pathi (1964, 41 n. 5) notes that ‘Mokari’ in Mayuraśarman’s inscription may referto ‘the existence of a people or locality called Mukari on the Pon. n. i or Kaverı.’Pires 1934, 26 thinks that they are the Maukharis of Magadha (see below, p. 71).

41

42 Hans Bakker

sons whom King Aśvapati obtained thanks to a boon given by Vaiva-svata (i.e. Yama).112 This clearly refers to the Savitrı legend as told inMahabharata 3.277–283.113 According to this legend, Aśvapati is a kingof the Madras and, after his daughter Savitrı had obtained this boon fromYama, his wife Malavı bore him a hundred sons,114 who, according toMahabharata 3.281.58, are known as the Malavas, ks.atriyas who equalledthe gods.115

Apart from the rather common name Harivarman,116 there is thereforeapparently little that links the Shankarpur Copperplate to the Maukhariepigraphs and seals. Chronology as well as geography, however, imply aconnection with the later Maukharis of Kanauj. If such a connection ex-ists, we have to assume that the Maukhari rulers, once they had achievedhegemony in the Gangetic Plain and had supplanted their former over-lords, the Imperial Guptas, drafted a new, more prestigious genealogy. Asthe seals of Harivarman’s great-great-grandson Śarvavarman intimate, afirst step in the direction of the Gangetic Plain may have been that Hari-varman arranged a marriage between his son Adityavarman and a princessof the house of the ‘Later Guptas’ (lg) of Magadha—at that time stillfeudatories of the Imperial Guptas —viz. with a princess called Hars.a-gupta, whose name suggests that she was a sister or daughter of the lgHars.agupta, son of Kr.s.n. agupta.117

Before continuing the history of the Maukhari lineage of Harivarman,however, we have to consider a branch of the family that apparently livedin the same region as these Later Guptas, on the far, eastern, side of theRiver Son, viz. in Magadha.

112 Haraha Inscription verse 3 (EI 14, 115):sutaśatam. lebhe nr.po ’śvapatir vaivasvatad yad gun. oditam |tatprasuta duritavr. ttirudho mukharah. ks. itıśah. ks.atarayah. ‖ 3 ‖‘The Mukhara kings, who destroyed their enemies and stopped evil-doing, de-scended from that eminent century of sons which King Aśvapati obtained fromVaivasvata.’

Nowhere in the inscriptions and seals of the Maukharis does this dynasty claimdescent from Vaivasvata, which would imply descent from the latter’s father Vivas-vat, i.e. the Sun (Surya). Although Ban. a in HC 4 p. 244 [16] refers to the lineageof the Maukharis as the Suryavam. śa, the following iva means that this should notbe taken literally (see below, p. 78).

113 Cf. MtP 208–214.114 MBh 3.281.57, 3.283.13.115 Cf. MtP 213.16. The Br.hatsam. hita 14.27 mentions the Madraka and Malava to-

gether and locate these people along with many others, among whom the Hun. as,in the north.

116 At least two other kings between ad 450 and 550 carried the name Harivarman:the Kadamba king Harivarman and the Western Ganga Harivarman, feudatory ofthe Pallavas.

117 Cf. Ettinghausen 1906, 27; Thaplyal 1985, 37, 147f. See Figures 5 and 7.

The Maukharis 43

The Maukharis of the Gaya RegionWe possess three undated inscriptions of a certain Anantavarman, whodeclares himself to be an ‘ornament of the Maukhari family.’ From theseinscriptions we derive the pedigree presented in Figure 4.

Bhupanam. Maukharın. am. Kulam

Nr.pa Yajnavarman

Nr.pa (samanta) Śardulavarman

Anantavarman

Figure 4

The Maukharis of the Gaya Region

The open-minded Anantavarman records in these three inscriptions hisinstallation of 1) an image (murti) of Kr.s.n. a in a cave in the Barabar Hillcalled Pravaragiri, 2) an image (bimba) of Bhutapati and Devı (Śiva andParvatı) in a cave in the adjacent Nagarjuni Hill, and 3) an image of Katya-yanı in another cave in the Nagarjuni Hill, said to be part of the Vindhyamountains.118 In view of the mangala verse opening the last inscription,this should probably be taken as an image of Devı Mahis.asuramardinı.119

118 CII III (1888), 222, 225, 227. The Barabar cave in which an image of Kr.s.n. a wasinstalled is known as the ‘Lomas Rishi Cave’ and is most recently described asB3 in Falk 2006, 258–269 (cf. Patil 1963, 16). The inscription is found under thearch of the doorway (ibid. Fig. 26). About 1.5 km to the east are the NagarjuniHill caves. The images of Śiva and Parvatı may have been installed in the Vad. a-t.hika Cave (Falk op. cit. N2; cf. Patil 1963, 296f.), since the inscription is ‘on thesmooth and polished surface of the granite rock, on the right hand in the entranceto the cave’ (CII III (1888), 224). Kittoe 1847 (quoted in Falk op. cit. p. 275)observed: ‘From the fragments found scattered, I conclude that there must havebeen a very handsome temple here of very early date’ (cf. Patil 1963, 297). Falkdoes not mention the Maukhari inscription here. The cave in which the image ofKatyayanı was installed is Falk’s cave N1 (‘a wonder of the world,’ Falk op. cit. p.257), known as Gopika Cave (Patil 1963, 294ff.). The inscription is in the doorwayof N1 (Falk op. cit. pp. 270–79, Fig. 9).

119 CII III (1888), 227 vs. 1 (Śardulavikrıd. ita):unnidrasya saroruhasya sakalam aks. ipya śobham. ruca,savajnam. mahis. asurasya śirasi nyastah. kvan. annupurah. |devya vah. sthirabhaktivadasadr. śım. yunjan phalenarthitam. ,diśyad acchanakhansujalajat.ilah. padah. padam. sam. padam ‖

44 Hans Bakker

Only the inscription dedicated to Katyayanı/Mahis.asuramardinı locatesthe cave explicitly in the Vindhya mountains.120 As Yokochi has observed,this may point to the fact that the installed goddess Katyayanı ‘was re-garded not only as Mahis.asuramardinı, but also as Vindhyavasinı.’121

These inscriptions prove that the ancient caves, which in Maurya timeshad been donated to the sect of the Ajıvikas, had returned to mainstreamHinduism by the sixth century.

Apart from the caves themelves, little of this sixth-century materialapparently remains today, but the hills have not been properly explored,since most attention has gone to the caves.

When her foot was disdainfully puton the head of the Buffalo De-mon which made its anklet tin-kle, its lustre eclipsed completelythe splendour of a full-blown lo-tus flower; may that foot of theGoddess, crowned by an aureoleof shining toenails, show you therealm of fortune, when it unites(your) request, concurrent with(this) prayer of staunch devotion,with (its) reward.

Nagarjuni Hill Cave Inscription ofAnantavarman122

Plate 3Devı Mahis. asuramardinı, Kauva-Dol

Many Hindu images, among which Mahis.asuramardinı (Plate 3), are foundcarved into the rocks and boulders on the northern face of a hill calledKauva-Dol, southwest of the Barabar and Nagarjuni Hills.123 Moreover,

120 CII III (1888), 227, l. 8: adbhutavindhyabhudharaguham.121 Yokochi 2004b, 143. Yokochi also argues that it was probably this Gopika Cave

that was visited by Yaśovarman, as decribed in the Gaud. avaho vv. 285–338 (op.cit. pp. 149f.).

122 CII III (1888), 227 vs. 1. Sanskrit text in n. 119 on p. 43. This image represents theNorthern subtype in Yokochi’s classification (Yokochi 2004b, 134f.).

123 Yokochi 2004b, 142. Photos of this site have been kindly sent to me by Dr Michael

The Maukharis 45

at the back of the Lomas Rishi Cave (Pravaragiri) on a boulder at rightangles to the entrance to the Karn. a Caupar Cave (Falk: B1), there are‘three images in relief, very much defaced. One is evidently a linga. Theothers seem to have been males with two arms and standing.’124 The twostanding males seem identical, both clearly wearing a mala, whereas theright and left hands rest on what could be a cakra and a gada respectively,which would make them images of Vis.n. u (Plate 4).

Finally, there are extensive archaeological remains on the top of and onthe pathway leading towards the highest peak of the Barabar Hill, calledSurjangiri or Surjangarh. A rather modern temple replaces an older one,known by the name of Siddheśvara Mahadeva, and there are numerous old(Hindu) images around.125 These images have been (partly) inventoriedby Susan Huntington, who assigned them dates in the late seventh centuryand later.126 They therefore bear no direct relation to the religious activity

Willis. The site has been described, inter alia, in Cunningham ASI Report 1861–62,vol. 1, 41, and Patil 1963, 198f. The Mahis.asuramardinı (MAM) image illustratedin Plate 3 shows that Cunningham’s description is not precise: the goddess holdsthe left hind leg of the buffalo and not its tail. Therefore this MAM image be-longs to what Yokochi has described as ‘the Northern subtype of the Gupta IconicType’ (Subtype A1), which is different from the description of this icon in theSkandapuran. a (above, p. 6; Yokochi 2004b, 134f.). Yokochi 2004b, 142 remarks:‘These rude Mahis.asuramardinıs on the hillside are very likely to have been mod-elled on an authorative image available in the vicinity, which must have been theimage dedicated by Anantavarman.’ In addition to the MAM image of Plate 3, thesite contains a series of nine reliefs of the MAM icon in a row. Concerning theseimages, Yokochi wrote to me (e-mail dated 20-7-2006): ‘According to Seshadri’spaper [1963] and my photo collection of MAM images from the AIIS, the imagesof this type are usually found in Andhra Pradesh. I have two photos of this type:one dated to the late 8th cent., and the other to the 10th cent., the latter of whichis referred to in Seshadri’s paper.’

124 Buchanan quoted in Falk 2006, 260. See ibid. Fig. 8 ‘The idols’ rock’; Patil 1963,15.

125 Amar 2012, 173: ‘The Siddheśvara shrine at the top of Surjangiri, the highest hillin the Barabar range, can be dated to about the seventh century on the basis of twoinscriptions from the Vapi cave at the Nagarjunı Hill which mention Siddheśva-ra. One of the records mentions a preceptor or acarya with the name Yoganandabowing to Siddheśvara’ (based on Patil 1963, 298). The graffito referred to byAmar and Patil is found in Falk’s N3, Vahiyaka cave, at the entrance and reads:acaryaśrıyogananda〈h. 〉 pran. amati si〈d〉dheśvar(e). The script resembles that of theinscriptions of the Later Guptas (ad 650–750). Next to the entrance of this caveis a fragment of a doorjamb with, in low relief, two dancing, naked, male figures,whose raised right hands seem to hold their (matted) hair. It may also belong tothe period of the Later Guptas.

126 Huntington 1984, 90–92. Along the steep path leading up to the Siddheśvara Tem-ple, rock boulders are covered with mostly lingas and occasionally weathered im-ages, placed in artificial niches. Among the scattered loose images are a few in-scribed pieces of sculpture, several Uma-Maheśvara images, Varaha, Gan. eśa, Vis.n. u,Surya and a number of unidentified deities. The entrance to the temple contains

46 Hans Bakker

Plate 4Images near the Caupar Cave (Barabar)

of the Maukharis, but instead seem to testify to religious activity underthe Later Guptas of Magadha after King Hars.a.

The palæography of the three Anantavarman inscriptions has mostrecently been investigated by Jurgen Neuss (2003), who draws attention tothe fact that the Barabar Cave inscription does not yet, and the two Naga-rjuni Hill Caves inscriptions with increasing frequency do use a specifictype of a-matra, viz. ‘the small hook attached to the right end of the tophorizontal stroke of a consonant.’ According to Neuss, this type of matrais first found in the Mandasor Stone Inscription of Yaśodharman of ad532.127 Anantavarman’s inscriptions thus seem to belong to a transitionalperiod, since the new type of a-matra becomes customary in eastern Indiain the third quarter of the sixth century. Neuss consequently dates Ana-ntavarman to the second quarter of the sixth century ad, a date that is in

images of Uma, Maheśvara, Surya and Vis.n. u. Cf. Patil 1963, 17.127 Neuss 2003, 537. CII III (1888), 150–58. Neuss’s argument is open to question,

though. This type of a ‘horizontal medial a with a tick on the right’ (Dani 1986,274) is already common in the Deccan in the early fifth century and is, for in-stance, found in the Poona Plates of Prabhavatı Gupta. It seems therefore doubt-ful whether this type of medial a is a reliable criterion for dating Anatavarman’sinscriptions.

The Maukharis 47

keeping with Dani’s (1986, 126) observation that ‘the Magadha Maukha-ris were earlier than the Kanauj Maukharis.’ That the two Nagarjuni Hillinscriptions are later than the one in the Lomas Rishi Cave also followsfrom the content, since only in the Barabar inscription does Anantavarmancall his father a samanta, while he styles himself in that inscription asa good son (suputra),128 who executes his father’s wrath. In the Naga-rjuni Hill inscriptions, his father, called nr.pa, is obviously no more, butAnantavarman refrains from calling himself king. Strangely, the last twoinscriptions no longer mention the (Maukhari) kula either, the reason forwhich is hard to fathom. Did Anantavarman lose his status after the deathof his father?

Despite these uncertainties, we cautiously date the flourishing of thisbranch of the Maukhari family to the first half of the sixth century,129

that is the period in which the senior branch of the Maukharis was onits way—probably with the help of the Magadha branch—to establishingitself as the paramount rulers in Kanauj.

Although Śardulavarman is called samanta in the earliest inscriptionof this set of three, the inscriptions do not reveal who the overlords of theselocal rulers may have been, nor do they identify the enemies against whomthey fought so heroically. The overlords (and enemies) might have beeneither what remained of the Imperial Guptas in the east (Vis.n. ugupta),or the early kings of the Later Guptas before these were driven out ofMagadha by Iśanavarman of Kanauj (see below), or their overlords couldhave been their own (distant) relatives, who traced their pedigree back toHarivarman. This last possibility, which seems most likely to me, wouldconform with our tentative suggestion made above to the effect that, afterthe breaking up of the Gupta Empire, the Maukhari family worked hardto establish an independent homeland in the northern Vindhyas, coveringa stretch of hilly land on both sides of the River Son (Plate 5) of altogetherc. 400 km, from Rewa in the west to Gaya in the east, from where theyencroached upon the territories still held by the Imperial Guptas.

The Maukharis of Kanauj

In addition to the well-known Haraha Stone Inscription of Iśanavarmanof Vikrama Samvat 611 (ad 554), in which Iśanavarman’s son Surya-varman records his renovation of a dilapidated Śiva temple, we possessonly one other inscription that undisputably belongs to that branch ofthe family which eventually turned the ancient Kanyakubja into theirnew capital: the Jaunpur Inscription, which is merely ‘a fragment of a very

128 For the concept of a satputra/suputra see SP II B, 104 ad SP 52.4.129 Cf. also Kielhorn in EI VI (1900–01), 3.

48 Hans Bakker

Plate 5The River Son near Arrah

much larger inscription.’130 I have discussed the structure and contents ofthis inscription elsewhere (Bakker 2009b). This discussion showed that theJaunpur inscription follows a pattern that is not altogether different fromthe Haraha Stone Inscription. It is to be ascribed to Iśanavarman or hissuccessor, not, as is usually done, to his father Iśvaravarman.

The tradition recorded in both inscriptions assigns to Harivarmangreat military prowess, to Adityavarman piety, to Iśvaravarman virtuous-ness, and to Iśanavarman emperorship. The latter was credited with thetitle Maharajadhiraja in the seals of his successors, and in the inscriptionshe is said to be seated on the lion-throne. His son Suryavarman was ayoung man in ad 554, the date of the Haraha Inscription. Iśanavarman’s

130 CII III (1888), 228. As far as I am aware, there is no direct evidence linking theMaukharis to Kanauj. There is circumstantial evidence, though, which has ledthe majority of historians to accept it as the most plausible hypothesis (not soPires 1934, 152f.): Tripathi 1964, 32–36; Majumdar in HCI III, 69f.; Goyal 1967,363; Thaplyal 1985, 19f. Among this evidence is the clay seal-matrix or sealing,probably to be ascribed to Avantivarman, that was found in Kanauj (Thaplyal1985, 153ff.). The major argument in favour of Kanauj as the Maukhari capital, atleast at the time of Grahavarman, is based on the evidence of Ban. a’s Hars.acarita(see Tripathi 1964, 32–36).

The Maukharis 49

The Maukharis

maharaja Harivarmanx Jayasvaminı

maharaja Adityavarmanx Hars.agupta

maharaja Iśvaravarmanx Upagupta

maharajadhiraja Iśanavarmanx Laks.mıvatı

maharajadhiraja Śarvavarmanx Indrabhat.t.arika Suryavarman (ad 554)

Maharajadhiraja Avantivarman Vasat.a Bhaskaravarmanx Hars.agupta (of D. Kosala)

maharajadhiraja Grahavarman Su(va). . .x Rajaśrı (of Sthaneśvara)

Figure 5

The Maukharis of Kanauj

50 Hans Bakker

exploits may therefore have taken place in the fourth and fifth decadesof the sixth century. As observed above, he may have been helped by hisrelatives from the Gaya region.Just as the power of the Imperial Guptas crumbled in the western partof their empire under pressure from the Huns and was gradually takenover by the stronger Aulikara forces of Daśapura (Western Malwa) in thefirst and second decades of the sixth century—Bhanugupta being the last,it seems, who represented the Guptas against Toraman. a—so in a similarpattern was the power of the Guptas gradually taken over in the centre,i.e. in Uttar Pradesh and Bihar, by the Maukharis who emerged from theirhomeland in the valley of the Son.

Regarding the ‘Imperial Guptas’ of the heartland, we have the isolatedevidence of an incomplete seal found in Nalanda of a Paramabhagava-ta, Maharajadhiraja Vainyagupta son of Purugupta, and an inscriptionof a Maharaja Vainyagupta whose rule was favoured by Mahadeva (i.e.Śiva) found in Gunaigarh (East Bengal) and dated to ad 507.131 However,the titles and religious persuasions indicate that the inscription and sealbelonged to different people.132 Vainyagupta’s power was only a shadowof that of his ancestors. The same applies to Vis.n. ugupta, son of Kuma-ragupta II, of whom we also possess a (fragmentary) Nalanda seal.133

These kings, whose territory seems to have been restricted to Magadhaand Gaud. a, could certainly no longer make a stand against the Huns. Aswe shall argue below, this role was reserved for the Maukharis.

When we try to understand what happened in the second and thirddecades of the sixth century in the Gupta heartlands, we encounter thedynasty of the Later Guptas, whose position in the power politics of thisperiod is poorly documented. The first testimony to this dynasty datesfrom one-and-half centuries later (ad 672) and belongs to the time whenits ruler Adityasena assumed imperial titles in Magadha in an attemptto fill the vacuum that had been left by the death of Hars.avardhana (see

131 Gunaigarh inscription l. 1: bhagavanmahadevapadanudhyato maharajaśrıvainya-guptah. (Sircar in SI I, 341). The phrase padanudhyata is used by the ImperialGuptas only to express the legitimate succession of father to son (e.g. CII III(1888), 50, 53). See Ferrier & Torzsok 2008, 100ff.; Mirnig 2013, 328f.

132 CII III (1981), 319–20; Sircar SI I, 341. P.L. Gupta identifies both Vainyaguptasand ascribes the gold coins with the legend (biruda) Dvadaśaditya to them (Gupta1974–79 I, 361f.; see also Majumdar 1971, 40). A copperplate recently discoveredin Dhaka puts an end to the controversy. It describes the ‘Bengal’ Vainyagupta aspancadhikaran. oparika-mahapratıhara-maharaja śrı Vainyagupta who is favouredby the his sovereign (paramabhat.t.arakapadanudhyata). See Furui forthcoming.

133 CII III (1981), 364. The Damodarpur Inscription No. 5 is possibly to be ascribedto this Vis.n. ugupta (CII III (1981), 360–64). For a discussion about this see below,p. 58.

The Maukharis 51

below, p. 130). As we already have seen, in the early sixth century therelationship between the Maukharis and the Later Guptas may still havebeen rather close. Not only was Harivarman’s son and successor, Aditya-varman, apparently married to a sister or daughter of lg Hars.agupta, Adi-tyavarman’s own son Iśvaravarman was also married to a Gupta princess:Upagupta. Matters are made even more complex by the fact that both theMaukharis of the Gaya region and the Later Guptas are to be located inthe same territory: Magadha (South Bihar). The relationship between thetwo houses may have turned sour after the Maukharis achieved successagainst Mihirakula in the third and fourth decades of the sixth centuryand they, consequently, were no longer prepared to accept a subservientrole. This development may also have caused the putative loss of statusof Anantavarman (see above, p. 47).

Iśvaravarman

The (final) confrontation between the Maukharis and Mihirakula tookplace under Iśvaravarman between ad 530 and 532. In the Aphsad Inscrip-tion of the lg Adityasena, reporting on the subsequent conflict betweenthe Maukharis and the Later Guptas, the elephant troops of the formerare said to have thrown aloft the army of the Hun. as.134 The troops ofwhich Maukhari king is not specified, and this vagueness may indicatethat the event had taken place in the more remote past but ever since hadbeen the pride of the Maukhari army.

We find support for the supposition that it was Iśvaravarman who wasthe one who was confronted with Hun. a power in the Jaunpur Inscription,which says of him in verses 6–7:

The son of this (king), whose spotless fame spread worldwide, wasIśvaravarman . . . (6)

. . . who, indeed, . . . of the virtuous people 〈would equal him in〉virtue, (virtues) which brought happiness into the world and alle-viated the distress caused by the arrival of cruel (people) throughcompassion and love? (7)135

134 CII III (1888), 203 ll. 8–9 (vs. 11). See below, p. 61.135 Bakker 2009b, 211. Cf. the Haraha Inscription verse 8:

tenapıśvaravarman. ah. ks. itipateh. ks.atraprabhavaptaye,janmakari kr. tatmanah. kratugan. es.v ahutavr. tradvis.ah. |yasyotkhatakalisvabhavacaritasyacaramargam. nr.pa,yatnenapi yayatitulyayaśaso nanye ’nugantum. ks.amah. ‖ 8 ‖‘By him (i.e. Adityavarman) was begotten, destined to gain the power of nobil-ity, the lord of the earth Iśvaravarman, whose soul was well-disciplined and whoinvoked the Enemy of Vr.tra (i.e. Indra) in many sacrifices; other kings, no mat-ter how they tried, were unable to match his way of conduct, he whose deeds

52 Hans Bakker

Regardless of whether Mihirakula and his men actually deserved the blameheaped upon them in Indian (and Chinese) sources—in these sources‘Hun. a’ is virtually a byword for cruelty—there is a certain plausibilityin taking the phrase ‘the arrival of cruel people (kruragama)’ as a refer-ence to Mihirakula and his Huns.136 The Maukharis were, in all likelihood,the ‘eastern ally’ of the king of Daśapura, as we suggested above, and theking who was ‘won over by peaceful means’ must have been Iśvaravarman.

The Gwalior Stone Inscription indirectly bears witness to the jointmilitary operations of Yaśodharman and Iśvaravarman. In the 15th yearof his reign (c. ad 530), Toraman. a’s son Mihirakula was still recognized assovereign in the stronghold of Gwalior, 250 km to the north of Eran, thebattlefield of his father. Obviously the Huns at this time held a broad cor-ridor from their home base in the northern Panjab and Himachal Pradeshto central India and Eastern Malwa in particular. This corridor borderedin the south-west on the kingdom of Yaśodharman (Western Malwa) andin the north-east on the territories of the Maukharis, who had already an-nexed parts of the Ganges-Yamuna Plain by this time, driving the remainsof Imperial Gupta power further to the east. The military operation of theAulikara–Maukhari forces may thus have resembled a successful pinchermovement. To attack the Hun. a position in the corridor, the Aulikaras hadto march from Daśapura through the Vindhyas, to the valley of the Betwa,a march, it seems, which is alluded to in the same inscription that refersto the ‘eastern allies’:

Its rays were dimmed, its disk clouded, the orb of the sun resembledthe eye in a peacock’s tail turned round, which was caused by thedust, grey like the skin of a donkey, that was thrown up by his (i.e.Yaśodharman’s) army which marched through the jungle with itsbanners raised: the infuriated elephants of the army tossed aboutthe lodhra trees with their tusks and the noise it (produced) madethe crevices of the Vindhya mountains resound.137

We have already seen that after the victory, Yaśodharman declared himselfsovereign of the whole of India, but his eastern allies may equally havegained in self-confidence and have acted as the new emerging sovereignpower of the Yamuna-Ganges Doab. The rise of Kanauj and the enmity

eradicated the very nature of the Kali (age) and who equalled Yayati in fame.’136 Cf. Virkus 2004, 87.137 Mandasor Stone Inscription of Yaśodharman Year 589 (CII III (1888), 153; SI I,

414):yasyotketubhir unmadadvipakaravyaviddhalodhradrumair,uddhutena vanadhvani dhvaninadadvindhyadrirandhrair balaih. |baleyacchavidhumaren. a rajasa mandam. śu sam. laks.yate,paryavr. ttaśikhan. d. icandraka iva dhyamam. raver man. d. alam ‖ 9 ‖

The Maukharis 53

with the Later Guptas (lg) of Magadha may date back to those days ofvictory, whereas the seeds of a conflict with the Aulikaras were sown atthe same time.

IśanavarmanIśvaravarman was succeeded by his son Iśanavarman who further strength-ened the sovereign position of the Maukharis in the fourth and fifthdecades of the sixth century. Most of the information we have about thisking comes from the two Maukhari inscriptions. Our historical reconstruc-tion will be guided by the three enemies whom Iśanavarman is said to havedefeated: the Sulikas, the Andhras and the Gaud. as.

The SulikasThe Sulikas are mentioned in the Haraha Inscription in between the en-emies in the south and the east; consequently, they are probably to belooked for in the west or north-west, a direction also indicated, it wouldseem, by the great number of horses at their disposal. Iśanavarman issaid to have ‘routed in battle the Sulikas and their galloping arrays ofinnumerable horses.’138 These Sulikas have been variously identified.139

Chattopadhyaya 1968 argued in favour of the Hun. as, of which we knowthat they had been the adversaries of the Maukharis, but B.P. Sinha ob-jected that there ‘does not seem any justification as to why the same powershould be referred to as Śulikas in the Haraha Inscription but as Hun. asin the Aphsad Inscription.’140 The last feat of arms of the Hun. as undertheir king Mihirakula in central India that is known to us is the battlewith and defeat by the alliance between Yaśodharman and Iśvaravarman.

The Br.hatsam. hita (9.21) confirms the (north-)western location, if weare allowed to read Śulika instead of Sulika;141 it groups the Śulikas with

138 Haraha Inscription verse 13 (EI 14, 117):jitvandhradhipatim. sahasragan. itatredhaks.aradvaran. am. ,vyavalganniyutatisam. khyaturagan bhanktva ran. e sulikan |kr. tva cayatim(o)citasthalabhuvo gaud. an samudraśrayan,adhyasis. t.a nataks. itıśacaran. ah. sim. hasanam. yo jitı ‖ 13 ‖‘[Iśanavarman] is victorious, he who has ascended the lion-throne, his feet bowedto by the rulers of the earth, after having vanquished the sovereign of Andhra ofwhom the elephants, counted by the thousands, were passing triple fluid (scil.,temple must secretion, urine and faeces, due to excitement), after having routedin battle the Sulikas and their galloping arrays of innumerable horses, and afterhaving made the Gaud. as, driven from their territory by (his) might, take refugeat the ocean.’

139 This discussion is summarized by Thaplyal 1985, 23.140 Quoted from Thaplyal 1985, 23.141 The inscription clearly seems to read sulikan (see also Sircar SI I, 388), not

54 Hans Bakker

the Gandharas.142 This might refer to a tribe of Central Asian origin whosettled somewhere within north-west India. It is conceivable that thesehorsemen entered the service of Indian kings as mercenaries,143 whichcould explain the geographically scattered presence allotted to them inour sources. If the Sulikas were thus allied to a bigger western power, thispower may have been the former ally of Iśanavarman’s father, the powerfulAulikara king Yaśodharman, or his (unknown) successor.144

We have a lead, admittedly vague and inconclusive, to Yaśodharmanwho, as we have seen above (p. 37), wrote on his victory column that hewent to war under ‘the flying banner of Śulapan. i.’ The stone on whichthe Mandasor inscription of Year 589 (= ad 532) is engraved, describingYaśodharman’s campaign (see above, p. 52) and proclaiming his supremesovereignty (rajadhirajaparameśvara), contains an illustration on its back,described by Fleet as follows.

On the back, which is divided into two compartments by what seemsto be either a spear [śula, H.T.B] with a curved handle , or a shep-herd’s crook, there are engraved in outline, very roughly,—at theupper corners, the sun on the proper right, and the moon on theproper left; and, lower down, on each side, a man on horseback, fac-ing toward the central dividing figure; the horseman on the properright side carries in his left hand either a chaurı or fly-flap made ofthe bushy tail of the Bos Grunniens, or a śankha or conch-shell; theother horse-man caries something in his right hand, but I could notdistinguish the object.145

śulikam as read by Śastri (EI XIV, 117), although the Śulika, the ‘lancer,’ mightbe intended.

142 The Br.hatsam. hita (BS) 10.7 gives the following list: Bahlıkas, Cınas, Gandharas,Śulikas and Paratas, but Br.hatsam. hita 14.8 assigns a country named Śaulika (sic)to the south-east along with, inter alia, Kosala, Kalinga, Vidarbha, Andhra, etc.The Matsyapuran. a 121.45–46 associates the Śulikas with other barbaric peoplethrough whose country the river Caks.u flows, according to Dey 1971 s.v. to beidentified with the river Oxus (however, cf. Kane IV, 742 s.v. ‘Caks.us’). Cf. alsoRamakrishna Bhat’s comments ad BS 14.8: ‘The Śulikas are called Culikas in thePuran. as. In the Matsyapuran. a this country is said to be watered by the Caks.u(Oxus). A branch of the Śulikas that migrated to Orissa is to be identified withthe Śulkis. P.C. Bagchi connects the Śulikas with the Calukyas of the Deccan.’

143 ‘Groups living by the profession of arms’ (ayudhajıvisangha) are mentioned in P5.3.114.

144 A fragmentary inscription of Kumaravarman found in Mandasor may point to adynastic shift from the ‘Later Aulikara’ line of Yaśodharman/Vis.n. uvardhana backto the lineage of the ‘Early Aulikaras’ of Bandhuvarman in the middle of the sixthcentury (see above, Figure 3. Salomon 1989, 23, 29f.).

145 Fleet in CII III (1888), 150. Unfortunately, the present whereabouts of this slab ofstone are unknown.

The Maukharis 55

Not elephants, but cavalry! The lancers of the Aulikaras may have beatenthe Huns at their own game. The image seems to resemble, if we areallowed to speculate, a coat of arms of the lancers’ battalion, the Śulikas,and the aulikaralanchana, the emblem of the Aulikaras that is attributedto Yaśodharman-Vis.n. uvardhana in the inscription, may be represented bythe sun and moon in the upper corners of the depiction, that is, if theseare not merely auspicious signs.146

What we may derive from all this fragmentary information is thatYaśodharman in the fourth decade of the sixth century was a powerfulbut troublesome neighbour, who may easily have broken the intermittantcoalition at the very moment that he declared himself the paramountsovereign of the Subcontinent. An armed conflict with the rising powerof the Maukharis of Kanauj may therefore have been inevitable, once theintervening Hun. as were out the way.

The Andhras

The Jaunpur and Haraha inscriptions are unanimous in ascribing toIśanavarman a victory over the king of Andhra, whose army seems tohave contained an exceptionally large number of elephants.147 The kingof Andhra must have been the Vis.n. ukun.d. in ruler Indravarman, aliasIndrabhat.t.arakavarman, who reigned roughly from ad 526 to 555.148 This

146 Cf. Fleet in CII III (1888), 151 n. 4; Sircar in SI I, 413 n. 3. The etymology of theword aulikara is unknown, but kara could signify ‘ray of light.’

147 See above, n. 138 on p. 53. S.R. Goyal 2005 argues against this interpretation. Heascribes the Jaunpur inscription to Iśvaravarman and argues that the war with theAndhra king could have been conducted by Iśanavarman when still crown prince(this view was first proposed by N.G. Majumdar in Indian Antiquary 1917, 127).Much weight is given by Goyal to the syntactic structure of verse 13 of the HarahaInscription, in which the absolutives jitva, bhanktva, and kr. tva precede the aoristadhyasis. t.a . . . (sim. hasanam). The Andhra adversary of Iśvaravarman/Iśanava-rman was, according to Goyal, the Vis.n. ukun.d. in King Madhavavarman II Janaśrayaand the battle must therefore have taken place before ad 518. However, this andthe other victories ascribed to him by Goyal would have made Iśvaravarman aMaharajadhiraja, which title is explicitly denied to him by his grandson Śarva-varman (who calls him Maharaja in his seals). Chronologically speaking it wouldplace the Maukhari kings Harivarman, Adityavarman, Iśvaravarman and the yu-varaja Iśanavarman all between ad 487 and 518 (31 years), followed by a gap of34 years, after which we hear again of a ruling king Iśanavarman in ad 554 (inthe Haraha Inscription), which, although not impossible, is not the most obviousinterpretation. Goyal’s arguments fail to carry conviction. After his victories, Iśa-navarman held the throne (sim. hasana), acknowledged as such by the bowing of hisenemies; the aorist adhyasis. t.a may express the actuality of this feat at the momentthe praśasti was written. But the basic mistake is that Goyal, like his predecessors,fails to collate the verse pattern in the Jaunpur Inscription with the one in theHaraha Inscription (see Bakker 2009b).

148 Sankaranarayanan 1977, 65f.

56 Hans Bakker

king took great pride in the successes of his elephant army as ‘he hadwon in numerous elephant battles hundreds of thousands of conflicts,’ ahyperbole repeated by his successor Vikramendravarman II.149 The clashbetween Iśanavarman and Indravarman may have taken place in or nearthe Vindhya mountains, the natural border between the two powers, sinceIndravarman is said to have taken shelter in the caves of this mountainrange.150 As has been argued by Sinha (1977, 118f.), the Maukhari kingmay have been helped by the Pan.d. uvam. śin king of Śrıpura (Daks.in. aKosala), Tıvaradeva. The relationship between the Maukharis and theearly Pan. d.uvam. śins of Śrıpura was an old one, as the origin of the latterdynasty was Mekala, the region between the homeland of the Maukharisand Daks.in. a Kosala.151

As a matter of fact, the good relationship between the Maukharisand the Pan. d. uvam. śins was confirmed in the next generation, when Iśa-navarman’s granddaughter Vasat.a was given in marriage to Tıvaradeva’snephew Hars.agupta.152 The same Tıvaradeva may have been Iśanava-rman’s natural ally in his conflict with the Gaud. as, who, no doubt, con-tested Tıvaradeva’s claims to Utkala (Orissa).153

The peace between Iśanavarman and the Vis.n. ukun. d. in ruler of Andhra,Indravarman, may have been sealed by Indravarman giving his daughterto Iśanavarman’s son Śarvavarman, as the name of the latter’s wife, Indra-

149 Ramatirtham Plates of Indravarman, Year 27 ll. 5f.: anekacaturdantasamaraśata-sahasrasam. ghat.t.avijayı; cf. e.g. Chikkulla Plates of Vikramendravarman II, Year10, ll. 12f.: anekacaturdantasamarasam. ghat.t.adviradagan. avipulavijayasya (Sankara-narayanan 1977, 66f., 166, 169).

150 A lively description of a great elephant battle in the Vindhyas is given in Skanda-puran. a 66, where Kauśikı Vindhyavasinı destroys the demon army of Nisumbhaand Sumbha (SP III, 130f., 318–25).

151 See further below, p. 202.152 EI XI, 191 vss. 15–16 (see below, n. 162 on p. 59). See Bakker 1994, 15ff. The rela-

tionship between the two families, the Pan. d.uvam. śins of D. Kosala and the Varmansof Kanauj/Magadha, remained close during at least the early years of the reign ofVasat.a’s son Mahaśivagupta Balarjuna (ad 590–650), who in his undated MallarPlates complied with the request of his maternal uncle (matula) Bhaskaravarmanto give a village to the small Buddhist monastery at Tarad. am. śaka, probably inthe neighbourhood of Mallar (Shastri 1995, II, 135f.). Obviously, this MaukhariBhaskaravarman is different from the homonymous king of Kamarupa, who washis younger contemporary. Whether or not this relationship remained good afterKing Hars.avardhana had ascended the throne of Kanyakubja is less certain. Seebelow, p. 103. However, contact and exchange between Kanauj and Daks.in. a Ko-sala during the reign of Mahaśivagupta seems to be indicated by the influentialpositions acquired by religious office holders, such as a Paśupata ascetic namedUdbhavaraśi (below, p. 139, n. 465 on p. 149, Bakker 2007, 5f.) and a Samavedinbrahmin of the Kauśika-gotra hailing from Kanyakubja, named Jejjat.a (Bosma2013, 250f.).

153 See Bakker 1994, 16f.

The Maukharis 57

Plate 6Territory of the Imperial Guptas around ad 530

bhat.t.arika, suggests.

The Gaud. as

The ‘Gaud. as’ of the Haraha Inscription seem to be a global reference tothe eastern neighbours, the rulers of Magadha and beyond.154 That thekingdom of Magadha included the country of Gaud. a proper, or part of it,is virtually proven by the Damodarpur Copperplate Inscription (No. 5) ofthe Gupta year 224 (= ad 543/4). This inscription is one of a set of fivecopperplates found in Damodarpur in the Dinajpur District (Bangladesh),i.e. Gaud. a proper. It reports a land transaction in the Kot.ivars.a vis.aya ofthe Pun.d. ravardhana bhukti (i.e. Paun. d. ra). It was made ‘on the 5th dayof Bhadra in the year 224, while the paramadaivata, paramabhat.t.araka,maharajadhiraja Śrı. . . ◦gupta was ruling the earth.’155 Various restora-tions have been proposed for the missing Gupta name: Budha◦, Vis.n. u◦,Upa◦, Bhanu◦ or Kumara◦.156 Michael Willis (2005, 145) argues that ‘the

154 The toponym Gaud. a in 6th and 7th-century sources is vague. If it is not usedto indicate the east as such, Gaud. a proper may be best circumscribed as theborderlands of present-day West Bengal and Bangladesh, i.e. south of the Ganga,the land around Berhampur (Murshidabad District) with the capital Karn. asu-varn. a on the Bhagırathı (Hooghly) River, and, north of the Ganga, the Malda andDinajpur Districts, in which Damodarpur is situated. See Plate 6.

155 CII III (1981), 361f., Sircar SI I, 347. For further discussion of the contents of thisinscription see below, p. 242.

156 Gupta 1974–79 I, 48f.; CII III (1981), 361; Sircar SI I, 346 n. 5. Willis 2005, 145n. 68.

58 Hans Bakker

opening three lines’ and ‘the pattern of endowment’ are similar and con-form to the style of charters of the Imperial Guptas. This argument, com-bined with the date of Gupta Era 224,157 makes Vis.n. ugupta, son of Kuma-ragupta II, grandson of Narasim. hagupta, a possible candidate. He is thelast king of Magadha known to us who claimed in an (undated) seal foundat Nalanda to belong to the Imperial Guptas.158 If Vis.n. ugupta was stillaround as late as the fourth or fifth decade of the sixth century, he musthave been in his seventies and lived long enough to become a contempo-rary of a ruler who belonged to a new dynasty that has been styled the‘Later Guptas,’ viz. Kumaragupta. As observed above (p. 42), we assumethat the Later Guptas started as feudatories under the Imperial Guptas,in the days when the latter had been reduced to their Magadha and Gaud. adominions.159

Although nominal power in Magadha and Gaud. a thus still seem tohave rested with the Imperial Guptas, this power was somehow sharedwith the lg Kumaragupta. No doubt the Imperial Gupta court will havemade things look as if nothing had changed for as long as possible, andwill have clung to imperial titles till its last breath. Such is the inertia ofhistory.

Another, admittedly vague, indication that at least parts of Gaud. awere included in the Magadha kingdom in the fourth and fifth decadesof the sixth century may be the praise of Kumaragupta’s father, lg Jı-vitagupta I, in the Aphsad Stone Inscription of Adityasena, in which in-scription it is said that Jıvitagupta struck terror into his enemy who wasliving at the borders of the ocean.160 This may be seen as an indication

157 Cf. EI XV (1919–20), 142 where Basak reads 10 instead of 20, which would givethe date ad 533/34. The date 224 was first read by Dikshit in EI XVII, 193.

158 CII III (1981), 364. Willis 2005, 145 thinks that ‘his reign may have extended tocirca 515 CE,’ but he gives no arguments to support this date. Cf. Devahuti 1998,16. For coins of Vis.n. ugupta see Allan 1914, 145f. and Raven 1994, 186 (‘ruledtowards the middle of the sixth century’).

159 The dynasty of the Later Guptas never claimed descent from the Imperial House,nor did they use imperial titles before the second half of the seventh century. Thisclaim is found for the first time when the power of Kanauj, after King Hars.a, wastemporarily at a low ebb, namely in the Mandar Hill Rock Inscription of the lgking Adityasena, who is there called paramabhat.t.arakamaharajadhiraja, just as hisdescendants give themselves imperial titles in their inscriptions. They were, in thewords of R.C. Majumdar, ‘the residuary legatee of the Gupta empire’ (Majumdar& Altekar 1967, 209). The still living memory of their failed bid for the successionto the imperial throne may be one of the factors that prevented lg Adityasenafrom assigning high-flown titles to the earlier kings of his lineage, or to himself forthat matter, in his Aphsad inscription.

160 CII III (1888), 203, ll. 4–5. That Gaud. a belonged to Magadha seems also to emergefrom the Aphsad Stone Inscription itself, which was composed (or engraved) bya native of Gaud. a, Suks.maśiva (CII III (1888), 205). Sircar 1971, 125 observes:‘About a century after the establishment of the powerful kingdom of Gaud. a, its

The Maukharis 59

that the rulers of Magadha in the first half of the sixth century had greatdifficulties in standing their ground against, on the one hand, their easternneighbours on the Bay of Bengal, and, on the other, the Maukharis in thewest. If we follow the Maukhari report of these events, the Guptas them-selves were soon forced ‘to take refuge at the ocean’: the Later Guptas,that is; for the ‘imperial house’ of Vis.n. ugupta, or what still remained ofit, this defeat at the hands of the Maukharis was the final blow.

Let us return to the Haraha Inscription. If we level the hyperbole, we maynow explain the expression ‘and after having made the Gaud. as, drivenfrom their territory by (his) might, to take refuge at the ocean,’161 as areference to the fact that the Maukhari king reduced the Magadha-Gaud. akingdom to Gaud. a proper. This is in conformity with the claim made in theinscription of Iśanavarman’s granddaughter to the effect that her fatherSuryavarman was born in the dynasty of the Varmans who were overlordsof Magadha (magadhadhipatya).162 And this claim reveals precisely why

name began to be used in a general sense to indicate the countries of EasternIndia. In Dan.d. in’s Kavyadarśa (seventh century) one of the two principal styles ofSanskrit composition is described as Gaud. a or pracya, i.e. eastern.’ The use of theword ‘Gaud. a’ in the Haraha Inscription may indicate that this process had alreadyset in a century earlier. Cf. also Thaplyal’s (1985, 48f., 81f.) discussion of Yaśo-varman’s conquest of Magadha described in Vakpati’s Gaud. avaho (cf. Majumdar1971, 86f.).

161 See above, n. 138 on p. 53. Various interpretations of the reading cayatim(o)cita-sthalabhuvo have been proposed. See Thaplyal 1985, 23f. Sircar 1964, 129 trans-lates: ‘after having made the Gaud. as [compelled] to leave dry land for all time tocome and to take shelter in [the waters of] the sea.’ My translation of ayati with‘might’ takes its cue from Hemacandra’s Anekarthasam. graha 3.261: ayatih. syatprabhavottarakalayoh. (I am obliged to Alexis Sanderson for pointing this out tome). Maybe a pun was intended.

162 The Maukharis were referred to by their allies, the Pan. d. uvam. śins of Daks.in. a Ko-sala, as the House (kula) of the Varmans, which had acquired great prestige dueto their sovereignty over Magadha. The Vasantatilaka and Śardulavikrıd. ita verses15–16 in the Sirpur Laks.man. a Temple Stone Inscription of the time of ŚivaguptaBalarjuna run (Shastri 1995, 143; EI XI, 191):

tasyorujanyajayino jananı jananam, ıśasya śailata[na]yeva mayu[rake]toh. |vismapanı vibudhalokadhiyam. babhuva, śrıvasat.eti narasinhatanoh. sat.eva ‖ 15 ‖nis.panke magadhadhipatyamahatam. jatah. kule varman. am. ,pun. yabhih. kr. tibhih. kr. tı kr. tamanah. kampah. [sudha]bhojinam |yam asadya sutam. himacala iva śrısuryavarma nr.pah. ,prapa prakparameśvaraśvasuratagarvanikharvam. padam ‖ 16 ‖

The daughter, named Vasat.a, of Suryavarman, brother of the Maukhari sovereignŚarvavarman, was married to the Pan. d. uvam. śin king Hars.agupta (in the verse in-dicated as prakparameśvara, Lord of the East), and became the mother of KingMahaśivagupta Balarjuna (the contemporary of Hars.avardhana). These verses jux-tapose Mahaśivagupta, his father (Hars.agupta) and his mother (Vasat.a) to Ska-nda/Karttikeya (mayuraketu), Śiva (parameśvara) and Parvatı (śailatanaya), andpoint out that just as Parvatı’s father, Himacala, took pride in being the father-in-

60 Hans Bakker

the Maukharis spoke of their eastern enemies as the Gaud. as and not as theLords of Magadha: the latter would have had the unwanted implicationthat the Maukharis did not possess the birthright to that much-wanted,prestigious title.

The exact role of the Maukharis of the Gaya region in driving theGuptas out of Magadha remains unclear, but the most likely scenario isthat they acted in alliance with their relatives from Baghelkhand. Ana-ntavarman marked his newly won territory with three sanctuaries in theage-old caves of the Malabar-Nagarjuna Hills. It was probably from thistime onwards, the moment when the Maukharis, that is Iśanavarman, suc-ceeded in ousting the Guptas from Magadha in the fifth decade of the sixthcentury, that the dynasty assumed the imperial title of Maharajadhiraja.The battle had been all about Magadha. Kanauj, as a political centre, wasstill in its infancy in those days. The real trophy was sovereignty over thecradle of empire, the ancient land of the Buddha.163

There is more evidence that the territory of the Later Guptas was re-duced to Gaud. a in the second half of the sixth century. The Suman. d. alaCopperplate Inscription of the time of Prithivıvigraha-bhat.t. araka, rulerof Kalinga, is dated in the Gupta Year 250 (ad 569).164 As to this, Sircarremarks:165 ‘This shows that, even though the Guptas had been oustedbefore this date from Bihar and U.P. by the Maukharis, they were contin-uing their hold on Bengal and Orissa.’ Together with their allies in Orissa,these Later Gupta kings became sandwiched between the mighty allianceof the Maukharis and the Pan.d. uvam. śin kings of Daks.in. a Kosala on theone hand, and the hereditary ruler of Vanga (Bengal), Samacaradeva, onthe other.166

The Aphsad Inscription informs us about the Maukhari-Gupta conflictfrom the other, Gupta side. King Adityasena reports in it the exploits ofhis ancestors, among which is lg Kumaragupta’s inconclusive encounterwith Iśanavarman.

[Kumaragupta] who, after he had become Mount Mandara (adasher) as it were, quickly churned the formidable milk-ocean thatwas the army of the illustrious moon among kings Iśanavarman, in

law of Śiva, the Maukhari prince Suryavarman was proud of his daughter’s maritalbond with the House of the Pan. d.uvam. śins.

163 Kaut.ilya’s Arthaśastra (2.19.10) mentions Magadha as the region from whichmetal/iron (ayas) is procured, and in AS 2.12.37 it observes: akaraprabhah. kośah.kośad dan. d. ah. prajayate | pr. thivı kośadan. d. abhyam. prapyate kośabhus.an. a ‖

164 SI I, 490: . . . varttamanaguptarajye vars.aśatadvaye pancaśaduttare kalingaras. t.ramanuśasati śrıpr. thivıvigrahabhat.t.arake. . . .

165 SI I, 490, n. 4.166 Majumdar 1971, 42–44.

The Maukharis 61

which the spreading rings of ripples were (like) banners, easily wav-ing in the rising wind, and in which the rocks were the massive, ag-gressive war-elephants, which were played round by the water-massthat was the dust that flew up—(this ocean) that was the sourcefrom which fortune/Laks.mı was obtained.167

The same inscription suggests, however, that the final defeat of the LaterGuptas at the hands of the Maukharis was suffered, not by lg Kuma-ragupta, whose pious death at Prayaga is mentioned,168 but by his sonDamodaragupta, who fell in battle:

While he was rendering asunder the stepping array of the mightyelephants of the Maukhari, which had in (earlier) battles thrownaloft the army of the Hun. as, [Damodaragupta] passed out and, bythe sweet caresses of the lotus-hands of the ladies of heaven, regainedconsciousness, taking his choice: ‘mine!’169

As already observed, Adityasena, the first of the Later Guptas (of Maga-dha and Gaud. a) of whom we possess inscriptions, shied away from linkinghis pedigree to the Imperial Guptas. And his tone in describing his great-great-grandfather’s war with the Maukharis is mild on the whole, when hecalls the illustrious adversary Iśanavarman ‘the very moon among kings.’As it happens, the political situation had changed dramatically in thehundred-and-thirty odd years since this battle. Adityasena had given hisdaughter in marriage to a descendant of the Maukharis, Bhogavarman,and he no longer needed to be afraid of their power, as this had dwindledby the time of Hars.avardhana’s demise (see Figure 6).

167 CII III (1888), 203 ll. 6–7 (vs. 8):

utsarpadvatahelacalitakadalikavıcimalavitanah. ,prodyaddhulıjalaughabhramitagurumahamattamatangaśailah. |bhımah. śrıśanavarmaks.itipatiśaśinah. sainyadugdhodasindhur,laks.mısam. praptihetuh. sapadi vimathito mandarıbhuya yena ‖8 ‖

168 CII III (1888), 203 ll. 7–8: śauryasatyavratadharo yah. prayagagato dhane* | ambha-sıva karıs. agnau magnah. sa pus.papujitah. ‖ * Sircar SI II, 47 conjectures dhanı, i.e.‘as a wealthy man.’ ‘[Kumaragupta] who, faithful to his word and prowess, hasgone to Prayaga in state, and, plunging (himself) into a fire of dry cow-dung, as ifinto water, became adorned with flowers.’ Devahuti 1998, 20 suggests that Kuma-ragupta’s failure against the Maukharis might have made him commit suicide.

169 CII III (1888), 203 ll. 8–9 (vs. 11):

yo maukhareh. samitis. uddhatahun. asainya,valgadghat.a vighat.ayann uruvaran. anam |sam. murchitah. suravadhur varayan mameti,tatpan. 〈i〉pankajasukhaspar〈i〉śad vibuddhah. ‖ 11 ‖

62 Hans Bakker

Figure 6

Marital alliances of the royal houses of North Indiain the 6th and 7th centuries

Summarizing, we can say that, backed by the junior branch of his family inthe Gaya region and in alliance with the Pan.d. uvam. śins of Daks.in. a Kosala(Chhattisgarh), Iśanavarman, after having secured his position vis-a-visthe Aulikaras in Malwa in the west and the Vis.n. ukun.d. ins in Andhraand eastern Maharashtra in the south, succeeded in driving the Guptas,‘Imperial’ and ‘Later,’ out of their home base in Magadha before themiddle of the sixth century.

A Period of Stability and Cultural Bloom

Śarvavarman and Avantivarman

The power of the Hun. as having been curbed by his father, the son Iśanava-

The Maukharis 63

rman thus managed to keep the major forces of his time at bay, throughtactical alliances and military prowess. As we have suggested above, apeace treaty was agreed with the Vis.n. ukun. d. ins by accepting a marriagebetween his son Śarvavarman and the Vis.n. ukun. d. in princess Indrabhat.t.a-rika, and the friendship with the Pan. d.uvam. śins was reinforced by themarriage of his granddaughter Vasat.a to Hars.agupta. This strong positionallowed Iśanavarman to become sovereign of Magadha. The Later Gu-ptas would not return there before the beginning of the seventh century,when the Maukharis had run into trouble. For the first time since thecrumbling of the Gupta empire, Iśanavarman’s successful policy createdstability in Madhyadeśa, which lasted for the greater part of the secondhalf of the sixth century. He assumed the imperial title and began issuinghis own coinage. His son and grandson, Śarvavarman and Avantivarman,are known not so much for their military successes as for the culturalachievements of their reigns.170 Under them Kanauj developed into thepolitical and cultural capital of North India.171

This city had a long history before it entered into the limelight. Itsremains are covered today by a mound ‘known as Qila’ (‘40-ft. thick de-posit’), which was subjected to a small-scale excavation in 1955 by K.K.Sinha, very summarily reported in IAR 1955–56, 19–20. The finding ofPainted Grey Ware (PGW) proves that this site reaches back to the Vedicperiod. ‘Several stone sculptures,’ Hindu and Jain, ‘are datable from the4th century ad to the medieval times.’ A great number of terracotta fig-urines have been assigned ‘to a date bracket between 450 and 650.’172

If we assume that Kanauj did not significantly change between Mau-khari rule and that of King Hars.a, we may utilize Xuanzang’s descriptionof it in order to get an idea of the city.

The mound (qila) partly excavated and identified as the ancient city ‘issituated on the s. bank of the Bhagirathi near the confluence of the Gangaand Kali.’173 Xuanzang situates the town on the right or western banksof the Ganga River,174 and says that it ‘on the west borders on the river

170 Seals of Śarvavarman, Avantivarman, and the enigmatic Su. . . have been found atthe University of Nalanda (Patna District), i.e. in Magadha.

171 Wink 1992.172 An Encyclopaedia of Indian Archaeology (EIA) II (Ghosh 1989), 199. J.S. Nigam,

the author of the lemma on Kanauj in this Encyclopaedia, repeats what had beenreported in IAR 1955–56. This is preceded by a summary of earlier (surface) find-ings, also lacking in precision and system. In 1977 Nigam had already publishedsome terracottas in the Archaeological Museum at Kanauj, which had been recov-ered from the mound by the local people. What we need are new excavations thatare properly recorded and published.

173 Ghosh in EIA II, 199.174 Si-yu-ki I, 218.

64 Hans Bakker

Ganges,175 which, if we solve the apparent contradiction, probably refersto the tribute Kalı River. In other words, the ancient city lay betweenthe Ganga in the east and the Kalı in the west, in the doab formed bythe confluence of both rivers.176 Xuanzang reports that the city stretchedover a length about 6 km (20 li). Its breadth was c. 1.5 km (4/5 li) andwas defended by a ‘dry ditch,’ which, no doubt, in the event of a siege,could be flooded with Ganges water. The ‘lofty towers facing one another’were connected by a wall to complete the defences. Within these defences‘the houses were rich and well founded.’ Its inhabitants were ‘well off andcontented,’ and the layout of the city seems to have provided enough roomfor gardens, parks and ponds. That Kanauj was a centre of cultural lifefollows also from the words of the Chinese pilgrim,177 and the opulenceof Kanauj did not escape the attention of the Chinese pilgrim either:‘Valuable merchandise is collected here in great quantities.’178

The Persian poet Firdawsı reports in his Shahname the precious giftssent to the Sasanian king Khusrau Anushirvan (ad 531–579) by the ‘Rayof Qannuj,’ such as ‘jewels, shades, gold and silver, musk, amber and freshincense, rubies and silk fabrics.’ Among these presents there may also havebeen a gift to mankind as observed by Renate Syed (Plate 7):

Plate 7King Khusrau Anushirvan sits before the chess board

while his vazir and the envoy of Qannuj are playing chess 179

175 Si-yu-ki I, 206.176 This interpretation is supported by the Japanese translation of the Si-yu-ki by

Mizutani, who, in a footnote, remarks that the ‘Ganges’ here probably refers to itstribute river Kalı (information kindly provided by Yuko Yokochi). The ‘Bhagirathi’mentioned by Ghosh (above) may indicate an old channel of the Ganga.

177 Si-yu-ki I, 206f. See above, p. 15.178 Si-yu-ki I, 206.

The Maukharis 65

The Indian envoy presented the Sasanian king with the game ofchess, saying:

‘May you live as long as the skies endure! Bid those who have beenmost engaged in the pursuit of science to place the chequerboardbefore you and let each man express his opinion as to how this subtlegame is played. Let them identify each piece by name, declare howit must be moved and upon which square. They must be able todistinguish the foot-soldier [pawn], the elephant [bishop], and the restof the army, such as the rook [castle] and the horse [knight], and themovements of the vizier [queen] and Shah [king]. If they discover howthis subtle game is played, they will have surpassed all other sagesand I will then gladly send to your court the impost and tribute whichyou exact.If, on the other hand, the council of notable men of Iran fail utterlyin this science and prove themselves to be unequal with us in it, youwill no longer be able to exact from this land and territory of oursany kind of tribute or impost. You, on the other hand, will submit tothe payment of tribute; for science is superior to any wealth howevernoteworthy.’180

The caturanga thus reached Persia as a game; it had a gameboardand a set number of figures, which moved and struck according tocertain rules.181

This ‘Ray of Qannuj,’ Renate Syed argues, may have been the Maukhariking Śarvavarman.182

179 St Petersburg National Library of Russia: MS Dorn 329, 307r : ‘La date de notremanuscrit, a filets rouges et violets, est de l’annee 733 = 1332/33. Il est ecrit en car-acteres neshky et rempli de peintures d’un tres mauvais style.’ The picture has beenretrieved from http://shahnama.caret.cam.ac.uk/shahnama/faces/cardview.

180 This translation is taken from Reuben Levy’s translation of the Shahname, 327f. Itreplaces the one given by Syed. Finally the Vazir Bozorgmehr succeeds in penetrat-ing the rules of the game, much to the mortification of the envoy from Hind. He issent back to his king with a rival of chess, the game of backgammon (Nard) alongwith a similar message as the one with which he had come. The same vazir, accord-ing to the Shahname, rendered the Pancatantra (Kalıla and Dimna) into Pahlavı,a book that had secretly been copied/translated and sent from the king’s libraryin Kanauj to the court of Khusrau Anushirvan by his physician Borzuy. Both sto-ries point to an intensive cultural exchange between the Persian and Indian royalcourts.

181 Syed 2008, 365. Renate Syed discusses three terracotta figurines of ‘warriors’ (Figs.32.1–3) found at Kanauj (taken from Nigam 1977, 214–18). She argues that thewarrior figurines could be chess pieces and compares them to similar pieces ofhorsemen and elephants found elsewhere in central, northern India (Figs. 32.4–8).

182 Syed, op. cit. pp. 366f. For a different opinion see Elliot & Dowson I, 409f., whoascribe the origin of the game to King Cac of Sindh (see below, p. 114). The firstreference to the game of chess (caturanga) in Indian sources is Ban. a’s HC 2 p. 127[35]: as.t.apadanam. caturangakalpana (Śam. kara: ◦phalakanam): ‘[Under the benign

66 Hans Bakker

To substantiate this hypothesis, Syed refers to the seventh-century Nir-mand Copperplate Inscription of Samudrasena, which has been found inNirman. d. a, an agrahara, on the right bank of the Sutlej River near Rampur,c. 60 km north of Simla, c. 170 km north of Sthaneśvara, in the foothills ofthe Himalaya (Himachal Pradesh). The inscription records the donationof the village Sulisa to the local community of Atharvan. a brahmins by alocal and otherwise unknown king Samudrasena, and reinforces an earlierdonation of a plot of land that had been granted by Maharaja Śarvavarmanon the occasion of the installation of the god Kapaleśvara.183

If this Śarvavarman is the Maukhari king of Kanauj, the land grantwould indicate that Śarvavarman held suzerainty over the land belongingto the Pus.yabhutis of Sthaneśvara, whose king, at the time of Śarva-varman, may have been Adityavardhana, Hars.a’s grandfather. The formaltitle of this king, Maharaja, in Hars.a’s dynastic pedigree may point tohis feudatory status.184 The presentation of the king of Hind, the Ray ofQannuj, in the Shahname as the eastern neighbour of the Shah of Iranmay therefore be roughly correct, if we disregard, with Firdawsı, minorkingdoms, such as that of Sindh, which existed in between the two realms.

On the other hand, a king of Kanauj like Śarvavarman did not need topossess suzerainty over the countries between Kanauj and the Hindu Kushin order to enter into exchange with the King-of-Kings and the Shahnamemay simply be inaccurate when it uses the term ‘neighbour.’ A king ofKanauj making a land grant in a petty hill state more than 500 km to thenorth-west is problematic and the Śarvavarman mentioned in the NirmandCopperplate may have been another king of that name.

A similar, more likely case in which reference is made to a grant madeearlier by the Maukhari Śarvavarman seems to be presented by the BarahCopperplate Inscription (ad 836). In this inscription the Pratihara kingBhoja I of Kanauj informs the residents of the village of Valaka in theUdumbara vis.aya in the Kalanjara man. d. ala of the Kanyakubja bhuktithat he reinforces an agrahara grant that had originally been made by the

reign of King Hars.a] the (only) formation of an army is that of the eight squares.’Cf. HC 1 p. 14 [3] referring to the chequerboard: as.t.apada (= caturangaphalaka,according to Śam. kara ad loc.).

183 CII III (1888), 289 l. 9: kapaleśvaradevasya purvapratis. t.hayam. maharajaśarva-varmen. a bhumı datta. For a Kapalasthana, a siddhasthana, in connection withthe two sanctuaries of Śuleśvara and Mahakaya see SP I, 73 (SP 7.26–36) andn. 517 on p. 167.

184 Devahuti 1998, 33 comes to the same conclusion, though on different grounds. Sheargues that Śarvavarman may have passed through the territory of the Vardhanasin order to repel the Hun. as. We have argued that the Maukhari fight with theHun. as had taken place two generations earlier, but it is not unlikely that fromthat time onwards the Pus.yabhuti kings of Sthaneśvara were feudatories of theMaukharis.

The Maukharis 67

Plate 8Pillar at Deo-Baranark

Inscription of Jıvitagupta

Parameśvara Śrı Śarvavarmadeva.185 Kala-njara and its district belonged to theheartlands of Maukhari power and Śarva-varman’s engagement here is only natu-ral.186

Like his father, Śarvavarman thus pro-moted the cause of religion, in particularŚaivism. But not exclusively Śaivism, asthe eighth-century Deo-Baranark Inscrip-tion of Jıvitagupta II proves.187 In this in-scription Śarvavarman and his son Avanti-varman are named as kings who had re-newed an earlier grant made by an (as yetunidentified) Parameśvara Baladityadevato the temple of bhagavacchrı-Varun. avasi-bhat.t.araka.188

The god Varun. avasin is to be identified with the sun god Surya.189 The

185 SI II, 233–35.186 Kalanjara, the present-day hillfort Kalinjar in Bundelkhand, has very old Śaiva

connections. See further below, p. 202.187 CII III (1888), 213–18.188 Von Stietencron 1966, 242 remarks regarding the identity of this King Baladitya:

‘Versuche ihn mit Baladitya, dem Besieger des Mihirakula im Anfang des 6. Jahrh.zu identifizieren oder als Narasim. hagupta Baladitya, Sohn des Purugupta undVater des Kumaragupta um 470 n. Chr. zu datieren, mussen als fehlgeschlagenbetrachtet werden.’ The three successive generations of Bhojaka priests (for whichsee next note)—Suryamitra, Ham. samitra, R. s.imitra, donees of Baladitya, Śarva-varman, and Avantivarman respectively (CII III (1888), 216)—suggest that Bala-ditya, the founder of the temple, lived in the time of Iśanavarman of Kanauj, i.e.in the middle of the sixth century. This agrees more or less with von Stietencron’sconclusion that ‘die Bhojaka nicht vor der 2. Halfte des 6. Jahrhunderts nach Indiengekommen sind’ (op. cit. 243). We should perhaps look for a minor King Baladi-tya, whose commitment to the Sun temple was endorsed by the Maukharis. Thefragmentary Sarnath Stone Inscription of Prakat.aditya, found 150 km to the west,records the building of a temple of Kr.s.n. a by Prakat.aditya, who was the son of a patiBaladitya and his wife Dhavala. Among his ancestors was another nr.pa Baladitya.Fleet dates this inscription to the end of the seventh century on palaeographicalgrounds (CII III (1888), 285).

189 The identification of Varun. avasin with Surya is evident from the priests who werethe beneficiaries of the (reconfirmed) grant: the bhojakas. A bhojaka is an officiantof the Surya cult and this inscription appears to be the earliest attestation ofthis class of temple priests (Von Stietencron 1966, 242). Von Stietencron (op. cit.248ff.) distinguishes between the Magas, whom he considers to have been non-Zoro-astrian worshippers of the Iranian sun god Mithra (Mihira), and the ZoroastrianBhojakas. Mayrhofer, EWA 2. Teil s.v. bhoja rejects an Iranian origin of the word(against Scheftelowitz, Die Mithra-Religion der Indoskythen und ihre Beziehungzum Saura und Mithras-Kult, in: Acta Orientalia 11 (1933), p. 306: from Middle

68 Hans Bakker

Plate 9Emblem of a seal of Śarvavarman

site of this temple, ‘Varun. ika,’ is in the present village Dev or Deo-Bara-nark on the left bank of the Son in today’s Bhojpur District (the formerShahabad District) of Bihar, c. 40 km south of Arrah, i.e. in a region thatbelonged to the Maukharis from an early date.

The inscription is found ‘on two contiguous faces of a pillar in theentrance-hall of a temple on the west side of the village, which has appar-ently been adapted in modern times as a temple of the god Vishn.u.’190

According to Krishna Dev, this pillar (Plate 8) ‘is stylistically assignableto c. A.D. 500–550.’191

Broad-minded advancement of a harmonious religious and social orderseems also to be expressed in the seals of the Maukhari kings, Śarva-varman and Avantivarman. If we are allowed to interpret their beautifullycarved emblems (Plate 9) as a historical allegory, we see the cause of theDharma, described as a four-footed bull by Manu and later sources,192

Iranian *bozak, ‘saviour’).190 CII III (1888), 214.191 EITA II.1, 122. This places the original Sun temple built by Baladitya squarely

in the Maukhari period. Obviously the temple was refurbished again by the lgJıvitagupta, who had his inscription engraved on an original pillar. The place isa rich archaeological site with sculptures from the seventh century and later (seePatil 1963, 99–103; Huntington 1984, 85f.; EITA II.1, 122). For one of a dozenSurya images from Deo-Baranark see Huntington op. cit. Fig. 88.

192 As four-footed in Manu 1.81 (catus.pat sakalo dharmah. ) and as a bull in Manu

The Maukharis 69

happily making headway—the Bull attended respectfully by a guardianwith a chowry in his right hand and a plough (?) in his left,193 and led bya man in front, carrying the double battle-axe in his right and the royalstandard, topped with the cakra, in his left hand—the Law accompaniedby the allegorical figures of the king and his countryman. The Bull is led bythe cakradhara, the wielder of the cakra, who in the seal inscription belowthe emblem is likened to the Maukhari king in person, whose rule (cakra)furthers the cause of the dharmic order.194 And the double entendres donot end here. The reader may also see in the image the thriving Śaivareligion, represented by the Bull, the vahana of Śiva, who is aniconicallyrepresented by the parasol. These seals of the Maukhari kings thus seemto express through subtle metaphor the prospering under their rule of thecombined cause of the Śaiva religion and the Dharma.

The topicality of the Śaiva accommodation of the Dharma in the sec-ond half of the sixth century is nicely illustrated by a myth found in theoriginal Skandapuran. a. In SP 33.102–129 it is told how the uncontrollable,wild bull (vr.s.a) is domesticated by Śiva’s Gan. apa Prabhakara. Its frontpart is said to accommodate the Dharma and the hind part the Vedicsacrifice (Yajna); together they come to constitute the body of the bull.All deities are further said to take up residence in this bull at their specificplaces. In this way the bull is transformed into Śiva’s vehicle (vahana).195

The commitment to the Śaiva religion is no less eloquently worded inthe inscription of Śarvavarman’s brother, Suryavarman, where the princereports the restoration work he ordered on a dilapidated temple of the Foeof Andhaka, which he had seen on one of his hunting parties.

The story of the asura Andhaka in Śaiva mythology is comparativelylate. Although this myth is known to the Mahabharata, which calls Śiva‘Andhakaghatin,’196 the first extant text in which it is told in extenso is ourSkandapuran. a.197 This ‘Andhaka Cycle’ is inserted into the middle of theSkanda Cycle of the Skandapuran. a, occupying 384 pages (SPBh 73–157)

8.16 (vr.s.o hi bhagavan dharmas), and thus in the Vis.n. udharmottarapuran. a as afour-footed bull: vr.s.o hi bhagavan dharmaś catus.padah. prakırtitah. (VDhP 3.48.18).

193 Sircar in SI II, 214 sees in this an ‘ordinary long-handed double-axe,’ the axe inthe hand of the the man in front as ‘a curved double-axe on a short transversed(sic) handle.’

194 CII III (1888), 220: varn. aśramavyavasthapanapravr. ttacakrah. .195 SP 33.116–17.196 MBh 7.57.50, 12 App. I No. 28, l. 177, 13 App. I No. 6, l. 51. Cf. also MBh 7.130.38,

7.131.53, 7.172.65.197 I base myself on a presentation by Yuko Yokochi at the XIVth World Sanskrit

Conference in Kyoto in 2009, titled How to Incorporate Vais.n. ava Myths into theŚaiva Mythology, which is forthcoming in the publications series Studies in theSkandapuran. a.

70 Hans Bakker

of the total of 981 pages of text in Bhat.t.araı’s edition.198 Andhaka is theŚaiva counterpart of Vis.n. u’s demon-devotee Prahlada. The Skandapuran. amakes Andhaka the issue of Śiva, born from the total darkness that ensuedwhen Parvatı blindfolded him in jest.199 Śiva gives Andhaka to the demonHiran. yaks.a as his son, who replaces the five sons known from the Puran. a-pancalaks.an. a.200 It is a strategy by the composers of the Skandapuran. ato incorporate originally Vais.n. ava mythology into their Śaiva world (seeabove, p. 5). Andhaka is finally defeated by Śiva and, purified by his lance,becomes Śiva’s devotee and is adopted as his son (SPBh 155–156).201 Theimportant place allotted to the Andhaka mythology in the Skandapura-n. a may thus be seen as signifying a shift of religious paradigm, from areligious world dominated by Vis.n. u to one in which the Śaiva religion takespole position. This first epigraphical attestation of the figure of Andhakasubstantiates this significance.

The temple restored by Suryavarman, known as Ks.emeśvara, became‘an ornament of the earth’ and the praśasti composer, Raviśanti, opens theinscription with two mangala verses, proclaiming an accomplished Śaivitetheology and expressing the wish that the reinstated temple god, the ‘Foeof Andhaka,’ may provide the devotee with a safe haven.

He who is the cause of the acting deities that effect the manifes-tation, destruction and preservation of the world; the one on whomyogis, for whom the mass of darkness (i.e. tamas) has been destroyedand who have overcome their passions (rajas), meditate; the one inwhose heart passion has no place, although a woman occupies half(his body); he, the soul of (all) living beings, the destroyer of theTriple City, the fountainhead of bliss on earth, is victorious.202

198 To arrive at the number of pages that encompass the Andhaka Cycle proper inthe text, we have to subtract 55 pages of chapters SPBh 113–129, which deal withvarious battles between gods and demons, among which Vis.n. u’s fight with Bali(the Vamana-Trivikrama myth, SPBh 116.13–117.20) and the myth cycle of Ra-ma-Jamadagnya (Paraśurama, SPBh 121.23–124.18).

199 Cf. MBh 13.127.40–45 where this blindfolding explains Śiva’s third eye, but is notrelated to the birth of Andhaka.

200 PPL pp. 166, 196, 215.201 The Stotra spoken by Andhaka in praise of Śiva (SPBh 155.20–39) is borrowed by

the Avantyakhan. d. a of the later Skandapuran. a (1.38.11–23) and Sr.s.t. ikhan. d. a of thePadmapuran. a (33.153–171, Venk. Ed.; 28.157–175, ASS Ed.). For a study of theseparallels see Yokochi 2004a.

202 EI XIV, 115 vss. 1–2 (Metre Śardulavikrıd. ita):

lokavis.kr. tisam. ks.ayasthitikr. tam. yah. karan. am. vedhasam. ,dhvastadhvantacayah. parastarajaso dhyayanti yam. yoginah. |yasyardhasthitayos. ito ’pi hr.daye nasthayi cetobhuva,bhutatma tripurantakah. sa jayati śreyah. prasutir bhuvah. ‖ 1 ‖aśon. am. phan. inah. phan. opalaruca saim. hım. vasanam. tvacam. ,

The Maukharis 71

Suryavarman did not only restore temples, he took pride in his contribu-tion to the arts and learning as well.203 That this patronage carried withit more than mere rhetoric may be deduced from the fact that conditionswere created in which literary talent could flourish.

In his Kadambarı (1) Ban. a acknowledges that his teacher, Bhatsu,was honoured by royal Maukharis (saśekharaih. maukharibhih. kr. tarcanam. ),and one of India’s main playwrights, Viśakhadatta, worked at the courtof Avantivarman in the last quarter of the sixth century. The latter con-tention has been disputed and requires some further consideration. Butbefore doing so, some words need to be said about another, anonymousplaywright,204 the author of the Kaumudımahotsava.

The Kaumudımahotsava

Because of the simplicity of its style, this play has been compared withthose ascribed to Bhasa and is accordingly dated by Warder (1989–1992IV, 428) to the third century ad. Its protagonist is Kalyan. avarman, sonof Sundaravarman, who had been ousted from his Magadha kingdom byhis adopted son, Can. d. asena. Can. d. asena had married a Licchavi princessand had wrested Pat.aliputra from his foster father with the help of hisin-laws. As Warder (op. cit. 428) remarks, ‘the play thus purports to behistorical.’

The historicity of the play has been taken seriously by Edward Pires,who, on account of the names ending in ◦varman and their association witha Magadha dynasty (Magadhakula), takes Sundaravarman and Kalyan. a-varman as two historical, fourth-century Maukhari rulers of Magadha. Tosupport this identification Pires refers to the Chandravalli Inscription ofthe Kadamba king Mayuraśarman.205 The Maukharis/Mokaris mentioned

śubhram. locanajanmana kapiśayad bhasa kapalavalım |tanvım. dhvantanudam. mr.gakr. tibhr. to bibhrat kalam. maulina,diśyad andhakavidvis.ah. sphuradahi stheyah. padam. vo vapuh. ‖ 2 ‖

For the translation of the second verse see Introduction p. 14.203 EI XIV, 117 vs. 17 (metre Śardulavikrıd. ita):

yo balendusakanti kr. tsnabhuvanapreyo dadhad yauvanam. ,śantah. śastravicaran. ahitamanah. param. kalanam. gatah. |laks.mıkırtisarasvatıprabhr. tayo yam. spardhayevaśrita,loke kamitakamibhavarasikah. kantajano bhuyasa ‖ 17 ‖

For the translation of verse 17 see Introduction p. 14.204 The editors of this play, which is found in a single palmleaf manuscript preserved in

Kerala, conjecture that its author may have been a woman: the first two aks.arasof the supposed name have been eaten away by worms, and the remaining twolegible aks.aras point to a feminine instrumental, ◦kaya), possibly of the author’sname Vijjika. This hypothesis was further investigated and rejected by Warder(1989–1992 IV, 429).

205 See above, n. 111 on p. 41.

72 Hans Bakker

in this inscription are, according to Pires, the Maukharis of Magadha,ruling there at the end of the third and beginning of the fourth century.According to him, the play was commissioned by Kalyan. avarman after hehad regained the throne, for the occasion of his coronation, which Piresdates to about ad 326.206 Pires’s arguments are ingenious—he identifiesthe adopted son Can.d. asena with Candragupta I, who indeed married aLicchavi princes—but fail to convince in the end, due to a lack of indepen-dent historical corroboration.207 The word ‘Maukhari’ does not feature inthe play and Maukhari rulers, let alone ones who carry these names, areunknown from other sources pertaining to the history of Magadha in thethird and fourth centuries.

However, just as the Trivandrum plays (formerly ascribed integrallyto Bhasa) have been dated variously, but on the whole considerably later,by modern scholarship,208 the Kaumudımahotsava may also belong to alater period. Although essentially fictional, the Magadha setting and thefact that the Varman dynasty is hailed as its legitimate sovereign mayhave made the court of the Maukharis, after they had safeguarded theirsovereignty over Magadha, the most natural theatre for the performanceof this play.209 A Prahars.in. ı verse in the play, referring to Prince Kalya-n. avarman at his victorious entry into Pat.aliputra, would have receivedenthusiastic acclaim, if spoken on a Maukhari stage, especially after Iśa-navarman’s expulsion of the Later Guptas from Magadha and his captureof Pat.aliputra.210

Viśakhadatta

If we now return to Viśakhadatta, we have to admit that the association ofthis playwright with the Maukhari court is equally hypothetical. For thedate of this author is uncertain. This is mainly because in some MSS ofViśakhadatta’s play Mudraraks.asa (in which Candragupta Maurya is one

206 Pires 1934, 26, 168.207 Cf. Warder 1989–1992 IV, 428.208 Tieken 1993; van der Geer 1998; Bruckner 1999–2000; Salomon 2010.209 Part of the homeland of the Maukharis was, as we have seen above (p. 47), in

Magadha, to the east of the River Son, in the Gaya region. The date of the play,however, remains uncertain; it could be considerably later, written, like some ofthe Trivandrum plays, by a (south Indian) author who wished to give his play acredible historical setting.

210 Kaumudımahotsava Act 4 vs. 18:śrımadbhih. sacivasataih. * samanvito ’sau,laks.mıvan ud. univahair ivod. urajah. |prahlanam. nayanamahotsavah. prajanam. ,samprapto magadhakulankurah. kumarah. ‖* read ◦śataih.

Translation given above, p. 14.

The Maukharis 73

of the protagonists), the final blessing spoken by the chief actor featuresthe name ‘Candragupta.’ On account of this, the play has been ascribedto the time of the Gupta king Candragupta II.211 However, other MSSread instead of Candragupta: ‘Avantivarman’ (certainly the lectio diffici-lior), the name of the Maukhari king. The matter has been discussed byDhruva, the editor of the text (pp. ix–xv), who infers from the mention ofa successful campaign against the Mlecchas in the final verse—Dhruva hasno doubt that these Mlecchas are the Hun. as—that Viśakhadatta is to bedated to the time of Avantivarman in the last quarter of the sixth century.His conclusions have been adopted by Warder and we follow Dhruva andWarder in this matter.212

May that King Avantivarman guard the world for a long time tocome, while his servants and relatives enjoy prosperity—the one tothe tip of whose tusk the Earth earlier, when she was engulfed bya destructive flood, had been fastened when he, as the Self-bornGod, had assumed the body of a boar appropriate to a mode ofprotecting (the world), and the (same) one, now in the form of aking, to whose arms she has (again) been driven, terrified as she isby the Mlecchas.213

The Mlecchas, we presume with Dhruva and Warder, are the Hun. as andtheir confederates, who had seriously threatened the reigns of Avanti-varman’s ancestors and who had not yet appeared in India during Candra-gupta II’s rule.214 Moreover, as we have earlier argued (Bakker 2006,167ff.), in another, unfortunately lost but reconstructed play of Viśakhada-tta, the Devıcandragupta, this playright presented an embarrasing pictureof the internal relations at the court of the Guptas at the time of Candra-gupta, a play which it is hard to imagine being staged at the Gupta court

211 Most recently by Michael Willis 2009, 60f., who argues in favour of CandraguptaII as the subject of the blessing.

212 Warder 1989–92 III, 257. Earlier the matter was discussed to the same effect byWinternitz 1920, 210, especially n. 3.

213 Mudraraks.asa (final verse):varahım atmayones tanum avanavidhav asthitasyanurupam,yasya prag dantakot.im. pralayaparigata śiśriye bhutadhatrı |mlecchair udvejyamana bhujayugam adhuna sam. śrita rajamurteh. ,sa śrımadbandhubhr. tyaś ciram avatu mahım. parthivo ’vantivarma ‖

Other MSS finish the verse with: parthivaś candraguptah. .214 Warder 1989–1992 III, 206: ‘. . . the conclusion becomes irresistible that at the time

of the triumphant end of the Hunnish wars Viśakhadatta had a message for Indiaon the subject of the barbarian invasions. He does not oppose “Indians” or Aryansto barbarians, however. He is interested in the defence, not of a people or a nationbut of a civilisation. It is wise and just government which has to be establishedand maintained. It is the Earth which is to be protected from oppresssion by thebarbarians, as we read in the final benediction of the play.’

74 Hans Bakker

itself. Performance at the Maukhari court, on the other hand, may accountfor the living tradition reported by Ban. a in his Hars.acarita, in which, morethan 250 years later, Hars.a’s commander of the elephant troops, Skanda-gupta, warns the king of the vileness of men by telling him, among manyother examples, that the king of the Śakas, who desired another man’swife, was cut down by Candragupta who had disguised himself in herdress, the theme of the Devıcandragupta.215

As another example, the minister of elephantry had just beforereminded Hars.a that, in the not so distant past, the otherwise un-known Maukhari ‘fool’ (maukharimurkha) and lover of panegyrists,Ks.atravarman, had been assassinated by his enemy’s emissaries whohad disguised themselves as bards (mankha). This is one more, albeitsomewhat ironic, indication of the high regard in which the arts wereheld at the Maukhari court.216

It is during this period, as we will argue below (see p. 137), that thedecision was taken, now that the composition of the Mahabharata andthe Puran. a-Pancalaks.an. a had come to an end, to compose the first greatPuran. a that would do full justice to the expanding corpus of Śaiva mythol-ogy, a text that could serve the whole community of Maheśvaras, a swiftlygrowing community of laymen and ascetics. This seems a likely scenario,though it has to be admitted that there exists no direct proof that such avast literary endeavour, the composition of the Skandapuran. a, was knownto and supported by the Maukharis and their court.

To return to political history, we can observe that the threat of the Mle-cchas had not yet completely vanished during Avantivarman’s reign. Anindication to this effect, in addition to Viśakhadatta’s evidence, can bededuced from the Hars.acarita which tells us that after the wedding ofhis daughter Rajyaśrı to Grahavarman, the king of Thanesar wanted todeliver a final blow to the Hun. as by sending Rajyaśrı’s elder brother intothe field:

Then, one day, the king summoned Rajyavardhana, who had reachedthe age of bearing arms, and sent him, after he had placed him at thehead of an immense force, joined by reliable feudatories and time-serving advisors, to the north, in order to destroy the Hun. as, as thelion, master of the deer, (dispatches) his cubs against the deer.217

215 Hars.acarita (HC) 6, p. 346f. [51]: aripure ca parakalatrakamukam. kaminıves.aguptah.candraguptah. śakapatim aśatayat. For a reconstruction of the lost Devıcandraguptasee Raghavan 1963, 858–880; cf. Bakker 2006, 167f.

216 HC 6, p. 346 [51].217 HC 5, p. 250 [19]: atha kadacid raja rajyavardhanam. kavacaharam ahuya hun. an

hantum. harin. an iva harir harin. eśakiśora(ka)m aparimitabalanuyatam. cirantanairamatyair anuraktaiś ca mahasamantaih. kr. tva sabhisaram uttarapatham. prahin. ot |

The Maukharis 75

The newlywed Grahavarman, on the other hand, remained in Thanesarfor ten more days, enjoying his marital bliss, before, according to Ban. a,collecting his dowry gifts and leading his young bride and her cortege toKanyakubja. An idea of the lustre and grandeur of the capitals of thosedays may be derived from Ban. a’s description of Sthan. vıśvara at the timeof the wedding. And the Mudraraks.asa gives us an insight into the politicalskills and sophistication that were cultivated at the royal court of Kanya-kubja. But all this grandeur and sophistication would soon no longer standup to the forces of resentment that conspired against them, and they couldnot prevent Skanda—the god of war whose birth was being sung in oneof the greatest literary productions of the time—from demanding urgentattention.

Plate 10Image of Skanda (Karttikeya) found in Kanauj

A Reshuffle of Power in Madhyadeśa

The Thanesar-Kanauj Axis

Preamble

When the Maukharis established their power in Madhyadeśa by the mid-dle of the sixth century, they did so by accommodating the remains ofthe central political structures of the Gupta empire and marginalizing itsremnants in the west and the east. To the south-west of their kingdom,descendants of the Gupta viceroys of Vidiśa may have continued playinga role in local politics, be it mainly as vassals and subordinates of new,more powerful rulers such as the Hun. as, Aulikaras and Kalacuris. To theeast of their kingdom in Gaud. a (West Bengal), the dynasty of the LaterGuptas held its ground in alliance with the rulers of Kalinga (Orissa),among whom was the afore mentioned Prithivıvigraha. These two oppo-site corners became hotbeds of resistence, which reached boiling pointafter the death of Avantivarman. In the events that followed, the kingdomof Thanesar played a key role.

Sthaneśvara

We possess two contemporaneous literary sources for the early historyof Sthaneśvara/Sthan. vıśvara (Thanesar), one mythological, the Skanda-puran. a, the other semi-historical, Ban. a’s Hars.acarita. In the second partof this study their congruity will be discussed in more detail. Here it willsuffice to observe that in his Hars.acarita Ban. a ingeniously played withthe tradition of the Bhargava brahmins by linking the story of his owndescent to that of Dadhıca, the founder of Sthaneśvara according to theSkandapuran. a.

The pedigree constructed by Ban. a (see Figure 10, below, p. 159) notonly linked the Vatsyayanas to the collateral Bhargava branch of Dadhıca,it also seems to have had a geo-political dimension in that it connectedtwo regions: the one along the Sarasvatı, the ‘Side of the Sarasvatı River,’and the other along the Śon. a river, the ‘Side of the Śon. a River.’ ThePus.yabhutis belonged to the former, the Maukharis to the latter. Hars.a

77

78 Hans Bakker

united both regions within his kingdom, a feat foreshadowed by his sister’smarriage, as succinctly observed by Ban. a, when he puts the followingwords in the mouth of the wise brahmin Gambhıra, who speaks to Hars.a’sbrother-in-law Grahavarman when he enters the bridal chambers:

At last, my dear, the brilliant Houses of the Pus.yabhutis and theMaukharis have been brought together by Rajyaśrı (royal glory),when she obtained you (as her husband), two lineages whose myriadsof qualities, like the two lineages of the Sun and Moon, evoke blissin the ears of the wise, when they are sung by all the world.218

About the origins of the Pus.yabhuti dynasty itself we know nothing, apartfrom what Ban. a tells us. The Hars.acarita in chapter 3 (pp. 135–94) gives alegendary account of how the Pus.yabhutis obtained power over the land oftheir kingdom in Śrıkan. t.ha (Kuruks.etra). The alleged founder of Hars.a’sdynasty, Pus.yabhuti, is depicted as a staunch Śaiva.

At the invitation and under the guidance of a Tantric teacher hail-ing from the South (daks. in. atya), the Mahaśaiva preceptor (bhuvanaguru),Bhairavacarya, Pus.yabhuti becomes deeply involved in a black magic rit-ual (Vetalasadhana) within the cremation ground (Mahaśmaśana).219 Inthis ritual the deity (vastunaga) of the land (janapada) of Śrıkan. t.ha, inwhich Sthan. vıśvara (Pus.yabhuti’s capital) is situated,220 the eponymousNaga Śrıkan. t.ha, is conjured up by this acarya,221 and forced into submis-sion by the king, who is about to kill him with the magic sword At.t.ahasa,

218 HC 4 p. 244 [16]: tata, tvam. prapya cirat khalu rajyaśriya ghat.itau tejomayau sa-kalajagadgıyamanabudhakarn. anandakarigun. agan. au somasuryavam. śav iva pus.ya-bhutimukharavam. śau | The Maukharis traced their pedigree back to the Sun (above,n. 112 on p. 42), though they shied away from claiming to belong to the Surya-vam. śa.

219 Ban. a gives a flowery description of this teacher (HC 3, pp. 169–73 [46f.]), who,wrapped in a black woollen cloth, sits on a tiger skin spread on the ground thatwas smeared with cow dung and encircled by a line of ashes (bhasmarekha), withthe whole of the Śaiva scriptures (Śaivasam. hita) at the tip of his tongue, his headpurified by the dust of Paśupati’s feet, with a following of a throng of Maheśvaras,being a temple of grace (prasadam. prasadasya), like another Lord (bhagavantam).This teacher, who resides temporarily in a Bilva-vat.ika north of an old Matr. tem-ple, does not seem to be smeared with ashes (apart from his forehead), nor isa skull one of his attributes. His sectarian affiliation is undefined, although themacabre ritual which he performs in the cremation ground—a fire sacrifice in thecavity of the skull/mouth of a corpse on which he is seated (śavasyorasy upaviśyajatajatavedasi mukhakuhare prarabdhagnikaryam)—suggests a Śaiva belonging tothe Tantric rather than the Paśupata tradition (HC 3 p. 182 [51]; cf. Sanderson2001, 11–13).

220 The Gaud. avaho refers to Śrıkan. t.ha, as Siriam. t.ha (v. 484). Vakpatiraja describesit as the site where Janamejaya’s serpent sacrifice took place (vv. 472–484); v. 485makes it clear that Kuruks.etra is meant.

221 The goblin/deity (vetala), whose submission is sought, is evoked by the japa of theMahamantra named ‘Mahakalahr.daya’ (HC 3 pp. 178 [49], 184ff. [51f.]).

A Reshuffle of Power in Madhyadeśa 79

a gift from Bhairavacarya. Then, when the king is prepared to let himgo because of his brahminhood, a goddess dressed in white, evidentlyLaks.mı,222 emerges from the sword, annoints the king and grants thatPus.yabhuti becomes the founder of a mighty dynasty. This dynasty istherefore believed to have acquired its legitimacy and authority over theland thanks to the magic of the Śaiva ascetic.223

Both our sources, the Skandapuran. a and the Hars.acarita, seem to reflecta historical situation around the turn of the seventh century, in which anewly emerging dynasty of rulers whose names ended in Vardhana, in arebuilt capital called Sthaneśvara, sought to provide itself with respectableantecedents that could legitimate its claims to grandeur and power.224 ThePuran. a composer gave the capital a divine origin, making Śiva its founder,through the good works of his devotee Dadhıca (see below, p. 161). Thelatter was acclaimed a member of his own ancestry by the Kavya authorand he furnished the Vardhanas with an equally respectable ancestor,Pus.yabhuti,225 who obtained the legitimate right to rule his country fromLaks.mı (Rajyaśrı) herself, due to the power of Śaiva ritual and the mightof his arm.

As we have seen above, this kingdom of Thanesar under Hars.a’s grand-father Adityavardhana may have been feudatory to Śarvavarman of Ka-nauj, and his son Prabhakaravardhana may also have begun his careeras a feudatory of Avantivarman. The marriage of the latter’s son andheir apparent Grahavarman to Prabhakaravardhana’s daughter Rajya-

222 HC 3 p. 189 [53]: viddhi mam. narayan. orah. sthalılılaviharaharin. ım.223 For the gradual takeover by Śaiva officiants of the function of the royal Purohita

see Sanderson 2004. Here we seem to be concerned with an early literary exampleof this historical process.

224 Habitation at the site of the ancient city of Thanesar goes back to the days ofPainted Grey Ware. During the Kus.an. a and Gupta periods the site already con-tained a fortified city. However, excavations have revealed that in the Vardhanaperiod ‘a massive building complex’ came into existence and the ‘fortification wallwas also renovated’ (IAR 1989–90, 29). Von Hinuber 2002, 80 thinks that thename ‘Sthaneśvara’ or ‘Sthan. vıśvara’ is not attested before ad 600. However, apartfrom the evidence of the Skandapuran. a, Pande 1994, 564 (quoting P.S. Shukla inKuruks.etra 1987) reports that ‘at a site called Daulatpur, about 15 kilometres fromThanesar, terracotta sealing bearing the legend sthan. eśvarasya (sic) in 4th–5th cen-tury ad characters was found’ (cf. Bisschop 2006, 197f.). Another seal that couldpossibly support this evidence was reportedly found in the Kashmir Smast andreads sthaneśvarasya in Brahmı characters (Rahman & Falk 2011, 169 (16.01.10)).A Puran. a text that we date later than the Skandapuran. a, the Matsyapuran. a, tellsus that Devı is known as Bhavanı in Sthaneśvara (MtP 13.31).

225 HC 3 p. 163 [44] reads pus.yabhutir (supported by Śam. kara’s commentary), thoughMS K in Fuhrer’s edition (p. 151) reads pus.pabhutir. The former reading must bethe correct one, as already observed by Buhler in EI I, 68 n. 5 (cf. Ettinghausen1906, 115).

80 Hans Bakker

śrı, however, seems to have opened the road to Kanauj, which broughtsovereignty to the most ambitious of the princess’s two brothers, Hars.a-vardhana. The fictional character of Pus.yabhuti can be inferred from thefact that nowhere in his own records does Hars.a refer to this ancestor,nor does any other source for that matter, and the dubious way in whichhe obtained power in Kanauj may be one of the reasons why the Chinesepilgrim Xuanzang, when he visited Hars.a’s court at the end of his reign,was left completely uninformed about the Thanesar origins of the greatking of Kanauj.

Malava

The wedding of the Thanesar princess Rajyaśrı and Grahavarman of Ka-nauj may have taken place in the beginning of the seventh century.226 Itis not clear whether at the time of the wedding the prince had alreadysucceeded Avantivarman to the throne. Ban. a introduces Grahavarmanas the eldest son of Avantivarman, whom he calls ‘an ornament’ (tilaka)of the dynasty of the Maukharas that ‘stands at the head of all kingsand is venerated by the whole world,’227 but, as Devahuti observes, Ba-n. a seems to let Grahavarman seek the hand of the princess himself anddoes not assign any role to his father.228 Possibly Grahavarman’s matri-monial alliance with Thanesar was a move in his conflict with his brotherSuva. . . , who seems to have had his power base in Magadha (see below,p. 85). Prabhakaravardhana, on the other hand, his political clout nowincreased, assumed the grand title Paramabhat.t.arakamaharajadhiraja,229

‘one to whom other kings bowed because of his valour and affection.’230

Malwa may have been a country that actually had come within hissphere of influence. Ban. a, calling Hars.a’s father ‘an axe to the creeperof Malava’s Fortune/glory,’231 names two princes, Kumaragupta and

226 Devahuti 1998, 79.227 HC 4 p. 235 [13]: dharan. ıdharan. am. ca murdhni sthito maheśvarah. padanyasa iva

sakalabhuvananamaskr. to maukharo vam. śah. |228 Devahuti 1998, 35.229 At least, this is the title assigned to him in Hars.a’s inscriptions (e.g. EI I, 72).230 pratapanuragopanatanyarajo, EI I, 72 (cf. Thaplyal 1985, pp. 177, 182, 187). For a

detailed discussion of all the wars that Prabhakara might have waged see Devahuti1998, 70–74. This discussion is based on a number of epithets that Ban. a gives toPrabhakaravardhana: ‘famed far and wide under the second name Pratapaśıla, alion to the Hun. a deer, a burning fever to the king of the Indus land (Sindhu), atroubler of the sleep of Gujarat, a bilious plague to that of the scent-elephant thelord of Gandhara, a looter to the lawlessness of the Lat.s, an axe to the creeperof Malwa’s glory’ (transl. Cowell & Thomas p. 101; HC 4 p. 197 [1]), which praiseboils down to the historical fact that the king of Thanesar obviously held outagainst the surrounding kingdoms, but to little more.

231 HC 4 p. 197 [1]: malavalaks.mılataparaśuh. .

A Reshuffle of Power in Madhyadeśa 81

Madhavagupta, sons of the king of Malava,232 who lived at the court inThanesar. These two Malava princes cannot easily be identified with theirnamesakes in the lineage of the Later Guptas of Magadha. lg Kuma-ragupta was a contemporary of the Maukhari king Iśanavarman in themiddle of the sixth century, as we have seen, and lg Madhavagupta wasa contemporary of King Hars.a (Aphsad Inscription), who was succeededby his son Adityasena, of whom we possess the date ad 672/73.233

The ‘Malava’ brothers were assigned to Rajyavardhana and Hars.a-vardhana by their father Prabhakaravardhana as their personal aides-de-camp (anucara). The suspicion seems justified that the role of these twoboys was that of guarantors of peace, if they were not simply princelyhostages.234 The elder brother Kumaragupta was about 18 years old, theyounger brother Madhavagupta was said by Ban. a to carry the farewellkiss of his father on his head in the guise of a low crest of Malatı flowers,being as it were an imprint of fame.235 Who the father was, Ban. a doesnot reveal. We presume that the two boys were brought to the court inThanesar to stabilize the relationship between the two kingdoms.

Instability increased, however, when not much later Prabhakara-vardhana died and Thanesar became, like Kanauj, prey to successiontroubles. Immediately after the death of Prabhakaravardhana, if weare allowed to believe the author of the Hars.acarita, the fragile peacewith Malwa was broken. The king of Malava attacked Kanauj, capturedHars.a’s sister Rajyaśrı and killed his brother-in-law, the Maukhari kingGrahavarman. As in the case of the Lord of Gaud. a, the personal name ofthis deadly foe was anathema to Ban. a, and so it is left to historians tospeculate about his origin.

Devahuti (1998), and many historians with her, have assumed that theLater Guptas of Magadha, under the rule of their King Mahasenagupta,were driven out of Magadha by the Maukharis and took refuge in Malava:‘the Later Gupta king was virtually obliged to flee to Malava’ (Devahuti1998, 23). This conjecture is based on only one piece of evidence, namelythat two princes whose names concur with two kings in the lineage of theLater Guptas are said in the Hars.acarita to be the sons of the king ofMalava. The conjecture does not stand up to scrutiny.

First, we have already seen that the Later Guptas had been driven outof Magadha before Mahasenagupta, not to the west, however, but to theeast, viz. Gaud. a, where Mahasenagupta was confronted with Susthita-varman, whom he apparently vanquished on the banks of the Lauhitya

232 HC 4 pp. 229f. [11]: malavarajaputrau.233 Shahpur Stone Image Inscription, CII III (1888), 210, cf. ibid. p. 209 n. 3.234 Cf. Ettinghausen 1906, 25.235 HC 4 p. 232 [12]: anulban. amalatıkusumaśekharanibhena nirjigamis.ata gurun. a śiraśi

cumbitam iva yaśasa.

82 Hans Bakker

River (i.e. Brahmaputra).236

Secondly, it is hard to believe that foreign rulers, who had had theirhomeland in Magadha for a century or so, could so easily move to a highlydeveloped country on the opposite borders of the territory of their arch-enemies, 800 km to the west, in order to set themselves up as kings there,a belief that is an unwarranted extrapolation of the historical fact that theImperial Gupta dynasty in the fifth century had managed to hold somegrip on both regions, though with great difficulty (see Bakker 2006).

Thirdly, as we argued above, the position of the two Malava princesdoes not really correspond with that of their namesakes in the genealogyof the Later Gupta dynasty (Figure 7).237 To us it seems more plausibleto assume that, just as in Magadha after the downfall of the ImperialGuptas, a minor lineage of the Gupta nobility managed to survive andeventually set up its own royal house again in Eastern Malwa. The lastGupta known to have ruled in Vidiśa is Bhanugupta (c. ad 510) and afterhim the local Guptas may have been eclipsed due to their feudatory statusduring most of the sixth century, first under the Hun. as, then under theAulikaras of Mandasor.

When the Later Guptas were being dislodged from their Magadha domin-ions by the Maukharis and were being forced to retreat into Gaud. a proper,they may have tried to drive a wedge between the Maukhari kingdom andthe rulers to the west of it, thus taking recourse to the time-honouredArthaśastra policy of encirclement (man. d. ala). Whether the Guptas of

236 Aphsad Inscription, CII III (1888), 203 ll. 10–11. In the ‘International Seminar onCoinage of the Imperial Guptas and Contemporaries: New Perspectives’ (Chandi-garh, 11/12–11-2010), Prashant Kulkarni showed one of two gold coins of Maha-senagupta found recently in Bangladesh.

237 As proof of the alleged identification of the Malava with the Later Gupta Ma-dhavagupta reference is made to the Aphsad Stone Inscription of Adityasena line15 which contains a partly illegible Vasantatilaka verse that runs as follows (CIIIII (1888), 204):

ajau maya vinihata balino dvis.antah. ,kr. tyam. na me ’sty aparam ity avadharya vırah. |śrıhars.adevanijasam. gamavanchaya ca, ˘ ˘ ˘ ˘ ˘ ˘ ˘ ‖‘Considering: “There is nothing left for me to do, as I vanquished my mightyenemies on the battlefield,” and wishing to align himself with the illustriousHars.adeva, the hero [Madhavagupta] . . . .’

What Madhavagupta did to align himself with Śrıhars.adeva is anybody’s guess, butthe next thing reported in the inscription is the birth of his son Adityasena. Thelatter’s mother is named Śrımatı (ibid. p. 204 line 24, cf. ibid. p. 215). A plausibleguess would be that lg Madhavagupta married a princess of the Vardhana house(a daughter or sister of Śrıhars.adeva?), named Śrımatı. If this conjecture is correct,it would indicate that the dynasty of the Later Guptas, after the Maukhari warsand the Śaśanka episode, had concluded a peace with the king of Kanauj, Hars.a-vardhana (see below, p. 92, n. 412 on p. 131).

A Reshuffle of Power in Madhyadeśa 83

The Later Guptas

nr.pa śrı Kr.s.n. agupta

deva śrı Hars.agupta

śrı Jıvitagupta (I)

śrı Kumaragupta

śrı Damodaragupta

śrı Mahasenagupta

[. . . ] śrı Madhavagupta x mahadevı Śrımatı

maharajadhiraja Adityasenadeva x mahadevı Kon. adevı

maharajadhiraja Devaguptadeva x mahadevı Kamaladevı

maharajadhiraja Vis.n. uguptadeva x mahadevı Ijjadevı

maharajadhiraja Jıvitaguptadeva (II)

Figure 7

The Later Guptas of Magadha

84 Hans Bakker

Malava were originally involved in this scheme we do not know, but thefact that the Maharaja of Thanesar, Hars.a’s grandfather Adityavardhana,had married a princess named Mahasenagupta is often seen in this light,since she is generally considered to be a sister of lg Mahasenagupta (seeFigure 8).238

However, the Maukharis seem to have kept their grip on the Pus.ya-bhuti kingdom, as we have seen above (p. 66), and responded in kind.239

The marriage of Adityavardhana’s granddaughter Rajyaśrı to the Mau-khari King Grahavarman indicates that the rulers of Thanesar of the nextgeneration preferred a strong ally in Kanauj to a weak one in the far east(Gaud. a). An attack on the Thanesar-Kanauj entente became imminentwhen the political situation in both capitals was seriously undermined bysuccession troubles.

The Malava–Gaud. a conspiracy

The above developments made the rulers of Malava and Gaud. a the nat-ural enemies of the axis of power in Madhyadeśa. At the beginning ofthe seventh century, when Prabhakaravardhana died, at a crucial phaseof the history of North India—dramatically reported in Ban. a’s history—they saw their chance. We possess a moving literary description of whatthe death of this king—great in the eyes of his contemporaries—broughtabout, how it was experienced by the court, and which rituals and cere-monies were performed to cope with the calamity. In Hars.acarita 5 Ban. atells us that, even before the actual demise of Hars.a’s father, the latter’sfirst wife, Queen Yaśovatı, was so overwhelmed by grief and the prospectof widowhood that she, against the express wish of her son, committedherself to the flames (i.e. became a satı), while her husband was still alive.After the death of his mother, Hars.a goes to his dying father and claspshis feet in despair. The latter comforts him, recommends him to his royalduties, and utters his last words: ‘enemies should be exterminated.’240

The body is brought to the banks of the Sarasvatı on a bier, a funeralpyre is built and the cremation ritual is performed at dusk. It is suggestedthat the remaining ladies of the king’s harem also committed satı. Hars.akeeps a vigil and next morning goes to the Sarasvatı to bathe, offers anoblation of water to his father and goes home in distress (p. 293). Abrahmin eats the first pin. d. as offered to the ghost (preta) and the periodof impurity passes. The deceased king’s paraphernalia are given to thebrahmins; the collected bones are brought to holy places. Then it is saidthat a ‘stab of pain’ was set up in the form of a monument at (the place

238 Devahuti 1998, 30.239 Cf. Thaplyal 1985, 31.240 Hars.acarita (HC 5 p. 288 [32]): niravaśes.atam. śatravo neyah. .

A Reshuffle of Power in Madhyadeśa 85

of) the funeral pyre made of a mass of bricks.241 Finally the royal elephantis set loose and gradually the lamentations subside. The metaphor usedby Ban. a, ‘a stab of pain’ (śokaśalya)—the primary meaning of śalya being‘dart’—suggests that the monument was needle-shaped.242

Regardless of this outburst of grief, the succession of Prabhakara-vardhana was a complicated affair. It would seem that the Pus.yabhutithrone was disputed by the two sons, who, according to Ban. a, were bothinitially ‘hesitating,’ in all likelihood a euphemism for ‘rivalling,’ indicat-ing a standoff.243 The Maukhari throne in Kanauj had also been weakenedby this time, as revealed by a Maukhari seal found in Nalanda testifyingto the existence of another claimant to the throne, Grahavarman’s brother‘para. . . raja[dhi]raja-śrı Suva. . . (or ◦Suca. . . ), who, Devahuti argues, wasa rebel.244

The rulers of Malava and Gaud. a launched an attack on Kanauj.Grahavarman was killed, his wife imprisoned by the Malava king.245

Hars.a’s brother Rajyavardhana marched on Malava in revenge and slewits king. In the inscriptions of Hars.a a verse is quoted in which it isrecorded that Rajyavardhana defeated a king named Devagupta,246 and

241 Hars.acarita (HC 6 p. 300 [36]): kalpitaśokaśalye sudhanicayacite citacaityacihne,which Śam. kara Kavi glosses: citayam. caityacihnas tadakaram. cihnam, śmaśanade-vagr.ham. va |

242 Cf. Anguttara-Nikaya III, 62: Sokasallaharan. a, which is the name of the discourse(dhammapariyaya) that takes away the grief of bereavement, and which is taughtto King Mun. d. a. Such a funerary monument is known in Sanskrit literature as aned. uka/aid. uka; see VDhP 3.84.1–15 and Bakker 2007a.

243 Devahuti 1998, 80f. See below, p. 165.244 Devahuti 1998, 35f.:

‘It is probable that ‘Suva’ had been appointed the governor of Magadha by hisfather [i.e. Avantivarman], on whose death he declared his independence. Thedivision of the ancestral domains at this juncture was fatal for the Maukharis,because the new dynasties that had fed themselves on the remains of the ImperialGuptas had by now achieved strength and stability in their respective territoriesin different parts of north India and the time was once again ripe for a strongand capable man to establish his paramountcy.’

245 HC 6 p. 314 [41]: yato yasminn ahany avanipatir uparata ity abhud vartta tasminneva devo grahavarma duratmana malavarajena jıvalokam atmanah. sukr. tena sahatyajitah. | bhartr.darikapi rajyaśrıh. kalayasanigad. ayugalacumbitacaran. a cauranga-neva sam. yata kanyakubje karayam. niks. ipta |

246 EI I, 72; cf. Thaplyal 1985, 178, 183:rajano yudhi dus.t.avajina iva śrıdevaguptadayah. ,kr. tva yena kaśapraharavimukhah. sarve samam. sam. yatah. |utkhaya dvis.ato vijitya vasudham. kr. tva prajanam. priyam. ,pran. an ujjhitavan aratibhavane ’satyanurodhena yah. ‖‘[Rajyavardhana], who, after he had brought all kings equally into submission,beginning with Devagupta, making their faces turn away (as they took to theirheels), on the battlefield, by applying his whip as if they were wicked horses,and who, after he had uprooted his enemies, had conquered the earth, and had

86 Hans Bakker

the odds are that this Devagupta was the father of Kumaragupta andMadhavagupta.247 The Malavas properly dealt with, it was the king ofGaud. a who remained to be shaken off. According to the Hars.acarita, thevictorious Rajyavardhana was lured into a trap set for him by the king ofGaud. a. Drawn into his camp under false pretences, he was treacherouslymurdered.248 His brother Hars.avardhana and the king of Gaud. a were theonly two antagonists who remained.

Hars.avardhana’s consecration

Hars.a was a young man of about fifteen years old when in c. ad 605 thenews was brought to him (in Thanesar) by an officer of cavalry, Kuntala,

pleased his subjects, gave up life in the house of his enemy due to the breakingof a promise.’

247 Cf. Buhler in EI I, 70; Ettinghausen 1906, 37. Devahuti is forced to indulge iningenious speculation to salvage her conviction that the two Malava princes weresons of the lg Mahasenagupta:

‘We have suggested earlier that Mahasenagupta, after being expelled from Maga-dha, chose to live in Malava because of his possible earlier connections with thatregion. A collateral branch of Mahasenagupta’s line may have existed in Malavafor a long period, and Devagupta may have been a member of that line. He mayhave sided with Śankaragan. a against Mahasenagupta and the former may haverewarded him with subordinate kingship over Malava. Though Mahasenaguptawas an old man and too weak to re-establish his authority over Malava, Prabha-karavardhana, with the ambition to increase his own power, seems to have cham-pioned the cause of his uncle [i.e. Mahasenagupta, supposed to be the brother ofPrabhakara’s mother, H.T.B.]. A war against Devagupta, though devoid of deci-sive results, seems to have formed the basis of Ban. a’s statement. Mahasenagupta,at this time of great stress, decided to send his two young sons to the court ofSthan. vıśvara’ (Devahuti 1998, 74; cf. Thaplyal 1985, 42).

We admit that it is quite possible that the local Guptas in eastern Malava (Vi-diśa) were reinstalled by the Kalacuri King Śam. karagan. a by the end of the sixthcentury, as this king issued a charter in ad 597 from his ‘victorious camp (vasaka)in Ujjayanı,’ in which he claimed ‘to reinstate royal dynasties that are long-timegone’ (cirotsannanam. nr.pativam. śanam. pratis. t.hapayita, CII IV.1, 41), but this hecould have done just as well or even better, in our view, if there were no Later Gu-ptas of Magadha in Malava. There is not a shred of evidence that Mahasenaguptaplayed an important role in the politics of Eastern Malwa in the last decade ofthe sixth century and ‘his possible earlier connections with that region’ are simplyunknown.

248 HC 6 p. 321 [43]: tasmac ca helanirjitamalavanıkam api gaud. adhipena mithyopa-caropacitaviśvasam. muktaśastram ekakinam. viśrabdham. svabhavana eva bhrata-ram. vyapaditam aśraus. ıt. Cf. Hars.a’s inscriptions quoted above, n. 246 on p. 85.In recent historiography the opinion has gained ground that the murder of Rajya-vardhana may not have been unwelcome to his younger brother and that the lattermay even have taken part in a conspiracy (e.g. Devahuti 1998, 80ff.). Most recentlyall the arguments in favour of such a conspiracy have been summed up by ShankarGoyal (2007, 200), the apologete of his father S(hri) R(am) Goyal. Without furtherevidence, however, none of this can transcend the status of historical speculation.

A Reshuffle of Power in Madhyadeśa 87

that his elder brother had been killed by the king of Gaud. a.249 The youngprince takes an oath, swearing to commit himself to the flames, unless hesucceeds in destroying the Gaud. as and all who make common cause withthem within a limited number of days.250

Preparations are made for a conquest of the world (digvijaya) and aceremony is performed. Ban. a describes the abhis.eka ritual in some detail(see below, p. 162). It is concluded by Hars.a coming forth from his palaceamid cries of victory. Rather than as a coronation ceremony,251 this shouldbe conceived of as a rite of passage, which consecrated a young prince, thekumara, as commander in chief of the army (senapati).252

Ban. a continues by telling us that the base for starting the military cam-paign was fixed outside the city, at the Sarasvatı River, where a makeshiftconstruction of bamboo was erected, provided with a high gateway andan altar, which was thronged by people dressed in white and brahminsreciting texts. Here Hars.a received the signet (mudra) containing the em-blem of the Bull, which, accidentally falling from his hand, stamped theearth on the Sarasvatı banks with the royal seal. Then, at the close of thethird watch of the night, the drums of war were beaten. Ban. a providesus with an elaborate review of the assembled army, which moved off atthe break of day, headed by the prince commander on his elephant to theaccompaniment of musical instruments.253

Though Ban. a’s description of Hars.a’s consecration as chief of the army,after earlier having declined kingship, seems to follow a literary conventionthat has its basis in a mythic archetype (see below, p. 162), we assumethat the historical reality underlying his literary imagination was thatHars.a’s sister had been captured, that his brother had been killed, thathe was installed as senapati in Sthaneśvara, and that he marched againstKing Moon of Bengal. The eventual recovery of his sister, the dowagerqueen Rajyaśrı or ‘Royal Glory,’ which concludes the Hars.acarita, is Ba-

249 HC 6 p. 321 [43].250 HC 6 p. 335 [47].251 Heesterman 1957, 50: ‘The ratnin episode is one of the few parts of the rajasuya that

are directly and exclusively related to kingship.’ These Ratnin offerings are absentfrom the ritual described by Ban. a. Moreover, a the Vedic Rajasuya ceremony takesone year to accomplish.

252 See for this also Xuanzang’s report in the Si-yu-ki I, 210–13, which is distorted,however, by Buddhist zeal (cf. Devahuti 1998, 92–94). Though Xuanzang situatesthese events in Kanauj, he does not describe the royal coronation, it seems, butHars.a’s installation as chief of the army. The advice of the Bodhisattva, ‘Ascendnot the lion-throne and call not yourself Maharaja’ (Beal I, 213), as well as thepilgrim’s report that: ‘He called himself the King’s Son (Kumara),’ fit the situationin Thanesar and the years of the campaign against Śaśanka. As Ban. a’s biographyand the king’s own epigraphs evince, however, these two conditions were removedby a coronation in Kanauj.

253 HC 7 pp. 353–60 [54–56].

88 Hans Bakker

n. a’s ingenious allegorical justification of King Hars.avardhana’s ascensionto the throne of Kanauj.

ŚaśankaThe enemy, the Lord of Gaud. a (gaud. adhipa), is, as we have noted above,not mentioned by name, but Ban. a seems to refer covertly to it once inthe form of a metaphor that describes the inauspicious portents seen byHars.a: the rising disk of the blotted moon (śaśanka).254 Śaśankadeva iscommonly identified with a feudatory (mahasamanta) in the territory ofthe Maukharis known from a seal matrix that is reportedly cut in the rockat the hillfort of Rohtasgarh or Rohitagiri on the river Son.255 If this is theŚaśanka who is alluded to by Ban. a, this author must have known him verywell, at least at the beginning of his career, since the seal is reportedlyfound at the Rohitagiri (Rohtasgarh), not too far from Prıtikut.a, Ban. a’shome village, which is said to be situated on the western bank of the Śon. aRiver where it flows out of the Vindhya.256

However, ever since its publication by Fleet in the Corpus Inscrip-tionum Indicarum 1888, this seal matrix has remained untraceable.257

The suspicion seems justified that the assignment of the seal to the hill-fort Rohtasgarh is a mistake, made either by Fleet himself or by Beglar

254 HC 6 p. 306 [38]: prakat.akalankam udayamanam. viśankat.avis. an. otkırn. apanka-sam. karaśam. karabarkuraśakvarakakudakut.asam. kaśam akaśatakaśe śaśankaman. d. a-lam | This interpretation is confirmed by Śam. kara, who in his commentary on the2nd opening verse of Ucchvasa 6 (ibid. p. 299) glosses ‘khalasya’ ( ‘the wickedone’ whose destruction is caused by the anger of the mighty one) as: khalo ’tragaud. apasadah. | nidrataskarah. śaśankah. | (‘The wicked one here is the degenerateGaud. a, the pain-in-the-neck Śaśanka’). Another tacit reference to Śaśanka maybe found in the speech of Hars.a’s general Sim. hanada, who says about the enemy:‘But how could Fortune (Laks.mı) stay with a coward longer than two nights, thecoward, who has the heart of a deer/has a deer inside (i.e. the moon), and whoseback/surface is pale, like the moon (śaśin)’ (HC 6 p. 329 [46]). Sam. kara ad loc.:‘pan. d. urapr.s. t.hasya’ (‘whose back is pale’) deśabhas.aya nirlajjasyapi | (shameless).

255 CII III (1888), 283f. Devahuti 1998, 38: ‘As the beginning of Śaśanka’s careermay be reasonably placed around A.D. 600, he must have been a feudatory of theMaukharis, not of the Later Guptas.’

256 HC 1 p. 30 [8]: [. . . ]apaśyac [subject: Sarasvatı] cambaratalasthitaiva haram iva va-run. asya amr.tanirjharam iva candracalasya śaśiman. inis.yandam* iva vindhyasyakarpuradrumadravapravaham iva dan. d. akaran. yasya [. . . ] svacchaśiśirasurasavari-purn. am. pitamahasyapatyam. hiran. yavaha**namanam. mahanadam yam. janah. śon. aiti kathayanti |We take this to mean that the village lay at the foot of the Vindhyahills, i.e. somewhere in the vicinity of Akbarpur (24o 31’ N, 83o 54’ E), where theŚon. a River enters into the Gangetic Plain (‘the ooze of the Vindhya’s moon-gem,’tr. Cowell and Thomas), and loses its turbulence (some MSS in Fuhrer’s ed. read*nis.pandam and this edition accepts the variant **hiran. yabahu as the name of theriver), not far from the Rohtasgarh hillfort, which lies about 2.5 km to the west ofAkbarpur.

257 I am grateful to Dr Gautam Sengupta (ASI) who brought this to my attention.

A Reshuffle of Power in Madhyadeśa 89

to whom Fleet ascribes its discovery.258 What is certain is that Fleet didnot see the seal himself, but based himself on ‘Mr Beglar’s rubbing.’ Ir-respective of whether Śaśanka left his trace on the Rohtasgarh Hill, allour sources confirm the picture that at the turn of the seventh century,when the political stability in Thanesar and Kanauj was put under heavypressure by the vicissitudes of succession, a feudatory named Śaśankausurped the Later Gupta domains in Bengal, i.e. Gaud. a, and conqueredMagadha.259 If he himself was not related to the Later Guptas—one manu-script of the Hars.acarita apparently names him Narendragupta260—fromthis time onwards he certainly joined their cause, or vice versa.

When the Later Guptas were driven out of Magadha into Gaud. a proper,they may have come into conflict with the rulers of that province, notablySamacaradeva, although we do not have direct evidence of this.261 TheAphsad Inscription tells us only of Mahasenagupta’s victory over the king

258 CII III (1888), 283f.259 This episode seems to be reported in the Aryamanjuśrımulakalpa (MMK)

53.657–660 (Ganapati Shastri’s edition 1925, e-text: http://gretil.sub.uni-goettingen.de/gretil/1_sanskr/4_rellit/buddh/bsu041_u.htm), whereŚaśanka is called Soma (text partly emended):

somakhyo ’pi tato raja ekavıro bhavis.yati |gangatırasya paryantam. * varan. asyam atah. param |naśayis.yati durmedhah. śastur bimban manoraman** ‖jinais tu kathitam. purvam. dharmasetum analpakam |dahapayati durmedhah. tırthikasya vace ratah. ‖tato ’sau kruddhalubdhas tu mitthyamanı hy asam. matah. |vihararamacaityam. ś ca nirgranthavasathan*** bhuvi |bhetsyate ca tada sarvan**** vr. ttirodhanakarakah. ***** ‖

Editions read: * gangatıraparyantam. (unmetrical); ** bimbam. manoramam; **nirgrantham. vasatham. ; **** sarvam. ; ***** vr. ttirodhamakaraka. Cf. Jayaswal &Sankr.ityayana 1934, vss. 715–18, and Jayaswal’s translation p. 49f. After detaileddiscussion of all available sources, Devahuti 1998, 45 concludes: ‘Having overrunthe greater part of Magadha, taking in his sweep Banaras, Kuśı-nagara, Gaya, andPat.ali-putra, Śaśanka turned towards Gaud. a, which appears to have been an easyprize on account of Jaya-naga’s timely removal from the scene [see below, n. 264on p. 90, H.T.B.]. Śaśanka was now definitely a power to be reckoned with.’

260 Buhler in EI I, 70 referring to the ‘Kasmir edition.’ This may refer to the editionthat was published in Jammu in ad 1879. A revised edition with the commentarySanketa of Śam. kara was published by Kaśınath Pan. d. urang Parab in Bombay inad 1897 (3rd ed. in Bombay 1912). These editions were not available to me andthe variant could not be traced in Fuher’s edition (which is based on the Jammuedition). Devahuti 1998, 37f: ‘Some scholars, on the grounds of similarities betweenthe coins of the Imperial Guptas and those of Śaśanka, as well as on the basis thatthe variant Narendra-gupta is used for Śaśanka in one manuscript of the Hars.a-carita, have tried to establish his relationship either with the Imperial Gupta orwith the Later Gupta dynasty. Both contentions are inadequately supported.’

261 Cf. Devahuti 1998, 22.

90 Hans Bakker

of Kamarupa, Susthitavarman. As we have seen above, Mahasenaguptamight have sought to strengthen his position by a matrimonial alliancewith the Pus.yabhutis of Sthaneśvara, but evidently this was of little avail.If indeed Mahasenagupta had been a contemporary of Hars.a’s grandfatherAdityavardhana (married to a Mahasenagupta), he may no longer havebeen alive when the succession troubles in Sthaneśvara and Kanauj beganin the early seventh century, troubles that offered Śaśanka his opportunityfor fame.

Before these events, however, one or two expeditions of a Gaud. a armyinto Kamarupa had taken place. S.R. Goyal (2005, 198) has argued thatMahasenagupta acted as a feudatory of Śaśanka in this expedition. Hetried to substantiate this thesis by juxtaposing the Aphsad Inscriptionand the Dubi Plates of Bhaskaravarman, arguing that both inscriptionsrefer to one and the same event, viz. Mahasenagupta’s invasion of Kama-rupa.262 However, in the Aphsad Inscription the adversary of Mahasena-gupta is said to have been Susthitavarman, whereas in the Dubi Platesthe Gaud. as are said to have fought his two sons, viz. Supratis.t.hita andBhaskara, after their father had died.263 The brothers lost, were capturedand taken to Gaud. a. They were soon released again, and returned to theirown kingdom, probably as feudatories of the Gaud. a ruler.264 Supratis.t.hi-

262 Goyal 2005, 192–98.263 Sircar 1950, 241ff., EI XXX (1953–54), 287–304 and SI II, 1–15. The phrase pitari

ks.mantarnilıne may mean that he had disappeared into the earth, i.e. had diedor ‘died when the Gaud. a invasion was going on’ (Goyal 2005, 194) and had beenburied. Verse 68 (p. 303, ll. 94f.) runs (emended):

yav etau prathame vayasy api pr. thupraspardhisattvodgamau,śakram. śam. vidhina pragatya pitari ks.mantarnilıne kramat |prapte gaud. abale baliny api jale viśrambhasam. rambhatah. ,stokair eva balacyutav iva balair yau lılayopasthitau ‖‘These two (princes), although they were still in their early youth, equalling Pr.thuin their ascent to virtue, inherited legally their share of power, in due order, whentheir father had disappeared into the earth (?). Then, when the army of the Gau-d. as had arrived, they, due to recklessness, met (the enemy) with a small armyonly, (as if) in sport, like Balarama and Kr.s.n. a, although (the Gaud. as) had comewith a strong fleet.’

264 Devahuti (1998, 44) thinks that this king of Gaud. a was Jayanaga. Devahuti arguesat length (1998, 38–45) that this king, known from the Vappaghos.avat.a Inscrip-tion (EI XVIII (1925–26), 60–64; Sanyal 2010, 115f.) and coins with the legendjaya and the cakra emblem (obverse) and śrıprakan. d. ayaśah. (reverse) (Allan 1914,150–51), ruled between Mahasenagupta and Śaśanka in Karn. asuvarn. a (Śaśanka’scapital). Her argument is based on very slim evidence indeed and seems to havebeen especially dictated by the fact that this scholar has Mahasenagupta put asidein Malava, thus leaving a vacuum in Gaud. a. Cf. R.C. Majumdar 1971, 73: ‘Thedate of Jayanaga cannot be ascertained with precision, but judging from his coinsand inscription, he may be placed within the period 550–650 A.D. On the basis ofthe tradition recorded in Manjuśrımulakalpa we may hold that after the anarchy

A Reshuffle of Power in Madhyadeśa 91

ta apparently died shortly after he had succeeded his father as the elderson. His younger brother Bhaskaravarman became king of Kamarupa andoffered allegiance to King Hars.a in his war against Śaśanka (HC 7).

The Aphsad Inscription mentions Madhavagupta as Mahasenagupta’s suc-cessor, of whom it is said that he later aligned himself with Hars.adeva(i.e. Hars.avardhana).265 As argued above, we do not believe that this Ma-dhavagupta, son of Mahasenagupta, was the prince of Malava that grewup at the court of Sthaneśvara. He may have taken part, on the otherhand, in the Gaud. a army that caused the downfall of the Maukharis andtemporarily occupied Kanauj, since this army seems to have been, ratherthan a homogeneous enemy, a confederation of all those who held a grudgeagainst the Maukharis, spearheaded by Śaśanka.

Backed by all those who had an axe to grind, and in alliance withthe ruler of Malava, Śaśanka had set out to dismantle the power of theMaukharis. Initially this campaign of the ‘Lord of Gaud. a’ was successful,and it would seem that his army temporarily occupied Kanauj after theelimination of Rajyavardhana. The Pus.yabhuti general Bhan. d. i informsHars.avardhana about what had happened in Kanyakubja after Rajya-vardhana had been murdered: ‘Your Majesty, “while Kuśasthala (Kanya-kubja) was controlled by a man named Gupta and King Rajyavardhanahad reached heaven, Princess Rajyaśrı escaped from her captivity andhas entered the Vindhya forests with her retinue.” Thus I heard peoplereport.’266 An imprisonment, we may add, that had earlier been imposedon her by the Malava king. And this makes it impossible that by ‘Gupta’the Malava king Devagupta is meant, because he was already dead, or atleast slain, by the time that Rajyavardhana was murdered.

The same event is narrated again at the end of the Hars.acarita (p. 438).Here Hars.a hears from Rajyaśrı’s own attendants that, during the Gaud. atroubles (gaud. asam. bhramam. ), his sister fled from Kanyakubja, being setfree by a nobleman named Gupta (guptanamna kulaputren. a). Obviouslythe nobleman named Gupta belonged to the party of the Gaud. as, butwho he may have been is less clear. Is it possible that this man of goodfamily was Madhavagupta, son of Mahasenagupta? After all, Rajyaśrı mayhave been a ‘kissing cousin’ (see Fig. 8), and at a certain point this lg

and confusion caused by the invasion of Bhaskaravarman had subsided, and a sonof Śaśanka had vainly tried to re-establish the fortunes of the family, the kingdompassed into the hands of Jayanaga.’

265 See above, n. 237 on p. 82.266 HC 7 p. 395 [67]: deva devabhuyam. gate deve rajyavardhane guptanamna ca gr. hıte

kuśasthale devı rajyaśrıh. paribhraśya bandhanad vindhyat.avım. saparivara pravis. t.etilokato varttam aśr.n. avam | Some manuscripts (Fuhrer’s ed. p. 302) read ‘while theGaud. as controlled’ (gaud. air gr. hıte).

92 Hans Bakker

Vardhanas of Thanesar Guptas of Magadha

Damodaragupta

Adityavardhana x Mahasenagupta

Prabhakaravardhana Mahasenagupta

?Rajya Hars.a Rajyaśrı Śrımatı x Madhavagupta

Adityasena (ad 672)

Figure 8

Matrimonial relationship of the Pus.yabhutis and Later Guptas

Madhavagupta aligned himself with Hars.a, who, once his rule had beenfirmly established, may have turned their enmity into an alliance, whichwas confirmed by matrimony (see above, n. 237 on p. 82). Contrary to whathas often been suggested, however, there is no evidence for the assumptionthat Hars.a made the lg Madhavagupta his governor in Magadha.267

We do not know whether Śaśanka himself briefly occupied Kanauj,but his sway over the city only lasted until Hars.avardhana took up thewar. Ban. a tells us that Hars.a entered into an alliance with the ruler ofPragjyotis.a (Kamarupa) Bhaskaravarman, a deadly enemy of the rulersof Gaud. a, as we have seen (p. 90), and he sent his army under GeneralBhan. d. i towards the East. Śaśanka was forced to retreat but by no meanssurrendered.268

In about ad 606 the political situation had stabilized enough to organ-ize a royal coronation ceremony. Hars.avardhana was enthroned in Kanauj,an event only hinted at by metaphor in the Hars.acarita (above, p. 88). Itwould take Hars.a six more years, however, if we follow Xuanzang’s re-port in this,269 to consolidate his sovereignty over the combined heredi-tary lands of the Pus.yabhutis and Maukharis, and before finally, to usethe closing metaphor of Hars.a’s Deeds, Fame, Glory and Force united to

267 Xuanzang ‘records the names of two kings, Purn. a-varman and Śaśanka, in connec-tion with Magadha, but not that of Madhava-gupta’ (Devahuti 1998, 19).

268 See below, n. 301 on p. 103.269 Beal I, 213, above, n. 252 on p. 87.

A Reshuffle of Power in Madhyadeśa 93

hand over to him the Moon.270 When Śaśanka was thus on the wane,seven years after the young prince had sworn his oath, King Hars.a as-sumed grand titles and an imperial lifestyle. It must have been aroundthis time that Ban. a was brought to the monarch in his encampment oneevening, an overwhelming experience that inspired the apotheosis withwhich he concluded his chef-d’oevre.

270 HC 8 p. 453 [86]: jvalatsam. dhyaragarajyamanajalapravahah. punar iva puran. apuru-s.apıvarorusam. put.apis. t.amadhukait.abharudhirapat.alapat.alavapur abhavad adhipatirarn. asam | samavasite ca sam. dhyasamaye samanantaram aparimitayaśah. panatr.s. i-taya muktaśailaśilacas.aka iva nijakulakırtya, kr. tayugakaran. odyatayadirajarajata-śasanamudraniveśa iva rajyaśriya, sakaladvıpajigıs. acalitaya śvetadvıpaduta iva ca-yatya, śvetabhanur upanıyata niśaya narendraya ‖ For a paraphrase see above,p. 17.

The Reign of Hars.avardhana (ad 612–647)

A literary and a Buddhist record

Ban. a’s introduction to the court

Once Hars.a had thus successfully driven back the Gaud. as and had es-tablished his authority in Madhyadeśa, the life-long project to keep hisenemies at bay began. The eastern frontier of his realm may initially havebeen somewhere between the present Gandak and Burhi Gandak rivers.According to the Hars.acarita, the army headquarters (skandhavara) werepitched on the river Ajiravatı, which has been identified with the old bed ofthe Gan.d. akı River, to the east of the present course of the Gandak, neara village called Man. itara (a variant reads Man. ipura). This village hasnot been identified, but it may have been somewhere within the presentMuzaffarpur District of North Bihar. It was here that the poet Ban. a sit-uates his meeting with the ‘Maharajadhiraja Parameśvara Śrıhars.adeva’(HC 2 p. 83 [23]), a camp that he reached after a three-day journey fromhis native village of Prıtikut.a on the River Son (see Plate 11).271 Ban. agives a lengthy description in ornate style of what it was like for a youngambitious poet to be received at the court of the emperor, Lord of theFour Oceans.

The seraglio surrounding the mobile royal residence (rajabhavana) wasteeming with princes and their retinues of servants, elephants, horses, and

271 The itinerary of the poet (described in the HC 2 p. 91f. [26] ), if historical, will havebeen partly by boat, downstream on the River Son. On the first day Ban. a passedthrough a forest, named Can. d. ikayatanakanana, and arrived at a village calledMallakut.a. On the second day he arrived at the Ganges, which he crossed. Theconfluence of the Son and Ganges was situated c. 18 km to the east of the presentconfluence, at the western outskirts of today’s Patna, according to a map preparedby the Department of Geology of the Patna University and kindly placed at mydisposal by Ms Daniela De Simone. The distance from Rohtasgarh (presumablynear Prıtikut.a) to this confluence is c. 150 km, which would imply that Ban. a couldcover about 75 km a day, which seems only possible by boat, going downstream.The poet stayed the second night in a forest village named Yas.t.igr.haka on the leftbank of the Ganges. On the third day he reached Hars.a’s headquaters at Man. itaraon a river called Ajiravatı.

95

96 Hans Bakker

camels—expectants of all sorts, who anxiously awaited the day that anaudience with the great king would be granted. Not only allies, feudatories,vassals and suppliants were waiting for darshan, but also holy men ofvarious persuasions, Jainas, Buddhists, Paśupatas, ascetics of the Para-śara sect,272 brahmin students, alongside folk from all regions of India andambassadors of foreign (mleccha) countries.

In the early evening after dinner, Ban. a, being there at the king’s owninvitation, was allowed to pass through the Palace Gate (Rajadvara) andentered the first of four courtyards, congratulating himself: ‘I am fortunate,that the king thinks me worthy of such an honour.’273 He visits the royalstables with horses from Persia and other countries to the north-west,he admires the king’s magnificent, proud elephant Darpaśata, and, beingushered through three further courtyards, is stunned by the sight of stillmore kings and their royal households. Then in the fourth, inner court,seated in front of a dining pavilion/audience hall, he sees the emperor(cakravartin) Hars.a, ‘the turn of the dharma,’ so to speak.274

The king is seated in the relaxed, gentle pose of the sovereign on athrone of white stone with feet of ivory; behind him his bodyguard is linedup and in front of him a scene is laid of enticing dancing girls, ‘who weresucking out of him completely, as it were, the desire for his own wife, bytheir large jar-like breasts, which were covered by a series of garlands madeout of large bakula flowers.’275 Yet, notwithstanding all the temptationsto which a mighty sovereign is exposed, Ban. a describes King Hars.a as ‘acelibate, kingly sage who, after having been embraced by Royal Laks.mı,’i.e. after he had succeeded to the throne, ‘had been steadfast in observingthe Asidhara vow which he had taken.’276 The young poet stands in awe,shedding tears of joy, thinking to himself: ‘This here is the paramountlord, his majesty Hars.a (‘Joy’), of noble birth, living up to his name, aheap of lustre, the holder of the field between the four oceans, the eater ofthe fruit of the world-tree, the wrestler whose victory surpasses the deedsof all earlier kings.’277

272 Agrawala 1969, 136f.273 HC 2 p. 100 [28]: dhanyo ’smi yad evam anugrahyam. mam. devo manyate.274 HC 2 p. 125 [35]: avartanam iva dharmasya. Cf. the concluding verse of Hars.a’s

Copperplates (the Madhuban, the Banskhera and the Kurukshetra-VaranasiCopperplates), which, according to Buhler (EI I, 71), should be ascribed tothe king himself: karman. a manasa vaca kartavyam. pran. ine hitam | hars.en. aitatsamakhyatam. dharmarjanam anuttamam ‖

275 HC 2 p. 121 [34]: vikat.abakulavalıvarat.akaves.t.itamukhair br.hadbhih. stanakalaśaih.svadarasam. tos.arasam ivaśes.am uddharantıbhih. .

276 HC 2 p. 113 [32]: gr. hıtabrahmacaryam alingitam. rajalaks.mya pratipannasidhara-dharan. avratam avisam. vadinam. rajars. im.

277 HC 2 p. 125 [35]: so ’yam. sujanma sugr. hıtanama tejasam. raśih. caturudadhikedara-kut.umbı bhokta brahmastambhaphalasya sakaladirajacaritajayajyes.t.hamallo devah.

The Reign of Hars.avardhana 97

Ban. a is treated coolly, hindered by his reputation. Hars.a lets himknow that ‘as long as he is not propitiated he will not see him,’278

and to his favourite courtier sitting behind him, the Malava prince, i.e.Madhavagupta, he adds: ‘he is a great libertine.’279 The poet speaks upfor himself, saying that he is now a respectable householder and thathe will trust in the king’s good judgement, in the expectation that hisgreatness makes him indifferent to gossip. ‘So we have heard,’ Hars.aanswers and is silent. He casts an affectionate glance at the poor poet,280

dismisses his courtiers, and enters his private apartments. Ban. a leavesfull of hope and makes a resolve to behave virtuously. This propitiatesthe king and the poet becomes a welcome and much esteemed visitor tothe royal court.

Hars.a’s campaigns against Gaud. aAbout thirty years after Ban. a first visited King Hars.a’s court at his head-quarters on the river Ajiravatı, another celebrated contemporary was wel-comed again to an army camp. Since Ban. a’s introduction to court Hars.ahad spread his power eastwards, but it was only after the death of Śaśa-nka that the road to Bengal and Orissa lay open. The Chinese pilgrimXuanzang has left us an account of his meeting, which took place 215 kmeast of the Ajiravatı, on the lower reaches of the Ganges at a place calledKajurgira (Kajinghara), identified with Kajangala,281 c. 150 km north-west of what had been Śaśanka’s capital, Karn. asuvarn. a (see Plate 11).282

Xuanzang, who had evidently not met Hars.a when he had stayed in theBhadravihara in Kanauj in ad 636,283 saw the king for the first time atthe end of ad 642 or the beginning of ad 643, not long before his depar-ture from India, on the occasion of Hars.a’s return from his campaign intoOrissa.284

parameśvaro hars.ah. |278 HC 2 p. 128 [36]: na tavad enam akr. taprasadah. paśyami.279 HC 2 p. 129 [36]: mahan ayam. bhujam. gah. .280 HC 2 p. 130f. [36]: kevalam amr.tavr.s. t.ibhih. snapayann iva snehagarbhen. a dr.s. t.ipata-

matren. antargatam. prıtim akathayat |281 Life p. 172, Si-yu-ki I, 215f. Watters 1904–05 II, 183. For Kajangala near present-

day Rajmahal see Schwartzberg Atlas, 28; Majumdar 1971, 54.282 Identified with the modern Jadupur (former Chiruti). Majumdar 1971, 7: Karn. asu-

varn. a, ‘the ruins of which have been recently discovered at Rajbarid. anga (near themodern Railway Station Chiruti), six miles south-west of Berhampur, headquartersof the Murshidabad district.’ Cf. Si-yu-ki II, 201–04; EIA II, 359f.

283 Life p. 84.284 The date of this meeting is tied in with the date of Xuanzang’s departure from

China (see Sen 2003, 250 n. 7). We think, with Devahuti, that the later date (i.e.ad 629/30) agrees best with the chronology of Hars.a’s reign. Xuanzang met theking after the latter returned from his expedition to Kongoda (see below, p. 103).This would imply that Hars.a had sent his first embassy to the Tang court (ad 641)

98 Hans Bakker

Incidentally, this gives us some insight into the way Hars.a lived andhow he held his empire together for all these years. Rather than stayingput in his capital Kanauj and operating through a centralized bureaucracy,the king resided mostly either in local strongholds or in a strengthened,but mobile court—beautifully described by Ban. a—which travelled at aslow pace throughout his kingdom (below, p. 99), attracting, wherever itcame, admiration and awe, making their sovereign real and tangible to allhis subjects, landowners and feudatories. The centre and fountainhead ofpower was, so to speak, omnipresent.285

In contrast with Ban. a, Xuanzang, who was in the company of the kingof Kamarupa, Bhaskaravarman, was welcomed with due respect. KingHars.a ‘with some twenty attendants came forth to meet them. Enteringthey sat down, when choice viands were set before them, accompanied bymusic and strewn flowers.’286

According to the Life, the night before his reception at court took placeXuanzang had received the honour of a visit of King Hars.a in person atthe pilgrim’s pavilion on the left bank of the Ganges. The spectacle of the

before he met the Chinese pilgrim in 642/643 (Devahuti 1998, 250f.). The accountof the meeting of Hars.a and Xuanzang explicitly states that the king had invitedthe pilgrim earlier: ‘why did you not comply with my request?’ (Life p. 173). Theinvitation may have reached Xuanzang during his stay in Nalanda (Si-yu-ki I, 215),where his rising fame may have come first to the notice of the emperor. Further,it becomes evident from his account that Hars.a already had some informationabout China (Si-yu-ki I, 216f.; Life p. 174). Tansen Sen, however, argues that ‘theChinese monk played some role in convincing the Indian king to explore diplomaticties with Tang China’ (Sen 2003, 19).

285 Cf. Sharma 2001, 24: ‘The state which appears in the time of Hars.avardhana innorth India is different in nature from the state which is to be found in pre-Guptatimes. The role of the bureaucracy was considerably important in the pre-Guptastates which generally paid their employees in cash. [. . . ] In Gupta times, and moreso later, religious and administrative services were remunerated through grants ofland, and even in the core area, administrative powers devolved on landed benefi-ciaries or on functionaries who were granted certain taxes for their maintenance.This naturally took away much authority from the head of state. The weaknessof the central authority is also indicated by the existence of several capitals.’ Al-though I subscribe to some extent to the view expressed above by Sharma, I amhesitant to ascribe ‘weakness’ to the ‘central authority.’ The ‘weakness of the cen-tre’ was counterbalanced by the ‘omnipresence’ of the royal authority, achieved byan active, non-sedentary style of government. It is the unique personality of KingHars.avardhana, apparently, that succeeded in sustaining such a lifestyle for overforty years. It would go beyond the scope of this book to discuss here Sharma’sseminal ideas of ‘urban decline’ and ‘feudalisation,’ but his theory of ‘urban de-cline’ has recently been seriously questioned from an archaeologist’s point of viewby Derek Kennet, who concedes, though, that the century of transition from theEarly Historic to the Early Medieval, broadly speaking the sixth century, ‘was aperiod of change rather than one of stability’ (Kennet 2013, 348).

286 Life p. 175.

The Reign of Hars.avardhana 99

king crossing the river is reported:

On the river there were several thousand lighted torches, and theyheard the sound of beating drums. [. . . ] As Śıladitya marched hewas always accompanied by several hundred persons with goldendrums, who beat one stroke for every step taken; they called thesethe ‘music-pace-drums.’287

Xuanzang informed the king about his home country and showed him‘a Śastra’ he had composed himself,288 which so impressed Hars.a and hissister Rajyaśrı, who used to sit behind him in court, that he invited the pil-grim to come with them to Kanauj to glory in the religious assembly thatwould be convoked there.289 Thus a triumphant Hars.avardhana, accompa-nied by his feudatory Bhaskaravarman and the Chinese pilgrim, set out forKanyakubja in a procession of great pomp. They reached the capital after90 days.290 It could have been on this occasion that another of Hars.a’scourt poets, the can. d. ala Divakara, composed the following Mandakrantaverse:

Why tell stories that pertain to another man’s house? Yet, I can-not keep it to myself any longer, being, as a southerner, talkativeby nature: everywhere, in shops, in the market-place and at drink-ing parties, your lover roams about, like a wanton woman; lo, it isFame!291

Thanks to the two eyewitness reports discussed above, we know that Hars.ahad pitched his camps at the Burhi Gandak (the river Ajiravatı), after hehad gained sovereignty over the kingdoms of Thanesar and Kanauj inabout ad 612–13, and in Kajangala at the Ganga, 215 km further to the

287 Life p. 173. This story of King Hars.a crossing the Ganges to welcome the Chinesepilgrim in person seems historically very unlikely. The same pilgrim was apparentlynever received by the king during his stay in Kanauj in ad 636! See above, p. 97.The story may have entered Chinese sources in order to please the Chinese emperorTaizong (ad 626–49) ‘the principal audience of his work’ (Sen 2003, 17).

288 This may be the Destruction of Heresy of 1600 ślokas which Xuanzang composedin Nalanda (Life p. 165).

289 Life p. 176.290 Si-yu-ki I, 217f.; Life p. 176. If true, this implies that the average speed of the

mobile court was 9 km per day.291 Divakara quoted in Vidyakara’s Subhas. itaratnakos.a (No. 996) and Vallabhadeva’s

Subhas. itavalı (No. 2544):

kim. vr. ttantaih. paragr.hagataih. kim. tu naham. samarthas,tus.n. ım. sthatum. prakr. timukharo daks. in. atyasvabhavah. |deśe deśe vipan. is.u tatha catvare panagos.t.hyam,unmatteva bhramati bhavato vallabha hanta kırtih. ‖

100 Hans Bakker

east thirty years later, after his campaign into Orissa in ad 642. Betweenthese two headquarters lay the countries of Magadha and (part of) Gaud. a,the kingdom of Śaśanka. There can be little doubt that Gaud. a remainedlargely under the control of Śaśanka until his death (between ad 620 andad 636),292 but whether he also ruled Magadha in the period from ad612 till his death, is uncertain. Did King Hars.a conquer south Bihar fromŚaśanka, or was it only after the latter had died that Hars.a could establishhis authority beyond the Son and Gandak Rivers?

Plate 11The army camps of Hars.a at Man. itara and at Kajurgira

With respect to this question, the same set of historical sources has beenscrutinized by Majumdar and Devahuti, and they have come to oppositeconclusions. Devahuti’s hero is King Hars.a, Majumdar’s King Śaśanka.293

The argument hinges on the most problematic source of all, the Arya-manjuśrımulakalpa.

And the younger brother of that (king) [viz. Rajyavardhana], calledHa[rs.avardhana], a great and unique man, will rule—a hero, pos-sessed of a great army, whose power reaches far. Now, (this) King

292 Xuanzang visited Magadha in ad 637–38, staying there nearly two years altogether.By that time Śaśanka had died, but not long before, it would seem (Si-yu-ki II,118, 121–22; Life 171). On account of Xuanzang’s report, Majumdar (1971, 56)dates the death of Śaśanka to ad 636. Devahuti (1998, 102) places his death inc. 620, one year after our last epigraphical testimony to his rule (EI VI, 143–46).

293 Majumdar 1971, 52–73; Devahuti 1998, 99–102.

The Reign of Hars.avardhana 101

Ha[rs.avardhana], who belongs to the Vaiśya class, being determined(nirdharaye) against a king known as ‘Moon’ (Soma), on a certainday (tada) marched eastwards on a major city Pun. d. ra, with a greatarmy and great force. Turned to his military duty, pride and anger,[though] being himself tender-hearted, righteous and wise, he sentto perdition many people of wicked conduct. Sympathizing with thesuffering of the good and resolved on stopping (that arrogant con-duct), he defeated Soma, who led a life of wicked deeds. Then, after(that king) named Soma had been curbed, he has stayed within hisown country; Ha[rs.avardhana] returned, [since] he was not welcomein that kingdom of Barbarians. (The king) called Ha[r.s.avardhana],being prosperous and lawful, has accomplished his aim within hisown country, by fortune or reaching the destination that he had as-pired after. Only by those who are associated with the joyful andkingly (hars. in and rajya) an achievement is brought about.294

The author clearly shares the anti-Śaśanka bias of Xuanzang, but althoughhe credits Hars.a with a victorious campaign against Śaśanka, he admitsthat the latter remained an independent king within his own country.Hars.a’s campaign against him, therefore, seems to have had only limitedeffect and probably never reached Pun. d. ra(vardhana). Xuanzang reportsthat, while Śaśanka was still alive and committed his acts of sacrilege inBodh Gaya, the king of Magadha was Purn. avarman, whom the pilgrimcalls ‘the last of the race of Aśoka-raja’ (Si-yu-ki II, 118), but who morelikely was a scion of the Maukhari family.295

294 MMK 53.661–69 (Ganapati Shastri’s edition 1925, e-texthttp://gretil.sub.uni-goettingen.de/gretil/1_sanskr/4_rellit/buddh/bsu041_u.htm):tasyapy anujo hakarakhya ekavıro bhavis.yati |mahasainyasamayuktah. śurah. krantavikramah. ‖nirdharaye hakarakhyo nr.patim. somaviśrutam |vaiśyavr. ttis tato raja mahasainyo mahabalah. |purvadeśam. tada jagmuh. pun. d. rakhyam. puram uttamam ‖ks.atradharmam. samaśr. tya manaros.am aśılinah. |ghr.n. ı dharmarthako vidvan kuryat pran. ivadham. bahun ‖sattvanupıd. anaparo nigrahayaiva so ratah. |parajayamasa somakhyam. dus.t.akarmanucarin. am ‖tato nis. iddhah. somakhyo svadeśenavatis. t.hata* |nivartayamasa hakarakhyah. mleccharajye-m-apujitah. ‖tus.t.akarma hakarakhyah. ** śreyasa carthadharmin. ah. |svadeśenaiva prayatah. yathes.t.agatinapi va |tair eva karitam. karma rajyahars. isamanvitaih. *** ‖

Editions read: * svadeśenavatis. t.hatah. (I read an augmentless imperfect (m.c.) in-stead); ** I omit nr.pah. ; *** rajyahars. ısamanvitaih. . Cf. Jayaswal & Sankr.ityayana1934, vss. 721–27, and Jayaswal’s translation p. 50.

295 I find the suggestion by Cunningham (ASI Report vol. XV (1882), 166) to the

102 Hans Bakker

After the Hars.a–Śaśanka war, he may have pledged allegiance to Hars.a-vardhana, rather than to Śaśanka. King Purn. avarman, who was no longeralive when Xuanzang was in Magadha,296 is reported to have repaired theBodhi Tree in Gaya and to have built a pavilion for a copper image ofthe Buddha of 80 feet (24 m) high, which was seen by Xuanzang whenhe stayed at Nalanda in ad 637.297 That Hars.a had gained control overMagadha before the Chinese pilgrim came to live there, may be derivedfrom the pilgrim mentioning a ‘vihara of brass’ in Nalanda, which wascommissioned by Hars.a but was not yet finished.298 Hars.a’s presence inNalanda is further endorsed by an undated clay seal and fragments ofsimilar seals.299

To sum up, the most likely scenario that can be derived from oursources is that Śaśanka committed his sacrilege in Magadha in the first orsecond decade of the seventh century,300 but was driven out of that countryby the army of Hars.a in the second decade, after which Hars.a reinforcedthe rule of Purn. avarman; from that time onwards he had suzerainty overMagadha. What remained of Maukhari power after Hars.a’s takeover inKanauj thus became concentrated in Magadha; Purn. avarman may havebeen descended from Anantavarman, the local Maukhari ruler of the Gayaregion in the first half of the sixth century (see above, p. 43), or from themysterious Suva. . . (above, p. 85). Hars.a’s Buddhist leanings and schol-arly interests made him naturally a patron of the Nalanda University. Theposition of the Later Guptas of Magadha during this period remains un-known, apart from the alleged fact that lg Madhavagupta was allied, ifnot related, to Hars.a’s court (above, p. 92).

The establishment of an empire

The East: Bengal and Orissa

After the death of Śaśanka and of Purn. avarman († before ad 636), Hars.abegan to extend his power further eastwards. His campaigns in Gaud. amay have begun in the mid or late 630s. Once Gaud. a had come under hiscontrol, Hars.a’s invasion of Orissa in ad 642 seems to have been directed

effect that Xuanzang may have confused Maurya and Maukhari/Maukharıya ratherplausible.

296 Life p. 154.297 Si-yu-ki II, 118, 174.298 Si-yu-ki II, 174; Life p. 158f.299 Thaplyal 1985, 186f.300 In the undated Panchrol (Egra) Copperplate lnscription of the Time of Śaśanka,

the Maharajadhiraja Śaśanka is called a paramamaheśvara (Furui 2011, 121).

The Reign of Hars.avardhana 103

against the old allies of the king of Karn. asuvarn. a,301 the Śailodbhavasof Konkoda. When Xuanzang met the king in the winter of ad 642/643,Hars.a had just come back from Kong-u-t’o (Kongyodha, Kongoda),302

which is to be located in the region around the Chilka Lake south ofPuri.303 Obviously, Hars.a had marched from Karn. asuvarn. a towards thesouth, along the coast, as far as Lake Chilka. There is no evidence that theemperor of Kanauj, at any time, held any control over Kalinga, furthertoward the south, or over the interior of Orissa, and his power in thenorthern coastal region (Utkala, Od. ra and Konkoda) may have lasted notmuch longer than the actual campaign. Hars.a’s grant, reported in the Life,of ‘the revenue of eight large towns of Orissa’ to the learned householderJayasena, who wisely declined—just as he had earlier declined a similaroffer by Purn. avarman—may have been a boast, if not entirely spurious.304

Thus his military ambitions brought Hars.a in the final decade of hiscareer as far as the coastal region of Orissa. This left untouched thePan.d. uvam. śin kingdom in Daks.in. a Kosala (Chhattisgarh) of Mahaśiva-gupta Balarjuna, whose rule in Śrıpura (Sirpur) from c. ad 590 to 650coincided with the duration of Hars.a’s life. The Houses of Kanauj andSirpur were related by matrimony, and it seems not unlikely that this al-liance guaranteed that both kings left each other in peace, though theirhearts may not have been in it.

The fictitious battle between the ruler of Vatsa, Udayana, and theLord of Kosala as told in Śrıhars.a’s play the Ratnavalı does not point to awarm relationship between the two kingdoms and is not very flattering tothe ruler of Kosala. In Act IV we are told that the legendary King Uda-yana had sent his minister Ruman.vat against the Lord of Kosala, who hadentrenched himself in ‘a fort in the Vindhyas.’305 The anonymous king of

301 The Ganjam Plates of Madhavaraja II of the Śailodbhava dynasty (EI VI, 143–46) are dated to year 300 of the Gupta Era, i.e. ad 619–20. In this inscriptionMadhavaraja acknowledges the suzerainty of Maharajadhiraja Śaśanka, which in-dicates that Śaśanka had safeguarded his sovereignty over greater parts of Bengaland Orissa at least until ad 620 (see Devahuti 1998, 49–53).

302 Life p. 172. For Xuanzang’s description of Kongoda in Orissa see the Si-yu-ki (BealII, 206f.).

303 The Ganjam Plates of Madhavaraja II had been issued in Vijaya-Kongoda on thebanks of the Śalima river. ‘The river is now identified with the present Salia whichruns past the township of Banpur in Puri district and flows into Chilka Lake.Banpur is situated only four kilometers north-west of this lake’ (Tripathy 1998,83). Majumdar (1971, 50) locates the kingdom of Kongoda in the ‘region roundthe Chilka Lake in Orissa, extending probably to the Ganjam district in the south.’

304 Life 153f. From this text it is impossible to deduce when exactly this offering wasthought to have been made.

305 Ratnavalı (ed. Cappeller/Bohtlingk): ito devadeśat katipayair eva vasarair aneka-karituragapadatidurnivaren. a mahata balasamuhena gatva ruman. van vindhyadurga-vasthitasya kosalapater dvaram avas.t.abhya samavasayitum arabdhavan. See also

104 Hans Bakker

Kosala is induced to leave the fortification and to give battle, in which heis slain:

After (his) main force had broken down, the king of Kosala waschallenged to the frontline—there where heads, whose helmets hadbeen thrown down by missiles, were constantly chopped off by strik-ing swords, where rivers of blood drew lines of division, where sparksissued from the armour by sounding blows—and there that king,seated on a mad elephant, was killed by Ruman.vat on his own withhundreds of arrows.306

Whatever the historic value of this play, nearly unanimously ascribed toKing Hars.avardhana, there might have been pressing political reasons forthe ‘marriage of convenience’ of both Houses. The rising power of the Ca-lukyas in Badami in the Deccan ensured that they had a dangerous enemyin common.

The South: Hars.a’s struggle with the Calukyas

In the period between Hars.avardhana’s first campaign in Gaud. a, in c.ad 612, and his last one in 642, the king’s major challenge lay in thesouth-west. At stake was access to the Western Ocean or Arabian Sea,through the important trade route that led from Kanyakubja via Mathu-ra and Ujjayanı, to the Gulf of Cambay and its ports, such as Bhr.gukaccha(Bharukaccha/Broach). After the successful campaign of his elder brother,Rajyavardhana, against the ‘wicked king of Malava,’ Devagupta, Hars.a-vardhana may have kept some sort of control over Eastern Malava and itscapital Vidiśa from the very beginning of his reign. However, similar tothe initial situation of his great predecessor, Candragupta II, possessionof Vidiśa did not automatically bring control over Western Malava and itscapital Ujjain; this required political acumen and military pressure.

The Kalacuri king Śam. karagan. a of Mahis.matı had been succeeded byhis son Buddharaja in c. ad 601. The latter, along with his vassals, theGuptas of Vidiśa, may have lost his hold on Eastern Malava due to thecampaign of Hars.a’s brother, but he probably kept control over westernMalava and its capital Ujjain. The greatest danger for Buddharaja initially

below, p. 200.306 Ratnavalı 69 (ed. Cappeller/Bohtlingk):

astravyastaśirastraśastrakas.an. otkr. ttottamange* muhur,vyud. hasr.ksariti svanatpraharan. air varmodvamadvahnini** |ahuyajimukhe sa kosalapatir bhagne pradhane bale,ekenaiva ruman. vata śaraśatair mattadvipastho hatah. ‖ 69 ‖

* I follow Kale’s reading ◦kas.an. otkr. tto◦ instead of ◦kas.an. air kr. tto◦ and ** varmo◦

instead of gharmo◦ (Cappeller/Bohtlingk).

The Reign of Hars.avardhana 105

came not from the Pus.yabhutis, but from the rising power in the south.In ad 602 Buddharaja suffered a defeat at the hands of the Calukya kingMangaleśa.307 Despite this defeat, Buddharaja managed to some extentto hold the kingdom of his father intact by campaigning in Malava andGujarat. In ad 610, when Hars.a was locked in his life-and-death strugglewith Śaśanka, Buddharaja issued two charters from his army camps atVidiśa and Anandapura.308 However, soon after Hars.a had established hissovereignty in Northern India, fortune deserted the Kalacuris of Mahis.ma-tı.

The Maitrakas of Valabhı

Dharasena II

Kharagraha I Śıladitya I ‘Dharmaditya’ [Hars.avardhana]

Dharasena III Dhruvasena II Baladitya x daughter

Dharasena IV

Figure 9

The Maitrakas of Valabhı and Hars.avardhana of Kanauj

The Virdi Grant of the Maitraka ruler of Valabhı (Saurashtra), Khara-graha I, issued from his ‘victorious camp’ in Ujjayanı in ad 616, provesthat this city had come under the control of the former feudatories ofthe Imperial Guptas,309 the Maitrakas of Valabhı.310 Kharagraha I, who,although his military prowess is sung in his praśasti, is actually with-out royal titles,311 had succeeded his elder brother, Maharaja Śıladitya I

307 Fleet 1882, 21f.; Mirashi in CII IV.1, xlviii. For the Kalacuris also see below, p. 204.308 CII IV.1 nos. 14 & 15. Anandapura is ‘either Anand in the Kaira District or

Vad. nagar in the Mehsana District’ (both in Gujarat) (Mirashi in CII IV, 53).309 CII III (1888), 59; Devahuti 1998, 53.310 Gadre 1935, 667. Cf. Devahuti 1998, 58; HCI III, 30f., 60–63.311 Gadre 1935, 669 ll. 22–23; repeated in later inscriptions, e.g. the Alina Copper-

plate Inscription of Siladitya VII in CII (1888), 175 ll. 18–19: gan. atithavipaks.aks. iti-patilaks.mısvayam. grahaprakaśitapravırapurus.aprathamasam. khyadhigamah. parama-

106 Hans Bakker

‘Dharmaditya,’ whom Xuanzang glorifies as the pious king of Mo-la-p’o,i.e. Malava (Figure 9).312 Evidently, between ad 610 and 616, Malava,west and probably also east, had passed into the hands of the Maitra-kas, who may have been the main target of Buddharaja’s campaign in ad610. Although they thus successfully turned the tables on the Kalacuris,the position of these new rulers of Ujjain was delicate, since the Maitra-kas were, like the Kalacuris, sandwiched between the expanding empiresof Hars.a in the north and the successor of Mangaleśa, the Calukya kingPulakeśin II in the south.

Kharagraha I was succeeded by his son Dharasena III, of whom wepossess two inscriptions, dated to ad 623 and 624, in which he, like hisfather, is without royal titles.313 He in turn was succeeded by his youngerbrother, Dhruvasena II Baladitya, who is equally without royal titles in hisinscriptions, the first and last of which date from ad 629 and 642 respec-tively.314 The two land grants made by Dhruvasena II in ad 639 and 640,recorded in the inscriptions found in Nogawa (= Navagaon),315 seem toprove that this province fell under Maitraka jurisdiction, since the donatedvillages are said to lie within the province (bhukti) of Malava.316 Evidently

maheśvarah. śrıkharagrahah. . ‘The Paramamaheśvara, the illustrious Kharagraha,whose attainment of being counted first among heroes is evident from the factthat he was chosen by Laks.mı herself from the assembled hostile kings.’

312 Beal II, 261f. The relationship between the brothers seems not to have been un-problematic, in view of the long sentence that first appeared in Kharagraha I’sUjjain inscription and was repeated in the inscriptions of the later members of theMaitraka dynasty (see Gadre 1935, 669 ll. 17–23; CII III (1888), 174f. ll. 14–19).Maybe Śıladitya I ‘Dharmaditya’ had become disabled (suggested in particular bythe gerundive samabhilas.an. ıyam): tasyanujas tatpadanudhyatah. svayam upendra-gurun. eva gurun. atyadaravata samabhilas.an. ıyam api rajalaks.mım. skandhasaktam.paramabhadra iva dhuryas tadajnasampadanaikarasatayaivodvahan khedasukha-ratibhyam anayasitasattvasampattih. . . . . ‘His (i.e. Śıladitya I “Dharmaditya”’s)younger brother (the “illustrious Kharagraha” I) possessed a wealth of virtuethat was neither exhausted by afflictions nor by enjoyment of happiness, whenhe, favoured by the feet of him [that is to say, with the blessing of Śıladitya I“Dharmaditya”], like a choice beast of burden, was carrying royal Fortune restingon his shoulders, even though she was (still) desirable to (his) elder (brother)—whowas extremely respectful (nevertheless), as if he were the elder (brother) of Upe-ndra himself (i.e. Indra)—(a task accomplished) simply and with the sole intentionof carrying out that (elder brother’s) commands . . . .’

313 Bhavnagar Plates of Dharasena III—[Valabhi-] Sam. vat 304 in: EI XXI (1931–32),181–184 (ad 623); a second inscription published in the Journal of the Universityof Bombay XIX.4 (ad 624, unavailable to me) is mentioned in Shastri 2000, 54ff.,427.

314 IA VI (1877), 12–16 (ad 629); inscription published in the Journal of the OrientalInstitute (Baroda) X, 123 (ad 642) mentioned in Shastri 2000, 56ff.

315 16 km north of Ratlam and 80 km north-west of Ujjain, i.e. in Western Malwa.316 EI VIII, 188–99. About the expression malavake ucyamanabhuktau and malavake

ucyamanavis.aye (ll. 41f. in Nogawa Plates A and l. 44 in Nogawa Plates B),

The Reign of Hars.avardhana 107

the Maitrakas were feudatories during the third and fourth decades of theseventh century, but the question is of whom, Hars.a in the north or Pula-keśin in the south?

The situation may have been complex. Welcome light on it is shedby the Chinese pilgrim Xuanzang, who visited the capital Valabhı (Fa-la-pi) under the Maitraka king Dhruvasena II (T’u-lu-h’o-po-tu) near theend of his reign (c. ad 641–42).317 The pilgrim describes this king as ‘ofa quick and impulsive nature, and his manners are heavy and dull, butyet he esteems virtue and advances learning.’318 In his Si-yu-ki (Beal II,267) the pilgrim says that this king is ‘the nephew of Śıladitya-raja ofMalava,’ who can be no one else than the Maitraka king Śıladitya, brotherof Kharagraha, who was indeed his uncle and ruled over (western) Malava(see above). More importantly, the Chinese pilgrim adds that the king wasalso the ‘son-in-law of Śıladitya, the present king of Kanyakubja.’ (Figure9).319

Xuanzang is our only source for this. Neither Hars.a nor Dhruvasena IIand his son Dharasena IV ever allude explicitly to this relationship in theirinscriptions. Nevertheless, in view of the fact that Xuanzang is so specificand consistent on this point, which he could hardly have invented himself,especially not because he stayed at the courts of both kings, I am inclinedto attach credit to his testimony; it may explain why Dharasena IV paysrespect to his maternal grandfather in his earliest inscription (see below,n. 340 on p. 112). The reason why Dhruvasena and his son are silent ontheir relationship to Hars.a’s House could be that they anxiously avoidedreferring to Hars.a’s suzerainty; for Hars.a, who hardly refers to historicalfacts in his inscriptions at all, apart from those concerning his accessionto the throne, it may have been too insignificant.

We may consider it as certain, that King Hars.a would not have givenhis daughter to Dhruvasena, if the latter had been a feudatory of the Calu-kyas at the time. If Xuanzang is correct, we have to conclude that, beforethe Maitrakas came under the sway of Pulakeśin, Hars.avardhana had mar-ried off his daughter to Dhruvasena, possibly at a time when his brother

Hultzsch remarks: ‘Consequently the word ucyamane before bhuktau and vis.ayecan only refer to the preceding locative malavake’ (EI VIII (1905–06), 189).

317 Beal II, 267; Life, 149318 Life, 149; cf. Si-yu-ki (Beal II, 267): ‘He is of a lively and hasty disposition, his

wisdom and statecraft are shallow. Quite recently he has attached himself sincerelyto faith in the three “precious ones”.’

319 Beal II, 267 actually reads ‘son-in-law of the son of Śıladitya . . . ’, but the Lifereads just ‘son-in-law.’ Watters 1904–05 II, 246 discusses the matter saying: ‘Herethe words rendered “son-in-law” are tzu-hsu and Julien took these to mean “sonin law, of son”, but the phrase means simply “son-in-law”, and the Life employsthe common term nu-hsu.’ See also Life, 149: ‘he is son-in-law (nu sai) of Śıladityaraja of the kingdom of Kanyakubja; his name is Dhruvabhat.a.’

108 Hans Bakker

Dharasena III was still reigning, that is in the middle of the third decadeof the seventh century, when Hars.a himself was in his late thirties.320 Ob-viously Hars.a had established influence over the Maitraka kingdom, andthis achievement was crowned, it would seem, by a matrimonial alliance.In this way the empire of Kanauj hoped to have secured its access to theWestern Ocean.

However, this situation would not last. The Calukya king PulakeśinII (ad 610–42)—who had emerged victorious from a bloody civil war andwho, after securing his position at home, had initially waged war on hissouthern neighbours, the Kadambas, Gangas and Al.upas321—could nottolerate that a northern imperial power had taken control over what thesovereign of the Deccan must have considered to be his natural birthright,namely, unlimited access to the ocean ports of the Gulf of Cambay, such asBharukaccha.322 After having ‘swept away’ the Mauryas in the Konkan,323

steps were taken to subject to his rule the three regional powers thatcontrolled the Gulf of Cambay, viz. 1) Lat.a, the country to the south ofBharukaccha, 2) the principality in which the port was actually situatedand that was ruled by a branch of the Gurjaras, and 3) the Maitrakas,referred to as the rulers of Malava.

Whether this implied war, or whether diplomatic or military pressurealone sufficed, we do not know,324 but in his Aihole Inscription Pula-keśin claims the successful accomplishment of this policy, which thereforemust have happened before ad 634 and probably after Dhruvasena II hadsupplanted his brother.

Subjected to whose [i.e. Pulakeśin’s] power the Lat.as, Malavas andGurjaras became teachers, as it were, of how vassals, subjected bythe rod, ought to behave.325

It may be speculated that the transfer of power from Dharasena III toDhruvasena II was the outcome of this conflict with the Calukyas. If thiswere the case, it could possibly explain the later tradition that a king ofValabhı had fled to the court of the Gurjara ruler of Broach, Dadda II (ad

320 Cf. Shastri 2000, 58.321 EI VI, 5f. (Aihole Inscription), vss. 18–19; HCI III, 235. Ramesh 1984, 78f.322 This north-south conflict over access to the Gulf of Cambay was not unique. It

resembles, for instance, the first-century conflict between the Mahaks.atrapa kingNahapana and his Satavahana contemporaries (see Brancaccio 2011, 61f.).

323 EI VI, 5f. (Aihole Inscription), vss. 20–21.324 In Skandapuran. a 61.106–09 four methods (upaya) of foreign policy are discussed:

conciliation (saman), gifts (dana), dissension (bheda) and war (yuddha). Cf. Artha-śastra 7.16.4–8. See SP III, 110.

325 SI II, 447, vs. 22:pratapopanata yasya lat.amalavagurjarah. |dan. d. opanatasamantacaryac[a]rya ivabhavan ‖

The Reign of Hars.avardhana 109

629–642).326 Hars.avardhana may have intervened in favour of DhruvasenaII to secure the loyalty of the Maitrakas in his conflict with Pulakeśin. Ithas to be admitted, though, that Dhruvasena II mentions his elder brotherin all his inscriptions with due respect.

Pulakeśin’s advance against the kingdoms around the Gulf of Cambaymay have been a casus belli for Hars.a. It would seem that the AiholeInscription itself refers to it in the next Malinı verse:

And due to whom (i.e. Pulakeśin), Hars.a’s joys were thwarted byfear, as he was annoyed by the array of majestic elephants that hadfallen in battle—Hars.a, whose lotus feet had been flooded with thelustre of the crest jewels of hosts of feudatories, who were rolling inimmeasurable wealth.327

Hars.a and Pulakeśin did clash, we have the testimony of Xuanzang and theAihole Inscription for this, but whether this was the epoch-making battle

326 Mirashi in CII IV, liii. The great-grandson of this Dadda II, Jayabhat.t.a III(ad 690–715) is the first in this family who ascribes to his great-grandfatherthe achievement of having provided protection to the king of Valabhı when thelatter had been attacked by Hars.adeva: parameśvaraśrıhars.adevabhibhutavalabhı-patiparitran. opajatabhramadadabhraśubhrabhravibhramayaśovitanah. , ‘[Jayabhat.t.a,]who possessed a canopy of glory—the grace of which was that of a moving, notsmall, white cloud—that had originated from his rescuing the Lord of Valabhıwhen the latter had been defeated by the supreme lord, the illustrious Hars.adeva.’Mirashi (loc. cit.) observes that ‘It has been recognized that Dadda II, the ruler ofa petty state, a mere samanta, could not have, unaided, given protection to theruler of Valabhı against the armies of the most powerful ruler of the time with theresources of a vast empire at his command. Besides, it is noteworthy that Daddahimself is silent about this glorious achievement ascribed to him, for there is nomention of it even in his grants of K. 392 (641–42 A.C.), which were made towardsthe end of his reign and before which the event must certainly have happened.’Why, we may ask, would Hars.a attack his son-in-law, drive him into exile, andthen invite him again as his guest of honour at his great assembly in Prayaga? Itis obvious that this story, not corroborated by any other historical source, is spuri-ous. Nevertheless, many historians have taken it at face value and have construedout of it a military attack by Hars.a on Valabhı. Devahuti, for instance, (1998, 62)speaks of ‘Hars.a’s march on Valabhı.’ How is this to be reconciled with anotherobservation on the same page to the effect that ‘Throughout his reign Harsha pos-sessed the friendship and loyalty of the Maitrakas and was assured of the safety ofthe western extremity of his empire’? A solution to this conundrum is available,however, if we assume that it was not Dhruvasena II who went into exile, buthis elder brother Dharasena III, after Hars.a’s intervention in favour of Dhruva-sena. This would indicate that the house of the Maitrakas had been divided overwhich side to choose, Pulakeśin or Hars.avardhana. With the exile of Dharasenaand the confirmation of Dhruvasena’s loyalty to Kanauj the conflict may have beenresolved.

327 SI II, 447, vs. 23:aparimitavibhutisphıtasamantasenamakut.aman. imayukhakrantapadaravindah. |yudhi patitaga[je]ndranıkabıbhatsabhuto bhayavigalitahars.o yena cakari hars.ah. ‖

110 Hans Bakker

that later sources and historians want us to believe is very uncertain.328

The three inscriptions of Hars.a that we have are silent about it, but, itshould be observed, they all date from before ad 632.329

Xuanzang tells us that after his victory over the Gaud. as (ad 612),Hars.a ‘increased his forces; he had 60,000 war elephants and 100,000 cav-alry. After thirty years his arms reposed, and he governed everywhere inpeace.’330 Watters (1904–05 I, 346) has contested Beal’s rendering of theChinese: ‘Our pilgrim has expressly stated that the king’s conquests werecompleted within six years, and it is against text and context to makehim represent the king as fighting continuously for thirty or thirty-sixyears.’331 The ‘thirty years’ clearly refers to the period between Hars.a’sfirst war with Gaud. a (Śaśanka), which was concluded in about ad 612,and ad 642, when another expedition to Gaud. a and Konkoda had cometo an end and the Chinese pilgrim met the king in his army camp atKajurgira. The conflict with the Calukyas shows that Watters’ renderingof ‘a thirty-year period of peace’ is historically untenable, as is obviousalso from the following passage of the same Si-yu-ki.

At the present time [c. ad 640–42(?), when Xuanzang visited theDeccan] Śıladitya Maharaja (i.e. Hars.a) has conquered the nationsfrom east to west, and carried his arms to remote districts, butthe people of this country alone have not submitted to him.332 Hehas gathered troops from the five Indies, and summoned the bestleaders from all countries, and himself gone at the head of his armyto punish and subdue these people, but he has not yet conquered

328 See most recently Kulke 2012, 58, who is following a historiographic tradition: ‘Itwas the year 630 that proved to be a turning point in the history of India. In theattempt to emulate Samudragupta and to expand his domains up to central India,Hars.a’s army was defeated by Pulakeśin II, the Chalukya king of Badami (Vatapi)in western central India’ [my italics]. However, there was no ‘turning point,’ rathera status quo ante. Before and after Hars.a, as well as during his reign, northernimperial kingdoms were bound by the Narmada river in the south. This is alsotrue for Samudragupta, whose southern exploits, like those of Hars.a, were mainlylimited to an expedition along the east coast through Orissa into Andhra, butwhich did not result in control over territory south of the Narmada. The Guptasacquired some influence in the Deccan through a matrimonial alliance with theVakat.akas and such an alliance was evidently denied to Hars.a, but can we see thisas a turning point in Indian history?

329 Cf. HCI III, 105f.330 Beal I, 213. This interpretation of the passage is confirmed by Deeg’s rendering

(forthcoming): ‘After almost thirty years no weapons were raised any more, gov-ernment and education were peaceful [. . . ].’ Cf. HCI III, 108f.

331 Watters (1904–05 I, 343) translates: ‘[Hars.a] reigned in peace for thirty years with-out raising a weapon.’

332 The country called Mo-ho-la-ch’a, Maharas.t.ra, i.e. the kingdom of Pulakeśin.

The Reign of Hars.avardhana 111

their troops.333

Although Watters thus appears to be wrong about the 30 years withoutwar, his general impression from the pilgrim’s report to the effect thatHars.a’s rule on the whole was not characterized by incessant, large-scalewarfare may be right.

The Hars.a–Pulakeśin encounter may have taken place between ad 631and 634, and its outcome seems to be related in the next Malinı verse ofthe Aihole Inscription:

While he (i.e. Pulakeśin) was ruling the earth with his extensivearmies, the flanks of the Vindhya Range, whose beauty is enclosedby the various sandbanks of the Reva River, shone forth all the moredue to the greatness of his power, (now) being free of (war)elephantswhose height had eclipsed their peaks, as it were.334

I can only read this to mean that Hars.a’s elephant army no longer threat-ened to cross the Reva, i.e. Narmada River, after it had once been forcedto retreat with heavy losses, and that consequently this river had becomethe effective border between the two powers. What became of Bharu-kaccha/Broach on the north bank of the Narmada’s estuary is unknown,but it probably remained within the Calukya sphere of influence underthe feudatory Gurjara king Dadda II (above, p. 108). When Xuanzang,coming from Maharas.t.ra, visited it in c. ad 640, it seems to have been anunpleasant place.335

Calukya power thus came to a halt at the Narmada River in the north,but the Aihole Inscription presents the conquests of Pulakeśin as a digvi-jaya and goes on to tell us that, after the victories in the west, north-westand north, and the establishment of his power in Maharas.t.ra,

The (rulers of) Kalinga and Kosala, who excelled in the three aimsof life thanks to the intrinsic virtues of their householders and whohad contrived to explode the arrogance of other kings, began (now)to show signs of fear, due to his (i.e. Pulakeśin’s) might.336

333 Beal II, 256f.334 SI II, 447, vs. 24:

bhuvam urubhir anıkaiś śasato yasya rev(a)-vividhapulin. aśobhabandhyavindhyopakan. t.ha[h. ] |adhikataram arajat svena tejomahimna,śikharibhir ibhavarjy(o) var[s. ]man. a spardhayeva ‖ 24 ‖

335 Si-yu-ki II, 259f. See below, p. 116.336 Aihole Inscription in SI II, 447 vs. 26 (Aupacchandasika):

gr.hin. am. svagun. ais trivargatunga, vihitanyaks. itipalamanabhanga[h. ] |abhavann upajatabhıtilinga, yadanıkena sako[sa]lah. kalingah. ‖ 26 ‖

112 Hans Bakker

Rather than describing a conquest of Kosala and Kalinga, as this verse hasoften been taken to mean,337 it should be interpreted literally: the rulersof these two kingdoms anxiously watched the growing Calukya power, butwere not subdued by it. Kosala’s alliance with Kanauj may have heldPulakeśin back from a war with Śrıpura.338

A similar deterrence may have worked in the case of the Maitrakasof Valabhı. If they had come within the immediate sphere of influence ofPulakeśin at all, as the Aihole Inscription suggests, they were certainlyback within Hars.a’s fold after the battle at the Narmada River. Xuanzangreports that their king Dhruvabhat.t.a, i.e. Dhruvasena II, Hars.a’s assumedson-in-law, was a guest of honour at Hars.a’s last quinquennial assemblyin Prayaga in ad 643, at a time when the two kings were nearing theend of their lives.339 The omission of royal titles in the charters of theMaitraka kings Kharagraha I, Dharasena III, and Dhruvasena II provesbeyond doubt that these kings should be considered feudatories, but inthe same inscriptions they themselves carefully avoid acknowledging theiroverlord.340 In view of their delicate situation, they may have bowed tothe south and bowed to the north, according to the political wind, inan attempt to safeguard as much of their independence as possible, buttheir allegiance and long-term political interests certainly lay in the north.Tensions raised by this awkward predicament may, as observed above, have

337 e.g. Devahuti 1998, 113.338 The Aihole Inscription continues with Pulakeśin’s advance into Andhra and his

occupation of Pis.t.apura, where he installed his brother Vis.n. uvardhana (Calukya)on the throne (Ramesh 1984, 81f.).

339 Life p. 185f. Dhruvasena II also accompanied Hars.a when he took his final leaveof the Chinese pilgrim (op. cit. p. 189).

340 Buhler in IA VI (1877), 13 acutely observes that ‘it may be that the omission isdue to an accident; but, considering the habitual grandiloquence of Indian princes,the case is spurious, and it would not be surprising if it were found eventuallythat Dhruvasena II had some cogent reasons for being silent about his magnifi-cence.’ Since the feature observed by Buhler has been confirmed by charters pub-lished afterwards, the possibility of an accident may safely be excluded. Moreover,Dhruvasena II’s son, Dharasena IV (c. ad 645–50), the assumed grandson of Hars.a,immediately adopted grand titles, such as Paramabhat.t.araka, Maharajadhiraja,Parameśvara, and Cakravartin in his earliest inscriptions of ad 645/6. The En-glish translations of these two grants, dated to Samvat 326, are published in IA I(1872), 14–18, 45–46. R.G. Bhandarkar translates: ‘who meditates on his grandfa-ther’s feet’ (op. cit. p. 16, 45). Shastri (2000, p. 62), who obviously had access tothe Sanskrit texts of these two inscriptions, notes that ‘Dharasena pays respect toŚrı Dhruvasena and Śrı Ajjaka as well, while the phrase pertaining to Śrı Ajjakais omitted in his later edicts dated year 330 (648 A.C.). As ajjaka [Skt. aryaka],means a maternal grandfather and as Hars.a is known to have died in 646 or 647A.C., the reference to Ajjaka in Dharasena’s earlier edicts seems to apply to Hars.a.’Judging by Dharasena’s exorbitant pretensions, this grandfather had evidently losthis grip on his grandson in the last year of his life (cf. Devahuti 1998, 64).

The Reign of Hars.avardhana 113

led to the exile of Dharasena III (p. 108).

A looming threat from the WestAfter the inconclusive tug-of-war between Kanauj and Badami in the early630s, the following years may have been relatively peaceful, seen fromthe perspective of the Emperor Hars.avardhana—at least this is what oursources allow us to infer. At the three extremities of his realm, the north-east, the south-west, and the south-east, Hars.a had ensured himself ofreliable allies, the king of Kamarupa, Bhaskaravarman, the king of Vala-bhı, Dhruvasena II, and the king of Śrıpura, Mahaśivagupta Balarjuna.The conflict in the east was not yet decided, but Magadha was under thecontrol of his appointee, King Purn. avarman, until the middle of the fourthdecade. The western part of his realm, no doubt, consisted of his homecountry, the kingdom of Thanesar. Beyond, to the west thereof, lay thekingdoms of the Panjab, Kashmir, and Sindh.

Among the petty feudatory states that fell within Thanesar’s sphereof influence and that may have belonged to Hars.a’s dominion from thestart is the kingdom of Jalandhara, i.e. the Doab between the rivers Beas(Vipaśa) and Sutlej (Śatadru) around the present-day city of Jalandharin the eastern Panjab.341 Its strategic importance, which it shared withthe more western kingdoms of the Panjab, was above all economic, as itwas a storehouse of rice and wheat.342 Beyond Jalandhara, from the Beasto the Indus river, i.e. in the western Panjab, was the kingdom of T. akka,which Xuanzang called Tseh-kia, in which lay the ancient city of Śakala(Sialkot), once the stronghold of the Hun. as under Mihirakula.343 It seemslikely that Hars.a maintained good relations with its rulers, since we do notpossess indications that he carried arms again against this country, afterhe had initially joined the campaign of his elder brother Rajyavardhana inhis early youth.344 Its king may have acknowledged Hars.a’s supremacy.345

However, this supremacy seems to have been contested by the king ofSindh, whose kingdom was said to stretch as far north as Kashmir (seebelow). The information regarding the kingdom of Sindh in this periodstems mainly from two sources, Xuanzang and the Arab/Persian chronicleCac-nama.346 Xuanzang reports having visited the country in c. ad 640,

341 Si-yu-ki I, 175–77. Devahuti 1998, 103f.342 Si-yu-ki I, 165, 175f. Cf. Arrian, Anabasis of Alexander V. 25.343 Si-yu-ki I, 167. The name Tseh-kia/T. akka may be related to Skt. tunga, ‘mountain’.

See Turner CDIAL, s.v. *t.akka-3.344 Hars.acarita, chapter 5; Rajyavardhana returns victoriously in chapter 6.345 Devahuti 1998, 104.346 The Cac-nama is a work translated into Persian by Muh. ammad ↪Alı ibn H. amid ibn

Abı Bakr Kufı in the early 13th century ad from an Arabic manuscript. According

114 Hans Bakker

when it was said to have been ruled by a king ‘of the Śudra (Shu-t’o-lo)caste.’347 This ruler must have been King Cac (Chach)—known from theArabic chronicle as a brahmin, who usurped the throne after the deathof Sahasi II, son of Sahiras—or his successor, his brother Chandar.348

From Xuanzang’s description it emerges that Sindh was a rich countrythat derived its wealth from the production of salt and the breeding ofcattle, both important export products.349 When Ban. a was introducedto Hars.a’s mobile court he saw in its stables horses imported from thewestern countries, among which Sindhudeśa and Paraśıka.350

A wordplay in Ban. a’s eulogy of the emperor in his Hars.acarita351 is

to Sir Henry Elliot the Arabic original must date from before ad 753 (Elliot &Dowson I, 136). Elliot & Dowson only published a few extracts from this work,but a full English translation was published by Mirza Kalichbeg Fredunbeg in1900. This English translation is included in Persian Literature in Translation ofthe Packard Humanities Institute, which may be consulted on the website: http://persian.packhum.org/persian/main. This source gives the following geographicaldescription of the kingdom of Sindh in the seventh century:

‘The town of Alór [Al-Rur, modern Rohri, H.T.B] was the capital city of Hindand Sind. It was a town adorned with various kinds of royal buildings, villas,gardens, fountains, streams, meadows and trees (and was) situated on the bankof a river called the Mehran [Mihran, H.T.B.]. In this beautiful and splendidcity, there lived a king whose name was Sahiras son of Sahasi Rai. This king hadinnumerable riches and immense buried treasures. His justice was well knownin the world, and his liberality and bravery (were) widely spread. The limits ofhis dominions extended on the east to the boundary of Kashmir, on the west toMakran, on the south to the coast of the sea and Debal, and on the north tothe mountains of Kurdan and Kıkanan (the mountainous region south of Kabul:Paktia, Waziristan).’

347 Si-yu-ki II, 272. Xuanzang’s account of Sindh may have been second-hand. Histravel routes in the Si-yu-ki and the Life diverge. Watters is sceptical too, quotingapprovingly General Haig (The Indus Delta Country : a memoir chiefly on itsancient geography and history and topography, London 1887): ‘Hiuen Tsang’sSindh, in fact, is not the Sindh of any period known to history, and his descriptionof it is wholly irreconcilable with the facts which we gather from the contemporaryhistory embodied in the “Tarıkh Hind-wa-Sind” [i.e. the Tarıkh al-Hind wa-l-Sind= Cac-nama].’ See also Devahuti 1998, 71 n. 5.

348 Elliot & Dowson I, 411. King Cac, his name and descent, has been discussed byElliot and Dowson (Vol. I, 409–14). Elliot thinks that ‘Chach’ may be the Persianvariant of the Arabic ‘Sassa’ (op. cit. 410 = Sasa?). For a discussion of Cac’sdynasty see also Devahuti 1998, 71 n. 5. The Cac-nama tells us that at the timeal-Mughıra attacked Debal, i.e. in ad 636, King Cac was reigning in his 35th year(see below, p. 117; Elliot & Dowson I, 412). The same chronicle informs us that hereigned for forty years (Elliot & Dowson I, 152), which means that he died in c.ad 640/41, precisely when Xuanzang is supposed to have visited Sindh.

349 Si-yu-ki II, 272–74.350 HC 2 pp. 100f. [28f.] See above, p. 96.351 HC 3 p. 147f. [40]: atra purus.ottamena sindhurajam. pramathya laks.mır atmıkr. ta |

‘He (i.e. King Hars.a), as the foremost of man (Purus.ottama, i.e. Vis.n. u) gainedpossession of wealth (Laks.mı), after he had churned the Lord of the Rivers, i.e.

The Reign of Hars.avardhana 115

often taken as referring to a conflict between Hars.a and the country ofSindh. Devahuti even speaks of an ‘occupation of Sindh.’352 However, ifthere was a conflict between the kingdoms of Kanauj and Sindh, it wasprobably over Jalandhara.353 The Arabic chronicle Cac-nama corroboratesthis surmise, as it reports the long march northwards of King Cac. Fromthe Cac-nama we have the impression that this expedition took placerather early in his career, which may mean in the second or third decadeof the seventh century.

He (i.e. Cac) sent trustworthy men to different directions and called uponthe rulers of divisions to join him in his undertaking, which was to lead anarmy up to the limits of Hindustan, where it adjoined the country of theTartars [Turk]. Then, at an auspicious time, found out by astrologers aftercarefully consulting the stars, he gave marching orders. After doing many aday’s journey and passing several stages, he arrived at the fortified town ofBabiah [Papiya], which is situated on the southern bank of the river Beas.The governor of that fort commenced hostilities with him. (But) after agreat and bloody battle the ruler of Babiah fled, and threw himself into thefort, and Rai Chach was the victor. . . . (The governor escaped from the fortand entrenched himself in another fort named ‘Iskandah’ [Askalanda, = al-Iskandariyya?]). . . Now, there was a brave chief in the fort of Iskandah,who had always been obedient and well disposed to Chach. . . . Chach senta messenger to him, and offered to make him sole governor of the fort,and directed a royal order to be inscribed, containing a solemn promiseto bestow the governorship on him, and also to hand over Babiah to him,as soon as he killed Jetar (or Chitra) [Chatera/Chatar/Jatra], the king ofBabiah, or took him prisoner. . . . (Eventually this chief) killed King Jetarand sent his head to Chach. Rai Chach showed him great favour, andgranted him a robe of honour, with many other presents, and ordered thecastle to be entrusted to him as its governor.354

The country of the Tartars may refer to the descendants of the Hun. as inthe country of T. akka, and the limits of Hindustan may refer to the limitsof Hars.a’s dominion. The fortified town of Babiah (Babiah) has not beenidentified, but given its location on the southern bank of the Beas River,it cannot have been too far away from the present city of Jalandhara. Thefort Iskandah (Alexandria) may have been somewhere to the north-west of

the ocean/the king of Sindhu (Sindhuraja).’352 Devahuti 1998, 105. Apart from Ban. a’s pun, she does not adduce any evidence for

this contention. The Chinese sources to which she refers are silent regarding theprovince of Sindh. For a more nuanced view see Majumdar in HCI III, 106.

353 For a description of the religious significance of Jalandhara see Barkhuis in Bakker& Entwistle 1983, 58–69.

354 Quoted from Fredunbeg’s translation of the Cac-nama at http://persian.packhum.org/persian. Names and toponyms between square brackets have beentaken from Elliot & Dowson’s translation (volume I, 140f.). Passages in italicsbetween round brackets are my summary.

116 Hans Bakker

the River Beas (Hypasis), the furthest point reached by Alexander.355 Thehead sent to King Cac may have been that of a governor of Hars.avardha-na. Cac’s control over Jalandhara, being so far from his own power base,was short-lived. Hars.a regained control over that region. The Life relatesthat Xuanzang was given an escort by Hars.a when he was preparing toleave the country by the end of ad 643, and this task was entrusted to‘the king of North India Udhita (Udita),’ who reigned from ‘Che-lan-ta(Jalandhara), the royal city of North India.’356

To the north of ‘the country of the Tartars’ (i.e. Sialkot), beyondthe borders claimed by King Cac, lay the kingdom of Kashmir. Deva-huti speaks of Hars.a’s ‘Kashmir exploit’ and argues that the king ‘cowed’its ruler (Durlabhavardhana) ‘into submission,’357 just to acquire the re-putedly light-emitting relic of the Buddha, a tooth, to be precise. Theonly source for this story is Xuanzang.358 We may safely relegate it to therealm of Buddhist fables, taking it as testimony to the credulity of thepious Chinese pilgrim (and Devahuti). There is no evidence whatsoeverthat Hars.a ever set foot in Kashmir.

Let us return to Sindh. This rich trading country between east and westwas among the first regions of South Asia to make the acquaintance ofthe unleashed Arab forces. Within four years of the death of the prophetMuh. ammad, under the second caliph ↪Umar, ‘a military expedition set outfrom ↪Uman, to pillage the coasts of India’ in ad 636.359 The tapping of thewealth of India had begun.360 The Futuh. al-Buldan of Ah.mad ibn-Yah. yaal-Baladhurı († 892) reports that under the caliphate of ↪Umar, ↪Uthmanwas appointed to Oman, and his brother al-H. akam to Bahrain. ↪Uthmansent a force to Tana (Thane near Bombay), while his brother al-H. akam‘despatched a force to Barauz [Broach]; he also sent to the Bay of Debalhis brother Mughıra [al-Mughıra], who met and defeated the enemy.’361

This report is contradicted by that of the Cac-nama, which deals with thislast event more elaborately and is worth quoting in full.

↪Usman [↪Uthman] himself started in the direction of Amman [i.e. ↪Uman]

355 Arrian, Anabasis of Alexander V. 24. The eastern-most Alexandria that we know ofwas on the River Acesines, i.e. the Chenab. See Arrian op. cit. V. 29. This chapteralso relates how Alexander had altars set up at the Hypasis.

356 Life 189f. It remains an open question whether this King Wu-ti-to (Udita) is thesame as ‘the former king of this land,’ who was appointed by the ‘king of Mid-India’ as ‘sole inspector of the affairs of religion (the three gems)’ (Si-yu-ki I, 176),as proposed by Watters (I, 297) and adopted by Devahuti 1998, 103f.

357 Devahuti 1998, 111f.358 Si-yu-ki I, 158f.; Life, 181–83.359 Elliot & Dowson I, 415.360 Wink 1991, 39.361 Elliot & Dowson I, 116.

The Reign of Hars.avardhana 117

or the high sea, and his army was sent in a small fleet of boats, by wayof the sea under the command of Mughairah [al-Mughıra] the father ofA’s [al-↪As.] to Bahrain, and thence to Debal. At that time, the country ofSind was in the hands of Chach son of Selaij, 35 years of whose reign hadalready passed. The people of Debal were mostly merchants. Samah son ofDewaij was the ruler of the place, on behalf of Chachrai. When the Arabarmy arrived at Debal, he issued out of the fort and engaged with them infight. It is related by one of the intelligent men who were with him, thatwhen the two armies stood opposite each other, Mughairah the father ofA’s drew his sword, (and attacked the enemy’s force) crying: ‘In the nameof God and in the cause of God’ and fought till he was killed.

Cac-nama quoted from http://persian.packhum.org/persian/main?url=pf%3Ffile%3D12701030%26ct%3D0+.

Debal is to be situated in the Sindhian Delta at the mouth of the IndusRiver,362 and Barauz is the port of Bharukaccha in the Gulf of Cambay.

Excavations at Banbhore/Daybul 52 km south-east of Karachi, throwa new light on the identification of Debal. ‘In the pre-Muslim layers abuilding was exposed which obviously was a Śiva shrine judging by thediscovery of two lingas, one of which was still standing on its yoni.’363

Van Lohuizen remarks that it is tempting to identify this structure withthe temple described in the Cac-nama, which reports that: ‘In the heartof Debal there was a magnificent idol-temple, of a towering height, witha dome built on its top, and a green silken flag fixed on it. The buildingof the temple was 40 cubits high, and the dome on it was also 40 cubitsin height.’364 The identity of Banbhore and Debal has long been debatedamong archaeologists, but convincing arguments for such an identity havebeen advanced by Monique Kervran.365 The excavator of Banbhore, F.A.

362 Elliot & Dowson I, 374ff. discuss the original location of Debal (cf. Lambrick 1964,153). It has been identified with Tatta, 100 km east of Karachi, but Elliot ‘enter-tain[ed] little doubt that Karachı itself represents the site of Debal’ (op. cit. 375).For the probable location see below.

363 Van Lohuizen 1979, 171, referring to Pakistan Archaeology No. 1 (1964), p. 53.364 Cac-nama quoted from http://persian.packhum.org/persian/main?url=pf.365 Kervran 1999, 113ff. Pakistan Archaeology 1 (1964), 50f. reports: ‘Banbhore, an

early Islamic site, is situated on the north bank of the Gharo Creek, near theKarachi–Hyderabad highway about 40 miles east of Karachi. It conceals the re-mains of a considerable ancient settlement divided into two parts: the fortifiedcitadel area of the mound itself measuring 2,000 by 1,000 feet; and the lower un-walled city round the lake at the foot of the mound, extending far outside, whichincludes an industrial area and also an ancient graveyard. It is well situated to havebeen a port of some importance, once connected with the interior of the countryby a branch of the Indus. Hydrographical studies have established that the GharoCreek once formed the westernmost branch of the great river, which has shiftedits course and is now flowing by Thatta, 25 miles east of the site. Its identificationwith Debal, the famous Hindu port which fell to the young Arab general, Muham-mad bin Quasim, has more than once been suggested.’ Khan 2005, 45 concludes:

118 Hans Bakker

Khan, describes the temple found as follows.

The most interesting building uncovered in the Pre-Muslim levels is a SivaTemple in the western part of the citadel. It has thick mud brick wallscovered with several coats of red paint on lime plaster, an ornamentalpedestal, also thickly coated with lime plaster, and other architecturaldecorations. A considerable area round the temple is covered by a fineconcrete floor decorated wih sea-shell. The plan of the temple was clearedfully after removing the superimposed building remains of later period fromthe area.

Near the pedestal, which was probably meant for a large deity, two votiveSiva Lingams—one complete with Yuni and pedestal—were found lying onthe floor. This is unmistakable evidence of Hindu occupation of this site.It may be mentioned here that a number of large monolithic Siva Lingamshave been recovered from other parts of the site as well, and this cultappears to have been popular in the period immediately before the adventof the Muslims.366

In view of its location in the Indus Delta, the Śiva temple found inBanbhore may indeed be the famous Śiva temple named Śankukarn. eśvara,which the Skandapuran. a locates at the confluence of the Sindhu (Indus)River and the ocean, describing it as the place where Hiran. yakaśipu prac-tised his asceticism.367 Śankukarn. eśvara, identified with the great templein Debal as described in the Cac-nama, may thus be one of the very fewtemples mentioned in the Skandapuran. a of which we have a contempora-neous description.

One year after the attack on Debal, in June 637, the victorious com-mander of the Arab forces Sa↪d ibn Abı Waqqas. entered the Persian capitalCtesiphon in triumph. The King of Kings, Shahanshah, the Sasanid em-peror Yazdagird III and his court were on the run. The final defeat, whichthe Sasanid army suffered at Nahavand in 642, sealed the fate of the Per-sian empire. With the conquest of the Makran coast in 643, the Arab landforces reached the borders of India, eleven years after the death of theProphet. For the time being the river Indus served as the frontline.368

‘The question of the identification of Banbhore with Debal has now been solvedbeyond reasonable doubt.’

366 Khan 2005, 7–9. Cf. Pakistan Archaeology 1 (1964) in which Plate XVI-B illus-trates the linga on the yoni found ‘in situ on the floor of a Hindu temple underthe mound,’ and Plate XVII-A shows a large Śiva linga ‘re-used in the area of theGreat Mosque.’

367 SPBh 73.6: sa śam. kukarn. am. samprapya sindhusagarasam. game | tapas tepe diteh.putro devam. dras.t.um. pinakinam ‖ 6 ‖ After all, ‘Debal’ may derive from devalaya,i.e. ‘Temple.’ According to SPBh 73.69–71, after Śiva had practised tapas at thissite, Brahma installed the linga there along with creating other tırthas, such asDaśaśvamedhika and Pancaśvamedhika. See also n. 471 on p. 151 and n. 534 onp. 173.

368 Hitti 1970, 156f.

The Reign of Hars.avardhana 119

The sudden collapse of its big western neighbour, the Persian empire,which had been a steady factor throughout the rule of Kanauj in the firstcentury of its existence, must have sent a shock wave through Hars.a’srealm and beyond, though it has to be admitted that we have little pal-pable evidence for this. It must have meant, at least for some period, adisruption in commercial activities, which, in the Arabian Sea, had beenlargely under Sasanid control.369 The small ‘trading diasporas of the Zoro-astrians’ seem to have survived this disruption. Contrary to what has oftenbeen assumed, however, the Muslim conquest of Iran did not immediatelylead to an exodus of Zoroastrians.370

Foreign Affairs: The empire in the international arena

The political stability in Northern India during the last seven years ofHars.avardhana’s rule contrasts sharply with the momentous events thatwere taking place in the outside world. As we have seen, the major powerin the west, the Sasanid kingdom, collapsed in ad 642 and the conquestof the Makran coast began in the following year.

In the north Hars.a’s empire bordered on the kingdom of Nepal. Here aturnover took place between ad 641–42, when the Licchavi king Narendra-deva returned to the Kathmandu valley from Tibet. His father Udayadevahad earlier fled there after a coup d’etat by Jis.n. ugupta (c. ad 622–23) and

369 Wink 1991, 68. Xuanzang’s report of his visit to Bharukaccha in c. ad 640, afterthis first Arab raid, depicts a town that can hardly be called thriving (Beal II,259f.).

370 Wink 1991, 104 refers to the Buddhist historian Taranatha (without citing hisprecise source), to whom he ascribes that ‘the Buddist emperor Harsha [. . . ] en-tertained a community of 12,000 Zoroastrians for some time.’ This statement ismisleading as it stands, or at least a tendentious representation of the followingfull text (in the English translation by Lama Chimpa and Alaka Chattopadhyaya):‘Among them, King Śrıhars.a was incomparable as a king. He wanted to wreck thisreligion of the mleccha-s. In a small place near *Maul-tan, he built a *masita,that is a big monastery of the mleccha-s. The whole of it was made only of wood.He invited all the mleccha teachers there, lavishly offered gifts to them for severalmonths and made them collect all their scriptual works there. Then he set fire to it,and, as a result, twelve thousand experts of the doctrine of the mleccha-s perished.At that time, there lived in *Khorosan only a weaver well-versed in the mlecchareligion. From him the mleccha-s of the later period gradually grew in number.As a result of this wreck [of the mleccha religion] by this king, there remained forabout one hundred years only a few to follow the religion of the Persians and theTurus.kas.’ (Taranatha’s History of Buddhism, p. 178 (= fol. 64B)). The secondhalf of this quote seems to make it clear that this ‘wreck’ can hardly apply to theZoroastrians. The whole story seems to be spurious. It has been suggested that themleccha religion at issue may have been Manichaeism, but the story goes so muchagainst the spirit of King Hars.avardhana as we know it from other sources, that Ihave little doubt that this account goes back to a historical misunderstanding orconfusion. Moreover, Multan lay way beyond Hars.a’s sphere of influence.

120 Hans Bakker

had given a Nepalese princess, Bhr.kut.ı, to his Tibetan host, King Srong-brtsan sgam-po. Twenty years after their flight, however, in c. ad 641,the Licchavis came back and the independent rule of the Mahasamantaseffectively came to an end when the Tibetan king Srong-brtsan sgam-porestored Udayadeva’s son Narendradeva to the throne and made Nepalindebted to Tibet. The Mahasamanta Vis.n. ugupta (successor of Jis.n. u-gupta) and the puppet king Bhımarjunadeva quitted the scene.371

The Tibetan king had also aligned himself with the Tang emperor bymarrying the Chinese princess Wencheng in ad 641. These events pavedthe way for a direct exchange between Hars.a and the Chinese emperorTaizong via the Nepal-Tibet route.372 It is this development, we wouldlike to argue, that prompted Hars.a to send an embassy to the Tang court,rather than the encouragement of the Chinese pilgrim Xuanzang (above,n. 284 on p. 98). By ad 643 Hars.a thus saw himself confronted by two new,expansionist world powers: the Arab forces in the west and the Chinese-Tibetan coalition in the north.

If the above argument is accepted, it becomes evident why Hars.a sawa need to send a delegation up north in ad 641. There was a new politicalreality at his northern border which was potentially threatening, that is tosay, endangering the good and peaceful relationship that had existed untilthen. It is another sign of Hars.a’s good statesmanship that he immediatelynoticed this potential and wanted to keep a finger on the pulse.

Devahuti has shown that there is no evidence that Nepal was tributaryto Hars.a,373 and she is equally correct when she observes that ‘in viewof Vardhana prestige and power, the regular use of the Nepal route byChinese and Indian secular and religious missions, and the close historicaland cultural connections between India and Nepal, the relations betweenthe two are likely to have been very intimate and purposeful.’374 Thisapplies, we would like to argue, in particular to the period in which theRegents (Mahasamantas) Am. śuvarman and his successors Jis.n. ugupta andVis.n. ugupta were the de facto rulers of Nepal (ad 605–641).

This new departure in Nepalese history is marked, among other things,by a change in era. Whereas the older Licchavis dated their reigns in the(Indian) Śaka Era, Am. śuvarman introduced a Nepalese era, which has its

371 Slusser 1998 I, 32. Despite this change of regime, the restored Licchavi king Naren-dra continued the patronage of the Śr.nkhalika Paśupatas in the PaśupatinathaTemple (see below, p. 149), a patronage which was initiated, it would seem, byJis.n. ugupta. Actually Narendradeva, and all Licchavi kings after him, embracedPaśupati as the deity of the kingdom/state, an idea promoted by the ousted Maha-samantas (see Mirnig 2013, 337f.).

372 Sen 2003, 22.373 Devahuti 1998, 122–127.374 Devahuti 1998,127f.; cf. Slusser 1998 I, 5.

The Reign of Hars.avardhana 121

start at an unknown event in ad 576. It is known as the ‘Manadeva Era,’ asit is sometimes ascribed to an otherwise unknown predecessor of Am. śu-varman, Manadeva II,375 or as the ‘Am. śuvarman Era,’ after the regentwho first introduced this era in his edicts, the earliest occurence of whichis the year 29 (ad 605).376 We venture the hypothesis that the start ofthis new era refers to Am. śuvarman’s year of birth,377 which would entailhim entering the service of the Licchavi king Śivadeva I as a young man atthe age of about eighteen and seizing power when he was twenty-nine.378

Intense religious acculturation of Indian forms of religion in Nepal hadno doubt been a significant process since the early Gupta empire, but therule of the Mahasamantas in the first four decades of the seventh centurycontributed something new. Under their rule, Paśupati became the na-tional deity and his temple a state sanctuary. While still ruling under theLicchavi king Śivadeva, Am. śuvarman declared his devotion to Śiva (re-ferred to under his name Bhava) as follows: ‘[Am. śuvarman,] who securedthe well-being and bliss in the future by means of being fully intent onhumbly worshipping the lotus feet of Lord Bhava . . . .’379 In his BunmatıInscription of ad 605, the first inscription in his own name, the Mahasama-nta derived the legitimation of his power directly from Paśupati himself,and this would become the standard for Nepalese kings: ‘The illustriousMahasamanta Am. śuvarman, who is blessed by his venerable Master and

375 It seems that the name ‘Manadeva Era’ was invented by Baburam Acharya (1940;not available to me), referred to by Slusser 1998 I, 385. In Nepalese sources thisera is referred to as simply ‘Sam. vat’ and the hypothesized Manadeva II has notleft any inscriptions.

376 Am. śuvarman’s Bunmatı Inscription, Vajracarya 1973 (LA 71); Slusser 1998 I, 114.The inscription ends with . . . sam. vat 29 jyes.t.haśukladaśamyam | (Vajracarya 1973(LA), p. 291).

377 Slusser 1998 I, 385–89 discusses this era and its possible origin, without coming to asatisfactory conclusion. Earlier, Petech 1984, 12f. established the epoch year of ad576 without discovering, however, the event that gave rise to this new reckoning.But Petech 1984, 13 makes a very interesting observation: ‘We can, at any rate,suppose that the expression Śrı-Paśupati-bhat.t.araka-samvatsara implies that thefoundation of the era was due to some religious event connected with the nationalshrine of Paśupati Nath.’ Whether the nexus is causative or concomitant remainsto be shown. I conjecture alternatively that the origin of this era might have beenpolitical rather than astronomical or religious. Before this era was recognized,its years were assigned to the Hars.a Era and, on the basis thereof, it has beenconjectured that Am. śuvarman was a vassal of Hars.avardhana (see Devahuti 1998,122–126).

378 The first dated inscription in which he is mentioned as regent under Śivadeva isfrom Śaka year 516 (= ad 594). Slusser 1998 I, 25; Vajracarya 1973 (LA 60). Thefirst edict in Am. śuvarman’s own name dates to ad 605. Vajracarya 1973 (LA 71).

379 Vajracarya 1973 (LA 68), p. 274 ll. 5–6: bhagavadbhavapadapankajapran. amanu-s.t.hanatatparyopayattayatihitaśreyasa.

122 Hans Bakker

Lord Paśupati and favoured by his venerable father, . . . .’380 And in theHam. d. igaun Inscription dated to Sam. vat 32 (ad 608) he appointed Bhaga-vat Paśupati (i.e. the temple) as the foremost recipient of redistributedtaxes.381

Unfortunately we have no records that throw more light on the rela-tionship between King Hars.a and the Mahasamantas of Nepal. In view ofthe flourishing of Hinduism, and Śaivism in particular, in Nepal duringthis period, we assume that these relations were manifold and intensive.

The Chinese, on the other hand, have recorded their political history,and from that we know that Hars.a’s embassy in ad 641 met with con-siderable success. We are informed about this by the New History of theTang (Xin Tang shu), which reports that in response to Hars.a’s mission,the Tang emperor sent Liang Huaijing (‘a mid-level official’), who arrivedin Kanauj in the company of Hars.a’s envoy. I quote the New History ofthe Tang in Tansen Sen’s translation.

Hars.a asked the people of his country: ‘Have any envoys from Maha-cına come to my country in the past?’ They all answered ‘No.’ [Theking, then,] came out [of the palace]. Prostrating himself, he receivedthe imperial letter and placed it on his head. He again sent an envoy[with Liang Huaijing] to the Chinese court. (Sen 2003, 20)

All this is supposed to have happened in ad 641, a year before Hars.a’sOrissa campaign at the end of which he met Xuanzang. Sen describes onemore embassy from the Tang court to India during Hars.a’s lifetime, ledby Li Yibiao, arriving in Magadha in early ad 644, just after Xuanzanghad left Kanauj. Wang Xuance formed part of this mission and his records(the Fayuan shulin) provide precious information about the Chinese mo-tives.382 According to Tansen Sen, the ‘Chinese envoys first attended aBuddhist ceremony organized by Hars.a, where they also gained an audi-ence with the king of Kamarupa. In early 645, the mission reached thecity of Rajagr.ha, where they placed an inscription tablet at the foot of

380 Am. śuvarman’s Bunmatı Inscription, Vajracarya 1973 (LA 71); Slusser 1998 I,114. The inscription opens with: svasti kailasakut.abhavanad bhagavatpaśupati-bhat.t.arakapadanugr. hıto bappapadanudhyatah. śrımahasamantam. śuvarma kuśalı |(Vajracarya 1973, p. 290 ll. 1–2). A detailed study of this development is foundin Mirnig 2013, 328f., 331. Cf. Torzsok & Ferrier 2008, 101–03 (on the meaning ofpadanudhyata).

381 Vajracarya 1973 (LA 77), p. 320 ll. 6–7. The first lingas that we know to have beeninstalled in the Paśupatiks.etra were presented along with a land donation by themerchant Pratihara Dhruvasangha Vartta in ad 533 (cf. Mirnig 2013, 339f.).

382 Cf. Devahuti’s (op. cit. p. 241) rendering of the records of Wang Xuance: ‘In thetwelfth month of this year they arrived in the kingdom of Magadha. They trav-elled over it and visited the lands of the Buddha, and contemplated the survivingremains of the Buddha.’

The Reign of Hars.avardhana 123

the sacred Gr.dhrakut.a mountain. A month later, the envoy visited theMahabodhi Monastery in Bodh Gaya and placed an inscription beneaththe Bodhi Tree . . . .’ In addition to diplomatic and spiritual aims, how-ever, the Chinese seem to have taken a keen interest in ‘acquiring thetechnology of making sugar.’383

Fragmentary as they may be, these testimonies to Hars.a’s internationalpolicy show that the king of Kanauj had an open attitude towards theworld outside India. We are lucky to have the Chinese sources for this.Despite the absence of similar evidence of Hars.a’s dealings with his newArab neighbours in Iran, it cannot easily be concluded that such contactsdid not exist. They simply may not have been recorded, although theabsence of a shared religious interest will no doubt have limited diplomacyto the spheres of power and trade. What strikes us in all the Chineserecords that pertain to Hars.a’s India is the stability of his kingdom andhis apparently unchallenged sovereignty. One wonders how many of themen living in northern India in those last seven years, with the surroundingworld changing at such a rapid pace, could have divined that their strongand prosperous country would fall apart very soon indeed.

The ‘Arena of Charitable Offerings’

In ad 643 Hars.avardhana organized his sixth quinquennial assembly. Itwould be his last. This meeting in Prayaga was attended by the Mai-traka king Dhruvasena II (his presumed son-in-law, see above, p. 107),his life-long friend and ally, the king of Kamarupa, Bhaskaravarman, andthe Chinese pilgrim Xuanzang. The latter’s biographer Huili relates thefollowing.

The king said: ‘Your disciple (i.e. King Hars.a), succeeding to theroyal authority, has been lord of India for thirty years and more: Ihave constantly regretted the small increase to my religious merit,resulting from a want of previous good deeds. In consequence of thisI have accumulated every kind of treasure and precious substance inthe kingdom of Prayaga, and between the banks of the two rivers,I have established a great religious convocation every five years, toattend which all the Śramans and Brahmans of the five Indies are in-vited; and besides these the poor and the orphans and the destitute;on this occassion during seventy-five days the great distribution ofalms called the Moksha is attended to; I have completed five of theseassemblies and am now about to celebrate the sixth: why does not

383 Sen 2003, 38; ibid. pp. 146–150 on its supposed impact on Sino-Indian trade.

124 Hans Bakker

the Master (i.e. Xuanzang) delay his departure till then, and so, bywitnessing the spectacle, rejoice with us?’ (Life p. 183f.)

The pilgrim decided to postpone his departure and to attend the spectacle.In a highly organized and hierarchical society such a potlatch ceremonyis of course only possible if the coffers are quickly refilled. The concludingremark in Xuanzang’s Records is therefore revealing, though it does notcome as a real surprise: ‘After this the rulers of the different countries offertheir jewels and robes to the king, so that his treasury is replenished.’384

Yet it would have been a great tamaśa. The order of precedence wouldhave been established. In the charity arena, status and prestige wouldhave been confirmed or denounced, climbers and losers would have re-joiced and resented. Hars.a’s gain was not only the confirmation of hissovereignty; his magnanimity would have greatly enhanced his prestigeand consequently the consensus by which he ruled. And in the religioussphere it would, in the eyes of himself and his contemporaries, have en-hanced his balance of merit. The fact that the invitation was extended tothe brahman. as and śraman. as would have significantly contributed to thataim and it would have made this happening a cultural and religious eventof the first order, probably rather similar to the present-day Kumbh Mela.But while the present-day gathering in Prayaga is completely Hindu incharacter, Hars.a’s Buddhist leanings would have guaranteed an œcumeni-cal exchange. And, Hars.avardhana’s mobile court and its entourage, on itsway towards or in Prayaga, provided a perfect audience for the recitationof new Sanskrit compositions—the Skandapuran. a may have been a casein point, sung by a Singer of Ancient Lore.385

As will be argued below (p. 137), we consider it most likely that theprocess of composition of this text, i.e. the finalizing of its first redaction,took place before or in the middle of Hars.a’s reign. It seems unthinkablethat the king, a literary man himself of no mean qualities as well as a para-mamaheśvara by his own account,386 would not have been eager to hearthe myths and wonders of Maheśvara as contained in this newly availablePuran. a, which relates, among so many other things, the foundation mythof his own native city (below, p. 161). This hypothesis finds support in theopening verses of the Skandapuran. a, which had its premiere not in theusual Naimis.a, but in Prayaga:

The sages, assembling in Prayaga to bathe in the confluence ofGanga and Yamuna on the day of full moon, see the Singer of An-

384 Si-yu-ki I, 233. The Life p. 187 suggests that the vassal kings bought back thedistributed riches and restored them to the king.

385 Cf. Ban. a HC 3 p. 140 [39]. Below, p. 156.386 E.g. The Madhunan Copperplate of Harsha, dated Samvat 25, EI I, 72 l. 8: parama-

maheśvaro maheśvara iva sarvasattvanukampı.

The Reign of Hars.avardhana 125

cient Lore (pauran. ikasuta) coming towards them to pay his respects.A seat is prepared for him. After he is seated they ask him aboutthe birth of Karttikeya (Skanda), a story that equals the story ofBharata (i.e. Mahabharata) and surpasses the Puran. a (i.e. the textreferred to as the Puran. apancalaks.an. a), both of which he, like an-other Son of Satyavatı (Vyasa), already told the sages in the Naimis.aforest on the occasion of a continuous recital of Sacred Lore (brahma-sattra). The Singer announces the story of the birth of Skanda, ofhis friendliness towards brahmins, his glory and his heroism, greaterthan (that of) the gods, after which the Singer begins his recital ofthe Skandapuran. a.387

First and foremost, however, Hars.a’s mobile court was the perfect settingfor the performance of the king’s own literary compositions, the Ratnavalıor the Priyadarśika, in which Śiva and Gaurı are invoked in turn as theguarding deity.388 Our Chinese sources throw some light on the chronologyof events.

Xuanzang took leave of Hars.a after this event in Prayaga, in April643, or a bit later.389 The Life (p. 176) records that Hars.a arrived in Ka-nauj after a march of ninety days from his camp at Kajurgira in ‘the lastmonth of the year,’ whereas the Records tell us that the procession fromKanauj to Prayaga was on its way in ‘the second month of spring-time’(Si-yu-ki I, 218). Though the ‘second month’ seems too late, it may safelybe inferred from this scattered information that the New Year in Hars.a’srealm began in spring, i.e. in Caitra, that is in March,390 and that the Pra-yaga assembly was held in Spring ad 643; it is therefore to be consideredas part of the Spring Festival (Vasantotsava).391

We have observed that the quinquennial assembly in Prayaga mayhave offered an opportunity for staging theatrical productions.392 The

387 Synopsis of SP 1.4–13.388 Cf. the fourth pada of the third Sragdhara verse of the Ratnavalı: śam. sann ity

at.t.ahaso makhamathanavidhau patu devyai śivo vah. , or the fourth pada of the firstŚardulavikrıd. ita verse of the Priyadarśika: sparśad utpulaka karagrahavidhau gaurıśivayastu vah. . See also below, n. 398 on p. 127.

389 Watters 1904–5 II, 336.390 Kane V, 659 remarks: ‘At present the year begins in Caitra or Karttika in differ-

ent parts of India. But in ancient times it began in different months in differentcountries and for different purposes.’

391 For the Vasantotsava see Anderson 1988. Anderson (op. cit. p. 76) argues that ‘theVasantotsava was not, in fact, a specific Spring festival composed of specific ritualpractices to be celebrated at a specific time but, rather, that the term Vasantotsa-va was a generic term meaning simply and literally “Spring festival.” That is, theVasantotsava is a term which may include a larger number of festivals celebratedin spring and it may not itself be treated as a festival separate from these otherfestivals.’

392 At occasions like this the audience would actually match the one addressed by the

126 Hans Bakker

Sutradhara (stage director) in the Ratnavalı and Priyadarśika introducesHars.a’s plays with a request to the king, to the effect that the occasion ofthe Festival of Spring would afford the right opportunity to see the newplay performed, if he pleases.393

Xuanzang reports that, after the gathering of feudatories, Śramans andBrahmans in Kanauj preceding the quinquennial assembly, King Śıladitya(i.e. Hars.a) rode in procession (towards Prayaga, we presume), scatteringpearls and other riches, while being ‘dressed as Śakra,’ i.e. as Indra.394

According to the Chinese sources, the preliminary assembly in Kanaujtook place at the end of the year. The procession to Prayaga may thereforehave begun on the eve of the New Year. It could thus, in addition to beingpart of a spring festival, be seen as a New Year Festival. Regarding theNew Year Festival, Kuiper has remarked the following.

According to the Taittirıya Sam. hita Indra was born on new year’seve ekas. t.aka but his festival lasted only five to seven days, at least inthe account of Puran. ic texts, and the Indra Tree, which symbolizedthe cosmic centre and the world pillar, was at the end of the festivalremoved and thrown into a river. Though nothing is known abouta vedic Indra Festival, it is quite likely that the vedic contests of apotlatch-like character took place about the beginning of a new yearand stood under the patronage of Indra.’395

Sutradhara (below). Cf. Ali 2004: ‘In King Hars.a’s Ratnavalı and Priyadarśika theputative audience of subordinate kings attending the play are addressed as beingappreciative of poetic qualities’ (Ali 2004, 190; italics mine).

393 Ratnavalı (ed. Cappeller, p. 327) and Priyadarśika (ed. Kale): adyaham vasa-ntotsave sabahumanam ahuya nanadigdeśad agatena rajnah. śrıhars.adevasya pada-padmopajıvina rajasamuhenokto yatha | asmatsvamina śrıhars.adevenapurvavastu-racanalam. kr. ta ratnavalı/priyadarśika nama nat.ika kr. tety asmabhih. śrotraparampa-raya śruta na tu prayogato dr.s. t.a ‖ The Priyadarśika and Ratnavalı are said tobelong to the genre of the nat.ika or ‘light play.’ For a translation see above, p. 18.

394 Si-yu-ki I, 218ff. It was during this procession, in which a golden Buddha imagewas carried along on an elephant, that fire broke out and an attack on the lifeof Hars.a was foiled. Xuanzang also reports: ‘At this time there were but abouttwenty Śraman. as following in the procession, the kings of the various countriesforming the escort. After the feast they assembled the different men of learning,who discussed in elegant language on the most abstruse subjects.’

395 Kuiper 1979, 30. For Indra’s ancient connection with the New Year, see also Kuiper1961, 182: ‘In the preceding pages the word parya- has been studied in the lightof the theory that the nucleus of the Rigveda, and in particularly the hymns toIndra, are concerned with the New Year’s festival, and with ceremonial contestswhich aimed at the acquisition of wealth and social prestige, at prolongation of lifeand new vital energy.’ That the Indra festival—Indradhvajotthanotsava or Indro-tsava or Indramahotsava or Indramaha or Śakradhvajotsava—was traditionallycelebrated at the end of the year is attested in MBh 1.57.17–19:yas.t.im. ca vain. avım. tasmai dadau vr. tranis. udanah. |is. t.apradanam uddiśya śis. t.anam. paripalinım ‖17 ‖

The Reign of Hars.avardhana 127

Although most Sanskrit sources place the Indra Festival in autumn, notin spring,396 elements of this ancient festival may have been retained inthe New Year celebrations of Hars.a’s times, which took place around thebeginning of the month Caitra, and some may have been included in thefollowing spring festivities. It cannot therefore be ruled out that the thirdof Hars.a’s plays, the Nagananda, which is said to have been staged at theIndrotsava,397 was also performed on this occasion in spring—in partic-ular if we take seriously Xuanzang’s report saying that a statue of theBuddha was the main object of the procession, which caused the envy ofthe brahmins and heretics and led to an attack on the life of the king.398

The Nagananda is a play in which Hars.a ‘has made the attainment of em-pire by the hero the objective . . . , though the wish of the hero is only tosacrifice himself for others.’399 This could dovetail very well with Hars.a’sself-image and the ideology that he would like to broadcast at such aprestigious gathering of all his subjects and feudatories.

Finally Xuanzang reports the erection of a pillar in Prayaga, but itseems unlikely that this has any connection with the celebration of char-

tasyah. śakrasya pujartham. bhumau bhumipatis tada |praveśam. karayamasa gate sam. vatsare tada ‖18 ‖tatah. prabhr. ti cadyapi yas.t.yah. ks. itipasattamaih. |praveśah. kriyate rajan yatha tena pravartitah. ‖19 ‖

396 The issue is discussed in detail by Meyer 1937 III, 112–18; Kuiper 1979, 133–142.Varahamihira’s Br.hatsam. hita 43 and Atharvavedapariśis.t.a 19.1 place the festivalin Bhadrapada. The Indrotsava does not feature in the Skandapuran. a, but it ismentioned in connection with the immediately following Udakasevika-BhairavaFestival. This festival is only described in the RA recension, which is (partly)quoted by Laks.mıdhara and Can. d. eśvara (Harimoto 2006, 34). The Śakradhvajo-tsava is said to take place just before the full moon of autumn is visible, that is whenthe month Karttika begins and the Udakasevika-Bhairava Festival is celebrated.According to the RA recension of the SP, the Indrotsava is consequently celebratedin the bright half of the preceding month, viz. Aśvina. For the reconstruction ofthis passage and its interpretation see Harimoto 2006, 32f.

397 The nat.aka (drama) Nagananda begins (Steiner 1997, 122): adyaham indrotsave sa-bahumanam ahuya nanadigdeśagatena rajnah. śrıhars.adevasya padapadmopajıvinarajasamuhenokto yatha | It is the only play of which a source is explicitly mentioned,the vidyadharajataka (see Steiner 1997, 127). The Nagananda was thus staged onthe Indrotsava. As such it was a re-enactment of the primordial performance ofSanskrit drama and, concomitantly, of the cosmogonical strife between Indra andthe Asuras, as we learn from the Bharatıya-Nat.yaśastra (Kuiper 1979, 128f., 142–44). For the Indrotsava see, inter alia, Meyer 1937 III, 87–118; Kane II, 824–826,V, 274; Gonda 1967; Kuiper 1979, 129–42; Steiner 1997,123–26.

398 Si-yu-ki I, 219–221. Not Śiva or Parvatı but the Buddha is invoked in the tripartiteNandı or Blessing that precedes the play. See Hahn’s General Introduction to theNorth-Indian recension of the Nagananda 1991, VI–X. The fourth pada of thesecond Śardulavikrıd. ita verse runs: sers.yam. maravadhubhir ity abhihito bodhau ji-nah. patu vah. .

399 Warder 1989–1992 IV, 52; cf. Devahuti 1998, 179f.

128 Hans Bakker

ity; the Chinese pilgrim may have meant the famous Aśokan Allahabadpillar.400

To sum up, the potlatch character of the Prayaga meeting, the comingtogether of the feudatories at the end of the year and the assembly inspringtime (Caitra), the (possible) erection and climbing of a pillar, andHars.a’s reported masquerade as Indra—all these elements suggest thatHars.a’s charity arena was linked with a Spring Festival equal or similar tothe famous Indrotsava. The theatre of largesse may have had a counterpartin theatrical performances. The blending of both theatres offered KingHars.avardhana the perfect stage to show off his power, liberality, learning,literary and artistic qualities, and religious acumen; in other words, it wastotal theatre avant la lettre.

The end of an epochFor the period between Hars.a’s last quinquennial assembly in Spring ad643 and his death in ad 647 we have only the Chinese sources to go by.As we have seen above, the second Chinese embassy led by Li Yibiao wasin Magadha from 644 to 645. This mission had returned to China by late645 or early 646. It brought to China a sapling of the Bodhi Tree.401 Athird Tang embassy was sent to India in early 648, this time headed byWang Xuance, that is after Xuanzang had presented his travelogue to theemperor Taizong (ad 646). By the time it arrived in India, Hars.a haddied and his empire was in a state of disintegration. Devahuti has madean assessment of the Chinese sources. She gives the following rendering.

On arrival Wang Hsuan-ts‘e clashed with A-lo-na-shun, the ruler ofNa-fu (stated variously as Na-fu-ti or Ti-na-fu-ti, which we surmisestands for Tıra-bhukti on the northern borders of Bihar), who it isstated was a subordinate of Harsha and had usurped the throne. Thebattle of Cha-po-ho-lo (Champaran?) on the banks of the Chien-t‘o-wei (Gan.d. akı?) took place in the fifth month of the twenty-secondyear of Chen-kuan and lasted three days. A-lo-na-shun was takencaptive and carried off to China to be presented to the Emperor.402

400 Si-yu-ki I, 234: ‘The heretics who practise asceticism have raised a high columnin the middle of the river; when the sun is about to go down they immediatelyclimb up the pillar; then clinging on the pillar with one hand and one foot, theywonderfully hold themselves out with one foot and one arm; and so they keepthemselves streched out in the air with their eyes fixed on the sun, and their headsturning with it to the right as it sets. When the evening has darkened, they comedown. There are many dozens of ascetics who practise this rite.’

401 Sen 2003, 45.402 Devahuti 1998, 252f.; cf. 260f. Cf. Sen 2003, 23.

The Reign of Hars.avardhana 129

A-lo-na-shun is usually interpreted as Arun. aśva or Arun. aśa, evidently afeudatory ruling in the principality of Tirhut, who had seized the oppor-tunity presented by Hars.a’s death to declare his independence. Why heattacked the Chinese embassy is unknown; plunder seems to have beena likely motive.403 Wang Xuance, however, fled to Tibet (or Nepal?) andreturned with reinforcements from Tibet and Nepal and made Arun. aśvaa prisoner of war. No doubt the Chinese sources are greatly exaggeratingwhen they report that, in addition to the royal family, ‘[Wang Xuance]took captive 13,000 people, male and female, together with more than30,000 head of cattle and horses. Upon this India was overawed. Thecities, towns, and villages which submitted numbered more than 580.’404

The one most likely to have been overawed, though, is Hars.a’s oldfriend and ally Bhaskaravarman. Bereft of his patron, and with the neigh-bouring Tirhut kingdom overrun by Chinese and Tibetan invaders, theking of Kamarupa is reported to have sent to Wang Xuance ‘30,000 oxenand horses as provisions for the army and also bows, swords and spears,’405

in other words, he may have felt forced to surrender a contingent of troopsin order to avoid further trouble. Bhaskaravarman must have been anold man, and with him the ancient Varman dynasty of Kamarupa cameto an end. It appears that he was succeeded by Śalastambha (HCI III,141). However, not only the Tibetans may have contributed to this regimechange; the death of Hars.a offered new opportunities to the archenemiesof Assam, the rulers of Gaud. a.

The picture in the western part of Hars.a’s kingdom is less clear due toa lack of sources. It may a priori be assumed that the collapse of centralauthority offered new chances for local rulers to declare independence,leading to a further fragmentation of the empire. The Mauryas or MoriRajputs of Chitor may have temporarily taken control of Malwa.406 Fur-ther to the south-west, the Maitrakas were quick to take advantage ofthe situation. Dhruvasena II, who had still been guest of honour in Pra-yaga in 643, was succeeded by his son Dharasena IV (c. ad 645–50), thepresumed grandson of Hars.a, who also may have considered himself a le-gitimate heir, a dauhitra, since no other male offspring of Hars.a is known.He certainly assumed his grandfather’s titles, including Maharajadhirajaand Cakravartin, apparently even while Hars.a was still alive (see above,n. 340 on p. 112).

403 Devahuti 1998, 260; Sen 2003, 23f.404 The T‘ung-tien as rendered in Devahuti 1998, 242.405 The Hsin t‘ang-shu as rendered in Devahuti 1998, 247. According to Levi the

territory of the Tibetan king Srong-brtsan sgam-po comprised ‘le Tibet entier, lebassin du Koukou-nor, les marches occidentales de la Chine, l’Assam, et le Nepal’(Levi 1905 II, 148).

406 Jain 1972, 319; HCI III, 161f.

130 Hans Bakker

What was happening in the meantime in Kanauj is anybody’s guess.Four decades of historical research have not been able to make the follow-ing assessment by V.V. Mirashi obsolete: ‘The period of about sixty yearsfrom A.D. 647, the date of Harsha’s death, is regarded as the Dark Agein the history of North India; for we have no information about the rulersof Kanauj in this period.’407 To many in the capital it must have seemedas if the king’s own words had come true, spoken at the death of the heroJımutavahana in Hars.adeva’s Nagananda:

O son, the earth has become empty, now you have gone to theother world! Firmness has become without firm base, whom shouldDecency take as her refuge? Able to exercise Patience, who here?The pinacle of Generosity is no more. Truth, truly, has died. Whereshould poor Compassion go now?408

The aftermath: the rise of the Gupta kings of MagadhaIn our reconstruction of the history of northern India in the sixth andearly seventh centuries we have frequently encountered the dynasty of theLater Guptas. I have argued that this dynasty was driven out of Maga-dha by the Maukhari king Iśanavarman, but survived in Gaud. a proper.It joined Śaśanka’s attack on Kanauj, was beaten back again, and finallyconcluded a peace with Kanauj, when lg Madhavagupta ‘aligned himselfforever with the illustrious Hars.adeva’ (above, n. 237 on p. 82, p. 92). Thisreconstruction is based substantially on a single inscription, viz. the Aph-sad Stone Inscription of Adityasena, son of lg Madhavagupta, which tellsus the dynastic history of the Later Guptas. A donative inscription relat-ing to Nalanda provides us with the only date for Adityasena, Sam. vat 66,which, if it is taken to refer to the Hars.a era, as it usually is, would yieldthe date of ad 672.409

Either lg Madhavagupta or his son Aditya availed himself of the op-portunity created by the chaos in Hars.a’s former empire and the weak-ening of the archenemy, the Varman dynasty of Kamarupa, to establishthis Gupta dynasty in Magadha. Adityasena regarded himself as the truesuccessor of Hars.avardhana, with whom he may have been distantly re-lated (see Figures 7 and 8). In the bombastic language of the Aphsad

407 Mirashi 1974, 18. Cf. Mishra 1977, 36ff.408 Nagananda V, vs. 106 (metre: Śikharin. ı):

niradharam. dhairyam. kam iva śaran. am. yatu vinayah. ,ks.amah. ks. antim. vad. hum. ka iha virata danaparata |hatam. satyam. satyam. vrajatu ca kr.pa kvadya kr.pan. a,jagaj jatam. śunyam. tvayi tanaya lokantaragate ‖

409 CII III (1888), 208f.: Shahpur Stone Image Inscription of Adityasena.

The Reign of Hars.avardhana 131

inscription, composed by Suks.maśiva of Gaud. a,410 he boasted of havingobtained royal glory by his prowess on the battlefield.

Whose arms are swollen and twitch due to the splitting of the tem-ples of rutting elephants in war, whose halo of fame [. . . ] by thepower of the various enemies whom he had destroyed, whose feet arelike a wild and sparkling fire (as they are touched) by the crownedheads of all the kings that he had cast down, whose renowned andspotless glory is his pride in battle, (that) king is possessed of For-tune.411

The religious affiliation of Aditya is not expressed clearly in his own in-scriptions, but the Aphsad stone records the building of a temple (bha-vana) for Vis.n. u by the king, a mat.ha for the pious ones (dharmika) by hismother Śrımatı,412 and a pond (saras) by his wife Kon. adevı. The samequeen is reported in two rock inscriptions found on the Mandaragiri,413

to have constructed a lotus pond (pus.karin. ı). In these inscriptions, herhusband Aditya is given imperial titles for the first time, such as Parama-bhat.t.araka Maharajadhiraja.414 Finally, we know from the Deo-BaranarkInscription that his great-grandson Jıvitagupta II guaranteed the contin-uation of the worship of Surya in Deo-Baranark in west Bihar (see above,p. 67). In this inscription Adityasena is called a paramabhagavata. Despitethe fact that his court poet seems to have been an initiated Śaiva, tojudge by his name ending in ◦śiva, it is obvious that Aditya distancedhimself from the Saivism of his predecessors in this region and returnedto the faith of his illustrious namesakes, the Imperial Guptas. Aditya’ssuccessors, however, Devagupta and Vis.n. ugupta, are again called parama-maheśvaras.415

410 Above, n. 160 on p. 58. Cf. Fleet in CII III (1888), 202: ‘It offers about the earliestinstance of the hyperbolical expressions and mythological allusions with whichthe later inscriptions abound, distinguishing them so completely from the artistic,concise, dignified, and frequently really poetical, style of the more ancient records.’

411 CII III (1888), 204 ll. 22f. (Sircar SI II, 48):

mattagajendrakumbhadalanasphıtasphuraddoryugo,dhvastanekaripuprabhava ˘ ˘ yaśoman. d. alah. |nyastaśes.anarendramaulicaran. aspharapratapanalo,laks.mıvan samarabhimanavimalaprakhyatakırtir nr.pah. ‖

412 One wonders who are meant by dharmika. A plausible conjecture would identifythem as Buddhists. If, as we have suggested above (n. 237 on p. 82), the motherof Adityasena, Śrımatı, came from King Hars.a’s family, her Buddhist orientationcannot come as a surprise.

413 Modern Mandar Hill, 44 km south of Bhagalpur in east Bihar near the border ofJharkhand.

414 CII III (1888), 212; see also above, n. 159 on p. 58.415 CII III (1888), 215 ll. 2–3 (Sircar in SI II, 50).

132 Hans Bakker

This Devagupta had a sister who was married off to a descendant ofthe Maukharis, the family’s former enemies, namely to a certain Bhoga-varman.416 This is reported in the Paśupati Temple Inscription of Jaya-deva II of ad 732/33 (Am. śuvarman Sam. vat 157).417 A daughter born ofthis marriage, Vatsadevı, married the Licchavi king Śivadeva II of Ne-pal (see Figure 6), and she, as a widow, together with her reigning sonJayadeva II, recorded this parentage in the Paśupati Temple Inscription.

That King (Śivadeva II) married the illustrious Vatsadevı with duerespect, as if she were another Śrı; she, namely, was the daughter ofthe illustrious Bhogavarman, who by his fame had put hostile kingsto shame and who was the crest jewel of the illustrious Varmans, i.e.of the mighty Maukhari family, and she was the granddaughter ofthe great overlord of Magadha, Adityasena.418

The inscription, ipso facto, proves that the family of the Maukharis out-lived the reign of Hars.a, one way or the other. When Princess Vatsadevımarried her Nepalese husband somewhere in the ninth or last decade ofthe seventh century, the Magadha Country was ruled by her maternaluncle Devagupta and his son and successor Vis.n. ugupta, who was, unlikehis name suggests, but just like his father, a paramamaheśvara, as wasVatsadevı’s husband Śivadeva. This king is known from seven inscriptions,among which the Laganat.ola Inscription (ad 694/95), where he recordshis patronage of the acaryas of the Paśupata sect.419

416 This Bhogavarman cannot be the same as the son of Śurasena and nephew ofAm. śuvarman of Nepal, who is mentioned in the Deopat.ana Inscription, Sam. vat 39[ad 615]. Verma & Singh 1994, 140; Vajracarya 1973 (LA 85), p. 357, l. 9.

417 Buhler & Indraji 1880 read 153, Gnoli 1956 I, 115ff. and Sircar in SI II, 429 read159; Verma & Singh 1994, 241 (see ibid. p. xlix f.) and Vajracarya 1973, 552 read157. We follow Vajracarya.

418 Vajracarya 1973 (LA 148), p. 550, vs. 13 (Śardulavikrıd. ita):

devı bahubalad. hyamaukharikulaśrıvarmacud. aman. i-khyatihrepitavairibhupatigan. aśrıbhogavarmodbhava |dauhitrı magadhadhipasya mahatah. śryadityasenasya ya,vyud. ha śrır iva tena sa ks. itibhuja śrıvatsadevy adarat ‖ 13 ‖

419 Śivadeva handed over to them the village of Vaidya as an agrahara for themaintenance of the temple of Śivadeveśvara Bhat.t.araka, which the king hadbuilt himself: ayam. gramah. śarırakot.t.amaryado(papanna)ś cat.abhat.anam apra-veśyenacandrarkavanikaliko bhumicchidranyayenagraharataya matapitror atmanaśca vipulapun. yopacayahetor asmabhih. svakaritaśrıśivadeveśvaram. bhat.t.arakam. ni-mittıkr. tya taddevakulakhan. d. asphut.itasam. skarakaran. aya vaśapaśupatacaryebhyah.pratipaditas. Vajracarya 1973 (LA 139), p. 514, ll. 5–9; Verma & Singh 1994,no. 147; cf. Slusser 1998 I, 235. The phrase vaśapaśupata◦ is unclear. DiwakarAcharya reports to me in an e-mail dated 9-8-12: ‘The reading of vaśapaśupatalooks not so certain. va and śa are clear, but there is a slanted line on śa. If it isnot an accidental scratch, the letter comes closer to śe. Before va there is a thick

The Reign of Hars.avardhana 133

This ties in with the Mangraon Stone Inscription of Vis.n. ugupta’sTime, Year 17, in which we are told that the pilgrim Avimuktajja (Skt.Avimuktarya) hailing from Kut.ukka Country, who had visited numer-ous Śaiva holy places inhabited by Siddhas (anekaśivasiddhayatana◦), wasdevoted to ‘his own doctrine’ (svasiddhantabhirata).420 As we have ar-gued elsewhere, the term svasiddhanta, which in the Skandapuran. a (SPS

167.129–30) refers explicitly to the Pancartha doctrine of Lagud. i, mayhave been a code word meaning ‘His own doctrine,’ an implied claim thatthe Pancartha doctrine of this Paśupata lineage was the only true one, re-vealed directly by the Lord himself.421 Avimuktajja may thus have been aPaśupata ascetic following the Pancartha. His gift of lamp oil was made tothe Subhadreśvara Temple, evidently in the village of Mangraon.422 Thefact that the ascetic’s history and his purchase/donation were recorded ina stone tablet by a composer and scribe points to the prominent positionhe must have held within this community. In this capacity Avimukta-jja travelled around the Magadha of Vis.n. ugupta, and possibly beyond.It seems likely that among the ‘Śaiva holy places’ he mentions, the eightplaces in the country of Magadha were, those where, according to an earlyinterpolation in the Skandapuran. a, Lagud. ıśvara (Lakulıśa) roamed aboutwith his pupils.423 Avimukajja is also likely to have visited the eponymousAvimukteśvara Temple in Avimukta, i.e. Varan. ası, and it is a distinct pos-sibility that he went to Nepala to see Paśupatinatha, another sanctuarymentioned in the same interpolation.424

The Skandapuran. a reaches NepalAs far as our limited epigraphical sources allow us to draw general con-clusions, the evidence they provide seems to prove that, after the collapseof Hars.a’s empire, Magadha was the first country in which stability wasrestored under the authority of the Later Guptas. An alliance with whatremained of the former mighty Maukharis may have contributed to this.While the kingdoms of Kanauj, Paun.d. ra and Kamarupa were still in dis-

angular shape but it does not look like a letter, neither aligned to the headmark norto va.’ In a subsequent mail (13-8-12) Acharya adds: ‘I think it is possible to read◦karan. ayavam. śapaśupatacaryebhyah. .’ This would make perfect sense and seemsconfirmed by a similar qualification of early Paśupata ascetics as ‘being withoutfamily,’ which is found in a graffito on Mount Kalanjara: nirva(m. )śalakud. ah. (seebelow, p. 147).

420 EI XXVI, 246; Thaplyal 1985, 168.421 Bakker 2007, 6 and 2011, 32. See below, p. 145.422 EI XXVI, 215. Mangraon is a village 23 km south-west of Buxar.423 SPS 167.169: magadhasu smr.tany as.t.au sthanani śaśimaulinah. | śis.yaih. parivr. to

yani babhrama lagud. ıśvarah. | tani dr.s. t.va bhavet sadyah. puman papavivarjitah. ‖169 ‖ (Bisschop 2006, 108). See below, p. 137.

424 SPS 167.186–187; see Bisschop 2006, 15–17, 222f. Below, p. 137.

134 Hans Bakker

array, Magadha could consolidate its power by re-establishing its goodrelations with the Licchavi kingdom of Nepal. The marriage of Vatsadevıshows that these relations must have been very close, particularly duringthe last two decades of the seventh century, when her uncle and cousin,Devagupta and Vis.n. ugupta, ruled in Magadha and her husband ŚivadevaII in Nepal. They resulted in intensive cultural exchange, of which thePaśupata religion may have formed a substantial part. We will see below(p. 138) that it was probably during this period that the hyparchetype ofthe S recension of the Skandapuran. a came into being, including the above-mentioned interpolation, and that an early version thereof designated asa, the ancestor of our MS S1, reached Nepal.

II

The World of Culture

The Skandapuran. aits composition and position

within the literary and religious landscape

The Paśupata Movement and the Skandapuran. a

d Benares d

The Composition and Transmission of the Skandapuran.a

The historical time frame

In her Introduction to the Skandapuran. a edition Volume III, Yuko Yokochigives a brilliant philological analysis of the evolution and transmission ofthe text, from its inception to the point when a major second redactioncreated the new RA recension, hyparchetype of the R and A recensions asattested by our manuscript evidence. She presents a plausible chronologi-cal framework in which this evolution is supposed to have taken place, butrefrains from giving a date for the second redaction. We take Yokochi’sframework as our point of departure, but in order to integrate the historyof the text into the history of North India, we must be less careful than sheis and accentuate a somewhat more precise time frame. This time frame isa working hypothesis which, in our present state of knowledge, cannot becorroborated by philological means but must prove itself in the historicalreconstruction.

1 We presume that the inception of this Puran. a occurred during thereign of the Maukhari kings Śarvavarman or Avantivarman of Kanaujin the second half of the sixth century (see above, p. 74), and that theclosing of this composing activity, that is to say the first redaction, wasaccomplished during the reign of King Hars.avardhana of Kanauj. Thecomposition of a text of such length and complexity was a process thatmay easily have taken a man’s lifetime, or one or two generations of activework within a team of pandits. A period of ad 570 to 620 therefore seemsreasonable.

2 As argued by Yokochi, it must have taken at least half a centuryfor the original composition to develop into the hyparchetype of the Srecension. Apart from small alterations and corruptions, one of the majorchanges to the original composition was the inclusion in this hyparche-type of the long interpolation of SPS 167.163–87. In this section the eightsanctuaries connected with the perigrinations of Lagud. ıśvara (Lakulıśa)

137

138 Hans Bakker

and his pupils in Magadha, the temple of Prahasiteśvara in Pat.alıputra(SPS 167.168–69), and the temple of Paśupati in Nepala (SPS 167.186) arethe most conspicuous additions to the stock of holy places known to theSkandapuran. a. This indicates that our text was transmitted in Paśupatacircles with close links to Magadha.

Either immediately or after a short period, the S hyparchetype came toexist in at least two slightly different versions, one of which is called a byYokochi. This a version was contained in a manuscript that was broughtto Nepal and became the ancestor of MS S1. In Yokochi’s frame this musthave happened before ad 700, and we hypothesize that this transmissiontook place when the Later Guptas of Magadha had restored their kingdomand their good relationship with the Licchavi kingdom of Nepal betweenad 670 and 700 (see above, p. 133).

3 The other version of the S hyparchetype evolved further in NorthIndia into what is called the b version, and at least one instance of this lineof transmission was brought to Nepal, the ancestor of our MSS S2, S3/S4.Yokochi argues that this b version may be a century or more later thana. It is therefore assumed that this ancestor b entered Nepal between ad770–800, that is, when the Pala king Dharmapala controlled the greaterpart of Eastern India (see p. 260). This is the same king who is reportedto have performed holy rites at Kedara, Gokarn. a, and the mouth of theGanga.425

4 The pandit or pandits who made the new redaction, the ancestorof the RA recension, made use, according to Yokochi, of exemplars ofboth, the a and b versions, and in addition seem to have had access toone or more other lines of early transmission. Therefore this redactionmost likely took place after b had come into existence. It is best dated tothe 9th century.426 In view of its later transmission and survival, thisredaction was made in Eastern India, either in Bihar (Magadha) or Bengal(Gaud. a), that is, within the Pala kingdom of the ninth century.

Varan. ası

On the basis of the above time frame we can say that the composition ofthe Skandapuran. a was in full swing by the turn of the seventh century.But where did this activity take place? In the Introduction to Volume IIa

425 Mungir Plate of Devapaladeva, EI XVIII, 305 vs. 7; Petech 1984, 29. Little histori-cal value can be attached to this attribution, since the inscription seems to describethe traditional digvijaya rather than historical events, mentioning the extremitiesin the north-west, south-west and east.

426 This is a slight modification of the date proposed for this redaction in the Intro-duction to SP II A, p. 54, where we argued that it ‘may have taken place in theeighth, possibly first half of the ninth century.’

The Paśupata Movement and the Skandapuran. a 139

we argued that ‘it has a certain probability per se that the Skandapura-n. a was composed either in Varan. ası, or in a (Paśupata) centre that hadclose contacts with this city’ (SP II A, 52 n. 174). Since this was written,our joint efforts have not yielded any evidence that could negate thisstatement or question it, and no alternative has emerged. In about thistime Benares was firmly included in the kingdom of Kanauj, the politicalcentre where, according to our text, a lineage of the Pancartha Paśupatashad established itself that traced its origins back to an ascetic from Kuru-ks.etra through the guru Raśıkara; it may have been the fountainhead ofthe lineage of Paśupatas whose initiation names ended in Raśi.427 Theseascetics apparently referred to their doctrine as the Svasiddhanta (seebelow, p. 145).

Since the publication of SP II A, a number of studies have appearedthat deepen our understanding of the early Paśupata movement. It seemstherefore to be opportune to reconsider the position of our text withinthis movement, taking as our starting point Varan. ası in around 600 ad.

By this time the city had developed into a holy place revolving aroundits major sanctuary, the Avimukteśvara Temple. The fame of this templehad risen in the preceding century thanks to the settlement of holy men ingreat numbers, the majority of whom belonged to the Paśupata movement.This historical process is described in the Skandapuran. a itself, naturallycouched in a mythic vocabulary. After Śiva had established himself in theholy city the following is said to have happened:

Having thus spoken, Mahadeva looks around in all directions. Thenthat land all of a sudden becomes as if it were on fire, while lightningflashes and thunder roars; it becomes filled with lustres, resembling(the lustre) of the sun and fire in one—as if a million suns are rising.Thereupon numerous Paśupata masters (siddhas) suddenly arrive—shining whitely as their bodies are smeared with ashes, lords of theyogins (yogıśvara)—and also appear those who follow (Vedic) sacri-ficial observances, and those whose sectarian marks are not visible(avyaktalingin), blazing in their union with Śiva (śivayoga)—andthey make obeisances to Maheśvara.428

Significant in this passage is the lumping together of, 1) common ash-smeared Paśupata ascetics, 2) orthodox brahmins who maintain the sac-rificial fires (vaitanikavrata) and who consequently belong to the class ofhouseholders, and 3) advanced sadhakas, no longer recognizable by sectar-ian signs.429 The Skandapuran. a thus presents the Paśupata as a composite

427 See below, n. 445 on p. 145 and p. 175.428 Synopsis of SP 29.60–63.429 This may refer to the second stage of the Pancartha practice (see Acharya 2011,

460).

140 Hans Bakker

movement and may itself be considered as an exponent of this broad reli-gious community of Śaivas in Benares. That Paśupata Saivism in ad 600comprised more than just ascetics pursuing the Pancartha (as taught inthe Paśupatasutras and Kaun. d. inya’s commentary thereon) seems also tofollow from the sectarian history related in SPS 167.110–38, which directlyprecedes another eulogy of Varan. ası (SPS 167.139–62).

The Paśupata history as told in SPS 167

This history makes a distinction between, on the one hand, four pupilsof Bhagavat, Kauśika etc., and, on the other, a brahmin named Soma-śarman, a descendant of Atri, to whose house in Karohan. a (Karvan) thesame Bhagavat is said to have resorted after he had assumed a white (i.e.ash-smeared) body (rupam. kr. tva sitangam. ), that is, after he had adoptedthe appearance of a Paśupata ascetic. SPS 167.125–26 twice explicitlysays that this embodiment of the Lord initiated Somaśarman along withhis family (sakulam), something not said of Kauśika and the other threepupils.430 This feature, which has so far gone unnoticed, may be signifi-cant, since it indicates that the Paśupata branch of which Somaśarmanwas believed to be the fountainhead was conceived of as including morethan only male ascetics (sadhakas) in pursuit of the aims of the Panca-rtha.431 The Bhagavat is said to have bestowed favour upon the members

430 Cf. the Lakud. as, who are without family (below, p. 147).431 Somaśarman is mentioned once more in SPBh 131.2, namely as a gan. anayaka,

along with Din. d. i and Bharabhuti, who in later literature are considered to beearlier incarnations of Śiva (cf. below, p. 214 and Bisschop 2006, 44). The nameSomaśarman is also known from the Vayu◦ and Lingapuran. as and he is evidentlythe same as the Someśa named in the Jayadrathayamala 4.453 as Lakulıśa’s pre-decessor: someśo lakulıśaś ca hy as.t.avim. śaty amı śivah. | vyakhyananugrahakarah.praman. ajnanabhedatah. ‖4.453 ‖ (edition by Sanderson based on a MS in the Na-tional Archives in Kathmandu, Catalogue Vol. 5, MS No. 4650). VaPVenk 1.23.202–05 (= VaPAA 23.214–17 ≈ LiP 1.24.120cd–124ab) describes him as the 27th incar-nation of Śiva, born in Prabhasatırtha, a holy place in Saurashtra near the ArabianSea, famous for its Somanatha Temple. Bisschop 2006, 43 argues that this chapterof the VaP postdates our Skandapuran. a (see below, p. 214).The Somnath Paśupata tradition is recorded in the Somnathpattan Praśasti ofBhava Br.haspati, dated Valabhı Sam. vat 850 = c. ad 1169 (Ozha 1889; SP II A,62 n. 191). In this inscription the holy site is said, in a mysterious Śardulavikrıd. itaverse, to have been given by Soma, who was purified by Śiva, to the Paśupataswho in the Kr.tayuga had gone to Vr.ndavana due to a curse of Gaurı, after hehad established his own (i.e. Soma’s) path (paddhati) by order of Śiva: somah.so ’stu jayı smarangadahano yam. nirma[la]m. nirmame, gauryah. śapa[balena vaikr. t]ayuge vr.ndava[n]opeyus. am | pradat paśupataryasadhusudhiyam. yah. sthanam. ,kr. tva svam atha paddhatim. śa[śibhr. ]to devasya tasyajnaya ‖ 3 ‖ (cf. MBh 9.34.38–81, 12.329.45f., SPBh 70.20–22; Davis 1994). The inscription records that the tem-ple was, after its destruction, restored by a Paśupata saint, born in Varan. ası as anincarnation of Nandıśvara, by the name of Bhava Br.haspati, who belonged to the

The Paśupata Movement and the Skandapuran. a 141

of the House of Atri by granting yogasiddhi, ‘perfection in yoga.’ Thisseems to be a state of perfection that is brought about along lines some-what different from the path of the Pancartha ascetic, who is said to reachthe ‘end of suffering’ (duh. khanta) in the cremation ground after a life ofhardship, a path that amounts to complete union with Śiva (śivayoga) atthe moment of death.

Yogasiddhi is a perfection attained in so-called Maheśvara or Paśupa-ta yoga, which Skandapuran. a 174–182 defines as knowledge (jnana).432 Itis a yoga structure overlying the Sam. khya philosophical framework thatleads towards yogipratyaks.a, a system in which the subtle channels callednad. ıs play an important role.433

After one has adopted a sitting posture and withdrawn all limbs,one should become motionless and meditate, while one directs one’sthought on the twenty-sixth reality. Gradually one should move thewind/breath in the trachea through the channels. When one seesthe disc of the moon appear in the heart(-lotus), one should remainconcentrated, being without emotions due to the establishment ofdeep meditation. Seeing the light fixed in the body, one realizesyogic mastery. Then, due to continuously exercising in this way andthe realization of perfection in yoga, yogic perception sets in, i.e.knowledge of the past, present and future.434

The attainment of a divine body and ultimately the realisation of completesovereignty (aiśvarya) enables the practitioner to commit yogic suicide (ut-kranti), and to merge with the 26th reality, viz. God.435 This path does

vam. śa of Gargya. It remains unclear whether this foundation myth has anything todo with Somaśarman (cf. the Stone Inscription at Veraval , Peterson 1895, 208ff.).

432 SPBh 174.2 and 4cd–5ab: yan mam. pr.cchasi taj jnanam. yogam. maheśvaram. param |sarvasam. yogavidyanam. yat tattvam. parikırtyate ‖ ‘What you have asked me, thatis knowledge, viz., the Maheśvara yoga (which is) supreme, because it is said to bethe true reality of all yoga doctrines.’

433 This system with a serpent-shaped power in the nad. ıs in three states (jagrat,svapna, sus.upta) is described in SPBh 181.

434 SPBh 179.28–31:ekam asanam asthaya s.ad. vim. śam. samanusmaran |sarvangani samahr. tya nis.pando dhyanam acaret ‖ 28 ‖śanaih. sam. carayed vayum. nad. ıbhih. srotasi sthitam |tavat tis. t.heta yuktatma dhyanayogad gataspr.hah. ‖ 29 ‖somaman. d. alasam. kaśam. yavat paśyed dhr.di sthitam |jyotir dehaśritam. dr.s. t.va yogaiśvaryam. pravartate ‖ 30 ‖tata evam. sadabhyasad yogasiddhipravartanat |trikalavis.ayam. jnanam. pratyaks.am. yogino bhavet ‖ 31 ‖

435 SPBh 182.10–11:śr.n. u vyasa param. jnanam ıśvaradhanakaran. am |svecchaya yena mucyante yogino jnanatatparah. ‖ 10 ‖

142 Hans Bakker

not seem intrinsically different from other, theistic yoga schools of thetime.436 This perfection was reached by Parvatı on the Śrıparvata after agreat effort (tapas).437 It was granted by Śiva and made her omniscient(pararthajna) and equal to the accomplished yogis (yogasiddhas) living onthat mountain (SPBh 70.54), who are described as follows:

The excellent men who worship me here, O Goddess, before theygive up their lives, these are great men, who are steeped in Paśu-pata yoga and have attained perfection therein; they have reachedcomplete mastery over their bodies and are moving about to what-ever destination/rebirth they will, (and eventually) they shall goto a state of highest perfection of complete freedom, the supremestation.438

Somaśarman and his family were thus blessed by the Lord with this per-fection in yoga. The importance of the figure of Somaśarman in the reli-gious history of the time may also be deduced from an early attestationin two pre-Angkorean inscriptions. The inscriptions are not easy to in-terpret but they may be referring to our Somaśarman. The first speaksabout the installation of a linga called śrıtribhuvaneśvaram, which wasśrısomaśarmmarkayutam. Dominic Goodall, who kindly drew our atten-tion to these inscriptions, takes the latter to mean ‘together with a statueof Śrı-Somaśarman.’439 Statues of Somaśarman have not come to light, orhave not been recognised as such yet.

tridhotkrantim imam. jnatva muhur abhyasya caiva hi |svadeham. ripuvat tyaktva s.ad. vim. śe layam apnuyat ‖ 11 ‖

436 Oberhammer 1977, 57f.: ‘Demgegenuber (i.e. the ‘sam. khyistische Meditation’)scheint der theistische Yoga, wie er fur die religiosen Traditionen des Hinduis-mus faßbar wird (dies gilt jedenfalls fur das Paśupata- und Pancaratra-System,vermutlich aber ganz allgemein), selbst dort, wo er in der Meditation eine gewissestrukturelle Ubereinstimmung mit dem sam. khyistischen Meditationsweg zeigt, inder grundsatzlichen Orientierung und Gestimmtheit einen ganz anderen Typus derMeditation zu verwirklichen.’

437 SPBh 70.51–54. Cf. MBh 3.83.16f. This Śrıparvata or Śrıgiri is an old Paśupataayatana, identified with Śrıśaila in the Kurnool District of Andhra Pradesh (seeSP I, 107; Bisschop 2006, 201). Its Mahatmya is told in SPBh 70.42–59. SP 25.22tells us that he who throws himself off the cliff (bhr.gu) will become a Gan. apa.

438 SPBh 70.48–49:yogasiddha mahatmano yoge paśupate sthitah. |ya iharadhya mam. devi jahuh. pran. an narottamah. ‖ 48 ‖ta ete siddhadeharthah. svacchandagaticarin. ah. |moks.asiddheh. param. nis. t.ham. gantarah. paramam. padam ‖ 49 ‖

Cf. the Paśupatasutras in which, after having summed up the powers obtained bythe yogin (PS 1.21–37), the descriptions ends with: ity etair gun. air yukto bhagavatomahadevasya mahagan. apatir bhavati ‖ 1.38 ‖

439 K.359 (Cœdes 1937–66) of c. ad 600:śrıvıravarmmaduhita svasa śrıbhavavarmman. ah. |

The Paśupata Movement and the Skandapuran. a 143

The Junvanı Copperplate Inscription of MahaśivaguptaWe have another, Indian source regarding this mysterious figure of Soma-śarman, viz. the Junvanı (Malhar) Copperplate Inscription of Mahaśiva-gupta, Year 57. This Pan. d. ava king of Sirpur (Daks.in. a Kosala) reignedfrom about ad 590–650, so the inscription dates from c. ad 647. The textof the inscription has been discussed and re-edited by myself in Bakker2000.440 Contrary to what I said in the latter article, however, I now thinkthat the two sources, the Skandapuran. a and the inscription, are not con-tradictory, though in the former text the Lord is said to resort to, in thelatter to be born into the house of Somaśarman.441 To underscore thisstatement I here give the relevant text again as well as my new transla-tion/interpretation.442

pativrata dharmmarata dvitıyarundhatıva ya ‖hiran. yavarmmajananım. yas tam patnım upavahat |dvijendur akr. tisvamı samavedavidagran. ıh. ‖śrısomaśarmmarkayutam. sa śrıtribhuvaneśvaram |atis. t.hipan mahapujam atipus.kaladaks. in. am ‖‘She was daughter of Vıravarman, sister of Bhavavarman, devoted to her husband,devoted to duty, like a second Arundhatı, whom the moon among brahmins,Akr.tisvamin, foremost among those knowing the Samaveda, married, she themother of Hiran. yavarma.He installed [here the linga called] Śrı-Tribhuvaneśvara, together with a statue(arka) of Śrı-Somaśarman, along with elaborate worship and extremely generousbenefactions.’ (Translation by Dominic Goodall)

The second inscription K.54/K.55, [629 ad] (Cœdes 1937–66 III, pp. 157–163,Inscriptions de Kdei An) runs [personal communication Dominic Goodall]:

acaryavidyavinayahvayena maya punah. sam. skr. tam atra bhaktya |samastadayasthiram astu sarvalokaikanathasya śivasya lingam ‖ 1 ‖khapancendriyage śake rohin. yam. śaśini sthite | [550 śaka = 629 ad]śivalingam. tada tena devah. sam. skriyate punah. ‖ 2 ‖sarvasvam. bharyaya sardham. yajnadattasya bhojakah. |śivadattad avapyaitat śivalingaya dattavan ‖ 3 ‖ [Khmer portion omitted]nanatarugan. akırn. am. devayatanam ıdr.śam |kr. tam. namabhavat tena rudraśrama iti smr.tam ‖ 4 ‖punah. sam. skr. tya tenaiva śrımadamratakeśvare |yojitaśes.avibhavam. śivalingadvayam. kr. tam ‖ 5 ‖śomaśarmajat.alingam. hariś caite tatha ˘ |tes. am. tena ca dattam. yo devasvam. hartum icchati ‖ 6 ‖

440 My edition of 2000 was based on Rayakwar and Singh 1994 and a (draft of a)new edition by A.M. Shastri, which was published in 2001. Since then the originaledition of Rayakwar & Singh has been republished in Rayakwar & Singh 2005,188–192. The inscription was again edited by Susmita Majumdar in 2007.

441 Torzsok 2011, 356ff. follows my earlier interpretation.442 I now think that the conjectured haplography and emendation proposed by H.

Isaacson (mahavrate [te]na) is not warranted. The phrase as it stands probablymeans that the incarnation followed the mahavrata and as such became the firstconsecrated Paśupata, whose white appearance (cf. above, p. 140) elicited the titlejagadindu. This tallies with the moon symbolism of the param. para (parallel to the

144 Hans Bakker

adhuna kalikalam asadya,śrımal-lakulıśanatho ’vatırya,somaśarma-khyabrahman. akule bhutva,mahavratena dıks. ito jagadindus,tenapi mugalisas,tatah. somadiparam. paryakramen. a,sthanaguruśrı-rudrasoma–praśis.yaśrı-tejasoma-śis.yebhyah.śrımad-bhımasoma-padebhyah. [. . . ]

Now the Kali age has come and Lakulıśanatha has descended; Hewas born in the family of a brahmin named Somaśarman. Conse-crated by the Great Vow (Mahavrata), He (i.e. Lakulıśanatha) be-came a ‘Moon on Earth’ (Jagadindu). By Him again Mugalisa [wasinitiated]. Then, in due succession of the lineage that started withthe Moon, the aforementioned Bhımasoma—the pupil of the illustri-ous Tejasoma and grand-pupil of the illustrious Sthanaguru Rudra-soma— etc.

There are good reasons for identifying Mugalisa with Musala, which wouldmake the lineage of Bhımasoma that of the Mausala.443 As I have arguedin Bakker 2000, the doctrine of this branch came to be referred to as theSoma Siddhanta.

Bhımasoma was in charge of the protection of the tapovana attached tothe Baleśvara-bhat.t.araka Temple in Sirpur,444 and the inscription furtherattests that these Mausala Paśupatas received the grant for worship cere-monies (yaga), initiations (dıks. a), teaching (vyakhyana), housing (vasati)of pupils and grand-pupils, and for repairs to the temple. In other words,these Paśupatas acted as acaryas and temple priests. Some of them mayhave been ascetics, but the performance of ceremonies (yaga), for instance,will have been done by brahmins who maintained the sacrificial fires, the

dynastic pedigree of the Pan.d. ava/Soma-vam. śa, to which King Mahaśivagupta sayshe belongs: Bakker 2000, 16): Somaśarman, the ‘Shelter of the Moon’ is known fromSPS 167.124 to have belonged to the ‘House of Atri’ (atrivam. śa), whose son wasSoma (MBh 7.119.4: atreh. putro ’bhavat somah. ). Lakulıśanatha is thus conceivedas another Moon, to whom the ◦ soma ascetics traced their pedigree.I accept Majumdar’s conjecture to read sthanaguru◦ (instead of sthane guru◦), inthe sense of guru presiding over the foundation at issue (Majumdar 2007, 286,292). The qualification sthanaguru may be taken to apply to every member of theparam. para (for sthanaguru, ‘guru in charge,’ also see Bosma 2013, 250, 252).

443 Bakker 2000; Acharya 2011, 462.444 Bakker 2000, 8: śrıbaleśvarabhat.t.arakatapovanapratipalanartham aropitebhyah. . In

this office he seems to have succeeded the Siddhantin Astraśiva who had becomehead of the Baleśvara Temple complex, unless we assume that the Śaiva Siddha-ntins and the Soma Siddhantins each had their own tapovana connected to theBaleśvara Temple (Bosma 2013, 252).

The Paśupata Movement and the Skandapuran. a 145

vaitanikavrata of the Skandapuran. a. We think that the Paśupatas repre-sented by our Skandapuran. a may have included similar groups, in additionto sadhakas who traced their guru-param. para back to Lagud. i through oneof his four pupils. Something more needs to be said about this latter tra-dition.

The Lagud. i traditionAfter initiating Somaśarman and his family, the Lord’s embodiment is saidto have gone on foot to Ujjayanı (Ujjain), where he settled in the localcremation ground and initiated his first student, Kauśika. It is generallyaccepted that this Kauśika or Kuśika refers to a historical person. Hisactivities are attested in the Mathura Pillar Inscription and Kaun.d. inya’sPancarthabhas.ya and he may have founded the Paśupata movement aswe know it in the second century ad. By the time the Skandapuran. a wascomposed, Kauśika was considered as the first of four students altogether.The other three, Gargya, Mitra, and ‘a man born in the Kuru Country’(Kuruks.etra), were initiated in Jambumarga (near Pus.kara), Mathura,and Kanyakubja respectively.445

In this context, and in this context only (apart from the interpolationin SPS 167), the Lord’s embodiment is named Lagud. i in the Skandapura-n. a. He is said to have pronouced to these four students ‘His own doctrine,’

445 See also below, p. 170. Some information about these four lineages seems to havesurvived the ages. The (Jaina) tradition, preserved in Gun. aratna’s (fourteenthcentury) commentary on the S. ad. darśanasamuccaya of Haribhadra (ninth century)and Rajaśekharasuri’s S. ad. darśanasamuccaya (fourteenth century), records the firstfour names of the teachers in each of the four lineages that are said to derive fromLakulıśa. The fourth teacher, here named Kaurus.a, was succeeded by Manus.yaka,Pus.paka and Raśıkara. Gun. aratna ad Haribhadra’s S. ad. darśanasamuccaya (Dalal1966, 28):

tasya cas. t.adaśavatara amı | nakulıśo ’tha,[1.1] kauśikah. , [1.2] gargyah. , [1.3] maitryah. , [1.4] kaurus.ah. ,[2.1] ıśanah. , [2.2] paragargyah. , [2.3] kapilan. d. ah. , [2.4] man. us.yakah. ,[3.1] kuśikah. , [3.2] atrih. , [3.3] pingalah. , [3.4] pus.pakah. ,[4.1] br.hadaryah. , [4.2] agastih. , [4.3] santanah. , [4.4] raśıkarah. ,vidyaguruś ca |

Cf. Rajaśekharasuri’s S. ad. darśanasamuccaya, ibid. p. 35. For an elaborate discus-sion of this param. para see Bisschop 2006, 48ff.The commentator of the Paśupatasutra, Kaun. d. inya, apparently placed himself inthe lineage of Kuśika and Iśana (Kaun.d. inya ad PS 4.10: kuśikeśanasambandha).No doubt the lineage of Kuśika is the oldest Paśupata tradition (Indraji 1881–82, 322f., 327; Sircar: SI I, 278), a fact corroborated by the Skandapuran. a itself:ujjayanyam. gurujyes.t.hah. kauśiko nama namatah. | (SPS 167.122ab). The lineage ofRaśıkara went back to Nakulıśa/Lakulıśa through Kaurus.a, i.e. the ‘man born inthe Kuru Country,’ who was initiated in Kanyakubja. The Paśupata ascetics whosename ended in ◦raśi may have originated from this branch. See Bakker 2007, 4f.

146 Hans Bakker

Svasiddhanta, which is explicitly said to be the Pancartha (the doctrineexpounded in Kaun. d. inya’s commentary), and which is evidently to bedistinguished from what was imparted to Somaśarman. The Skandapura-n. a describes the four disciples, after they had been initiated (anugr.hya),as apostles when it says:

Lagud. i granted union (yoga) according to His own teaching (Sva-siddhanta) and spoke:

This is the final mystical teaching known as Pancartha. It hasbeen proclaimed to you in order that you liberate the brahminsfrom the fetters of death. You should make the brahmins reachthe highest station by initiating them (in this teaching). Yourdwelling places shall be on sacred river banks, in holy sanctu-aries, as well as in deserted houses and forests, excluded fromsociety.

Having heard this word of Lord Paśupati Himself, all these fourdisciples, being enlightened by God, did as they were told.446

As proven by the Junvanı Inscription and in conformity with the accountof the Skandapuran. a, however, the followers of Somaśarman also recog-nized Lakulıśa/Lagud. i as the divine initiator of their tradition.

To sum up, all our sources acknowledge the existence in around ad 600of a major division in the Paśupata movement: 1) one practising andpreaching the orthodox Pancartha, and 2) a (socially) more comprehensiveone, the origin of which was traced back, in retrospect, by one lineage atleast, to Somaśarman and Musala. Kaun. d. inya wrote for the first branch,at an earlier stage, when the ramifications of four param. paras and thefigure of ‘Lagud. i’ as such had apparently not yet been conceived; we maytentatively date him therefore to the fourth century.447

It would seem to me that both divisions may have originated anddeveloped side by side from the second century onwards and distinctionsbetween them may often have become blurred. The earliest testimony tothe Pancartha branch, if we leave the graffito on Pad. ana Hill readingkosikaya aside,448 is the already mentioned Mathura Pillar Inscription ofad 380/381, which refers to a group of acaryas whose names end in ◦vimalaand who traced their param. para back to Kuśika. Though we assume thatthis may refer to the first pupil who was initiated in the Pancartha inUjjain, the gurus of the lineage specified in the inscription were in charge

446 SPS 167.129cd–132. See Bisschop 2006, 104.447 Cf. Acharya 2011, s.v. Paśupatas p. 459.448 For this see Indraji 1881–82, 322, Bakker 1991, 23, and Bisschop 2006, 47.

The Paśupata Movement and the Skandapuran. a 147

of a shrine (gurvayatana) in which memorial (kırti) lingas were installed,and worshippers thereof were addressed as Maheśvaras.449

From the fifth century onwards we come across images of a teacher car-rying a lagud. a or club, sometimes surrounded by two, later by four disciples(see Bakker 2011). Some of these followers of the ‘club-bearer’ Lagud. in/La-gud. i were called Lagud. as or Lakulas, and their earliest epigraphical attes-tation is a recently discovered fifth-century graffito on Mount Kalanjarareading: ‘The Lakud. as, who are without family, coming from the neigh-bourhood of Vidiśa.’450

If we leave the graffiti found on Pad. ana Hill reading sadhamusala (i.e.siddhamusala) and musaladatta aside again,451 the earliest attestation ofthe other, comprehensive, tradition is the corpus of inscriptions found inBagh, which features Paśupatas such as the acarya Bhagavat Lokoda-dhi,452 who are said ‘to possess rights to enjoy, cultivate, and inhabit thetemple lands granted.’453

Diversity and coherence of the Paśupata movementInitiation (dıks. a) definitely applied to all Paśupata ascetics, and all group-ings may have included teachers, acaryas, as well as supporting laymen.The differentiation between the communities was instead brought about bythe practice of various initiations (and post-initiatory observances) and therecipients thereof; in other words, the differentiation was achieved withinthe broad spectrum of Maheśvaras that existed between strict Pancarthasadhakas and the community of ordinary devotees, the laukikas.454 To allappearances the Lagud. as were the most orthodox, but they also includedteachers who abstained from the ascetic path, as is attested in a passage ofthe 10th-century Gan. akarikaratnat. ıka, which I quote here in Sanderson’stranslation.

449 SI I, 278: naitat khyatyartham abhilekhyate | [atha] maheśvaran. am. vijnaptih. kriyatesambodhanam. ca |

450 Below p. 196. See above, n. 419 on p. 133 for the qualification of Paśupata asceticsas ‘without family.’

451 Indraji 1881–82, 323–25; Bisschop 2006, 47.452 Bagh grant No. 10, issued in year 56 (ad 376) by Maharaja Bhulun. d. a (Ramesh &

Tewari 1990, 22).453 Sanderson 2013, 225. Ramesh & Tewari 1990, Inscriptions Nos. 3, 5, 10, 12, and

13. See also below, p. 171.454 A Paśupata manual (Paddhati) dealing with the Sam. skaravidhi, ‘Transformatory

Rite,’ has recently been discovered and published by Acharya 2007. It is one of fourViddhi texts which are ascribed to Gargya and dated by Acharya ‘between Kau-n. d. inya and Bhasarvajna,’ i.e. 4th and 10th centuries (Acharya 2011, 459). It addssubstantially to the initiation ceremony of the Pancartha tradition as described inKaun. d. inya’s commentary on PS 1.1 (see Bakker 2004b, 124 n. 28) and is thereforemost likely to be assigned to the comprehensive strand of the Paśupatas, probablypost ad 600.

148 Hans Bakker

O Lord, is the observance of all injunctions of the Pancartha theonly means of attaining the end of suffering? No, that is not the onlymeans. It is also possible for a person to attain it even though hedoes not have the capacity to put all those injunctions into practice,if [as a holder of the office of Acarya] he properly favours [throughinitiation and the rest] such outstanding brahmins as approach himas candidates. Why? because he is [thereby] guarding the tradition.For by doing so he enables many who seek to attain the end ofsuffering through the power of that tradition to achieve their goal.By this means he accumulates merit that will bestow infinite reward.It is through this [merit] that he will attain union [with Rudra] andthence, through [Rudra’s] favour, the end of suffering.455

The two divisions had much in common. They shared, in addition to atext collection that would develop into the Paśupatasutras as it existstoday, a group of texts called ‘Praman. as,’456 with which the Lakulas inparticular are associated (to distinguish them from the older Pancarthikatradition represented by Kaun.d. inya). The increasing differentiation of thebranches may have induced the Mausala to add their own Praman. as tothis corpus. The early Śaiva-siddhanta texts, such as the Niśvasa corpus,may be descended from the latter. After all, the main, immediate rivalsof Bhımasoma of Sirpur and his Mausala group, vying for MahaśivaguptaBalarjuna’s favours, were the Śaiva-siddhanta priests.457

Moreover, within both varieties, it seems, an iconic form of Rudrabecame venerated in the fifth and sixth centuries: the ithyphallic figurewith the club; ‘Mausala’ and ‘Lagud. a’ derive from musalin and lagud. in,both meaning ‘armed with a club (musala/lagud. a).’ Although this icon,generally referred to as Lakulıśa, was probably emblematic for all Paśu-patas of the sixth century, the ‘Somaśarman branch’ may have added tothis a distinctive icon of its own, which has, admittedly, not so far beenrecognised as such.

In the course of the seventh century, the Mausala or Somas mayhave been eclipsed partly (the domesticated wing) by the emerging ŚaivaSiddhanta and may have become partly closely associated with a group

455 Sanderson 2013, 233: translation of Gan. akarikaratnat. ıka p. 2 ll. 7–12,kim. nu bhagavan pancarthoktasamastaniyoganupalanad eva duh. khantah. pra-pyata iti | ucyate | na kevalam. tatah. kim. tu samastaniyoganus.t.hanaśakti-vikalenapi brahman. aviśes. an. am. śis.yatvenopagatanam. samyaganugrahakaran. ad apiduh. khantah. prapyate | kasmat | sam. pradayaraks.an. at | sam. pradayam. palayata hitatsam. pradayasamarthyena duh. khantam. gamis.yatam. bahunam. api duh. khantah.sam. padito bhavati | tato ’nantaphalapun. yopacayah. | tato yogapraptau prasadad duh. -khanta iti |

456 Sanderson 2006, 171, 173, 176–77; Acharya 2011, 459, 462.457 Bakker 2000, 15f.

The Paśupata Movement and the Skandapuran. a 149

that produced the non-Saiddhantika texts such as the Jayadrathayamala.The latter, incidentally, is the only extant text known to us that recog-nises the doctrine of sixty-six Bhavas or Rudras mentioned in the JunvanıInscription.458 They became associated with a type of ascetic called Kapa-likas, ‘skull-bearers,’ with whom the Soma Siddhanta came to be identifiedin later sources.459

The Śr.nkhalika Paśupatas are sometimes seen as representatives ofthese skull-bearers on account of their name, śr. nkhalika, which is usuallytaken to mean ‘one who wears a chaplet’ (scil. of skulls).460 We meetthem for the first time in a seventh-century Licchavi inscription from Ne-pal, dating from the reign of Jis.n. ugupta (624–632 ad). This inscripton isfound in the Paśupati temple, within the shrine of Chatracan. d. eśvara, ona pillar next to the image. It records a donation by a (Paśupata) teacher(acarya) called Bhagavat Pranardanapran. a Kauśika.461 Pranardana (‘theRoarer’) informs us in another inscription on the pedestal of the sameimage that he is living outside the varn. aśrama confines of society.462 Thedonation concerns a few fields for the sake of maintaining the shrine ofChatracan.d. eśvara and a water conduit in some village and is made tosome ascetics of the mun. d. aśr. nkhalikapaśupatacaryapars.ad . The Śr.nkhali-ka Paśupatas are met again as donees in an inscription of Narendradeva,Year 69 (ad 645).463

The lineage name of the donor Pranardanapran. a ‘Kauśika,’ seems toindicate a Pancartha affiliation, and such an affiliation is apparently con-firmed by the Indragarh Stone Inscription of vs 767 (ad 710/11) foundin the Mandasor District, which records the erection of a Śiva temple,Guheśvara, by a Paśupata named Danaraśi, who is said to be a discipleof the Paśupata acarya and foremost of the Rudra–Śr.nkhalikas, Vinıta-raśi.464 The Raśi branch adhered to the Pancartha (above, p. 139), andthe Rudra–Śr.nkhalikas may have done so too.465

458 Bakker 2000, 11; see below, p. 214.459 Sanderson 1988, 668.460 E.g. Buhler & Indraji in IA 9 (1880), 174 n. 39: ‘mun. d. aśr. nkhalika literally “wearing

a chain of skulls” is the name of a sub-division of the Paśupatas.’ In view ofthe other sub-division of Rudra-Śr.nkhalikas (below), we prefer to interpret thiscompound as a Karmadharaya, saying that these Śr.nkhalikas shaved their heads,(mun. d. a), instead of having matted hair (see below, n. 469 on p. 150).

461 Vajracarya 1973, LA 112 (= Verma & Singh 1994, no. 119).462 Vajracarya 1973, LA 113 (= Verma & Singh 1994, no. 120): varn. aśramodvasita.463 Vajracarya 1973, LA 125 (= Verma & Singh 1994, no. 132)). Diwakar Acharya

informed me (e-mail d.d. 13-8-2012) that we should read danam. in ll. 3 and 4instead of dana◦. The gift was made to these Śr.nkhalika Paśupatas in order tocare for the sick: danam. śr. nkhalikapaśupatanam. glanabhais.ajyartham. dattam. (LA125, ll. 3f.).

464 EI XXXII, 116 vss. 5–7.465 The designation of a Raśi ascetic as ‘Rudra’ occurs also in an inscription of the

150 Hans Bakker

However, the complexity of the Paśupata tradition is illustrated by thesame Paśupati Chatracan. d. eśvara inscription, which tells us that withinthe assembly (pars.ad) of the donees—the Mun.d. a–Śr.nkhalika Paśupatas—acaryas figure whose names end in ◦soma—Umasoma [sic], [Ca]krasoma—and who are qualified as khad. ukas, i.e. priests (?).466 Were the Mun. d. a–Śr.-nkhalikas ‘Somas,’ the Rudra–Śr.nkhalikas ‘Raśis,’ and Pranardana himselfa ‘Pran. a,’ i.e. initiated in the lineage of Kuśika (Kauśika)?467

Whatever may have been the case, the seventh and eighth centuriesmay have seen a tendency in many Śaiva ascetic lineages to wear a skull-chaplet, thus making the term ‘Kapalika’ a rather generic one.468 Theywere classified as lokatıtavratins or Mahapaśupatas in the Mukha sectionof the Niśvasatattvasam. hita.469

It is good to keep in mind that the above is, without any doubt, a muchsimplified picture of a very complicated historical reality. Many lineagesand branches may have existed that we do not know at all or of whichwe have no more information than just their names, such as the Karukas.This complex reality may explain why we find all sorts of, sometimes ide-ologically motivated, classifications in seventh-century and later sources.

time of King Mahaśivagupta Balarjuna (ad 590–650), found in the Gandharveśatemple in Sirpur (Śrıpura). Here a teacher, Udbhavaraśi Rudra, is called ‘an oceanfor the streams of his own doctrine’ (svasiddhantadhunıpayo(dhi); EI XXXIX, 151,vs. 3; see also Bakker 2007, 5).

466 Vajracarya 1973, LA 112, ll. 16–17. Buhler & Indraji 1880, 174 n. 38: ‘Khad. d. ukaoccurs on other inscriptions as a name of certain priests of Śiva.’

467 An initiation name ‘Pran. a’ of ascetics belonging to the Kuśika/Kauśika lineage isapparently not attested elsewhere, so it might not be an initiation name at all.Buhler & Indraji 1880, 174 n. 37 interpret this name as follows: ‘The meaning ofthis name seems to be equivalent to the modern Śivaprasada or Śivalala, and tomean “dear as life of the roarer” (Rudra).’

468 See for an early example the iconography of Andhaka that we find in the HarahaStone Inscription of Iśanavarman vs. 2 (EI XIV, 115; see above, p. 70 and p. 14).Torzsok 2011, 355: ‘Given the usages, one could distinguish between two religiousmeanings of Kapalika. First, in a sticter sense, it denotes a particular Śaiva order,closely related to the Lakulas and the Paśupatas. Second, in a wider meaning, itrefers to a (usually Śakta) Tantric practitioner who adopts the observance andpossibly other practices of the original Kapalikas.’

469 Mukha 4.88f. (cf. Acharya 2011, 461):

alabdhah. pancabhir guhyair dıks. itaś caiva so bhramet |khat.vangı ca kapalı ca sa jat.ı mun. d. a-m-eva va ‖ 88 ‖valayajnopavıtı ca śiromun. d. aiś ca man. d. itah. |kaupınavaso bhasmangı divyabharan. a[bhus. itah. ] ‖ 89 ‖

The five guhyas are the five Brahmamantras. These Lokatıtavratins are calledMahapaśupatas in Mukha 4.128. This Mukha section is later than the Mula, theoldest part of the Niśvasatattvasam. hita. The latter has been dated very early byGoodall & Isaacson 2007, 6, viz. c. ad 450–550, on insufficient grounds, in my view(above, n. 36 on p. 9).

The Paśupata Movement and the Skandapuran. a 151

My methodological strategy has been to take the sources dating from be-fore or about ad 600 as the point of departure, since these are documentsthat belong to the Paśupata movement itself.

This brings us back to the Skandapuran. a. The question regarding thegrouping within the Paśupata or Atimarga of which this text may beconsidered to be an exponent cannot, in the present state of our knowledge,be answered with more precision than has been attempted above. The textin its first redaction does not yet seem familiar with the Śaiva Siddhantaor with the Kapalikas as a distinct class of its own, irrespective of the factthat the Kapala myth plays a foremost role in this Puran. a (SP 5–7).

The concluding chapter of the Skandapuran. a (SPBh 183) describes thehighest attainable goal of the layman devotee (laukika), viz. the world ofsakala Iśvara, the Śivapura.470 The devotee reaches there when he has seenthe twelve major holy lingas and dies at one of these places.471 But alsothose who observe the Paśupata vrata and the devotees of Bhava are saidto reach it after death (SPBh 183.50cd–52ab).

The composition, in short, may have been in the hands of a mixedgroup of learned, laukika pandits and initiated acaryas headed by a mainredactor, a team in which the whole gamut of views of the Paśupatamovement in the Varan. ası of the time was aknowledged. It could be soliberal because it had no particular sectarian agenda, but aimed to provideall the Maheśvaras of its age, in particular the uninitiated laity (lauki-kas), with an exoteric, mythological account of the cosmos as created andgoverned by Mahadeva.472 What counted more than sectarian partisanshipwas staunch Śaiva faith and a sound knowledge of Sanskrit and the Epicand Puranic traditions.

The Kapalikas of Magadha

As has been discussed above, there are indications that the hyparchetypeof the S recension came into being in Paśupata circles with close links toMagadha in the last three decades of the seventh century. By this timeKapalika ascetics were commonplace.473 In this respect it seems significant

470 See Bisschop 2007; below, p. 175.471 SPBh 183.47–50: Kedara, Madhyama, Mahalaya, Catuh. śr.ngeśvara (location un-

known), Śrıgiri (i.e. Śrıparvata), Karohan. a, Gangadvara, Bhadreśa (= Bhadreśva-ra?), Śankukarn. eśvara (situated at the mouth of the Indus River, SPBh 73.6; above,n. 362 on p. 117), Mahakaleśvara (i.e. Ujjain), both Gokarn. as, and Avimukteśvara(i.e. Varan. ası). For this and similar lists see Bisschop 2006, 15f.

472 Cf. Sanderson 2009, 52.473 Lorenzen 1991, 14.

152 Hans Bakker

that the only (indirect) reference to the Kapalikas in our text is possiblyfound in the interpolation that distinguishes the S hyparchetype fromthe first redaction, and that in connection with a famous, old holy site inMagadha, the Gr.dhrakut.a, Vulture Peak. SPS 167.166–167 runs as follows:

In Gaya there is another holy place of the bountiful Lord; it iscalled Gr.dhrakut.eśvara and it sets free from all evil. The excellentmen who have seen him (the Lord of Gr.dhrakut.a), they will not errat the time of death. A brahmin who smears himself with ashes isliberated there from all evil, and also he will gain the fruits of theTwelve-years vow (dvadaśabdavrata).

The Vayupuran. a mentions the sanctuary of Gr.dhreśvara on the Gr.dhra-kut.a as well as a Gr.dhravat.a there.474 The Mahabharata connectsthis Gr.dhravat.a with a bath in ashes and the Twelve-years vow.475

The Twelve-years vow is traditionally the chief penance for killing abrahmin,476 and a vow derived from it became specific to the Kapalika,who re-enacted Śiva’s behaviour (see below, p. 166).

A Paśupata adaptation of the traditional expiation rite is easily ef-fected by substituting ashes for water in the thrice-daily bath and ashesfor grass to sleep on in items 2 and 4 of this observance, as it is describedin the Vis.n. usmr.ti 50.1–15, which I follow here.477 This adaptation be-came part of the Mahavrata of the Kapalikas, who are generally calledMahavratins. They made the stick with a skull attached, the khat.vanga—mandatory in the expiation rite478—emblematic for their sect. The rela-tively late Vis.n. usmr.ti, although it describes the Dharmaśastra vow andnot the Paśupata version, seems tacitly to acknowledge this semantic shiftwhen it identifies this observance, if it is performed for twelve years to ex-piate the killing of a brahmin, as the ‘Mahavrata’: etan mahavratam ‖5 ‖brahman. am. hatva dvadaśasam. vatsaram. kuryat ‖ 6 ‖ 479

474 VaPAA 108.62–63 (= VaPVenk 2.46.65–66):ato girir gr.dhrakut.as tatra gr.dhreśvarah. sthitah. |dr.s. t.va gr.dhreśvaram. natva yayac chambhoh. padam. narah. ‖62 ‖tatra gr.dhre guhayam. ca pin. d. adah. śivalokabhak |tatra gr.dhre vat.am. natva praptakamo divam. vrajet ‖63 ‖

475 MBh 3.82.79–80: ato gr.dhravat.am. gacchet sthanam. devasya dhımatah. | snayıtabhasmana tatra abhigamya vr.s.adhvajam ‖79 ‖ brahman. ena bhavec cırn. am. vratam.dvadaśavars. ikam | itares. am. tu varn. anam. sarvapapam. pran. aśyati ‖80 ‖ Cf. PdPSvargakhan. d. a 38.11–12. See also Bisschop 2006, 218.

476 For a description see Kane IV, 88–91.477 Compare ViS 50.2 tris.avan. am. snayat with PS 1.2 bhasmana tris.avan. am. snayıta

and ViS 50.4 trr.n. aśayı ca syat with PS 1.3 bhasmani śayıta. See Lorenzen 1991,73–82.

478 ViS 50.15: sarves.u śavaśirodhvajı syat.479 ViS 50.5–6. Cf. Lorenzen 1991, 74, 76f.

The Paśupata Movement and the Skandapuran. a 153

The Gr.dhrakut.a, in view of its old connection with the Twelve-yearsvow, may thus have developed into a place that attracted Kapalika Paśu-patas and evidently was important enough to be included in the additionallist of holy places in the S recension, promising the brahmin ‘the merit ofthe twelve-years observance,’ whatever this may have been believed to be.

The MahavrataA final word may be neeeded concerning the term mahavrata. We haveseen above (p. 143) that in the Junvanı Inscription Lakulıśanatha wasconceived of as having adopted the appearance of a consecrated Paśu-pata by following the observance called Mahavrata, thanks to which hebecame a ‘Moon on Earth.’ This seems in accordance with the definitionof mahavrata that is given in Skandapuran. a 180:

After having gone through the Paśupata initiation, leaving alldharma (prescriptions) behind, they attain to Maheśvara thanks tothe power of the Transcendent One. The observance taught by Śivamerely consists of taking baths in ashes; it leads to omniscience andbeyond asceticism and dharma. Those, whether they are wicked orlawful, who are devoted to the Lord entirely, they reach liberationby practising it: know it as the Mahavrata.480

The Mahavrata is orginally a Vedic ritual connected with the vratyasand it may or may not have been a prototype underlying ascetic prac-tices.481 The Paśupata came to call their observance the ‘great vow,’ asthe Skandapuran. a attests, and so may other religious orders have done.Eventually the name came to be reserved especially for the dvadaśabda-vrata and the derivative vow, which had now lost its expiatory character.Though an early Paśupata adaptation of the expiation rite as describedabove is likely in view of the importance attached to the Kapala myth,we cannot conclude that wherever we find Paśupata and Mahavrata to-gether we are dealing with Kapalikas. The habit of carrying and utilizinghuman skulls in ritual may have developed gradually within the Paśupatafold, until finally a distinct group became recognized as Mahavratins orKapalikas.

480 SPBh 180.9–11 (emended)sarvadharmam. bahis.kr. tva nis.kalasya prabhavatah. |dıks. am. paśupatım. prapya aviśanti maheśvaram ‖ 9 ‖sarvajnanasamayuktam. tapodharmabahis.kr. tam |bhasmasadhanamatram. tu śivena kathitam. vratam ‖ 9 ‖papis. t.ha dharmavanto va bhaktas kevalam ıśvaram |sevanad yasya mucyante tan nibodha mahavratam ‖ 10 ‖

The text continues by equating this Mahavrata with the Paśupata vrata.481 Rolland 1973; Falk 1986; Parpola 1992. See Lorenzen 1991, 74.

The Skandapuran. a and Ban. a’s Hars.acarita

d The Sarasvatı, Kuruks.etra and Thanesar d

Preamble

Assuming that the time frame of the Skandapuran. a presented above iscorrect, we can observe that the composers of the Puran. a and the author ofthe Hars.acarita, Ban. a, were actors in and witnesses to the same historicalworld in northern India in around ad 600. This historical reality becamesublimated in the two works in two very different ways: on the one hand itis lifted up to the realm of the Śaiva mythological universe in the versified,pseudohistoric style of the Puran. a, in which historical details are eschewedas far as possible, and, on the other hand, transmuted into a historicalnovel in the kavya prose style, in which historical reality is glorified andits timeless essence is expressed through symbolism and literary metaphor.Moreover, the focus of each of the texts is intrinsically different: the Puran. aintends to sanctify geographical locations by associating them with theworld of Śaiva mythology, thus giving them a transcendent quality; thekavya aestheticizes the life and deeds of its historical protagonist, andthus transcends the contingency and idiosyncrasy of the human condition.The common ground of the two texts is found in their nature: they areboth idealized, comprehensive, literary imaginations, products of the samecultural soil, which explains the occurrence of interface.

According to Ban. a’s own account, the recitation of part of his chef-d’œvre, Hars.a’s deeds, had been prompted by a bard (bandin) with thefollowing two Arya verses, which were sung in accompaniment to a Puran. arecitation (see below):

Methinks that this Puran. a does not differ from the deeds of Hars.a(hars.acarita) after all! That too is sung by a sage and broad beyondboundaries/surpasses king Pr.thu, that too encompasses the world,and that too is holy/Wind-born(pavana).

This song is like the reign of joy/Hars.a itself: it issues from an excel-lent voice/Śrıkan. t.ha (Śiva), it follows the flute accompaniment/the

155

156 Hans Bakker

dynasty, it is without disharmony/undisputed, it articulates rhyth-mically/expands widely, and it is loaded with adoration of Bhara-ta’s/India’s path.482

For us, studying the interface of the two genres will enhance our under-standing of the literary genesis of both texts.

The Puran.a in Performance

Ban. a’s familiarity with the Puran. a literature of his times is expressedin a very interesting passage in the Hars.acarita, from which the aboveverses have been taken. At the beginning of Chapter 3 the poet tells usthat he went back to his home village, the one where Sarasvata and Vatsahad grown up together, Prıtikut.a on the River Śon. a (modern Son), wherehe heard a Puran. a recitation by the reader (pustakavacaka) Sudr.s.t.i—dressed in white silk made in Paun.d. ra (Bengal), his forehead marked by atilaka consisting of lines of orpiment on a white clay coating (tırthamr.dagorocanaya ca racitatilakah. ), his topknot ornamented by a small bunchof flowers, his lips reddened by betel, and his eyes beautified by linesof collyrium.483 He seated himself on a chair and then began his perfor-mance. Because of the rarity of such information, the passage deserves tobe quoted in full.

He pauses for a moment before he places, on a desk made of reedstalks that is put in front of him, a codex (pustaka), which, althoughits wrapping has been removed by that time, is still wrapped, as itwere, in the halo of his nails which shine softly like the fibres of alotus. Then, while he assigns two places behind him to two ‘flautists,’Madhukara (‘the bee’) and Paravata (‘the turtle-dove’), his closeassociates,484 he turns over the leaf marking the end of the chapterread in the morning session, takes a small bundle of some folios,and recites the ‘Puran. a spoken by the Wind.’ By his chanting heenchants the hearts of his audience with sweet intonations, (evoking)as it were, the tinkling of the anklets of Sarasvatı, as she presentsherself in his mouth, while it seems as if, by the sparkling of his teeth,

482 HC 3 vs. 3 (p. 140f. [39]):tad api munigıtam atipr. thu, tad api jagadvyapi pavanam. tad api |hars.acaritad abhinnam. , pratibhati hi me puran. am idam ‖ 3.3 ‖vam. śanugam avivadi sphut.akaran. am. bharatamargabhajanaguru |śrıkan. t.haviniryatam. gıtam idam. hars.arajyam iva ‖ 3.4 ‖

483 HC 3 p. 139 [39].484 Mentioned also in Ucchvasa 1 of HC (p. 67 [19]).

The Skandapuran. a and Ban. a’s Hars.acarita 157

he whitewashes the ink-stained syllables and worships the book withshowers of white flowers.485

The passage confirms that the Puran. a was composed in writing on palm-leaves and recited during a series of performances by professional singers.These performances were evidently open to a general public in villagesthroughout the country, in contrast to the kavya, which may have beenmainly recited to the educated audiences of select assemblies or the royalcourt.

The ‘Puran. a spoken by the Wind’ is usually taken to be a reference tothe Vayupuran. a.486 As we have argued in the Prolegomena to SP I (pp.20–22), the composers of the Skandapuran. a were familiar with a Puran. atext whose main contents are represented by the Pancalaks.an. a and whichis expounded by Vayu; in other words, which may have been an earlyform of the Vayupuran. a. There can be no doubt that, at the time Ban. awas writing, the Vayupuran. a was a well-known specimen of the Puran. agenre and this makes the standard interpretation perfectly plausible. Ban. amentions that Sarasvatı made her appearance in the recitation, which is,first and foremost, a literary metaphor for the learning that the singing ofthe Puran. a makes manifest.

However, it should be noted that the Skandapuran. a uses a similarmetaphor when, immediately after the reported recitation of the Puran. a-pancalaks.an. a by Vayu (SP 5.5–9), it says that the Goddess Sarasvatıappears in the middle of the seers, who all take a bath in her. This Sara-svatı not only personifies learning, it is also the river of that name, whosedivine origin is told in Mahatmya style in SP 7.4–14. The holy river isclosely associated with the Brahmasaras and the spot where Śiva installedthe skull (kapala) of Brahma. We will come back to this when we deal withKuruks.etra. What is important in the present context, though, is that thechapters covering the appearance of Sarasvatı in our text (SP 5–7) arealso spoken by Vayu, at the request of the seers.487

485 HC 3 p. 139f. [39]: sthitva ca muhurtam iva, tatkalapanıtasutraves.t.anam api nakha-kiran. air mr.dumr.n. alasutrair ivaves.t.itam. pustakam. puronihitaśaraśalakayantrakenidhaya, pr.s. t.hatah. sanıd. asam. nivis. t.abhyam. madhukaraparavatabhyam. vam. śika-bhyam. datte sthanake, prabhatikaprapat.hakacchedacihnıkr. tam antaram. pattramutks. ipya, gr. hıtva ca katipayapattralaghvım. kapat.ikam, ks. alayann iva mas. ımali-nany aks.aran. i dantakantibhih. arcayann iva sitakusumamuktibhir grantham. , mu-khasam. nihitasarasvatınupuraravair iva gamakair madhurair aks. ipan manam. si śro-tr.n. am. gıtya, pavamanaproktam. puran. am. papat.ha |

486 HC 3 p. 140 [39]: pavamanaproktam. puran. am. is glossed by Śam. kara (ad loc.): pava-mano vayuh. ; pavanam. tad api in pada b of v. 3: pavamanam. (read: pavanam. )vayuproktam api.

487 See SP I, pp. 132–149.

158 Hans Bakker

The location of the village of Prıtikut.a, Ban. a’s native village wherethe Puran. a recitation is said to have taken place, has been discussed inPart I (see n. 256 on p. 88); it lay in the homeland of the Maukharis.

In the early morning of the day following the Puran. a recitation, whenthe Sam. dhya rite had been performed, Ban. a, surrounded by his kinsmen,began to tell the story of Hars.a’s deeds’ (hars.acarita),488 actually his‘early’ deeds, leading up to the moment when he definitely establishedhis royal power, expressed metaphorically by his union with Rajyaśrı. Thecomposition is then already in chapter 3 of the Hars.acarita. In the secondchapter Ban. a gave us an account of his own first, not entirely satisfying,visit to the king’s headquarters (skandhavara) near the river Ajiravatı,which we have described above (see p. 95). Soon thereafter, however, hewas invited back to court where ‘the king, becoming most friendly afterjust a few days, in his loving kindness, bestowed upon him the greatesthonours and affection, and made him share entirely in his confidence,wealth, pleasures, and majesty.’489

Ban.a’s Family Tree and the Foundation of Sthaneśvara

The Myth of DadhıcaThe recitation of the Puran. a the previous day had evoked the goddess oflearning, Sarasvatı, or, in the world of the Puran. a, the Sarasvatı River,which must have touched the right chord with the audience, not onlybecause this consisted of learned brahmins, but because Ban. a and hiskinsmen, living alongside the Śon. a River, through their historic relation-ship with the goddess Sarasvatı and Dadhıca, had ties with the Sarasvatıside, the riverbanks on which King Hars.a had grown up. Ban. a works outthis intricate relationship in the first chapter of his book, and it providesus with another instance of significant interface between the Hars.acaritaand the Skandapuran. a.

In this first chapter Ban. a tells the story of his own descent from thesage Bhr.gu through the latter’s son Cyavana, who is not only a forefatherin his own family tree, but also the father of Dadhıca, who, as mythologyhas it, fell in love with the goddess Sarasvatı, when she was living in exileon the left bank of the River Son and he was searching for his father Cya-vana, whose hermitage was across the river, on its eastern bank. The sonof this (divine) couple, Sarasvata, became the foster brother (frere de lait)

488 HC 3 p. 153 [42].489 HC 2 p. 134 [37]: aviśac ca punar api narapatibhavanam | svalpair eva cahobhih. pa-

ramaprıtena prasadajanmano manasya premn. o visrambhasya dravin. asya narma-n. ah. prabhavasya ca param. kot.im anıyata narendren. a ‖

The Skandapuran. a and Ban. a’s Hars.acarita 159

of Vatsa, the other grandson of Cyavana,490 who became the father of theVatsyayanas from which the author of the Hars.acarita stemmed, in hisown account (see Figure 10).

Figure 10Ban. a’s Pedigree

490 I take the expression bhargavavam. śasam. bhutasya bhratur brahman. asya jayamaks.amalam (HC 1 p. 62 [17]) to mean that Aks.amala was the wife of a brahminof the Bhr.gu family who was the brother (bhratr.), viz. of Dadhıca. Kane ad loc.thinks it might have been his cousin. Who will tell?

160 Hans Bakker

When he embroidered on the story of Dadhıca’s mother Sukanya, told inMahabharata 3.121–25, Ban. a and his audience may have been thinkingof the mythology that attributed the foundation of Hars.a’s native cityThanesar to her son Dadhıca. This mythology is told in the Skandapura-n. a.

The Skandapuran. a tells us about the origin of Sthaneśvara in a Stha-neśvaramahatmya (SP 31.48-115), which is intricately positioned betweenthe prelude to the famous Daks.a myth (SP 31.36–47) and this myth itself(SP 32). The Mahatmya provides the cause (karan. a) of Śiva’s war andvictory over all the gods in the Daks.a myth,491 viz. the boon that is askedby the staunch Śaiva brahmin Dadhıca after his victory over the Vais.n. a-va king Ks.upa, to the effect that Śiva may be victorious over all godsincluding Vis.n. u.492

The Śaiva sage Dadhıca, whose aśrama is said to be on the SarasvatıRiver,493 had entered into a hot dispute with the Vais.n. ava king Ks.u-pa about the superiority of either the brahman or the ks.atra principle,which dispute turned into a fight in which Dadhıca eventually defeatedhis Vais.n. ava rival with Śiva’s help. To commemorate this victory, Śivaallowed the site (sthana) where the fight between Dadhıca and Ks.upa tookplace and ceased (sthita), to become known under the name of ‘Sthane-śvara’; it is described as one krośa in circumference, full of flowers (pus.pa)and creepers.494 This is evidently the foundation myth of Sthaneśvara,i.e. Thanesar. Dadhıca’s own aśrama is here called ‘Sthan.utırtha,’ wherethe famous Sthan.uvat.a is said to stand,495 both already known from theMahabharata.496 Sthan.utırtha was thought to have been established by

491 SP 31.4, 31.101–02.492 SP 31.101: bhagavan yadi tus.t.o ’si yadi deyo varaś ca me | icchami vis.n. una sardham.

sarvan devam. s tvaya jitan ‖101 ‖493 MBh 3.81.162–64 mentions a Dadhıcatırtha where one attains the Sarasvatı state

(sarasvatım. gatim) and where the ascetic Angiras Sarasvata lives; it is near theKanyaśrama and the Sam. nihitı, where Brahma etc. assemble every month (MBh3.81.165f.).

494 SP 31.105f.: yasmat sthitam idam. vairam. varadanat tava prabho | iha tasmat tavasthanam. namnaitena bhavatv aja ‖ 105 ‖ deva uvaca | sthaneśvaram iti khyatam.namnaitat sthanam uttamam | bhavitr. krośaparyantam. nanapus.palatakulam ‖ 106 ‖

495 SP 31.109–10:sthan. utırtham. ca bhavitr. tavaiva papanaśanam |aśvamedhaphalam. hy atra snatah. prapnoti pus.kalam ‖ 109 ‖ayam. capi vat.ah. śrıman sthito ’ham. yatra sam. pratam |varam. datum. madakhyato namna sthan. uvat.o mahan |bhavis.yati na sam. dehah. phalam. casyapi me śr.n. u ‖ 110 ‖

496 The fight (vaira) between Dadhıca and Ks.upa resembles in many respects thatbetween Vasis.t.ha and Viśvamitra, whose aśramas are also said to be in Sthan.u-tırtha (MBh 9.41.4).

The Skandapuran. a and Ban. a’s Hars.acarita 161

Lord Sthan.u (MBh 9.41.6) and is the place credited in the Mahabharatawith the birth of Skanda and his consecration (abhis.eka) as senapati (seebelow, p. 162). The Mahatmya in the Skandapuran. a thus reformulates thesignificance of Sthan.utırtha and adds to its glory the site of Sthaneśvara,whose foundation is thus ascribed to Dadhıca, who figures prominently inBan. a’s ancestry.

This great, divine and holy site was established by Dadhıca. It iscalled Sthaneśvara and has become famous throughout the ThreeWorlds.497

This mythology, though it remained untold in the Hars.acarita, must, aswe suggested above, have been known to Ban. a, his audience, and KingHars.a, and it may actually have been the very reason why the authorgave the entirely mythological figure of Dadhıca such a prominent role inthe first chapter of his history and why he linked his own pedigree to him.

The evocation of Sarasvatı in the Puran. a recitation thus served as asort of hub, a key passage in the middle of the composition, in whichBan. a’s family history (chapter 1), its link with the goddess of learning,evokes the eponymous river,498 on whose banks one of his own Bhargavaancestors had founded the native city of the hero of the Hars.acarita. Hisstory is told in the subsequent ‘exhalations’ (Ucchvasas 3–8), beginningon the day following the Puran. a recitation, in Sarasvatı’s place of exile,the banks of the river Śon. a.

The Abhis.eka of Skanda and Hars.avardhana

We shall now return to the original home of the Sarasvatı, the river run-ning through Kuruks.etra past the capital of the Pus.yabhutis, on whoseauspicious banks the god Skanda as well as the mortal Hars.avardhana

497 SPS 167.81: dadhıcena mahad divyam pun. yam ayatanam. kr. tam | sthaneśvaramiti khyatam. lokes.u tris.u viśrutam ‖ 81 ‖ Compare the version in SPRA 167.4.10,20: tapah. ks.etre kuruks.etre dharmaks.etre sanatane | dadhıcena mahad divyam.pun. yam ayatanam. kr. tam | dadhıcasyalayah. khyatah. sarvapapaharah. parah. ‖ 10 ‖[. . . ] dadhıcena yatas tatra kr. tam ayatanam. śubham | sthaneśvaram iti khyatam.tena lokes.u tris.v api ‖ 20 ‖

498 SP 7.7–8 gives a nirukti of the name Sarasvatı. Śiva Nılalohita pronounces a sacredutterance (brahman), viz. the syllable OM, which makes the goddess Vac manifest.Because this goddess resides in his mouth, which is the brahmasaras, therefore thisgoddess will be known to the world as Sarasvatı. Springing from the brahma-saras/Brahmasaras she will purify all the worlds with her waters. The descent ofthis celestial river from the Brahmaloka to the earth is said by Vayu to be due tothe great tapas of the seers, and is said to have taken place in order to accomplishtheir perfection (siddhi) (SP 5.17–19).

162 Hans Bakker

were installed as commanders of the army (senapati). The Skandapuran. adescribes the first mythological event, the Hars.acarita the second, histor-ical episode, but the two descriptions have much in common.

After Hars.a had received the news of the killing of his brother and hadsworn revenge (see above, p. 86), the Hars.acarita tells us that he bathedin golden and silver vessels, offered worship to Nılalohita, performed afire oblation, and bestowed cows and golden and silver vessels filled withjewels on the brahmins. Seated on an auspicious throne covered by a tigerskin, he anointed his bow and his body with sandalwood paste, put ona pair of silken garments, placed a chaplet of white flowers on his head,stuck a blade of young kuśa grass behind his ear, and wound an amuletribbon around his arm together with the royal armlet. Then the chaplain(purohita) sprinkled his head with drops of lustral (śanti) water.499 Hars.amakes large donations and the ceremony is concluded when a tremblingin his right arm augurs well for a conquest of the eighteen continents.500

Installed as the commander-in-chief of the army,

Like the Golden Foetus from the Cosmic Egg as it were, (Hars.a)issued forth from the palace to prepare the Golden Age, amid cheersof ‘victory’ rising from his excited subjects, (propelled) by the to-tality of auspicious omens pushing forward, like zealous servants intheir officiousness.501

When Ban. a, through the prism of his literary imagination, thus describesthe inauguration of Prince Hars.a as the leader of the army—a prince whowas, we should not forget, at that very moment not yet crowned—theauthor and his readers cannot but have been fully aware of the significancethat traditional lore lent to this occasion. And, assuming that Ban. a’sliterary fiction reflects a historical event, we may add that the historicalactors too—ready to march against an enemy who was named Śaśanka or‘Moon,’ whose recent successes threatened the world order—would haverealized its full import and experienced a similar sensation.

The template of a prince (kumara) installed to lead the army in aperilous situation evokes the mythic archetype of Skanda, god of war,consecrated in order to destroy the asura Taraka or ‘Star’—an archetypethat is virtually reified when the installation (abhis.eka) of the historical

499 HC 7 p. 351 [53]: paripujitaprahr.s. t.apurohitakaraprakıryaman. aśantisalilaśıkaranika-rabhyuks. itaśirah. . For the integration of astrological śanti (appeasement) rituals inthe consecration see Geslani 2012.

500 HC 7 p. 351f. [53]: niyujya tatkalasmaran. asphuran. ena kathitatmanam iva cas. t.ada-śadvıpajetavyadhikare daks. in. am. bhujastambham.

501 HC 7 p. 352 [53]: ahamahamikaya sevakair iva sunimittair api samagrair agratobhavadbhih. pramuditaprajajanyamanajayaśabdakolahalo hiran. yagarbha iva bra-hman. d. at kr. tayugakaran. aya bhavanan nirjagama |

The Skandapuran. a and Ban. a’s Hars.acarita 163

prince actually takes place in Sthan. vıśvara (Sthaneśvara), on the banksof the Sarasvatı River.

The myth is told in the Mahabharata thrice, and we find the Śalya-parvan version partly verbatim in the Skandapuran. a.502 The scene of theaction is Samantapancaka, i.e. Kuruks.etra, on the banks of the Sarasvatı.

Thereupon, taking Skanda with them, the gods of boundless splen-dour came to the holy Samantapancaka, headed by Brahma. There,on the banks of the Sarasvatı, in a wide, open field, free of holes,stones, obstacles and saline soil, directed towards the north-east, allthose eminent gods then took their seats, around Skanda, while theirhearts rejoiced and they shone like blazing fires.503

Compared to the popularity of this theme in our Sanskrit sources, itsarchaeological, i.e. iconographical, attestation is limited. I know of onlytwo images dating from this period (i.e. 6th–7th century) that depict the

502 MBh 9.41.6–7; MBh 9.43–46. SPBh 164.1–3 ≈MBh 9.43.50–52; SPBh 164.61–185 ≈MBh 9.44.24–45.47. For a thorough analysis of these three versions in comparisonwith other sources, such as the Kumarasam. bhava and the Skandapuran. a, see Mann2012. Skandapuran. a 164.52–53 runs as follows (translation next to Plate 12):

atha te hr.s. t.amanasah. prahr.s. t.am. pavakatmajam |abhyas. incan tada sarve senapatye mahaujasam ‖ 52 ‖jayaśabdam. tataś cakruh. samam. sarve divaukasah. |jayinah. śarvaputrasya r.s.ayaś ca samagatah. ‖ 53 ‖

Cf. MBh 9.44.17–19:

jagr.hus te tada rajan sarva eva divaukasah. |abhis.ecanikam. bhan. d. am. mangalani ca sarvaśah. ‖divyasam. bharasam. yuktaih. kalaśaih. kancanair nr.pa |sarasvatıbhih. pun. yabhir divyatoyabhir eva tu ‖abhyas. incan kumaram. vai sam. prahr.s. t.a divaukasah. |senapatim. mahatmanam asuran. am. bhayavaham ‖‘Thereupon all those celestials, O king, brought an implement for the lustrationceremony as well as every other auspicous article (needed). The jubilant celestialssprinkled the prince (kumara) with the holy, divine waters of the Sarasvatı fromgolden vessels, O king, which were provided with the divine requisites, (conse-crating him as) great commander of the army to strike terror into (the heartsof) the Asuras.’

Cf. MBh 3.218.23: so ’bhis. ikto maghavata sarvair devagan. aih. saha | atıva śuśubhetatra pujyamano mahars. ibhih. ‖

503 SPBh 164.1–3 (cf. MBh 9.43.50–52):atha te skandam adaya sura brahmapurah. sarah. |samantapancakam. pun. yam ajagmur amitaujasah. ‖ 1 ‖tasmim. s tıre sarasvatyah. pragudakpravan. e same |vistırn. e śvabhrapas. an. akan. t.akos.aravarjite ‖ 2 ‖skandasya paritah. sarve tada sam. hr.s. t.acetasah. |nis.edus te suraśres.t.ha jvalitanalavarcasah. ‖ 3 ‖

164 Hans Bakker

abhis.eka of Skanda, both from Uttar Pradesh, one of which was found inKanauj (Plate 12).504

Thereupon all the assembled celes-tials and seers, delighted at heart,consecrated by sprinkling the de-lighted Son of Fire (i.e. Skanda), thevery strong one, as the commanderof the army, and then all togethercheered the victorious son of Śarva,shouting: ‘Victory!’

Skandapuran. a 164.52–53cf. MBh 9.44.17–19

Plate 12The abhis.eka of Kumara/Skanda

In addition to this general theme of the abhis.eka on the banks of theSarasvatı, Ban. a seems to have adopted other tropes from Epic and Pu-ranic counterparts. In the Skandapuran. a the impotence of the gods whendealing with Taraka is described.505 It begins with the lament of Indra

504 Agrawala 1967, Pl. XVIIIa. Agrawala dates this image to the 6th–7th century (ibid.p. 89), but the caption says ‘7th–8th century.’ Cf. Sharma 1971–72, p. 110: ‘Datableto the early Pratihara period, c. 7th century A.D.’ Sharma (ibid.) observes: ‘Inthis image too Brahma stands as usual on the right side of Skanda in the actof performing his abhis.eka. But the figure carved on the left side of Skanda hasbeen wrongly identified by P.K. Agrawala as that of Śiva. A minute observation ofthis figure would clearly reveal that it represents Vis.n. u and not Śiva, who besidesa kirıt.amukut.a and vanamala also holds a cakra in his rear left hand while hiscorresponding right hand is not visible.’ Sharma (ibid.) describes ‘another almostcontemporary sculpture of the deity from Uttar Pradesh, wherein the left sidefigure of Vis.n. u, though headless and damaged, is also shown wearing an ekavalı andvaijayantımala.’ This image is presently kept in the National Museum in New Delhi.A third stone image dating from the Gupta period is preserved in the MathuraMuseum (see Mann 2012, Figure 42).

505 SPBh 163.66–72. This conformity has been brought to our notice by Martine P.

The Skandapuran. a and Ban. a’s Hars.acarita 165

to Mahadeva, saying, ‘O Lord, our Laks.mı has been violated by someonecalled Taraka,’506 just after Indra had offered the kingship of the worldto Skanda, who had kindly declined.507 A reader of Sanskrit in the sev-enth century will not have failed to notice the agreement between themyth and Ban. a’s literary biography when, after the conversation betweenRajyavardhana and Hars.a, in which the former had offered the kingshipto his younger brother with the words: ‘Hence, though it is not muchthought-of, you should take from me the care of kingship, which annuls thejoys of youth,’508 a servant comes in and reports that, ‘Even the princess,Rajyaśrı, has been confined like a robber’s wife and thrown into prison inKanyakubja, her feet kissed by a pair of iron chains.’509

We can conclude from this comparison that Ban. a’s description ofHars.a’s despondency when his brother offers him the throne may owemore to literary convention than to historical reality (cf. p. 85).

Nılalohita and Kuruks.etra

Above we have noted (p. 162) that Prince Hars.a, as part of the ceremonythat inaugurated him as commander-in-chief, is said by Ban. a to haveperformed, with utmost devotion, a worship ritual for Lord Nılalohita.510

This may come as a surprise, since the figure of Nılalohita is not mentionedagain in the Hars.acarita, nor does he seem to play a very prominent rolein early Hinduism or Saivism. Nevertheless, the mention of Nılalohita inthis specific context is significant, a significance that we may derive fromthe Skandapuran. a, which builds on the Mahabharata and earlier Sanskritliterature.

In MBh 7.74 Nılalohita is said to be a figure who emerges from Śiva’sside—very strong, with tawny eyes, a brahmacarin, the embodiment oftapas—who demonstrates the use of the Paśupata weapon to Arjuna.In the following Karn. aparvan, Nılalohita is the manifestation of Śiva asthe berserk warrior, described as wearing only a skin, smoke-coloured

Kropman in an unpublished essay titled: The Consecration of Kumara.506 SPBh 163.66ab: tarakakhyena no laks.mır amr.s. t.a prasabham. prabho | Cf. Kalidasa’s

Kumarasam. bhava 2.31, in which Indra addresses Brahma: evam. yad attha bha-gavan amr.s. t.am. nah. paraih. padam | pratyekam. viniyuktatma katham. na jnasyasiprabho ‖ 31 ‖

507 SPBh 163.33: gr.han. a rajyam api etat prarthanıyam. manasvinam | Cf. MBh3.218.18–19: tatra tvam. mam. ran. e tata yathaśraddham. vijes.yasi | tasmad indrobhavan adya bhavita ma vicaraya ‖ skanda uvaca | tvam eva raja bhadram. te trai-lokyasya mamaiva ca | karomi kim. ca te śakra śasanam. tad bravıhi me ‖

508 HC 6 p. 309 [39]: yatas tvam antaritayauvanasukham anabhimatam api . . . gr.han. ame rajyacintam |

509 HC 6 p. 314 [41]: bhartr.darikapi rajyaśrıh. kalayasanigad. ayugalacumbitacaran. a cau-ranganeva sam. yata kanyakubje karayam. niks. ipta |

510 HC 7 p. 350 [53]: viracayya paramaya bhaktya bhagavato nılalohitasyarcam.

166 Hans Bakker

(dhumra), wrapped in the flames of his tejas, blazing like ten thousandsuns, and inciting fear.511 He is identified with Sthan.u, Mahadeva andRudra, and with the help of this warrior, the gods finally manage to con-quer the demons of the Triple City. MBh 13.14.154 calls him the foremostof Rudras, and MBh 13.16.47, finally, tells us that, in the past, Prajapatihad wished him, i.e. Śiva, to be his son with the name of Nılalohita.512

This brings us to the Skandapuran. a, when it describes in SP 4 that thedeluded Brahma wishes Śiva to be his son. Śiva says: ‘Because you wantyour father to be your son, you shall have a son in my form (murti); mygan. apa Rudra shall incarnate and he will teach you a lesson’ (SP 4.6–7). Itis then described how this Nılalohita is born from Śiva’s tejas in Brahma’ssacrificial fire. In this way the warrior figure of Nılalohita is connected tothe Vedic myth of Rudra’s birth. Just as in the Brahman. a sources, thisRudra Nılalohita is conceived of as Śiva’s manifestation in space and time,immanent in the phenomenal world, (symbolically) expressed by his eightnames and eight forms, as.t.amurti.513 Nılalohita, as an ectype of Śiva, thusseems to represent him in this world, in particular in situations where anabuse is to be corrected by violent means. This is also his main role in theSkandapuran. a.514 For this we shall again take a closer look at the Maha-tmya of the Sarasvatı as told by Vayu, which forms part of the KapalaCycle in SP I.

In the fifth chapter of the Skandapuran. a it is told how Brahma quarrelswith Yajna and Veda about his putative superiority, but when he goes asfar as to claim that he is the creator and leader of all Revelation, the timehas come to teach him a lesson. Brahma falls from the sky and lands in abower on the Himavat mountain. When Brahma is unable to discern whois doing this to him, since he is hidden in a sun-like disc in the sky, hesprouts a fifth, aggresive head. This head is chopped off by Śiva’s factotumNılalohita with the nail of his left thumb (SP 5.43). It is Nılalohita thenwho goes around with Brahma’s skull (kapala) as his begging bowl (SP5.64).

We have referred to this myth above (p. 152) in the context of thepractice of the Paśupata-Kapalikas. It was observed there that the Skanda-puran. a, with the exception of one early interpolation, seems free of Kapa-

511 MBh 8.24.87: sa nılalohito dhumrah. kr. ttivasa bhayam. karah. | adityayutasam. kaśastejojvalavr. to jvalan ‖

512 MBh 13.16.47: evam. prajapatih. purvam aradhya bahubhih. stavaih. | varayamasa pu-tratve nılalohitasam. jnitam ‖

513 SP 4.18: śarvadyair namabhir brahma tanubhiś ca jaladibhih. | stutva tam. sarvagam.devam nılalohitam avyayam ‖ 18 ‖ Kaus. ıtaki Brahman. a 6.1–3 (cf. ŚBr 6.1.3.8–18;PPL 121–127; Bakker 1996, 9ff.).

514 The figure of Nılalohita in the Skandapuran. a and his role in later Puran. a literatureis discussed by Granoff 2003, 96–107; Granoff 2006, 90f.

The Skandapuran. a and Ban. a’s Hars.acarita 167

lika influences as such. That does not alter the fact that the myth ofBrahma’s skull, as told here, probably relates to an existing practice byŚaiva ascetics of wandering around with a skull in hand which could notbe ignored. It accommodates this practice to Paśupata Saivism by tellingan ætiological myth, in which the original meaning of this observance,expiating the evil of brahmin slaughter, is no longer relevant. On thecontrary, Brahma asks Nılalohita to carry his fifth head/skull for ever, sothat he shall remember his own decapitation (SP 5.62). Nılalohita assumesa disguise, coils his hair into a kaparda, and goes around begging with theskull. The gods are not willing to give, but Vis.n. u tries in vain to fill theskull with his own blood. From this blood Nara is born (SP 6).

Nılalohita reaches the World of Brahma. Brahma asks Nılalohita thefollowing.

O Devadeveśa, I wish that this mark of mine be made by you insuch a way that this world may be marked by this mark, O Lord ofthe World.515

The Lord fulfils this wish and begins by creating Sarasvatı in the brahma-saras that is his mouth when he pronounces ‘OM’ (see above, n. 498 onp. 161). Sarasvatı creates the Brahmasaras in the Brahmaloka. Then shedescends to earth and Nılalohita follows her (anuprapya) in order to in-stall the skull in the foremost spot of this world.516 The skull installed atthat spot is known as Mahakapala and, once installed, it becomes knownas Śiva’s Pond (śivatad. aga). To date, our text continues, that great lake(saras) Mahakapala can still be seen there.517

A hermeneutic problem is created by the fact that our text does notsay explicitly where this ‘there’ (tatra), ‘the foremost spot in this world,’ is

515 SP 7.4: icchami devadeveśa tvaya cihnam idam. kr. tam | yena cihnena loko ’yam.cihnitah. syaj jagatpate ‖ 4 ‖

516 SP 7.13: tam. gr. hıtva mahadevah. kapalam amitaujasam | imam. lokam anuprapyadeśe śres.t.he ’vatis. t.hata ‖ 13 ‖

517 SP 7.24–25, 36:kapalam. sthapitam. yasmat tasmin deśe pinakina |mahakapalam. tat tasmat tris.u lokes.u gadyate ‖ 24 ‖sthapitasya kapalasya yathoktam abhavat tada |khyatam. śivatad. agam. tat sarvapapapramocanam ‖ 25 ‖tad adyapi mahad divyam. saras tatra pradr.śyate |mahakapalam. viprendrah. svargas tatraks.ayah. smr.tah. ‖ 36 ‖

This text passage has a complicated history and was borrowed by the composerof the Avantyakhan. d. a (SkP Avantyakhan. d. a 1.9.17–19). It has been analysed byYokochi 2004a. We agree with Yokochi that the intermediate verses SP 7.26–35are an interpolation (see also SP I, 73 n. 41). In these verses the Mahakapala siteis called a Siddhaks.etra, which is connected with a cremation ground (śmaśana).Here are two images of Deva installed: Śuleśvara and Mahakaya.

168 Hans Bakker

situated. From the text as we have it we can derive that we are concernedwith a great lake (saras), called Mahakapala, ‘the Great Skull,’ otherwiseknown as ‘Śiva’s Pond’ (Śivatad. aga), somewhere in the neighbourhood ofthe Sarasvatı River, a pond which is conceived of as the hallmark (cihna)of Brahma and as the most holy spot on earth.518

If we take the latter indication as our lead, we arrive at the land be-tween the Sarasvatı and Dr.s.advatı rivers, which is defined as Brahmavartaby Manu,519 whereas the Mahabharata calls the same Kuruks.etra.520 Theepic continues by saying that the holy Kuruks.etra is Brahma’s sacrifi-cial altar and within it is the Ramahrada,521 which is confirmed by theSaromahatmya in the Vamanapuran. a.522

This Mahatmya identifies this Ramahrada lake with the Brahma-saras.523 It is located near the Sthan. utırtha on the Sarasvatı. A traceof the myth as told in the Skandapuran. a has been preserved in thisMahatmya when it tells us that Brahma resorted to this lake when hishead had been crushed, as a result of his lust for his daughter. Afterhaving worshipped Śiva there, Lord Nılalohita himself comes to him,524

and tells him that, a long time ago, in the Varahakalpa, he had taken hishead (śiras), and this has become four-faced (caturmukha) and will neverperish.525 Brahma is freed from his sin (papa) and installs a linga on the

518 MBh 3.81.118 mentions a Kapalamocana near Saptasarasvata and Auśanasa inKuruks.etra, where Mankan. aka had his vision of Mahadeva (see below n. 538 onp. 174). A big modern reservoir called Kapalamocana is found near a hamlet 3.5 kmnorth of the village of Bilaspur (50 km north-east of Thanesar). There is a littleCan. d. ı temple on its banks, where some archaeological remains (lingas and anamalaka) indicate that this is an old site. However, it seems unlikely that it is thetırtha mentioned in the Mahabharata or the Skandapuran. a.

519 MaS 2.17: sarasvatıdr.s.advatyor devanadyor yad antaram | tam. devanirmitam. de-śam. brahmavartam. pracaks.ate ‖

520 MBh 3.81.175 daks. in. ena sarasvatya uttaren. a dr.s.advatım | ye vasanti kuruks.etre tevasanti trivis. t.ape. Cf. VamP Saromahatmya 1.1.

521 MBh 3.81.177f.: brahmavedı kuruks.etram. pun. yam. brahmars. isevitam | tad avasantiye rajan na te śocyah. katham. cana ‖ tarantukarantukayor yad antaram. ramahra-danam. ca macakrukasya | etat kuruks.etrasamantapancakam. pitamahasyottaravedirucyate ‖ Cf. VamP Saromahatmya 1.14. See for these verses Kane IV, 683 n. 1551.

522 VamP Saromahatmya 1.13: adyais. a brahman. o vedis tato ramahradah. smr.tah. |kurun. a ca yatah. kr.s. t.am. kuruks.etram. tatah. smr.tam ‖ The Ramahrada and theSamantapancaka refer to the myth of Rama-Jamadagnya (Paraśurama) who ex-tinguishes the ks.atriyas. This myth is also briefly told in SPBh 123.1–29; SPBh

123.21 mentions Samantapancaka as the place where this Bhargava Rama filledseven (sic) hradas with their blood. There are no connections between this mythand the Kapala Cycle in the Skandapuran. a.

523 It is also called Sannihatya/Sam. nihita-saras, after the river Sam. nihitı. VamPSaromahatmya 1.4–5; Kane IV, 686; above, n. 493 on p. 160; ad SP 29.79a.

524 VamP Saromahatmya 28.9: tasyaivam. bhaktiyuktasya śivapujaparasya ca | svayamevajagamatha bhagavan nılalohitah. ‖

525 VamP Saromahatmya 28.20: pura varahakalpe te yan mayapahr. tam. śirah. | catur-

The Skandapuran. a and Ban. a’s Hars.acarita 169

banks of this lake.526

On the basis of the combined evidence, we can conclude that the Maha-kapala or Śiva’s Pond of the Skandapuran. a was regarded as the terrestrialcounterpart of the Brahmasaras in the Brahmaloka created by Sarasvatıand springing from Śiva’s mouth. The myth of Nılalohita and Brahma’sskull served as an ætiology for this big holy pond in Kuruks.etra, which istoday known as Brahma Sarovar.

Plate 13The Mahakapala or Brahma Sarovar in Kuruks.etra

Kuruks.etra

The Kapala Cycle, which contains the Skandapuran. a’s first Mahatmya,viz. the glorification of (the) Sarasvatı, may therefore also be seen as theMahatmya of Kuruks.etra. In addition to the Sarasvatı, the Sthan.utırthaon its banks, and this great holy pond Mahakapala, its main deity seemsto have been Śiva in his form as Nılalohita. In this way the Skandapuran. agave a powerful Śaiva twist to the respectable Sanskrit lore that existedregarding this ancient site (see Kane IV, 680–86).

mukham. ca tad abhun na kadacin naśis.yati ‖526 VamP Saromahatmya 28.37–38.

170 Hans Bakker

That this twist was in conformity with the religious situation by theend of the sixth century is evidenced by the Hars.acarita, in which Ban. adepicts Thanesar (Sthan. vıśvara) under the legendary king Pus.yabhuti asa country completely devoted to Maheśvara.527 A seal reading ‘śrırudrah. ’in ‘first-second century characters’ was reportedly found in the Kushanalayers of the so-called Hars.a ka T. ıla in Thanesar (IAR 1987–88, 29), towhich another terracotta seal (kept in the site museum of the Hars.a kaT. ıla) reading śrıbhairava in seventh century characters may be added. Inview of this significant Śaiva evidence, it can no longer come as a surprisethat Hars.a is said to have worshipped Nılalohita during his inaugurationceremony, not only because this ectype of Śiva represents his warrior side,evoked in the face of a dangerous threat, as we have seen, but also becausethe Paśupata tradition connected this figure in particular with Kuruks.etraand its central holy place, the Mahakapala or Śiva’s pond. Whether ornot an independent cult of Nılalohita existed in Kuruks.etra is difficult todetermine in the absence of archaeological evidence.528

The Skandapuran. a’s report, finally, saying that the founder of the Pa-śupata lineage in Kanyakubja (Kanauj) came from the country of theKurus, corroborates our observation that Kuruks.etra in about this timewas important Paśupata territory.529

527 HC 3 p. 164 [45]: gr.he gr.he bhagavan apujyata khan. d. aparaśuh. | That with khan. d. a-paraśuh. (‘the lord with the cleaving axe’) Śiva is meant follows from the precedinglines, in which Pus.yabhuti’s devotion is said to have been exclusively directed toBhava, who sustains the universe, creates the beings and cuts through the chain ofworldly existences: bhagavati bhaktisulabhe bhuvanabhr. ti bhutabhavane bhavacchidibhave bhuyası bhaktir abhut |

528 SP 8.30–35 describes an epiphany of Nılalohita on Mount Mainaka. The gods,who have been given a divine eye, see the following figure encircled by flames ina vimana: his gaping mouth showing dreadful fangs, his eyes like fire, his mattedhair glowing like the triple sacrificial fires, his girdle formed by snakes, his earringsgleaming, his hands holding the lance (śula), the Pinaka bow, rod, hammer, thun-derbolt, sword, halberd, and disc, with erect penis, the rosary in his hand, wearingmoon, sun and planets in a wreath. (Synopsis of SP 8.30–35; Skandapuran. a Vol.I, 75, 154).

529 SPS 167.123cd: brahmacarı caturthas tu kurus.v eva sugotrajah. ‖; SPS 167.129:kanyakubje tataś canyam anugr.hya jagatpatih. | svasiddhantam. dadau yogam uva-cedam. ca lagud. ih. ‖129 ‖ The reading of SPS 167.123d is uncertain. The syllableskuru are relatively certain as they are attested in all manuscripts ad loc. (S1 andS2 and all SPRA MSS (R before correction)). SPS reads the corrupt ◦s.vava insteadof ◦s.veva, but the latter reading is supported by all A MSS. The latter MSS readsa gotrajah. instead of sugotrajah. . We deduce from this evidence that the name ofthe fourth pupil was either unknown to the composer of the SP or, for reasons un-known, he did not wish to reveal it, but the tradition reported connects him withthe Kurus. In order to supply a name for the founder of this lineage, this traditionlater invented the name Kaurus.ya (LiP 1.7.51, 1.24.131, ŚiP Śatarudrasam. hita5.49) or Kaurus.a.

The Skandapuran. a and Ban. a’s Hars.acarita 171

We may thus conclude that the region around the city of Sthaneśvara, i.e.Kuruks.etra, was an important centre of Saivism in the sixth and seventhcenturies, from where Śaiva teachers originated, and which attracted suchfigures. We encountered one such figure in Part I (p. 78) of this study, viz.the Mahaśaiva teacher Bhairavacarya, who hailed from the South andpersonified, as it were, the one who destroyed Daks.a’s sacrifice.530 Thisteacher is reported to have resided in a Bel-tree garden, north of an oldMatr. temple.531

The Skandapuran. a provides many connections between the Paśupa-ta movement and the cult of goddesses or ‘Mothers.’ This is a very oldrelationship, as follows from the Grant of Bhulunda, Year 56 (ad 374/5),in which it is said that the Paśupata acarya, Bhagavat Lokodadhi, hadfounded a temple at the site of the Mother Goddesses.532 In the presentcontext, however, it is relevant to draw attention again to the myth thatexplains the origin of the Mahakapala.

In verses SP 7.15–23, preceding the ones about the Mahakapala quotedabove (n. 517 on p. 167), we are told how the ‘Mothers of the Skull,’ Kapa-lamatr.s, came into being in that neighbourhood. They were Piśacıs, whohad come there to eat the buffalo demon (asura) Halahala. Halahala andhis fellow demons had been attracted by the roar set up by the Gan. aswhen Nılalohita installed the skull of Brahma. They were killed by theGan. as who defended the skull and the Piśacıs were invited to feast onthe remains of the buffalo. A connection between the Piśaca cult and thePaśupatas is also attested by the Bagh Inscriptions, in which Paśupatasare said to serve in the temple of Bappa Piśaca-deva.533

On account of this, and in view of the ritual practised by Bhairava-carya in the cremation ground (see above, p. 78), the accommodation ofthis Mahaśaiva acarya near a Matr. temple is not surprising. Although thehistoricity of Bhairavacarya may be doubted, the presence of a temple ofthe Mothers in the vicinity of the Mahakapala pond is, on the basis of thisjoint evidence from the Skandapuran. a and the Hars.acarita, a historicallyplausible proposition. It may have been an unpleasant place. We will comeback to this cult of Mothers or Matr.s below (p. 257).

530 HC 3 p. 165 [45]: . . . bhagavantam aparam iva saks. ad daks.amakhamathanam.daks. in. atyam. bahuvidhavidyaprabhavaprakhyatair gun. aih. śis.yair ivanekasahasra-sam. khyair vyaptamartyalokam. bhairavacaryanamanam. mahaśaivam |

531 HC 3 p. 169 [46]: ‘asya jırn. amatr.gr.hasyottaren. a bilvavat.ikam adhyaste’ iti |532 Ramesh & Tewari 1990, 22: . . . bhagavallokodadhipaśupatacaryapratis. t.hapitakapi-

ncchikanakagramamatr.sthanadevakulasya . . . .533 Ramesh & Tewari 1990, 10f., 13, 26.

172 Hans Bakker

Conclusion

It transpires from the above analysis that the interface between theSkandapuran. a and the Hars.acarita is substantial. This, however, doesnot imply that there is intertextuality between the two texts, such asthere is between the Skandapuran. a and the Mahabharata or Vayupuran. a,for instance. Given the fact that the two texts belong to very differentgenres of Sanskrit literature, this was not to be expected from the outset.Consequently, the question of whether the composers of both texts mayhave known each other’s work should be addressed. Since we may asumethat composers at this level will have been acquainted with the majorSanskrit works of their times, this amounts to a question of chronology.

As we have discussed above (p. 95), Ban. a situates his first meetingwith King Hars.a in the latter’s army camp on the River Ajiravatı, wherethe king resided after his war with Śaśanka. This must have taken placein around ad 612. After first having been snubbed, Ban. a tells us that hesoon after became a regular visitor to the court (above, p. 158). The com-position of the Hars.acarita thus seems to have been begun after ad 612.When one day (atha kadacid, the opening words of the third Ucchvasa),the poet returned to his native village, where he recited a part of what wasto become his Hars.acarita, it seems that Ban. a’s favour with the king wascommon knowledge and his literary enterprise was known to his cousins.Irrespective of whether this recitation actually took place or not, the over-all structure of the composition, which includes the recitation scene itself,suggests that the work was completed after Ban. a’s return to his village.

From these admittedly inconclusive indications, we may tentativelydeduce that the Hars.acarita cannot have been finished much before theend of the second decade of the seventh century. That is, its compositionmay have begun once that of the Skandapuran. a was concluded. Ban. athus may have known some or all of the Puran. a text and it cannot beruled out that in the Arya verses quoted above (p. 155) our Skandapuran. aresounds, in particular its Sarasvatı Mahatmya, which was indeed sungby Vayu, the God of the Wind—or, in the words of Sucıvan. a: ‘Methinksthat this Puran. a does not differ from the deeds of Hars.a after all!’

The Myth of Daks.a’s Sacrifice

d Hardwar and Rishikesh d

Preamble

The myth cycle that opens the Puran. a, that of Rudra–Nılalohita and theKapala (SP 3–7), thus leads to the Mahatmya of the Sarasvatı River andthe holy field of Kuruks.etra. It is completed by the Mahatmya of theSthan.utırtha on the Sarasvatı and the neighbouring town of Sthaneśvarain SP 31.48–115. The great majority of the Mahatmyas that follow in theSkandapuran. a are situated to the east of this.534

In between the Mahatmya of Kuruks.etra and its completion, the Ma-hatmya of Sthaneśvara in SP 31, the composers of the Skandapuran. a haveinserted the Mahatmya of what they must have considered to be the majorPaśupata holy place, viz. that of Varan. ası (SP 26–31.14).535 We have givenan elaborate analysis of this Mahatmya in the Introduction to VolumeII A of our edition, and there seems to be no need to repeat it here. Ithas been shown there that of all the Paśupata settlements, Varan. ası wasthe most successful in the sense that the Paśupata preoccupation withcremation grounds was accommodated by the Hindu community at large,a gradual process, which seems to have gained momentum in the seventhcentury. The original Paśupata sanctuary near the śmaśana, Avimukte-śvara, dating back to at least the Gupta period, developed into one of theholiest temples of pre-Muslim India.

534 An exception is the Mahatmya of Śankukarn. eśvara at the mouth of the Indus Rivertold in SPBh 73. For this Mahatmya see Granoff 2004, 114f. and Bisschop 2006, 220.See also above, n. 362 on p. 117 and n. 471 on p. 151. Another possible exceptionmay be the Mahatmya of Japyeśvara and the Pancanada in SP 22, which is toldin the context of the Nandin Cycle (SP 20–25), and whose location we do notknow (see annotation ad SP 22 (synopsis) and Bisschop 2006, 191). Later sourcesassign these holy places to Kamarupa, but nowhere else in the Skandapuran. a arethere references to that far-eastern area. The easternmost and last ayatana thatis described in the Skandapuran. a is Kot.ıvars.a (SPBh 171.78–137). The Pancanadanaturally suggests the Panjab rather than Assam.

535 See ad SP 31.15 (SP II B, pp. 4, 32 n. 59).

173

174 Hans Bakker

We will turn to another site of great significance for the sixth cen-tury Śaiva community, which, in contrast to Varan. ası, is now completelyforgotten, by continuing with the Mahatmya of Bhadreśvara. This formspart of the myth cycle of Daks.a’s sacrifice and its destruction, told inSkandapuran. a 32.1–200.536

Bhadreśvara

The initiation of the gods in the Paśupata vrata

Bathing in ashes is the hallmark of the Paśupata movement. Not onlyascetics, but also gods are transformed by this practice, as we will seeshortly. The RA recension of the Skandapuran. a relates the myth of theorigin of the bath in ashes (bhasman/bhuti), which may be briefly reportedhere.537

An ascetic, frenzied by his mortifications (tapomada), cuts his finger.When vegetable juice (śakarasa) pours out he twists and shouts. Śiva,in disguise, approaches him and tells him to calm down and advises himto take a bath in the ashes that he pours out of his body (instead ofblood).538 The sage, taken aback, takes a bath and cools down. He recog-nizes Śiva, praises him, and asks for the ‘grace of ashes’ (bhuter anugraham.kuru). Śiva complies by taking a bath in ashes. Ever since, bathing inashes (bhutisnana) is the greatest of all gifts; it makes effective the merit(saphalam) that is gained in (Śaiva) holy places, first and foremost inKanakhala.539

536 The myth of Daks.a’s sacrifice is ubiquitous in Sanskrit literature. The basic textsare MBh 12.274 and MBh 12 App. I No. 28. For the development of the Daks.amyth see Mertens 1998. Klostermaier 1991, 111–14 argues that the Daks.a mythreflects the historical occupation of the holy site of Kanakhala by the Paśupatasin their drive against their Vais.n. ava rivals.

537 The Anukraman. ika in SP 2.23a announces the myth of how the bath in ashes cameabout (bhasmasomodbhava). The myth is not found in the S recension but is toldin RA, immediately after the myth of the origin of the crescent on Śiva’s head (forwhich see below, n. 631 on p. 212). I am basing myself here on the synopsis given ofthe RA recension by Judit Torzsok (Groningen 2001; not published) and Torzsok2004, 24–31. For a discussion on how to explain the fact that myths are listed inthe Anukraman. ika but are missing in the oldest S recension see Torzsok op. cit.,26ff. and above, p. 16.

538 This is similar to the myth of the sage Mankan. aka told in MBh 3.81.97–118 and9.37.33–50, which myth there, however, serves to recommend a bath in the Saptasa-rasvata in Kuruks.etra, instead of a bath in ashes (above, n. 518 on p. 168; Torzsok2004, 25).

539 This is my interpretation of the verse as found in the R and A manuscripts:pun. ya(m. ) kanakhale yac ca prayage yad athapi ca | tat phalam. saphalam. devibhutisnane vidhıyate ‖ It is one of the verses only occuring in the RA recension that

The Myth of Daks.a’s Sacrifice 175

Kanakhala is the place where Daks.a performed his great sacrifice. Itis close to Gangadvara (modern Hardwar) and Mount Golden Peak.540 Inorder to witness its destruction by Haribhadra, Bhadrakalı and the Gan. as,Śiva and Parvatı hide themselves in the vicinity, in the hermitage of Rai-bhya (SP 32.23). When the sacrifice has been destroyed, the despondentgods discover Śiva nearby and propitiate him. Śiva laughs and lets thepoor gods enter his body, where the whole universe is shown to them asfollows.

The first stage is the universe (jagat) of the seven lokas, which maystand for the traditional seven bhuvanas/lokas of the Cosmic Egg. Thehigher, eighth, to which the gods are brought by a Gan. apa named Panca-ks.a, contains seven concentric, increasingly paradisiacal cities (nagaras),where endless happiness reigns. This eighth loka may be equivalent to theŚivapura, the paradise where the laymen Śaivas (laukikas) go, situated atthe top of the Brahman. d. a.541 The hard-core Paśupata realm, however, isbeyond (ati) the shell of the Cosmic Egg, seemingly represented by a closedcircuit of fire.542 It is the gods’ ultimate destination.543 Pancaks.a forces anentrance with a lance (śula) and the gods are allowed to have a glimpseof it, but since they are not yet initiated, they only experience Śiva’sangers (krodhas) in the form of ferocious lions. The gods are overwhelmedby fear. Thereupon that final city and the terrifying delusions disappear.The revelation ends here.544 The gods are allowed to see Śiva and Parvatıagain in their natural appearance, as they reside in Raibhya’s aśrama, andthey take refuge with them. Śiva has reined in his anger, but Devı’s anger,in its embodiment as Kalakarn. ı, attacks the gods.

Thereupon the gods, shaken by fear of Kalakarn. ı, frightfully entera heap (raśi) of ashes (bhasman) at Deva’s side. When Devı seesthat the gods have taken shelter in a mound of ashes and are (con-sequently) smeared with ashes, she holds back (that) goddess andsays: ‘O Kalakarn. ı, stop! Don’t kill the eminent gods; they have be-come Paśupatas, since their bodies are smeared with ashes. In the

is quoted by Laks.mıdhara. In his Niyatakalakan. d. a (p. 54) this verse runs: pun. yam.kanakhale yac ca prayage yac ca sundari | tat phalam. sakalam. devi bhutisnane dinedine ‖ (other MSS read vidhıyate instead of dine dine, and so do all R and A MSS).Cf. Torzsok 2004, 28f.

540 See below, n. 571 on p. 185. SP 32.24:vr.ks. ah. kanakhala yatra gangadvarasamıpagah. |suvarn. aśr. ngaś ca girir meruparvatasam. nibhah. |tasmin pradeśe daks.asya yajno ’yam abhavat tada ‖ 24 ‖

541 See above, p. 151. It should be noted that the name Śivapura is not used in thispassage.

542 SP 32.86ab–87cd.543 SP 32.89cd.544 SP 32.95.

176 Hans Bakker

past this Paśupata observance has been proclaimed by Paśupati,545

for a bath in purifying ashes is the best of baths.’546

So it is the fear of feminine fury that drives the gods into the Paśupatafold. Since this fury Kalakarn. ı turns out to be a personification of Death(Mr.tyu), it is ultimately Death itself that is transcended, the raison d’etreof all religion.

This is a key Paśupata myth, which illustrates the power of the prac-tice of bathing in ashes. It may also be an ætiological myth accountingfor the practice of giving initiation names ending in raśi (above, p. 139).We have not found this myth in other sources. However, the Śalinı versequoted from an unknown source by Kaun.d. inya in the Yamaprakaran. a ofhis commentary expresses the same idea:

Having drunk alcohol, having slept with one’s guru’s wife, havingstolen or killed a brahmin, one who is covered by ashes, who lies ina heap of ashes, and studies Rudra, he is freed from (those) sins.547

When the gods have come to their senses, have praised Devı and have seenthat all the world including themselves are originating from Śiva’s bodythey say:

This highest mystery (guhya) is revealed to us by You, O God. Themortal who knows it attains to imperishable worlds. This place, OLord, shall be famous as Bhadreśvara; it is the best place on earth,as it is the most beneficial of (all holy) places. It is an eternal holyfield of Rudra measuring one yojana in circumference. When onedies here one shall become a Gan. apa and favourite of Yours.548

545 Cf. MBh 12 App. I No. 28 ll. 407–08: maya paśupatam. daks.a yogam utpaditam.pura | tasya cırn. asya tat samyak phalam. bhavati pus.kalam ‖ According to Paśupa-tasutra 4.10, Indra was the first practitioner of the observance. Cf. Kaun.d. inya adPS 1.1, 4.10 etc. SPBh 122.71–74 tells us that Vis.n. u and all the other gods wereinstructed in the Paśupata vrata by Devadeva on Mount Śailendra.

546 SP 32.104–07 (cf. SPBh 180.1–4):tato vyathitacittas te kalakarn. ya bhayat surah. |bhasmaraśim. sthitam. parśve devasya viviśur bhayat ‖ 104 ‖tan dr.s. t.va bhasmakut.am. tu pravis. t.an charan. arthinah. ‖suran bhasmaviliptangan devım. devı nyas.edhayat ‖ 105 ‖devy uvaca |kalakarn. i nivartasva ma vadhıh. surasattaman |ete paśupatıbhuta bhasmana digdhamurtayah. ‖ 106 ‖etat paśupatiproktam. vratam. paśupatam. pura |yad bhasmana pavitren. a snanam. snanebhya uttamam ‖107 ‖

547 Pancarthabhas.ya ad PS 1.9:madyam. pıtva g(u)rudaram. ś ca gatva, steyam. kr. tva brahmahatyam. ca kr. tva |bhasmoddhvasto bhasmaraśau śayano, rudradhyayı mucyate patakebhyah. ‖

548 SP 32.134–36:

The Myth of Daks.a’s Sacrifice 177

If we follow the narrative of Skandapuran. a 32, this place must be the samespot where Śiva and Parvatı were hiding in order to watch the destruc-tion of Daks.a’s sacrifice, viz. Raibhya’s hermitage (Raibhyaśrama), nearKanakhala where the sacrifice took place.549 It would thus follow thatRaibhyaśrama and Bhadreśvara are either identical, or that the sacredcomplex of the latter included the former. The text continues by sayingthat here is the Bhadrakarn. ahrada (SP 32.138), and the spot where Śivaaddressed the gods is called ‘Haribhadra’ (SP 32.140). Then it is said thatin this holy site Brahma has installed a linga (SP 32.153).550 It is evidentthat Bhadreśvara was an important Paśupata holy place in the vicinity ofHardwar (Kanakhala) at the time that our text was composed and it istherefore all the more curious that this tırtha has apparently fallen intooblivion.551

Plate 14Bird’s-eye view of the geography of

site VBA (Bhadreśvara), Gangadvara, and Kanakhala (Daks.eśvara)

guhyam etat param. deva tvayasmakam. nidarśitam |evam. yo vetsyate martyah. sa lokan prapsyate ’vyayan ‖ 134 ‖idam. ca bhagavan sthanam. bhadreśvaram iti śrutam |bhavis.yati jagacchres.t.ham. sthanebhyah. pun. yakr. ttamam ‖135 ‖samantad yojanam. caiva rudraks.etram. sanatanam |mr.to ’tra gan. apo deva vallabhas te bhavis.yati ‖ 136 ‖

549 That this hermitage is not far (natidure) from Kanakhala seems also to follow fromSP 32.50–52.

550 Bhadreśvara is mentioned again in SPS 167.182 (see Bisschop 2006, p. 221f. adloc.). It is also mentioned in SP 29.83d and SPBh 183.49c.

551 Laks.mıdhara, quoting a Vamanapuran. a (not identified), refers to Bhadreśvara inthe land of Pancala and says that ‘it is protected/concealed by Gan. as’: bhadreśva-ram. gan. air guptam (TVK 267). Cf. KuP 2.39.4, MtP 22.25, 32, LiP 1.92.136.

178 Hans Bakker

Bhadreśvara: the site

It is obvious that for Bhadreśvara we have to look in the neighbourhood ofHardwar, 130 km due east of Kuruks.etra. The historicity of the site seemsto be attested by a seal-die found in the Ganges canal between Hardwarand Kanakhala, depicting a Gaja-Laks.mı above, the legend in 6th/7th-century script, reading śrı(bha)dranupavis.ayadhikaran. asya, i.e. ‘From theoffice of the river-bank district of Śrı Bhadra.’552

Plate 15Hardwar: Seal-die

A first brief notice of the discoveryof a promising archaeological sitein this area was published in IAR1963–64 (p. 45). It reports that,‘Near the confluence of the Rambhaand Ganga outside the [Antibiotic]Factory area [at Rishikesh], ruinsof a massive brick fortification werenoticed.’ This led to exploration ofthe site ‘Virabhadra,’ VBA-I andVBA-II, under the supervision ofN.C. Ghosh in 1973–74 and 1974–75. This site, on the high bank of theRambha, near its confluence withthe Ganga, is situated 20 km north-east and upstream of the Daks.eśva-ra Temple, i.e. Kanakhala, the spotwhere Daks.a’s sacrifice is supposedto have taken place.

If we are allowed to look with the divine sight of the myth-maker, it couldbe viewed as the perfect vantage point for Śiva and his spouse to watchunnoticed the spectacle of the destruction of the sacrifice, which took placein the plain, 60 m lower.

Excavations revealed a site that was occupied continuously ‘from circathe second century A.D. to circa eighth century.’553 The excavation of siteVBA-II exposed two temples. Structure 1 belongs ‘to the early level ofthe middle phase,’554 that is the Gupta period (4th to 5th centuries), andremained in use during the rest of the period. In Structure 1 was founda ‘Siva-linga resting on a bhadra-pitha,’ ‘octagonal at the base.’555 It is

552 The seal is preserved in the Victoria & Albert Museum (London), No. IM.496–1924. According to the V&A website, characters 2–4 of the legend are read byGauriśwar Battacharya (2006) as ◦vahanta◦, rather than ◦bhadranu◦.

553 IAR 1973–74, 28.554 IAR 1974–75, 41.555 IAR 1974–75, 41.

The Myth of Daks.a’s Sacrifice 179

Plate 16Confluence of Rambha and Ganga; from site VBA

illustrated in Plate 17.Another structure (2) was found, in the report said to be ‘about 20 m

south of structure 1.’ Reportedly, this structure ‘was added to Structure 1.’It is said to be in alignment with it, oriented east-west with approach fromthe east.556 Although the report speaks of ‘massive religious structuresand a vast complex of residential establishments attached to the abovestructures,’557 the nature of Structure 2 is not clear.

When we visited the site in January 2012, we found a second linga(Plate 18), similar to the one noticed above, which is, however, not men-tioned in the three reports made by N.C. Ghosh.558 It remains as yetunclear to which of the two structures or temples this linga originallybelonged.

The identification of the archaeological site VBA with the holy placeBhadreśvara of the Skandapuran. a finds further support in the presenceof a (modern) temple, 50 m north of the VBA mound, called Vırabhadre-śvaramahadeva. On the basis of our fieldwork data and the excavationreports, it may provisionally be concluded that the site VBA developedinto an important Śaiva complex in the fifth to eighth centuries, containingat least two linga temples or shrines. The complex as a whole, as well as

556 IAR 1973–74, 30.557 IAR 1973–74, 30.558 IAR 1973–74, 28–30; IAR 1974–75, 41f.; Puratattva 1975, 180f.

180 Hans Bakker

possibly one of the lingas, were known to the composers of the Skanda-puran. a as Bhadreśvara. The other temple/linga may have been known asHaribhadra, as suggested by our Puran. a.

Plate 17Excavation site VBA-II, structure 1: temple with linga (1)

Plate 18Excavation site VBA-II, linga (2)

The Myth of Daks.a’s Sacrifice 181

Adjacent to these temples was a monastic settlement of Śaiva ascetics.According to the Skandapuran. a, it was a Rudraks.etra, measuring oneyojana in circumference (SP 32.136). It contained a Bhadrakarn. ahrada.559

The excavators of VBA I and II date the first phase of this site tothe Kus.an. a period (2nd to 3rd cent. ad), on account of pottery and acoin ‘belonging to the Kushan copper coinage of circa first-second centuryA.D.’560

KubjamrakaThe presence of the Śaiva religion in the neighbourhood during this earlyperiod is corroborated by local finds that are preserved in the modest,but rich gallery that is attached to the Hr.s.ikeśa Narayan. a Śrı BharataMandir. This temple itself is within the precincts of the Śam. karacaryaAśrama and is situated 5 km to the north of VBA, at the confluence ofthe Candrabhaga and Ganga rivers in Rishikesh.

Plate 19Saivite figure

Bharat Mandir Museum

The museum preserves a standing, ithy-phallic figure, measuring 167 x 60 cm,holding a kaman. d. alu in his left hand andshowing abhayamudra with his right. Hehas a kaparda, is ornamented by a neck-lace and wrapped in a sash, of which oneend is on the right shoulder and the otheris draped over the left wrist. This imagein Mathura stone is datable to the Kus.a-n. a period (2nd–3rd cent. ad).561

According to local knowledge, thissite is associated with Raibhya’s her-mitage. This tradition is also foundin the so-called Kedarakhan. d. a (KeKh),a work ascribed to the Skandapuran. a,but which is different from, and prob-ably later and more extensive than, theKedarakhan. d. a that is published by theVenkateshvara press as part of the edi-tion of the Skandapuran. a. In chapter 116

559 SP 32.138. MBh 3.82.35 knows a Bhadrakarn. eśvara. Laks.mıdhara, quoting a Ma-tsyapuran. a (not identified), mentions a Bhadrakarn. a (TVK 241). Further see e.g.KuP 1.29.45–46, KuP 1.33.15, BrP 25.50. Bhadrakarn. a features in the pancas. t.aka,the five ogdoads of the early Agama literature, which list Śaiva sanctuaries; seeBisschop 2006, 32.

560 IAR 1973–74, 29.561 The museum preserves an image of a female figure (123 x 41 cm) made of Mathura

stone, styled Yaks.ı, of the same style and date.

182 Hans Bakker

it tells the story of Saint Raibhya, who, on seeing Vis.n. u, became crooked(kubja), after he had taken refuge at a mango tree (amra).562 Vis.n. u wassatisfied and allowed the spot to become the holy place Kubjamraka, wherehe will be present. In the Kali age he, i.e. Vis.n. u, would be installed byŚam. kara under the name of Bharata.563 This is obviously the Mahatmyaof the present Bharata Mandir, which temple, however, is an ancient one,to judge by some of the sculptures on its pratis. t.hana.564

An intriguing fragment of an image in the museum (36 x 25 x 33 cm)points stylistically to the fifth or sixth century. It consists of an attribute,the cakra with left forefinger still on it, and side frame. The main image ismissing.

Plate 20Fragment of Vis.n. u image: cakra

Bharat Mandir Museum

The frame above this cakra has an imageof a couple sitting on a bull. The goddessapparently turns her back to the male de-ity and is sitting astride his left leg. Heseems to be embracing her with his lowerleft hand. His upper left hand seems toholding a rudraks.a. Next to the coupleis a standing male figure, who wields aclub and whose right foot is resting on thetrunk of a makara. There is something inbetween this figure and the goddess, whomay be bestowing something on him; amala?

This fragment of a supposedly Vis.n. uimage brings us back to our Skandapura-n. a, which reports, as an exception, the ex-

562 KeKh 116.11f., 116.31: yasmad amram. samaśritya kubjarupen. a vai tvaya | dr.s. t.o’smi raibhya tasmad vai kubjamrakam iti smr.tam ‖ 31 ‖ This goes back (partlyverbatim) to the Mahatmya of Kubjamraka as it is told in VarP 126.1–28, whichis quoted by Laks.mıdhara (TVK 206). It is called Kubjamraka because Raibhyabecame crooked (kubja) on seeing Varaha (VarP 126.14). Cf. MBh 3.82.36, KuP1.29.46, KuP 2.34.33–35, NsP 65.11 (TVK 252), ViS 85.11.

563 KeKh 116.41: śam. karah. śam. karah. saks. at punar mam. sthapayis.yati | kalau bharata-namanam. vadis.yanti mahıtale ‖ 41 ‖ This Śam. kara is believed to be Śam. karacarya,the Vedanta philosopher, who reinstalled ‘the idol of Lord Hrishikesh Bharatji inthe temple on the pious day of Basant Panchami,’ ‘during the first decade of theninth century’ (Sharma s.d., 10).

564 Ghosh & Sharma in Puratattva 1975, 181 date the temple to c. ad 800. An in-scription found behind the Hr.s.ıkeśa Narayan. a image (= Bharata Bhagavan) alongthe pradaks. in. apatha seems to testify to a renovation of the temple in the 12thcentury by an unknown king. The outer structure of the temple was reconstructedby Maharaja Yaśavanta Sim. ha Bahadur of Nabha State in ad 1832 (Sharma s.d.,6f.).

The Myth of Daks.a’s Sacrifice 183

istence of a ks.etra of Vis.n. u, (measuring) eight nivartanas,565 located notfar from Bhadreśvara/Haribhadra. According to our Puran. a this placeoriginated when Śiva left for Mount Mandara, after he had confirmed thestatement made by the gods to the effect that Bhadreśvara would be amost holy Rudraks.etra. The gods and king (Daks.a) stay behind at thatplace, but Brahma and Vis.n. u follow Śiva.

The Lord, after having gone a small distance with these two, spokenot far from there [i.e. from Bhadreśvara]: ‘Stop, O mighty Vis.n. u.’Purus.ottama bowed to the feet of Hara and stood still, holding (abranch) of a mango tree; that (tree) became crooked (kubja).

Because Hari stood at that site holding a mango tree (amra)while staring at Deveśa, that place became known as Kubjamraka:a wealthy ks.etra of Vis.n. u, holy, measuring eight nivartanas, andyielding the results of a (donation) of a thousand cows.566

There can be little doubt that the place described corresponds with theBharata Mandir in Rishikesh and testifies to the importance of this Vai-s.n. ava site in the sixth century.567 Its Mahatmya in the Varahapuran. a andKedarakhan. d. a derives from the one in the Skandapuran. a.

Brahmavarta

The Skandapuran. a continues by specifying a place connected to Brahma.

Śiva, having gone a little further, said to Brahma ‘stop’ (nivarta)and entered into the sky. After that god had gone into the sky,Brahma folded his hands, turned his head upwards, and made acircumambulation; then, making an obeisance, he proceeded.

That site is known as Brahmavarta and is pure, because he (i.e.Brahma), by performing the holy act of circumambulating Hara,has made his own body turn around (avarta) there. One receivesthe reward of an Aśvamedha by bathing there, and one reaches the

565 A nivartana is an area measure, equal to 20 dan. d. as (rods), c. 36 m2, the size of asmall temple.

566 SP 32.145–48:

sa gatva stokam adhvanam ubhabhyam. sahitah. prabhuh. |natidure tatah. praha tis. t.ha vis.n. o mahabala ‖ 145 ‖sa pran. amya tatah. padau harasya purus.ottamah. |tasthav amram. samalambya sa kubjah. samapadyata ‖ 146 ‖yasmad amram. samalambya tasmin deśe sthito harih. |nirıks.aman. o deveśam. deśas tasmad abhud asau ‖ 147 ‖kubjamraka iti khyato vis.n. oh. ks.etram. samr.ddhimat |pun. yam. nivartanany as.t.au gosahasraphalapradam ‖ 148 ‖

567 Cf. Tırthank, 65.

184 Hans Bakker

Plate 21Rishikesh: Shyampur Garhi (Brahmavarta (in Rishikesh)?),

VBA (Bhadreśvara), and Bharat Mandir (Kubjamraka)

world of Brahma by dying there, after preparing a porridge andhaving fed it to a brahmin.568

We have not found a site in the vicinity of Bhadreśvara in Rishikesh thatcan be positively identified with Brahmavarta,569 but there is a distinctpossiblity that it corresponds with the early historical mound at ShyampurGarhi, c. 6 km west of VBA on the Golapani (Goila Nala), a small trib-utary of the Ganges. This site was first explored in 1974–75. The mound

568 SP 32.149–52:

natiduram. tato gatva bhuyo devah. pitamaham |nivartety abravıd vyasa gaganam. ca samaviśat ‖ 149 ‖tasmin viyadgate deve brahma pranjalir unmukhah. |pradaks. in. am. samavr. tya pran. amya prayayau tatah. ‖ 150 ‖yasmat tatra haram. tena kurvata vai pradaks. in. am |avartah. svaśarırasya prakr. tah. pun. yakarman. a |tasmat sa deśo vikhyato brahmavarteti śobhanah. ‖ 151 ‖aśvamedhaphalam. tatra snatah. prapnoti manavah. |sadhayitva carum. catra bhojayitva tatha dvijam |pran. an parityajya tato brahmalokam avapnuyat ‖ 152 ‖

569 Cf. MaS 2.17 (above, n. 519 on p. 168). Manu clearly defines Brahmavarta as aregion, partly or wholly coinciding with Kuruks.etra, and so the term is mostly used(cf. Kalidasa Meghaduta 48: brahmavartam. janapadam). SP 32.151–52, though itspecifies it as deśa, seems to describe a particular holy spot where one can takea bath. This is in keeping with MBh 3.81.43, MBh 3.82.38; cf. MtP 22.69 (v.l.brahmatırtha), BrP 25.39. The recommendation to die at this spot may point to atradition of (religious) suicide there.

The Myth of Daks.a’s Sacrifice 185

Plate 22Shyampur Garhi excavation site

measures 300 x 75 m. ‘The ceramics obtained from this site belong to theearly historical period similar to that found from phases I and II at Vira-bhadra.’570 The site was subject to further excavations for three seasons,from 2003–04 to 2005–6, by Manoj Kumar, curator at the ArchaeologicalMuseum in Hardwar. He reports the finding of Painted Grey Ware at adepth of 18 m (oral communication).

Kanakhala

After the epiphany in Bhadreśvara, Daks.a returns to the place wherehe performed the sacrifice, 20 km downstream, installs a linga there (notmentioned by name, but no doubt referred to in later sources as Daks.e-śvara), and worships Rudra (SP 32.157). Śiva grants that the site becomesholy as ‘Kanakhala.’571 He appoints Haribhadra to this place and gives

570 IAR 1974–75, 41.571 SP 32.188f. SPS 167.58, 172. See Bisschop 2006, 188ff. Cf. MBh 3.82.16, MBh

3.88.19, and MBh 13.26.12. MBh 3.135.5 refers to the Kanakhala (literally meaning‘little threshing floor’) as the ‘seers’ favourite trees’ (nagah. ), wrongly taken in itsmeaning ‘mountain’ by van Buitenen in his translation. SP 32.188f. gives a ratherforced nirukti. It is said that there are two daughters of a sage here. Since thesetwo blessed girls, after practising asceticism for a full year, have prepared food for

186 Hans Bakker

him Bhadrakalı as his wife.Present Kankhal is a village 5.3 km south-west of Gangadvara, near

the junction of the Ganga and Nıladhara. The trees at Kanakhala are saidto be of gold.572 Due to human behaviour these trees have now turned intowood (SPS 167.58), which was indeed the condition in which we saw them.

The Skandapuran. a further declares that the tears shed by Daks.a whenhe was allowed a vision of Śiva and the Goddess formed a holy rivercalled ‘Arjuna,’ said ‘to flow in manifold beds’ (parva parva samasadyavahis.yasi).573 This may refer to the complicated hydrological environmentin which the present Daks.eśvara temple is found. Once having passed theGangadvara, the Ganges River fans out in shifting riverbeds, nowadayscontroled by a sophisticated system of dams, dikes and canals. This untilrecently unstable situation may be one of the main reasons why no ancientremains have been found within the precincts of the modern Daks.eśvaraMandir. The sixth-century situation as described in the Skandapuran. acannot be further confirmed or elucidated by the contemporary state ofaffairs.

Conclusion

When the Paśupatas arrived in Hardwar/Rishikesh in the late fifth orsixth century, they found a region in which religious positions had longsince been staked out. Excavations at Virabhadra (VBA = Bhadreśvara)and Shyampur testify to settlements from at least the beginning of theCommon Era, if not earlier. Gangadvara itself belongs to the oldest layerof Hindu holy places and so does Kanakhala. Neither is of an intrinsicallyŚaiva nature, but an early Śaiva presence is attested by the Kus.an. a imagein the Bharat Mandir Museum. The importance of the Vais.n. ava presencein the region may be deduced from the fact that the composers of theSkandapuran. a thought it necessary to specify a Vis.n. uks.etra named Kubja-mraka, though Vais.n. ava holy sites are normally ignored in our text (e.g.not one Vais.n. ava place is mentioned in the detailed Varan. ası Mahatmya).

Two factors may be adduced to comprehend the process of Paśupatasettlement in this area. The first is political. The Hardwar region probablybelonged to the kingdom of the Vardhanas (Pus.yabhutis) of Thanesar inthe sixth century and, as we have already seen, this kingdom appearsto have functioned as a hub in the spread of the Paśupata movement.

him, with grains (kan. a) and a threshing-floor (khala), and since Daks.a is present(sthita) there, therefore this shall be the holy place (sthana) Kanakhala.

572 SPBh 74.2. The river near Kanakhala is here called Hairan. vatı, i.e. ‘Gold River.’573 SP 32.179–81.

The Myth of Daks.a’s Sacrifice 187

The other factor is of a cultural nature, viz. the power of the popularDaks.a myth, which so clearly defines the superiority of Saivism over otherforms of early Hinduism or Brahmanism. To be sure, this myth was notan invention of the Paśupatas themselves, but it provided them with anideological backing. They settled at the ancient site of Bhadreśvara (VBA)and probably founded a mat.ha there, next to the rebuilt or newly builtBhadreśvara temple. To underpin their claims, the myth of Śiva’s epiphanyand the initiation of the gods in the Paśupata vrata by an (accidental) bathin ashes was designed. Authority was given to it by inserting this mythinto the Daks.a Cycle and including it in the Skandapuran. a.

In the next chapter we will see another instance of how the composersof the Puran. a made use of existing mythology to further their aims. Forthis we move eastwards into the kingdom of Kanauj, 240 km south of thatcapital.

The Seven Brahmins, King Udayana,and Śveta’s Struggle with Death

d Kalanjara Hill d

The Legend of the Seven Brahmins

Historical evidence

Kalanjara, modern Kalinjar, is the name of a famous hill-fort and ancientarchaeological site on the northern spurs of the Vindhyas in Bundelkhand,95 km south-west of the ancient Kauśambı, on the borders of Madhya andUttar Pradesh.574

Plate 23Kalanjara Hill and surroundings

574 Although the hill is part of Madhya Pradesh, archaeologically speaking it fallsunder the State Archaeology of Uttar Pradesh (Lucknow).

189

190 Hans Bakker

The hill rises steeply c. 250 m above the plain. Its nearly unassailable for-tifications have given it its military reputation, won in numerous siegesfrom time immemorial till the beginning of the 19th century, when in 1812the fort was at last surrendered to Colonel Martindell.575

Plate 24Kalinjar Hill seen from the North

The best historical survey of the hill to date is still that by General Cun-ningham, who based himself on a description by Lieutenant F.C. Maiseyfrom 1848.576 The hill is replete with archaeological remains from theGupta period onwards and contains countless inscriptions and graffiti.577

The first historical attestations of this site are two seals in northernGupta characters found in Bhıt.a, one reading kalanjara, containing a lingaon a pedestal and a representation of the hill and a trident-axe, the otherreading kalanjarabhat.t.arakasya, containing a linga with an umbrella andtriśula on either side.578

575 For this history see General Cunningham’s Report of 1883–84 (ASI Reports XXI(1884–85), 20–29).

576 Description of the antiquities at Kalinjar by F.C. Maisey. Printed for the Govern-ment at the Baptist Mission Press, Calcutta 1848. Cunningham in ASI ReportsXXI of 1884–85, 20–45.

577 A recent survey by Rajendra Yadav counted ‘Eleven pillars and five pilasters [. . . ]at Kalanjar fort which can be assigned to Gupta period’ (Yadav 2009–10, 66). Seefurther below, n. 579 on p. 191.

578 John Marshall in ASI Annual Report 1911–12, 49. A seal-die found in exca-vations at Sanchankot, 220 km north of Kalanjara and 41 km east of Kanauj(26o 59’ 10” N and 80o 19’ 15” E), should be compared with these seals. This diereads: ‘kalanjara va(?)ta in northern Gupta characters of circa fourth century C.E.’

Seven Brahmins, King Udayana and Śveta’s Struggle with Death 191

k[a*]la[n*]jara k(a)lanjarabhat.t.arakasya

Plate 25Two seals found at Bhıt.a

Below this is a river (presumably the Ken) and what is usually taken tobe a representation of the hill, but which, according to Ellen Raven, mayinstead represent a clustering of caves of ascetics. The hill is full of suchcaves.

On top of the hill in the centre of the fort is the Kot.itırtha, on whosewestern bank stands the palace of Raja Aman Singh, which serves as thesite museum. It contains a store of images, among which Gupta piecessuch as a fifth-century ekamukhalinga (Plate 26).579 Other historical in-

(Thaplyal & Tewari 2006, 452). Its lettering and iconography resemble the Bhı-t.a Kalanjarabhat.t.araka seal. Thaplyal and Tewari argue, not fully convincingly,that, because it concerns a die and not a seal imprint, ‘there was a temple of thename Kalanjara at Sachankot’ [. . . ] and ‘that some devout resourceful person (orpersons) of the Gupta period stationed at Sachankot made a pilgrimage to theŚaiva shrine at Kalanjara Hill and was (or were) so much impressed and devotion-ally moved, that on return to native place, got built a shrine named after thatshrine’ (op. cit. p. 453f.), and, we could add, had a replica made of the seal of themother-temple, which is, however, a less plausible corollary in our view.

579 Doc. No. 545 (not listed by Yadav). Yadav writes: ‘Among the large collection ofsculptures housed in Aman Singh Palace, eight pieces have been identified as apart of door-jambs (dvaraśakhas) of Gupta temples (Doc. Nos. 86, 102, 116, 227,274, 661, 784 and 909)’ (Yadav 2009–10, 70). Further, Yadav notes among thefragments two candraśalas (Doc. Nos. 613, 666) and ‘a large number of rock-cutand independent sculptures at Kalanjar, some of which are śivalingas, mukha-lingas

192 Hans Bakker

Plate 265th-century Ekamukhalinga found on Kalanjara Hill

formation, already noted above (p. 66), is contained in the Barah Copper-plate Inscription (ad 836), which reports that the Pratihara king Bhoja Iof Kanauj informed the residents of the Kalanjara man. d. ala that he wasreinforcing an agrahara grant that originally had been made by the Para-meśvara Śrı Śarvavarmadeva, who can hardly be anyone other than thesixth-century Maukhari king of Kanauj of that name.580

The archaeological evidence thus proves that the hill was a well-knownreligious centre and had been associated with Saivism from at least Guptatimes. It contained a linga sanctuary that was important enough to bedepicted on seals that were issued from Kalanjara.

Literary evidenceThe holiness of the mountain is confirmed by literary evidence: the Maha-bharata refers to Kalanjara twice in the Tırthayatraparvan and once in alist of tırthas in 13.26.34. MBh 3.83.53–54 mentions Kalanjara as a sitewhere one should bathe in the Devahrada and, by accomplishing oneself(i.e. ending one’s life), one will rejoice in heaven.581 As we will see shortly,

(Doc. Nos. 946, 1094, 323), Karttikeya (Doc. No. 1069), Gan. eśa (Doc. Nos. 1080,1129) and Lakulıśa (Doc. No. 1086)’ (Yadav 2009–10, 74), all of which he ascribesto the Gupta period (‘5th/6th century’) and which, we assume, are housed in thesite museum, if they have been given a Doc. No.

580 EI 19 (1927–28), 17f. SI II, 233–35.581 MBh 3.83.53–54:

Seven Brahmins, King Udayana and Śveta’s Struggle with Death 193

suicide at Kalanjara is a recurring theme in other descriptions of the hillas well.

The seven deer of Kalanjara

Another story that features Kalanjara Hill is epitomized in a famous ślokawhich is widespread in Sanskrit literature. The story belongs to the contextof the pitr.kalpa, and illustrates the merits of devotion to the ancestors.The earliest version of the verse at issue, as has been demonstrated byYuko Yokochi (2000), is found in Harivam. śa 19.18. It is also found in theSkandapuran. a 57.39 and runs as follows:

In Daśarn. a seven hunters, deer on Mount Kalanjara, and cakravakabirds in Saridvıpa—of those (seven), you (three) have sunk down.582

The three singled out and addressed in the fourth pada are King Brahma-datta of Kampilya and his two ministers, Pancala and Kan.d. arıka (Bra-hmadhanvan in the SP version).583

This śloka is a key verse, epitomizing the legend of the seven brahminswho, after they had slaughtered and eaten a cow, go through four succes-sive births before reaching final emancipation. The seven who committedthe sin are said in the Harivam. śa to be the sons of Kauśika and pupils ofGargya living in Kuruks.etra (HV 16.3–5), whereas in the Skandapuran. athey are said to be the pupils of Kauśika Suparvan, living in Daśarn. a, theland around the city of Vidiśa, i.e. Eastern Malwa. After first having beenreborn as hunters in Daśarn. a, they are reborn again as deer on Kalanjara,where they eventually commit suicide by abstaining from food and drink,according to the Harivam. śa, or throwing themselves off a cliff, accordingto the Skandapuran. a,584 thus acting on a local practice, it seems. This

tatah. kalam. jaram. gatva parvatam. lokaviśrutam |tatra devahrade snatva gosahasraphalam. labhet ‖ 53 ‖atmanam. sadhayet tatra girau kalam. jare nr.pa |svargaloke mahıyeta naro nasty atra sam. śayah. ‖ 54 ‖

MBh 3.85.15 situates Agastya’s hermitage named Hiran. yabindu on this hill.582 For this verse see Yokochi 2000, 534–39 and 2013b ad SP 57.39 (SP III, 96f.). HV

19.18 = SP 57.39:sapta vyadha daśarn. es.u mr.gah. kalanjare girau |cakravakah. saridvıpe yuyam. tebhyo ’vasıdatha ‖

583 Saridvıpa is said to be a lake in the vicinity of the capital of south Pancala, Ka-mpilya/ Kampilya, which may be identified with modern Kampil on the Ganges(UP) (see Ghosh in EIA II, s.v. Kampil). The three who had ‘sunken down’ were intheir former lives the three cakravaka birds who had expressed a wish for sensualpleasures. The other four birds became masters of yoga after rebirth, and, to remindtheir three ‘fallen’ brothers of their former births, they had this verse sent to theking, while they themselves committed utkranti, yogic suicide (SP 57.28–32). Forthe whole story see SP III, chapters 56–57.

584 SP 56.88cd: sahitah. sahasa pran. an marutprapatanaj jahuh. . Cf. HV 16.25.

194 Hans Bakker

Plate 27Plan of Kalanjara Hill (Cunningham)

legend is first and foremost engaged with yoga (yogadharma) and ances-tors and has no obvious sectarian agenda, although the names Kauśika andGargya, as we have seen (p. 145), both feature prominently in the Paśupa-ta param. paras and in other evidence relating to this form of early Saivism.The composers of the Skandapuran. a did not alter the original purport ofthe legend, but they thought the story fit for inclusion by linking it to themyth cycle of the Goddess, through the figure of Brahmadhanvan, whois transformed into Parvatı’s tiger Somanandin.585 For the Śaiva asceticsliving on Mount Kalanjara, there appears to have been no hindrance toaccommodating the legend of the seven brahmins into their own socialand religious world. For this contention we have found material evidence.

Mr.gadhara

On the southern edge of the fort is a place called Mr.gadhara, or ‘Deer-spring.’ It consists of a small building, reconstructed with the help ofolder material, that houses a natural spring. Immediately to the west of

585 SP 57.102–04, 69.23–32. This matter is discussed in detail in Yokochi’s IntroductionSP III, 15–22.

Seven Brahmins, King Udayana and Śveta’s Struggle with Death 195

this building, on a promontory platform above the steep slope, is anothersmall water reservoir (Plate 28). The dressed stone above this reservoircontains an image of a carrier (see also below, p. 205). Next to it are foundthree short inscriptions or graffiti in fifth-century Gupta characters. Thisproves that the place has been in use since the fifth century, and, in view ofthe promontory on which it is situated it may occasionally have witnessedsome suicides, in line with the reputation of the Kalanjara Hill.586

Plate 28Kalanjara: Mr.gadhara promontory platform with graffiti and carrier

The first two graffiti have apparently little religious significance. The firstone runs: . . . ti-vis.ayapati[m. ] pran. amati, that is, ‘. . . makes an obeisanceto the head of the district of . . . .’587 The second one runs: adityad(a)sa[h. ]t(ri)śul(a)dityam pra(n. a)[ma](ti), that is, ‘Adityadasa makes an obeisanceto Triśuladitya.’ We have no further information about this Adityadasaand Triśuladitya. The latter may have been the governor of the fort oranother high official in the fifth century.588

586 The Paśupata practice of ending one’s life by throwing oneself off a cliff is attestedin the case of the Śrıparvata (see above, n. 437 on p. 142) and the hill-top Hariś-candra (SPBh 122.104–07), modern Hariścandragarh (see Bisschop 2006, 203). Formany ascetics it might have been an easy alternative to yogic suicide (utkranti).

587 This inscription is rendered by Krishna Kumar as vidisa gahapati pranamata inIAR 1976–77, 63.

588 This inscription is rendered by Krishna Kumar as Sri Vyasa matragunaditya–ppra-(nama)ta in IAR 1976–77, 63.

196 Hans Bakker

The third graffito is more signifcant (Plate 29). The reading of the lastword is uncertain, but we are confident that the following is not far offthe mark:

vaidiśaprativasina[h. ] ne[ni*](r)vva(m. )śala(kud. )ah.‘The Lakud. as, who are without family, coming from the neighbour-hood of Vidiśa.’589

vaidiśaprativasinah. nirvam. śalakud. ah.Plate 29

Kalanjara: Mr.gadhara graffito

The ‘neighbourhood of Vidiśa’ is the country of Daśarn. a, the originalhome, according to the Skandapuran. a, of the seven brahmins who wereborn as deer on Kalanjara. If we have read the last word correctly—theu and half of the k are visible and the contour of the damage to its rightsuggests a (cerebral) d. : ‘lakud. ah. ’—this would be the earliest attestationof a sectarian branch within the Paśupata movement called Lakud. as orLakulas, i.e. followers of Lagud. i (see above, p. 147). These ascetics andothers may have lived in caves, such as, for instance, the Sıtasej (‘Bedof Sıta’) on the northern slope, the walls of which are full of unexploredgraffiti and inscriptions (Plate 30).

The little building at Mr.gadhara next to the platform preserves theoriginal foundation stone of the spring. This stone shows a depiction ofseven deer (Plate 31).590 Iconographically or art-historically, these fineimages of seven deer are difficult to date with any precision, but combinedwith the adjacent 5th-century graffiti, a fifth or sixth-century date seemsplausible. Obviously the site was locally regarded as the place where theseven deer ‘threw themselves off the cliff.’591

589 This inscription is rendered by Krishna Kumar as Vedisa prativasana ne(tra)palain IAR 1976–77, 63.

590 The Kalanjaramahatmya (KM), which is presently being studied by Peter Bisschopand to whom I owe this information, refers to these deer: mr.gan. am. darśanam. kr. tvasnatva tatra jale śucih. | saptasagarakasnanat phalam. prapnoti manavah. ‖1.19 ‖ (seeBisschop 2013, 281 n. 6).

591 The relief shows a division within the group of seven, setting three deer apart.This may be accidental, but it could well indicate that the sculptor was aware ofthe continuation of the legend, in which three of the seven will deviate from theascending path (above, p. 193).

Seven Brahmins, King Udayana and Śveta’s Struggle with Death 197

Plate 30Kalanjara: Sıtasej Cave

When the composers of the Skandapuran. a adopted the legend of the sevenbrahmins from the Harivam. śa and inserted it into the Vindhyavasinı Cy-cle, they might have been familiar with this place. This may have beenone of the reasons why the legend was included in the Puran. a.592

592 Immediately to the east of the Mr.gadhara building are the remains of a pavilionwhich dates from Candella times. In front of this are more architectonic remains,among which two fragments containing the depiction of two and eight deer respec-tively, all facing right (the larger fragment is given the number: KF 262). The twofragments formed part of a bigger frieze. The number ten poses a problem, since wedo not know of any version of the legend in which the seven deer are augmented toten. However, since the left portion of the fragment with eight deer and the left andright portion of the fragment with two deer are broken off, it might well be that thefrieze originally contained more deer and that their number is insignificant. Eachof the depicted deer has a caption in post-twelfth-century script above its back.We can read the last seven of the larger fragment with certainty; the captions tothe remaining three remain illegible. The second and third captions of the seriesof eight deer read: ‘Unmukha’ and ‘Nityavitrasta.’ These and the following namescorrespond to the names of the deer in the Harivam. śa 16.22–23.

śubhena karman. a tena jata jatismara mr.gah. |trasodvegena sam. vigna ramye kalam. jare girau ‖ 22 ‖unmukho nityavitrastah. stabdhakarn. o vilocanah. |pan. d. ito ghasmaro nadı namabhis te ’bhavan mr.gah. ‖23 ‖‘Due to that meritorious behaviour they are born again on beautiful Mount Kala-njara as deer who possess a memory of their former births, (but) who are afflictedby trembling of fear: Unmukha, Nityavitrasta, Stabdhakarn. a, Vilocana, Pan. d. ita,Ghasmara, and Nadin.’

198 Hans Bakker

Plate 31Kalanjara: Mr.gadhara, image of the seven deer

Further proof of the presence of an early Lakud. a community on the hillis found in a relief (Plate 32), which is described by Yadav as follows:

A fragment of sculpture (Doc. No. 1086) measuring 0.31 m x 0.40 m x0.09 m showing an image of Lakulıśa, seated over a chaukı (asandı).He is shown with curly hair, urdhvaretas, plain round halo behindhis head and his right hand is raised up to the shoulder while leftone holding a dan. d. a which is obliterated. Two male ascetics hold-ing a beaded garland in their left hands are shown to the right ofLakulıśa. [They] seem to be his pupils, while the left portion is miss-ing. These ascetic figures might have been out of his four pupilsof Lakulıśa named Kauśik, Gargeya, Maitreya and Kaurus. A stripof scroll carved below is disrupted by a label inscription engravedbelow chaukı. This two line Brahmı inscription is palaeographicallyassignable to early fifth century ad.593

In view of the position of Lakulıśa vis-a-vis his pupils, we may classify thisimage as a Daks.in. amurti (see Bakker 2004c). He seems to be seated on

Harivam. śa MS D3 reads dıno instead of nadı in 23c, and so reads the caption ofthe last deer.

593 Yadav 2009–10, 75. The whereabouts of this important image are not given, buton account of the fact that it carries a Doc. No., we surmise that it is kept in RajaAman Singh Palace or Site Museum.

Seven Brahmins, King Udayana and Śveta’s Struggle with Death 199

Plate 32

Lakulıśa with disciples

an animal (tiger) skin, the head of which lies on his left thigh. The sacredthread over his left shoulder may be represented by a snake. Yadav haskindly sent me a photo of the image, which allows us to read the inscrip-tion, though the name of the deity remains tentative. The script could beslightly later, sixth century, which would make the image assignable to theMaukhari period. It concerns a donative inscription, recording the ‘(gift)of a Śuddhidevaka who is submitted to Bhadra.’594 The designation of thedeity as Bhadra suggests a connection with the sanctuary of Bhadreśvaraon the hill, which is attested in another inscription (see below, p. 201).

594 [1] bhaddrasya prasanna[2] su[śu]ddhidevakasyaFor prasanna meaning ‘submitted to’ see ad SP 29.145b, SP 31.17, and SP 32.163.I take the apparent ‘horizontal hooked medial a’ in the first and the ‘slant of themedial a’ on top of the second syllable both as accidental scratches in the stone,not belonging to the proper aks.aras. The second aks.ara shows gemination of dbefore r.

200 Hans Bakker

King Udayana

In spite of the lack of hard evidence, there can be little doubt that Kala-njara Hill became part of Hars.avardhana’s kingdom; it may have served,in addition to its religious function, as a military stronghold. In Hars.a’splay the Ratnavalı, Ruman.vat, minister of the legendary king Udayanaof Kauśambı, attacks the king of Kosala, who had entrenched himself ina vindhyadurga, a fort in the Vindhyas (see above, p. 103). Now there arenumerous forts in the Vindhyas and the fort of the play may well have beenimaginary from the start, but in his description of the conflict, in whichthe Lord of Kosala could only be vanquished after he had been forced toa sortie to give battle, Hars.a may have been thinking of the impregnablefortifications of Kalanjara, although, unlike the later versions of the Uda-yana legend discussed below, none of Hars.a’s plays connects Udayana withor mentions the Kalanjara Hill.

We find a different situation in the later tradition of King Udayana,as represented by Somadeva’s Kathasaritsagara. In this tradition the kingbecame explicitly connected to Kalanjara, as it is said to be the site wherehe and his two wives, in old age, terminated their earthly existence bythrowing themselves off the cliff.595

Thereupon the king, along with his queens Vasavadatta andPadmavatı, mounted an elephant and set off together with hisministers. When he left Kauśambı, citizens, including women,children and old men, followed him, lamenting and weeping. Afterhe had reassured them that Gopalaka would protect them andthey had returned reluctantly, that king of Vatsa went to MountKalanjara.

Having reached it, he climbed it, bowed before Vr.s.adhvaja andtook his life-long friend, the lute Ghos.avatı, in his hand; with histwo queens at either side, and joined by Yaugandharayan. a and theother ministers, he jumped off the cliff. As he was falling, a shiningchariot came along and brought that king and his train, radiantly,toward heaven.’596

There is an elephant path leading up the mountain. On this ramp, whichis lined by many weathered, sculptured votive reliefs, a pre-Candella in-scription can be found in the third of the seven gates, the Chandi Darwaza(Plates 27 & 33).

595 Cf. Ks.emendra’s Br.hatkathamanjarı XVIII, 31: gopalakam. niveśyaśu bhrubhange-natmanah. pade | kalanjarabhr.gutsange divam. vatseśvaro yayau ‖

596 KSS 16.1.79–84. Cf. Adaval 1970, 226ff.

Seven Brahmins, King Udayana and Śveta’s Struggle with Death 201

This inscription has been edited byPeter Bisschop (2013). It is datedto the ‘ninth century’ by Kielhorn,but may be a century earlier. Thepilaster fragment on which it is en-graved belongs to the Gupta pe-riod, in view of the beautifullysculptured kırtimukha topped by ahalf medallion containing a ham. sa,which crowns the inscription (Plate33).597 The situation resembles thepilaster at Deo-Baranark on whichJıvitagupta II had his inscription in-scribed (see above, p. 67, Plate 8).The Malinı verse that opens the in-scription introduces Udayana. Plate 33

Inscription on a pilasterat the Chandi Darwaza

There was a king in the lineage of the Pan.d. avas, Udayana by name;by that (king), as he was extremely devout, was caused to be built inthe past, the most excellent temple of this Lord of the universe, Bha-dreśvara—a temple made out of bricks, whose beauty is delightfuland whose banners are flying in the wind.598

The inscription goes on to praise King Udayana and the temple. TheBhadreśvara temple, as represented by the pilaster, is earlier than theinscription and may have been constructed in the fifth or sixth century.599

597 Bisschop (2013, 284) draws our attention to two very similar fifth-century pilastersfrom Bhıt.a and Kauśambı in the Allahabad Museum, which have a blank at thespot where the Chandi Darwaza inscription is engraved. See Chandra 1970, Plates122 and 200.

598 The first verse runs (Bisschop 2013, 286):

udayana iti raja yah. kule pan. d. avanam. ,sakalabhuvananathasyasya bhadreśvarasya |pavanalulitacihnam. ramyakantıs. t.akabhir,gr.havaram atibhaktya karitam. tena purvam ‖

Cf. Cunningham ASI XXI (1884–85), 40, Plate IX l; Kielhorn in JRAS 1905, 621;Shastri 1995 I, 158.

599 A seal found in Bhıt.a reads bhadreśvara (ASI Annual Report 1911–12, p. 50 (No17)). It shows a two-armed male figure seated on a pedestal. The objects in hishands are unclear. Above his head are the hoods of a Naga (‘foliage or flames’according to John Marshall). Cf. Bisschop 2013, 289 n. 46.

202 Hans Bakker

Its association with King Udayana came, it seems, at a later date, that is,post-Hars.avardhana.

Apart from this pilaster, so far no evidence pertaining specifically tothe Bhadreśvara temple has been recognized among the rich archaeologicalremains on the hill, but the caption to the Lakulıśa image (above, p. 199)may refer to it.

However, Rajendra Yadav describes and illustrates two pillars whichdepict:

Plate 34

Kalanjara: Pillar KF 251

River goddesses Ganga (KF 251) andYamuna (KF 252) at the bottom. Tothe left of the river Ganga, a four-armed seated Gan. eśa is also carvedon the bottom of the pillar [see Plate34]. The pillar KF 252 has depicteda seated devotee in anjali-mudra tothe right of River Yamuna and a la-bel inscription carved over the headof devotee in Brahmı character whichis palaeographically assignable to lateGupta period or around mid-sixthcentury ad.600

Styllistically, these two pillars come veryclose to the inscribed pilaster and they mayhave been part of the same Bhadreśvaratemple. Yadav (op. cit. 77) concludes that‘at least three temples [of the Gupta pe-riod] were present at Kalanjar, and two ofthem seem to be dedicated to Śiva and oneto Vis.n. u.’

It has been suggested by me in an earlierpublication (Bakker 1994, 12) that the al-leged founder of the Pan. d.uvam. śin dynastyof Śrıpura, who is named in two inscrip-

tions as Udayana, may have come from Bundelkhand.601 This suggestion

600 Yadav 2009–10, 69, Plates 3 & 4. These pillars were fixed ‘on either side of de-scending steps of western embankment [of the Kot.itırtha], now shifted to AmanSingh Palace.’

601 In the Arang Inscription of Bhavadeva Ran. akesarin and Nannaraja the exact re-lationship between this Udayana (in verse 16) and subsequent kings is uncertain,since the Arya verse 17 is lost, apart from the last five syllables featuring the word‘son’: . . . sya tanujanma (Kielhorn in JRAS 1905, 626; Shastri 1995 II, 97). TheSirpur Stone Inscription of Śivagupta Balarjuna declares: ‘There once was a king

Seven Brahmins, King Udayana and Śveta’s Struggle with Death 203

was based on the identification of this Udayana with the homonymous kingwho is the subject of the Chandi Darwaza inscription. This suggestion nowseems to me naive. The Kalanjara inscription and the two Pan. d.uvam. śininscriptions from Kosala refer to one and the same legendary Pan.d. ava kingUdayana of Vatsa, son of Śatanıka, descendant of Bharata through Pan.d.u,Arjuna, Abhimanyu, Parıks.it, Janamejaya [. . . ] and Śatanıka II.602

Thus the lore of King Udayana—believed to be a descendant of thePan.d. ava Arjuna—served, on the one hand, to provide the new rulers inŚrıpura, starting with Indrabala (see Bakker 1994, 11f.), with a venerablepedigree, and, on the other hand, to bestow respectable antiquity on theBhadreśvara temple on Mount Kalanjara. So who were the rulers of Kala-njara after Hars.avardhana that brought the Udayana lore to the hill?

The Kalacuris of UjjainApart from the Barah Copperplate Inscription, which, as we have notedabove (p. 192), testifies to the control of the Gurjara-Pratiharas of Ka-nauj over the hill in the ninth century, the three centuries between thedeath of Hars.avardhana and the conquest of Kalanjara by the Candellaking Yaśovarman, reported in the Khajuraho Stone Inscription of Yaśo-varman, vs 1011 (= ad 954),603 are, as far as the historiography of Kala-njara is concerned, at present little more than a blank. It is unclear fromwhom the Candellas conquered the mountain, but one possible candidate,a local branch of the Kalacuris, remains unmentioned in Trivedi’s discus-sion of the matter.604 There are still numerous unexplored epigraphs andgraffiti on the mountain that date from before the Candella period, andtheir study may eventually solve the problem. This study, though a greatdesideratum, has to be postponed, since it would go beyond the scope ofthis book. In the present context we will limit ourselves to pointing outsome evidence that is indicative of Kalacuri possession of the mountainfor, at least, a certain period within the three centuries mentioned.

The later Kalacuri king Sod. hadeva reports in his Kahla Plates, vs1135 (= ad 1077) that his ancestor, the elder bother of Laks.man. araja,conquered Kalanjara.

Udayana by name who belonged to the Lunar Dynasty; from him arose the powerfulKing Indrabala who equalled Indra’ (asıd udayano nama nr.pah. śaśadharanvitah. |abhud balabhida tulyas tasmad indrabalo balı ‖ 2 ‖, Shastri 1995 II, 150).

602 ViP 4.21.3, 14–15. Cf. HCI I, 323–26, HCI II, 9–11. See Hars.a’s Ratnavalı (Cap-peller/Bohtlingk (ed.) p. 365), in which Babhravya speaks to King Udayana: hamaharaja, kim idam akaran. a eva bharatakulam. sam. śayatulam aropitam (‘Ah, greatking, why putting the House of Bharata in jeopardy for just a trifle’); cf. Adaval1970, 19ff.

603 CII VII.3, p. 345 vs. 31. The first half of the verse is partly illegible, but it is clearlystated that the mountain of Kalanjara was the dwelling place of Nılakan. t.ha.

604 CII VII.1, 89–91, VII.3 p. 341f.

204 Hans Bakker

In the dynasty descending from him [i.e. Arjuna Kartavırya] therewas [a king] whose pair of lotus-feet was scattering dust/pollen overthe heads of the kings that were lined up before him in obeisance;who was a forest fire to the withered woods that was the tremblingarmy of his enemies, and by whose conquest of Ayomukha, as a(first) step, the reduction of Kalanjara was accomplished.

After he had conquered his enemies, (this) ornament of the Kala-curi gave the kingdom to his younger brother; he, the illustriousLaks.man. araja, conquered Śveta’s Station (Śvetapada) again.605

As we have seen in the first part of this study (p. 104), the Kalacuris ofMahis.matı had been squeezed between the two emerging powers of theseventh century, those of Hars.avardhana in the North and the Calukyaking Mangaleśa in the South. They were stauch Paśupata followers ontheir own account. Śam. karagan. a, in the charter he issued from Ujjainin ad 597, described himself as a paramamaheśvara and declared thathis father, Kr.s.n. araja, had been a devotee of Paśupati all his life.606 Hisson Buddharaja issued a charter on behalf of his queen Anantamahayı,who is said to have also been a Paśupata (paśupatarajnı).607 The king’s

605 CII IV.2, 386 vss. 5–6:tatah. prabhr. ti sam. tate pran. atarajarajıśiraś-caraccaran. apankajadvitayaren. ur asıt kule |ayomukhajayakramakraman. asiddhakalanjarah. ,sphuratparavaruthinıjaradaran. yadavanalah. ‖ 5 ‖kalacuritilakah. śatrun jitva rajyam. dadau laghubhratuh. |sa śrılaks.man. arajah. śvetapadam. yah. punar jitavan ‖ 6 ‖

The locality of Ayomukha is unknown. A mountain of that name, said to be ‘richin ore,’ is mentioned in Ram. 4.40.13, HV App. 1 No. 42a l. 448, MtP 163.71. Seealso the discussion in CII IV.2, 385. śvetapada in pada d of v. 6 is a pun. Its pri-mary meaning is that Laks.man. araja attained Śvetas Station, namely immortality,which actually means that he died (cf. below, p. 211). The word punar in thesame pada announces the double entendre by alluding to the mentioned Kalanjaraconquered by his elder brother, the place where Śveta struggled with Kala andattained immortality (see below). For this interpretation I am obliged to YukoYokochi. I interpret the compound in 5c differently from Mirashi (op. cit., 383),who paraphrases: ‘who established himself in Kalanjara, from where he graduallyoverran Ayomukha’ (cf. Kielhorn in EI VII, 86: ‘who by conquering Ayomukhaand subduing the Krathas (reading kratha◦ instead of krama◦) possessed himselfof Kalanjara.’Mirashi argued that the elder brother of Laks.man. araja was Vamaraja who, ac-cording to him, is mentioned as the founder of the Kalacuris of Tripurı in theirlater inscriptions (CII IV.1, lxviii–lxx.). However, see Sircar in EI 30, 46f. andSanderson 2009, 272 n. 649, who argue that the Vamadeva mentioned was a Śaivaguru rather than a king. If our interpretation of these two verses is correct, it isLaks.man. araja’s elder brother who conquered Kalanjara.

606 CII IV, 41 l. 4, 16; Schmiedchen 2013, 351. (see above, n. 247 on p. 86).607 CII IV, 50 l. 32. Above, p. 104.

Seven Brahmins, King Udayana and Śveta’s Struggle with Death 205

name Śam. karagan. a may itself have been a Śaiva initiation name endingin gan. a.608

We have argued that Buddharaja may initially still have controlledUjjain, from where his father Śam. karagan. a had issued the charter, butafter ad 610 this branch of the Kalacuris falls silent. The hypothesis thatwe would like to venture is that their descendants, Laks.man. araja andhis elder brother, settled on Kalanjara after Hars.a’s kingdom had fallenapart.609

As the former rulers of Ujjain, these Kalacuris naturally took thatcity’s traditional lore with them—the story of the legendary princess ofUjjayinı, Vasavadatta, who became the queen of Vatsa (Kauśambı). If thishypothesis is correct, it will follow that it might have been these Kalacuriswho brought the legend of Udayana to the hill.610 To this hypothesis aspeculation may be added.

Udayana, Vasanta, and the carrierAll over the Kalanjara Hill we find carved into the rocks an image of afigure carrying two buckets on a yoke (viham. gika or ‘bangy’), but none ofwhose iconographic details are exactly the same.611 The latter applies inparticular to the figure’s right arm and the object it holds.612 An instance

608 Sanderson 2009, 291f.609 This hypothesis was earlier advanced by V.V. Mirashi. Cf. Fleet 1882, 57ff. Unlike

Mirashi, we do not believe that: ‘In the beginning of the sixth century A.C., it(i.e. Kalanjara) was in the occupation of Udayana of the Somavam. śı dynasty’ (CIIIV.1, lxviii; cf. Krishna Deva in EITA II.2, 90).

610 Cf. Kalidasa’s Meghaduta 30: prapyavantın udayanakathakovidagramavr.ddhan, pu-rvoddis. t.am anusara purım. śrıviśalam. viśalam | svalpıbhute sucaritaphale svargin. am.gam. gatanam. , śes.aih. pun. yair hr. tam iva divah. kantimat khan. d. am ekam ‖ 30 ‖

611 The viham. gika (Amarakośa 2.10.29f.) is an attribute associated with brahminicalactivity from the earliest visual representations in Sanchi (eastern gate of theGreat Stupa, in situ) and on a vedika coming from the Bharhut stupa (north-eastquadrant), photographed by Beglar in situ, ad 1874 (British Library, Shelfmark:Photo 1003/(1502)). The BL on-line catalogue describes this scene as: ‘Illustrationsfrom the Cammasat.aka-jataka which narrate the story of a pilgrim who arrives atan arena where rams were fighting.’ This ‘pilgrim’ is described as a paribbajaka(Skt. parivrajaka, i.e. wandering brahmin mendicant), clad in a skin garment (ca-mmasat.aka) (Jatakas IV.4 (no. 324)). Buddhaghos.a in his commentary on the Bra-hmajalasutra 1 (PTS D. I.1.1 p. 41) associates the parivrajaka ascetics with a yoke(kacadan. d. aka) too, while they are carrying all sorts of junk, such as stools, tridents,peacock’s tails, or pots, etc.—in contrast with Buddhist mendicants, who wanderaround without dragging all these paraphernalia with them (I am grateful to PhyllisGranoff for this reference). In SPBh 119.78 Indra appears to the seven seers in theguise of a parivrajaka. He is accompanied by a dog and ridiculed by the seersbecause of his fatness (SPBh 119.80–82). For the close relationship between theparivrajaka and the Paśupata ascetic see Oberlies 2000, 176 n. 6.

612 In the specimens of this image found at Mr.gadhara, Balkandeswara, and KhamborShiva (Plates 28, 35, 38), it seems that the right hand holds a kalaśa or gourd from

206 Hans Bakker

Plate 35Kalanjara: Balkandeswara, linga and carrier

with inscription of Vasanta

of this image is found on the promontory platform at Mr.gadhara, as wehave seen above (Plate 28). Another similar figure is found below theelephant path, halfway up the northern slope of the mountain, at a placethat according to Cunningham is called ‘Balkandeswara.’613 There is asmall rebuilt temple, around which some ascetics are living.614

samadhigata-pancamaha(ś)abdas(a)manta-śrıvasantah.Plate 36

Kalanjara: Balkandeswara, Inscription of Vasanta

On the platform in front of this temple, hewn in a rock boulder, is anekamukhalinga and a carrier next to it (Plate 35), above which is an

which a stalk rises that curls over the right shoulder of the carrier. The figure atthe Gangasagara tank (Plate 37) holds a straight stick of which the top portion isbroken off and which has a knob at the bottom.

613 ASI Reports Vol. XXI (1884–85), 41.614 It seems that this Śiva temple is called ‘Vanakhan. d. eśvara’ in the Tırthank p. 124.

Seven Brahmins, King Udayana and Śveta’s Struggle with Death 207

inscription reading (Plate 36): samadhigata-pancamaha(ś)abdas(a)manta-śrıvasantah. , ‘The illustrious Minister/Vassal Vasanta, who has obtainedthe five great titles.’615

Below Balkandeswara, near the Gangasagara tank at the foot of the north-ern slope of the hill, is a free-standing image of this carrier, which oncehad his own shrine (Plate 37).

[1]samadhigatapa[2]ncamaha(ś)abda . . .[3]ntaśrıvasantah.

Plate 37Carrier image near the Gangasagara tank with inscription

This figure is clearly furnished with Śaiva paraphernalia such as a braceletand necklace made of rudraks.a and matted hair. In one of the bucketsflowers are visible, apparantly covering the water that we suppose to bein it, and in his right hand he carries a stick or khat.vanga (?), broken offat the top. To the left of his head is an inscription which, as far as it islegible, reads the same as the Balkandeswara inscription: samadhigatapa-ncamaha(ś)abda . . . ntaśrıvasantah. .616

615 See also ASI Report XXI (1884–85), Plate IX Q. For this formula and the panca-mahaśabda see Sircar 1965, 341f. and Sircar 1966, 288; Salomon & Willis 1990, 148,153; below, n. 632 on p. 213. For samanta see Gopal 1963; Mirnig 2013, 329–31.

616 See also ASI Report XXI (1884–85), Plate IX P for a rubbing.

208 Hans Bakker

Above Balkandeswara, on the northern slope, at a site called ‘KhambhorShiva’ (Hi.: kham. bhar śiva, Śiva of Concern’), we find another carrier fig-ure in the facade of the rock above a panel that features seven mothers,dancing on small pedestals to which their respective vahanas are attachedon the proper left, flanked by Vırabhadra (dancing on the bull) and Ga-n. eśa, to the right of which is a linga worshipped by two figures on eitherside. To the right of the carrier the contours of another figure are visible,the carving of which remained unfinished.

Plate 38Kalanjara: Rock facade at Khambhor Shiva

Top left of the carrier, in a niche in this facade, is another, more lengthyinscription, which features the name ‘Vasanta’ again.617 To judge by thepalaeography of the three inscriptions (at Balkandeswara, Gangasagara,and Khambhor Shiva), these three groups may all be assigned to the ninthcentury.

There are still more images of this carrier on the hill, which we have toleave out of account here. They deserve a study of their own. We have no

617 Information kindly provided by Peter Bisschop. IAR 1976–77, 63 reports yet an-other inscription in the ‘rock-bed of a cave’ known as ‘Bhagwana Sej,’ below theSıtasej, which reads according to Krishna Kumar: ‘om samadhigata pancha maha-sabda mahasamanta sri Vasantaka.’

Seven Brahmins, King Udayana and Śveta’s Struggle with Death 209

information about the meaning of this figure. According to Cunninghamhe is generally called ‘Sarwan (Skt. saravaha?) Baba,’ and the Generaldiscusses extensively the mythology of this figure, which, however, to usseems to be spurious.618 As to the identity of the accomplished samantaVasanta, we are also groping in the dark. Vasanta and the carrier belongtogether, that much is clear, but their exact relationship is less clear. Themost natural, but certainly not the only possible, interpretation wouldassume the carrier to be Vasanta; but who is he?

We have no information concerning a historical feudatory or ministernamed Vasanta, ‘Spring,’ and the question of whether we are concernedwith a historical person at all seems justified.619 As we have seen above,the inscription in the Chandi Darwaza—about one century earlier than theVasanta inscriptions—ascribes the construction of the Bhadreśvara templeto the legendary king Udayana. It may be more than just coincidence thatthe confidential friend and minister of this king, who acts as vidus.aka inthe Udayana plays, is called Vasantaka, ‘Minister Spring.’620

The Story of Śveta

The above reconstruction indicates an influx of new ideas in Kalanjaraafter the fall of Hars.a’s empire, and this brings us back to the Puran. as andthe Śaiva tradition, in this case to the myth cycle of Śveta. Śveta’s mythis told in Kurmapuran. a 2.35.11–38. Here Śveta is said to be a rajars. i whoworships Śiva devoutly on Mount Kalanjara. When death (kala) threatensto take him, he prays to Rudra, who comes to his rescue and beats (jari-tavan) Kala.621 Śiva grants that the king can become his gan. apati and, at

618 Cunningham in ASI Report XXI (1884–85), pp. 41–45.619 The Saundaryalaharı 6 calls ‘Spring’ (vasanta) the Samanta of Ananga (Kama).

The flowers carried in the bucket also seem to point to the Spring season.620 In the 3rd act of the Pratijnayaugandharayan. a the three ministers of Udayana

secretly meet in Ujjain to prepare the escape of their lord. Vasantaka is said toenter the stage in the guise of a D. in. d. ika (d. in. d. ikaves.o vidus.akah. ), who is looking forhis moda-mal.l.a (Skt. modaka-mallaka or ‘vessel with sweets’), a metaphor for KingUdayana. He is called a bahman. ausa (Skt. brahman. opasaka, a brahmin layman)by Ruman. vat, who is disguised as a śraman. aka. Although the viham. gika of ourcarrier is obviously not a modaka-mallaka, a Din. d. ika/D. in. d. ika is sometimes takento be a roguish Śaiva ascetic associated with the Paśupatas (Schokker 1966, 143–45). Śiva’s incarnation in the Treta yuga is called Din.d. imun. d. a (SPS 167.116 andBisschop 2006 ad loc.). A Gan. apa Din. d. i is described in SP 23.6–9.

621 This nirukti is found in KuP 2.35.11: kalanjaramahatırtham. loke rudro maheśvarah. |kalam. jaritavan devo yatra bhaktapriyo harah. ‖ 11 ‖ It endorses the connectionbetween the Śveta myth and Kala-m. -jara (see below, p. 211). The nirukti mayhave been implied in MBh 12.238.9 (see Nılakan. t.ha’s commentary ad loc.; Bisschop2013, 280 n. 1).

210 Hans Bakker

Brahma’s request, Śiva brings Kala back to life.The myth of Śveta also occurs in the Skandapuran. a, but here the

written tradition is more complex.

Śveta in the Skandapuran. aIn addition to the reference in the story of the seven deer, Kalanjarafeatures in the Skandapuran. a in a list of Śaiva ayatanas in chapter 167:‘One who has seen the linga on Mount Kalanjara, the dwelling of Umapati,he obtains the result of an Aśvamedha sacrifice and reaches the Rudra-loka.’622 Contrary to what is usual, this linga is not given a name. We couldhave expected ‘Bhadreśvara,’ in accordance with the inscription and seal,if a linga of that name had featured in the Skandapuran. a in connectionwith Kalanjara, which is, however, not the case.623

On the other hand, it is a distinct possibility that in SPS 167.87 aKrodheśvara linga is meant, the manifestation of Śiva that features inthe Śveta myth as it occurs in the S recension of Skandapuran. a 166. Inthe story as told in this recension,624 Śiva’s intervention on behalf of hisdevotee (bhakta), the muni Śveta, son of Kuśadhvaja and grandson ofDarbhadhvaja, is said to have taken place on a mountain in the Vindhyas,where he became known as Krodheśvara. Echoing the Mahabharata andother sources, the place is recommended for suicide.

Because I was angry (kruddha) with the Son of the Sun (i.e. Ka-la/Yama)—this (spot) on this hill became most holy as a result—and since I shall continue to reside here on the top of the foremost ofmountains, therefore it shall be (known) in this (world) as Krodhe-śvara.

People who, freed from delusion, abandon their body on (this)holy mountain, at Krodheśvara in the Vindhyas, which is visited by

622 SPS 167.87 (Bisschop 2006, 100): girau kalanjare lingam. dr.s. t.va sthanam umapateh. |aśvamedhaphalam. prapya rudralokam avapnuyat ‖

623 As we have seen above (p. 174), the Skandapuran. a knows of an abode of Śiva called‘Bhadreśvara’ (SP 32.134) near Gangadvara in Rishikesh.

624 This story runs in brief as follows. Śveta has a sister, whom their father refusesto give away to a Danava Vipr.thu, since she was to be betrothed to Vis.n. u. Vi-pr.thu kills the father as well as Darbhadhvaja. Brother and sister both resort toasceticism. Śveta is said to go first to the bank of the Narmada (SPBh 166.43).When he sees Yama’s messengers coming, he takes refuge with Śam. kara, who issaid to live on a mountain top (girındraśikharasthita) (SPBh 166.49). After Nandı-śvara, who was sent by Śiva, fails to persuade Yama to set Śveta free, they decideall together to leave the spot where this dispute took place and to resort to amountain in the Vindhyas (vindhyaparvata) in order to see Śiva himself (SPBh

166.84). There Yama refuses again to obey Śiva, when he orders him to deliverŚveta. Enraged Śiva burns him with his third eye (cf. SPBh 155.36; Yokochi 2004a,107).

Seven Brahmins, King Udayana and Śveta’s Struggle with Death 211

immortals and holy men (siddhas), their destination is declared byme to be the same as that of Śveta.625

The Skandapuran. a continues by specifying a (linga) known as Śveteśvaraand a river (nadı) called the Śvetaganga (SPBh 166.110). All these holyplaces are as yet unknown from other sources. The nirukti of Kalanjara(see below) is unknown to the S recension, and therefore it remains uncer-tain whether by Krodheśvara the Kalanjara mountain is meant.626 Thisuncertainty can only be resolved if we find independent evidence, textualor archaeological, for a sixth/seventh-century Krodheśvara sanctuary onthe Kalanjara mountain, but the uncertainty is alleviated to some extentby the RA recension.

SPRA 167.5.3–5 (Bisschop 2006, 154) introduces Śveta, whose tormentby death is stopped by Hara, and with reference to this, although thepassage is very corrupt, it gives the following nirukti:

There, on that hill, the Three-eyed God has beaten (jaritavan)Kala (death); for that reason that most excellent mountain is calledKalanjara.627

The connection between Krodheśvara and Kalanjara is also strongly sug-gested by Skandapuran. a chapter 166 according to the RA recension. Thischapter, which tells the myth of Śveta as given above, is extended withfour more chapters in the RA recension. This recension starts from thestory of Śveta corresponding with the S recension (though the wording is

625 SPBh 166.108–09:kruddhaś ca yasmad aham arkasunor, asmin girau pun. yatamas tato ’yam |vatsyami cehacalarajamurdhni, krodheśvaras tena bhavis.yatıha ‖ 108 ‖vindhyapradeśe ’marasiddhajus.t.e, krodheśvare pun. yagirau pumam. sah. |tyaks.yanti ye deham apetamohas, tes. am. gatih. śvetasama smr.ta me ‖ 109 ‖

626 Cf. LiP 1.30 and VDhP 1.236, in both of which the Śveta myth is apparently notconnected to Kalanjara.

627 SPRA 167.5.5 (cf. KuP 2.35.11 in n. 621 on p. 209):kalam. jaritavam. s tatra tasmin chaile trilocanah. |tena kalanjara iti sa khyatah. parvatottamah. ‖ 5 ‖

The problem with this passage is that the same nirukti appears twice, also in theimmediately preceding SPRA 167.5.1: narmadaprabhavo yo ’sti girih. kr.s.n. aśilata-lah. | yatra kalam. jaritavan † kalanvitasya † śam. karah. ‖ 1 ‖ (Bisschop conjectureskalanvıtasya metri causa, but it does not make sense either). The mountain inthis verse is said to be the source of the Narmada, which cannot possibly refer toKalanjara, since this source is generally known to be Mount Amarakan. t.aka. Weassume that the duplication in SPRA 167.5.1–5 is caused by a confusion and thatthe second nirukti in SPRA 167.5.5 is the correct one (see also Bisschop ad loc.cit.). The confusion may have been caused by a wrong reading of the Śveta mythas told in SPBh 166, in which Śveta, after his father has been killed, first practisestapas on the banks of the Narmada and then resorts to a mountain in the Vindhya(see above, n. 624 on p. 210).

212 Hans Bakker

different), including Krodheśvara etc. (SPBh 166.34–121). It departs fromthe S recension by explaining the way Śveta worships Śiva, viz. by danc-ing, singing and laughing, in other words by behaving according to thePaśupata mode of worship. There is a detailed description of the fightbetween Śiva and Death, and the nirukti of Kalanjara is given after Kalahas been burnt by Śiva’s bellowing (hum. kara).628

In sum, the RA recension ties the Śveta myth, more firmly than S,to the Paśupata tradition on the one hand, and Mount Kalanjara on theother. In this context it recognizes Krodheśvara and thereby reinforcesour assumption that the myth of Śveta and Krodheśvara as told in the Srecension is also related to Mount Kalanjara.

The incongruity of our sources concerning Kalanjara in the period un-der discussion would be resolved, if we could assume that the (Gupta)temple known from the Chandi Darwaza inscription and the Bhıt.a sealas ‘Bhadreśvara’ became known to the authors of the Skandapuran. a asKrodheśvara on account of the Śveta myth. This is not an entirely im-possible assumption. Admittedly, the semantics of bhadra (‘friendly’) andkrodha (‘anger’) are opposed, but in Śaiva mythology, rather than exclud-ing, this may prompt their juxtaposition or euphemistic exchange. Theembodied anger (krodha) of Śiva in the Daks.a myth is called ‘Haribhadra’or ‘Vırabhadra’ and that of Devı ‘Bhadrakalı.’629 The place where Śivaand Parvatı were hiding in order to watch the destruction of Daks.a’s sac-rifice by these two ‘Bhadras’ is called Bhadreśvara.630

Whatever might have been the case, Mount Kalanjara became thor-oughly accommodated to the Śaiva world in the three centuries betweenthe beginning of the Skandapuran. a composition and the new redactionthat created the RA recension.631 We surmise that the Kalacuris played

628 I am basing myself here on Torzsok 2001. The place where this fight took place iscalled Śvetaran. ya. The gods, led by Brahma, praise Śiva there, which leads to theresurrection of Kala/Yama.

629 SP 32.11–19. Especially SP 32.11 which describes the birth of Bhadrakalı: lala-t.e bhr.kut.ım. kr. tva tato devy ayateks.an. a | krodhat karen. a nasagram. sam. mamardaśucismita ‖ 11 ‖

630 SP 32.135 (above, p. 176).631 There is another myth connected with Kalanjara in the RA recension. In this

recension it is explained why Śiva wears the crescent. Though this story is possiblymentioned in the Anukraman. ika in SP 2.23a (bhasmasomodbhava), the myth itselfis not found in the S recension. In RA (A2 ff. 406r–407v, A4 ff. 212r–213r, R ff. 240v–241v) it is told that after the death of Satı, Śiva’s tejas is scorching the earth. Thegods approach him with one pot filled with amr.ta and another with poison. Śivadraws a crescent of amr.ta with his fingernail on his head. Thanks to this moon(soma/amr.ta) cresent, his tejas is cooled down (tejo me mr.dutam. jatam). Whenhe takes the poison his throat turns dark (nıla). This myth is located on MountKalanjara (Torzsok 2001). Further research is needed to investigate whether themyth-makers of RA were thinking of the origin of the Nılakan. t.ha complex on the

Seven Brahmins, King Udayana and Śveta’s Struggle with Death 213

a prominent role in this.

The Kalacuris of KalanjaraThe Kalacuris of Mahis.matı had been Paśupatas, but in the centuriesunder discussion Saivism developed into new forms. The later Kalacurisof Tripurı are known in particular for their patronage not of the Paśupata,but of the Śaiva Siddhanta.

Śaivism became the paramount cult throughout the extensive do-minion of the Kalachuris from the 8th to the 12th century A.C. Itreceived a great fillip during the reign of Yuvarajadeva I, who, un-der the influence of his queen Nohala, invited several Acaryas of theMattamayura clan to the Chedi Country and built magnificent tem-ples of Śiva and monasteries for them in Gurgi, Masaun, Chandrehe,Bilhari, Bhera-Ghat., and other places.632

This clan of acaryas came from Mattamayura, identified with the village ofKadwaha in the Ashok Nagar District of Madhya Pradesh, c. 270 km westof Kalanjara, 145 km north of Vidiśa in Eastern Malwa—in other words,it came from the area under the control of the Kalacuris at the beginningof the seventh century.633 It became one of the main branches of the ŚaivaSiddhanta. This clan, or its precursor, may have introduced the new formof Tantric Saivism to Kalanjara. The hill’s significance within the earlyŚaiva Siddhanta appears from the Guhyasutra of the Niśvasatattvasam. -hita (1.33), which mentions Kalanjara as one of the great Siddhaks.etras,as well as from the following verse in the Rauravasutrasam. graha:

I pay obeisance to Sadaśiva, the one who creates and knows time,the one who is master of time, Mahakala, who lives on Kalanjara.634

hill, which, as we will show (below, p. 215), has been a major religious centre sinceGupta times, developing into something spectacular in the Kalacuri period.

632 Mirashi in CII IV.1, cli. Cf. Introduction of SP II A, p. 63; Sanderson 2009, 263ff.Sanskrit-Kannada inscriptions of the 12th century feature also chiefs of the Kala-curi clan (see Fleet 1882, 57ff.). The inscription from the Pran. aveśvara Temple inTal.agunda (ad 1157) contains the following phrase: samadhigatapancamahaśabda-mahaman. d. aleśvara-kalanjarapuravaradhıśvara◦ [. . . ] ◦kal.acuryakul.a

◦ [. . . ] śrımad-bhujabal.acakravarti-bijjan. adevarasara prabhavapratapam (from the hand-out pre-sented by Florinda De Simini, Paris 27-3-2014).

633 See Sears 2004, 2008.634 Rauravasutrasam. graha, upodghata 7:

kaleśvaram. mahakalam. kalanjaranivasinam |kalakr. tkalavettaram. pran. ato ’smi sadaśivam ‖

Bisschop 2013, 282 notes: ‘The name Kalanjara is also included in the so-calledpancas. t.aka lists, [. . . ] which are found in many tantric (mainly Śaiva Siddha-nta) sources, but also already in the pre-tantric Śivadharmaśastra, a work of layŚaivism.’ See above, p. 9.

214 Hans Bakker

The twenty-eight incarnationsA list of twenty-eight avataras of Śiva is found in several Puran. as, thoughnot in the Skandapuran. a. Bisschop has argued convincingly that this list,including the one that is contained in the Vayupuran. a 23, postdates thecomposition of the Skandapuran. a.635 The twenty-eight incarnations beginwith Śveta and end with the last two gurus, Someśa/Somaśarman andLakulin/Lakulıśa. However, Śveta figures twice. The first one is said tohave descended on the Chagala mountain in the Himalaya. Another Śveta,the 23rd incarnation, is connected to the struggle with Kala, and this mythmay have prompted his assignment to Kalanjara, endorsed by the newlyinvented etymology (nirukti): kalam. jaritavan.

An early historical, though indirect, attestation of the concept of mul-tiple incarnations of Śiva is the Junvanı Copperplate Inscription of Maha-śivagupta (year 57), which belongs to the Mausala/Soma Siddhanta tradi-tion (above, p. 143). This inscription informs us that: ‘The embodimentsof Gahaneśa are the sixty-six Rudras, who bestow grace (initiation) andwho roam about in each yuga.’636 The concept of sixty-six manifestationsis apparently only known to the Jayadrathayamala, which text dividesthem into two lines of gurus (gurupanktis), a set of twenty-eight Śivas andone of thirty-eight Rudras.637

The idea of twenty-eight avataras of Śiva in twenty-eight kalpas seemsto have developed parallel to the concept of sixty-six Rudras, but, unlikethe latter, it was a laukika represention without sectarian agenda andtherefore became widely accepted. It took its cue from and supplanted thePaśupata idea found in the Skandapuran. a, viz. that there had been fourincarnations of Śiva in Karohan. a in the four yugas (of the present kalpa):Bharabhuti, Din. d. imun. d. a, As.ad. hi and Lagud. i, each with four pupils.638

New ideas like these may have been introduced into Kalanjara by theKalacuris and their contemporaries in the second half of the seventh oreighth century, along with forms of institutionalized Saivism. As we haveseen above (n. 605 on p. 204), the later Kalacuris of Tripurı referred, equiv-ocally, to the Mount Kalanjara conquered by their ancestors as ‘Śveta’sStation’ (Śvetapada).

635 For a detailed discussion of this list see Bisschop 2006, 41ff. Śveta’s descent ismentioned, e.g. in VaPAA 23.202–205 (= VaPVenk 23.191–93).

636 Bakker 2000, 8 (ll. 14–16): gahaneśasya murtayo rudrah. s.at.s.as.t.y anugrahaka yugeyuge parivartamanah. .

637 Bakker 2000, 11. The Jayadrathayamala belongs to the Vidya Corpus of TantricSaivism and is devoted to the cults of Goddesses (Sanderson 2009, 45f.). It speaksof sixty-six ‘Bhavas,’ who are associated with different levels of Śaiva teaching andwho are divided into two sets. The first set of twenty-eight Śivas is said to bestowboth exegesis of the scriptures and, occasionally, initiation (text and informationkindly provided by Alexis Sanderson).

638 SPS 167.110–18 and SPS 167.124–29.

Seven Brahmins, King Udayana and Śveta’s Struggle with Death 215

In keeping with other Puran. as, such as the Kurmapuran. a (quotedabove, n. 621 on p. 209), the redactors of the RA recension in the ninth cen-tury, when they elaborated SP 166, naturally combined the Śveta mythol-ogy as found in the S recension with the nirukti that by then had becomestandard, and in this way removed an obscurity from the first redactionof the text. They preserved the conception of the four incarnations in Ka-rohan. a though, and, in order to avoid contradiction, it seems that theydecided not to include the concept of the twenty-eight incarnations andŚveta’s role therein.

The Paśupatas and Śaivas of the Nılakan.t.ha Temple complex

To judge by the exceptionally rich archaeological remains on and aroundKalanjara, the seventh to twelve centuries must have seen an outburst ofreligious and other activity on the hill. The reason that so many questionscould not be answered satisfactorily above is therefore not that there isa lack of sources, but rather a lack of scholarly investigations of thesesources. When our team visited the hill in January 2012 (see frontispiece),we were truly amazed by the abundance of material, images, epigraphs andbuilding constructions alike, but hardly any of these have been describedand explored since General Cunningham, though an exception must bemade for Rajendra Yadav, who reports his findings in two brief articlesin Pragdhara (2002–03, 2009–10). The situation is slightly better for thetenth to twelfth centuries when the hill was ruled by the Candellas. Someof the inscriptions pertaining to this period, but by no means all, havebeen collected in the Corpus Inscriptionum Indicarum VII.

On the west side of the hill, along the staircase leading towards theNılakan. t.ha Temple, within the fausse-braye of the fortification, we find theface of the rock and the wall of the fort embellished by sculptured votive re-liefs, images and minor shrines.639 Several images of Lakulıśa testify to theprominence of the Paśupatas on the hill. An instance thereof is presentedin Plate 39. Above the image is an inscription in eighth/ninth-centurycharacters. The inscription reads: śrılaku(l)ıs.a[śam. *] krayana(ga)rı pra-n. amati.640 It thus testifies to the presence of a Paśupata community onthe hill in the eighth century, which was sponsored by a collective of mer-chants (krayanagarı).

639 Two of these rock-cut reliefs, (1) showing worshippers before a linga on a pedestalflanked by Brahma and Vis.n. u, and (2) worshippers before a linga carried on theback of the Bull, have been discussed by Vats 2002.

640 Yuko Yokochi, in an e-mail to P.C. Bisschop d.d. 29-6-12, reads (first line): ‘śrıla-ku(l)ı(ma)(ka?)yanagarı pran. amati; (second line): para(ma)la(ga).’

216 Hans Bakker

Plate 39Kalanjara: Nılakan. t.ha, Lakulıśa

The important role of this Paśupata community is confirmed by theKalanjar Stone Inscription of the time of Kırtivarman of ad 1090,found in the Nılakan. t.ha Temple itself, in which inscription the guru ofKırtivarman, Vasudeva—the builder of the man. d. apa, who had receivedenlightenment (prabodhamahima) thanks to Śiva’s grace—extends arequest for approval to the royal superintendants, the Śaivas, whom Itake to refer to the Śaivasiddhantins, as well as the foremost temple andother officials and teachers of the Paśupata.641

In this period the hill and fort were part of the kingdom of the Cande-llas and it was under their rule in the eleventh century that the majesticarchitecture of the Nılakan. t.ha Temple complex was accomplished, a re-construction of a preceding edifice. The latter becomes clear from thefollowing observations by Rajendra Yadav:

Findings of twenty-six Śiva-lingas, both iconic and non-iconic types,[which] are carved in living rock in a row behind the sanctum of

641 CII VII.3, 370 ll. 9–10: tad etad anumanyantam. rajaśrıkaran. adhipah. | śaivah. paśu-patacaryavarikapramukhaś ca ye ‖ 6 ‖ This verse has been interpreted differently.The word varika means ‘beggars,’ i.e. mendicants according to Mirashi, (op. cit., II,285 n. 6), or ‘temple superintendents’ according to Sircar (quoted in op. cit. III, 370n.,5). Unlike Mirashi, but following Katare in EI XXXI, 163, I take the compoundending with ◦pramukhaś not as a Bahuvrıhi with śaivah. , but as a Tatpurus.a.

Seven Brahmins, King Udayana and Śveta’s Struggle with Death 217

Nılakan. t.ha Temple.642

[. . . ]

A non-iconic Śiva-linga is carved in the centre of row depictingbrahmasutra on the rudra portion and a four line Brahmı inscrip-tion engraved below. A devotee in anjalımudra is seated to the leftof the inscription below Śiva-linga (KF 106). [. . . ] The findings ofthis label inscription is very important so far as the dating of Śiva-lingas and antiquarian remains of Kalanjar fort are concerned. Onpalaeographic grounds, it can be assigned to late fourth to mid-fifthcentury ad. The findings of a group of rock-cut Śiva-lingas at Nıla-kan. t.ha Temple suggest that this place was an important centre ofall religious activity in Kalanjar for at least right from the Guptaperiod.643

Plate 40Kalanjara: Lingas in the cave behind the sanctum of the Nılakan. t.ha Temple

Thanks to a photo kindly sent to me by Rajendra Yadav, I could read,partly at least, the fifth-century inscription. It names a devotee Harideva

642 Yadav 2009–10, 77. These lingas in Cave 6 are numbered KF 94 to KF 119.643 Yadav 2009–10, 74, Pl. 12 (KF 106). Yadav does not give the inscription.

218 Hans Bakker

and records that the construction (?) was made for the sake of merit anddharma.644

Summary

After having acquired a reputation as a suitable place for suicide in thepre-Gupta era (MBh), the steep hill of Kalanjara in Gupta times wasincluded in the legend of the seven brahmins, who, reborn on the hillas deer, ended their lives by starvation (HV). A site on the southernside of the hill became associated with this event, known today as Mr.ga-dhara, and, to judge by the fifth-century graffito found on the face of therock next to it, it attracted a community of Lakud. a ascetics from Vidiśa,followers of the Paśupata saviour Lagud. i, who may have lived in one ormore of the many caves together with other ascetics. Śaiva presence onthe hill in the fifth century is further attested by several ekamukhalingas,a relief of Lakulıśa, a row of twenty-six lingas in cave 6 in the Nılakan. t.haTemple, and seals depicting a linga sanctuary. One of the fifth-centuryŚiva temples may, according to the (later) Chandi Darwaza inscription,have been known as Bhadreśvara.

Sixth-century sources confirm the continuing importance of the hill byconfirming the legend of the suicide of the seven deer, who in the Vindhya-vasinı Cycle of the Skandapuran. a are said to have thrown themselves offthe cliff. The Maukhari king of Kanauj, Śarvavarman, donated rent-freeland (agrahara) to unknown brahmins living in the Kalanjara circle, agrant that was reconfirmed by a king of Kanauj in the ninth century.

The Skandapuran. a connects the hill for the first time, it seems, withthe myth of Śiva’s devotee Śveta. The latter is saved from Death (Ka-la), by Śiva’s intervention, which in later versions of the myth yielded thestandard nirukti: kalam. jaritavan. In this connection the Skandapuran. amentions the holy site of Krodheśvara, which may or may not have beenidentical to the fifth-century Bhadreśvara. The figure of Śveta developsinto an incarnation of Śiva in later Puran. a sources.

After the first redaction of the Skandapuran. a had been closed, that is postad 650, the hill seems to have undergone an influx of new religious ideas. Asubstantial part of the rich archaeological remains on the hill date from theperiod between Hars.avardhana and the arrival of the Candellas in the sec-ond half of the tenth century. They testify to a strong Śaiva presence—thePaśupata probably alongside the Śaiva Siddhanta. The second redaction of

644 The inscription on the linga in the front runs:[1] hari〈de〉va〈h. 〉 (pra)n. a〈mati〉 [2] palite . . . [3] karitam. pun. yadha [4]〈r〉(ma)yoh.

Seven Brahmins, King Udayana and Śveta’s Struggle with Death 219

the Skandapuran. a (RA) confirms explicitly Śveta’s association with Kala-njara, making him a singing and dancing Paśupata saint, and it confirmsthe presence of a Krodheśvara sanctuary on the hill.

The Kalacuri tradition claims the conquest of the hill, obliquely re-ferred to as Śveta’s abode, by their ancestors, and these Kalacuris (comingfrom Malwa) may have brought not only new forms of institutionalizedTantric Saivism, but also the ancient legend of King Udayana and hiswife, the Ujjayinı princess Vasavadatta, who in this period came to beconnected to Kalanjara. This emerges from the Br.hatkatha version of thelegend, in which Udayana is said to have committed suicide together withhis two wives by throwing himself off the cliff.

To this pre-Candella period we also ascribe the enigmatic figure of thecarrier, found all over the hill, who is connected with the samanta namedVasanta, registered at least three times in accompanying epigraphs. TheChandi Darwaza inscription confirms the belief that King Udayana ofVatsa once ruled Kalanjara, as he is credited with the foundation of theBhadreśvara temple on the hill. His Minister Vasantaka, known from theUdayana plays as the king’s vidus.aka, may have been thought to have hadsomething to do with the representation of the carrier, but in the presentstate of our knowledge this is no more than speculaton.

The Myth of Tilottama and Man.d. aleśvara

d Mun. d. eśvarı Hill d

Man.d.aleśvarasvamin and the Skandapuran.a

The myth of Tilottama

Passing Varan. ası (for which see above, p. 138), we travel 310 km due eastfrom Kalinjar in order again to find sixth-century material evidence ofanother story contained in the Vindhyavasinı Cycle of the Skandapuran. a(SP III, 259–267), in addition to the legend of the seven deer on KalanjaraHill, namely the Man. d. aleśvara Mahatmya (SP 62.1–53). This Mahatmyatells the story of the demons Sunda and Nisunda and the nymph Tilotta-ma, also known from the Mahabharata,645 which serves as an ætiologicalmyth of the origin of the caturmukhalinga.

Tilottama, her creator Brahma, and the other gods assemble in theVindhyas to worship Śiva, but he is invisible at first. Then Śiva appearson the scene in the form of a linga. When Tilottama makes her pradaks. in. aaround him, four faces emerge on the four sides of the linga, which givesher the power of sexual desire, the tejas, in fact, that had been storedwithin Śiva ever since he had burnt Madana. Thus the caturmukhalingacame into being (along with sexual desire).646 Then Śiva says that,

Because she [Tilottama], as well as the gods, has circumambulatedHim in a circle (man. d. ala), this place on Mt Vindhya, which is sacredto Him, will be called Man.d. aleśvara, and He will always be presenthere.647

645 MBh 1.200.18–204.26; Mbh 13.128.1–6. See Bakker 2001.646 SP 62.8–26.647 Yokochi’s synopsis of SP 62.27–28:

abravıc ca suran sarvam. s tatredam. vacanam. śivah. |yasmad iyam. mam. yuyam. ca man. d. alena pradaks. in. am ‖ 27 ‖cakre sarve suraśres.t.hah. sthanam. tasmad idam. mama |bhavis.yati girau vindhye man. d. aleśvarasam. jnitam |sam. nidhyam. sarvada hy asmin karis.yami varapradam ‖ 28 ‖

221

222 Hans Bakker

The story continues, again in the words of Yokochi.

After sending Tilottama to the demons, all the gods huddle together.They see Parameśvara in their midst, [in the form of another linga].Because He appeared in the midst of the gods while they were hud-dling together (pin. d. yamana), He was called Pin. d. areśvara there.648

A holy place called Pin. d. araka is known to the Mahabharata (MBh13.26.54), and other places with this name are known to several Puran. as.The Pin.d. areśvara at issue, however, must be located in the vicinityof Man. d. aleśvara. The word pin. d. ara means (Śaiva) mendicant, and theætiological myth told in the Skandapuran. a probably refers to a locallinga that was worshipped by Śaiva ascetics.

The place where our text locates these two lingas is in the northernpart of the Vindhyas, since Tilottama is said to go to the south to findthe two demons in the foothills of the Vindhya (vindhyapades.u).649 Thehypothesis I would like to advance, in which I follow the earlier suggestionby Yokochi, is that both lingas are to be situated on the Mun. d. eśvarı Hill—the Man. d. aleśvara being the ancient caturmukhalinga preserved on the sitenear the main temple, and the Pin. d. areśvara being one of the dozens of5th–7th-century lingas that are scattered all over the hill.650 In order tosubstantiate this hypothesis we should take a closer look at the remainson Mun.d. eśvarı Hill.

Mun. d. eśvarı Hill

The hill near the village of Ramgarh, on whose top (height 184 m) theso-called ‘Mun.d. eśvarı Temple’ is found,651 belongs to the northern spursof the Kaimur Range, the eastern branch of the Vindhyas. It is situatedin the Kaimur District of south-west Bihar, 70 km SE of Varan. ası, 170 kmSW of Patna, 60 km NW of Rohtasgarh on the River Son, and 8 km SW

648 Yokochi’s synopsis of SP 62.36–38:atha sampres.ayitva tu devatas tam. tilottamam |atmanam. pin. d. ayamasur devas te sarva eva hi ‖ 36 ‖kah. kah. katama ayata ihadyeti surars.abhah. |tato madhye sthitam. bhuyas te ’paśyan parameśvaram ‖ 37 ‖pin. d. yamanes.u deves.u yasman madhye samasthitah. |pin. d. areśvara ityeva tatrasav abhavat tatah. ‖ 38 ‖

649 SP 62.41cd, 62.42a; Yokochi 2013b, 116f.650 Yokochi (2004b, 88 n. 28) writes in her thesis: ‘Man. d. aleśvara and Pin.d. areśvara

have not yet been located in the Vindhya mountains. One of the candidates is theMun.d. eśvarı Temple built atop of a hill situated close to the village of Ramgarh tothe southwest of Bhabua, District Kaimur, Bihar.’ Cf. Yokochi 2013b, 115ff.

651 The present-day main object of worship in this temple is a fragmentary imageof the Goddess, named Mun.d. eśvarı, which has given the temple and the hill itsmodern name: Mun. d. eśvarı Hill. This image is a later accretion to the temple.

The Myth of Tilottama and Man. d. aleśvara 223

Plate 41Mun. d. eśvarı Hill and surroundings

of Bhabua. The present octogonal temple has attracted the attention ofscholars since the days of Francis Buchanan, whose description was used(plagiarized would be a better description)652 by Montgomery Martin inhis Eastern India, which appeared in 1838. Buchanan, who surveyed thedistricts of Bihar between ad 1807 and 1814, reports that his descriptionis based on the drawings of a painter whom he had sent up to the place,which might have been in January 1813.653 A print based on this wasincluded in Eastern India (Plate 42).654 Earlier, a drawing of the templewas made by Thomas Daniell, who probably visited the site in February1790,655 and whose engraving, published in 1808, might have been knownto Buchanan (Plate 43).656

As can be derived from both pictures, the temple was largely coveredby a mound of rubble overgrown with plants and trees. The site was ex-cavated by the ASI after a visit by Theodor Bloch in 1902.657 Bloch wasalso the first to report on the inscription found here.

652 See Neuss 2003, 533 n. 7.653 Nuess 2003, 533 (based on Oldham 1926). Martin Eastern India I, 456: ‘I could not

conveniently visit the place, but sent a painter, who drew the most remarkable part.The temple, now very ruinous, has been an octagon supported by four columns(E).’

654 Vol. I, opposite p. 474 (Pl. V).655 Nuess 2003, 533 (based on Archer 1969).656 The drawing was engraved by Thomas and his brother William and published in

Antiquities of India, December 1808 (Archer 1980, 235).657 Bloch in ASI Annual Report 1902–03, 42f. Neuss 2003, 533f.

224 Hans Bakker

Plate 42 Plate 43Buchanan: sketch ad 1813 Daniell: drawing ad 1790

[. . . ] Among the debris the second half of an inscribed stone hasbeen found, the first half of which had been sent to the Indian Mu-seum, Calcutta in 1891. This stone is dated sam. vatsare trim. śatitamekarttika-divase dvavim. śatitame (year 30, the 22nd of Karttika) whichmust be referred to the Hars.a Era, thus corresponding with 635A.D.658

The two halves of the stone have been combined and were first editedand published by R.D. Banerji in Epigraphia Indica IX (1907–08). ThisMun. d. eśvarı Inscription of the Time of Udayasena: the Year 30 has gonethrough five editions altogether,659 but in spite of the work done by theseexcellent epigraphists, there is room for improvement. I have presented anew edition, translation and interpretation of this inscription in Bakker2013.

Before discussing the inscription in more detail, however, a few wordsshould be said about the temple. Joanna Williams gives the followingassessment:

Fragments at the site extend from the sixth century until the eighthor later. On stylistic grounds the surviving octagonal shrine shouldbelong to the first half of the seventh century. [. . . ] The octagonal

658 Bloch in ASI Annual Report 1902–03, 43.659 R.D. Banerji 1907–08, N.G. Majumdar 1920, P.K. Agrawala 1987, P.R. Srinivasan

1991, J. Neuss 2003.

The Myth of Tilottama and Man. d. aleśvara 225

plan of Mun.d. eśvarı may illustrate the same search for variety thattakes the form of pancaratha and stellate plans at Sirpur.660

Plate 44 Plate 45Daniell: drawing ad 1790 Man. d. aleśvara linga

The interior of the temple consists of a central pavillion supported by fourrucaka pillars and encircled by a pradaks. in. apatha, which enshrines a tall,c. 13th-century caturmukhalinga.661

The Insciption of Udayasena, Year 30

The hill, temple, surrounding debris and the inscription have been thor-oughly examined most recently by Jurgen Neuss (2003). Neuss subscribesto the almost universally accepted theory that dates the architecture ofthe present temple to the first half of the seventh century.662 However,

660 Williams 1982, 166f. Cf. EITA II.1, 119: ‘If the foundation inscription refers to theHars.a era, yielding a date of A.D. 636, the temple stylistically can be assigned tothe same period.’

661 EITA II.1, 119.662 Neuss 2003, 543.

226 Hans Bakker

Neuss rejects one piece of evidence used to endorse this theory, viz. thealleged contemporaneity of the temple and the inscription.663 Neuss’s ar-gument is twofold. 1) Refering to Asher, he argues ‘that there is thus nobasis to connect the date of our inscription with any known era.’664 2)A detailed palæographic investigation leads him to the final observationthat,

As the Mun.d. eśvarı record still uses the older tripartite ya, I wouldsuggest that the Mun. d. eśvarı Inscription must have been engravedbetween A.D. 570 and A.D. 590, and that the script shows influencesfrom writing habits developed in Western India.665

I subscribe tentatively to Neuss’s analysis and would therefore proposeto date the inscription to the second half of the sixth or early seventhcentury. It follows that the otherwise completely unknown MahasamantaUdayasena was a feudatory, not of Hars.avardhana, but of the Maukharisof Kanauj, and, most likely, ruled during the reign of Avantivarman. Theinscription is not the foundation charter of the main temple itself, but ittestifies to a historical process in the second half of the sixth century, inwhich the building of a new religious complex on the hill was in full swing.In order to facilitate further discussion, I give here my translation of theinscription. For the Sanskrit text and a philological analysis the reader isreferred to Bakker 2013.

Success!In the thirtieth year, at the twenty-second day of Karttika; here, at theabove year, month and day—within the reign of the illustrious Mahasa-manta, Mahapratıhara, Maharaja Udayasena—after having requested thepermission of the head of the community (kulapati), Bhanudayin (?),666

and the Temple Council, General Gomibhat.a has commissioned the build-ing of this accommodation as part of the Vinıteśvaramat.ha, in order toincrease the merit of his father, mother and himself.

He has made an assignment to the temple of the illustrious Narayan. a of apala of lamp oil and two prasthas of rice for the daily food-offering (to thedeity), permanently, for as long as sun and moon exist, to be taken fromthe storeroom that belongs to the temple of the illustrious Man.d. aleśvara-svamin.

663 Neuss 2003, 535.664 Neuss 2003, 534f. Cf. Asher 1985, 133: ‘In the ancient written documents of eastern

India, time was almost invariably marked according to the regnal year of a currentmonarch.’

665 Neuss 2003, 538.666 For this reading see Bakker 2013, 269f. As has been remarked by Sanderson (e-mail

15-2-2012), the beginning with bha suggests a Paśupata name, but no satisfactoryinterpretation has suggested itself.

The Myth of Tilottama and Man. d. aleśvara 227

(And) to the Temple Council he has given the means, such as cows, offer-ings, garlands and food, at the value of fifty dınaras, meant collectively forthe regular set of various observances for the sake of the illustrious feet ofMan.d. aleśvarasvamin.

Having learned so, there should not be a disruption of what has thus beenassigned, neither by future incumbents nor by the ascetics.

The one who, having thus been informed, acts otherwise, will live in hellwith great sins.

Thus [. . . ]

And it has been declared,

O Yudhis.t.hara, zealously protect the land, whether it has beengiven by yourself or by others. O foremost of the landlords, protectionis superior to giving.

General Gomibhat.a made three donations altogether, for which he had tomake two formal requests, one to the head (kulapati) of the Vinıteśvara-(mat.ha), and the other to the Temple Council (devanikaya).

The task of the Devanikaya may have been to supervise and administerthe temple properties on the hill and that of the Man. d. aleśvarasvamintemple complex (devakula) in particular. General Gomibhat.a may himselfhave been a member of this council, if we consider the people who arereported to sit on such a council. To this effect the Taleśvara CopperplateGrant of Dyutivarman provides information.667 The latter grant was madeat the request of the Mahasattrapati Trata and the Devanikaya, ‘whichwas headed by the minister Bhadravis.n. u, the governor of the fort, theofficer in charge of the king’s betel-box, the custodian of the sacrificialfire, and the head of the royal guard.’668

Gomibhat.a needed the permission of kulapati Bhanudayin for his firstdonation, viz. accomodation or housing (mat.ha) for the ascetics, sincethat was to be built on the premises of the Vinıteśvaramat.ha. The wordkulapati signifies ‘head of a community.’ Because the community here isdefined as that of a mat.ha, a likely interpretation is that the kulapati is

667 EI XIII No. 7, 109–21: Two Talesvara Copperplates by Y.R. Gupte (1915–16).These two copperplates were found in Taleśvara in the Almore District, UnitedProvinces (modern Uttarakhand). Though they appear to be forgeries, they areactually documents pertaining to the period ‘between the middle of the 6th andsecond quarter of the 7th century’ (Gupte op. cit. 113), i.e. to the same period towhich the Mun. d. eśvarı Inscription belongs.

668 EI XIII, 115 ll. 6–8: rajadauvarikagnisvamikarankikakot.adhikaran. ikamatyabhadra-vis.n. upurah. saren. a ca devanikayena vijnapitam. , which is rendered by Gupte as: ‘pre-ceded by royal doorkeepers, the attendants of the sacred fire, the karankikas, thesuperintendent of the female (temple) slaves, the minister Bhadravis.n. u.’

228 Hans Bakker

the Mahant or abbot of the community of ascetics (tapovanikas) of theVinıteśvaramat.ha.669

He needed the approval of the Temple Council for his second donation,viz. a daily offering of lamp oil and rice meant for the Narayan. a Temple(devakula), since this was to be taken from the storeroom (kos.t.hika) con-nected to the Man.d. aleśvarasvamin Temple. How this was paid for remainsunclear, unless we assume that the third donation to the Temple Councilwas considered to be sufficient to cover the costs of this as well.

His third donation to the Temple Council, for which he also neededapproval, was meant for the worship of Man. d. aleśvarasvamin himself andis said to have had a value of 50 dınaras. For this amount he provided themeans needed for worshipping the deity, such as milk (a cow), food andgarlands.670

The inscription attests to the existence, by the end of the sixth cen-tury, of a main Śaiva temple complex, called Man.d. aleśvara. Connected tothis temple was a storeroom, which could also supply the daily offeringsfor a minor Vais.n. ava sanctuary dedicated to Narayan. a. In addition, andpossibly closely connected to the Man. d. aleśvara complex, was a monasteryof Śaiva ascetics, called Vinıteśvaramat.ha. Whether there was a templeor shrine of Vinıteśvara cannot be deduced from this inscription with cer-tainty, but the existence at the time of such a temple is very likely indeed.The monastery or/and its temple was headed by Bhanudayin.

The Sites of Mun.d. eśvarı Hill

The Temple of Man. d. aleśvara

Although the great majority of the remains on the Mun.d. eśvarı Hill datefrom the sixth to the eighth century, as observed by Joanna Williams(quoted above), there is a significant number of artefacts that can bedated to the fifth to sixth centuries. Krishna Deva observes in EITA II,119:

Among the architectural members lying loose around the temple arefragments of four pillars, the tallest about 11 ft. high; these are plainRucaka pillars carved with an unornamented semicircular medallionand a short octagonal section on the bharan. a. These need not be

669 Cf. Vallabhadeva ad Raghuvam. śa 1.94, who glosses kulapatina as aśramagurun. a.Parallels for this usage are also found in Cambodian inscriptions (Cœdes 1937–66)K.95, K.309, K.362.

670 For discussion of our conjected reading vicitrakarmantatantra◦ (regular set of var-ious observances) in l. 10, see Bakker 2013, 271.

The Myth of Tilottama and Man. d. aleśvara 229

Plate 46Mun. d. eśvarı Hill seen from the South

of the same date as the temple inscription [taken to be ad 636 byKrishna Deva, H.T.B.].

The same may apply to the Rucaka pillars that form the central pavillionwithin the temple, in which the present caturmukhalinga with four uniformfaces is installed.

Among the original pieces Neuss adduced are ‘an elaborately carveddoor lintel and another lintel with a lat.abimba showing a head of Śiva.’671

To these we may add some fragments which, in our view, could well belongto the fifth to sixth centuries: a fragment of a doorway with guardian anda fragment of an ornamented pilaster (Plate 47), as wel as a Candraśalawith Skanda holding a cup and a cock (?), and another niche with a seatedascetic figure holding a cup and a khat.vanga (Plate 48).

Most important to our present concern, however, is that the site haspreserved an old caturmukhalinga (Plate 49), which resembles in size andshape the present one in the temple (Plates 44, 45), and which may havebeen its predecessor.672 The image is extremely worn and therefore im-possible to date exactly, but the contours of four distinct heads are still

671 Neuss 2003, 543 Plates 6 and 7.672 Deva 1990, 161. Cf. Neuss 2003, 546 n. 37: ‘This caturmukhalinga seems to be

the model for the one which is at present in the temple and which most likelyis a substitute for the former.’ Unlike the linga presently installed in the temple,however, the old caturmukhalinga shows four different faces, significant for its earlydate and in conformity with the description in SP 62.16–24 (see also Yokochi 2013b,114–17).

230 Hans Bakker

Doorframe with dvarapala Ornamented pilaster

Plate 47Mun. d. eśvarı: architectural fragments

Skanda with cock? Figure with khat.vanga

Plate 48Mun. d. eśvarı: architectural decorations

The Myth of Tilottama and Man. d. aleśvara 231

discernible. We conjecture that this is the Man. d. aleśvarasvamin mentionedin the inscription.

Plate 49Mun. d. eśvarı: old caturmukhalinga & detail

Vinıteśvara

The hill contains other 5th–6th century remains, of which the most im-portant ones seem to have belonged to what Neuss calls site A, ‘abouthalf way up the hill just a few meters to the north of the stairway’ (seePlate 46). Neuss gives the following description.

The entrance to this open space is flanked by two small pillars. At theeastern end of this platform-like space there is a pit in which severalrock-cut architectural members, either heavy pillars or lintels of anold temple, are lying buried. At other places stone slabs which couldonce have been parts of a roof are found. On the hill slopes to thenorth and the east a number of dressed rocks as well as numerousold bricks are strewn. All these remnants indicate that once a stonetemple and some brick building must have existed in this place.673

The relatively early date of this site is warranted by a linga and an in-scription. The ekamukhalinga is today stored in the main temple on topof the hill (Plate 50), but Neuss reports that on his first visit in January2001 he still saw it in situ.674

673 Neuss 2003, 559.674 Neuss 2003, 546.

232 Hans Bakker

Plate 50Mun. d. eśvarı: Ekamukhalinga of site A

I agree with Neuss (op. cit. 547 and Plates 9 to 12) that: ‘Stylistically, theekamukhalinga at site A doubtlessly reflects the Gupta school of art.’

Close to site A, along the staircase, is a large image hewn into the rockof a six-armed Bhairava in striding pose, holding a elephant’s hide abovehis head (Plate 51). At his left leg lies a dog and further on to his leftsits Parvatı, who turns her head away in fright.675 A terminus ante quemfor this sculpture is provided by a graffito written below the figure (Plate52), reading priyangu pran. amati, which is dated by Neuss to the middleof the seventh century.676

It is very likely that this site A represents what remains of themonastery of Śaiva ascetics, Vinıteśvaramat.ha and the little temple ofVinıteśvara, in which the ekamukhalinga may have been enshrined.677

675 For this type of Bhairava image and its (political) significance see Mirnig 2013,334f.

676 Neuss 2003, 560. Cf. Salomon 1976, 105: ‘The characters are late Brahmı of aboutthe 7th century A.D.’

677 Cf. Neuss 2003, 547.

The Myth of Tilottama and Man. d. aleśvara 233

Plate 51Bhairava near site A

priyangu pran. amati

Plate 52Inscription below the Bhairava image

234 Hans Bakker

Plate 53The Skandapuran. a research team at site B

The Tapovana

On the opposite slope a steep footpath leads ‘to a large flat rock out-cropping on the west side of the hill, about three quarters of the way upoverlooking the plain below’ (Plate 53).678 This is Salomon’s and Neuss’ssite B (see Plate 46).

upacarutapaPlate 54

Graffito at site B

The proximity of the main sanc-tuaries on the one hand, and theinaccessibility of the site on theother, made this spectacular out-crop with adjacent (shallow) rockshelters an ideal hide-out for as-cetics. It may have functioned as thetapovana where the tapovanikas ofthe Vinıteśvaramat.ha retreated fortheir meditation and left their graf-

fiti, datable ‘to about the late sixth or early seventh century A.D.,’679 suchas this specimen, reading upacarutapa (Plate 54).

678 Salomon 1983, 73. Cf. Meister 1981, 77.679 Salomon 1983, 74f. Ibid. 73f.: ‘The inscriptions at site B are in two sub-groups,

one on a triangular rock on the inner portion of the outcropping, the other ona large flat rock farther out on the outcropping. Altogether at site B there are9 inscriptions in shell characters, 14 in ornate Brahmı, and a few fragments andgraffiti in standard Brahmı.’

The Myth of Tilottama and Man. d. aleśvara 235

Narayan. adevakula

The material remains on Mun.d. eśvarı Hill are predominantly Śaiva. Thereference to a Narayan. a temple (devakula) on the hill in the inscriptionhas therefore created much confusion. Due to a wrong construction ofthe Sanskrit in the inscription, Neuss even speaks of a Narayan. adevakulamat.ha (op. cit. 540), which leads him to conceive of the main temple as ‘apancayatana complex,’ which would turn this ‘mat.ha’ into ‘a subsidiaryshrine of the main temple’—all this in order to solve the alleged problem‘that we know of no vais.n. ava artefact found on the hill’ (ibid. 549). WhyGomibhat.a would make a special donation to a subsidiary shrine of themain temple is herewith not explained.

We can alleviate the problem to some extent by recording that, inaddition to the Vis.n. u figures in the two sarvatobhadra-stambhas on theentrance path to the main temple,680 we have found two images of Vis.n. uwhich may date to the period of the inscription. One of them is amongthe debris within the premises of the main temple; it leans on a Śiva linga(Plate 55). The image is four armed, both upper hands have broken off,but the lower hands hold the ayudha figures of cakra and gada.

Cakra Vis.n. u leaning on a linga Gada

Plate 55Mun. d. eśvarı Hill: Vis.n. u Image

680 See Neuss 2003, 549, Pls. 15–17.

236 Hans Bakker

Another four-armed Vis.n. u was found near site A. Both upper arms arebroken off completely, but the lower arms rest on the figures of Cakra andGada, both on their knees, in devotional pose. On the front of the base isa kneeling devotee (Plate 56).

Vis.n. u Figure at the base of the Vis.n. u image

Plate 56Vis.n. u image near Mun. d. eśvarı Hill site A

On the left side of the base is a figure seated in European posture, whois holding something in front of him. We conjecture that this sculpturemight have been a votive offering or the donation of a grateful devotee.

On the basis of the combined evidence of inscription and images, weconclude that the hill contained a Vis.n. u or Narayan. a sanctuary, either onthe top near the main temple, or on the eastern slope near site A.

Conclusion

This is as far as epigraphy and archaeology can take us. It cannot tell uswhy ascetics and pilgrims flocked to this hill in great numbers, or whythe site was considered holy. The answers to these questions are partlyprovided by the Man. d. aleśvara Mahatmya, which we quoted above (p. 221).

The textual and archaeological evidence presented indicate that theMun. d. eśvarı Hill was a thriving centre of early Śaiva activities from the

The Myth of Tilottama and Man. d. aleśvara 237

Plate 57Mun. d. eśvarı: Lakulıśa on lintel 681

second half of the sixth century onwards. A charming old myth was givena sectarian Śaiva twist, and its dissemination by way of the Skandapuran. amay have contributed to the fame of the hill and its central sanctuary,Man. d. aleśvara. An increasing number of pilgrims visited the holy site, andfrom the income generated the sacred complex could expand, so much sothat a storehouse and a governing body were required. The population ofascetics increased accordingly and was organized in a monastery of somesort, a mat.ha. At least one temple of Narayan. a was integrated into thecomplex.

The Śaiva ascetics in the sixth and seventh centuries would for thegreater part have been Paśupatas, and their patron saint, Lakulıśa ispresent in a lintel to welcome them (Plate 57). This lintel, like many othermajor images, has been removed from the site and is presently stored inthe Patna Museum.

Once the hill had established its reputation it began to attract the pat-ronage of the elite. Gomibhat.a exemplifies this development, and we havesuggested elsewhere that the Vinıteśvara temple may have been foundedby someone named Vinıta, who may have ruled from the nearby Rohita-giri, Rohtasgarh.682

681 The website of the AIIS, s.v. Mundesvari, gives the following specifications: ‘ca.600-625 CE, buff sandstone, 27 x 54 cm, architectural fragment, presently locatedat: Patna Museum, Acc. No 54094, Negative No 485.94.’

682 See Bakker 2010. In inscriptions of the Tunga dynasty, ruling in the tenth centuryas feudatories of the Bauma-Kara kings of Orissa in the Yamagarta-man. d. ala inthe Dhenkanal District of Central Orissa, this dynasty of petty kings claims thatits ancestor, named Vinıta, came originally from the Rohitagiri.Another possibility is that the Mat.ha was named after the Paśupata ascetic whofounded it. Above we came across a Raśi ascetic called Vinıta who was active at

238 Hans Bakker

Plate 58 Plate 59Mun. d. eśvarı: Skandamatr.

683 Mun. d. eśvarı: Skanda 684

The inscription and earliest archaeological material on the hill belong tothe period in which we date the composition and first redaction of theSkandapuran. a, that is, when North India became united again, for the firsttime since the fall of the Gupta empire, under the Maukhari rulers, whosehomeland was in the nearby valley of the River Son. It is the period inwhich the Śaiva movement is beginning to organize itself through sectariansettlements or mat.has in centres of pilgrimage that are on the rise, suchas Varan. ası, Kalanjara, or the western Bhadreśvara (Rishikesh). Man.d. a-leśvara may have been another hub in the emerging network of Paśupataascetics, important enough to deserve a Mahatmya of its own in one of themovement’s greatest literary productions, the Puran. a of Skanda. There are

the end of the seventh century (p. 149).683 The AIIS website s.v. Mundesvari, gives the following specifications: ‘ca. 600-625

CE, Buff sandstone Standing figure, 88 x 63 cm, presently located at: Patna Mu-seum, Accession No 54048, Negative No 486.5.’ The other (adopted) son of Parvatı,Gan. eśa, is sitting at her feet. Next to him is an unidentified figure who may beNandin, who in SP 22.20–22 is adopted by Parvatı as her son. The mother handsover to her son a toy instrument.

684 The Patna Museum Catalogue gives the following specifications (p. 44f.): ‘Kartti-keya seated on peacock, light buff sand-stone, Arch. 6003, 65 x 42 cm, seventhcentury ad.

The Myth of Tilottama and Man. d. aleśvara 239

several images of this god on the hill, such as the one in which he sits onthe arm of his mother (Plate 58), and another where he proudly rides hispeacock (Plate 59).

The success of the holy place attracted more patronage during thereign of King Hars.avardhana of Kanauj, witness the remains of the octago-nal temple, which replaced the original Man.d. aleśvara temple in the firsthalf of the seventh century.

Plate 60Mun. d. eśvarı: Tilottama?

What we have presented above are ar-guments in favour of a theory. Thistheory would gain in strength consider-ably if we could find out more aboutthe Mahasamanta Udayasena of the in-scription on the one hand, or the em-bedding of the Tilottama myth of theSkandapuran. a in the local religious build-up of the Mun.d. eśarı Hill on the other.I have searched for sculptural evidencethat could endorse such embedding, inaddition to the ancient caturmukhalinganamed Man. d. aleśvara. Could the architec-tural fragment shown in Plate 60 be in-terpreted as the beautiful nymph Tilo-ttama while gracefully circumambulatingŚiva?685

685 This is part of an architectural fragment of which the website of the AIIS gives thefollowing specifications: ‘ca. 600-625 CE, buff sandstone standing figure, 30 x 38cm, presently located at: Patna Museum, Acc. No 54047, Negative No 486.4.’ Thepanel next to this image, which seems to be unconnected, contains a Naga deity.

Goddesses from the East

d Bangarh or Kot.ivars.a d

in cooperation with Yuko Yokochi

Introduction

We conclude our journey through the world of the Skandapuran. a in theeast. After a brief note on the possible place of composition of our text,Varan. ası, our survey began in Kuruks.etra, the westernmost area coveredby the Puran. a. Via Hardwar and Kalinjar, we reached Mun.d. eśvarı Hill inBihar in the previous chapter. From Tilottama’s Man.d. aleśvara, we travelc. 500 km further eastwards until we come to the present-day village ofGangarampur in the Dakshin Dinajpur District of West Bengal (India).As will be argued below, a historic site near this village can be identified asKot.ivars.a. The glorification of Kot.ivars.a and its goddesses is the last Ma-hatmya in the Skandapuran. a, found in the second half of chapter 171.686

Herewith we reach the eastern limit of the geographical area encompassedby our text, situated 81 km north-east of the army camp of King Hars.ain Kajurgira on the Ganges, the site where he reportedly met the Chinesepilgrim Xuanzang (see above, p. 97).

Bangarh or Kot. ivars.a: the material evidence

The Damodarpur Copperplate Inscriptions

In April 1915 a set of five copperplates was discovered while constructinga road near the village of Damodarpur, 13 km west of Phulbari Railway

686 SPBh 171.78–137. This Mahatmya is inserted after the myth of Mount Kraunca.SPBh 171.1–77 tells how Skanda runs a race with Indra around Kraunca. He de-stroys the mountain and brings it back to life. Indra acknowledges Skanda’s great-ness. The Mahatmya answers Vyasa’s question as to how Śam. kara became theleader of the Matr.s and who these Mothers are. SPBh 172 continues with the mythof Indra clipping the wings of the mountains.

241

242 Hans Bakker & Yuko Yokochi

Station in the Dinajpur District of what is now Bangladesh (see Plate6).687 These documents record the legal execution of land transactionsunder the Imperial Gupta dynasty, two under Kumaragupta I (Nos. 1–2),two under Budhagupta (Nos. 3–4), and the last one, No. 5, issued under aGupta king whose name (of two syllables) has been lost, and which hasbeen, consequently, much debated. The opening of this 5th inscriptionruns as follows.

On 5 Bhadra in the year 224—when the Paramadaivata, Parama-bhat.t.araka, Maharajadhiraja, the illustrious . . . gupta was reigningover the earth, and when in the Kot.ivars.a District—thriving thanksto the government of Devabhat.t.araka Maharajaputra, command-ing the infantry, cavalry and elephants, who had been elected bythat venerable (king) to the position of viceroy in the province ofPun. d. ravardhana—(when thus in the Kot.ivars.a District) the prefectSvayambhudeva, who had been appointed in this (office) by that(viceroy), was leading the administration of the capital together withthe venerable mayor . . . pala, the merchant Sthan. udatta, the chiefartisan Matidatta, and the first secretary Skandapala—a request wasmade by a man of good family hailing from Ayodhya, Amr.tadeva.688

Various restorations have been proposed for the illegible Gupta name:Budha◦, Vis.n. u◦, Upa◦, Bhanu◦, or Kumara◦,689 but we have argued

687 Basak in EI XV (1919–20), 113; Sircar in SI I, 332. Confusion about the findspotis caused by the editors of the revised edition of CII III (1981), when they describethe findspot as lying in ‘West Dinajpur District, West Bengal’ (p. 282), whichwould imply India.

688 Text based on EI XV, 142f. and SI I, 347f. (emended). Basak (EI XV, 142) readsthis date as G.D. 214 (= ad 533/34).sa〈mva〉 200 20 4 bhadra di 5,paramadaivataparamabhat.t.arakama〈ha〉rajadhirajaśrı ˘ ˘ gupte pr. thivıpatau,tatpadaparigr. hıt(a)〈sya〉 pun. d. ravardhanabhuktav upari〈kamaharaja〉sya 〈maha〉-rajaputradevabhat.t.arakasya hastyaśvajanabhogenanuvahama〈na〉ke ko〈t.iva〉rs.avi-s.a〈ye〉 ca tanniyuktak(a) 〈i〉ha vis.ayapatisvayambhudeve adhis. t.hanadhikaran. a〈m〉arya〈na〉gara〈śres.t.hi〉 ˘ ˘ pala-sarthavahasthan. udatta-prathamakulikamatidatta-prathamakayasthaskandapala-puroge 〈sam. 〉vya〈vaha〉rati,ayodhyakakulaputraka-amr.tadevena vijnapitam〈 |〉

689 Gupta 1974–79 I, 48f.; CII III (1981), 361; Sircar SI I, 346 n. 5. Willis 2005, 145 n. 68assigns the inscription to Budhagupta and argues for an earlier date for this in-scription. He tentatively proposes the reading ‘154,’ but admits that this date ‘doesnot fit the case anymore than 224.’ Another argument advanced by Willis to assignthe Damodarpur Inscription No. 5 to Budhagupta is that the title paramadaivata(found in the inscription at issue) is ‘favoured by Budhagupta.’ This argument isnullified by the fact that, on the one hand, this title is not found, to the best of myknowledge, in any other Budhagupta inscription from elsewhere (in the ShankarpurCopperplate Budhagupta is called paramadeva), and, on the other hand, this titleis also bestowed on the Imperial Kumaragupta in the two Damodarpur inscriptions

Goddesses from the East 243

(above, p. 58) that the most likely candidate would be Vis.n. ugupta, beingthe last king of Magadha of the Imperial Gupta dynasty known to us.

These five documents offer a unique opportunity to monitor continuityand change in the administrative bureaucracy and (religious) economicactivity on the local and regional levels over a period of one century (ad444–543). The administrative continuity can be rendered as follows.

The Imperial Gupta administration in Bengal

pr. thivı Paun. d. ra bhukti Kot.ivars.a vis.aya adhis. t.hanaNo. G.S./A.D. king governor prefect mayor

1 124/444 Kumaragupta Ciratadatta Vetravarman Dhr.tipala2 128/447 Kumaragupta Ciratadatta Vetravarman Dhr.tipala3 163690/482 Budhagupta Brahmadatta – –4 – Budhagupta Jayadatta Śan. d. aka Ribhupala5 224/543 . . . ◦gupta maharajaputra Svayambhudeva . . . ◦pala

Figure 11Administrative machinery in Gaud. a

This survey shows that under Kumara and Budhagupta the office ofuparika-maharaja (governor/viceroy) in Pun. d. ravardhana remained in thehands of the Dattas. During the reign of the last Gupta, however, this of-fice was given to the maharajaputra, replacing the Datta family who hadheld this office for at least half a century. It tells us that by the time ofthe last of the Imperial Guptas the empire had shrunk so much that thecrown prince (maharajaputra) or ‘viceroy’ (devabhat.t.araka) had to con-tent himself with a provincial post.691 The office of vis.ayapati (prefect) ofKot.ivars.a District seems to have changed hands with every new emperor.

On the local level, on the other hand, in the eponymous district capitalKot.ivars.a itself, power remained in the hands of the Pala clan, which heldthe office of nagaraśres.t.hin (mayor) for about one century. The adminis-tration of the town also remained remarkably stable. In addition to themayor, its three-headed governing board consisted during one century ofa distinguished merchant (sarthavaha), artisan (kulika) and administrator

that belong to this ruler. This title is therefore characteristic of the Damodarpurinscriptions rather than of Budhagupta (cf. Virkus 2004, 154 n. 394).

690 The date of this inscription is just Sircar’s conjecture (SI I, 332). Only the figure 3is clearly visible. Basak conjectured 60 (EI XV, 135). These conjectures have beenaccepted in CII III (1981), 336.

691 Cf. Gupta 1974–79 I, 369.

244 Hans Bakker & Yuko Yokochi

The Administration in Kot. ivars.a

nagaraśres.t.hin sarthavaha kulika kayastha pustapalaNo. mayor merchant first artisan first secretary registrar

1 Dhr.tipala Bandumitra Dhr.timitra Śambapala R. s.idatta2 Dhr.tipala Bandumitra Dhr.timitra Śambapala R. s.idatta34 R. bhupala Vasumitra Varadatta Viprapala Vis.n. udatta5 . . . ◦pala692 Sthan.udatta Matidatta Skandapala Gopadatta

Figure 12kot.ivars. adhis. t.hanadhikaran. a

(kayastha). The latter office remained in the Pala family, which clan alsoproduced the nagaraśres. t.hins for a hundred years, whereas in the repre-sentation of the merchants and artisans we see a gradual shift from theMitras to the Dattas. The Dattas served the city administration also on alower level, since they kept the register office (pustapala) for one century,sharing it with the Nandins.693

Finally, constancy is also suggested by the price of fallow (khila) gov-ernment land, which remained remarkably stable during one century ata rate of three gold coins (dınaras) for one kulyavapa.694 Even if we al-low for a decrease of gold content in the coinage of the Gupta kings afterSkandagupta (i.e. ‘inflation’),695 this fixed price testifies to a remarkable

692 Although all authors since Basak (EI) have filled in this name as Ribhupala (R. bhu-pala), this is not very plausible, see below, p. 246.

693 The registry in the five Damodarpur inscriptions: No. 1: R. s.idatta, Vibhudatta,Jayanandin; No. 2: R. s.idatta, Jayanandin; No. 3: Patradasa; No. 4: Vis.n. udatta, Vi-jayanandin, Sthan. unandin; No. 5: Naranandin, Bhat.anandin, Gopadatta.

694 Sircar 1966, 165: ‘An area of land requiring one kulya measure of seed grains tobe sown.’ In Damodarpur inscription No. 3 the price seems to have been loweredto two dınaras, but here there is uncertainty. Basak remarks (EI XV, 136 n. 2)that ‘the word dvaya was at first omitted, but was inscribed probably afterwardsat the bottom.’ Could this have been caused by the fact that the registry officerin charge, Patradasa (see above, n. 693 on p. 244), was new to the profession andlacked the backup of a family tradition? Cf. Virkus 2004, 157.

695 The analysis by Maity 1960 (266–68) shows that the percentage of gold in the coinsof Narasm. hagupta is 65.0–78.8%, Kumaragupta II 70.5% and Vis.n. ugupta 63.6%(gold coins in the Indian Museum in Kolkata). See also Maity 1961 (cf. Gupta1981, 4: Vis.n. ugupta coins with 43% gold in the Bharat Kala Bhavan). The goldcoin ascribed to a ‘Kumaragupta III’ with 50.9% gold falls outside the series givenby Maity. We think the existence of an imperial Kumaragupta III spurious (seeWillis 2005, 140ff.). We are grateful to Ellen Raven for her support in this difficultmatter.

Goddesses from the East 245

local stability.The Damodarpur copperplates examined show few signs of an ap-

proaching end to Imperial Gupta power. Close reading, however, revealssmall variations that seem to indicate great changes in the overall polit-ical reality of Northern India and to point to the perilous confinementof the Imperial Gupta dynasty to the provinces of Magadha and Gaud. a.One such indication is the replacement of the Dattas as governors of thePun.d. ravardhana bhukti by the emperor’s son (maharajaputra). Along withthis shift goes a militarization of the post of governor/viceroy, as appearsfrom the fact that in the 5th Damodarpur inscription it is explicitly statedfor the first time that the viceroy exercised his powers by commanding theinfantry, cavalry and elephants (hastyaśvajanabhogena).

In all five documents the purchase of land is made for religious purposes.The payment in gold coins (dınaras) indicates that we are concerned withthe religious activities of the elite. But here we can observe a conspicuousshift in objective. The two purchases under Kumaragupta’s reign weremade by brahmins to perform their ritual obligations, the Agnihotra andthe five Mahayajnas.696 The third purchase, under Budhagupta, was madeby a village headman (gramika) for the sake of providing residence tosome brahmins. However, the fourth document, dating also from the reignof Budhagupta, records a transaction made by the lord mayor of Kot.i-vars.a himself, Ribhupala (R. bhupala), who purchased land in order to buildthereon two temples (devakula) and appendant storerooms (kos.t.hika), onefor Kokamukhasvamin,697 and one for Śvetavarahasvamin. These (Vais.n. a-va) deities were imported from Nepal, where Ribhupala had earlier madedonations to them.698 The wide appeal of these two deities is not onlyproven by Ribhupala’s devotion, which, after it had earlier brought himto Nepal, made him bring both deities to the neighbourhood of his hometown in Gaud. a,699 but also from the fifth document, in which a man ofgood family (ayodhyaka-kulaputraka) from Ayodhya, the former city ofSaketa,700 Amr.tadeva was allowed to purchase 5 kulyavapas of land for 15dınaras for the sake of one of these two deities, Śvetavarahasvamin andhis temple.701

696 Willis 2009, 97ff.697 Kokamukha (SPBh 110.5), or Kokavaktra, is a servant of Skanda who fights with

Varaha’s son Vr.ka; the latter is a boar in the myth as told in SPBh 109–10, sinceit is addressed as Śukareśvara in SPBh 109.16. For a possible relation of this Koka-mukha and the tırtha Kokamukha see Granoff 2004, 135f.

698 Sircar 1971, 275–281; Willis 2009, 92f.699 The two temples, it seems, were located somewhere in the district, but outside

Kot.ivars.a town (atraran. ye).700 Bakker 1986, 1ff., 33.701 Willis 2009, 92–94.

246 Hans Bakker & Yuko Yokochi

Just like the name of the reigning Gupta king, the name of the naga-raśres. t.hin in the Damodarpur inscription No. 5 is uncertain. The familyname, Pala, is just readable, but the preceding two syllables are hard todecipher. Basak and Sircar (followed by Michael Willis) conjecture ribhu◦,which would make this mayor the same as the nagaraśres.t.hin Ribhupalawho is mentioned as the founder of the temple of Śvetavarahasvamin in theDamodarpur inscription No. 4.702 However, the names of all other officialsin inscriptions Nos. 4 and 5 are different and the nagaraśres.t.hin in theinscription of the time of Budhagupta (No. 4) is not called Arya. Therefore,and in view of the date which, however we read it, is considerably later, it isa dubious matter to fill in the illegible part of the Pala name in the No. 5inscription with a name taken from the Budhagupta inscription No. 4.The arya-nagaraśres.t.hin of the Damodarpur Copperplate Inscription No. 5may have been a son or other member of the Pala family in Kot.ivars.a,thus continuing a family tradition over one century.

Who was this deity Śvetavarahasvamin? This question has been ad-dressed by Sircar and Willis: he is an iconographic type of Varaha, Vi-s.n. u’s incarnation as the Boar.703 The original (adya) temple of this deitystood, as Damodarpur Inscription No. 5 informs us, in Nepal (himavac-chikhare).704 Amr.tadeva’s land donation was meant for repairs to andworship ceremonies at the temple in the Kot.ivars.a district.705

However, another question remains: What’s Śvetavarahasvamin tohim, a kulaputraka from Ayodhya, or he to Śvetavarahasvamin, that heshould provide for him, a deity in a foreign country? Had he the motiveand the cue for passion?

The family of Amr.tadeva may have migrated from Ayodhya to Kot.i-vars.a earlier and have settled there for reasons of its own. One of thereasons may have been that Gupta officials were forced to leave the newly

702 EI XV, 138f., 142; Sircar SI I, 337, 347; Willis 2005, 145 n. 68.703 Sircar 1971,275ff., Willis 2009, 95ff., 282 n. 58: ‘In its early history at least, the

white form of Varaha would appear linked with the Vajasaneya.’ The White Boarmay be identical to Yajnavaraha, described by Rao 1914 I, 135 (on the authority ofthe Vaikhanasagama): ‘This form of Varahamurti has to be of white complexion,and should have four arms, two of which are to carry the śankha and the ca-kra; the figure should be seated upon a sim. hasana, with the right leg hangingdown [lalitasana] and the left resting upon the seat; it should be clothed in yellowgarments and adorned with various ornaments.’ The cosmic boar, Yajnavaraha, issaid to embody all sorts of divinities in the Harivam. śa 31.21–30 (cf. MtP 148.67–77, and SPBh 97.39). The idea goes back to the (Vedic) cosmic boar (emus.a), beinga form of Prajapati in ŚBr. 14.1.2.11, and TaiSa. 7.1.5.1 (see Gonda 1969, 138f.).Agrawala 1963, 29 refers to a concept of Panca-Varahas in which Śveta-Varaha isidentified with Surya.

704 See Sircar 1971,275ff.705 These are discussed in Willis 2009, 92ff.

Goddesses from the East 247

established Maukhari kingdom, because, at the time of Amr.tadeva, Ayo-dhya in all likelihood no longer formed part of the Gupta empire.

Or, maybe the answer should be looked for in what connected the twocities: a system of navigable rivers, the Sarayu–Ganga–Punarbhava basin.Kot.ivars.a may have functioned as the gateway to Gaud. a, and Pun. d. ra-vardhana in particular. Commercial relations may have continued to exist,regardless of the new political reality. An office in the local administrationremained reserved for the merchants (sarthavahas) till the very end.

Another, admittedly less tangible, aspect of Amr.tadeva’s devotion toVis.n. u as the divine Boar may be found in the symbolic power of the Vara-ha imagery. In the face of disaster, when it looks as if the existing orderis falling apart, the rescue of the earth by the Boar is an article of faiththat provides hope. This motivated Dhanyavis.n. u to execute the will of hisfallen brother Matr.vis.n. u and to erect the great Boar on the battlefield ofEran, even under the very eyes of Toraman. a the Hun, as we have discussedabove (p. 31). An ‘Empire of Faith,’ so to speak, although Faith could notsave the Empire.

Similar faith may have motivated Amr.tadeva, when he saw that thekingdom of the Guptas had virtually come to an end and the throneof the personification of the Boar, the Gupta emperor Vis.n. ugupta, wastottering. After the fall of the Guptas, official faith shifted from Vaisna-vism to Saivism, and Kot.ivars.a developed into a famous centre of Goddessworship, as we will see below.

The Bangarh excavationsAs to the identification of Pun. d. ravardhana with Mahasthangarh in theBogra District of Bangladesh, scholars have unanimously followed Cun-ningham’s early suggestion.706

Kot.ivars.a is identified with a site called Bangarh to the north of thevillage of Gangarampur, situated on the river Punarbhava, a tributaryof the Ganges, in the South Dinajpur District of West Bengal (India)—an identification also secured by archaeology.707 The ancient city is known

706 E.g. CII III (1981), 284. Cunningham in ASI Report 1879–89, Vol. XV (1882),110. This is the Pun-na-fa-tan-na of Xuanzang (Beal II, 194f.). The identificationis corroborated by the Mahasthan Fragmentary Stone Plague Inscripion found atMahasthan and featuring the (Prakrit) name ‘Pud. anagala’ (Skt. Pun.d. ranagara)(SI I, 79f.). For excavations in Mahasthan see e.g. Gill 1999; Rahman 2000; Salles2001, Lefevre & Boussac 2007. Pun. d. ravardhana (like Kot.ivars.a) figures in earlylists of the Śakta sacred sites (see Sanderson 2009, 225 n. 507).

707 Goswami 1948, 2; CII III (1981), 285. A copperplate of the 9th regnal year of Mahı-pala I (c. ad 997) found at this site (EI 14, 1917–18, No.23), which records a grantof a village in the Gokalika Man. d. ala in the Kot.ivars.a vis.aya in the Pun. d. ravardha-na bhukti, underscores this identification (cf. Yokochi 2013a, 320). Two copperplateinscriptions of Madanapala have also been discovered from Rajibpur (Bangarh),

248 Hans Bakker & Yuko Yokochi

Plate 61Rajabar. ı Mound (Devıkot.a Citadel) within the fortified area of ancient

Bangarh/Kot.ivars.a on the Punarbhava River

by various names, such as Kot.ivars.a/Kot.ıvars.a, Śon. itapura, Ban. apura andUmavana.708 The old city, the so-called ‘fortress area,’ was explored byAlexander Cunningham (1879–80), who observed:

The fortified portion of the old city consisted of three distinct parts sepa-rated by broad ditches, and surrounded by massive earthen ramparts. Thecitadel, which the people call Devikot, is about 2,000 feet square [= 186 m2],and is so filled with dense jungle that it is quite impossible to penetrateany distance inside except in the very hottest weather, when the grass andunderwood have been burnt, and the tigers and leopards have sought shel-ter elsewhere. [. . . ] To the north of Devikot is a walled enclosure about1,000 feet square [= 93 m2], and to the north of this is a second enclosureof about the same size. Both of these are surrounded by massive earthenramparts and broad ditches. At the north-western corner of the northernenclosure are the ruins of the shrine Shah Bokhari, standing on the top ofa lofty mound, which was certainly the site of a great Hindu temple. HereI found brick walls and many fragments of moulded bricks, together with

which record land grants in the Kot.ivars.a vis.aya (Mukherji 1999, 63; Sanyal 2010,111f.). Also see Sanderson 2001, 7 n. 4.

708 Purus.ottamadeva’s Trikan. d. aśes.a 2.1.17abc: devıkot.o ban. apuram. kot.ıvars.am us. a-vanam | syac chon. itapuram. catha; Hemacandra’s Abhidhanacintaman. i 4.43bcd:devıkot.a umavanam | kot.ivars.am. ban. apuram. syac chon. itapuram. ca tat ‖ Purus.otta-madeva flourished during the first half of the 12th century, contemporaneous toHemacandra (1088–1172) (Vogel 1979, 331f.). According to Vogel, ‘there can belittle doubt about his having been a Bengali by birth and a Buddhist by creed’(ibid., 332). Cf. Goswami 1948, 2.

Goddesses from the East 249

Plate 62Excavations at the Rajabar. ı Mound

six stone pillars of Hindu workmanship.709

The so-called Rajabar.ı Mound or ‘Devıkot.a’ (Citadel) was the subjectof excavations conducted by K.G. Goswami from 1938 to 1941.710 ‘TheRajbari mound, measuring 106 m N-S and 91 m E-W and surrounded bya brick fortification wall, was seen to have five occupation levels (called“strata” by the excavator, dating from the 3rd century BC to medievaltimes.’711 The archaeological examination of Kot.ivars.a (Bangarh), asrecorded in the excavation report by Goswami, further shows that the sitewas occupied from early historic stratum IV (‘Śunga period’), throughstrata III and II to the ‘medieval period’ (stratum I).712 It does not show

709 ASI Report 1879–89, Vol. XV (1882), 95f. Cf. Buchanan’s report in Martin 1838III, 661f.: ‘I first examined the citadel, which is a quadrangle of about 1800 feet[=549 m] by 1500 [= 457 m], surrounded by a high rampart of bricks, and on thesouth and east by a ditch.’ It seems that Buchanan calls the ancient city ‘thecitadel,’ whereas Cunningham’s ‘citadel’ or Devikot apparently corresponds to theRajabar.ı Mound in the centre, called by Buchanan ‘a large heap of bricks, said tohave been the Raja’s house’ (op. cit. 662). See Plate 61.

710 Goswami 1948; EIA II, 47.711 EIA II, 47.712 Goswami 1948; EIA II, 47.

250 Hans Bakker & Yuko Yokochi

a break after the Gupta layers (Stratum III), but rather a continuityfrom the Śunga to the Pala periods, thus corroborating the stability thatwe found at the administrative level of the city. The oldest ‘stratum’represented by ‘a ring-well, stone beads, silver punch-marked and coppercast coins and terracotta figurines,’ goes back to the 3rd century BC(stratum V).

Plate 63Varahı image found near the

Shib Bari site (Bangarh)

For our present purpose it is importantto observe that a substantial number ofterracotta plaques depicting female fig-ures (goddesses) were found, mainly instratum IV (some cast from the samemould),713 as well as terracotta femalefigures in the round, ‘in different lev-els, higher as well as lower,’ describedas ‘very crude and of a primitive na-ture.’714 ‘The fourth stratum of Ban-garh shows distinct signs of the pros-perous condition of the town. The cityhad by then a brick built wide rampartwall (10’ 8” wide). It had drains, cess-pits, and residential buildings, made ofburnt bricks of a very large size.’715

In brief, though the dating is im-precise and the excavations are clearlypreliminary, they evince the origins ofa prosperous town that is represented

in the Damodarpur inscriptions of the fifth and sixth centuries.716

The religious centre of the town, however, seems to have been located900 m north-east of the Rajbari, outside the ‘citadel area,’ around themound called today Shib Bari (Śivabar.ı), where many images have beenfound. They are collected in a shed at the site. The great majority ofthose images are datable to the Pala period, among which are several

713 Goswami op. cit. pp. 20f. See Plates XIX a & b, XX 2a, XXI b 1–5.714 Goswami op. cit. pp. 22f. See Plates XXI a 2, 7. Cf. Bautze 1995; Ahuja 2005.715 Goswami op. cit. p. 34.716 Goswami assigns stratum III to the Gupta Period, ‘which is manifest from the

find of terracotta and other objects in association with it. The peculiarity of thebuildings of this level is that the houses were built of thin walls with roofs of tiles.Another class of massive structure of this period was found in Tr. 1. But that wasprobably meant for some other purpose. Ordinarily the rickety variety was morein general use. The special associated objects of this stratum are the tiles of thethatch or roof for the house, stamped pottery with the design of lotus and conch-shell, and the oval shaped round bottomed pottery vases of medium size etc.’ (op.cit. p. 38)

Goddesses from the East 251

many-armed (ferocious) goddesses, but a slighty older image shows agoddess with animal head, probably Varahı, holding a sword and shieldin her right and left upper hands, a cup in her left lower hand, and whatseems to be a fish in her right lower hand (Plate 63).

Kot. ivars.a in Sanskrit Literature

Vais.n. ava sources of the Gupta age

As we observed above, support for Vaisnavism in Kot.ivars.a, more in par-ticular for the cult of Śvetavaraha, may have been politically motivatedunder Imperial Gupta rule. As such it was in competition with local god-dess worship and emerging Saivism. Kot.ivars.a appears for the first timein the Sanskrit literature of the Gupta age under the name Śon. itapura,‘Blood City.’ In this context the town is connected with a conflict be-tween Śaiva and Vais.n. ava forces, centring around the mythological Ban. a,the demon king ruling the city. In the episode found in the Harivam. śa andVis.n. upuran. a,717 it is told how Kr.s.n. a and his companions attack Ban. aand his demons in this city, in order to rescue the captured Aniruddha,Kr.s.n. a’s grandson. Since Ban. a was a devotee of Śiva and had attained thestate of his Gan. apati (HV 106.6), Śiva, Skanda, and their retinue foughtin league with Ban. a against Kr.s.n. a.718

Furthermore, a goddess Kot.avı appears in the episode in order to pro-tect the demon Ban. a from Kr.s.n. a’s fatal attack, which indicates that Kot.a-vı should be regarded as the tutelary deity of the city Śon. itapura/Kot.i-vars.a.719 She appears naked in front of Kr.s.n. a. Kr.s.n. a reproaches Ban. a and,shutting his eyes, casts his cakra onto Ban. a and cuts off a thousand of his

717 HV 106113, ViP 5.3233; cf. BhagP 10.59–60.718 This episode resembles a similar attack on the city of Varan. ası by Kr.s.n. a as told

in the ViP 5.34 and BhagP 10.63. As in the latter case, this fight between Ban. aand Kr.s.n. a seems clearly to reflect historical rivalry in the late Gupta age betweenVaisnavism and Saivism (see SP II A, 26f.).

719 HV 112.97; she is called daiteyavidya in ViP 5.33.36. In BhagP 10.60.18 her nameis Kot.ara ( but ‘Kot.avı’ in several variant readings) and she is called mata. Kot.avıappears twice in the episode. The reference to her first appearance in order to pro-tect Skanda from a fatal blow by the cakra of Kr.s.n. a (HV 112.49) is problematic.Her appearance is very abrupt and it is difficult to understand what is happen-ing; many manuscripts add several different passages around this verse. Moreover,Pradyumna fights with Skanda and puts him to flight without this goddess’s in-tervention in the ViP and BhagP. The passage referring to her second appearance(HV 112.97–99), though also containing some textual problems, fits better into theimmediate context and has corresponding descriptions in ViP 5.33.36 and BhagP10.60.18.

252 Hans Bakker & Yuko Yokochi

arms (HV 112.100–105). Then Śiva intervenes to save Ban. a’s life. Localgoddess worship is, as we have seen, attested by archaeological evidencegoing as far back as the Śunga period. It will be further corroborated forthe sixth and seventh centuries by our Skandapuran. a.

Kot.ivars.a and the Skandapuran. aSkandapuran. a 68.1–9 describes how a series of goddesses, who had sprungfrom her body in SP 64.19–29, are placed by Kauśikı (an ectype of Parva-tı to be identified with Vindhyavasinı and the main subject of SP III) invarious countries and cities. One of them is the goddess Bahumam. sa, whois stationed in Kot.ıvars.a (SP 68.7c). This Bahumam. sa plays a prominentrole in the Kot. ıvars.a Mahatmya as found in chapter 171 of Bhat.t.araı’sedition of the Skandapuran. a (SPBh 171.78–137). The textual transmis-sion of this Mahatmya is complicated and discussed in Yokochi 2013a.Here we will limit ourselves to a brief outline, in which some repetitionis unavoidable. A critical edition of the two recensions of this Mahatmyaprepared by Yuko Yokochi is presented for the first time in an Appendixto this chapter.

The Kot. ıvars.a Mahatmya (KoM) is preserved in the two main re-censions, S and RA, although only S2 is available for this part of the Srecension. The text, comprising the second half of chapter 171, belongsto the last part of the Skandapuran. a, which was substantially altered andexpanded at the time of the major revision that brought into being thearchetype of the RA recension. This alteration comes to the fore in thefirst half of the Mahatmya, which contains the foundation myth.720 Thetext found in the RA recension is about twice as long as that in the Srecension. In the second half of the Mahatmya, which contains the Ma-hatmya proper, the RA recension runs parallel with S.721 According toour dating of both recensions,722 the S version of the Mahatmya reflectsthe post-Gupta situation around ad 600, whereas the altered RA versionrepresents the period in which Kot.ivars.a was part of the Pala kingdom ofthe ninth century.723

The foundation myth in the first half of the Kot. ıvars.a Mahatmya inthe S recension has one element in common with the story of Śon. itapurain HV and ViP, namely that the town is said to have been occupied bydemons (daityas). These demons, however, are not led by Ban. a, but by

720 KoM(S) 1–23 and KoM(RA) 1–56; SPBh 171.78–116ab.721 KoM 24–44; SPBh 171.116cd–137. The verses after KoM 42c, including the

colophon, are missing in the RA recension. This omission is probably due to an ac-cidental loss of some folios, since verse 42ab continues in the middle of the followingPrahlada myth cycle in this recension.

722 Discussed above, p. 137 and p. 138.723 Majumdar 1971, 101–31.

Goddesses from the East 253

Haimakun. d. a, and their expulsion from the city is effected not by Kr.s.n. a,but by the Mothers under the authority of Śiva. The story is briefly asfollows (KoM(S) 1–23).In reply to Vyasa’s question as to how Śiva had become Lord of the Mothers andwho these Mothers are, Sanatkumara tells of the genesis of the Mothers.

Brahma once performed the Sam. dhya ritual by the side of a vast mass ofwater in the east (purvamahodadhi) during one crore of years (vars.akot.i).724

After finishing his ritual, he founded a city in that beautiful region (tasmin ramyepradeśe) and declared that the people who will live there will become invincibleand remain forever young. After Brahma had gone home, demons headed byHaimakun.d. a came to live in this city. They oppressed the brahmins and destroyedtheir rituals. The gods complained to Brahma and together they go to a forestin the Himalaya mountains in order to see Bhava (i.e. Śiva).

Here they see Gaurı (i.e. Parvatı) practising tapas. As a result of the powerof her asceticism, all the gods are transformed into women. Asked by Śiva whythey have come, the gods complain about the oppression by the demons andbeg him to reinstate their manhood. Śiva orders them to approach the demonsas women, because the demons can only be killed by women. Their manhoodwill be reinstated after having created embodiments of their female forms, whichembodiments will become the Mothers.

The gods asked him to create a female embodiment of himself, whereuponŚiva creates the Mother Rudran. ı and, assuming a deformed form, the secondMother Bahumam. sa. At Śiva’s command the following Mothers come into being:Vis.n. u creates (1) Varahı and (2) Vais.n. avı. (3) Brahmı was born from Brahma, (4)Śarvan. ı (= Rudran. ı) from Śam. kara, (5) Kaumarı from Skanda, and (6) Mahendrıfrom Indra. The greatest of the Mothers was born from Vr.s.adhvaja (i.e. Śiva),viz. (7) Bahumam. sa. Subsequently, thousands of Mothers—Vayavya, Varun. ı,Yamya, Kauberı, Mahakalı, Agneyı, and so forth—issue from the bodies of theother gods, each being the female counterpart of their male origin.

Thereupon these goddesses go to the town (of Kotıvars.a) and kill the demons.

When we collate the S version of the foundation myth with the story astold in RA the following changes can be noted.

(1) The gods are transformed into women at Umavana ‘Uma’s forest,’which is mentioned as one of the names of Kot.ivars.a by Hemacandra(n. 708 on p. 248). There is no reference to Parvatı’s tapas. (2) The godswho are transformed into women enter into Śiva’s belly and are emittedfrom Śiva by means of Angiras’s intervention. The introduction of Angirasinto the story suggests an influence of the Atharvavedins in the region.725

724 Udadhi or mahodadhi is usually translated as ‘the sea, the ocean,’ and purvamaho-dadhi may mean ‘the eastern ocean,’ i.e. the Bay of Bengal. The reason to take theword literally here is that the Brahmaputra River, whose epithet is Lohita, is alsodesignated as udadhi in MBh 3.215.22. See below, n. 727 on p. 254 and Chauduryand Sircar in EI 29, 151. Yokochi 2013a, 304–06 discusses the possibility thatthe daughter (kanya) of Lohita, who is mentioned as one of the Mothers of Skanda(MBh 3.215.22, 3.219.39), may have been a forerunner of the goddess Bahumam. sa.

725 See Schmiedchen 2007, 361f., for the copperplate inscriptions that record some land

254 Hans Bakker & Yuko Yokochi

(3) The standard seven Mothers in the S recension become the unusual fiveMothers (Brahmı, Rudran. ı, Kaumarı, Vais.n. avı and Varahı). Bahumam. sadisappears not only in the foundation myth, but also in the following Ma-hatmya.726 On the other hand, the total number of the Mothers seems tohave increased, including several names that are not derived from malecounterparts, such as Vr.ddhapramatr., Mr.s.ika, and, last but not least,Camun.d. a (KoM(RA) 50–51). (4) Śiva as the husband of the Mothersis described in a Bhairava form with many arms, legs, faces and bodies,carrying a skull-bowl and wearing a wreath of severed hands, feet andheads (KoM(RA) 46–47).

In the following second half of the Mahatmya, Śiva comes to bestowfavours on the Mothers. He himself will remain present in Kot.ivars.a asHetukeśvara. Bahumam. sa, who had been leading the Mothers and killedthe demons with her lance (śula), is represented by the Śulakun. d. a. Chaga-la, a goat deity, and Kumbhakarn. a are the doorkeepers of the Mothers,and the female devotees who worship the Mothers in secret rites (rahasya)will become Yogeśvarıs.

From the text summarized above and presented in the Appendixit can be concluded that in around ad 600 the main (male) deity ofKot.ivars.a is Śiva Hetukeśvara, who might have been represented by alinga,727 but the worship of goddesses, called the Mothers, is paramount.

grants to the Atharvavedin Brahmins in the Pala period. An account of the trans-mission or revelation of the Brahmayamala scripture (see below, p. 257), narratedin its first chapter, may be relevant to this element. This account incorporatestwenty-eight individuals from various castes and regions, of which four are Atha-rvan. a brahmins; these four are all disciples of Padmabhairava and three of themcome from Madhyadeśa (Hatley 2007, 228–232, Table 4.8).

726 The two references to Bahumam. sa in KoM 31b and 32b are changed into differentreadings in the A MSS; they are lost in R, along with many verses before andafter KoM 31 and 32. The variant readings of A might be due to an accidentalcorruption but, taking into account the disappearance of this name in Part 1, theyare more likely due to intentional alterations made at the time of the compositionof the hyparchetype of the RA recension. Cf. Yokochi 2013a, 314.

727 As he is grouped with the Mothers and called their leader, it is also possible thathe was represented in an iconic form, flanking the image of the seven Mothers (cf.Hatley 2007, 58). The sanctuary of Hetukeśvara was rebuilt several times in thePala period, as seems to be attested in (1) Parbatiya Plates of Vanamalavarmanof Pragjyotis.a in Assam, middle of the 9th century (EI 29, No. 20 by Chaudhury& Sircar, vs. 24). The king, who was a Paramamaheśvara, rebuilt the temple ofHetuka-Śulin that had fallen down, patita (EI 29, 154 vs. 24). However, compareYokochi 2013a, 315, who doubts whether the temple in Kot.ivars.a is implied. Theinscription contains a reference to the river Brahmaputra as Lohityasindhu (EI 29,135). (2) Syan stone slab inscription of Nayapala, c. ad 1027–43 (EI 39, No. 7, bySircar). The inscription mentions many temple constructions by a king (probablyNayapala). It testifies to the establishment of the Śaiva Siddhanta in Kot.ivars.a(Yokochi 2013a, 315).

Goddesses from the East 255

Plate 64Camun. d. a type of image

(Bahumam. sa/Carcika?) found nearthe Shib Bari site (Bangarh).

Pala period

We find the births of the seven Moth-ers in accordance with the standardSaptamatr.kas: Brahmı, Śarvan. ı, (i.e.Maheśvarı), Kaumarı, Vais.n. avı, Vara-hı, Mahendrı (i.e. Aindrı), and Bahu-mam. sa who takes the position of Ca-mun. d. a. These seven are put in versesKoM(S) 19–20 in an order that cor-responds to the regular iconography,which, again, shows that this descrip-tion is consciously based on the stand-ard icon of the Seven Mothers. Wethink that the passage owes its signifi-cance to the fact that it is the earliesttextual reference to the standard SevenMothers, notwithstanding the fact thatimages testifying to this iconographyhave made their appearance since thefifth century, having their roots in ear-lier groups of goddess images of theKus.an. a period.728

The figure of Bahumam. sa deservessome further attention. Her prominentstatus in the Kot. ıvars.a Mahatmya sug-gests that Bahumam. sa represents agoddess worshipped locally as a tute-lary deity of Kotıvars.a. Her positioncorresponds to Camun.d. a in the stand-ard Seven Mothers and she must have been a goddess similar to Camun. d. a,

No doubt there would have been other Śaiva sanctuaries in the sixth and seventhcenturies. The list of twenty-eight incarnations of Śiva (see above, p. 214) men-tions Dan. d. in (probably a corruption of Din. d. in) Mun. d. ıśvara (cf. SP 34.44a) as thetwenty-fifth incarnation, who is said to have been born in Kot.ıvars.a in VaPAA

23.208–211 (≈ LiP 1.24.114c–117b, ŚiP Śatarudrasam. hita 5.37–38). This passagementions Chagala and Kumbhakarn. a, known from KoM 39, as two of the four sonsof this incarnation who will attain maheśvaram. yogam here.

728 Van Lohuizen 1990, 13. Three panels of the Seven Mothers located in UdayagiriCaves IV and VI, in which some emblems can be recognized in two of the panels, aredatable with relative certainty to about ad 400. See Harper 1989, 75–79, Figs. 16–30. A group of five images, probably some of the Seven Mothers of this type, housedin the Baghbhairava Temple at Kirtipur in Nepal, may be attributed to the fourthcentury (Pal 1974, 129, Fig. 215; Harper 1989, 74–75). For the development ofthe idea of the standard Seven or sometimes Eight Mothers, see Harper 1989 andPannikkar 1997.

256 Hans Bakker & Yuko Yokochi

who is usually represented as having an emaciated and dreadful appear-ance and as being horrific in nature. The statement that Śiva assumeda deformed shape to create her underlines her ugliness, and her dreadfulnature and appearance are apparent from an adjective qualifying her asan embodiment of the destruction of the world in KoM(S) 17e. The ap-pellation Bahumam. sa may be taken to mean ‘Well-fleshed Lady’ or ‘Ladywho eats much meat,’ but it seems very likely that this name was given toher as a euphemism, referring to her emaciated appearance and the factthat she received animal sacrifices on a large scale. She must have been awild goddess.729

Bahumam. sa may thus have been the local variant of a popular type ofMother Goddess that became known as ‘Camun. d. a’ in later iconographictexts by convention. This observation is based on the fact that the name‘Bahumam. sa’ is completely omitted from the later RA version of the Ma-hatmya, while, on the other hand, the name of Camun.d. a figures therein,although not in the same position.730

As Alexis Sanderson observes, Kotıvars.a is ‘one of the set of eightKs.etras, or primary sites of the Goddess, venerated in our earliest sources’and also, in some texts, one of the eight Śmaśana sites.731 In conformitywith this observation, the Kotıvars.a Mahatmya describes the place as ‘dearto the Mothers’ and ‘an eminent Śmaśana’ (KoM 40).

729 In KoM 33 Śiva promises Bahumam. sa: ‘The lovely Mandakinı that is called Prati-kula [i.e. the Punarbhava], that river will flow here carrying a stream of blood(rudhiraughavatı) owing to you.’ Geographically speaking, the red colour of theriver is probably due to the river flowing back at high tide in the rainy season, whichis sometimes (in South India) regarded as the menstruation of the river goddess.The composer of the Mahatmya may have attributed the colour to Bahumam. sa’skilling of the demons. The practice of animal sacrifices may have reinforced thisimage.

730 The name Bahumam. sa disappears from textual and archaeological sources. TheBangarh stone inscription of the time of Nayapala (Sircar 1973–74, 1980–82) refersto the goddess as ‘Carcika,’ whereas in many Śaiva texts the goddess in Kot.ivars.a iscalled ‘Karn. amot.ı’ (see Yokochi 2013a, 312f.). The research of Gudrun Melzer hasconvincingly shown that: ‘The Dinajpur region (Uttar Dinajpur, Dakshin Dinajpurand the Dinajpur district in Bangladesh) seems to have been a stronghold of Ca-mun. d. a worship. It is also the region where several multi-headed and multi-armedBhairava images have been discovered. The provenance of at least 24 Camun. d. aimages can be traced to this region’ (Melzer 2009, 142). Melzer’s use of the name‘Camun. d. a’ for these images, however, is based on the later iconographic convention;for most of these images we do not know how the deity that they represent wasaddressed.

731 Sanderson 2001, 7 n. 4. Hatley 2007, 63 (BraYa 84.81): prayaga varun. a kolla at.t.a-hasa jayantika | caritraikamrakan caiva kot.ivars.am. tathas. t.amam ‖ 81 ‖ For identi-fication of these places see Sanderson 2001, 7 n. 4 and Hatley 2007, 232–34.

Goddesses from the East 257

The Matr. tantras or the seven Yamalas

That goddesses play an important role in the Skandapuran. a is known fromSP III. And that the Paśupata movement had close links with goddess wor-ship from an early beginning has been discussed above, p. 171. However,that the Skandapuran. a makes explicit mention of a genre of texts out-side the epic/puranic tradition, texts that are concerned with the ‘secret’(rahasya) rituals of Mother worshippers and are, by and large, at oddswith the Paśupata tradition, that is surprising in its uniqueness. In theProlegomena to the first volume of the Skandapuran. a edition we notedthe following: ‘The SP does however show awareness of Tantric Śaivism ofsome sort. See especially the intriguing reference to and list of what arecalled matr. tantras in SPBh 171.127–131ab.’732

Here follows a translation of Śiva’s address to Bahumam. sa and theother goddesses (KoM 34–38).

I along with Brahma, Vis.n. u and the sages rich in asceticism,we will compose divine and holy Matr.tantras, (which teach) thesupreme procedure of sacrifice to the Mothers, according to whichyou will be worshipped. These Tantras of yours, such as the Brahma,Svayambhuva, Kumara-yamala, Sarasvata, Gandhara, Aiśan. a, andNandi-yamala,733 along with thousands of others, will come into be-ing, and people will make oblations to you in accordance with them.

You, O goddesses, will fulfil the wishes of and grant divine powers tothose who worship you, and you will possess divine yoga. And thosewomen who always make offerings to you in esoteric rituals (sara-hasyatah. ), they will become Yogeśvarıs, beautiful and possessed ofdivine power.

We follow Sanderson in his interpretation of this passage to the effect thatwe take the seven Matr.tantras mentioned to be seven Yamala-tantras, ofwhich five are known by name from later Tantric literature.734 It seemsthat at least one of the texts mentioned is still extant, though probablynot in its most original form, namely the Brahmayamala, which is also

732 SP I, 4 n. 9.733 The two references to Yamala in KoM 35b and d are both changed or corrupted

in the R and A recensions: vayukam. in R and vamanam. in A in 35b; ◦vardhanamin both R and A in 35d.

734 SP I, 4 n. 9. Sanderson 2001, 7 n. 4: ‘I propose that yamalam. in 128d [= KoM35b] is to be understood with all (sarvaśes.ataya), so that the titles indicated areBrahmayamala, Svayambhuvayamala, Kumarayamala (= Skandayamala), Saras-vatıyamala, Gandharayamala, Iśan. ayamala and Nandiyamala. Of these seven, allbut the Svayambhuvayamala and the Gandharayamala are found in the scripturallists of Yamala-tantras known to me.’

258 Hans Bakker & Yuko Yokochi

known as the Picumata.735 There are various texts going by the name ofBrahmayamala, but the oldest of them is the one which is preserved in aNepalese manuscript dating from ad 1052.736

This Yamala has been the subject of a thorough study, and has partlybeen edited by Shaman Hatley in an unpublished dissertation.737 Centralto this Yamala is the cult of Yoginıs, and as such the text claims tobelong to the Vidyapıt.ha division of the Bhairavatantras. With this sortof text we leave the Atimarga tradition of the Paśupata and enter a newphase in the development of Saivism, the Mantramarga. Evidently theregion around Kot.ivars.a, i.e. Gaud. a, was a hotbed in which this newdivision could develop, when local variants of goddess cult accommodatedto burgeoning forms of Tantric Saivism.

This poses the question of the relative date of the Brahmayamala vis-a-vis our Skandapuran. a. This question has been carefully investigated byHatley, who comes to the following conclusion.

Significant uncertainties thus surround both the absolute and rela-tive chronologies of the Brahmayamala, further complicated by thefact that the text appears to have multiple strata. It does seem highlylikely that the BraYa, in a form close to that preserved in its oldestNepalese codex, existed at some point in the eighth century, whileI see no grounds for ruling out a seventh-century dating, especiallyfor the older portions of the text. This period in fact appears quiteplausible. The possibility that some form of the BraYa existed in thesixth century also merits consideration, especially given its mentionin the old Skandapuran. a, while it is also not impossible that theBraYa continued to develop into the early ninth century; this possi-bility depends upon the dating of sources such as the Laghuśam. varaand Tantrasadbhava.738

Regarding the provenance of the Brahmayamala Hatley is equally cau-tious, admitting that: ‘At the present juncture, I would assert little morethan that the text seems unlikely to herald from the far south or far northand northwest of the subcontinent.’739

In view of the evidence of the Kotıvars.a Mahatmya in the Skandapura-n. a, it appears possible to go one step further and advance the hypothesisthat the early beginnings of this and similar Matr.tantras is to be sought

735 For the meaning and significance of this name see Hatley 2007, 243ff.736 National Archives in Kathmandu accession no. 3–370. Petech 1984, 44. NGMPP

Reel no. A 42/6. Palm-leaf. Date verified as Sunday, 12 January, 1052.737 University of Pennsylvania, 2007. The dissertation is accessible through

UMI/ProQuest (http://www.proquest.com/en-US/default.shtml).738 Hatley 2007, 228. For this text see also Torzsok 2011, 359f.739 Hatley 2007, 228.

Goddesses from the East 259

for in Gaud. a, more or less contemporaneous with the composition of theSkandapuran. a. In the days when Gaud. a was asserting itself in the politicalarena of Northern India at the turn of the 7th century, while Śaśanka wasbeing settled in his capital Karn. asuvarn. a 150 km south of Kot.ivars.a—afterhaving broken away from the rule of Kanauj, having usurped the LaterGupta domains in Bengal, and having conquered Magadha (see above,p. 89)—it seems plausible in our view that at this juncture of history anew self-awareness gave a boost to domestic religious trends, which wouldcome to dominate the centuries to come. It may have been the reasonwhy the Aryamanjuśrımulakalpa referred to this regon as mleccharajya(above, p. 101). Somehow the authors of the Skandapuran. a thought itwas warranted to acknowledge this new departure at the end of theircomposition.

This is not to say that the Tantric cult of Mother Goddesses wasrestricted to Bengal. As we have seen above (p. 171), the Tantric teacherBhairavacarya, who in Ban. a’s account was at the root of the power of theVardhana dynasty of Thanesar, had taken up his residence near a Matr.temple.740 Nor did this goddess cult flourish without a central male deity.Both the Kot. ıvars.a Mahatmya and the Brahmayamala acknowledge thepivotal role of Śiva-Bhairava in his manifestation as Hetukeśvara in Kot.ı-vars.a. As such he is found in the north-eastern corner of the Mahaman. d. alathat is described in the Brahmayamala, surrounded by six Yoginıs.741 Thesame passage mentions Chagala, Kumbhakarn. a and the pond called Śulo-daka, thus showing familiarity with the local tradition as known from KoM31, 39, though the association with Bahumam. sa appears to be missing, asis the case in the RA recension of the Mahatmya.

A date for the part of the Brahmayamala under discussion that isslightly later than the composition and first redaction of the Skandapuran. acould possibly explain this omission, although, admittedly, other reasonsfor the disappearance of the local tutelary goddess in the Picumata, suchas the difference in context, cannot be excluded.742 The matter needsfurther research, but for the time being our working hypothesis is thatat least this part of the Brahmayamala came into being in the secondhalf of the seventh century, when the Later Gupta dynasty was tryingto restore authority in the region after the collapse of Hars.avardhana’sempire, though the inception of the Yamala tradition seems to have beencontemporaneous with the composition of the Skandapuran. a. The RArecension of our text, on the other hand, shows some influences from this

740 The ritual described by Ban. a (see above, p. 78) may be compared to the onedescribed in the Brahmayamala as the Mahamanthana (see Torzsok 2011, 360).

741 BraYa 3.119–127 presented in Yokochi 2013a, 311f. (based on a preliminary editionby Alexis Sanderson). See below, n. 743 on p. 260.

742 Cf. Yokochi 2013a, 313f.

260 Hans Bakker & Yuko Yokochi

tradition, not only in its reference to the Atharvavedins (above, n. 725on p. 254), but especially in its depiction of the Bhairava form of Śiva asbeing surrounded by goddesses, found in the foundation myth of the RArecension (KoM(RA) 44–48ab), which seems to echo the concept of theMan. d. ala centred by Bhairava in the Brahmayamala.743

The later Skandapuran.a transmission and the Pala kingdom

Kot.ivars.a remained a town with a strong religious profile during the eighthand ninth centuries. This not only follows from the rich layers of stratumII in Goswami’s excavation, but in particular from the wealth of imagesfound in the environs of the Shib Bari site that can be related to the Palaperiod, images that do not exclusively belong to the Śakta-oriented Śaivafold, but also include some Bauddha and Vais.n. ava ones.

The dynasty that gave its name to this period rose to prominence inthe second half of the eighth century, after a century in which, in the wordsof the great Indian historian R.C. Majumdar, ‘anarchy and confusion [had]prevailed in Bengal.’744 As we have argued above (p. 138), it might havebeen under the first great Pala ruler, Dharmapala, that the b version ofthe S hyparchetype was brought to Nepal, the ancestor of our MSS S2 andS3/S4.

The provenance of this Pala dynasty is uncertain: Majumdar is in-clined to lend credence to a tradition recorded by Taranatha that Gopala(Dharmapala’s father) was born near Pun. d. ravardhana in the middle of theeighth century.745 We venture the conjecture that the Kot.ivars.a Districtcould be a good candidate for this honour, in view of the uninterrupted,century-long presence of a Pala clan there, at least until the last days ofthe Imperial Gupta age, and the imperturbable way in which it apparentlywielded power in the capital of this district, even when the political worldaround it was falling apart.

Be that as it may, operating from Gaud. a, Dharmapala expanded histerritory to Magadha, and both provinces would remain united in thePala kingdom during most of the following century.746 From this period

743 See Yokochi 2013a, 313. BraYa 3.123–26ab: karn. ikayam. mahadevi mahabhairava-rupin. am | rudras. t.akasamopetam. purvavad devi calikhet ‖ rudran. am. bahyato deviyoginyapadmam alikhet | hetukı kampanı caiva ks.obhin. ı bhımavikrama ‖ kros.t.ukıbhadrakalı ca yoginya s.at. samalikhet |

744 Majumdar 1971, 94.745 Majumdar 1971, 168.746 Majumdar 1971, p. 123: ‘Three inscriptions of Narayan. apala (B 16, 17, 18) dated

in the years 7, 9 and 17, and found in Bihar, seem to prove that the kingdom ofMagadha was in his possession at least up to his 17th year, i.e., c. 870 A.D.’

Goddesses from the East 261

onwards, the transmission of the Skandapuran. a held its ground in thispart of Northern India, as the provenance of the surviving manuscriptsoutside Nepal prove.747 On the basis thereof, we proposed the hypothesisthat the redaction of the RA recension was made within the Pala king-dom of the ninth century (p. 138). It may have been no longer an entirelyup-to-date reflection of the religious conditions of its time, considering thegeneral tendency of this recension to compensate for a lack of geograph-ical precision by an increase in mythological material—a tendency firstobserved by Peter Bisschop (2004, 70)—but it preserved for many futuregenerations, including our own, the world in which the Skandapuran. a wascomposed and spread.

vyasa uvaca | katham. matr.gan. a jatah. . . .

Plate 65Folio (322v) of A7, Dhaka University Library MS 3376.

Paper, Bengali script. Opening words (highlighted) of theKot.ıvars.a Mahatmya in the RA recension

747 Yokochi 2013b, 3f.

Appendix

d The Kot.ıvars.a Mahatmya of the Skandapuran. a d

Edited by Yuko Yokochi (Kyoto)

Introduction

The present edition is based on the following manuscripts: S2 for the S recension(S1 and S3 are lost for this part) and R, A3, A4 and A7 for the RA recension. Thesiglum A represents all three manuscripts of the A recension, and the siglum Bhthe reading in Bhat.t.araı’s edition (SPBh 171.78–137). Variants in the manuscriptsare recorded only when they are significant, even in cases where the acceptedreading is an emendation/conjecture by either Bhat.t.araı or the present editor.

Although Bhat.t.araı follows S2, in principle, the introductory part of his edi-tion (SPBh 172.78–96ab) is considerably different from S2, because there he mixesup S2 and A3, which results in a conflated, new text. These variants are notreported in our edition here. For the description of the manuscripts, see theProlegomena of SP I (pp. 31–38), the Introduction of SP II A (p. 10f), and theIntroduction of SP III, 33–66, 144.

Part I (S recension): KoM(S) 1–23

vyasa uvaca |bhagavan parameśano matr.n. am. nayakah. katham |kah. eta mataraś capi śam. saitat pr.cchato mama ‖ 1sanatkumara uvaca |atra te vartayis.yami matr.n. am udbhavam. śubham |pura hi bhagavan brahma tıre purvamahodadheh. |ramye cakara sam. dhyam. hi vars.akot.im atandritah. ‖ 2kr.tasam. dhyas tato devah. kamalasanasam. sthitah. |tasmin ramye pradeśe tu cakre puram anuttamam ‖ 3tat sr.s.t.va tu puram. brahma sauvarn. am. man. ibhus.itam |provaca vacanam. vyasa paramadbhutam arthavat ‖ 4ramye puravare ye ’smin vatsyanti sukr.taśinah. |avadhyas te jarahınas tejobalasamanvitah. |bhavis.yanti mahabhaga matprasadan na sam. śayah. ‖ 5evam. sa bhagavan uktva brahma lokapitamahah. |

263

264 Yuko Yokochi (ed.)

agacchat svasabham. divyam. naikasiddhanis.evitam ‖ 6atha daityah. samagamya haimakun.d. apurah. sarah. |tasmin puravare sthitva pıd. am. cakrur anekaśah. ‖ 7dves.t.aro jagatas te hi jaghnur vipran sahasraśah. |yajnahomadikaś chinnah. kriyas tair asuradhamaih. ‖ 8tan evam. karin. o jnatva sarva eva divaukasah. |brahmalokam athabhyetya sarvam aśravayan vibhoh. ‖ 9vaco divaukasam. śrutva jnatva daityavices.t.itam |naitac chakyam. mayety uktva brahma sardham. surottamaih. |ajagama bhavam. dras.t.um. ramyam. himagirer vanam ‖ 10yatra gaurı tapas tepe bhartr.dehardhadharin. ı |pravis.t.as tatra te deva brahmadyah. sarva eva hi |striya evabhavan turn. am. parvatyas tapaso balat ‖ 11tatas tan devadeveśah. strıbhutan hi surottaman |kim artham agata yuyam iti papraccha śuladhr.k ‖ 12tatas te sahitah. sarve purvam. daityavimardanam |nivedya śambhave paścat kr.cchram. strıbhavam atmanah. |yathaivam. purus.ah. purvam. bhavama iti te ’bruvan ‖ 13tan uvacatha deveśah. suran strıbhavam agatan |yata yuyam. surah. sarve striyo bhutvasurantikam |strın. am eva hi te vadhya danava baladarpitah. ‖ 14tanur murtimatıh. kr.tva parato matarah. śubhah. |yuyam eva yathapurvam. bhavis.yatha surottamah. ‖ 15atha bhuyah. pran. amyeśam. devata idam abruvan |atmano ’pi tanum. deva kuru murtimatım. striyam |yaya sardham. balad daityan ham. syama parameśvara ‖ 16tato devo ’sr.jad devım. rudran. ım. mataram. śubham |vikr.tam. rupam asthaya dvitıyam api mataram |namna tu bahumam. sam. tam. jagatsam. hararupin. ım ‖ 17niyogad devadevasya tato vis.n. ur api prabhuh. |matarav asr.jad dve tu varahım. vais.n. avım api ‖ 18abhut pitamahad brahmı śarvan. ı śam. karad api |kaumarı s.ad.mukhac capi vis.n. or api ca vais.n. avı |varahı madhavad devı mahendrı ca puram. darat ‖ 19sarvatejomayı devı matr.n. am. pravara śubha |bahumam. sa mahavidya babhuva vr.s.abhadhvajat ‖ 20sarves.am. devatanam. ca dehebhyo matarah. śubhah. |svarupabaladharin. yo nirjagmur daityanaśanah. ‖ 21vayavya varun. ı yamya kauberı ca mahabala |mahakalı tathagneyı anyaś caiva sahasraśah. ‖ 22ta gatva tat puram. ramyam. daityan bhımaparakraman |jaghnur bahuvidham. devyo ghoranadair vibhıs.an. aih. |daityahınam. ca tac cakruh. puragryam. hemabhus.itam ‖ 23

Kot. ıvars.a Mahatmya 265

Part I (RA recension): KoM(RA) 1–56748

vyasa uvaca |katham. matr.gan. a jatah. kim. nimittam athapi ca |kim. vıryah. kimprabhavaś ca sambhutas te gan. ah. prabho ‖ 1kaś casam aśrayo yena vartante tadgan. ah. śubhah. |kena sr.s.t.ah. svayam. vapi etad icchami veditum ‖ 2sanatkumara uvaca |atra te vartayis.yami itihasam. puratanam |yac chrutva papakarmapi gacched dhi paramam. gatim ‖ 3pura brahma prajadhyaks.ah. sr.s.t.vema vividhah. prajah. |aveks.aman. ah. kaleya cacara vibudheśvarah. ‖ 4sa kadacit prajadhyaks.ah. sam. dhyakala upagate |deśam. ramyam. samasadya sam. dhyam. cakre mahamatih. ‖ 5sopasya tam. mahayogı bhagavan sarvasr.k prabhuh. |cintayamasa ramyo ’yam. deśo dr.s.t.imanoharah. ‖ 6sraks.yamy atra puram. divyam. sarvakamasamanvitam |iti sam. cintya sa vyasa puram. tatra pracakrame ‖ 7prakarais tribhir atyusrair vistırn. air man. ibhus.an. aih. |lohatamrakasauvarn. aiś caturasram atiprabham ‖ 8

8ab] om. A

at.t.aiś cat.t.alakaiś caiva khatakaiś ca samanvitam |śr.ngat.akaiś catus.kaiś ca as.t.apadam iva sthitam ‖ 9vaprakhatakasam. yuktair gr.haiś capi mahadhanaih. |˘ ˘ ˘ ˘ ˘ ˘ ˘ ˘ ˘ ˘ ˘ ˘ ˘ ˘ ˘ ‖ 10

10cd] loss of two padas conjectured.

madhye mandakinı capi sarid divya manohara |nalinıvanasam. channa prahr.s.t.akhaganadita ‖ 11tat puram. sa tada sr.s.t.va devah. sarvasurottamah. |provaca ye ’tra vatsyanti te ’ks.ayah. sukhino narah. |bhavis.yanti avadhyaś ca purn. aśaś ca na sam. śayah. ‖ 1212d narah. ] conj., varah. R, ’marah. A; f purn. aśaś ca] conj., purn. n. o martya R, purn. a martya A

tatah. sa bhagavan deva uvaca punar eva ha |asmin sadaiva kamanam is.t.anam. surasattamah. |devo vars.ati kot.im. vai tenasmin sukhino janah. ‖ 13tasmin gate tato vyasa asurair devabhıs.itaih. |adhis.t.hitam. puram. tad vai sarvair brahman. akan. t.akaih. ‖ 14yes.am. patir abhut krodhah. sarvaśastrabhr.tam. varah. |tato halahalaś caiva dan. d. apan. iś ca viśrutah. ‖ 15

15cd] om. A

brahmaha yajnaha caiva devantakanarantakau |sim. hanado bhımanado mada ity evamadayah. ‖ 16te tat puram. samasadya sahadevan mahasurah. |anacara duracara badhante sarvaśo janan ‖ 17badhyamanes.u lokes.u deva brahman. am abruvan |

748 The passages preserved only in one of the sub-recensions are in a smaller font. Thesymbol † denotes an irreparable corruption.

266 Yuko Yokochi (ed.)

bhavata sa kr.to ’nartho yena badhyamahe vayam ‖ 18upayam. nah. prayacchadya yena mucyema sam. śayat |tvam. hi nah. paramo devah. sarvadevapitamaha ‖ 19sa evam uktah. kaleya brahma lokagurus tada |saha devair mahateja rudram. śaran. am agatah. ‖ 20tan apatata evatha bhagavan nılalohitah. |umavane sa deveśah. suran sarvan avaiks.ata ‖ 21te ’pi deva mahasattvah. satvaram. samupagatah. |strıbhavam. samanuprapta idam. rudram. tadocire ‖ 22

22a–23b] om. Avayam. lokahitarthaya bhavantam. śaran. am. gatah. |strıbhavam. samanupraptas trahi nah. sarvatah. prabho ‖ 23sa evam uktah. kaleya devaih. strıbhavam agataih. |uvaca mam. viśadhvam. vai tatah. śreyo hy avapsyatha ‖ 24tatas te vis.n. usahitah. sayamah. sahabhaskarah. |devasyodaram aviśya tasthur vars.agan. an bahun ‖ 25atha te danavah. sarve kot.ivars.anivasinah. |abhidravanta lokes.u vipran kruren. a karman. a ‖ 26te vadhyamana viprendrah. śaran. atran. avarjitah. |ucur angirasam. vyasa danavan sam. nisudaya ‖ 27sa tais tapodhanaih. sarvaih. samapuritamanasah. |maheśvaram. gan. adhyaks.am avahayati kalija ‖ 28

28c gan. adhyaks.am] conj., sanadhyaks.am R, prajadhyaks.am A

sa tena vidhina mantrair ahutah. kalinandana |uvaca kim. karomy adya bruhi mam. dvijasattama ‖ 29evam uktas tada vyasa bhagavan angiras tada |uvaca trahi viprendram. l lokam. ś ca sakalan prabho ‖ 30evam uktas tada vyasa bhagavan ambikapatih. |cakarapratimam. rupam. yugantagnisamaprabham ‖ 31tadrupam. sa tada kr.tva vyapya sarvadiśaś ca ha |at.t.ahasam. prakurvan. o vaktran naryo ’sr.jat prabhuh. ‖ 32ye te devas tada vyasa strıbhuta devakos.t.hagah. |ta eva nih. sr.ta bhuyo devadevam upasthitah. ‖ 33matr.tam. gacchata praha †balam. capratimam. bhavah. † |loke ca pujanıyatvam. svastyartham. capy ayatnatah. ‖ 34punar aha puram. gatva kot.ıvars.am atandritah. |danavan bhaks.ayadhvam. vai etad vah. kr.tyam udyatam ‖ 35

35b kot.ı◦] A, kot.i◦ R

ta evam ukta devena ucuh. sarva maheśvaram |atmano yaccha no deva balam. capratimam. yudhi |tejaś ca sumahad ghoram. yena tan sudayemahi ‖ 36tatah. sa bhagavan devas trinetras tripurardanah. |tr.tıyanayanan naryah. pancaiva sa cakara ha ‖ 37dan.d. ahasta triśulastra śaktyastra cakradharin. ı |varaharupa pancaitas tr.tıyanayanodbhavah. ‖ 38tasam. ya devata vyasa tatha strırupam aśritah. |apatyatam. samajagmus tenaśam. sanatam. gatah. ‖ 39†dan.d. am. dadau mahadevah. sapadmam. yan mahayudham† |

Kot. ıvars.a Mahatmya 267

dan. d. ahasta yatah. suta brahmı brahmıti cocyate ‖ 40suta śulakara caiva rudran. ıti tatas tu sa |kaumarı śaktina suta kaumarı sa tato ’bhavat ‖ 41

41b rudran. ı◦] conj., brahman. ı◦ R, ◦mindran. ı◦ A

vais.n. avıputritam. yata cakrin. ı kalinandana |vais.n. avı tena sa prokta ta vai rudrat tu matarah. ‖ 42†raks.asaś ca piśacaś ca gandharvah. pannagasurah. |varaharupa sajayat pancam. laks.an. am† ‖ 43atha tah. panca kaleya ucur devam. sureśvaram |atmanam. purus.am. dehi prasavemahi yena vai ‖ 44tato devah. svayam. tasam. sarvair devagan. air vr.tah. |pran. ato bahumanena abhut patir amitrajit ‖ 45mahakapalam. sam. gr.hya śirasam. malaya prabhuh. |karapadaiś ca vadanaih. kr.tam. sragdama ves.t.ayan ‖ 46

46b śirasam. ] em., śirasi R, śirasa A; 46cd] om. A; 46c karapadaiś ca] conj., karayane ca R;46d kr.tam. sragdama] conj., kr.tamvagnama R

anekacaran. o bhutva anekabhujadharakah. |anekamukhakayaś ca tasam asıt purah. sarah. ‖ 47

47c ◦mukhakayaś ca] conj., ◦śvaśakayaś ca R, ◦rupam apanna◦ A

tatah. sa tabhih. sam. gatya janayamasa suvratah. |†malah. supadmah. sutapah. † aindryah. kauberya eva ca ‖ 48

48c malah. supadmah. sutapah. ] R, lomah. padmaks.ata aindrı A

vayavyaś caiva yamyaś ca agneyaś ca mahabalah. |varun. yaś caiva guhyaś ca †tatha caiva gava ca sa† ‖ 49

49d gava ca sa] R, gavabravıt Avr.ddhapramata †tathaivakhyamr.tita mr.s.ika tatha† |camun. d. a caiva vikhyata bhaks.abhojyahara para ‖ 50

50–56] om. Raran. ımataraś caiva antariks.asya matarah. |etaś canyaś ca rudrat ta matarah. samprasuvire |kot.yo matr.gan. anam. tu tasam. s.as.t.im. tadasr.jat ‖ 51tato balena mahata matr.n. am. sa tadeśvarah. |ajagama purım. tam. vai vividhayudhadharin. a ‖ 52tato ’maragan. ah. sarve ˘ ˘ ˘ ˘ ˘ ˘ ˘ |˘ ˘ ˘ ˘ ˘ ˘ ˘ ˘ sarvam. s tan vipradudruvuh. ‖ 53

53bc] loss of two padas conjectured.tan apatata evatha daityan daivatakan. t.akan |chitva chitvatha khadanti śataśo ’tha sahasraśah. ‖ 54te bhaks.yaman. ah. kaleya puram. viviśur anv atha |tatrapi tah. suyuddhena bhaks.ayanti yathabalam ‖ 55ye gr.hes.u bhavalılas tatra tatrasureśvarah. |tes.am. striyah. samaviśya te ’pi tabhis tu bhaks.itah. ‖ 56 749

749 After this R and A have the text contained in Part II, but some verses are probablylost at the end of Part I.

268 Yuko Yokochi (ed.)

Part II (S & RA recensions): KoM 24–44

atha daityair hataih. sarvair devadeva umapatih. |ajagama pradeśam. tam. matr.n. am. varaditsaya ‖ 24

24a–34b] om. R; 24c–25b] om. A

tan dr.s.t.va nihatan sarvan daityan amaravidvis.ah. |paritus.t.as tada tasam. varan pradad vr.s.adhvajah. ‖ 25jagato mataro yuyam. matr.bhuta bhavis.yatha |yus.makam. ye bhavis.yanti bhaktah. purus.apum. gavah. ‖ 26

26c–29d] om. A

striyo vapi mahabhaga na tan him. santi him. sakah. |mr.ta mama gan. aś caiva bhavis.yanty ajaramarah. ‖ 27bhavantınam idam. sthanam. kot.ıvars.am iti śrutam |bhavis.yati jagatkhyatam. sarvapapapramocanam ‖ 28aham. hetur hi yus.makam. yasmat sr.s.t.a mayaiva ca |hetukeśvaranamaham. sthasyamy atra varapradah. |yus.mabhih. saha vatsyami nayakatve vyavasthitah. ‖ 29yas tu yus.man maya sardham. vidhivat pujayis.yati |sarvapapavimuktatma sa param. gatim apsyati ‖ 30danava nihata yasmac chulena bahumam. saya |śulakun.d. am idam. namna khyatam. tırtham. bhavis.yati ‖ 31

31b bahumam. saya] S2Bh, bahumayaya A; 31c śulakun. d. am] S2Bh, śuladan. d. am A

iha śulodakam. pıtva bahumam. sam. pran. amya ca |avadhyah. sarvahim. sran. am. bhavis.yati narottamah. ‖ 32

32b bahumam. sam. pran. amya ca] S2Bh, pran. amya bahulodbhavam. A

pratikuleti vikhyata ramya mandakinı nadı |rudhiraughavatı seha bhavatınam. bhavis.yati ‖ 33aham. brahma ca vis.n. uś ca r.s.ayaś ca tapodhanah. |matr.tantran. i divyani matr.yajnavidhim. param |pun. yani prakaris.yama yajanam. yair avapsyatha ‖ 34brahmam. svayambhuvam. caiva kaumaram. yamalam. tatha |sarasvatam. sagandharam aiśanam. nandiyamalam ‖ 35

35b yamalam. ] S2Bh, vayukan R, vamanam. A; 35d ◦yamalam] S2Bh, ◦vardhanam RA

tantran. y etani yus.makam. tathanyani sahasraśah. |bhavis.yanti nara yais tu yus.man yaks.yanti bhaktitah. ‖ 36

36d yaks.yanti] em. Bh, ya+ks.a+nti S2, raks.yasti R, paśyanti A

naran. am. yajamananam. varan yuyam. pradasyatha |divyasiddhiprada devyo divyayoga bhavis.yatha ‖ 37

37c–38d] om.A

yaś ca naryah. sada yus.man yaks.yante sarahasyatah. |yogeśvaryo bhavis.yanti rama divyaparakramah. ‖ 38

38b yaks.yante sarahasyatah. ] S2Bh, raks.yanti ca vida niśi R

chagalah. kumbhakarn. aś ca madıyau gan. anayakau |yus.makam. dvarapalatve sthasyatas tau mamajnaya ‖ 39

39a chagalah. ] S2Bh, kokilo RA; kumbha◦] S2Bh, ghot.a◦ R, ghat.a◦ A4A7, vat.a◦ A3kot.ıvars.am idam. sthanam. matr.n. am. priyam uttamam |śmaśanam. pravaram. divyam. bhavis.yati sukhapradam ‖ 40

40a kot.ı◦] S2A7Bh, kot.i◦ R, kot.ım. A4

Kot. ıvars.a Mahatmya 269

varan evam. hi tam. l labdhva mataro lokamatarah. |bhaktya pran. amya deveśam. mumudur bhr.śam arditah. ‖ 41tatah. prabhr.ti tah. sarvah. sahitah. śaśimaulina |kot.ıvars.am. vasanti sma sarvalokabhayapradah. ‖ 42

42c–colophon] om. RA

evam. sa bhagavan devo matr.n. am. gan. anayakah. |abhavac cham. karo vyasa yan mam. tvam. pr.s.t.avan api ‖ 43utpattim etam. jagatıśvarın. am. brahmendraśakranilavanditanam |pat.hen manus.yah. śr.n. uyac ca nityam. yah. svargalokam. sa mr.to labheta ‖ 44

Colophon of chapter 171colophon] skandapuran. e dhyaya | 171 (in letter numerals) ‖ S2pc, ... | skandapuran. e ... ‖ S2ac

‘

References

Figures

Figure 1 Genealogy of the Imperial Guptas (after Kumaragupta I). . . . . . . . . . 27

Figure 2 The pedigree of Harivarman given in the Shankarpur CopperplateInscription. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

Figure 3 The Guptas, the Early and Later Aulikaras, and their ministers/generalsin Mandasor. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

Figure 4 The Maukharis of the Gaya Region. . . . . . . . . . . . . . . . . . . . 43

Figure 5 The Maukharis of Kanauj. . . . . . . . . . . . . . . . . . . . . . . . . 49

Figure 6 Marital alliances of the royal houses of North India in the 6th and 7thcenturies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62

Figure 7 The Later Guptas of Magadha. . . . . . . . . . . . . . . . . . . . . . . 83

Figure 8 Matrimonial relationship of the Pus.yabhutis and Later Guptas. . . . . . . 92

Figure 9 The Maitrakas of Valabhı and Hars.avardhana of Kanauj. . . . . . . . . 105

Figure 10 Ban. a’s Pedigree. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159

Figure 11 Administrative machinery in Gaud. a. . . . . . . . . . . . . . . . . . . 243

Figure 12 kot.ivars. adhis. t.hanadhikaran. a. . . . . . . . . . . . . . . . . . . . . . . 244

Plates(All photos were taken by the author, unless stated otherwise)

Plate 0 Mahis.asuramardinı. From Nachna. Ramvan Museum (Satna District)No. 83. (35 x 33 cm). . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

Plate 1 Third panel on the left of the reverse side of Nagarı lintel.Photo courtesy Joanna Williams. . . . . . . . . . . . . . . . . . . . . . 37

Plate 2 Gwalior Stone Inscription of Mihirakula & Fleet’s rubbing. Photo courtesy ofthe British Museum & Ink impression published by Fleet in CII III (1888),retrieved from:http://commons.wikimedia.org/wiki/File:Mihirakula_inscription.jpg 38

Plate 3 Devı Mahis.asuramardinı, Kauva-Dol. Photo courtesy Michael Willis. . . . 44

Plate 4 Images near the Karn. a Caupar Cave (Barabar). . . . . . . . . . . . . . 46

Plate 5 The River Son near Arrah. . . . . . . . . . . . . . . . . . . . . . . . . 48

Plate 6 Territory of the Imperial Guptas around ad 530. Courtesy Google Earth. . 57

273

274 References

Plate 7 King Khusrau Anushirvan sits before the chess board while his vazir and theenvoy of Qannuj are playing chess.Courtesy St. Petersburg National Library of Russia (MS Dorn 329, 307r). . 64

Plate 8 Right half of a pillar at Deo Baranark with an inscription of Jıvitagupta II.Courtesy British Library: No. 1003458.Photo: Garrick, Henry Baily Wade, ad 1880. . . . . . . . . . . . . . . . 67

Plate 9 Emblem of a seal of Śarvavarman, modern impression.Courtesy British Museum: Asia 1896. 2–1. 106. . . . . . . . . . . . . . . 68

Plate 10 Image of Skanda (Karttikeya) found in Kanauj. Allahabad Municipal Museum.Photo courtesy Huntington Archive, Ohio. . . . . . . . . . . . . . . . . 75

Plate 11 The army camps of Hars.a at Man. itara and at Kajurgira.Courtesy Google Earth. . . . . . . . . . . . . . . . . . . . . . . . . . 100

Plate 12 The abhis.eka of Kumara/Skanda. Kanauj. Courtesy Wikimedia Commons. 164

Plate 13 The Mahakapala or Brahma Sarovar in Kuruks.etra. . . . . . . . . . . . 169

Plate 14 Bird’s-eye view of the geography of site VBA (Bhadreśvara), Gangadvara, andKanakhala (Daks.eśvara). Courtesy Google Earth. . . . . . . . . . . . . 177

Plate 15 Seal-die found in Hardwar (mirorred). Courtesy Victoria & Albert Museum,London (No. IM.496–1924). . . . . . . . . . . . . . . . . . . . . . . . 178

Plate 16 Confluence of Rambha and Ganga. Photo taken from site VBA-II. . . . . 179

Plate 17 Excavation site VBA-II, structure 1: temple with linga (1). . . . . . . . 180

Plate 18 Excavation site VBA-II, linga (2). . . . . . . . . . . . . . . . . . . . . 180

Plate 19 Saivite figure (Kus.an. a period).Courtesy Bharat Mandir Museum (Rishikesh). . . . . . . . . . . . . . . 181

Plate 20 Fragment of Vis.n. u image: cakra.Courtesy Bharat Mandir Museum (Rishikesh). . . . . . . . . . . . . . . 182

Plate 21 Bird’s-eye’s view of Rishikesh: Shyampur Garhi (Brahmavarta?), VBA(Bhadreśvara), and Bharat Mandir (Kubjamraka). Courtesy Google Earth. 184

Plate 22 Garhi Shyampur excavation site (Rishikesh). . . . . . . . . . . . . . . 185

Plate 23 Kalinjar Hill and surroundings. Courtesy Google Earth. . . . . . . . . . 189

Plate 24 Kalinjar Hill seen from the North. Courtesy Google Earth. . . . . . . . . 190

Plate 25 Two seals found at Bhıt.a. John Marshall in ASIAnnual Report 1911–12, 49, Plate XVIII. . . . . . . . . . . . . . . . . 191

Plate 26 5th-century Ekamukhalinga found on Kalanjara Hill. Kalanjara Hill sitemuseum (Raja Aman Singh Palace) Doc. No. 545.Photo courtesy D. Stadtner. . . . . . . . . . . . . . . . . . . . . . . 192

Plate 27 Plan of Kalanjara Hill (Cunningham). ASI Reports 1883–84Vol. XXI Plate VIII. . . . . . . . . . . . . . . . . . . . . . . . . . . 194

Plate 28 Kalanjara, Mr.gadhara promontary with graffiti and carrier. . . . . . . . 195

Plate 29 Kalanjara, Mr.gadhara graffito. . . . . . . . . . . . . . . . . . . . . . 196

Plate 30 Kalanjara, Sıtasej, ascetic cave at the top of the northern slope. . . . . . 197

Plate 31 Kalanjara, Mr.gadhara. Image of the seven deer. . . . . . . . . . . . . . 198

Plate 32 Kalanjara, Lakulıśa with disciples. Kalanjara Hill, site museum (Raja AmanSingh Palace) No. 1086. Cf. Yadav 2009–10 Pl. 13. Courtesy of the U.P. StateArchaeology Department (Lucknow). Photo courtesy Rajendra Yadav. . . 199

References 275

Plate 33 Kalanjara, Inscription on pilaster in the Chandi Darwaza.Photo courtesy D. Stadtner. . . . . . . . . . . . . . . . . . . . . . . 201

Plate 34 Kalanjara, Pillar KF 251 (late Gupta period). Photo courtesy D. Stadtner. 202

Plate 35 Kalanjara, Balkandeswara linga and carrier with inscription of Vasanta. . 206

Plate 36 Kalanjara, Balkandeswara. Inscription of Vasanta. . . . . . . . . . . . . 206

Plate 37 Kalanjara, Carrier image near the Gangasagara tank with inscription. . . 207

Plate 38 Kalanjara, Rock facade at Khambor Shiva. Photo P.C. Bisschop 2012. . . 208

Plate 39 Kalanjara, Nılakan. t.ha, Lakulıśa. . . . . . . . . . . . . . . . . . . . . 216

Plate 40 Kalanjara, Lingas from a group of twenty-six in cave 6, behind the sanctum ofthe Nılakan. t.ha Temple. Cf. Yadav 2009–10 Pl. 9.Photo courtesy Rajendra Yadav. . . . . . . . . . . . . . . . . . . . . 217

Plate 41 Mun. d. eśvarı Hill and surroundings. Courtesy Google Earth. . . . . . . . 223

Plate 42 Buchanan: sketch ad 1813. From Martin 1838, Vol. I, opposite p. 474(Pl. V No. 1, ‘Temple of Mundeswari.’). . . . . . . . . . . . . . . . . . 224

Plate 43 Daniell: drawing ad 1790: ‘The Temple of Mandeswara near Chaynpore,Bahar.’ Courtesy British Library. Cf. Archer 1980, Plate 76. . . . . . . . 224

Plate 44 Daniell: drawing ad 1790: ‘Interior of the temple of Mandeswara nearChaynpore, Bahar.’ Courtesy British Library. Cf. Archer 1980, Plate 77. . 225

Plate 45 Man. d. aleśvara linga in the octagonal temple. Photo Yuko Yokochi 2008. . 225

Plate 46 Mun. d. eśvarı Hill seen from the South. Courtesy Google Earth. . . . . . . 229

Plate 47 Mun. d. eśvarı architectural fragments: doorframe with dvarapala & ornamentedpilaster. Photos F. Buckee 2008. . . . . . . . . . . . . . . . . . . . . 230

Plate 48 Mun. d. eśvarı architectural decorations: Skanda with cock? & figure withkhat.vanga. Photos F. Buckee 2008. . . . . . . . . . . . . . . . . . . . 230

Plate 49 Mun. d. eśvarı: old caturmukhalinga & detail. . . . . . . . . . . . . . . . 231

Plate 50 Mun. d. eśvarı: Ekamukhalinga of site A. Photo F. Buckee 2008. . . . . . . 232

Plate 51 Bhairava near Mun. d. eśvarı Hill site A. . . . . . . . . . . . . . . . . . . 233

Plate 52 Inscription below the Bhairava image (site A). . . . . . . . . . . . . . . 233

Plate 53 The Skandapuran. a research team at Mun. d. eśvarı Hill site B.Photo P.C. Bisschop 2012. . . . . . . . . . . . . . . . . . . . . . . . 234

Plate 54 Graffito at Mun. d. eśvarı Hill site B. Photo P.C. Bisschop 2012. . . . . . . 234

Plate 55 Mun. d. eśvarı Hill: Vis.n. u Image. . . . . . . . . . . . . . . . . . . . . . 235

Plate 56 Vis.n. u image near Mun. d. eśvarı Hill site A. Photo Yuko Yokochi 2012. . . . 236

Plate 57 Mun. d. eśvarı: Lakulıśa on lintel. Patna Museum, Acc. No 54094.Photo courtesy American Institute of Indian Studies. . . . . . . . . . . 237

Plate 58 Mun. d. eśvarı: Skandamatr.. Patna Museum, Acc. No 54048.Photo courtesy American Institute of Indian Studies. . . . . . . . . . . 238

Plate 59 Mun. d. eśvarı: Skanda. Courtesy Patna Museum. . . . . . . . . . . . . . 238

Plate 60 Mun. d. eśvarı: Tilottama? Patna Museum, Acc. No 54047.Photo courtesy American Institute of Indian Studies. . . . . . . . . . . 239

Plate 61 Rajabar.ı Mound (Devıkot.a Citadel) within the fortified area of ancientBangarh/Kot.ivars.a on the Punarbhava River. Courtesy Google Earth. . . 248

276 References

Plate 62 Excavations at the Rajabar.ı Mound. Photo Yuko Yokochi 2006. . . . . . 249

Plate 63 Varahı image found near the Shib Bari site (Bangarh).Photo Yuko Yokochi 2006. . . . . . . . . . . . . . . . . . . . . . . . 250

Plate 64 Camun.d. a type of image (Bahumam. sa/Carcika?) found near the Shib Bari site(Bangarh). Pala period. Photo Yuko Yokochi 2006. . . . . . . . . . . . 255

Plate 65 Folio 322v of A7. Dhaka University Library MS 3376. Paper, Bengali script. 261

Bibliography

Acharya, Diwakar2007 The Sam. skaravidhi: A Manual on the Transformatory Rite of the

Lakulıśa–Paśupatas. in: Goodall, Dominic & Andre Padoux (eds.),Melanges tantriques a la memoire d’Helene Brunner. Pondichery/Paris.pp. 27–48.

2011 Paśupatas. in: Brill’s Encyclopedia of Hinduism, Vol. III, 458–66. Leiden.

Adaval, Niti1970 The Story of King Udayana as gleaned from Sanskrit, Pali & Prakrit

Sources. Varanasi. The Chowkhamba Sanskrit Studies LXXIV.Agrawala, Prithvi K.

1963 Solar Symbolism of the Boar (Yajna-Varaha)—an Interpretation.Varanasi.

1967 Skanda-Karttikeya. A Study in the Origin and Development. BanarasHindu University, Varanasi 1967. Monographs of the Department ofAncient Indian History, Culture and Archaeology No. 3.

1987 Observations on the Mundeshvari Temple. in: Rao, Nagaraja (ed.),Kusumanjali. New Interpretation of Indian Art & Culture. Sh.C.Sivaramamurti Commemoration Volume. Vol. I. Delhi. pp. 179–83.

Agrawala, Vasudeva S.1947–48 Terracotta Figurines of Ahichchhatra, District Bareilly, U.P. in: Ancient

India 4 (1947–48), 104–179.

1969 The Deeds of Harsha [being a cultural study of Ban. a’s Harshacharita].Redacted and edited by P.K. Agrawala. Varanasi.

Ahuja, Naman2005 Changing Gods, Enduring Rituals: Observations on Early Indian

Religion as seen through Terracotta Imagery c. 200 BC–ad 100. in:South Asian Archaeology 2001, Paris. Vol. II, 345–354.

Ali, Daud2004 Courtly Culture and Political Life in Early Medieval India. Cambridge.

Allan, John1914 Catalogue of the Coins of the Gupta Dynasties and of Śaśanka, King of

Gaud. a. London.Amar, Abhishek S.

2012 Buddhist Responses to Brahman. a Challenges in Medieval India:Bodhgaya and Gaya. in: Journal of the Royal Asiatic Society 22.1(2012), 155185.

Amarasim. ha(AK) Amarakośa with the unpublished South Indian Commentaries

Amarapadavivr.ti of Lingayasurin and the Amarapadaparijata ofMallinatha. Critically edited with Introduction by A. A. Ramanathan.Vol. 1. Madras 1971. The Adyar Library Series 101.

277

278 References

Anderson, Leona1988 The Indian Spring festival (Vasantotsava): One or many? in: ABORI 69,

63–76.Archer, Mildred

1969 British Drawings in the India Office Library. London. 2 vols.

1980 Early Views of India. The Picturesque Journeys of Thomas and WilliamDaniell 1786–1794. The Complete Aquatints. London.

ArrianDer Alexanderzug (>Alex�ndrou >An�basic) und Indische Geschichte(>Indik�). Griechisch und deutsch. Herausgegeben und ubersetzt vonGerhard Wirth und Oskar von Hinuber. Munchen–Zurich 1985.Sammlung Tusculum.

Arthaśastra(AS) The Kaut.ilıya Arthaśastra, edit. and transl., with a study, by R.P.

Kangle. Bombay 1960–65. 3 vols. University of Bombay Studies,Sanskrit, Prakrit and Pali, No. 1–3.

Aryamanjuśrımulakalpa(MMK) E-text by members of the Sanskrit Buddhist Input Project. Based on

the edition by M.M.T. Ganapati Shastri, 1925. 3 vols. (reprint: Delhi :Sri Satguru Publications 1989, pp. 1-722) and on theAryamanjuśrımulakalpa in: Mahayanasutrasam. graha, part II, ed. byP.L. Vaidya, Darbhanga 1964 (Buddhist Sanskrit Texts, 18).http://www.sub.uni-goettingen.de/ebene\_1/fiindolo/gretil.htm

See Jayaswal 1934.

Asher, Frederick M.1985 Eastern Indian Inscriptions: The problem of Era. in: Asher, F.M. & G.S.

Gai (eds.), Indian Epigraphy. Its bearing on the History of Art. NewDelhi 1985, 133–37.

Atharvavedapariśis.t.aThe Pariśis.t.as of the Atharvaveda. Edit. by George Melville Bolling andJulius von Negelein. With Hindi notes by Ram Kumar Rai. Varanasi1976. Caukhamba Pracyavidya Granthamala 1.

Bakker, Hans T.1986 Ayodhya. Pt. I The History of Ayodhya from the 7th century BC to the

middle of the 18th century. Its development into a sacred centre withspecial reference to the Ayodhyamahatmya and the worship of Ramaaccording to the Agastyasam. hita. Pt. II Ayodhyamahatmya.Introduction, Edition, and Annotation. Pt. III Appendices,Concordances, Bibliography, Indexes, and Maps. Groningen. GroningenOriental Studies I.

1991 The Footprints of the Lord. in: Diana L. Eck and Francoise Mallison(eds.), Devotion Divine. Bhakti Traditions from the Regions of India.Studies in Honour of Charlotte Vaudeville. Groningen Oriental StudiesVIII. Groningen / Paris, pp. 19–38.

1992 Memorials, Temples, Gods, and Kings. An attempt to unravel thesymbolic texture of Vakat.aka kingship. in: A.W. van den Hoek, D.H.A.Kolff, M.S. Oort (eds.), Ritual, State and History in South Asia. Essaysin Honour of J.C. Heesterman. Leiden. pp. 7–19.

1994 Observations on the History and Culture of Daks.in. a Kosala (5th to 7thcenturies ad). in: Nalini Balbir & Joachim K. Bautze (eds.), Festschrift

References 279

Klaus Bruhn, zur Vollendung des 65. Lebensjahres dargebracht vonSchulern, Freunden und Kollegen. Reinbek 1994 (1995), pp. 1-66.

1996 Parvatı’s Svayam. vara. Studies in the Skandapuran. a I. in: WienerZeitschrift fur die Kunde Sudasiens 40 (1996), 5–43.

1997 The Vakat.akas. An Essay in Hindu Iconology. Groningen. GondaIndological Studies 5.

2000 Somaśarman, Somavam. a and Somasiddhanta. A Paśupata tradition inseventh-century Daks.in. a Kosala. Studies in the Skandapuran. a III. in:Ryutaro Tsuchida and Albrecht Wezler (eds.), Haranandalaharı. Volumein Honour of Professor Minoru Hara on his Seventieth Birthday.Reinbek, pp. 1–19.

2001 Sources for Reconstructing Ancient Forms of Śiva Worship. in: FrancoisGrimal (ed.), Les sources et le temps/Sources and Time. A colloquium,Pondicherry, 11–13 January 1997, Pondichery 2001. pp. 397–412.

2004a Mansar. in: Bakker, Hans T. (ed.), The Vakat.aka Heritage. IndianCulture at the Crossroads. Groningen. pp. 71–85.

2004b Origin and Growth of the Puran. ic Text Corpus with Special Referenceto the Skandapuran. a. Papers of the 12th World Sanskrit Conference Vol.3.2, edited by Hans T. Bakker. Delhi.

2004c At the Right Side of the Teacher: Imagination, Imagery, and Image inVedic and Śaiva Initiation. in: Granoff, Phyllis & Koichi Shinohara(eds.), Images in Asian Religions. Toronto 2004. pp. 117–148.

2005 Commemorating the Dead. A note on Skandagupta’s Bhitarı Inscription,vss. 8–12. in: Sharma, R.K. & Devendra Handa (eds.), Revealing thePast: recent trends in art and archaeology. Prof. Ajay Mitra ShastriCommemoration Volume. Aryan Books, New Delhi. pp. 248–51.

2006 A Theatre of Broken Dreams. Vidiśa in the Days of Gupta Hegemony.in: Brandtner, Martin & Shishir Kumar Panda (eds.), InterrogatingHistory. Essays for Hermann Kulke, Chapter 9. New Delhi. pp. 165–187.

2007 Thanesar, the Paśupata Order and the Skandapuran. a. Studies in theSkandapuran. a IX. in: Journal of Indological Studies 19 (2007), 1–16.

2007a Monuments to the Dead in Ancient North India. in: Indo-IranianJournal 50.1 (2007), 11–47.

2009a Purus.amedha, Manasarapurus.a, Vastupurus.a. The image of Man in theSacrificial Context. in: Journal of Indological Studies 20 & 21 (2008–09),1–23.

2009b The So-called ‘Jaunpur Stone Inscription of Iśvaravarman.’ in:Indo-Iranian Journal 52.2–3 (2009), 207–216.

2010 Rohitagiri. in: Caracchi, Pinuccia et al. (eds.), Tırthayatra. Essays inHonour of Stefano Piano. Edizioni dell’Orso, Alessandria. pp. 15–26.

2011 Origin and Spread of the Paśupata Movement. About Heracles, Lakulıśaand Symbols of Masculinity. in: Tikkanen, Bertil & Albion M. Butters(eds.), Purvaparaprajnabhinandanam, East and West, Past and Present:Indological and Other Essays in Honour of Klaus Karttunen. Helsinki.Studia Orientalia 110. pp. 21–37.

2013 The Temple of Man. d. aleśvarasvamin. The Mun. d. eśvarı Inscription of theTime of Udayasena Reconsidered. in: Bosma, N. & Nina Mirnig (eds.)2013, 263–277.

280 References

Bakker, Hans & Alan Entwistle1983 Devı. The worship of the Goddess and its contribution to Indian

pilgrimage. Groningen.Bakker, Hans & Harunaga Isaacson

1993 The Ramtek Inscriptions II. The Vakat.aka Inscription in theKevala-Narasim. ha Temple. in: Bulletin of the School of Oriental andAfrican Studies LVI (1993), 46–74.

Ban. a1909 Śrıhars.acaritamahakavyam: Ban. abhat.t.a’s biography of King

Harshavardhana of Sthan. vıśvara with Sankara’s commentary Sanketa,edited with critical notes by A. A. Fuhrer. Government Central Press,Bombay 1909. Bombay Sanskrit and Prakrit Series 66.

(HC) Hars.acaritam, śrı-Śankaraviracita ‘Sanketa’ vyakhyopetam.Hindıvyakhyakarah. śrı Jagannatha Pat.hakah. . CaukhambaVidyabhavan, Varan. ası 1958. Vidyabhavana Sam. skr.ta Granthamala 36.[Figures between [ ] refer to Kane’s edition, below].

The Hars.acarita of Ban. abhat.t.a (Text of Ucchvasa I–VIII). Edit. with anIntroduction and Notes by P.V. Kane. Delhi etc. 1965 (2nd ed.). [Pagereferences between [ ]].

1897 The Hars.a-carita of Ban. a translated by E.B. Cowell and F.W. Thomas.London. Oriental Translation Fund New Series II.

Kadambarı [compl. by Bhus.an. abhat.t.a]. Edit. by Peter Peterson.Bombay 1883. Bombay Sanskrit Series 24.

Banerji, R.D.1907–08 Mundesvari Inscription of Udayasena. The [Harsha] Year 30. in: EI IX

(1907–08), 289f.Basak, Radhagovinda

1919–20 The Five Damodarpur Copper-Plate Inscriptions of the Gupta Period.in: EI XV (1919–20), 113–45.

Bautze, Joachim Karl1995 Early Indian Terracottas. Leiden. Iconography of Religions, Section

XIII: Indian Religions, Fascicle 17.Beal

See Xuanzang Si-yu-ki.Bhagavatapuran. a

(BhagP) Śrıbhagavatam (Śrımad Bhagavata Mahapuran. a). Critically edited forthe first time by H.G. Shastri et al. Vol. I–IV. B.J. Institute of Learningand Research, Ahmedabad 1996–98. 4 vols.

Bharavi(KA) The Kiratarjunıya of Bharavi with the commentary (Ghan. t. apatha) of

Mallinatha and various readings. Ed. Pan.d. it Durgaprasad and KaśınathPan.d. urang Parab, revised by Wasudev Laxman. Śastrı Pan. śıkar.Bombay 1933.

BhasarvajnaThe Gan. akarika of Acarya Bhasarvajna. Edited by Chimanlal D. Dalal,[along with the Gan. akarikaratnat. ıka] and four appendices including theKaravan. -Mahatmya. Baroda 1920 (reprint, 1966).

Bhattacharyya, Pranab Kumar1977 Historical Geography of Madhya Pradesh from Early Records. Delhi.

References 281

Bisschop, Peter C.2004 Śiva’s Ayatanas in the Various Recensions of Skandapuran. a 167. in:

Bakker 2004b, pp. 65–78.

2006 Early Śaivism and the Skandapuran. a. Sects and Centres. Groningen.Groningen Oriental Studies XXI.

2007 The description of Śivapura in the early Vayu- and Skandapuran. a. in:Goodall, Dominic & Andre Padoux (eds.), Melanges Tantriques a lamemoire d’Helene Brunner. Pondichery. Collection Indologie 106. pp.49–71.

2013 Two Pre-Chandella Inscriptions from Kalanjara. in: Bosma, N. & NinaMirnig (eds.) 2013, 279–294.

Bisschop, Peter & Arlo Griffiths2003 The Paśupata Observance (Atharvavedapariśis.t.a 40). in: IIJ 46 (2003),

315–348.Bloch, T.

1902–03 The Temple of Mun. d. eśvarı. in: ASI Annual Report 1902–03, 42f.Bosma, Natasja

2013 The Baleśvara Temple Complex of Śivagupta. Epigraphical Evidence forthe Śaiva Siddhanta and Soma Siddhanta Traditions in Daks.in. a Kosala.in: Bosma, N. & Nina Mirnig (eds.) 2013, 245–262.

Bosma, Natasja & Nina Mirnig (eds.)2013 Epigraphical Evidence for the Formation and Rise of Early Śaivism.

[= Indo-Iranian Journal 56.3–4 (2013) (special issue), 201–440] Brill,Leiden.

Brahmapuran. a(BrP) Vol. 1, Sanskrit Indices and text of the Brahmapuran. a by Peter

Schreiner and Renate Sohnen. Vol. 2, Brahmapuran. a. Summary ofContents, with Index of Names and Motifs by Renate Sohnen and PeterSchreiner. Wiesbaden. 2 vols. Puran. a Research Publications, Tubingen 1and 2.

Brancaccio, Pia2011 The Buddhist Caves at Aurangabad: Transformations in Art and

Religion. Leiden/Boston. Brill’s Indological Library 34.Bruckner, Heidrun

1999–2000 Manuscripts and performance traditions of the so-called“Trivandrum-Plays” ascribed to Bhasa—A report on work in progress.in: Bulletin d’etudes indiennes 17–18 (1999–2000), 501–50.

Buddhaghos.aSumangala-Vilasinı, Buddhaghosa’s commentary on the Dıgha Nikaya.Edit. by T.W. Rhys Davids and J. Estlin Carpenter. Pali text Society,London 1968 (second edition).

Buhler, G. & Bhagvanlal Indraji1880 Inscriptions of Nepal. in: Indian Antiquary 9 (1880),163–94.

Burn, R.1906 Some Coins of the Maukharis, and of the Thanesar Line. in: JRAS

(1906), 843–50.Cac-nama

The Chachnama. An ancient history of Sind, giving the Hindu perioddown to the Arab conquest. Translated from the Persian by MirzaKalichbeg Fredunbeg. The Commissioner’s Press, Barachi 1900.http://persian.packhum.org/persian.

282 References

Casile, Anne2009 Temple et expansion d’une centre religieux en Inde centrale. Lectures du

paysage archeologique de Badoh-Pat.hari du 5e au 10e siecle de notre ere.Paris (These de Doctorat). 3 vols.

CatalogueCatalogue des manuscripts et xylographes orientaux de la Bibliothequeimperiale publique de St. Petersbourg [dresse par Bernhard Dorn etReinhold Rast]. St. Petersbourg [1852]. [Reprint, Leipzig 1978].

Chakravarti, Ranabir2008 Three Copper Plates of the Sixth Century ad. Glimpses of

Socio-Economic and Cultural Life in Western India. in: Raven, Ellen M.(ed.), South Asian Archaeology 1999. Groningen. pp. 395–399.

Chandra, Pramod1970 Stone Sculpture in the Allahabad Museum. A descriptive catalogue.

Poona.Cœdes, George

1937–66 Inscriptions du Cambodge. 8 vols. Paris: EFEO 1937 (vol. 1), 1942 (vol.2), 1951 (vol. 3), 1952 (vol. 4), 1953 (vol. 5), 1954 (vol. 6), 1964 (vol. 7),1966 (vol. 8). [K = Khmer inscription, numbered as in Cœdes 1937–66].

Corpus Inscriptionum Indicarum(CII II A) Kharosht.hı Inscriptions with the exception of those of Aśoka edited by

Sten Konow. Reprint, Delhi 1991. Corpus Inscriptionum Indicarum Vol.II, Part I.

( CII II B) Barhut Inscriptions edit. by H. Luders. Revised by E. Waldschmidt andM.A. Mehendale. Reprint, Delhi 1998. Corpus Inscriptionum IndicarumVol. II, Part II.

(CII III (1888)) Inscriptions of the Early Gupta Kings and their Successors. Edit. byJ.F. Fleet, Calcutta 1888.

(CII III (1981)) Inscriptions of the Early Gupta Kings. Revised by DevadattaRamakrishna Bhandarkar. Edit. by Bahadurchand Chhabra & GovindSwamirao Gai. New Delhi 1981.

(CII IV) Inscriptions of the Kalacuri-Chedi Era. Edited by V.V. Mirashi.Ootacamund 1955. 2 vols. Archaeological Survey of India.

(CII V) Inscriptions of the Vakat.akas. Edited by V.V. Mirashi. Ootacamund1963. Archaeological Survey of India.

(CII VII) Inscriptions of the Paramaras, Chandellas, Kachchhapaghatas and twoMinor Dynasties. Edit. by Harihar Vitthal Trivedi. New Delhi 1991,1989. 3 vols.

Cunningham, A.1885 Reports of a Tour in Bundelkhand and Rewa in 1883–84 and of a Tour

in Rewa, Bundelkhand, Malwa, and Gwalior, in 1884–85. ArchaeologicalSurvey of India XXI, Pts. I and II. Calcutta.

Dalal, Chimanlal D.1966 Gan. akarika of Acarya Bhasarvajna (with four appendices including the

Karavan. a-Mahatmya). Baroda. Gaekwad’s Oriental Series No. 15.Dan. d. in

Avantisundarı of Acarya Dan.d. in. Published by Śuranad. Kunjan Pillai.Trivandrum 1954. Trivandrum Sanskrit Series 172.

Dani, Ahmad Hasan1986 Indian Palaeography. New Delhi.

References 283

Davis, Richard H.1994 The Rebuilding of a Hindu Temple. in: Lopez Jr., Donald S. (ed.),

Religions of India in Practice, Princeton 1994. pp. 627–36.Deva, Krishna

1985 Mun.d. eśvarı Temple, Ramgarh. in: Asher, F.M. & G.S. Gai (eds.), IndianEpigraphy. Its bearing on the History of Art. New Delhi 1985, 125-27.

1988 Maukharis and Pus.pabhutis of Kanyakubja. in: Meister, Michael W. etal. (eds.), Encyclopaedia of Indian Temple Architecture (EITA). NorthIndia. Foundations of North Indian Style c. 250 B.C.—A.D. 1100. Vol.II.1. Delhi 1988.

1990 Mun.d. eśvarı Temple. in: Sam. skr.ti Sandhana, Journal of the NationalResearch Institute of Human Culture, Vol. 3, 156–64.

Devahuti, D.1998 Harsha. A Political Study. Third revised edition, Delhi etc.

Dey, Nundo Lal1971 The Geographical Dictionary of Ancient and Mediæval India. Delhi.

(3rd edition)Dundas, Paul

2006 A Non-Imperial Religion? Jainism in Its “Dark Age”. in: Olivelle,Patrick (ed.), Between the Empires, Oxford. pp. 383–414.

Edgerton, Franklin(BHSD) Buddhist Hybrid Sanskrit Grammar and Dictionary. Volume II:

Dictionary. Indian reprint, Delhi 1970.Elliot, H.M. & Dowson

1867–77 The History of India as told by its own Historians. The MuhammadanPeriod, edited from the posthumous papers of the late H.M. Elliot byJohn Dowson. London. 8 vols.

Encyclopedia of Indian Temple Architecture(EITA) North India. Foundations of North Indian Style (c. 250 B.C. – A.D.

1100. Edited by M.W. Meister & M.A. Dhaky & Krishna Deva. Vol.II.1–2. Princeton/New Delhi 1988.

Epigraphia Indica(EI) A collection of inscriptions supplementary to the Corpus Inscriptionum

Indicarum of the Archaeological Survey, translated by several orientalscholars; ed. by Jas. Burgess [et al.], Calcutta/Bombay 1892–1969. Vol.1–37.

Ettinghausen, Maurice L.1906 Hars.a Vardhana, empereur et poete de l’Inde septentrionale (606–648

A.D.). Etude sur sa vie et son temps. Londres, Paris, Louvain.Falk, Harry

1986 Bruderschaft und Wurfelspiel. Untersuchungen zurEntwicklungsgeschichte des vedischen Opfers. Freiburg.

2006 Aśokan Sites and Artefacts. A source-book with bibliography. Mainz.Monographien zur indischen Archaologie Kunst und Philologie 18.

Ferrier, Cedric & Judit Torzsok2008 Meditating on the king’s feet? Some remarks on the expression

padanudhyata. in: IIJ 51 (2008), 93–113.Firdawsı

The Epic of the Kings. Shah-nama. The national epic of Persia byFerdowsi, translated by Reuben Levy. Chicago 1967. Persian HeritageSeries.

284 References

Fleet, J.F.1882 The Dynasties of the Kanarese Districts of the Bombay Presidency.

Bombay Gazetteer, Bombay.Furui, Ryosuke

2011 Panchrol (Egra) Copperplate Inscription of the Time of Śaśanka. ARe-edition. in: Pratna Samiksha: A Journal of Archaeology, New Series 2(2011), 119–130.

forthcoming Ajıvikas, Man. ibhadra and Early History of Eastern Bengal: A NewCopper Plate Inscription of Vainyagupta and its Implications.

Gadre, A.S.1935 The Virdi Copper-Plates of Sam. . 297: the first known grant of

Kharagraha I. in: Proceedings and Transactions of the Seventh All-IndiaOriental Conference, Baroda, December 1933. Oriental Institute, Baroda1935. pp. 659–76.

Geer, Alexandra Anna Enrica van der1998 The Bhasa Problem. A statistical research into its solution. Thesis,

Leiden.Geslani, Marko

2012 Śanti Rites in the Development of the Puran. ic Rajyabhis.eka. in:Indo-Iranian Journal 55 (2012), 321–77.

Ghosh, A.(EIA) An Encyclopaedia of Indian archaeology. Vol. I: Subjects. Vol. II: A

Gazetteer of Explored and Excavated Sites in India. Delhi 1989.Gill, Sandrine

1999 Mahasthangarh: A Riverine Port in Ancient Bengal. in Ray 1999, pp.154–72.

Gnoli, Raniero1956 Nepalese inscriptions in Gupta characters. [Edit. by] Raniero Gnoli, Part

I: Text, Part II: plates. Roma. 2 vols. Serie Orientale Roma 10.Gonda, J

1967 The Indra festival according to the Atharvavedins. in: JAOS 87 (1967),413–429.

1969 Aspects of Early Vis.n. uism. 2nd ed., Delhi.

Goodall, Dominic2004 The Parakhyatantra. A scripture of the Śaiva Siddhanta. A critical

edition and annotated translation. Pondichery. Collection Indologie 98.Goodall, Dominic and Harunaga Isaacson.

2007 Workshop on the Niśvasatattvasam. hita: The Earliest Surviving ŚaivaTantra? in: Newsletter of the NGMCP No. 3, 4–6.

Gopal, Lallanji1963 Samanta—its vaying significance in ancient India. in: JRAS 95 (1963),

21–37.Goswami, Kunja Gobinda

1948 Excavations at Bangarh (1938–41). University of Calcutta. AsutoshMuseum Memoir No. 1.

Goyal, S. R.1967 A History of the Imperial Guptas. Allahabad.

2005 Ancient Indian Inscriptions: recent Finds and New Interpretations.Jodhpur.

References 285

Goyal, Shankar1992 History and Historiography of the Age of Harsha. With a foreword by

K.D. Bajpai. Jodhpur.

2007 The Recently Discovered Kurukshetra-Varanasi Grant of Hars.a: Year 23.in: East and West 57 (2007), 193–203.

Granoff, Phyllis2003 Mahakala’s Journey: from Gan. a to God. in: Rivista degli Studi Orientali

LXXVII (2003), 95–114.

2004 Saving the Saviour: Śiva and the Vais.n. ava Avataras in the EarlySkandapuran. a. in: Bakker, Hans T. (ed.), Origin and Growth of thePuran. ic Text Corpus. Delhi. 2004. pp. 111–138.

2006 Śiva and his Gan. as: Techniques of Narrative Distancing in Puran. icStories. in: Panda, Raghunatha & Madhusudan Mishra (eds.), Voice ofthe Orient. A tribute to Prof. Upendranath Dhal. Delhi. pp. 77–102.

Gupta, Parmeshwari Lal1974–79 The Imperial Guptas. Vol. I: Sources, historiography & political history

(1974). Vol. II: Cultural history (1979). Varanasi.Gupta, Parmeshwari Lal & Sarojini Srivastava

1981 Gupta gold coins in Bharat Kala Bhavan. Bharat Kala Bhavan,Varanasi.

Gupte, Y.R.1915–16 Two Talesvara Copperplates. in: EI XIII (1915–16) No. 7, 109–21.

Harimoto, Kengo2006 The original Skandapuran. a, Laks.mıdhara and Can. d. eśvara. Studies in

the Skandapuran. a VIII. in: Indo-Iranian Journal 49 (2006), 23–38.Harivam. śa

(HV) The Harivam. śa being the Khila or supplement to the Mahabharata. Forthe first time critically edited by P. L. Vaidya. Poona 1969–71. 2 vols.

Harper, Katherine Anne1989 Seven Hindu Goddesses of Spiritual Tradition. The Iconography of the

Saptamatrikas. Lewiston/Queenston/Lampeter. Studies in Women andReligion, Vol. 28.

Hars.adevaRatnavalı [edit. by C. Cappeller] in: Otto Bohtlingk,Sanskrit-Chrestomathie, 3. verbesserte und vermehrte Auflage [von R.Garbe], Leipzig 1909. pp. 326–369.

The Ratnavalı of Śri Harsha-Deva, edit. with an introduction, a Sanskritcommentary, a literal English translation, critical notes and appendicesby M.R. Kale. Bombay 1964 (3rd ed.).

The Nagananda by Hars.adeva edit. by Madhava Candra Ghos.a assistedby Kr.s.n. a Kamala Bhat.t.acarya (Calcutta 1864), with a generalintroduction by Michael Hahn and a preface and a bibliography of theeditions and translations of the Nagananda by Roland Steiner. NewDelhi 1991. The recensions of the Nagananda by Hars.adeva. Vol. 1: TheNorth-Indian recension.

Priyadarśika of Śrı Hars.adeva. Edit. by M.R. Kale. Delhi 1928 (reprintDelhi 1977).

Hatley, Shaman2007 The Brahmayamalatantra and Early Śaiva Cult of Yoginıs. University of

Pennsylvania (Dissertation).

286 References

Heesterman, Johannes Cornelis1957 The Ancient Indian Royal Consecration. The rajasuya described

according to the Yajus texts and annotated. ’s Gravenhage. (Thesis)Hemacandra

Abhidhanacintaman. i of Śrı Hemachandracharya. Edited by NemichandraŚastrı and Śrı Haragovinda Śastrı. Varanasi 1996 (2nd ed.).

Anekarthasam. graho nama kośah. . Edited by Jagannatha Śastrı Hośinga,with an Alphabetical Index prepared by Ghanananda Pan. d. eya &Janardana Joshi. Benares 1929. The Kashi Sanskrit Series No. 68(Lexicography Section, No. 2).

Hinuber, Oskar von2002 Book review of H. Bechert et al., der Buddhismus I. in Indo-Iranian

Journal 45 (2002), 77–86.Hitti, Philip K.

1970 History of the Arabs, from the earliest times to the present. TenthEdition. London etc.

Huili(Life) The Life of Hiuen-Tsiang by the shaman Hwui Li, with an introduction

containing an account of the works of I-Tsing by Samuel Beal. With apreface by L. Cranmer-Byng. London 1884. Popular re-issue 1914(Indian edition, Delhi 1973 (enlarged)).

Huntington, Susan L1984 The “Pala-sena” Schools of Sculpture. Leiden. Studies in South Asian

Culture 10.Indian Antiquary

(IA) Journal of Oriental Research in Archaeology, History, Literature,Languages, Folklore, etc., edited by Jas. Burgess [et al.], Bombay1872–1933, Vol. I – 62.

Indian Archaeology – A Review(IAR) Archaeological Survey of India, New Delhi.

Indraji, Bhagvanlal1881-82 Antiquarian Remains at Sopara and Padan. a. in: Journal of the Bombay

Branch of the Royal Asiatic Society 15 (1881–82). [Published in 1883]Jain, Balchandra

1977 Shankarpur Plate of Budhagupta and Harivarman: Gupta Year 166. inJournal of the Epigraphical Society of India (JESI) IV (1977), 62–66.[Comments on this edition in footnotes by the editor of the journal,K.V. Ramesh].

Jain, Kailash Chand1972 Malwa Through the Ages (from the Earliest Time to 1305 A.D.). Delhi.

JatakasThe Jataka together with its commentary: being tales of the anteriorbirths of Gotama Buddha, for the first time edited in the original Paliby V. Fausbøll. Index to the Jataka and its commentary by DinesAndersen. London 1877–97. 7 vols.

Jayaswal, K.P.1934 An Imperial History of India in a Sanskrit Text [c. 700 B.C. – c. 770

A.D.]. With a Special Commentary on Later Gupta Period by K. P.Jayaswal. With the Sanskrit Text Revised by Rahula Sankr.ityayana.Motilal Banarsi Dass, Lahore.

References 287

Kalanjaramahatmya(KM) Kalanjaramahatmya. Manuscript India Office Library 2688a. Paper,

Devanagarı script.Kalidasa

Vallabhadeva’s Kommentar (Śarada-Version) zum Kumarasam. bhava desKalidasa hrsg. von M. S. Narayana Murti, unter der Mitarbeit vonKlaus L. Janert. Wiesbaden 1980.

Kalidasa’s Meghaduta. Edit. from the Manuscripts with theCommentary of Vallabhadeva and provided with a completeSanskrit–English Vocabulary [by] E. Hultzsch. With a Foreword byAlbrecht Wezler. New Delhi 1998.

The Raghupancika of Vallabhadeva, being the earliest commentary onthe Raghuvam. śa of Kalidasa. Vol. I. Critical Edition with Introductionand Notes by Dominic Goodall & Harunaga Isaacson. Groningen 2003.Groningen Oriental Studies 17.

Kane, P. V.1930–62 History of Dharmaśastra (Ancient and Mediaeval Religious and Civil

Law in India). Poona. 5 vols. Government Oriental Series, Class B,No. 6. [= Kane I to V]

KaumudımahotsavaKaumudı Mahotsava. Edit. by M. Ramakrishna Kavi and S.K.Ramanatha Sastri. Trivandrum 1929. Andhra Oriental Series 1.

Kaus. ıtaki-Brahman. a(Kaus.Br) The Kaus. ıtaki-Brahman. a [with the] Vyakhya of Udaya. Edit. by E. R.

Sreekrishna Sarma. Wiesbaden 1968–76. 3 vols. Verzeichnis derorientalischen Handschriften in Deutschland. Supplementband 9.1–3.

Kedarakhan. d. a(KeKh) Mahars.ivyasapran. ıtah. Skandapuran. antargatah. Kedarakhan. d. ah.

(Ratnaprabhabhas.avyakhyasahita). Anuvadaka Vrajaratna Bhat.t.acarya.Caukhamba Sam. skr.ta Pratis.t.hana, Delhi 2007.

Kennet, Derek2013 Reconsidering the decline of urbanism in late Early Historic and Early

Medieval South Asia. in: Robin, Christian Julien & JeremieSchiettecatte (eds.), Les preludes de l’Islam. Paris. pp. 331–53.

Kervran, Monique1999 Multiple Ports at the Mouth of the River Indus: Barbarike, Deb,

Daybul, Lahori Bandar, Diul Sinde. in: Ray 1999, pp. 70–153.Khan, F.A.

2005 Banbhore. A preliminary Report on the Recent ArchaeologicalExcavations at Banbhore. Department of Archaeology and Museums,Government of Pakistan. Fifth edition 2005.

Kielhorn, F.1905 Nagpur Museum Buddhist Inscription of Bhavadeva Ranakesarin. in:

JRAS 1905, 617–633.Kintaert, Thomas

2011–12 On the Role of the Lotus leaf in South Asian Cosmography. in: WZKSA54 (2011–12), 85–120.

Kirfel, Willibald1927 Das Puran. a Pancalaks.an. a. Versuch einer Textgeschichte. Leiden.

1967 Die Kosmographie der Inder nach Quellen dargestellt. Reprint,Darmstadt.

288 References

Klostermaier, Klaus1991 The Original Daks.a Saga. in: Sharma, Arvind (ed.), Essays on the

Mahabharata, Leiden. pp. 110–129.Ks.emendra

Br.hatkathamanjarı. Edit. by Śivadatta & Kaśinath Pan. d. urang Parab.Bombay 1931 (2nd ed.). Kavyamala 69.

Kuiper, F.B.J.1961 Rigvedic parye divi. in: Indo-Iranian Journal 5.3 (1961), 169–183.

1979 Varun. a and Vidus.aka. On the origin of the Sanskrit drama. Amsterdametc. Verh. v. d. KNAW, Afd. Let. N.R. 100.

Kulke, Hermann2012 Hindu Medieval Regional Kingdoms (600–1526 CE). in: Brill’s

Encyclopedia of Hinduism Volume IV, 51–72.Kurmapuran. a

(KuP) The Kurmapuran. a, crit. edit. by Anand Swarup Gupta. Varanasi 1971.All-India Kashiraj Trust.

Laks.mıdhara Kr.tyakalpataru(NK) Tr.tıyo bhagah. . Niyatakalakan. d. am edit. by K. V. Rangaswami Aiyangar.

Baroda 1950. Gaekwad’s Oriental Series CXI.

(TVK) As.t.amo bhagah. . Tırthavivecanakan. d. am, edit. by K. V. RangaswamiAiyangar. Baroda 1942. Gaekwad’s Oriental Series XCVIII.

Lambrick, H.T.1964 Sind. A General Introduction. Hyderabad. History of Sind Series I.

Larson, Gerald James1987 Encyclopedia of Indian Philosophies Vol. IV: Sam. khya. A Dualist

Tradition in Indian Philosophy, edit. by G.J. Larson & Ram ShankarBhattacharya. Delhi.

Lefevre, Vincent & Marie-Francoise Boussac2007 Chefs-d’œuvre du delta du Gange. Catalogue realise sous la direction de

Vincent Lefevre et Marie-Francoise Boussac. Etablissement public duMusee des arts asiatiques Guimet (2007–08), Paris.

Levi, Sylvain1905 Le Nepal. Etude historique d’un royaume hindou. Paris. 2 vols.

Licchavikalakabhileka(LA) [Licchavi Inscriptions, according to the edition of Vajracarya 1973]

Lienhard, Siegfried198 A History of Classical Poetry. Sanskrit–Pali–Prakrit. Wiesbaden. A

History of Indian Literature III.1.Limaye, V. P. & R. D. Vadekar (eds.)

1958 Eighteen Principal Upanis.ads. Vol. I. Upanis.adic Text with Parallelsfrom extant Vedic Literature, Exegetical and Grammatical Notes.Poona. Gandhi Memorial Edition.

Lingapuran. a(LiP) Śrı-Vyasa-mahars.iproktam. Śrı-Lingamahapuran. am, with the Sanskrit

commentary Śivatos.in. ı by Gan. eśa Natu. [Edit. by] Gangavis.n. u (son ofKr.s.n. adasa). Venkatesvara Press, Bombay V.S. 1981 [= ad 1924].Reprinted, with a Ślokanukraman. ı by Nagaśaran. a Sim. ha, by NagPublishers, Delhi 1989 (2nd ed. 1996).

Lohuizen, J.E. van1979 The Pre-Muslim Antiquities of Sind. in: South Asian Archaeology 1975.

Leiden. pp. 151–74.

References 289

1990 Mother Goddesses in Ancient India. in: Folia Indica by J.E. vanLohuizen-de Leeuw. [Edit. by Maurizio Taddei.] Association of SouthAsian Archaeologists in Western Europe, Naples.

Lorenzen, David N.1991 The Kapalikas and Kalamukhas. Two lost Śaivite sects. Second revised

edition, Delhi.Mahabharata

(MBh) The Mahabharata. For the first time critically edited by V. S. Suktankarand others. Poona 1927–59. 19 vols.

Maity, S.K.1960 The gold content of the later Imperial Gupta coins. in: The Journal of

the Numismatic Society of India 22.2 (1960), 266-268.

1961 Metrological study of the gold coins of early India. in: The Journal ofthe Numismatic Society of India 23 (1961), 259-266.

1970 Economic Life in Northern India in the Gupta Period (c. A.D. 300–550).Dehli (1st ed. Calcutta 1957).

Majumdar. N.G.1920 The Mun. d. eśvarı Inscription of the Time of Udayasena: The Year 30. in:

Indian Antiquary XLIX (1920), 21–29.Majumdar, R.C.

1971 History of Ancient Bengal. Calcutta.Majumdar, R. C. & A.S. Altekar (eds.)

1967 The Vakat.aka–Gupta Age (circa 200–550 ad). Delhi.Majumdar, R. C. et al. (eds.)

(HCI) The History and Culture of the Indian People. Bombay 1951–77. 11 vols.Bharatiya Vidya Bhavan.

Majumdar, Susmita Basu2007 Re-editing the Junwani Copper Plate Inscription of Mahaśivagupta

Balarjuna, Regnal Year 57. in: Bhattacharya, Gouriswar et al. (eds.),Kalhar: Studies in Art, Iconography, Architecture, and Archaeology ofIndia and Bangladesh. New Delhi pp. 286–295.

Mann, Richard D.2012 The Rise of Mahasena. The Transformation of Skanda-Karttikeya in

North India from the Kus.an. a to Gupta Empires. Leiden. Brill’sIndological Library.

Manusmr.ti(MaS) Manusmr.ti with the Sanskrit Commentary Manvantara-Muktavalı of

Kulluka Bhat.t.a, edit. by J. L. Shastri with English Introduction by S. C.Banerji. Delhi etc. 1983.

Martin, Montgomery1838 The History, Antiquities, Topography, and Statistics of Eastern India.

Volume I, Behar (Patna City) and Shahabad. London. Indian Reprint,Delhi 1976.

Matsyapuran. a(MtP) Śrımad-Dvaipayanamuni-pran. ıtam. Matsyapuran. am, etad pustakam

Anandaśramasthapan. d. itaih. sam. śodhitam. [Poona] 1981. AnandaAshrama Sanskrit Series 54. [Reprint of the ed. of H. N. Apte ad 1907]

Mayrhofer, Manfred(EWA) Etymologisches Worterbuch des Altindoarischen. I. & II. Band,

Heidelberg 1992–96. III. Band, Heidelberg 2001. IndogermanischeBibliothek. II. Reihe Worterbucher.

290 References

Mehta, R.N. & A.M. Thakkar1978 M.S. University Copper Plates of the Time of Toraman. a. Baroda.

Maharaja Sayajirao University Archaeology Series 14.Meister, Michael W.

1981 Mun. d. eśvarı: Ambiguity and Certainty in the Analysis of a Temple Plan.in: Williams, Joanna G. (ed.), Kaladarśana. American Studies in the Artof India. AIIS, New Delhi etc. pp. 77–89.

Melzer, Gudrun2009 The Wrathful Śiva and the Terrifying Great Goddess in Eastern Indian

Art: Andhakari, Bhairava, and Camun. d. a. in: Journal of Ancient IndianHistory XXV (2008–09), 132–90.

Mertens, Annemarie1998 Der Daks.amythus in der episch-puran. ischen Literatur. Beobachtungen

zur religionsgeschichtlichen Entwicklung des Gottes Rudra-Śiva imHinduismus. Wiesbaden. Beitrage zur Indologie Band 29.

Meyer, J.J.1937 Trilogie altindicher Machte und Feste der Vegetation. Zurich/Leipzig.

Mirnig, Nina2013 “Favoured by the Venerable Lord Paśupati.” Tracing the Rise of a New

Tutelary Deity in Epigraphic Expressions of Power in Early MedievalNepal. in: Bosma, N. & Nina Mirnig (eds.) 2013, 325–347.

Mishra, Shyam Manohar1977 Yaśovarman of Kanauj. A study of political history, social and cultural

life of northern India during the reign of Yaśovarman. New Delhi.Mukherji, S. C.

1999 The Royal Charters of King Madanapala and the chronology of the Palakings of Bengal and Bihar. in: Journal of Bengal Art, 4, 61–65.

Narasim. hapuran. a(NsP) Edit. by Pus.pendra Kumar. With a Ślokanukraman. ı compiled by Nag

Śaran. Sim. ha. Nag Publishers, Delhi 1987.Neuss, Jurgen

2003 The Temple of Mun. d. eśvarı: Reconsidering the Evidence. in: BerlinerIndologische Studien 15/16/17 (2003), 531–585.

Nigam, J.S.1977 Some post-Gupta terracottas from Kanauj. in: JRAS 52 (1977), 214–18.

Niśvasatattvasam. hita(Guhya) The Guhyasutra. Transcribed by Dominic Goodall et al. from the

palm-leaf Nepalese MS No. 1–127, NGMPP A 41/14 supplemented withreadings from its Kathmandu apograph MS No. 5–2406, NGMPP A159/18.

(Mukha) The Mukhagama. Transcribed by Dominic Goodall et al. from thepalm-leaf Nepalese MS No. 1–127, NGMPP A 41/14, supplemented withreadings from its Kathmandu apograph MS No. 5–2406, NGMPP A159/18. Electronic text.

(Mula) The Mulasutra. Transcribed by Dominic Goodall from the palm-leafNepalese MS MS No. 1–127, NGMPP A 41/14, supplemented withreadings from its Kathmandu apograph MS No. 5–2406, NGMPP A159/18. Electronic text.

Oberhammer, Gerhard1977 Strukturen yogischer Meditation. Wien. Veroffentlichungen der

Kommission fur Sprachen und Kulturen Sudasiens 13.

References 291

Oberlies, Thomas2000 Krieglisten und ungeziemendes Benehmen: die Askesepraktiken der

Paśupatas. in: Tsuchida, R & A. Wezler (eds.), Haranandalaharı.Reinbek. pp. 175–191.

2003 A Grammar of Epic Sanskrit. Berlin/New York. Indian Philology andSouth Asian Studies 5.

Oldham, C.E.A.W. (ed.)1926 Journal of Francis Buchanan kept during the survey of the district of

Shahabad in 1812–13. Patna.Ozha, Vajeshankar G.

1889 The Somnathpattan Praśasti of Bhava Br.ihaspati. With an introductionby G. Buhler. in: WZKM 3, pp. 119.

Padmapuran. a(PdP) Śrı-Padmam. Mahapuran. am, [edit. by ] Ks.emaraja śrı-Kr.s.n. adasa and

Gangavis.n. u Śrıkr.s.n. adasa. Venkatesvara Press, Bombay V.S. 1984 (ad1927). 3 vols. [Reprinted by Nag Publishers, Delhi 1984 (2nd ed. 1996).Vol. 4, Ślokanukraman. ı by Nagaśaran. a Sim. ha. Nag Publishers, Delhi1984 (19962)]

Pakistan ArchaeologyPakistan Archaeology published by the Department of Archaeology,Ministry of Education, Government of Pakistan, Karachi. Number 1(1964) – .

Pal, Pratapaditya1974 The Arts of Nepal. Part 1: Sculpture. Leiden. Handbuch der

Orientalistik.Pande, B.M.

1994 Thanesar Excavations. in: South Asian Archaeology 1993. Proceedingsof the Twelfth International Conference of the European Association ofSouth Asian Archaeologists, Helsinki University 59 July 1993. Vol. II.Edited by Asko Parpola & Petteri Koskikallio. Helsinki, pp. 563–570.

Panikkar, Shivaji K.1997 Saptamatr.ka, Worship and Sculptures. An Iconological Interpretation of

Conflicts and Resolutions in the Storied Brahmanical Icons. New Delhi.Pan. ini

(P) The Vyakaran. a-Mahabhas.ya of Patanjali, edited by F. Kielhorn. 3rdedition, revised and furnished with additional readings, references andselect critical notes by K.V. Abhyankar. Poona 1962–72. 5 vols.

Parpola, Asko1992 The Metamorphoses of Mahis.a Asura and Prajapati. in: Hoek, A.W. van

den, D.H.A. Kolff, M.S. Oort (eds.), Ritual, State and History in SouthAsia. Essays in Honour of J.C. Heesterman. Leiden 1992. pp. 275–308.

Paśupatasutra(PS) The Paśupatasutra with the Pancarthabhas.ya of Kaun. d. inya. Edit. by R.

Ananthakrishna Sastri. Trivandrum 1940. Trivandrum Sanskrit SeriesCXLIII.

Patil, D.R.1963 The Antiquarian Remains in Bihar. Patna.

Patna Museum Catalogue2001 Patna Museum Catalogue. Stone Sculptures and other Antiquities.

Editor Naseem Akhtar. Patna.Petech, Luciano

1984 Mediaeval History of Nepal. Rome. Serie Orientale Roma LIV.

292 References

Peterson, Peter1895 A Collection of Prakrit and Sanskrit Inscriptions. Bhavnagar s.d.

[c. 1895]Pingree, David

1981 Jyotih. śastra. Astral and Mathematical Literature. Wiesbaden. A Historyof Indian Literature VI.4.

Pires, Edward E.1934 The Maukharis. Madras.

Pratijnayaugandharayan. aDigitalisierte Textkonstitution, Ubersetzung und Annotierung vonMatthias Ahlborn. Wurzburg 2007 (dissertation).

Puran. apancalaks.an. a(PPL) see Kirfel 1927.

PuratattvaPuratattva. Bulletin of the Indian Archaeological Society. IAS, NewDelhi 1967–

Purus.ottamadevaTrikan. d. aśes.ah. nama sam. skr.tapracınakos.ah. ,śrı-Purus.ottamadeva-nr.pativaren. a viracitah. ,śrı-Śılaskandha-mahanayakayativaren. a Sararthacandrika-namikayavyakhyaya samanvitah. . Ks.emaraja-śrı-Kr.s.n. a-Dasa. Bombay 1916.

Raghavan, V.1963 Bhoja’s Śr.ngara Prakaśa. Madras.

Rahman, Aman Ur & Harry Falk2011 Seals, Sealings and Tokens from Gandhara. Wiesbaden. Monographien

zur indischen Archaologie, Kunst und Philologie, Band 21. Studies inthe Aman ur Rahman Collection volume 1.

Rahman, S.S.M.2000 Archaeologcal Investigation in Bagura District. International Centre for

Study of Bengal Art, Dhaka.Ramayan. a

(Ram.) The Valmıki-Ramayan. a. Crit. edit. for the first time by a board ofeditors. Baroda 1960–75. 7 vols.

Ramesh, K.V.1984 Chalukyas of Vatapi. Delhi.

1986 Three early charters from Sanjeli in Gujarat. in EI 40 (1986), 175–186.

Ramesh, K.V. & S.P. Tewari1990 A Copper-plate Hoard of the Gupta Period from Bagh, Madhya

Pradesh. Archaeological Survey of India. New Delhi.Rao, Gopinath T.A.

1914 Elements of Hindu Iconography. Madras. [Reprinted by Paragon BookReprint Corp., New York 1968 in 2 volumes, 4 parts.]

Rauravasutrasam. grahaRauravagama. Vol. I. Edition critique par N.R. Batt. Introduction: LesAgamas Civaıtes par Jean Filliozat. Pondichery 1961. Publications del’Institut Francais d’Indologie 18/1.

Raven, Ellen M.1994 Gupta Gold Coins with a Garud. a-Banner. Groningen. 2 vols. Gonda

Indological Studies 1.Ray, Himanshu Prabha (ed.)

1999 Archaeology of Seafaring. The Indian Ocean in the Ancient Period.Delhi.

References 293

Rayakwar, G.L. & Rahul Kumar Singh1994 Mahaśivagupta Balarjuna ka 57 vem. rajya vars.a ka Junavanı (Malhar)

tamralekha. in: Puratana 9 (1994), 146–47.

2005 Utkırn. a Lekha. Lekhana evam. sam. padana: Balacandra Jain.Parivardhana: G.L. Rayakavara & R.K. Sim. ha. Raipur.

Rolland, Pierre1973 Le Mahavrata: Contribution a l’etude d’un rituel solennel vedique.

Gottingen. Nachrichten der Akademie der Wissenschaften in Gottingen.I: philologisch-historische Klasse, 1973.3.

Salles, Jean-Francois2001 France-Bangladesh Joint Venture Excavations at Mahasthangarh: first

interim report, 1993–99. Edited by Md. Shafiqul Alam & Jean-FrancoisSalles. Dept. of Archaeology, Ministry of Cultural Affairs, Dhaka.[French translation: Mahasthan I: Pundranagara, cite antique duBengale: Rapport preliminaire 1993–99. Turnhout 2007]

Salomon, Richard1976 New Inscriptions from the Mun. d. eśvarı Temple. in: Journal of the Bihar

Research Society LXII (1976), 104–08.

1983 More New Inscriptions from Mun. d. eśvarı Temple. in: Mukherjee, B.N. etal. (eds.), Sri Dinesacandrika. Sudies in Indology (Shri D.C. SircarFestschrift). Delhi. pp. 73–78.

1989 New Inscriptional Evidence for the History of the Aulikaras ofMandasor. in: IIJ 32 (1989),1–36.

2010 Like Father, Like Son: Poetic Strategies in “The Middle Brother”(Madhyama-vyayoga) Attributed to Bhasa. in: IIJ 53 (2010), 1–22.

Salomon, Richard & Michael D. Willis1990 A ninth-century Umamaheśvara image. in: Artibus Asiae 50 (1990),

148–155.Sam. khyakarika

Iśvarakr.s.n. a-viracita Sam. khyakarika. Rama-Śankara-Tripat.hi-viracitayaTattvaprabhakhyaya vyakhyaya ajnatakartr.kaya Yuktidıpikakhyayavivr.tya ca vibhus.ita. Varan. ası 1970.

Sanderson, Alexis1988 Śaivism and the Tantric Traditions. in: S. Sutherland et al. (eds.), The

World’s Religions. London.

2001 History through textual criticism in the study of Śaivism, thePancaratra and the Buddhist Yoginıtantras. in: Grimal, Francois (ed.),Les Sources et le Temps (Sources and Time). Pondicherry. pp. 1–47.

2003–04 The Śaiva Religion Among the Khmers, Part I. Bulletin de l’Ecolefrancaise d’Extreme-Orient 90 (2003–04), 352–464.

2004 Religion and the State: Śaiva Officiants in the Territory of the King’sBrahmanical Chaplain. in: IIJ 47 (2004), 229–300.

2006 The Lakulas: New Evidence of a System Intermediate betweenPancarthika Paśupatism and Agamic Śaivism. in: The IndianPhilosophical Annual 24 (2006), 143–217.

2009 The Śaiva Age — The Rise and Dominance of Śaivism during the EarlyMedieval Period. in: Genesis and Development of Tantrism. Edit. byShingo Einoo. Institute of Oriental Studies, University of Tokyo.pp. 41–349.

294 References

2013 The Impact of Inscriptions on the Interpretation of Early ŚaivaLiterature. in: Bosma, N. & Nina Mirnig (eds.) 2013, 211–244.

forthc. Śaivism and Brahmanism. [Revised Gonda lecture 2006].

Sankaranarayanan, S.1977 The Vis.n. ukun.d. is and Their Times (An Epigraphical Study). Foreword

T.V. Mahalingam. Delhi.Sanyal, Rayat

2010 Copperplate Inscriptions of West Bengal: Finding Find-spots andLocating Localities. in: Pratna Samiksha New Series 1 (2010), 107–134.

SaromahatmyaSee Vamanapuran. a.

Śatapathabrahman. a(ŚBr) The C. atapatha-Brahman. a in the Madhyandina-C. akha with extracts

from the commentaries of Sayan. a, Harisvamin and Dvivedaganga editedby Albrecht Weber. Reprint, Varanasi 1964.

SaundaryalaharıThe Saundaryalaharı or Flood of Beauty, traditionally ascribed toŚankaracarya. Edited, translated, and presented in photographs by W.Norman Brown. Cambridge, Massachusetts 1958. The Harvard OrientalSeries 43.

Schmiedchen, Annette2007 Epigraphical Evidence for the History of Atharvavedic Brahmins. in:

The Atharvaveda and its Paippaladaśakha: Historical and Philologicalpapers on a Vedic tradition. Edited by Arlo Griffiths and AnnetteSchmiedchen. Aachen. pp. 355–379.

2013 Patronage of Śaivism and Other Religious Groups in Western Indiaunder the Dynasties of the Kat.accuris, Gurjaras and Sendrakas from the5th to the 8th Centuries. in: Bosma, N. & Nina Mirnig (eds.),Epigraphical Evidence for the Formation and Rise of Early Śaivism. =Indo-Iranian Journal 56.3–4 (2013), 349–363.

Schokker, G. H.1966 See Śyamilaka

Sears, Tamara I.2004 Housing Asceticism: Tracing the Development of Mattamayura Śaiva

Monastic Architecture in Early Medieval Central India (c. 8th to 12thCenturies A.D.). University of Pennsylvania, thesis (UMI Number3138073).

2008 Constructing the Guru: Ritual Authority and Architectural Space inMedieval India. in: The Art Bulletin 90.1 (2008), 7–31.

Sen, Tansen2003 Buddhism, Diplomacy, and Trade. The Realignment of Sino-Indian

Relations, 600–1400. Honolulu. Asian Interactions and Comparisons.Sharma, Aśoka Prapanna

s.d. Hrishikesh Narayan Sri Bharat Mandir. Hrishikesh.Sharma, B.N.

1971–72 Abhis.eka in Indian Art. in: Journal of the Oriental Institute (Baroda),Volume XXI (1971–72), pp. 108–113.

Sharma, R.S.2001 Early Medieval Indian Society. A Study in Feudalisation. London.

References 295

Shastri, Ajay Mitra1995 Inscriptions of the Śarabhapurıyas, Pan.d. uvam. śins and Somavam. śins.

Delhi. 2 vols.

1997 Vakat.akas: Sources and History. New Delhi.

2001 Malhar Plates of Mahaśivagupta, year 57. in: Bharatiya PurabhilekhaPatrika (Studies in Epigraphy) XXVII (2001), 25–48.

Shastri, H.G.2000 Gujarat under the Maitrakas of Valabhı. (History and Culture of

Gujarat during the Maitraka Period—Circa 470–788 A.D.). Vadodara.Gaekwad’s Oriental Series 180.

Sinha, B.P.1977 Dynastic History of Magadha. Delhi.

Sircar, D. C.1950 Gaud. a–Kamarupa Struggle in the Sixth and Seventh centuries. in:

Indian Historical Quarterly 26 (1950), 241–46.

1964 Maukhari Inscription from Jaunpur. in: Journal of Indian HistoryXLII.1 (1964), 127–30.

1965 Indian Epigraphy. Delhi.

1966 Indian Epigraphical Glossary. Delhi etc.

1971 Studies in the Geography of Ancient and Medieval India. Second revisedand enlarged edition, Delhi.

1973–74 Ban. gad.h Stone Inscription of the time of Nayapala. in: Journal ofAncient Indian History 7, 135–158, Plate 1.

1980–82 Murtiśiva’s Bangarh prasasti of the time of Nayapala. in: Journal ofAncient Indian History 13, 34–56.

(SI) Select Inscriptions bearing on Indian History and Civilization. Volume I,Calcutta 1942 [2nd ed.: Delhi/Madras 1965]. Volume II, Delhi 1983.

Śivapuran. a(ŚiP) Śrı-Śivamahapuran. am, [edit. by] Khemaraja (son of Kr.s.n. adasa).

Venkatesvara Press, Bombay V.S. 2011 [= ad 1954].Skandapuran. a

(SkP) Śrı-Skandamahapuran. am, [edit. by] Ks.emaraja Śrıkr.s.n. adasa.Venkatesvara Press, Bombay V.S. 1967 [ad 1910]. 7 vols. [Reprinted byNag Publishers, Delhi 1986 (2nd ed. 1995). 7 vols. Vol. 8,Ślokanukraman. ı by Nagaśaran. a Sim. ha. Nag Publishers, Delhi 1990].

Skandapuran. a (SP)(SP I) The Skandapuran. a, Volume I. Adhyayas 1–25. Critically Edited with

Prolegomena and English Synopsis by R. Adriaensen, H.T. Bakker & H.Isaacson. Groningen 1998. Supplement to Groningen Oriental Studies.

(SP II A) The Skandapuran. a, Volume II A. Adhyayas 26–31.14: The Varan. asıCycle. Critical Edition with an Introduction, English Synopsis &Philological and Historical Commentary by Hans T. Bakker & HarunagaIsaacson. Groningen 2004. Supplement to Groningen Oriental Studies.

(SP II B) The Skandapuran. a, Volume II B. Adhyayas 31 – 52: The Vahana andNaraka Cycles. Critical Edition with an Introduction and AnnotatedEnglish Synopsis by Hans T. Bakker, Peter C. Bisschop, Yuko Yokochi,in cooperation with Nina Mirnig and Judit Torzsok. Brill, Leiden 2014.Supplement to Groningen Oriental Studies.

296 References

(SP III) The Skandapuran. a, Volume III. Adhyayas 34.1–61, 53–69: TheVindhyavasinı Cycle. Critical Edition with an Introduction & AnnotatedEnglish Synopsis by Yuko Yokochi. Groningen/Leiden 2013. Supplementto Groningen Oriental Studies.

(SP 167) [For edition of Adyaya 167 see Bisschop 2006.]

(SPBh ) Skandapuran. asya Ambikakhan. d. ah. , sam. padakah. Kr.s.n. aprasadaBhat.t.araı. Kathmandu 1988. Mahendraratnagranthamala 2.

Slusser, Mary Shepherd1998 Nepal Mandala. A cultural study of the Kathmandu Valley. Vol. 1 Text,

Vol. 2, Plates. Kathmandu (1st ed. Princeton 1982). 2 vols.Somadeva

(KSS) Kathasaritsagarah. , kaśmırapradeśavasina śrıramabhat.t.atanudbhavenamahakaviśrısomadevabhat.t.ena viracitah. . [With a Bhumika andAnukraman. ı edited by] Jagadıśa Lala Śastri. Delhi 1970.

Srinivasan, P.R.1991 Mundesvari Temple Inscription. in: Margabandhu, C. et al. (eds.),

Indian Archaeological Heritage. Shri K.V. Soundara Rajan Festschrift.Vol. I. Delhi.

Steiner, Roland1997 Untersuchungen zu Hars.adevas Nagananda und zum indischen

Schauspiel. Swisttal-Odendorf. Indica et Tibetica 31.Stietencron, Heinrich von

1966 Indische Sonnenpriester. Samba und die Śakadvıpıya-Brahman. a. Einetextkritische und religionsgeschichtliche Studie zum indischenSonnenkult. Wiesbaden. Schriftenreihe des Sudasien-Instituts derUniversitat Heidelberg Bd. 3.

ŚyamilakaPart 1. The Padatad. itaka of Śyamilaka. Edit. by G.H. Schokker. Part 2.A Translation by G.H. Schokker and P.J. Worsley, with a completeWord-index of the four ancient Sanskrit Bhanas by G.H. Schokker. TheHague/Paris 1966 (Indo-Iranian Monographs 9), Dordrecht/Boston1976. 2 vols.

Syed, Renate2008 Early Terracottas from Kanauj. Chessmen? in: Raven, Ellen M. (ed.),

South Asian Archaeology 1999. Groningen. pp. 363–74.Taittirıya-Sam. hita

(TaiSa) Taittirıya-Sam. hita, herausgegeben von Albrecht Weber. Leipzig1871–72. 2 vols. Indische Studien. Beitrage fur die Kunde des indischenAlterthums.

TaranathaTaranatha’s History of Buddhism in India. Translated from the Tibetanby Lama Chimpa [and] Alaka Chattopadhyaya. Simla 1970.

Thaplyal, Kiran Kumar & D.P. Tewari1985 Inscriptions of the Maukharıs, Later Guptas, Pus.pabhutis and

Yaśovarman of Kanauj. Delhi.

2006 Seal-die of the Śaiva shrine Kalanjara from Sanchankot. in: ResearchBulletin Vishveshvaranand Vedic Research Institute 4 & 5 (2006)(Jagannath Agrawal Centenary Volume), pp. 451–56.

Tieken, Herman1993 The So-called Trivandrum Plays Attributed to Bhasa. in: WZKSA

XXXVII (1993), 5–44.

References 297

TırthankTırthanka. in: Kalyan. a 31.1 (1957), 1–704. Gitapress, Gorakhpur.

Torzsok, Judit2001 Reading the Skandapuran. a with R and A4. Groningen. (unpublished)

2004 Three Chapters of Śaiva Material Added to the Earliest KnownRecension of the Skandapuran. a. in Bakker 2004, pp. 17–39.

2011 Kapalikas. in: Brill’s Encyclopedia of Hinduism, Vol. III, 355–61. Leiden.

Tripathi, Rama Shankar1964 History of Kanauj to the Moslim Conquest. With a foreword by L.D.

Barnett. Delhi.Tripathy, Snigdha

1998 Inscriptions of Orissa: Circa 5th–8th centuries A.D. Motilal Banarsidass,Delhi. Indian Council of Historical Research.

Turner, R.L.(CDIAL) Comparative Dictionary of the Indo-Aryan Languages. Oxford

University Press, London 1966.Vajracarya, Dhanavajra

1973 Licchavikalakabhileka (anuvada, aitihasika, vyakhyasahita) Kathmandu1973.

VakpatirajaGaudavaho. Edit. with an Introduction, Sanskrit Chaya, EnglishTranslation, Notes, Appendices, and Glossary by N.G. Suru.Ahmedabad/Varanasi 1975. Prakrit Text Society Series No. 18.

VallabhadevaThe Subhas. itavali of Vallabhadeva. Edit. by Peter Peterson and PanditDurgaprasada. Poona 1961 (original ed. Bombay 1886). BombaySanskrit and Prakrit Series 31.

Vamanapuran. a(VamP) The Vamana puran. a. Crit. edit. by Anand Swarup Gupta. All-India

Kashiraj Trust, Varanasi 1967.Van Put, Ineke

2007 The Names of Buddhist Hells in East Asian Buddhism. in: PacificWorld. Journal of the Institute of Buddhist Studies 9 (2007), 205–229.

Varahamihira(BS) Br.hat Sam. hita. With English Translation, Exhaustive Notes and

Literary Comments by M. Ramakrishna Bhat. Delhi 1981–82. 2 vols.Varahapuran. a

(VarP) The Varaha Puran. a, critically edit. by Anand Swarup Gupta. All-IndiaKashiraj Trust, Varanasi 1981.

Vats, M.S.2002 A Note on two Mediæval Rock-cut Sculptures in the Kalanjar Fort. in:

Annual Reports of the Archæological Survey of India for the Years1930–31, 1931–32, 1932–33, 1933–34. Edit. by C.L. Fabri. New Delhi(2002). pp. 317–18.

Vayupuran. a(VaPAA) Mahamuniśrımad-Vyasa-pran. ıtam. Vayupuran. am, etat pustakam

anandaśramasthapan.d. itaih. sam. śodhitam. [Poona] 1983. AnandaAshrama Sanskrit Series 49. [Reprint of the edition by H. N. Apte ad1905]

(VaPVenk) The Vayumahapuran. am. Edit. by Khemaraja. Delhi 1983. NagPublishers. [Reprint of the Venkatesvara edition of ad 1895.]

298 References

Verma, T.P. & Arvind Kumar Singh1994 A Corpus of Lichchhavi Inscriptions of Nepal. New Dehli.

VidyakaraThe Subhas. itaratnakos.a compiled by Vidyakara. Edit. by D.D. Kosambiand V.V. Gokhale. With an introduction by D.D. Kosambi. Cambridge(MA) 1957. Harvard Oriental Series 42.

Virkus, Fred2004 Politische Strukturen im Guptareich (300–550 n. Chr.). Wiesbaden.

Asien- und Afrika-Studien der Humboldt-Universitat zu Berlin 18.Viśakhadatta

Mudraraks.asa, or The Signet Ring: a Sanskrit drama in seven acts. Crit.ed. with copious notes, transl., introd. and appendices, Indexes etc. byK.H. Dhruva. Thoroughly revised and enlarged 3rd ed., Poona 1930.

Vis.n. udharmottarapuran. a(VDhP) Vis.n. udharmottarapuran. a. [edit. by] Ks.emaraja Śrıkr.s.n. adasa.

Venkatesvara Press, Bombay V.S. 1969 [1912].

Vis.n. udharmottarapuran. a. Third Khan. d. a. Vol. 1 Text, Critical Notesetc. Critically edited by Priyabala Shah. Vol. 2 Introduction,Appendixes, Indexes etc. A study on a Sanskrit Text and Ancient IndianArts. Vadodara. 2 vols. Gaekwad’s Oriental Series 130, 137.

Vis.n. upuran. a(ViP) The Vis.n. upuran. am, critically edited by M.M. Pathak. Pada-Index by

Peter Schreiner. Vadodara 1997–99. 2 vols.Vis.n. usmr.ti

(ViS) The Vis.n. usmr.ti with the commentary Keśavavaijayantı ofNandapan. d. ita. Edit. by V. Krishnamacharya. Madras 1964. 2 vols. TheAdyar Library Series 93.

Vogel, Claus1979 Indian Lexicography. A History of Indian Literature, 5.4. Otto

Harrassowitz, Wiesbaden.Warder, A.K.

1989–92 Indian Kavya Literature. Delhi. 6 vols.Watters, Thomas

1904–05 On Yuan Chwang’s Travels in India (ad 629–645). Edit. after his deathby T.W. Rhys Davids and S.W. Bushell with two maps and an itineraryby Vincent A. Smith. RAS, London 1904–05. 2 vols. (Reprint:Munshiram Manoharlal, Delhi 19732.)

Williams, Joanna Gottfried1982 The Art of Gupta India. Empire and Province. Princeton.

Willis, Michael2005 Later Gupta History: Inscriptions, Coins and Historical Ideology. in:

JRAS Series 3, vol. 15.2 (2005), 131–150.

2009 The Archaeology of Hindu Ritual. Temples and the Establishment of theGods. Cambridge.

Wink, Andre1991 Al-Hind. The Making of the Indo-Islamic World. Volume I: Early

Medieval India and the Expansion of Islam, 7th–11th Centuries. Leidenetc.

1992 Kanauj as the Religious and Political Capital of Early Medieval India.in: Bakker, Hans (ed.), The Sacred Centre as the Focus of PoliticalInterest. Groningen. pp. 101–117.

References 299

Winternitz, Moriz1920 Geschichte der indischen Literatur, Band 3. Leipzig.

Xuanzang(Si-yu-ki) The Si-yu-ki. Buddhist Records of the Western World. Translated from

the Chinese of Hiuen Tsiang (ad 629) by Samuel Beal. London 1884(Indian edition, Delhi 1969).

(Life) See Huili.

Yadav, Rajendra2002–03 Unpublished Gupta Remains from Kalanjar. in: Pragdhara. Journal of

the U.P. State Archaeology Department. No. 13 (2002–03), 223–26.

2009–10 Kalanjar: An Important Gupta Art Centre of Bundelkhand. in:Pragdhara. Journal of the U.P. State Archaeology Department. No. 20(2009–10), 65–78.

Yokochi, Yuko1999 Mahis.asuramardinı Myth and Icon. Studies in the Skandapuran. a II. in:

Studies in the History of Indian Thought (Indo Shisoshi Kenkyu) 11,65–103.

2000 The story of the seven brahmans in the Harivam. śa. Studies in theSkandapuran. a IV. in: Tsuchida, Ryutaro & A. Wezler (eds.),Haranandalaharı. Volume in Honour of Professor Minoru Hara on hisSeventieth Birthday. Reinbek. pp. 525–552.

2004a The Relation between the Skandapuran. a and the Avantyakhan. d. a. in:Bakker 2004b, 79–109.

2004b The Rise of the Warrior Goddess in Ancient India. A Study of the MythCycle of Kauśikı-Vindhyavasinı in the Skandapuran. a. Thesis, Groningen.

2013a The Development of Śaivism in Kot.ıvars.a, North Bengal, with specialreference to the Kot. ıvars.a-Mahatmya in the Skandapuran. a. in: Bosma,N. & Nina Mirnig (eds.) 2013, 295–324.

2013b See SP III.

Index

Abhidhanacintaman. i, 248abhis.eka, 3, 6, 87, 161, 162, 164acarya, 4, 8, 14, 20, 45, 78, 132, 144,

146, 147, 149–151, 171, 213Adityadasa, 195Adityasena, LG, 5, 50, 51, 58, 60, 61,

81, 82, 130–132Adityavardhana, 66, 79, 84, 90Adityavarman, 42, 48, 51, 55Agneyı, 253Agnihotra, 245agrahara, 30, 66, 132, 192, 218Ah.mad ibn-Yah. ya al-Baladhurı, 116Aihole Inscription, 108, 109, 111, 112Aindrı, 255Airikin. a, 31Ajiravatı, 95, 97, 99, 158, 172Ajıvakas, 8, 44Ajjaka, 112Ajjhitadevı, 28Akbarpur, 88Aks.amala, 159Alexander, 113, 116Alexandria, 115, 116Allahabad Museum, 201Allahabad pillar, 128Almore District, 227A-lo-na-shun, 128, 129Alór, 114Al.upas, 108Amarakan. t.aka, 211amr.ta, 212Amr.tadeva, 242, 245–247Am. śuvarman, 120–122, 132Am. śuvarman Era, 120, 121, 132Anandapura, 105Anantamahayı, 204Anantavarman, 43–47, 51, 60, 102Andhaka, 5, 14, 16, 69, 70, 150Andhaka Cycle (SP), 5, 10, 16, 69, 70Andhra, 3, 53–56, 62, 110, 112Andhra Pradesh, 2, 45, 142

Andhras, 53, 55Anekarthasam. graha, 59Angiras, 160, 253Angkor, 142Anguttara-Nikaya, 85Anukraman. ika, 16, 174, 212Aphsad Stone Inscription, 58, 82, 130Arab, 113, 116–120, 123Arabian Sea, 104, 119, 140Arabic, 113–115Arang, 202Archaeological Museum Hardwar, 185Arena of charitable offerings, 18, 123,

124, 128Arjuna, 13, 36, 37, 165, 203Arjuna (river), 186Arjuna Kartavırya, 204Arrah, 48, 68Arrian, 113, 116Arthaśastra, 60, 82, 108Arun. aśva, 129Arya, 246Aryamanjuśrımulakalpa, 89, 100, 259↪As., al-, 117As.ad. hi, 214ashes, 13, 78, 139, 152, 153, 174–176,

187Ashok Nagar District, 213Asidhara vow, 96aśrama, 160Assam, 129, 173, 254as.t.amurti, 166Astraśiva, 144Asuras, 11, 18, 127, 163Asuri, 8, 9Aśvapati, 29, 42Atharvan. a brahmins, 66, 254Atharvavedin, 253, 254, 260Atimarga, 151, 258Atri, 140, 141, 144At.t.ahasa, 78Aulikaras, 4, 32–36, 50–55, 62, 77, 82

301

302 Index

Auśanasa, 168Avantisundarıkatha, 36Avantivarman, 12–14, 16, 17, 48, 62,

63, 67, 68, 71, 73, 74, 77, 79,80, 85, 137, 226

Avantyakhan. d. a, 70, 167avataras (28), 214, 215, 255Avimukta, 12, 13, 133Avimuktajja, 133Avimuktaks.etra, 12Avimukteśvara, 12, 14, 133, 139, 151,

173ayatana, 10, 13, 18, 142, 161, 173, 210Ayodhya, 242, 245–247Ayomukha, 204

Babhravya, 203Babiah, 115Badami, 104, 110, 113Bagh, 147, 171Baghbhairava Temple, 255Baghelkhand, 29, 60Bahrain, 116, 117Bahumam. sa, 6, 252–257, 259, 264, 268Baladitya, 67, 68Baladitya (pati), 67Baleśvara-bhat.t.araka Temple, 144Bali (demon), 25, 70Balkandeswara, 205–208Ban. a, 15, 17, 19, 42, 48, 65, 71, 74, 75,

77, 78, 80, 81, 84–88, 92, 93,95–98, 114, 115, 124,155–162, 164, 165, 170, 172,259

Ban. a (demon), 251, 252Ban. apura, 248Banbhore, 117, 118Bandhuvarman, 34, 54bandin, 155Bangarh, 2, 241, 247–250, 255, 256Bangladesh, 57, 82, 242, 247, 256Barabar Hill, 43–45Barah Copperplate Inscription, 66,

192, 203Barauz, 116, 117Barwa, 41Bauma-Kara (kings), 237Beas, 113, 115, 116Bengal, 50, 60, 87, 89, 97, 102, 103,

138, 156, 243, 259, 260Berhampur, 57, 97Betwa, 31, 38, 52

bha◦, 36Bhabua, 222, 223Bhadrakalı, 175, 186, 212Bhadrakarn. a, 181Bhadrakarn. ahrada, 177, 181Bhadrakarn. eśvara, 181Bhadravihara, 97Bhadravis.n. u, 227Bhadreśa, 151Bhadreśvara, 13, 17, 151, 174,

176–180, 183–187, 210, 212,238

Bhadreśvara (Kalanjara), 199,201–203, 209, 210, 212, 218,219

Bhadreśvaramahatmya, 2, 174Bhagavaddos.a, 33Bhagavat, the, 122, 140Bhagavata religion, 35Bhagırathı, 57Bhagwana Sej, 208Bhairava, 232, 254, 256, 259, 260Bhairava (Mun. d. eśvarı), 232, 233Bhairavacarya, 78, 79, 171, 259Bhairavatantras, 258Bhan. d. i, 91, 92Bhanudayin, 226–228Bhanugupta, 32–34, 50, 57, 82, 242Bhanugupta, 33Bharabhuti, 140, 214Bharat Kala Bhavan, 244Bharat Mandir Museum, 181, 182, 186Bharata, 156, 203Bharata, 15, 125Bharavi, 36, 37Bhargava, 77, 161, 168Bharhut, 205Bhasa, 71, 72Bhasarvajna, 147Bhaskaravarman (Kamarupa), 90–92,

98, 99, 113, 123, 129Bhaskaravarman (Maukhari), 56bhasmasomodbhava, 174, 212Bhatsu, 15, 71Bhat.t.araı, 8, 70, 252, 263Bhava Br.haspati, 140Bhavadeva Ran. akesarin, 202Bhavas (66), 149, 214Bhımarjunadeva, 120Bhımasoma, 144, 148Bhıt.a, 190, 191, 201, 212Bhitargaon, 35

Index 303

Bhitarı, 27, 35Bhogavarman, 20, 61, 132Bhoja I, 66, 192Bhojakas, 67Bhr.gu, 158, 159Bhr.kut.ı, 120Bhulun. d. a, 147, 171Bhutapati, 43Bihar, 27, 28, 50, 51, 60, 68, 95, 100,

128, 131, 138, 222, 223, 241,260

Bilva-vat.ika, 78Bına River, 32Bodh Gaya, 101, 123Bodhi Tree, 102, 123, 128Brahma, 11, 118, 157, 160, 163–169,

171, 177, 183, 184, 210, 212,215, 221, 253, 257

brahmacarin, 165Brahmadatta, 193Brahmadhanvan, 193, 194Brahmaloka, 161, 167, 169Brahmamantras, 150brahman, 8, 160, 161Brahman. d. a, 175Brahman. d. apuran. a, 3Brahmanism, 7, 187Brahmaputra River, 40, 82, 253, 254Brahmasaras, 157, 161, 167–169

Brahma Sarovar, 169brahmasaras, 161, 167brahmasattra, 19, 125Brahmavarta, 168, 184Brahmavarta (in Rishikesh), 183, 184Brahmayamala, 254, 257–260Brahmı, 253–255Brahmı script, 79, 234Br.hatkatha, 219Br.hatkathamanjarı, 200Br.hatsam. hita, 42, 53, 54, 127Broach, 33, 34, 104, 108, 111, 116, 117,

119Buchanan, 45, 223, 224, 249Buddha, 13, 60, 102, 116, 122, 126, 127Buddhaghos.a, 205Buddharaja, 104–106, 204, 205Buddhist, 8, 10, 13, 14, 56, 87, 95, 96,

102, 116, 119, 122, 124, 131,205, 248

Budhagupta, Imp., 28, 29, 34, 57, 242,243, 245, 246

Buffalo Demon, 6, 7, 44, 45, 171

Bull, 13, 37, 69, 87, 215bull, 68, 69, 182vr.s.a, 69

Bundelkhand, 28, 67, 189, 202Burhi Gandak River, 95, 99

Cac, 65, 114–117Cac-nama, 113–118Caitra, 18, 21, 125, 127, 128cakra, 45, 69, 90, 164, 182, 235, 246,

251Cakrasoma, 150cakravaka (bird), 193cakravartin, 96, 112, 129Caks.u, 54Calukyas, 54, 104–108, 110–112, 204Cambodian, 228Camun. d. a, 254–256Can. d. asena, 71, 72Candella, 197, 200, 203, 215, 216, 218,

219Can. d. eśvara, 127Can. d. ı, 168Can. d. ikayatanakanana, 95Candrabhaga (river), 181Candragupta I, Imp., 72Candragupta II, Imp., 34, 73, 74, 104Candragupta Maurya, 72Carcika, 255, 256carrier (Kalanjara), 195, 205–209, 219Catuh. śr.ngeśvara, 151caturanga (chess), 65caturmukhalinga, 221, 222, 225, 229,

231, 239Central Asia, 31, 54Chagala, 254, 255, 259Chandar, 114Chandi Darwaza, 200, 201, 203, 209,

212, 218, 219Chatracan. d. eśvara, 149, 150Chedi Country, 213chess, see caturangaChhattisgarh, 62, 103Chilka Lake, 103China, 97, 98, 128, 129Chinese, 9, 15, 18, 52, 64, 80, 97–99,

102, 107, 110, 112, 115, 116,120, 122, 123, 125, 126, 128,129, 241

Chiruti, 97Chitor, 33, 129Citraghan. t.a, 6

304 Index

Citraratha, 3Citrarathasvamin, 3Cosmic Egg, 162, 175cremation ground, 13, 14, 78, 141, 145,

167, 173śmaśana, 167, 173, 256

Ctesiphon, 118Cunningham, General, 190, 206, 209,

215, 247, 248Cyavana, 158, 159

Dadda II, 108, 109, 111Dadhıca, 77, 79, 158–161Dadhıcatırtha, 160Daityas, 252Daks.a, 36, 160, 171, 173–175, 177, 178,

183, 185–187, 212Daks.a Cycle (SP), 187Daks.eśvara, 177, 178, 185, 186Dakshin Dinajpur District, 2, 241, 256Daks.in. a Kosala, 13, 54, 56, 59, 60, 62,

103, 104, 111, 112, 143, 200,203

daks. in. amurti, 36, 198Damodara, 36Damodaragupta, LG, 61Damodarpur, 57, 241, 243, 244Damodarpur Copperplate Inscriptions,

28, 50, 57, 241, 242,244–246, 250

Danaraśi, 149Dan. d. in, 36, 39, 54, 55, 59, 112, 255Daniell, Thomas, 223–225Darbhadhvaja, 210Darpaśata, 18, 96Daśapura, 32–36, 38, 50, 52Daśarn. a, 31, 32, 193, 196Daśaśvamedhika, 118Datta (family), 243–245Dattabhat.a, 34dauhitra, 129Debal, 114, 116–118

Daybul, 117Deccan, 2, 46, 54, 104, 108, 110Deo-Baranark, 67, 68, 131, 201Deo-Baranark Inscription, 67, 131Deogarh, 35Destruction of Heresy, 99Devagupta (Malwa), 85, 86, 91, 104Devagupta, LG, 5, 131, 132, 134Devahrada, 192devakula, 227, 228, 235, 245

devanikaya, 227Devas, 11, 18Devı, 5, 43, 44, 79, 175, 176, 212

Goddess, the, 7, 44, 142, 186,194, 256

Parvatı, 6, 37, 43, 59, 70, 127,142, 175, 177, 194, 212, 232,238, 252, 253

Devıcandragupta, 73, 74Devıkot.a, 248, 249Dhaka, 50, 261Dhanyavis.n. u, 31, 32, 247Dharasena III, 106, 108, 109, 112, 113Dharasena IV, 107, 112, 129dharma, 13, 68, 69, 96, 153, 218Dharmapala, 138, 260Dhruvasena II, 106–109, 112, 113, 123,

129digvijaya, 87, 138dıks. a, 144, 147Dinajpur District, 57, 242dınara, 227, 228, 244, 245Din.d. i, 140, 209Din.d. ika, 209Din.d. imun. d. a, 209, 214Divakara (poet), 99Dr.s.advatı, 168duh. khanta, 141Durlabhavardhana, 116dvadaśabdavrata, see Twelve-years vowDyutivarman, 227

East Bengal, 50ed. uka, 85ekamukhalinga, 191, 206, 218, 231, 232elephant’s hide, 232Ellore, 3Eran, 28, 30–32, 34, 35, 52, 247Eran Stone Pillar Inscription, 28, 30

Fayuan shulin, 122Firdawsı, 64, 66folk religion, 7Futuh. al-Buldan, 116

gada, 45, 235Gahaneśa, 214Gaja-Laks.mı, 178Gambhıra, 78Gan. akarikaratnat. ıka, 147, 148Gan. apa, 69, 142, 166, 175, 176, 209

gan. anayaka, 140Gan. apati, 12, 251

Index 305

Gan. as, 171, 175, 177Gandak River, 95, 100Gan. d. akı, 17, 95, 128Gandhara, 80Gandharas, 54Gandharayamala, 257Gan. eśa, 45, 192, 202, 208, 238Ganga, 18, 19, 57, 63, 64, 99, 124, 138,

178, 179, 181, 186, 247Ganges, 15, 18, 28, 33, 52, 64, 95,

97–99, 178, 184, 186, 193,241, 247

Gangadvara, 13, 151, 175, 177, 186,210

Gangarampur, 2, 241, 247Gangas, 108Gangasagara (tank), 206–208Ganj, 28, 29Gargya, 141, 145, 147, 193, 194Garhwal, 2Garud. a pillar, 31Gaud. a, 17, 18, 57–59, 77, 82, 84–86,

88–90, 100, 102, 110,129–131, 138, 243, 245, 247,258–260

Gaud. as, 13, 17, 53, 56, 57, 59, 60, 87,90, 91, 95, 110

Gaud. avaho, 44, 59, 78Gaurı, 125, 140, 253Gaya, 43, 47, 50, 51, 60, 62, 72, 89, 102Ghan. t.akarn. a, 2Gharo Creek, 117Ghasmara (deer), 197Ghat.otkaca, 26Ghos.avatı, 200Gıtavarman, 29, 30Goa, 2Gokarn. a, 2, 6, 138, 151Gokarn. a (North), 2Gokarn. a (South), 2Gokarn. eśvara, 2Golapani (Goila Nala), 184gold, 50, 64, 82, 99, 126, 162, 163, 186,

244, 245Golden Peak, 175Gomibhat.a, General, 226, 227, 235,

237Gopad River, 29Gopala, 260Gopalaka, 200Goparaja, 32Gopika Cave, 43

Govinda, 39Govindagupta, 34Grahavarman, 17, 48, 74, 75, 78–81,

84, 85Gr.dhrakut.a, 123, 152, 153Gr.dhrakut.eśvara, 152Gr.dhravat.a, 152Gr.dhreśvara, 152Guheśvara, 149Gujarat, 2, 34, 80, 105Gulf of Cambay, 33, 104, 108, 109, 117Gunaigarh, 50Gun. aratna, 145Gupta (nobleman), 91Gupta Era, 28, 29, 58, 60, 103Guptas, Imperial, 4, 7, 13, 15, 25–35,

40–42, 47, 50, 52, 57, 58,60–63, 73, 77, 82, 85, 89, 98,105, 110, 121, 131, 238, 242,243, 245–247, 251

Guptas, Later, 20, 42, 45–47, 50, 51,53, 58–61, 63, 72, 77, 81–83,86, 88, 89, 92, 102, 130, 133,138, 259

Guptas (Malwa), 82, 86Gurjara, 108, 111, 203Gurjaras, 108Gwalior, 31, 38, 39, 52

Haimakun. d. a, 253H. akam, al-, 116Halahala, 171Haraha Inscription, 41, 42, 47, 48, 51,

53, 55, 57, 59, 150Hardwar, 173, 175, 177, 178, 185, 186,

241Haribhadra, 175, 177, 180, 183, 185,

212Haribhadra (author), 145Harideva (devotee), 217Hariścandra Hill, 2, 195

Hariścandragarh, 2, 195Harivam. śa, 3, 15, 193, 197, 198, 246,

251Harivarman, 29, 30, 41Harivarman (Maukhari), 29, 41, 42,

47, 48, 51, 55Hars.a Era, 121, 224Hars.a ka T. ıla, 170Hars.acarita, 48, 74, 77–79, 81, 84–87,

89, 91, 92, 95, 113, 114, 155,

306 Index

156, 158, 159, 161, 162, 165,170–172

Hars.agupta, LG, 42, 51, 83Hars.agupta (Pan. d.uvam. śin), 56, 59, 63Hars.agupta, 42Hars.avardhana, 12, 17–19, 23, 46, 50,

56, 58, 59, 61, 63, 66, 74,77–82, 84–88, 90–93, 95–116,119–133, 137, 155, 158,160–162, 165, 170, 172, 200,202–205, 218, 225, 226, 239,241, 259

hell, 8–10, 227Hemacandra, 59, 253Hephtalites, 31Hetuka-Śulin, 254Hetukeśvara, 254, 259Himachal Pradesh, 52, 66Himalaya, 1, 2, 12, 37, 39, 40, 214, 253Himavat, 166, 246Hindu, 15, 44, 45, 63, 124, 173, 186,

248, 249Hindu Kush, 66Hinduism, 11, 13, 44, 122, 165, 187Hindustan, 114, 115Hiran. yakaśipu, 5, 118Hiran. yaks.a, 5, 70Hr.s.ikeśa Narayan. a Śrı Bharata

Mandir, 181–183Hsin t‘ang-shu, 129Huili, 123hum. kara, 212Hun. as, 4, 7, 12, 25, 26, 31, 32, 34, 35,

38–40, 42, 50–53, 55, 61, 62,66, 73, 74, 77, 80, 82, 113,115

Hypasis, 116

Indra, 4, 6, 11, 18, 25, 51, 106,126–128, 164, 165, 176, 203,205, 241, 253

Indra Festival, 126, 127Indra Tree, 126Indrabala, 203Indrabhat.t.arika, 57, 63Indravarman (Vis.n. ukun. d. in), 55, 56Indrotsava, 126–128Indus, 2, 80, 113, 114, 117, 118, 151,

173Indus River, 118Iran, 65, 66, 119, 123Iśana, 145

Iśanavarman, 12, 47, 48, 53–56, 59–63,67, 72, 130, 150

Iśan. ayamala, 257Iskandah, 115Iśvarakr.s.n. a, 9Iśvaravarman, 48, 51–53, 55

Jadupur, 97Jagadindu, 143, 144Jaina, 63, 96, 145Jalandhara, 113, 115, 116

Jalandhar, 113Jambumarga, 33, 36, 145Japyeśvara, 173Japyeśvaramahatmya, 173Jaunpur Inscription, 41, 47, 48, 51, 55Jayabhat.t.a III, 109Jayadeva II, 132Jayadrathayamala, 140, 149, 214Jayanaga, 89–91Jayanatha, 28Jayasena, 103Jejjat.a, 56Jetar, 115Jımutavahana, 130Jis.n. ugupta, 119, 120, 149Jıvitagupta I, LG, 58Jıvitagupta II, LG, 67, 68, 83, 131, 201Junvanı Copperplate Inscription, 143,

146, 149, 153, 214

Kabul, 114Kadamba, 41, 42, 71, 108Kadambarı, 71Kadwaha, 213Kaimur Range, 29, 222Kajangala, 97, 99Kajurgira, 97, 100, 110, 125, 241Kala, 204, 209–212, 214, 218Kalacuris, 4, 77, 86, 104–106, 203–205,

212–214, 219Kalakarn. ı, 175, 176Kalanjara, 66, 67, 133, 147, 189–198,

200, 202–206, 208–219, 221,238

Kalanjaramahatmya, 196Kalı River, 64Kalidasa, 165, 184, 205Kalinga, 54, 60, 77, 103, 111, 112Kalinjar, 67, 189, 190, 221, 241Kalyan. avarman, 71, 72Kamarupa, 20, 56, 90–92, 98, 113, 122,

123, 129, 130, 133, 173

Index 307

Kampilya, 193Kampil, 193

Kanakhala, 17, 174, 175, 177, 178, 185,186

Kankhal, 186Kanauj, 12, 29, 41, 42, 47–49, 52, 55,

56, 58, 60, 63–67, 75, 77,79–82, 84, 85, 87–92, 97–99,102, 103, 105, 108, 109, 112,113, 115, 119, 122, 123, 125,126, 130, 133, 137, 139, 164,170, 187, 190, 192, 203, 218,226, 239, 259

Kan. d. arıka, 193Kanyakubja, 12, 13, 15, 17, 18, 20, 47,

56, 66, 75, 91, 99, 104, 145,165, 170

Kanyaśrama, 160kapala, 157, 166Kapala Cycle (SP), 166, 168, 169, 173Kapalamatr.s, 171Kapalamocana, 168Kapalasthana, 13, 66Kapaleśvara, 66Kapalika, 149–153, 166, 167Kapila, 8, 9Karachi, 117Karn. a Caupar Cave, 45, 46Karn. asuvarn. a, 57, 90, 97, 103, 259Karohan. a, 2, 140, 151, 214, 215Karttikeya, 15, 16, 19, 59, 75, 125,

192, 238Karukas, 150Kashmir, 2, 113, 114, 116Kashmir Smast, 79Kaśmıra, 6Kathasaritsagara, 200Kathmandu, 2, 119, 140, 258Katyayanı, 43, 44Kauberı, 253Kaumarı, 253–255Kaumudımahotsava, 14, 71, 72Kaun. d. inya, 140, 145–148, 176Kaurus.a, 145, 170Kauśambı, 189, 200, 201, 205Kauśika, 140, 145, 149, 150, 193, 194Kauśika Suparvan, 193Kauśikı, 6, 7, 11, 56, 252Kauva-Dol, 44Kaverı River, 41Kavyadarśa, 59kayastha, 244

Kedara, 12, 138, 151Kedarakhan. d. a, 181, 183Ken (river), 191Kerala, 71Kevala Narasim. ha Temple Inscription,

26khad. uka, 150Khambhor Shiva, 208Kharagraha I, 105–107, 112khat.vanga, 152, 207, 229, 230Khoh, 29Khusrau Anushirvan, 64, 65Kiratarjunıya, 36, 37Kırti, 14Kirtipur, 255Kırtivarman, 216Kokamukha, 245Kokamukhasvamin, 245Kokamukha (tırtha), 245Kola, Mt, 6Kon. adevı, 131Kongoda, 97, 103, 110Kot.avı (goddess), 251Kot.itırtha, 191, 202Kot.ivars.a, 2, 4, 18, 241, 243–249,

251–254, 256, 258–260, 266Kot.ıvars.a, 2–4, 6, 173, 248, 252, 255,

259, 266, 268, 269Kot. ıvars.amahatmya, 2–4, 241, 252,

254–256, 258, 259, 261, 263Kot.ivars.a vis.aya, 57, 242, 243,

246–248, 260Koukou-nor, 129Kraunca, 241Krodheśvara, 210–212, 218, 219Kr.s.n. a, 15, 43, 67, 90, 251, 253Kr.s.n. agupta, LG, 42, 83Kr.s.n. araja, 204Ks.atravarman, 74ks.atriya, 42, 168Ks.emeśvara, 70Ks.upa, 160Kubjamraka, 181–184, 186Kubjamrakamahatmya, 182kulapati, 226, 227kulaputraka, 245, 246kulika, 243kulyavapa, 244, 245kumara, 17, 87, 162, 163Kumara (Skanda), 164, 165Kumaragupta I, Imp., 26, 27, 34, 67,

242, 243, 245

308 Index

Kumaragupta II, Imp., 27, 50, 58, 244Kumaragupta III, Imp. (spurious), 27,

244Kumaragupta, Imp., 57, 242Kumaragupta, LG, 58, 60, 61, 81Kumaragupta (Malwa), 80, 81, 86Kumarasam. bhava, 163, 165Kumaravarman, 33, 54Kumarayamala, 257Kumbh Mela, 124Kumbhakarn. a, 254, 255, 259Kuntala, 86Kurmapuran. a, 209, 215Kurnool District, 2, 142Kuru Country, 145Kuruks.etra, 1, 2, 13, 17, 78, 139, 145,

155, 157, 161, 163, 165,168–171, 173, 174, 178, 184,193, 241

Kuruks.etramahatmya, 2, 169, 173Kus.an. a, 79, 170, 181, 186, 255Kuśika, 145, 146, 150Kuśinagara, 89Kut.ukka Country, 133

Laghuśam. vara, 258lagud. a, 147, 148Laks.man. araja, 203–205Laks.mı, 6, 14, 25, 61, 79, 88, 96, 106,

114, 165Laks.mıdhara, 127, 175, 177, 181, 182Lakud. as, 140, 147, 196, 198, 218

Lakulas, 147, 148, 150, 196Lakulıśa, 9, 20, 133, 137, 140, 145,

146, 148, 192, 198, 199, 202,214–216, 218, 237

Lagud. i, 2, 133, 145–147, 170, 196,214, 218

Lagud. ıśvara, 133, 137Lakulıśanatha, 144, 153Nakulıśa, 145

Lat.a, 80, 108Lauhitya River, 40, 81laukika, 4, 14, 147, 151, 175, 214Law (dharma), 69Li Yibiao, 122, 128Liang Huaijing, 122Licchavis, 4, 20, 71, 72, 119–121, 132,

134, 138, 149Lingapuran. a, 140Lohita, 253Lohityasindhu, 254

lokatıtavratin, 150Lokodadhi, 147, 171Lomas Rishi Cave, 43, 45, 47Lucknow, 189

Madanapala, 247Madhavagupta, LG, 81, 82, 91, 92,

102, 130Madhavagupta (Malwa), 81, 86, 97Madhavaraja II, 103Madhavavarman II, 3, 55Madhukara, 19, 156Madhya Pradesh, 189, 213Madhyadeśa, 12, 63, 77, 84, 95, 254Madhyama, 12, 151Madhyameśvara, 13Madhyamika, 33, 36Madras, 42Magadha, 13, 14, 17, 20, 21, 41, 42,

46, 47, 50, 51, 53, 56–63, 71,72, 80–83, 86, 89, 92,100–102, 113, 122, 128, 130,132–134, 138, 151, 152, 243,245, 259, 260

magadhadhipatya, 59Magas, 67Mahabharata, 3, 15, 19, 42, 69, 74,

125, 152, 160, 161, 163, 165,168, 172, 192, 210, 221, 222

Mahabodhi Monastery, 123Mahacına, 122Mahakala, 14, 213Mahakalahr.daya, 78Mahakaleśvara, 151Mahakalı, 253Mahakapala, 167–171Mahakaya, 66, 167Mahaks.atrapa, 108Mahalaya, 2, 12, 151Mahaman. d. ala, 259Mahamanthana, 259Mahapadma Hell, 8Mahapaśupatas, 150Maharajaputra (imp. Gupta), 242,

243, 245Maharashtra, 62, 110, 111Mahasamanta (Nepal), 120–122Mahasenagupta, LG, 81, 82, 84, 86,

89–91Mahasenagupta, 84, 90Mahaśivagupta, 56, 59, 103, 113, 143,

144, 148, 150, 202, 214

Index 309

Mahaśmaśana, 78Mahasthan, 247Mahasthangarh, 247Mahatmya, 2, 4, 10, 16, 160, 161, 168,

173, 182, 183, 221, 241Mahavrata, 143, 144, 152, 153Mahavratin, 152, 153Mahayajnas (5), 245Mahendra Mountains, 40Mahendrı, 253, 255Maheśvara, 4, 9, 14–16, 18–20, 74, 78,

147, 151Maheśvara yoga, 141, 255Maheśvarapura, 34Maheśvarı, 255Mahıpala I, 247Mahis.asuramardinı, 6, 7, 43–45Mahis.matı, 104, 105, 204, 213Maitrakas, 4, 105–109, 112, 123, 129Makran, 118, 119Malava, 29, 42, 80–82, 84–86, 108

Malwa, 26, 31, 33, 50, 52, 62,80–82, 86, 106, 129, 193,213, 219

Malavas, 42, 86, 108Malavı, 42Mallakut.a, 95Mallar, 56Malwa, see MalavaManadeva II, 121Manadeva Era, 121Mandakinı River, 256Man. d. aleśvara, 221, 222, 225, 228,

237–239, 241Man. d. aleśvara Temple, 228, 239Man. d. aleśvaramahatmya, 221, 236, 238Man. d. aleśvarasvamin, 221, 227, 228,

231Man. d. aleśvarasvamin Temple, 226–228Mandara, 60, 183Mandaragiri, 131

Mandar Hill, 131Mandasor, 32–36, 38, 54, 82, 149Mandasor Stone Inscription, 33, 36,

39, 46, 52Mangaleśa, 105, 106, 204Mangraon, 133Manichaeism, 119Man. itara, 95, 100Mankan. aka, 168, 174mankha, 74Mantramarga, 4, 9, 10, 258

Manu, 68, 168, 184Martindell, Colonel, 190mat.ha, 13, 131, 187, 227, 235, 237, 238Mathura, 31, 33, 104, 145Mathura Pillar Inscription, 145, 146Matidatta, 242Matr. (Temple), 78, 171, 259Matr.ceta, 38Matr.tantras, 4, 7, 257, 258Matr.vis.n. u, 31, 32, 247Matsyapuran. a, 54, 79, 181Mattamayura, 213Maukharis, 4, 12, 13, 16, 17, 20, 29,

30, 35, 40–43, 46–53, 55, 56,59–61, 63, 65–69, 71–74, 77,78, 81, 82, 84, 85, 88, 91, 92,101, 102, 130, 132, 133, 137,158, 192, 199, 218, 226, 238,247

Mauryas (Chitor), 129Mauryas (Konkan), 108Mausala, 144, 148, 214Mayuraśarman, 41, 71Meghaduta, 184, 205Mehran (Mihran), 114Mekala, 56Mihirakula, 35, 37–39, 51–53, 67, 113Mithra, 67Mitra, 145Mitra (family), 244Mleccha, 7, 16, 25, 26, 73, 74, 96, 101,

119, 259Moon (Śaśanka), 17, 87, 93, 101, 162Mother Goddess, 171, 256, 259

matr. , 251Mothers, 6, 171, 241, 253–257

Mr.gadhara, 194–198, 205, 206, 218Mr.s.ika, 254Mr.tyu, 176Muh.ammad ↪Alı ibn H. amid ibn Abı

Bakr Kufı, 113Mudraraks.asa, 14, 72, 73, 75Mugalisa, 144Mughıra, al-, 114, 116, 117Muh.ammad (prophet), 116, 118Mukhara, 29, 41, 42Multan, 119Mun.d. a (king), 85mun. d. aśr. nkhalika, 149, 150Mun.d. eśvarı (goddess), 222Mun.d. eśvarı Hill, 221–223, 228, 229,

232, 235–239, 241

310 Index

Mun. d. eśvarı Inscription, 224, 226, 227Mun. d. eśvarı Temple, 222, 225, 230, 231Musala, 144, 146

Nabha State, 182Nachna, 6, 7, 28nad. ı, 141Nadin (deer), 197Naga (deity), 78, 201, 239Nagananda, 127, 130nagaraśres.t.hin, 243, 244, 246Nagarı, 33, 36, 37Nagarjuni Hill, 43, 44, 46, 47Nagas, 25Nahapana, 108Nahavand, 118Naigames.in, 2Naimis.a, 19, 124, 125Nalanda, 13, 27, 50, 58, 63, 85, 98, 99,

102, 130Nandin, 238Nandin Cycle (SP), 173Nandin (family), 244Nandıśvara, 140, 210Nandiyamala, 257Nannaraja, 202Nara, 167Narasim. hagupta, Imp., 27, 58, 67Narayan. a Temple, 226, 228, 235–237Narayan. apala, 260Narendradeva (Licchavi), 119, 120, 149Narendragupta, 89Narmada, 28, 110–112, 210, 211Navagaon, 106Nayapala, 254, 256Nepal, 2, 14, 20, 21, 119–122, 129,

132–134, 138, 149, 245, 246,255, 260, 261

Nepala, 21, 133, 138New History of the Tang, 122New Year, 18, 125–127New Year Festival, 126Nıladhara (river), 186Nılakan. t.ha, 203, 212, 216Nılakan. t.ha (commentator), 209Nılakan. t.ha Temple, 215–218Nılalohita, 161, 162, 165–171, 173Nirgranthas, 8Nirmand, 66nirukti, 161, 185, 209, 211, 212, 214,

215, 218Nisunda, 221

Niśvasaguhya, 9Guhyasutra, 213

Niśvasamukha, 9Niśvasamula, 9Niśvasatattvasam. hita, 9, 148, 150, 213Nityavitrasta (deer), 197nivartana, 183Niyatakalakan. d. a, 175Nohala, 213

OM, 161, 167Oman, 116Orissa, 20, 54, 56, 60, 77, 97, 100, 102,

103, 110, 122, 237Oxus, 54

Pad. ana Hill, 146, 147Padmapuran. a, 70Padmavatı, 200Painted Grey Ware, 63, 79, 185Pala, 138, 250, 252, 254, 255, 260, 261Pala (family), 243, 244, 246, 260Pancaks.a, 175Pancala, 193pancamahaśabda, 207Pancanada, 173Pancaratra, 142Pancartha, 2, 133, 139–141, 146–149Pancaśikha, 4, 8, 9pancas. t.aka, 10, 181, 213Pancaśvamedhika, 118Pan. d. ava, 143, 144, 201, 203Pan.d. ita (deer), 197Pan. d. uvam. śin, 4, 56, 59, 60, 62, 63,

103, 202, 203Panjab, 35, 52, 113, 173paramabhagavata, 5, 50, 131paramadaivata, 57, 242paramamaheśvara, 5, 18, 102, 106, 124,

131, 132, 204, 254Paramartha, 9param. para, 143–146, 194

guruparam. para, 145Paraśara, 96Paraśıka, 7, 114Paraśurama, 70, 168Paravata, 19, 156parivrajaka, 205Parn. adatta, 25Paśupata, 2–4, 8, 9, 13–17, 20, 34, 36,

56, 78, 96, 120, 132–134,137–140, 142–153, 166, 167,

Index 311

170, 171, 173–177, 186, 187,194–196, 204, 205, 209,212–216, 218, 219, 226, 237,238, 257, 258

Paśupatasutra, 140, 142, 145, 148, 176Paśupata vrata, 36, 151, 153, 174, 176,

187Paśupata weapon, 36, 165Paśupata yoga, 4, 8, 18, 141, 142Paśupati, 13, 36, 38, 39, 78, 120–122,

138, 146, 149, 176, 204Paśupatinatha, 14, 20, 21, 120–122,

133Paśupati Temple Inscription, 132Pat.aliputra, 71, 72, 89, 138Patna, 63, 95, 222Patna Museum, 237–239Paun.d. ra, 57, 133, 156Peddavegi, 3Pehelejpur, 32Persia, 65, 96Persian, 7, 64, 65, 113, 114, 118, 119Picumata, 258, 259Pin. d. araka, 222Pin. d. areśvara (Mun. d. eśvarı Hill), 222Piśaca, 171Piśaca-deva, 171Piśacıs, 171Pis.t.apura, 112pitr.kalpa, 193Pon. n. i, 41potlatch, 20, 124, 126, 128Prabha, 6Prabhakara (gan. a), 69Prabhakara (Mandasor), 34Prabhakaravardhana, 79–81, 84–86Prabhasatırtha, 140Prabhavatı Gupta, 46pradaks. in. a, 183, 221, 239pradaks. in. apatha, 182, 225Pradyumna, 251Pragjyotis.a, 92, 254Prahasiteśvara, 138Prahlada, 4, 8, 70, 252Prajapati, 166, 172, 246Prakaśadharman, 33–35Prakat.aditya, 67Praman. as, 148Pran. a, 150Pranardanapran. a, 149, 150Pran. aveśvara Temple (Tal.agunda),

213

Pratihara, 66, 122, 164, 192, 203Pravaragiri, 43, 45Pravarapura, 26Pravarasena II, 26Prayaga, 13, 18, 19, 61, 109, 112,

123–129Prithivıvigraha-bhat.t.araka, 60, 77Prıtikut.a, 88, 95, 156, 158Priyadarśika, 18, 125, 126Priyangu, 232, 233Pr.thivıs.en. a II, 28Pulakeśin II, 106–112Punarbhava River, 247Pun.d. ra, 18–20, 101Pun.d. ravardhana, 57, 101, 242, 243,

245, 247, 260Pune, 2Puran. a, the, 3, 15, 19, 125, 157Puran. apancalaks.an. a, 3, 70, 125, 157Puri, 103Puriścandra, 2Purn. avarman, 92, 101–103, 113purohita, 162Purugupta, Imp., 27, 28, 50, 67Pus.kara, 145pustaka, 18, 156pustakavacaka, 19, 156pustapala, 244Pus.yabhuti, 78–80, 84, 85, 170Pus.yabhutis, 13, 66, 77, 78, 90–92,

105, 161, 186

Qannuj (Kanauj), 64–66quinquennial assembly, 18, 112, 123,

125, 126, 128

Raghuvam. śa, 228Raibhya, 175, 177, 182Raibhyaśrama, 175, 177, 181Raja Aman Singh (palace/museum),

191, 198Rajabar.ı Mound, 248, 249Rajagr.ha, 122Rajaśekharasuri, 145Rajasthan, 33, 41Rajasuya sacrifice, 87Rajbarid. anga, 97Rajibpur, 247Rajmahal, 97Rajyaśrı, 74, 78–81, 84, 87, 91, 99,

158, 165Rajyavardhana, 74, 81, 85, 86, 91, 100,

104, 113, 165

312 Index

Rajyavardhana (Aulikara), 33Ramagiri, 31Ramahrada, 168Rama-Jamadagnya, 70, 168Ramayan. a, 15Rambha (river), 178, 179Ramgarh, 222raśi, 96, 139, 145, 149, 150, 175, 176,

237Raśıkara, 139, 145Ratnavalı, 18, 19, 103, 104, 125, 126,

200, 203Rauravasutrasam. graha, 213Ravikırti, 33Raviśanti, 70Reva River, 111Revatı, 6Rewa, 47Ribhupala, 244–246Rigveda, 126Rishikesh, 173, 178, 181, 183, 184, 186,

210, 238Rishtal Inscription, 34, 35Rohitagiri, 88, 237Rohri, 114Rohtasgarh, 88, 89, 95, 222, 237Rudra, 148–150, 166, 173, 176, 185,

209rudraks.a, 182, 207Rudraks.etra, 181, 183Rudraloka, 210Rudran. ı, 253, 254Rudras (38), 214Rudras (66), 149, 214Rudrasoma, 144rudraśr. nkhalika, 149, 150Ruman. vat, 103, 104, 200, 209

Sachankot, 191Sadanavarman, 30Sadaśiva, 213S. ad. darśanasamuccaya, 145sadhaka, 139, 140, 145, 147Sa↪d ibn Abı Waqqas., 118Sahasi II, 114Sahasi Rai, 114Sahiras, 114Śailendra, 176Śailodbhavas, 103Śaiva Age, 5, 37Śaiva Siddhanta, 9, 144, 148, 151, 213,

216, 218, 254

Saivism, passimŚaka Era, 120Śakala, 37, 113Śakas, 74Saketa, 245Śakradhvajotsava, 126, 127Śakta, 150, 247, 260śakti, 7Śalankayana, 3Samacaradeva, 60, 89Samah, 117samanta, 47, 109, 207, 209, 219Samantapancaka, 163, 168Śam. kara (the philosopher), 182Śam. karacarya Aśrama, 181Śam. karagan. a, 86, 104, 204, 205Sam. khya, 4, 8, 9, 141Sam. khyasiddhanta, 8Sam. nihitı, 160, 168Sam. skaravidhi, 147Samudragupta, Imp., 110Samudrasena, 66Sanatkumara, 253Sanchi, 205Sanjeli, 34Sanketa, 89Śankukarn. a, 2Śankukarn. eśvara, 2, 118, 151, 173Śankukarn. eśvaramahatmya, 2, 173Sanskrit, 7, 11, 14–17, 19, 20, 44, 59,

85, 124, 127, 151, 163, 165,169, 172, 174, 193, 226, 235,251

śanti, 162Saptamatr.kas, 208, 253–255Saptasarasvata, 168, 174Sarasvata, 156, 158, 160Sarasvatı, 14, 19, 156–158, 160, 161,

167, 169Sarasvatı (river), 17, 77, 84, 87, 155,

157, 158, 160, 161, 163, 164,168, 169, 173

Sarasvatımahatmya, 157, 166, 169, 173Sarasvatıyamala, 257Sarayu, 247Śardulavarman, 43, 47Saridvıpa, 193Sarnath Stone Inscription, 27, 28Saromahatmya, 168, 169sarthavaha, 243, 247Śarva, 164Śarvan. ı, 253, 255

Index 313

Śarvasvaminı, 29Śarvavarman, 42, 55, 56, 59, 62, 63,

65–69, 79, 137, 192, 218Śarvavarman (Maharaja), 66Sarwan Baba, 209Sasanid, 118, 119Śaśanka, 4, 17, 82, 87–93, 97, 100–103,

105, 110, 130, 162, 172, 259S. as.t.hı, 6Śatadru, 113Satavahana, 108satı, 32, 84Saundaryalaharı, 209Saurashtra, 25, 105, 140Savitrı, 42senapati, 18, 87, 161–163

mahasenapati, 41Senapati, see SkandaSeven Brahmins, the, 3, 17, 189, 193,

194, 196, 197, 218seven deer, the, 193, 196–198, 210,

218, 221Shahname, 64–66Shankarpur, 29Shankarpur Copperplate Inscription,

29, 30, 42shell script, 234Shib Bari, 250, 255, 260Shyampur Garhi, 184–186Sialkot, 37, 113, 116siddha, 139, 211Siddhaks.etra, 167, 213Siddhas, 11, 133siddhasthana, 13, 66Siddheśvara Temple, 45siddhi, 161Śıladitya I, 105–107Sim. hanada, 88Simla, 66Sindh, 65, 66, 113–117Sindhu, 80, 115, 118Sindhudeśa, 114Singer of Ancient Lore, 19, 124, 125Sirpur, 103, 143, 144, 148, 150, 225Sıtasej, 196, 197, 208site A (Mun.d. eśvarı), 231–233, 236site B (Mun. d. eśvarı), 234Śiva, 2, 6, 7, 9, 11, 13, 15, 16, 18, 35,

37, 43, 47, 50, 59, 60, 69, 70,79, 118, 121, 125, 127,139–142, 149, 150, 152, 153,155, 157, 160, 161, 164–170,

174–178, 183, 185–187, 202,206, 208–210, 212–214, 216,218, 221, 229, 239, 251–257,259, 260

Bhava, 121, 151, 170, 253Hara, 183, 211Mahadeva, 11, 45, 50, 139, 151,

165, 166, 168Maheśvara, 11, 46, 124, 139, 153,

170Śam. kara, 210, 241, 253Umapati, 210Vr.s.abhadhvaja, 35

Vr.s.adhvaja, 200, 253Śivadeva I, 121Śivadeva II, 20, 132, 134Śivadeveśvara Bhat.t.araka, 132Śivadharma, 9, 10Śivadharmasam. graha, 9Śivadharmaśastra, 213Śivadharmottara, 9Śivapura, 3, 4, 151, 175Śivas (28), 214Śivatad. aga, 167, 168śivayoga, 139, 141Skanda, 6, 8, 11, 17, 19, 59, 75, 125,

161–165, 229, 230, 238, 241,245, 251, 253

Senapati, 11Skanda Cycle (SP), 69Skandagupta (general), 74Skandagupta, Imp., 25–28, 31, 34, 244Skandapala, 242Skandapuran. a, passim

RA recension, 127, 137, 174, 211,212, 215, 252, 259–261, 263,265, 268

hyparchetype of the RArecension, 137, 138, 252, 254

S recensionhyparchetype of the S

recension, 134, 137, 138, 151,152, 260

Skandayamala, 257skull, 14, 78, 149, 150, 152, 153, 157,

166–169, 171, 254śmaśana, see cremation groundSod. hadeva, 203Soma, 140Soma Siddhanta, 144, 149, 214Somanandin, 194Somanatha Temple, 140

314 Index

Somas, 148, 150Somaśarman, 140–146, 148, 214Someśa, 140, 214Somnath, 140Son (river), 17, 29, 42, 47, 48, 50, 68,

72, 77, 88, 95, 100, 156, 158,161, 222, 238

Śon. itapura, 248, 251, 252Spring Festival, 18, 125, 128Śraman. as, 18, 123, 124, 126, 209Śrı Bhadra (anupavis.aya), 178Śrı-Hars.a, 18f., 126f., 130, 200, 203Śrıkan. t.ha, 78Śrıkan. t.ha (Śiva), 155Śrıkan. t.ha Naga, 78Śrımatı, 82, 83, 131Śrıparvata, 2, 3, 142, 151, 195

Śrıśaila, 2, 142Śrıparvatamahatmya, 142Śrıpura, 56, 103, 112, 113, 150, 202,

203Śr.nkhalika, 120, 149, 150Srong-brtsan sgam-po, 120, 129Stabdhakarn. a (deer), 197sthana, 13, 160, 186sthanaguru, 144Sthaneśvara, 12, 17, 66, 77, 79, 87, 90,

91, 158, 160, 161, 163, 171,173

Sthaneśvaramahatmya, 160, 173Sthan. u, 39, 161, 166Sthan. udatta, 242Sthan. utırtha, 160, 161, 168, 169, 173Sthan. utırthamahatmya, 173Sthan. uvat.a, 160Sthan. vıśvara, 75, 77–79, 86, 163, 170Subhadreśvara Temple, 133Sucıvan. a, 172Sudr.s.t.i, 156suicide, 61, 141, 184, 193, 195, 210,

218, 219Suks.maśiva, 58, 131śula, 54, 170, 175, 254Śulakun. d. a, 254Śulapan. i, 37, 54Śuleśvara, 66, 167Śulikas, 53–55Sulikas, 53, 54Sulisa, 66Śulodaka, 259Sumeru, 37Sunda, 221

Śunga period, 249, 250, 252Supratis.t.hita, 90, 91suputra, 47Śurasena, 132Suraśmicandra, 28, 30, 32, 34Surjangarh, 45Surjangiri, 45Surya, 42, 45, 46, 67, 68, 131, 246Suryavam. śa, 42, 78Suryavarman, 14, 47, 48, 59, 60, 69–71Susthitavarman, 81, 90Suta, 16, 18–20Sutlej River, 66, 113Sutradhara, 18, 126Suva. . . (Maukhari), 80, 85, 102svasiddhanta, 133, 139, 146Svayambhudeva, 242Svayambhuvayamala, 257Śveta, 189, 204, 209–212, 214, 215,

218, 219Śvetaganga, 211Śvetapada, 204, 214Śvetaran. ya, 212Śvetavarahasvamin, 245, 246, 251Śveteśvara, 211

Taittirıya Sam. hita, 126Taizong, 99, 120, 128T. akka, 113, 115Taks.akeśvara, 2Taleśvara, 227Tana, 116Tang Dynasty, 97, 98, 120, 122, 128Tantrasadbhava, 258Tantric, 3, 7, 10, 13, 78, 150, 213, 214,

219, 257–259tapas, 11, 32, 118, 142, 161, 165, 211,

253tapovana, 144, 234tapovanika, 228, 234Tarad. am. śaka, 56Taraka, 18, 162, 164, 165Taranatha, 119, 260Tarıkh al-Hind wa-l-Sind, 114Tartars, 115, 116Tejasoma, 144Temple Council, 226–228terracotta (figurine), 63, 65, 250Thane, 116Thanesar, 17, 74, 75, 77, 79–81, 84, 86,

87, 89, 99, 113, 155, 160,168, 170, 186, 259

Index 315

Tibet, 20, 119, 120, 129Tilottama, 221, 222, 239, 241tırtha, 5, 20, 118, 168, 177, 192, 245tırthayatra, 4Tıvaradeva, 56Toraman. a, 30–32, 34, 35, 38, 50, 52,

247Trata, Mahasattrapati, 227Trikan. d. aśes.a, 248Tripurı, 204, 213, 214triśula, 190Triśuladitya, 195Trivikrama, 70Tunga (dynasty), 237T‘ung-tien, 129Turus.kas, 119Twelve-years vow, 152, 153

Uccakalpa, 29ucchraya, 9Udakasevika-Bhairava Festival, 127Udayadeva (Licchavi), 119Udayagiri, 31, 35, 255Udayana, 103, 189, 200–203, 205, 209,

219Udayasena (Mahasamanta), 224–226,

239Udbhavaraśi, 56, 150Ujjain, 14, 33, 86, 104–106, 145, 146,

151, 203–205, 209, 219↪Uman, 116↪Umar, 116Umasoma, 150Umavana, 248, 253Unmukha (deer), 197Upagupta, 57, 242Upagupta, 51Upaka, 7Upendra, 106↪Uthman, 116Utkala, 56, 103utkranti, 141, 193, 195utsada, 9, 10Uttar Pradesh, 27, 28, 50, 164, 189Uttarakhand, 1, 227

Vr.s.asarasam. graha, 9Vad. at.hika Cave, 43vahana, 69, 208Vainyagupta, Imp., 50Vainyagupta, Maharaja, 50Vais.n. ava, 4, 5, 35, 69, 70, 160, 174,

183, 186, 228, 245, 251, 260

Vais.n. avı, 253–255Vaisnavism, 4, 5, 247, 251vaitanikavrata, 139, 145Vaivasvata, 42Vajasaneya, 246Vakat.akas, 26, 28, 110Vakpati, 59, 78Valabhı, 105, 107–109, 112, 113, 140Valaka, 66Vamana, 70Vamanapuran. a, 168, 177Vamaraja, 204Vanakhan. d. eśvara, 206Vanamalavarman, 254Vanga, 60Varaha, 31, 32, 45, 182, 245–247Varahapuran. a, 183Varahı, 250, 251, 253–255Varan. ası, 1, 2, 6, 11–17, 20, 89, 133,

137–140, 151, 173, 174, 221,222, 238, 241, 251

Varan. asımahatmya, 5, 11, 12, 173, 186Vardhanas, 4, 13, 17, 66, 79, 92, 186,

259varn. aśrama, 149Varun. avasin, 67Varun. ı, 253Varun. ika, 68vasaka, 86Vasanta, Minister/Vassal, 205–209,

219Vasantaka, 208, 209, 219Vasantotsava, 125Vasat.a, 56, 59, 63Vasavadatta, 200, 205, 219Vasis.t.ha, 160Vasudeva (guru), 216Vatapi, 110Vatsa, 156, 159Vatsa (region), 103, 200, 203, 205, 219Vatsadevı, 20, 132, 134Vatsagulma, 6Vatsyayanas, 77, 159Vayası, 7Vayavya, 253Vayu, 157, 161, 166, 172Vayupuran. a, 3, 140, 152, 157, 172, 214Vayuraks.ita, 34VBA-I, 178, 181VBA-II, 178, 180, 181Veda, 11, 166Vengıpura, 3

316 Index

vetala, 78Vidarbha, 54Vidiśa, 26, 28, 31, 34, 77, 82, 86, 104,

105, 147, 193, 196, 213, 218vidus.aka, 209, 219Vidyapıt.ha, 258viham. gika, 205, 209vihara, 102Vijayadevavarman, 3Vijayavarman, 29Vijjika, 71Vikramendravarman II, 56Vilocana (deer), 197◦vimala, 146Vinayaka, 8Vindhya, 1, 6, 7, 43, 44, 47, 52, 56, 88,

91, 103, 111, 189, 200, 210,211, 221, 222

Vindhyavasinı, 6, 7, 11, 44, 56, 252Vindhyavasinı Cycle (SP), 6, 17, 197,

218, 221Vinıta (ascetic), 237Vinıta (ruler), 237Vinıtaraśi, 149Vinıteśvara, 228, 231, 232, 237Vinıteśvaramat.ha, 226–228, 232, 234Vipaśa, 113Vipr.thu, 210Vırabhadra, 208, 212Virabhadra (VBA), 178, 185, 186Vırabhadreśvaramahadeva, 179Viśakhadatta, 14, 71–74vis.ayapati, 34, 243Vis.n. u, 13, 15, 25, 31, 45, 46, 70, 114,

131, 160, 164, 167, 176, 182,183, 202, 210, 215, 235, 236,246, 247, 253, 257

Janardana, 31Narayan. a, 31

Vis.n. ugupta, Imp., 47, 50, 57–59,242–244, 247

Vis.n. ugupta, LG, 5, 131–134Vis.n. ugupta (Nepal), 120Vis.n. uks.etra, 183, 186Vis.n. ukun. d. in, 2, 3, 55, 56, 62, 63Vis.n. upuran. a, 15, 251Vis.n. usmr.ti, 152Vis.n. uvardhana, 36Viśvamitra, 160vratyas, 153Vr.ddhapramatr., 254Vr.ka, 245

Vr.ndavana, 140vr.s.a, see BullVyaghradeva, 28, 29Vyasa, 4, 125, 241, 253

Wang Xuance, 122, 128, 129Warrior Goddess, 11Wencheng, 120West Bengal, 2, 28, 57, 77, 241, 242,

247

Xin Tang shu, 122Xuanzang, 15, 33, 63, 64, 80, 87, 92,

97–103, 106, 107, 109–114,116, 119, 120, 122–128, 241,247

Yajna, 69, 166Yajnavaraha, 246Yaks.ı, 181Yama, 42, 210, 212Yamagarta-man. d. ala, 237Yamalas (7), 257, 259Yamuna, 18, 19, 33, 52, 124, 202Yamya, 253Yaśavanta Sim. ha Bahadur, 182Yaśodharman, 33, 35–40, 46, 52–55Yaśovarman (Candella), 203Yaśovarman (Kanauj), 44, 59Yaśovatı, 84Yas.t.igr.haka, 95Yaugandharayan. a, 200Yavanas, 7Yazdagird III, 118yoga, 142, 146, 257yogadharma, 194yogasiddha, 142yogasiddhi, 141Yogeśvarıs, 254, 257Yoginıs, 258, 259yogipratyaks.a, 141Yuko Yokochi, 69, 137, 193, 204, 215,

241, 252, 263yupa, 41Yuvarajadeva I, 213

Zoroastrian, 67, 119