Mxlras Uolverrfty ffiatorical Series-No. 16
"A *
GENERAL EDITOR :
PROFESSOR K. A. NILAKANTA SASTRI
EARLY HISTORY OF THE ANDHRA COUNTRY
THE
SATAVAHANA EMPIRE
CM
CSl
*o
N
w w eo
O H OJ CO
r-t H r1 rH
THeais approved for the Degree of Doctor of Philosophy
PRINTED BY G. 0* fMBftft; MAMtU,
COOTENTS
PAGES
Foreword . . ix
Preface . . xi
Chapter I.
Introductory .. 1-3
Chapter IL ORIGINS OF SATAVAHANA POWER
Materials for a study of Satavahana history The old
theory Epigraphic, numismatic and literary evidence
against it 4- 27
Chapter III. THE EARLY SATAVAHANAS
The Starting point Founder of the dynasty : Simuka-
Kanha I Siri-Satakani I Satakarni : Meaning Sata-
kani's Imperial Position Satakani I Kharavela Syn-
chronism? Hakusiri Satisiri Skandastambhi Sata-
kani II Apflaka His reign: Extent of his kingdom
Hala Hala : meaning of the name Events of his
reign . . 28-43
Chapter IV. THE LATER SATAVAHANAS
(a) Gotamlputa Siri-Satakani (i) Metronymics (ii)
Gotamiputa Siri-Satakani and the Ksaharatas Ksaharata
inroads into Satavahana dominions Bhumaka Naha-
pana (iii) Comparative Chronology of Gotamlputa
Satakani and Nahapana The Satakarni of the Girnar
inscription (iv) Gotamiputa Satakani cannot be
identified with Gotamlputa Vilivayakura of the coins
(v) Gotamlputa's empire (b) Pulumavi II Messrs
R. G. and D. R. Bhandarkars' theory of the conjoint
rule of Gotamiputa Siri-Satakani and Pulumavi His
dominions (c) Siva Siri-Satakani (d) Sivamaka Sada
(Satakani) (e) Matfhariputa Sakasena Satakani (f)
Gotamiputa Siri-Yana Satakani (g) Vasithlputa Cada
Sfiti (h) Pulumavi m ..44-72
Chapter V. ADMINISTRATIVE ARRANGEMENTS IN THE
SATAVAHANA EMPIRE
Royal Succession The King Feudatories Nature of
the titles: MahSbhoja Mahabhoji Mahara{his Mahfi-
senapatis Officials and administrative divisions Appen-
dix A : The Alluru inscription . . 73-90
vi
CONTENTS
PAGES
Chapter VI. SOCIAL, ECONOMIC AND RELIGIOUS
CONDITIONS
I. Social conditions Foreign elements in Hindu and Bud-
dhist Society Yavanas and Sakas in Eastern Deccan
Family ^omen Luxuries Names Place Names
Military arrangements II. Economic conditions :
Foreign trade: Seaports: Eastern Ports Western Porte
Market-towns in the interior : Western Deccan
Eastern Deccan Merchants Coins and trade Indus-
trial and Commercial organisation Land III. Religion:
Buddhism in Eastern Deccan Western Deccan Sects :
Buddhist Sects in Western Deccan Eastern Deccan
Monks and Nuns Brahmanical Religion
Chapter VIL THE IK^VAKUS THIRD CENTURY A.D.
The Puranic label for the Iksvaku kings Duration of
the dynasty Home of the Iksvakus Rise of the
Iksvakus The founder of the line Siri Camtamula
"The Unosbtructed ' His reign His personality, sisters
and queens Conclusion Virapurisadata Alliance with
the powerful house of Castana Other queens Alliance
with the Cu^us The Buddhist monuments of his reign
Ehuvula Camtamula Buddhist monuments of his reign
Buddhism of the period Ceylon and Nagarjuriikon^a
Buddhist Canonical Books Mentioned ^Administration :
Administrative divisions Officials Official Titles: MahS-
senapati Mahatalavara Mahadan^anayaka Other con-
clusions
Chapter VHI. KINGS OF THE B^HATPHALAYANA GOTRA
Chronology The rise of the dynasty The capital of
Jayavarman Administrative organisation
Siapter IX. THE VAINGEYAKAS
Salankayana A Gotra and not a Dynastic Name
Salankayana not the name of a tribe Salankayana
of the inscriptions has nothing to do with the
bull banner of the Vaingeyakas The Dynastic Name
adopted here is 'Vaihgeyaka* Origins of the Dynasty
Devavarman : He was not the first king of the line
But was. predecessor of Hastivarman Date of Deva-
varman Events of his reign Extent of his kingdom
Importance of Devavarman's Grant Hastivarman
Nandivarman I: A misconception about him Nandi-
varman of the Kanteru grant is probably Nandivarman I
Extent of his kingdom Can<Javarman Nandi-
varman n Skandavartnan Contemporary Powers-
Section II: Government, Religion and Social Life:
Administrative units Officers Offices borrowed The
Citrarathasvimi Cult Religion
91-124
125-150
151-182
163-134
CONTENTS vii
PAGES
Chapter X.-THE KANDARAS
Dynastic Appellation Kandara: Meaning King
Kandara : His date Rise of the Dynasty Extent of his
kingdom Damodaravarman : His place in the Kandara
genealogy His date Attivarman His grant Atti-
vannan's statusReligion and Social life Appendix B .. 185-199
Chapter XI. THE VISNUKUNDINS
Visnukuntfin Genealogy .. 200-208
Additional Note to Chapters III and IV .. 209-216
Index . . 217-226
FOREWORD
Dr. K. Gopalachari's book comprises the results of the re-
search he carried on in this department in the years 1934-36.
The subject of his study, The Early History of the Andhra
Country, is well known in its outline but is full of many little
problems in epigraphy and archaeology that need elucidation.
Dr. Gopalachari's thesis does not claim in any way to revolutionise
our interpretation of the history of the period ; its value consists
in a large number of detailed suggestions confirming results now
generally accepted by stronger arguments or bringing forward
fresh points of view. Parts of the thesis may be found therefore
highly technical and possibly of little interest to the general reader;
but there is much in the thesis not merely in the selection of facts
but also in their elucidation and presentation that I hope will be
recognised to be of permanent value to all historians of India.
I have great pleasure in tendering on behalf of Dr. Gopala-
chari and myself our sincere thanks to the Syndicate of the Uni-
versity of Madras for their sanctioning the inclusion of the work
in the Departmental series.
University Buildings,
Triplicane, Madras, K. A. N.
15th Sept. 1941.
PREFACE
This book represents my work as a research scholar in the
Department of Indian History in the University of Madras from
1934 to 1936. It is an attempt to present a connected history of
the Andhras and the Andhra country from the earliest times to
the advent of the Eastern Calukyas. The Visj^ukim^ins had to
be brought into the picture as they close the epoch* A full account
of the dynasty would have involved a study of Ae many contempo-
rary dynasties and increased the bulk of the volume. So the last
chapter is a compromise) a treatment of the skirts and fringes of
the subject.
The first five chapters traverse a field covered long ago by
great scholars like Prof. Rapson, R. G. Bhandarkar and Bhaga-
wanlal Indraji and unruffled by startling discoveries. This has
saved me from pioneering work ; but I have had the difficult task
of challenging great names and accepted conclusions. An inde-
pendent study of inscriptions and monuments in situ has neces-
sitated my doing so in some cases. Palaeography and the dis-
covery of a few coins like the Apilaka coin and the silver coin
of Vasithlputa Sataka^i have enabled me to reconstruct Andhra
and K^atrapa chronology on less insecure foundations and ques-
tion Rapson's identification of Pulumavi with the son-in-law of
Rudradaman. The much neglected social, economic and cultural
conditions of the period, upon which a flood of light is thrown by
inscriptions and Buddhist remains, have been dealt with at length.
One of the conclusions which should not be lost sight of is that
the Satavahanas were Andhras but began their political career
in Western Deccaru
The second period in Andhra history beginning with the
Iksvakus, one of many short-lived dynasties, is a comparatively
unexplored field. The evidence is also scanty. I have built up
the chronological scheme with the help of palaeography. A fuller
chapter on the Iksvakus than anything written before, the date of
the Brhatphaliyanas, Kandara and Vaingeyaka genealogy and
chronology are some of the contributions to the subject D. C.
Sircar's monograph on The Successors of the Sdtavahanas in East-
ern Deccan was published while I was writing my thesis. I am
idi PREFACE
indebted to him only for the Vi^nukun^in genealogy, but even
here I have modified his conclusions with the help of palaeography.
The Akola hoard of S&tavahana coins discovered in 1939 does
not necessitate a modification or abandonment of the conclusions
reached in the thesis.
Some of the epigraphical notes in the thesis have been pub-
lished in VoL XXTV No. 6 of Epigraphia Jndica.
My task of reading the Alluru inscription has been greatly"*
facilitated by the article 6f Rao Bahadur C. R. Krishnamacharlu
waiting for publication in the Epigraphia Indica. The manuscript
was with me whej^i I was reading the inscription. Readings like
'ailasa* and Vadalabhikaro', to mention only a few and the transla-
tions of a few words I owe to the article. I am indebted to JProf ,
V. V. Mirashi of Nagpur for promptly supplying me a copy of the
photograph of the coins of the Akola hoard mentioned above as also
his readings.
Professor K. A. Nilakanta Sastri has more than guided me in
my work. Discussions with him have led me to new lines of
approach and new conclusions. Apart from specific suggestions, I
owe to him in no small measure the habit of minute attention to
details and of exactitude.
K. G.
LIST OF ABBREVIATIONS
EC
IA
JA
ASWI
ASSI
CTI
JRAS
JBBRAS
JBORS
JASB
YB of the
AS of B
CAI
CSI
ZDMG
CIC
ARE
ASR
BG
ABRI
SBE
ABIA
M. BH
IHQ
TSW
AGI
GOS
Mt
ERE
J R
GI
Epigraphia Indica
Epigraphia Carnatica
Indian Antiquary
Journal Asiatique
Archaeological Survey of Western India
Archaeological Survey of Southern India
Cave Temples of Western India Burgess and
Indraji
Journal of the Royal Asiatic Society
Journal of the Bombay Branch of the Royal Asia-
tic Society
Journal of the Bihar and Orissa Research Society
Journal of the Asiatic Society of Bengal
Year-Book of the Asiatic Society of Bengal
Coins of Ancient India, Cunningham
Coins of Southern India, Elliot
Zeitschrift der Deutschen Morgenlandischen
Gesellschaft
Catalogue of the Indian Coins in the British
Museum
Annual Report on Epigraphy (Madras)
Archaeological Survey of India, Annual Report
Bombay Gazetteer
Annals of the Bhandarkar Research Institute
>Sacred Books of the East
Annual Bibliography of Indian Archaeology
xMahabharata
Indian Historical Quarterly
Tree and Serpent Worship
x&ncient Geography of India, Cunningham
Gaekwad's Oriental Series
Matsya
Vayu
BrahmSn<Ja
Corpus Inscriptionum Indicarum
Encyclopaedia of Religion and Ethics, Hastings
Journal of Oriental Research
Gupta Inscriptions
LIST OF ^LUSTRATIONS
Frontispiece The Satavahana Empire.
Plate I. Fig. 1. .. A new inscription from the Caitya Cave Karla.
I. 2. .. A silver coin of Vasithiputa Siri-Satakani enlarged
2^ times.
II. .. Remains of the relievo figures and the inscriptions
above them Naneghat cave.
in. 1. .. Nagarjunikon^a Valley.
in. 2. . . Remains of the vihara on the Naharallabodu
mound.
IV. 1. .. Remains of the mandapa east of the Mahacaitya
founded in the fifteenth year of Virapurisadata.
TV. 2. .. A coin of Siva Siri-Apllaka.
V. 1. .. Monastery on a mound (Great Dhammagiri) to
the north west of Naharallabodu.
V. 2. .. The Mah&cetiya Nagarjunikon^a.
V . 3 . . . The Mahdcetiya Nagar junikon^a.
V. 4. ., Another view of the monastery on the mound
North- West of Naharallabodu.
VI. 1. .. Sculptured beams from Stupa No. 6 Nagarjuni-
kontfa.
VI. 2. .. Sculptured beams from Stupa No. 6 Nagarjuni-
konda.
VI . 3 . . . Apsidal temple by the side of the Mahacetiya
Nagar junikonda .
VI. 4. . . Another view of same.
Vn. .. Stupa slabs from Stupa No. & Nagar jumkontfa.
VII. .. Fragment of a statue of the Buddha Nagar juni-
konda.
VIII. 1. .. The Nanegha^ Cave in which the long sacrificial
inscription is incised.
VIII. 2. .. Karla Cetiya Cave.
IX. 1. .. Simhostambho^Karla Cetiya Cave.
IX. 2. .. Sculptures at the entrance to the Cetiya Cave
Karla.
X. 1. .. Three-storeyed vihara Karla.
X. 2. .. Mahayanist sculptures in Cave No. 24Nasik.
XI. 1. .. Medals and Medallions on a pillar in the Queen's
Cave- Nasik.
XI. 2. .. Queen's Cave Nasik.
XII. 1. .. Cave at the time of Kanha Satavahana Nasik.
XH. 2. .. Cetiya Cave Nasik. <l
CHAPTER I
INTRODUCTORY
The following pages represent an attempt to trace the fortunes,
political, social and religious, of the Andhras, a people whose hoary
antiquity is attested by pieces of evidence, literary, epigraphic
and numismatic. The period covered is that from the earliest
times to the advent of the Eastern Calukyas, Politically, socially
and culturally the Dravidian Andhras (condemned sons of Visva-
mitra) proved a tremendous success. Once their empire extended
from sea to sea. The Amaravati art is the most eloquent testimony
to the cultural achievements of the race. The activities, maritime
and colonial, of the people read like romance. No ancient tribe has
on record such a unique achievement in all branches.
Andhras as a people are mentioned as early as the fifth century
B.C. The Aitareya Brdhmana speaks of them as the exiled sons of
Visvamitra, as non-Aryans evidently. 1 The Jdtdkas speak of an
Andhakapura and Andhra country. 2 The inscriptions of Asoka
mention Andhras along with Pulindas as border peoples. 3
The Saptasatakam speaks of Pulindas. 4 The Mahdbhdrata and the
Ramayana speak of Andhras along with Colas, Ceras and Pan^yas 5
The Puranas speak of the Andhrajdtlyas. 6 They enjoyed the same
political status as Kambojas, Yavanas, and Gandharas in the north.
It will thus be seen that the earliest references to the Andhra are to
people or tribe and not to their country. The reference to their
country occurs first in the Mayidavolu inscription of Siva-Skanda-
varman (4th century) in which Dhamfiakataka is spoken of as the
headquarters of the Pallava province Amdkdpatha (Andhrtipatha) .
It is therefore clear that the country derived its name from the
people, an instance with many parallels in Indian History. The
1. VH,*.
2. The Jatakas, Cowell and Thomas,
Andhra Country, I, No. 80, p. 203.
Andhra city, ibid., I, 12.
Andhakas, ibid., V. pp. 10 and 138.
3. RE,Xin.
4. Webei, Das Saptasatakam des Hala.
5. M. Bh., SabhQparvan, XXXI; Ramayana, i?,41.
6. Pargiter, The Pur&na Text of the Dynasties vj the firii Age.
ILA.-1 '
2 EARLY HTSTOBY OF THE ANDHRA COUNTRY
name Andhradesa found in literature is also evidence in the same
direction. In the 5th and 6th centuries, the Pallava province
conquered from the Vaihgeyakas, came to be called Vengoratfra. 1
It is very probable that during the Vaihgeyaka period Andhradesa
got the name Vengidesa, or Vengind4 u or Vengimain4 a ^ am by
which it was commonly known during the period of the Eastern
Cijukyas. The Andhras are spoken of as Vadugar (Tamil), and
their country Vadugavali (northerners and the country of the
northerners). But the name Andhra lived through all these
changes. Andhra people are spoken of in the Chezarla inscrip-
tion of Kandara's grandson. The inscriptions of the Maukhari
kings Isvaravarman and Isanavarman speak of AndhrQdhipati.*
An inscription of the Vakataka king Harisena speaks of his con-
quest of the Kalihga and Andhra countries. 9 An inscription of the
14th century speaks of the Andhradesa.
To-day, Andhradesa is a linguistic and cultural unity. It may
be noted, that from the earliest times the Andhras were an entity,
ethnical and cultural. Megasthenes says that the Andhras were
a separate race. 10 The Bhattiprolu alphabet* the Yengi alphabet
as Burnell would call it, and the Telugu-Canarese script were
evolved in the Andhradesa. And the Krsna Prakrt of our
period, of which we know something, has peculiarities which we do
npt find elsewhere. 11 To-day the Andhras speak Telugu and
during the Middle Ages their country was known as Telihgana.
.The extent of the Andhradesa of our period is not however
easy of determination. As Asoka's inscriptions speak of the
Andhras and the Kalihga country, and as under Kharavela
Kalihga was a first-rate power, the Andhradesa of our study was
in the north limited by Kalihga. Since Ptolemy's Maisolia and
Periplus' Masalia refer to the Andhra country, the remark that
Masalia extended far into the interior, shows that not only the
seaboard between the Godavarl and the Krsija, but also a consi-
derable area in the interior was included in the Andhradesa. In
the south, Andhradesa did not extend far beyond tile northern part
of the modern Nellore District. For the Mayidavolu Amdbdpatha
7. MSngaJfir grant, IA, Vol. V.
8. CH, Vol. m, 6, 230.
Haraha Inscription, EL, XIV, 120.
9. JRAS, 1914, p. 137.
10. IA, Vol. VI, p. 339.
11. See Chap. XL
INTRODUCTORY 3
which refers to the region around Dhamnakataka was limited by
Karmardstra. 12 No doubt under the Satavahanas the Andhra
Empire extended from sea to sea, and from the Central Provinces
in the north to Cuddalore in the south and Mysore in the south-
west. 18 An inscription of Ca$a Sati is found in Kcwjavali near Pitha-
puram; and some of the dynasties that succeeded the Satavahanas
would seem to have annexed parts of Kalinga if only for a time.
In the Srirangam plates dated A.D. 1358 14 it is said that the Tllinga
country is bounded in the north, by Kanyakubja, on the west by
Mahara?tra, on the east by Kalinga, on the south by Pan^yaka.
The description of the Andhradefa is certainly reminiscent of the
old empire of the Satavahanas. But the Andhradesa of our period
is clearly only the territory bounded on the north by .Kalinga, on
the south by the southern part of the Nellore Dt., and extending
from the coast far into the mainland in the west.
12. Chapter on Kings of the Brhatphalayana gotra.
13. Coins of Pujumavi n bearing the device of ship with masts are
found on the Coromandel Coast as far south as Cuddalore.
14. Pa&c&t purattddapi yasya defau Khyatau
Samjnau \
Av&gudak Pantyaka-Kanyakubjau deias *a tatr&sti
EL, Vol- XIV, D. 90.
CHAPTER II
ORIGINS OF SATAVAHANA POWER
Materials for a study 6f Satavdhana history
It is some decades since some Prakrt inscriptions in Brahmi
characters of a line of kings called Satavahanas in lithic records
and in literature, and Andhras (Andhrajatlyah) in the Puranic
genealogies, were discovered. The first publication of their
western, inscriptions goes back to volume VII of the J.B.B.R.A.S.
Dr. R. G. Bhandarkar was the first to translate them in his
'Notices' published in the Transactions of the London Congress
of Orientalists (1874) pp. 306 ff. Buhler 1 and Bhagvanlal Indraji 2
improved upon Dr. R. G. Bhandarkar's readings and inter-
pretations ; in 1906 the French savant, Emile Senart, gave not
merely a modest gloss in the wake of his predecessors' learned
interpretations but a scholarly edition of the inscriptions. 3 The five
short but important Satavahana records from the eastern Deccan
have been edited by Buhler , 4 Burgess, 5 Sten Konow 6 and Sukthan-
kar 7 . Thanks to exceptionally skilled numismatists like General
Cunningham, F. W. Thomas, Prof. Rapson, Bhagvanlal Indraji
and the 3J,ev. H. R. Scott, we have as much information as could be
extracted from the Satavahana and Ksatrapa coins. The Puranic
material has been carefully studied and diligently collected by
F. E. Pargiter in his " Dynasties of the Kali Age ", though his con-
clusions on the history of Purana literature have been questioned
often.
Still it is true to say that the historian's task is made difficult
by the paucity of material. A great part of the Satavahana domi-
nions remains unexplored. Recently the archaeological depart-
ment of Hyderabad have begun excavations at Palthan. Only
a hoard of Satavahana coins has come to light so far. Twenty-four
1. ASWI, Vols. IV |ind V.
2. BG, Vol. XVI.
$. El, Vols. VU and VHI,
4. fil, Vol. I, pp. 95, 96.
5. ASSI, Vol. I, pp. 61, 100.
6. ZDMG, Vol. LXH, p. 592.
7. El, Vol. XIV, pp. 153-5$.
ORIGINS OF SATAVAHANA POWER 5
inscriptions (some of which are very short) for a line of 30 kings,
who held sway over the greater part of the peninsula for more than
three hundred years, are a disappointing number. A long histori-
cal night envelops kings Nos; 4 ta 22 8 in the Puranic list. Prior to the
discovery of the Jogalthembi hoard, we had no coins which could
with certainty be attributed to Gotamiputa^ Siri-Satakar;i. The
Puri$as do not tell us much. Neither does the Brhttikathd which,
according to tradition, was written in the court of a Satavahana
king, nor the Saptafatakam, an anthology of erotic verses attribut-
ed to Hala (Satakarjd) , nor even Ltldvatt, a Prfikpt work, 9 the theme
of which is the military transactions of Hala's reign, offer many
peep-holes into the dark period. In short, the historian has still
to call to aid his imagination to forge some of the missing links.
His enterprise even now is not unlike that of adding piece by piece
to the ends of the two arms of a cantilever bridge intended to
meet at the centre; the ends of the two structures are still, for all
we know, facing each other in the air. Until they have met and
been firmly and finally riveted they cannot offer a safe passage.
The old theory
Scholars who were assiduously collecting every scrap of infor-
mation on the Satavahana period found that the names gleaned
from inscriptions and coins as well as their order agreed with those
in the Puranic genealogies; and they straightaway identified the
Satavahanas of the epigraphic and numismatic records' with the
Andhras of the Puranas. The home of the Andhras was the next
question to be tackled. The early references 10 to the Andhras and
their country enabled them to fix the habitat of this people in the
country, the heart of which roughly comprised the present Goda-
varl, Krsija and Gurfyur districts. Scholars like Prof. Rapson,
V. A. Smith and Dr. Bhandarkar found no difficulty in building
on these postulates the theory of an eastern origin of Satavahana
power, i.e., in the AndhradeSa: 11 while V. A. Smith located the Sata-
vahana capital at grl-Kakulam, Dr. Bhandarkar saw it in Dhamiia-
kataka. 12
8. Recently, however, a copper coin of : v4 I^ri-Apflaka No. 8 in the
Matsya list has been discovered in die Central ftwinces.
9. Recently brought to light by Mr. M. Ramakr$na Kavi, ^feflratt
Vol. m, Part I, pp. 3 tf.
10. Vide ^ttpra.
11. ZDMG, 1&QS p. &7; CIC, -Andhras, and Western Kfrtrapas, etc., xvi
and xvii.
12. Vide infra.
6 EARLY HISTORY OF THE AKDHRA COUNTRY
Epigraphic, numismatic and literary evidence* against it
A careful revaluation? of the materials, epigraphic, numismatic
and literary, would throw in high relief the objections to the
orthodox theory of the expansion pf Satavahana power from the
Krna-GodvarI . valleys to western Deccan. It is of course
hard to break the cake of old theories. Except for a dissentient
note here and there 33 nothing was done tp disprove the old theory
till the year 1922 when Sukthankar took up the question. 14 His
spirited attack on the old theory, only marred by an erroneous
theory of the original habitat of the Satavahanas and the absence
of a sound constructive side, does not seem to have gained the
approval of later writers.
An inscription over a relievo figure, mentioning the founder of
the dynasty (Rya Simuka) , an inscription of the reign of Kajiha
(Krsna), his brother, and an inscription of queen Nayanika, the
widow of Siri-Stakani, son of Simuka, come from Naneghat and
Nasik in the western Deccan. The Amaravati Stupa has yield-
ed many inscriptions some of which, on palaeographical grounds,
can be ascribed to the 3rd century B.C., 15 some others to the 2nd or
1st century B.C., 16 and still others to the 1st century AD. 17 The
silence of these inscriptions about not only Simuka, Kanha and
Siri-Satakajji I, but also other early Satavahanas, put by the side
of the mention of two Satavahapa kings of the 2nd century A.D. in
two inscriptions, 18 tell& its own simple story. One would expect
the long record of queen Nayanika recording the numerous sacri-
fices performed during the minority of her son to be very near the
capital and not in a place on the farthest limits of the
empire, which would be the case if the theory of an
eastern capital is correct. It may be noted that JTi^eghat is
only 120 miles, as the crow flies, from Pratisthna,^Hte capital
according to tradition, of the early SatavShanas. 19 The
13. IA, 1913, pp. 281 #.
14. ABORI, ii, pp. 21ff.
15. El, Vol. XV, Some Unpublished Awardvatt Inscriptions, Not. 4, 5,
6, 9, 10 and 19; and ASSI, Vol. I, No. 4, p. 101.
16. M, Vol. XV, ibid., 1, 2, 3, 7, 8, Jl, 12, 13, 14, 15, 16, 17 and 1&.
If. Ibid., Nos. 25, 29, 33, 36, 40, 42, 43 and 44.
18. ASSI, Vol. I, p. 100, No. 1 and p. 61, No. 2.
19. Naneghat (Ghatghar) is a pass in live Western Ghats which was in
the direct line of communication from inland market-towns like Prati$-
fhana and Tagara to the western ports like Kalyin, Barygaza, etc.
GfctGlNS Off SATAVAJtAN A
prolu inscnptipns, ' probably only a faw decades later than Asoka's
edicts', mention a king Khubirako and his father Sa. 20 AndhradeSa
wouia, therefore, seem to have been ruled by a different line in
the 3rd and 2nd centuries B.C. A coin from the Andhradeta bear-
ing the legends ( ) B (i) ra or ( ) Vira is cor-
roborative evidence pointing in the same direction. 21 True, inscrip-
tions of kings Nos. 4 to 22 are not found either in the western
Deccan or in the Andhradesa. Gotamiputa Siri-Satakani's three
inscriptions come from Nasik and Karla. The Nasik record of
Gotami Balasiri recounting her son's political achievements, and
describing his empire, makes no reference to the AndJirade&i. 22
Only records engraved during the reigns of Vasithiputa sami Siri-
Pulumavi, son of Gotamiputa Siri-Satakani, and some of his suc-
cessors, i.e., during the latter half of the second century and the first
quarter of the third century A.D., come from the Andhradesa and
Kalinga. Of the 24 records of these kings, 8 come from Nasik, 5
from Kanheri, three from Karla, one from BhilSa, two from
Naneghat (besides there are 5 short ones over relievo figures) , one
from Myakadoni, one from Cinna Ganjam, two from Amaravat!
and one from Ko^avali. 23
The Hthigumph inscription of Kharavela, king of Kalinga 24
and a contemporary of the third or fifth king in the Satavahana
Gne, throws some welcome light on the question. In the inscription
Kharavela is said to have destroyed the city of Pithuriwja in the
eleventh year. 26 In the next line an expedition against the kings
of Uttardpatha in the twelfth year is spoken of. We must, there-
fore, look for the city elsewhere than in the North. The East is like-
wise excluded for the sea lies on that side. Since the destruction
of Pithum^a and the breaking up of 'the confederacy of the
T(r) amira (Damira or Tamil) countries of 113 years' are spoken of
in the same breath and achieved in the same year, the South has
greater claims than the West. Sylvain Levi has shown that the
20. El, Vol. H, p. 328, vi; p. 329 ix.
21. Vide infra.
22. Scholars like Dr. Bhandarkar have made attempts to identify some
mountains and countries mentioned in that record with those In the eastern
Deccan. Prof. Bapson is of opinion that the record only mentions Gofaml-
puta's conquests. For a discussion of these views, vide infra.
23. Ludersi List of Brthmi Jnacr., Nos^ 34$ $87, 994, KW1, 1002, 1024,
1105, 1106, 1110, 1112, 1122 o 1126, 1141, 1146 and 1147.
24. El, Vol. XX, p. fl-80.
25. Pithumfam gadabha naibgalena kfoayati, ibid., p. 79, t, 1. 11,
8 EABLY BJ$TORY OF THE ANP&RA commit
Pitundra of Ptolemy is a Greek transliteration of the Indian
Pi&um4a. He says : -" Ptolemy places Ujtundra in the hinterland,
between the mouths of the* Maisolos .and the Manadas, or in other
words, between the deltas of the Godavarl and Mahanadi, at an
equal distance from both. W^e must, therefore, look for the site of
the city between Chicacole and Kalingapatam, if Ptolemy's informa-
tion approximates to the truth." 26 We cannot very much rely on
Ptolemy's information here, as Ije has erred in placing to the south-
east of the mouth of the Maisolos a great peninsula which, however,
existed only in his imagination. 27 We have, moreover, to say with
Yule that Maisolos is tlje 3$r$$a and not the Godavarl as Lassen
and Sylvain Levi would have it. Ptolemy places a Kantakossula
near (latitude 13430' longitude 1140') and a Koddura not far
away from (latitude 135 longitude 1130') the mouth of the
Maisolos (latitude 134 longitude 1140'). 28 Koddura has been
identified with the mddern Gu$ur m the Bandar taluq of the Kr^a
district. 29 Kantakossula is the Kantakasila of a Nagarjunikoruja
inscription of the time of the Iksvaku Virapurisadata, 30 and Kod-
dura is the Kfidura of an Amaravati inscription of the second cen-
tury A.D. 31 They were, therefore, nearer to the Krsna than to the
Godavarl. It is possible to get a better clue. Ptolemy places
Koroungkala (identified with the modern Warangal) in longitudi
15 and more in the interior than Pitundra. Warangal is place
320' degrees and Pitundra %th of a degree north of the moutl
of the Mais61os (11/40 ). 32 Warangal is in the Godavari-Kr^ni
region (south of the Godavarl). Pitundra has, therefore, to be
sought for in the Andhrade6a and not in Kh&ravela's Kalinga. The
reference to the destruction of Pithum^a along with the reference
to the breaking up of the confederacy of Tamil powers is corrobo-
rative evidence in the same direction. 33 In such a case the destruc-
tion o Pithum^a (probably then, as in Ptolemy's days, the metro-
26. IA, LV, pp. 146-47.
27. Map appended to McCrindle's translation of Ptolemy's Geography,
IA, Vol. Xm, facing p. 353.
28. /bid., p. 333 (Ptolemy's Geography, Book VII, Chap. I, Sec. 15).
29. JbuveauDubreuil, Ancient History of the Deccan, p. 85.
$0. El, XX, p. 22, Ins., P, U.S.
31. Liiders, op. c&, No. 1295.
32. Op. c&, Bk. VH, Sec. 93.
33. the fact that &e destruction of the city is spoken of along with
his wars with the northern and smjtharfc powers makes it least probable
that the city was within the kingdom of Kalinga.
ORIGINS OF SATAVAHANA PQVQER 9
polte of the J&rihra country) 34 by Kfearavela would hav$ evoked,
ittiB&ediate attd tremendous hostilities (or Would have been a result
of such hostilities) between him and his Satavihana contemporary,
if reaily the early Satavahanas had been ruling over the Attdhra-
desa. 85 If such a conflict with 'the Lord of the Deccan' had taken
place, Kharavela would not have failed to make mentipn of it in
an inscription which speaks of his wars with the northern kings,
southern confederacies and Westei-n powers. The sileftce of the
Hathlgumpha inscription on this matter is, tfierefote, condlusive
proof that the early Satavahanas were not ruling over the land of
their birth in the third and second centuries B.C.
The association of the early S&tavahanas with the Mafoflrof his,
a class of officers who are mentioned in the western cave inscrip-
tions ofcly, is another piece of evidence that supports our conclu-
sion. Queen Nayanika, wife of Siri-Satakani the third king of the
dynasty, is the daughter of the Maharathi Tranakayiro of the
Angiya family.
Numismatics tells the same story. The earliest known coins
of the dynasty are two pieces, one of tapper and the other of lead^
bearing the legend 'Siri-Sdtosa/ Considering the angular ta, the
absence of the matra over ra and the nearly form of sa,
Rapson attributes these coins to Siri-Satakani of the Naneghat
inscriptions. 36 But the nailed heads of the so- of the copper coin
make its attribution to a later king, perhaps No. 5 in the Matsya
list, reasonable. 37 These coins were picked up in western India
with which they are connected by their Malwa fabric, i.e., the
Ujjain symbol, the standing man, the representation of a river with
fishes swimming in it, which reminds us of the representation of
34. In line 11 (El, Vol. XX, p. 79) K. P. Jayaswal reads 'ova r&ja ntoe-
sitam Pithutf4aih' for 'puva r&ja etc.,' his earlier reading. (Rthumfla
built by a former king) . While the upper and lower Hmbs of the usual a
of the inscription are not connected with eacH other, in the letter read as
a, they, are connected with each other even if the crease on the stone
can be taken 4o represent the lower curve. The curve taken- BB the tipper
limb is usually big. The letter may, therefore, be more correctly ead
as jw.
35. It is also highly improbable that the Andhradefa escaped the widely
thrown net of Kharavela's expeditions*
36. CIC, Andhraa and Weffter^^^rapoa, &c., p. 1.
37. The copper coin differs qgp^ie lead coin also in Ac representa-
tion of a man standing.
If EABLY JCOTOIflr Of THE AHBfetRA, CCWITRY
river Bii>a on the coins of Eran, 38 From western fiidla came $
cote bearing the legends (partly or fully) 'R&no Sfitofca^fia.'
The alphabetical characters of the legends seem to be later than
those of the Sata coins, but the elongated instead of the squat and
rounded form of to, on all the three coins, makes a very long inter-
val impossible.
Prof. Rapson brings ii^to the list of early Sfitavahana coins,
three coins coming from the Andhradesa; according to him two of
them bear the legends ' (Ra) no ( -- ) Vira ' and one, the legend
[ (gha) ] Sadasa? 9 The former are exceptionally large simha
coins "found in a deserted site at the village of Chittala, in the
Yemagudem Talook of the Godavari District." 40 V. A. Smith attri-
buted them provisionally to Siii-Yaiia Satakani (second century
AJ>i) - 41 True, the incomplete and indistinct nature of the legends
makes it impossible for us to rely on their palaeography for their
date, but according to Rapson, their early date seems to be indi-
cated by the fact that they are struck on one side only. We, do
not possess simha coins of Siri-Yaiia. 42 It is doubtful whether these
coins were issued by any member of the Stav&hana dynasty. We
have come across neither Satavahana names ending in *vira' or
'bira,' nor such unusually big Satavahana coins. The letter read
as vi may well be read as b (i) or b (e) . 'Vira' or 'bira' strongly
reminds us of king Khubiraka of the Bhattiprolu inscriptions. 48
Significantly enough* he is there called the head of the Simha
group (Sihagothlya pamukho). On the coins the term e rano 9
comes after the personal name. It does not do so on other coins
while in the Bhattiprolu inscriptions 'raja' comes after Khubirako. 44
The striker of these coins might presumably have belonged to this
line of kings of the Andhradefa. But the distance between Bha$i-
38. Cunningham, CAI, p. 100.
39. Op. cit., pp. 2, 28.
40. Sir Walter Elliot, CSI, p. 23 n.
41. 2BMG, 1903, p. 625.
42. On the other hand, Lion coins of Vasifluputa sfimi Siri-Pu}umSvi
and MS^harlputa Sakasena, have been picked up in the Andhradei^ A
small signet of lapis lazuli discovered among the Buddhist remains of Amarl-
vatf has on it the representation of a lion with open mouth and raised left
fore-leg together with the legend "Miutisa' written in Brahml characters
of the 3rd century B.C. The lion wpuld, therefore, seem to have been of
Buddhist origin. A8R, 1905-06, p. 166.
43. J2, Vol. H, pp. 328, vi; 329, ix,
44. Op. cit.
OftXGIWS 6r S Af AV AHAJ* A POWBfc li
prolu and the findspot of the coin oasts some doubt on this identi-
fication.
On th strength of the early form of da and the incomplete
legend read as [ (gha) } Sadasa,; Rapson attributes the third
coin to Meghasvati (Megha Satakarni), ninth in the Matsya list
(2nd or 1st century B.C.). 45 As ihs coin contains neither the
full legends nor the upper part of the first letter, we cannot be quite
sure of Prof. Rapson's reading. The letter read as gha by Rapson
might very well be read as na\ what appears as a vertical to the
proper right is a scratch (compared to the central vertical), and
does not start from the end of the horizontal, i We might
reconstruct the legend thus : (JRa)(o) SWosa. 46 IivJlnscriptions
Sada alternates with Sat a; and Sata, Sati and Sada (?) are abbrevia-
tions of Satakani (the Sanskrit form corresponding to it is Sata-
karni). The striker of this coin might therefore have been any
one of the numerous Satakarnis in the Puranic list. So far as
epigraphical evidence alone is concerned, the coin may be ascribed
to a period as late as the first century A.D., for da open to the left
occurs in some of the inscriptions of Usavadata and some epigraphs
from Amaravati which, on palaeographical considerations, have
been assigned to the first century B.C. or A.D. 47
It will do well to bear in mind the remark of Biihler that "the
contemporaneous employment of more advanced types and of more
archaic ones will have to be explained by a
desire to select archaic and monumental forms for epigraphic
purposes and a failure to completely carry out this intention." 48
The type is not that of a horse as Rapson would describe it in the
Catalogue, 49 but that of a bull whose hump and horns are visible.
Coins of the bull type tentatively attributed by Rapson to the
Satavahana dynasty, come from western India, especially from
45. According to V. A. Smith, Sangha is No. 9 and Meghasvati No. 16
in the Matsya list (ZDMG, 1902, p. 659) . The coin would seem to belong
to an early period in the history of the dynasty since the form of the
tttyoftt da is that found in the N&neghat inscription and in the Nasfk inscrip-
tion of Krsna Bfija: "So far as the evidence from epigraphy is concerned,
this coin might well be assigned to the first or second century B.C." Rap-
son, op. cit., Ixxvii.
46. In the Narreghat inscription of Catarapana Satakani we have rano
for rano. Lttdfcrs op. cit., No> 1120.
47. Nos. 36, 37, 38 and 49 in El, Vol. XV, plate facing p. m.
48. IA. xxxiii, Appendix, Ind, Pake. p. 43.
49. Op. cit., p. 28.
12 EARLY HISTORY OF THE ANDHRA COUNTRY
Ujjain and Eran. 50 The Sada coin is a square piece bearing the
impress of a round die. Cunningham notes that some square coins
with impressions made from round dies come from Ujjain and
Eran. 51 It is hazardous to conclude on the provenance of a single
coin that Che early Satavahanas ruled over the Andhradesa. It is
very probable that a coin of a Satavahana king of the first century
B.C. or A.D. found its way from his dominions in the western Dec-
can into the Andhradesa in the wake of commerce.
The next group of coins found in the Andhradesa belong to
Saka Sada (Sa^La?); 52 the name is an abbreviated and corrupted
form of Sakasena Satakani. Sometimes in the inscriptions, for
want of space or other reasons, titles and names are shortened. 53
Metronymics appear on some coins whilst in others of the same
kings they do not. 54 So Sakasena Satakani can be identified with
Madhariputa Sakasena Satakani of the Kanheri inscriptions. 55 As
the Andhradesa is not mentioned in the long record of Gotami
Balasirl 56 and as no coin or inscription of Gotamiputa Satakani has
been found in the Andhradesa, it is highly improbable that
Maolharlputa Sakasena preceded the former as Rapson would have
it. 57 Dr. Bhandarkar would place Madhariputa Sakasena late in
the Satavahana series. 58 Rapson remarks : " In the inscription
(of Madhariputa Sakasena) the later form seems to
50. Ibid., pp. 54-56.
51. Op. cit., pp. 99, 100.
52. At Gu^ivatfa and AmaravatI ; Rapson, op. cit., pp. 10-12.
53. Some coins of Nahapana (JBBRAS, XXIII, pp. 13 #.), coins of Cada
Satakani and the lead coins attributed to Siri-Satakani of the Naneghat
inscriptions are instances.
54. Rapson, op. cit., pp. 20-21, 30-33, 38-42.
55. Pandit Bhagwanlal Indraji identified him with siva Siri-Satakani,
successor of Pulumavi. He was probably led to it by his reading Sirisena
for Sakasena. This reading is incorrect. (JBBRAS, VII, p. 407). Rapson
remarks that (op. cit., Intro. Ixxv) for aka-Sata "no identification with
a similar name occurring on other coins or in inscriptions (of the Satava-
hanas) can be suggested." If Prof. Rapson and Dr. Bhandarkar (EHD, 3rd
ed. p. 61) have embarked on an ocean of conjectures and doubtfully identi-
fied Saka Sada of the coins with Madhariputa Sakasena of the Kanheri
inscriptions or tried to read the third letter as na (Rapson, op. cit., p. 11,
No. 38), it is because they did not look upon Saka Sada as an abbreviated
form.
56. Vide infra.
57. Op cit., Intro, xxviii.
58. After 202 A.D., EHD, 3rd ed. p. 61.
ORIGINS OF SATAVA&ANA POWE& ia
occur in the name while the earlier form is seen in other
words." It will be shown below that the alphabet of these inscrip-
tions resembles that of a Kajiheri inscription of Siri-Yana. 59 More-
over, the rather peculiar name Sakasena reminds us of the matri-
monial alliance contracted with the Sakas by a successor of Vasithi-
puta sami Siri-Pu}umavi, Names ending in sena are brone by the
Western Ksatrapas of the line of Castana* 60 We might not, there-
fore, be grudged the conjecture that the peculiar name is a result
of that matrimonial alliance. 61 Finally even according to Rapson's
assumption, the earliest king who ruled over the Andhradesa on
numismatic evidence would be No. 21 in the Satavahana series!
While discussing the inscriptions on the reverse of three coins
of Siri-Yana from Aparanta, Kathiawar and Baroda, Rapson re-
marks that the reverse inscription is substantially the same as the
obverse inscription, but in a different dialect and written in a
variety of the Branml alphabet which has not been found else-
where and which approaches most nearly to that of the Bhattiprolu
inscriptions. "It seems reasonable to suppose, then," he concludes,
"that the two varieties of alphabet used in the Kistna District were
associated with the use of two different dialects (1) the
'Lenaprakrt' of Prof. PischeJ, and (2) a local Prakrt, per-
haps containing Dravidian elements, peculiar to the Kistna Dis-
trict. Traces of this latter dialect are probably to be seen in
certain Andhra names, such as Haku=akii; HdlaSata, &c.; and
its occurrence like that of the alphabet associated with it, on coins
of Sri-Yajiia struck in Western India must, no doubt, be regarded
as a reminiscence of the old home of the race in the Telugu country,
", 62 So far as the alphabetical peculiarities are con-
cerned, it seems that we now have a nearer analogy than the Bhatti-
prolu inscriptions, to the characters on the Siri-Yana coins. And
this analogy is furnished by the inscription on the coin of (Hi)ru
Hatakani found in Sopara and now to be seen in the Prince of
Wales Museum, Bombay. And this coin is clearly modelled on
Ksatrapa coinage. The peculiarities exhibited in individual letters
like ha and sa by these coins would thus appear to be a develop-
59. ASWI, y, No. 15.
60. The Bhattiprolu $a has its tail turned to the left and not to the
right as on the coins of Siri-Yana and Vashhiputa Satakani; the fia of the
coins bears very little resemblance to the Bhattiprolu ha.
61. A predecessor of Gotamiputa Siri-Satakani bears in the Puranic lists
a name with -sena ending: Purindrasena .
62. Op. cit., Intro, xc, xcl.
14 EARLY HISTORY OF THE ANDHRA COtJNTRY
ment that took place in western India, a development which does
not seem to stand in any necessary or direct relation to the Bhatti-
prolu alphabet. The dialectical peculiarities exhibited by the
legends of Siri-Yana's coins are the use of ha for sa, and sa for
final sa. The Bhattiprolu inscriptions do not offer us a single
instance of the use of ha for sa. We cannot, therefore, be sure that
we have in names like Haku, Hala (the names of early Sata-
vahanas) and Hiru-Hatakani, traces of a dialect peculiar to the
Kr?na district. True in the use of sa for final sa, we seem to have
a trace of the Krsna dialect, but this occurs only on the coins of Hiru-
Hatakani and Siri-Yafia (second century A.D.) and not earlier.
Considering the distance in time and space between Bhattiprolu
and the Yana coins, and the parallels we find for all other features
shown by these coins in those of the Katrapas, we may hesitate to
accept the view that the use of sa for final sa is derived from the
influence, direct or remote, of an eastern dialect. I am not at
present able to offer an explanation of this feature.
Andhradesa is rich in stupas some of which date back to the
third and second century B.C., i.e., the Bhattiprolu and Amaravati
Stupas;* 3 it is really strange that these stupas should not have con-
tained coins of the early Satavahana kings whilst some of the later
stiipos, or old stupas which were decorated and enlarged in the
second century A.D., 64 should have yielded us numerous coins of
Vasithiputa sami Siri-Pulumavi and his successors, i.e., siva Siri-
Satakani, Cada Satakani, Rudra Satakarni, Siri-Yana Satakani and
Kanha Satakani. 65
The most characteristic titles of the Andhra kings are the
metronymics. Metronymics seem to have been purely local.
Gotiputa, Gagiputa and Vasithiputa (borne by a royal artisan)
occur in the Safici, Barhut and Bhilsa (Malwa) stiipa inscrip-
tions of the second century B.C. 66 In the Pitalkhora cave inscrip-
tions of the second century B.C. the royal physician Magila
bears the title Vachiputa (Vatslputra) . C7 Metronymics like those
borne by the Satavahanas are borne by their feudatories and
officers in their inscriptions in the western Deccan. 68 In the
63. El, H, p. 325.
64. Rapsoa, op. tit., Ixxi.
65. Rapson, op. ctt., pp. 20-52.
6. Liiders, op. cit., Nos. 680, 687 and 346.
67. Ibid,, Nos. 1189, 1191-03.
68. Ibid., Nos. 1088, 1100 and 1146.
ORIGINS OF SATAVAHANA POWER 15
numerous inscriptions in the Andhrade&a metronymics occur only
in three inscriptions of the second century A.D. 69 In the
Nagarjunikoiida inscriptions the Ikvakus and the high dignitaries
of state under them, like the Mahdtalavaras and Mahdsenapatis,
bear metronymics derived from Vedic gotras. It seems, therefore,
reasonable to suppose that the Satavahanas had nothing to do with
the AndhradeSa at first and that the practice of coupling metrony-
mics with personal names became common there after the Sata-
vahanas had overrun it; that the practice was not native to the
soil is shown by the fact that the successors of the Satavahanas in
the Andhradefa, i.e., the kings of the Brhatphalayana gotra,
the Vaingeyakas, the Kandaras and the Visnukimdins, do not
assume metronymics.
Many personal names, like alphabets and dialects, are local. 70
The queen of the third king of the dynasty bears a name ending
in 'anika' (anika) , 71 and names ending in 'anaka' and 'anika' (fern)
occur frequently in the western cave inscriptions. The earliest
known inscription in the eastern Deccan to mention a name with
such an ending is the Amaravati inscription dated in the regnal
years of Vasithiputa sami Siri-Pulumavi, the first inscription
on this side of South India to mention a SatavShana king. Such
names occur frequently in the Nagarjunikon^a inscriptions. Names
resembling ' Vedisiri ' and ' Bhaya ' of the Naneghat inscrip-
tions occur in the Ku<Ja and Mahad cave inscriptions; 72 a name
beginning in Bhaya (Bhayabhuti) occurs in an unpublished KarlS
inscription. 'Skanda' which enters into the composition of the
names of some of the early Satavahanas occurs in the Ku^a, Karla
and Nasik inscriptions of the first and second centuries A.D. and
in the eastern inscriptions only after the reign of Vasithiputa sami
Siri-Pulumavi. The Satfakara of a Kuda inscription 73 bears a
striking resemblance to the surname Satakani or Sadakani.
Thus, all available epigraphic and numismatic evidence proves
not only that undeniably the centre of gravity of the early Sata-
69. Ibid., Nos. 1248 and 1271. One of them (Gomdlputa) is quite unlike
the metronymics borne by the Satavahanas, another is the metronymic borne
by Pulumavi, son of Gotamiputa Siri-S&takani; a headless statue from the
Amaravati Stupa bears the inscription ' Gotami nama(o) '.
70. A glance at the long list of names in the Bhaftiprolu and the early
western cave inscriptions is enough to prove the truth of this statement.
71. Nayan'ka.
72. Luders, op. cit, Nos. 1050, 1073 and 1091.
73. Ibid., No. 1054.
16 EARLY HISTORY OF THE ANDHRA COUNTRY
vahana power lay in the western Deccan, but also that .the early
Satavahanas did not rule over the Andhradesa.
Jain literature furnishes corroborative evidence in the same
direction. In many versions of the Kdlakdcdryakatha, including
the Long Anonymous Version, the Kalaka who changed the
Paryusana, date is said to have gone to Pratisthana, the city of
Salivahana 74 and 'the ornament of the land of Maharastra.' The
evidence of this work is of course of doubtful value ; but it is
adduced because it corroborates evidence from other sources. In
Jain chronology the changing of the date is put at 993 Vira era
(446 A.D.). The Satavahanas as a political power pass out of
history in the first quarter of the 3rd century A.D. It seems pre-
ferable to accept the account of Yugapradhana-svarupa 75 according
to which it was Kalaka I (died Vira era 376 or 171 B.C.) who
changed the date, Kalaka III coming in to confirm the change 600
years later. Then, the Satavahana king mentioned must have been
ruling at Paithan during the first half of the second century
B,C. 76
An old gdthd taken from the Niyukti (50 B.C. 150 A.D.) and
cited in the commentary on the Avasyaka Sutra says that Bharukac-
cha is known for Paithana Salavahana and Nahavana. 77 Bharu-
kaccha, the Barygaza of the Pcriplus, is modern Broach. Obviously
the Nahavana and Salavahana were contemporaries. The Sanskrit
commentary on it extracted in the Abhidhdna Rdjendra makes
them contemporaries. The name Nahavana, corrupted into
Naravaha in Jinasena's Harivamsa Purdna, is a variation of Naha-
pana. 79 The only Nahapana so far known to history, is Rdjan
Ksatrapa Nahapana of the Ksaharata vamsa, 80 who dispossessed
his Satavahana contemporary of a part of Maharastra and
Aparanta. The * Naravahas ' of the Harivamsa Purdna may imply
not the existence of two or more Nahapanas, but Nahapana and his
74. Hemacandra in his grammar gives Salivahana as a Prakrt concep-
tion of Satavahana, 1, 8, 211; Kalakacaryakatha, Norman Brown: p. 1.
75. Ibid., p. 7.
76. An inscription in the Bhilsa Topes (2nd cen. B. C.) mentions a
bhikkhu by name Patithana. Bhilsa Topes, p. 255, No. 145.
77. JBORS, 1980, fc. 290.
78. Ibid., pp. 291-293.
79. For instances of the use of va for pa see Fischers Gram, der Pr.
Spra., Sec. 144.
80. The MambaneB of the Periplus.
ORIGINS OF SATAVAHANA POWER n
descendants. The old theory that dates in Uavadata's (governor
under Nahapana) inscription 81 and the Junnar inscription of
Ayama, a minister of Nahapana, 82 must be referred to the Saka
era is to be abandoned in favour of the theory that they are dated
either in the regnal years of Nahapana or in an era starting from
the end of the first century B.C. The capital of the Satavahanas in
the first century A.D. would, therefore, seem to have been
Paithan.
The theory of a second eastern capital of the Satavahanas also
rests upon unsafe foundations. The only source of the assertion
made by many writers that the capital of the early Satavahanas was
Dhannakataka, is the conjecture of Dr. Bhandarkar that the com-
pound Dhanakatasamanehi in Nasik No. S 83 may be taken as
Dhanakatasdminehi** Obsessed by the unproven and improbable
theory of the conjoint rule of Gotamiputa Satakani 85 and Vasithl-
puta Siri-Pulumavi, and taking for granted that the donation record-
ed in the beginning of Nasik No. 3 is identical with that recorded
in Balasiri's inscription, Dr. Bhandarkar says 86 that Dhanakatasami
[lord of Dhanakata (ka) ] is a title of Gotamiputa Satakani and
reads Benakatakasami of Nasik No. 4 as Dhanakatakasami. 87
In his valuable paper on the Nasik inscriptions Emile Senart
has pointed out the orthographical objection to the identification of
Dhanakata with Dhannakataka (equivalent to or near modern
Dharanikot) of the Amaravati inscriptions. Considering the gene-
ral similarity of b and dh, he would read Benakata for Dhana-
kata. 88 D. R. Bhandarkar, however, does not agree with
Senart. He says: 89 " What is read as Dhanakata can
81. El, Vol. VIII, Nasik, No. 12.
82. ASWI, Vol. IV, p. 103, No. 11.
83. Vide infra., El, Vol. VIII, p. 65, t. 1. 2; dated in the regnal years of
PulumavL
84. EHD, p. 30, n. 13; Dhannakataka is mentioned in two Amaravati in-
scrr. (El, Vol. XV, Nos. 4 and 5, pp. 262-63) as a market-town (nigama)
only. Ptolemy mentions Pitundra (Bk. VII, Chap. I, sec. 93) and not Dhan-
nakataka as the capital of the Maisolia region. The earliest mention of
Dhannakataka as the headquarters of a district is in the MayidavoJu plates ;
also ASSI, Vol. I, No. 53, Dhamnakafaka.
85. JRAS, 1926, pp. 644-650.
86. Op. cit., p. 30.
87. Transact. Second. Lend. Congr. Ori., p. 349.
88. Op. cit., p. 68.
89. IA, 1913, p. 280, n. 16.
H.A.-8
Id EARLY HISTORY OF THE ANDHRA COUNTRY
also be read as Dhamnakata (Dhannakata) ; and as, in
Nasik inscriptions n is used instead of n (compare e.g.. dnapayati
of the same Nasik inscription) , Dhamnakata can very well be taken
to be equivalent to Dhamnakataka. Sir R. G. Bhandarkar's view,
therefore, still stands incontrovertible." To draw a parallel
between the change of na in a Prakrt word into na in its Sanskrit
equivalent (djndpayati) and the change of na into na in different
forms of a Prakrt word is misleading. Moreover, the literary Pali
form, which occurs in the Nasik inscription also, is ' dnapayati 9 and
not ' dnapayati '. The instance cited by D. R. Bhandarkar is one of
the use of na for na. 90 There is not one clear instance of na used
for na in Prakrt. The only instance cited by Pischel is one of the
change of na into cina, (rand becomes rdcina, rdcino and rdcini)
and this is not relevant to the point at issue. 91 In the Mayidavolu
plates 92 edited after Pischel' s Prakrt Grammar was published 93 we
have the word ' ana ' (anna) the literary Pali form of which is
' anna ' and no support can be derived from this inscription for
Bhandarkar^ position regarding Dhanfiakataka because the Mayi-
davolu grant comes much later and from an altogether different
area. 94
Dhanakatasami could not have been the title of Gotamiputa
Siri-Satakani, for neither was he reigning when Nasik Nos. 2 and 3
were incised, nor are the donations recorded in the two inscriptions
identical; No. 2 records a non-official grant and No. 3 an official
grant. 95 Finally the reading Dhanakatasamiyehi must be aban-
doned for Buhler's and Senart's reading Dhanakatasamanehi.
Thus the title " Lord of Dhanakataka " is hypothetical.
V. A. Smith's and J. Burgess' 96 theory that SrI-Kakulam was the
capital of the early Satavahanas is based on a passage in the Tri-
lingdnusdsanam, translated and quoted by Campbell in his
90* We have a parallel in the Naneghat ins. of Catarapana Satakani
(rano for r<wo) .
91. Op. ctf., Sec. 237, p. 167.
92. El, Vol. VI, pp. 84-89.
93. The plates were, however, discovered a year before the book was
published. *
94. In Junnar No, 10, (ASWI, Vol. IV, PI. XLIX) we have Dkamnika-
seniya not Dhanikaseniya.
95. For a detailed discussion of these points, vide infra.
96. EHI, 2nd edn. p. 194; ASSI, Vol. I, pp. 3-4. The way in which gri-
Kakulam is marked on the map appended to the Catalogue of Indian Coins in
the British Museum shows that Rapson favours their view.
ORIGINS OF SATAVAHANA POWER 19
Telugu Grammar. 97 The thick fog of legendary matter
in the passage will be apparent to any reader. Andhra
Visnu, son of the first Andhra monarch Sucandra (Simuka of the
second century B.C.) , is said in the same work to have been a patron
of the first Telugu Grammarian Kanva. 98 We know that Telugu was
in the course of formation in the fifth century A.D., from the dis-
tinctly Telugu suffix in a Visnukundin record." Atharvanacarya
quotes from the Valmtki Sutras on Prakrt, and it has been shown
that the Sutras were composed by Trivikrama 100 who according to
Dr. Hultzsch must have lived between the twelfth and fifteenth
centuries A.D. 101 The testimony of a writer removed from the
early Satavahanas by more centuries than we are from him should
not have been made the basis of such a categorical statement.
The old theory has another weak link. The attempts of Rapson
and V. A. Smith to bring the epigraphic and numismatic evidence
in line with the Puranic testimony, have led them to postulate a
rapid expansion of the Satavahana empire from the lower Godavarl
and Krsna valleys, as far as Nasik before the end of Kanha's
reign, that is to say, within 20 years. 102 Yet Simuka and Kanha
in whom the Washington and the Napoleon are combined, are men-
tioned in very short inscriptions only and no deed whatever of
theirs is recorded. If they did in fact engage in wars of extensive
conquest, the vast military operations would have necessitated large
issues of coins. Not a single coin of Simuka or Kanha has been
picked up in the western Deccan or in the Andhradesa. Not even
a Candragupta Maurya could have accomplished the feat of liberat-
ing a people and building up, in such a short period, a huge and
well organised empire, that withstood the shocks from the Sakas
for a long period. Such a rapid expansion is not known to any
period of South Indian History. Expansion from the plains over
the tableland and the mountainous regions presents far greater diffi-
97. Intro, p. ii.
98. Footnote (Introduction viii) "He who speaks irreverently w of my
Grammar composed by the command of Andhra Vishnoo shall be considered
as guilty of irreverence to his priest."
99. El, Vol. IV, the Chikkulla plates 1. 26.
100. IA, XL, 219#.
101. Ibid., p. 221: "The time of Trivikrama can be settled only within
rather wide limits. He quotes Hemachandra, who lived in the 12th century,
and he is quoted in the Ratndpana of Kumarasvamin, who belonged to the
fifteenth or sixteenth century."
102. ZDMG, 1902, p. 657.
20 EARLY HISTORY OF THE ANDHRA COUNTRY
culties than expansion from the mountainous regions over the
plains. 103 With a powerful and jealous neighbour in Kalinga, which
would seem to have thrown off the Mauryan yoke along with the
Satavahanas, a westward expansion would well nigh have been
impossible. The inventive genius of the historian has not only
painted the glories of Simuka and Kanha whom inscriptions and
literature agree to treat in a singularly unimpressive manner, but
also brought about a travesty of justice in so far as the achieve-
ments of great conquerors like Gotamiputa Satakani and some
of his predecessors like Satakani I have been fathered upon dim
figures in history.
True, Satakani I, the third king of the line, is called 'Dakhina-
pathapati.' 104 But Dakhinapatha is an ambiguous term. In its
widest sense it includes the whole of the Peninsula south of the
Vindhyas; since a passage in the Vdyu Purdna excludes the Nar-
mada and the Tapti valleys, 105 the term seems to have been used
in a narrow sense, then, as now. To go to an earlier work than
the Puranas, the author of the Periplus (first century A.D.) men-
tions the market-towns of the Dachinabades separately from the
market-towns of Damirica, mistakenly called by him Limyrike, i.e.,
the extreme south of the Peninsula including particularly the Cera,
Cola and Pandya countries. 106 The extreme south is likewise ex-
cluded. ^ Since the Maisolos of Ptolemy is most probably the
Krsna, 107 and since the Maisolia of Ptolemy is the Masalia of the
Periplus, Masalia would seem to be the name of the lower Krsna-
Godavari region, i.e., the Andhradesa. The author of the Periplus
says that this region was studded with centres of trade and indus-
try. 108 Yet all the market-towns (of which Paithan and Tagara
identified with modern Junnar are the most important) of the
Dachinabades mentioned in the Periplus are in the western Dec-
can. 109 Thus it is clear that the Dachinabades of the Periplus
excludes the extreme east and south of the peninsula. 110
103. Ltiders, op. cit., Nos. 1112 and 1114.
104. ASWI, Vol. V, p. 60, PI. LI.
105. Chap. 45, Verse 104. Bibliotheca Indica ed.
106. Schoff, The Periplus of the Erythraean Sea. Sec. 51.
107. Vide supra.
108. Schoff, op. ctt., Sec. 62.
109. Ibid., Sec. 51, 52, and 53.
110. Suzerainty over the whole of the Peninsula is therefore to be ruled
out.
ORIGINS OF SATAVAHANA POWER 21
Even those who have propounded the theory of western origin
of Satavahana power have failed to correlate properly the Puranic
with epigraphic and numismatic evidence. Relying upon a passage
in the Aitareya Brdhmana which speaks of the Andhras as living
on the fringes of Aryan civilisation, 111 Mr. P. T. Srinivasa Ayyan-
gar places the Andhras in the Vindhyan regions. 112 But we do not
know the exact limits of Aryan civilisation in those days. It has
been proved beyond doubt that the ancient home of the Andhras
then as now was the lower valleys of the Godavari and the Krsna.
His theory of western origins of Satavahana power, not accom-
panied by arguments, looks like a shot in the air.
Sukthankar cuts the Gordian knot by challenging the Andhra
affinities of the Satavahanas. 113 According to him in the whole
range of epigraphic records, the Satavahanas are nowhere called
Andhras. The passages from Greek authors which mention the
Andhra country and people contain no reference to the Sata-
vahanas while those in which certain Satavahana kings are men-
tioned have nothing to say about the Andhras. The hopeless con-
fusion of the Puranas makes their evidence worthless. 114
All these objections would vanish if the available pieces of
evidence are properly weighed. Sukthankar treats 'AndTira-
bhrtya' as a Tatpurusa compound (Servants of the Andhras)
'having regard to the parallel phrase Sungabhrtya applied to the
Kanvas.' The Puranic words 'Andhrajdtiyah'' and 'Kdnvdyanams
tato bhrtyah Susarmdnah prasahya tarn' (Matsya) exclude the gram-
matical construction adopted by Sukthankar. The compound should,
therefore, be treated as a Karmadhdraya one, in which case it
would mean 'Andhra Servants.' Then the Satavahanas could have
been Andhras and Andhrabhrtyas. Satavahana is a family or a
dynastic name while Andhra is a tribal name (Andhrajdti) . In an
inscription we have the expression Satavahana kulam; 11 * in Prakrt
111. Vide supra.
112. IA, 1913, pp. 28 f.
113. Op. cit.
114. Some of the Puranas call these kings Andhras ; others call them
Andhrabhrtyas, and there are others that call them by both names. The
majority of the Puranas distinguish between Andhras and Andhrabhrtyas, and
state that the Ardhrabhrtyas succeeded the Andhras. Ibid., p. 29.
115. 'S&tav&hana kule Kanhe r&jdni etc'., Nasik, No. 22, El, Vol. Vm, p. 93;
and 'Satavdhawkula yasapatHh&panakarasa; Nasik, No. 2, ibid., 60 t. 1. 6.
22 EARLY HISTORY OF THE ANDHRA COUNTRY
a' essentially means 'family.' 116 The term 'jati* on the other
hand means *caste or tribe.' That the terms Satavahana and
Andhra are not identical is shown by the fact that in the grants of
Pallava Siva-Skandavarman Satahani-rattha (Satavahani-rattha)
and Amdhapatha (Andhrapatha) are the names given to two
provinces. 117 No wonder then that the inscriptions which give the
dynastic name considered the mention of the tribal name super-
fluous. 118 In the Pallava Kadamba and Calukya records the dy-
nastic appellation only is given and if literary evidence should
throw some light upon their tribal connections no one would
challenge them by saying that such connections are unknown to
epigraphic records. As for the Greek writers, Megasthenes does
not mention the dynastic name of the Magadhan, Kalingan, and
Andhra kings. Ptolemy mentions Polemaios (Vasithlputa sami
Siri-Pulumavi of the records) of Paithan, but does not give us his
dynastic name. Are we to hold that he did not belong to the
Satavahana kula ?
It will not do to ignore the Puranic testimony to the extent
to which Sukthankar has done. No doubt the Puranas have to
answer charges of defective chronology, incomplete lists of kings,
corruption in names and different readings of the same passage
in different manuscripts. Most of these defects are a result not of
ignorance of facts on the part of Puranic writers but of misreading
of manuscripts and bad copying; Pargiter thinks that the corrup-
tion in names must have occurred in the Sanskritization of Prakrt
names. 119 The earliest Purana, the Bhavisya, from which the
Matsya, Vdyu, Bhdgavata and Visnu derive their account, Sanskri-
tized earlier metrical accounts in literary Prakrt; the dynastic
portion terminates with the downfall of the Andhras and the rise
116. In the Mahdvamsa, Dharfimapadam, the Five Jatakas and Kuddhaka-
p&tha, it is used in this sense only. In the Talaguntfa ins. of Kakusthavarman,
(El., Vol. VIII p. 32, t. 1. 3), Kadambakula signifies the Kadamba family.
117. The Hlra-Ha^agalli and the Mayidavolu plates.
118. In Usavadata's Nasik and Karla inscrr. Nahapana is called a Ksaha-
rata, and we know from Nasik No. 2 that Ksaharata is a family name, (Kha-
kharatavasa). In a Kanheri ins. (Liiders, op. cit., No. 1021) MahaJcsatrapa
RudradSman's daughter is said to have belonged to the Karddamaka race or
family. From literary and other sources we know that Nahapana and Kudra-
daman belonged to the Pahlava and (Saka tribes.
119. Pargiter, Dynasties of the Kali Age, pp. 97 f. It must be noted, how-
ever, that Pargiter's conclusions have often been challenged by Keith, Kirfel,
and other writers.
OE1GINS OF SATAVAHANA POWER 28
of their servants. The Vdyu, Brahmaqda, Vi$nu and Bh&gavata
carry the narrative to the rise of the Guptas, but are silent about
the whirlwind campaigns of Samudragupta. Pargiter notes 120 that
between 325-340 A.D. the accounts were revised twice. The
Puranas were, therefore, redacted at a time when the Satavahana
dominion in South India was a thing of the immediate past.
Pargiter has shown that there is an indication that a compilation
was begun in the latter part of the second century A,D. in Siri-
Yana's reign, for five manuscripts of the Matsya, of which three
appear to be independent, 121 speak of him as reigning in his ninth
or tenth year. The nearness of the Puranas to the Satavahanas
makes their testimony about their tribal affinities unquestionable. 122
Having cut himself from the old moorings of Puranic testi-
mony, Mr. Sukthankar seeks to locate the habitat of the Sata-
vahanas in the modern Bellary District. The only source of his
assertion is the terms ' Sdtavahani-hdra ' and ' Sdtdhani-rattha '
(Sdtavdhani-rattha) occurring in inscriptions coming from the
small compass of the Bellary District. 123 On the analogy of in-
habitants lending their names to countries, he looks upon Sata-
vahani-hara corresponding to the modern Bellary District and per-
haps its neighbourhood as the original habitat of the Satavahanas.
120. Ibid., pp. xiii, g 23.
121. Ibid., p. 42, n. 8.
122. While Sukthankar accuses the Puranas of calling Andhrabhrtyas
(servants of the Andhras) Andhras, Ray Chaudhuri suggests that the name
Andhra "probably came to be applied to the kings in later times when they
lost their northern and western possessions and became purely an Andhra
power governing the territory at the mouth 6*f the river Krishna" (Pol Hi*.
of India, p. 280). Satavahana rule over the Andhradesa lasted for three
quarters of a century. Sami Siri-Yana, No. 27 in the Matsya list, ruled
over the western Deccan. The Satavahanas would seem to have become a
purely eastern power only. a few decades before their fall. It has been shown
that a compilation of the Puranas was begun in Siri-Yana's reign, at a time
when the Satavahanas were a western as well as an eastern power.
123. The Myakadoni inscription and the HIra-Hatfagalli plates. Myaka-
doni is a village in the Adoni taluk of the Bellary Dt.
Hemacandra gives Salahana and Salavahana as variations of Satavahana
(Pischel, op. cit.,) . True, visaya (HIra-Hatfagalli plates) and rostra (Colarat-
tha) denote sometimes a kingdom. But in the Uruvupalli grant Muntfarastra
is referred to at the end as a visaya (IA, Vol. V, p. 51 t. 11. 17, 28). The
Kudurahara of the Kontfamudi grant is called Kudrahara-visaya in the
Vairigeyaka grants and Kudrara- visaya in some Eastern Calukya grants*
Therefore Satavahana 'fl/ifira/ Pallava 'rostra' and Vaingeyaka 'viaya' would
denote the same territorial divisionnot bigger than a modern district.
24 EARLY HISTORY OF THE ANDHRA COUNTRY
If Satavahani-hara was the starting point of Satavahana power,
why are not inscriptions of the early Satavahanas found in this
territory? Worse still, dnly an inscription of the last king of the
line is found here; 124 and Sukthankar bases his conclusions on
the provenance of inscriptions! He gives instances of provinces
getting their names from their early inhabitants. But the term
in question is an instance of a dynasty lending its name to a part
of the kingdom and not of a people lending their name to the
whole kingdom. 125
We are prepared to say with Mr. Sukthankar that the pro-
vince must have been so called on account of "some intimate con-
nection " between the land and the dynasty. A tentative solution
may be proposed that under the later Satavahanas, a town in Sata-
vahani-hara became the seat of their capital which would have
been shifted to the east after the conquest of their western terri-
tories by the Western K^atrapas. True, during the reign of the
last king, the province is under a Mahasenapati. 126 Instances of the
headquarters of a district lending its name to the district are nume-
rous, e.g., Govadhana, Govadhanahara (Liiders, List No. 1124) ;
Kudura, Kudurahara (No. 1328); Patithana and Patithanapatha
(No. 988) and Dhanfiakataka and the kingdom of To-na-kie-tse-kia
which may be considered as the Chinese representative of
Dhannakataka, In the Talagunda inscription of Kakusthavarman,
the capital of the Pallavas is called Pallavapuri. Kandarapura at
which Maharaja Damodaravarman of the Ananda gotra is said to
have ruled 127 must have received its name from that prince Kan-
dara, who is mentioned as an ancestor of Attivarman. 128 The
Anandas and the Pallavas are not far removed from the Sata-
vahanas. The capital of the Satavahanas might have been called
Satavahanapura or Satavahanipura and the district in which it was
situated, Satavahani-hara ; 129 the Pallavas might have continued
the name.
124. The Myakadoni inscription of Pulumavi.
125. "The learned Parimellalagar is inclined to make Cola the name, like
the Pan^ya and Cera, of a ruling family or clan of immemorial antiquity and
renown." The Cdlas, Vol. I, p. 24. Cofamantfalaw would then be an instance
of a territorial designation formed on a dynastic name. It is not, however,
ah instance of a part of a kingdom getting its name from the dynasty to the
exclusion of the other parts.
126. The Myakadoni inscription of Pulumavi.
127. El, Vol. XVII, p. 328.
128. IA, Vol. IX, pp. 102-103.
129. Excavations of the type conducted at Nagarjunikontfa may bring to
ORIGINS OF SATAVAHANA POWEH 25
Now the Puranic, epigraphic and numismatic evidence can be
correlated in a way different from those so far considered. The
term Andhrabhrtya, 'Andhra Servant 7 gives the clue. Will the
facts of Satavahana history make it improbable that the Sata-
vahanas, undoubtedly Andhras by tribal connections, were high
officers of state under the Mauryas like the Kanvas, called the ser-
vants of the Sungas ? True, the Andhra territory while acknow-
ledging Mauryan suzerainty enjoyed some independence unlike
the 'King's Dominions.' This semi-independence need not have
been a bar to the Andhras (of the ruling family) accepting offices
under the suzerain. In Asoka's edicts, Yavanas are politically
classed with the Andhras; 130 and we find a Yavana serving as gover-
nor under Asoka. 131
It may still seem impossible to ascertain how these 'Servant
Andhras' of the eastern Deccan drifted into the western Deccan.
Asoka's edicts and the Girnar inscription of Rudradaman form
links in the chain of evidence, and suggest that the Mauryan domi-
nion in South India was the medium through which this drift
took place. In Asoka's edicts the Yavanas are placed with the
Gandharas and Kambojas in the north-west, and still Surastra was
governed by a Yavana king for Asoka. Indian History offers ujs
many clear instances of dynastic drifts like the Mauryas of Konkan,
the Guptas or Guttas of Guttal and the Colas of Renan^u. In the
reign of Pulakesin II "in the Konkanas, the watery stores of the
pools which were the Mauryas were quickly ejected by the great
wave which was Candadanda, who acted at his command." 132 A
prince, Dhavala, of the Maurya lineage is mentioned in the Kanaswa
inscription of A.D. 738-739, in the Kotah State, Rajaputana. 133 In
an inscription of Vaghli in the Khandesh District dated . 991,
princes of the Maurya clan, the original home of which is said to
have been the city of Valabhi in Surastra, are mentioned, 134 The
Guttas of the twelfth century A.D. with their capital at Guttavolal,
which may be safely identified with the modern Guttal in the Karaji
taluq of the Dharwar District (where all their records are found) ,
light the remains of the capital in the Bellary Dt. or its neighbourhood. In
the Adoni taluq there is a village called Satanuru.
130. RE. Xin.
131. The Girnar ins. of Rudradaman, El, VIII, p. 45, t. 1. 8.
132. The Aihole inscription of Pulakesin II, I A, VIII, p 244.
133. Ibid., XIX, p. 56.
134. El, Vol. U, pp. 220 ff.
H.A.-4
26 EARLY HISTORY OF THE ANDfiRA COUNTRY
trace their descent to Candragupta through a Vikramaditya who
is specified as a king of Ujjain. The earliest Telugu records
(eighth century A.D.) 'from the Cuddapah District including the
Malepa<Ju plates of Punyakumara, have brought to light a line of
kings claiming Cola descent, who had however their dominion in
Pallava territory. 135 The Velurpalayarh plates give us the clue; 136
there the Pallava Simhavi?nu is said to have "seized the country
of the Colas embellished by the daughter of Kavera whose orna-
ments are the forests of the paddy (fields) and where (are found)
brilliant groves of areca." When the power of the Colas fell to a
low ebb and Sirhhavisnu's sway extended over the Cola country,
the scions of the eclipsed Cola dynasty must have sought service
under their conqueror and so moved up north. 137 Epigraphical
records from the Maddaguri taluq of the Tumkur District refer to
a certain Dhanamjaya Eriga, a Cola. The Colas of the Tumkur
District may have been of a common stock with the Colas of
Renandu among whom we have a Dhanaihjayavarman. 1;> ' 8 The drift
of the Mauryas from Magadha to Konkan, Khandesh and Rajputana,
and of the Guptas (Guttas) from the north to Guttal may be ex-
plained in the same manner. Even as late as the sixteenth century,
Cd]a chiefs with traditionary descent from Karikala are found
as viceroys under Vijayanagara rulers. 139 The instances so far
cited support the theory that in the days of tribulation and rather
obscure existence under their Mauryan suzerains, scions of the
royal family in the Andhradesa might have passed into the service
of the Mauryan kings and so have gone to the western Deccan as
viceroys, thereby getting the Puranic appellation Andhrabhrtya .
A fragment of Rock Edict VIII discovered at Supara 140 makes it
certain that a part of the western Deccan was included in the
*King's Dominions/ When the strong arm of Asoka disappeared,
their shrewd and more fortunate descendants would have found
themselves in a position to strike a blow in their own interest, not
in the land of their birth which was far away, but in the land of
135. The Malepatfu plates, El, XI, p. 345.
136. SH, Vol. H, No. 98, pp. 507 ff.
137. The names of the first two princes mentioned in the MalepaqUi plates,
Nandivarman and his son Simhavisnu, bear striking resemblance to some
names in Pallava genealogy.
138. 380 of 1904.
139. ARE, 1909, p. 112.
140. CH,Vol. I.
ORIGINS OF SATAVAHANA POWER 27
their adoption. It is possible that in some such manner Simuka,
an Andhra, might have started the political power of the dynasty.
But at present we have no evidence in favour of this conjecture.
For all that we know, the ancestors of the Satavahanas of the
western Deccan might not have belonged to any royal family in
the Andhradesa. They might have been nobles or fortune hunters
who readily passed into the service of the Mauryan suzerains and
so moved up to western Deccan.
If the Jain legends which mention Paithan as the capital of the
first Satavahana king may be believed, it would seem to be the
starting-point of the Satavahana power. The close association of
the Satavahanas with Mahdrathis (matrimonial alliance) and the
office of Mahdrathi show the extent to which Simuka enlisted the
support of the powerful Rathikas of the west. This reminds one
of the Cutu-Pallava matrimonial alliance which would seem to
have, in the same measure, contributed to Pallava ascendancy in
the south (later Pallava inscriptions mention a Cuta-Pallava as
the founder of the dynasty). If the Puranic 'bhrtydh' and f $a-
jatiyah' are correct, it would seem that Simuka was also helped
by a number of faithful Andhras who like his ancestors had moved
up to the western Deccan. The early Satavahanas seem to have
been engaged in the first instance in the conquest of Maharasfra
north and south, Malwa and the modern Central Provinces.
CHAPTER III
THE EARLY SATAVAHANAS
The Starting point
Relying upon the supposed date in the Maurya era in the
Hathigumpha inscription of Kharavela, 1 and looking upon Khara-
vela as the third member of the Cedi dynasty of Kalinga 2 like Siri-
Satakani of the Satavahana dynasty, Rapson would place the
beginnings of the dynasty (170 B.C. + 41, i.e., Simuka 23, Kanha
18) somewhere between 220 and 211 B.C. 3 The chronological
Arrangements adopted here would place Simuka 384 years 4 before
150 A.D., i.e. 234 B.C. Though an edict later than Rock Edict VIII
dated in the tenth year of As oka 5 has not been found in western
Deccan, 6 it is improbable that Asoka's reign witnessed a break-up
of the empire; and Asoka's death would seem to have taken place
somewhere between 236 and 232 B.C. 7 The same conclusion can
be arrived at in another way. As Pargiter has pointed out, 8 the
1. Messrs. R. D. Banerji and K. P. Jayaswal remark: (El, Vol. XX,
p. 74) "It has been proved by repeated examinations of the rock that there is
no date in the Maurya era as supposed by Bhagwanlal Indraji and our-
selves formerly." The inscription reads "Muriya Kola vochinam ca coyathi
Amga-satika (m) turiyam upadayati " * causes to be completed the 11 Ahgas
of the 64 letters which had become lost (or fragmentary) with the time of
the Mauryas.'
2. Vrddharaja and Ksemaraja like Bhiksurdja are epithets applied to
King Kharavela, and not the names of his father and grandfather respec-
tively as suggested by Rapson. (CIC, Andhras and Western K?atrapas etc.,
xviii). The text has 'Khemaraja sa Vadharaja sa Bhikhuraja sa Dharmaraja
pasamt(o) Simat(o) anubhavato kaldnani rajasi Vasukula vinisrito mdha-
vijayo Raja Kharavela siri' El, Vol. XX, p. 80.
3. Op. cit.
4. It would be shown below that 6ivaskanda of the Puranas (No. 26
in the list) was the king defeated by Rudradaman twice before 150 A.D.
5. 10th year after his coronation.
6. A fragment found at Sopara.
7. The Puranic statements would place an interval of 49 years between
the accessions of Candragupta and Asoka. According to V. A. Smith's
scheme of chronology, Candragupta began to reign in 322-21 B.C. So Asoka
would have ascended the throne in 272 B.C., he is said to have ruled for 36
years and been anointed 4 years after his accession.
3. The Purana Text of the Dynasties of the Kali Age, p. 37,
THE EARLY SAT AV AHAN AS 29
Puranic total of the individual reigns (of the Satavahana kings)
excluding 24-a, is only 442% years even if we take the longest
periods, where there is a difference. But the whole duration is
said to have been 460 in the Matsya, though it is given as 411 in
Vdyu. The addition of Satakarni mentioned in e Vdyu only would
increase its total to 440. It would therefore appear, that the
total 442% years has much to be said in its favour. The end of
the Satavahana dynasty cannot be placed earlier than 207 A.D., 9
and 442% years before 207 A,D. would give us the same 234 B.C. 10
Founder of the dynasty : Simuka -
As the Puranas speak of 'Simuka Satavahana Sirimato' as
simply 'Sisuko 'ndhrah sa jafiyah' before the coup d'etat, and as
9. The Satavahanas ruled for 55 years after 150 A.D.
10. The Puranas place the Satavahanas after the Kanvas, i.e., (Mauryas
139, 6ungas 112 and Kanvas 45) 25 B.C. The Puranas treat contemporary
dynasties as successive. They say that 18 Sakas (Western Ksatrapas) came
after the Satavahanas. Some of the Western Ksatrapas of the Castana line
were certainly contemporaries of the later Satavahanas as inscriptions, coins
and Ptolemy's statement would show. It is not possible under the Puranic
scheme to place Gautamiputra Satakarni and Pulumavi who certainly
preceded Rudradaman of the Girnar inscription of 150 A.D., after that
date; for does not Ptolemy call Castana the grandfather of Rudradaman,
a contemporary of Pulumavi ?
In the memoirs of the Archaeological Survey of India, No. 1, Dates of
Votive Inscriptions from Sanci, R. P. Chanda argues on palaeographical
grounds that the earliest votive inscriptions are later than Asoka's and
Heliodorus' inscriptions by a century, and that the Siri-Satakani inscrip-
tion belongs to the later group, which he assigns to the period Between
75 and 20 B.C. He identifies the Siri-Satakani of the Sane! inscription with
No. 6 in the line. No. 6 in the line, according to the chronological scheme
adopted here, belongs to the years between 180 and 130 B.C. There is
no reason why he should not be one of the numerous Satakarnis found in
the Puranic lists after No. 6.
A comparative study of the palaeography of the Naneghat and Bhilsa
inscriptions enables us to reject the view of Buhler that the Satakanis of the
two inscriptions are identical. Bhilsa va with the shorter neck and rounded
body, the more ornate Bhilsa i sign, the less angular ta with the vertical at
the centre and the da with the more rounded back than the Naneghat inscrip-
tion, stamp the Bhilsa inscription as one later than the Naneghat inscription.
The Bhilsa Tope inscription under reference is carved on the bas-relief
of a tor ana in the middle of the upper architrave of the South Gateway.
It records the donation of a Vasitfiiputra Ananda, the foreman of the arti-
sans of Siri-Satakani. The plates published in the JBORS, 1917, make it
clear that Vasithiputra is the metronymic not of the king but of the arti-
san. For Rapson's view, op. c&., xlvii.
30 EABLY HISTORY OF THE ANDHRA COUNTRY
'Raji Simuka' after it, 11 it is certain that he was the founder of the
dynasty. But as his brother Kanha is also said to have belonged
to the Satavahana kula, 12 Simuka could not have given the name
of the dynasty. Then who gave the name to the dynasty ? The
question cannot at present be satisfactorily answered.
The meaning of 'Satavahana* is as obscure as those of 'Calu-
kya/ Tallava' and 'Vakataka.' For one thing the name is not
Satavahana as Rapson would have it. 13 Jinaprabhasuri, a Jain
monk of the fourteenth century A.D., derives the word thus :
'Sanoterdanarthatvdt lokaih Satavahana iti vypadesam lambhitah'
i.e. people call him Satavahana, because (the verb) 'sanoti* signi-
fies Ho give' and hence one by whom were given (sdtdni) con-
veyances (vahandni) was called Satavahana. 14 Another deriva-
tion of the name is given in the Kathdsaritsdgara which explains
it as meaning 'he who rode a yak$a named Sata (in the form of
a lion)'. 15 These fanciful explanations show that the origin of
the term was forgotten long before the fourteenth century. Recent-
ly M. Przyluski has given us an equally fanciful explanation. 16
11. Pargiter: op. cit., p. 38.
12. The Puranas give various readings : Matsya generally 'Sisuka'; d
Mt. Steudhrah; e Mt. isuka; e Vdyu Cismako; Visnu iprafca; j Vi?. Sudhra.
According to Pargiter (op. cit., p. 38, n. 17), Simuka was misread 'Sisuka* and
Sanskritized * 6isuka ' ; and Sisuka cannot be Sanskrit !rimukha (one with
a glorious face as Bhagwanlal and Buhler proposed (ASWI, Vol. V, p. 69).
Sans. *6r!' is invariably represented in Prakrt by 'siri.' In the relievo inscrip-
tion at Naneghat itself, Simuka bears the honorific prefix 'Sirimato'. 'Siva'
enters into the composition of some Satavahana names and is used as an
honorific prefix even by early Satavahanas. However, palaeography prevents
us from subscribing to Burgess' view that "Sivamaka (of an Amaravati
inscription) might possibly be the same as Simuka of the Naneghat ins-
cription No. 3." (ASSI, Vol. I, p. 62, n. 2).
13. Satavahana with the dental s occurs in Sana's verse, Hemacandra's
works and Somadeva's Kathdsarits&gara. However, Vatsyayana in his Kdma-
sutra spells, the word with palatal sibilant s. The words as given by
Bhandarkar are (EHD, p. 69, n. 7) 'Kartaryd Kuntaldh $&takarnih Sata-
v&hano Mahadevim Malayavatim (joghflna) '. Dr. Fleet remarks (JRAS,
1916, p. 818 n. 3): "It is, however not possible that Vatsyayana himself can
have used the palatal sibilant in these two names." As will be shown
below, Satakarni with the palatal sibilant s is a mistake for Satakarni with
the dental s. It then becomes easier to suppose that Satavahana with the
palatal sibilant s is a similar error in spelling.
14. JBBRAS, Vol. X, p. 132.
15. Kathdsaritsagara, trans. Vol. I, p. 37.
16. JRAS, 1929, p. 278,
1HE EAfcLY SAtAVAfiANAg &
According to him Sadam > saddm and sdddm in Muncja languages
mean 'horse/ vdhana is a Sanskritization of han or hapan meaning
'son,' Satavahana is rendered 'son of horse'. The explanation
given is, that princes born of the magical union between the chief
queen and the sacrificial horse (during the performance of the
Asvamedha) would have come to be called 'sons of the horse.' It
is not possible here to traverse the grounds, highly speculative, on
which Przyluski seeks to trace pre-Dravidian influences in post-
Aryan society and institutions in India. We must be content with
the observation that, according to the learned philologist's expla-
nation every ksatriya prince, whose father had performed the
Asvamedha would be a Satavahana or Satahapan. Yet history
knows of only one dynasty that went by that name. It is possible
to consider 'Sata' as the past participle of San, to obtain, to gain;
Satavahana would then mean one who obtained a 'vahana', perhaps
one who by his deeds secured a high position in Mauryan military
service; and the Satavahanas were according to the Puranas 'Ser-
vant Andhras.' The Silappadikaram refers to Purambanaiydn
vdlkottam and Pdsan^a Sdttan (ix, 11. 12 and 15). The commen-
taries explain Purambanaiydn by Mdsdttan and Sdtavdhanan ; I do
not think that these references to the village deity, the
guardian of the boundary of the village, and to his proficiency
in the heretical lore, have any place in the elucidation of the dy-
nastic name of the Satavahanas. For one thing Adiyarkkunallar,
the commentator, is only as old as Jinaprabhasuri. The spelling
in 'Sasta' is another argument. 17
The wife of Satakani I was versed in and performed numerous
sacrifices and worshipped Vasudeva, Krsna and Sarhkarsana.
Gotamiputa Siri-Satakani's mother led the life of a royal rsi. The
former prides himself over ' having stopped the contamination of
the four castes.' A later Satavahana king bears the name Yana
Satakani.
All Puranas are agreed that he ruled for 23 years. According
to Jain legend the first king, Satavahana by name (evidently
Simuka), built Jaina temples and cetiyas. But in the closing
years of his reign he became a wicked king and was dethroned
and killed. 18
17. The reference to S&ttan in ilappadik&ra i m was pointed out to me
by Mr. V. R. Ramachandra Dikshitar; but I am unable to follow his inter-
pretation for which see Jayanti Ramayya Pantulu Commemoration Volume,
pp. 156-8.
18. JBBRAS, Vol. X, p. 184.
EARLY HISTORY OF THE AN&riRA COUNTRY
Kanha 1
Simuka was succeeded by his brother Kanha. 19 In his time
the Satavahana kingdom extended as far west as Nasik if not fur-
ther. Evidence of the modelling of Satavahana administration on
Mauryan lines is furnished by the Nasik inscription of his time,
which mentions the construction of a cave by a Mdhdmdtra in
charge of the sramanas or monks (at Nasik) Maham&bras are a
class of officials mentioned in the Asokan inscriptions. The cave is
the earliest excavation in the series, and stands far below the other
caves. The cells on three sides are decorated with cetiya arches
at the top of the openings. Of the four pillars that support the
roof of the verandah, two are half -pillars and the others are square
at the top and bottom, and octagonal in the middle. They have
no capital. 20
According to Rapson, Kanha would have reigned for 18 years. 21
Matsya has generally o^tddasa. But some Mss. of Vdyu read
asmdddasa. 22 Pargiter has pointed out in the introduction that where
there are two readings, one asmdddasa and another astadasa?
dbdandasa would reconcile these different readings.
Siri-Sdtakani I
On epigraphic as well as literary (Puranic) evidence, the third
king of the line is Siri-Satakani according to Rapson the Siri-
Sata of the coins, the husband of Nayanika, the daughter of
Mdhdrathi Tranakayiro. 23 According to Rapson it is not possible
to reconcile the Puranic statement that Satakarni I was the son of
Krsna, the brother of Simuka, with the evidence supplied by the
19. The Mt., Fa., BcL, Bhdg. and Vs. are agreed in calling him the
brother of Simuka. According to Rapson this fact fully explains the
absence of his name in the Naneghat relievo inscriptions (op. cit., p. xix).
20. PI. IV, No. 4.
21. Op. cit., p. Ixvi.
22. Pargiter: op. cit., p. 39, n.28.
23. In a Nasik inscription (El, Vol. VIII, p. 88), Visnudata, daughter
of Saka Agnivarman calls herself a Sakanika. In the Poona plates of
Prabhavatigupta (El, Vol. XV, p. 41, t. 11. 7 and 8), Kuberanaga, wife of
Candragupta, is said to have belonged to the Naga tribe. On these analogies
the name Naganika may indicate the tribe to which she belonged. The
Naga alliance is partially preserved in the Jain legends which make Sata-
vihana the son of a Brahman girl and &esa, the king of serpents. (JBBRAS,
Vol. X, p. 132).
IttE fiARLY SATAVAHANAS 3*
relievo figures of Simuka, Siri-Satakani, and the latter 's family. 24
The relievo inscriptions mention Rdya Simuka, then Siri-Satakani
and his wife, then a Kumdra Bhaya. . , then Mdhdrathi Tranakayiro,
evidently the father of the queen, then Kumaras Hakusiri and
Satavahana. Most of the relievo figures are almost lost, only the
legs being partially visible; the rest are completely lost leav-
ing only the space. As will be seen presently, the space for two
figures between those of Nayanika and Kumdra Bhaya , . , was filled
by the figures of Vedisiri and Kumdra Satisiri, Then Kanha has
no place in the relievos.
Other results which Biihler and Rapson have arrived at by a
comparative study of the relievo figures and the sacrificial inscrip-
tion of Nayanika at Naneghat are that queen Nayanika was the
mother of Vedisiri and Satisiri, and that she governed the kingdom
during the minority of Vedisiri. According to them the Kumdra
Hakusiri of the relievos is the Satisirimat of the inscription. 25 True,
in the Dravidian Prakrt of the Satavahana epigraphs ha sometimes
takes the place of sa, e.g. Hamgha^Samgha ; 26 Hiru-Hatakani
~ Siri-Satakani, Hala Sata. But nowhere is ku or ka used for ti.
Moreover one would expect Hakuhiru rather than Hakusiri. 27 This
seems also to dispose of Biihler's identification of Satisiri with
Hakusiri. 28 Further, Kumdra Sati has ' sirimato ' and not * siri '
suffixed to his name. As Kumdras Bhaya. . , and Satavahana
are not mentioned in the sacrificial inscription ; 29 and as between
the representations of Kumdra Bhaya. . , and Mahdrathi Trana-
kayiro 30 two statues and their inscriptions have disappeared, 31
Nayanika would seem to have had more than two sons; it would
seem that Vedisiri and Satisiri were represented in the relievo
figures now lost and that the sacrificial inscription, which mentions
only two princes (neither of whom is the eldest son, i.e. Kumdra
Bhaya), is posterior to the relievo figures and the inscriptions over
them.
24. Op. ci*., xix, n. 4.
25. Rapson: op. cit., xx, n. 3.
26. Liiders, List Nos. 1210, 1271, 1272, 1281, etc.
27. Prof. Suniti Kumar Chatter jee assures me that the change of 'Sati'
into 'Haku' is not possible.
28. ASWI, Vol. V, p. 68, n.2.
29. Biihler would identify the latter with Vedisiri ASWI, Vol. V, p. 68;
but Rapson is more cautious, op. cit., xlvi.
30. PI. I, No. 1.
31. BG, Vol. XVI, p. 611.
H.A.-5
34 EARLY HISTORY OF THE ANDHRA COUNTRY
SUtakarni : Meaning
Many a prince in the Satavahana line bore the name Satakarrji,
sometimes along with a metronymic and another name, and some-
times without one or both of them : Siri-Satakani I, Cakora
Satakar^i, Mrgendra Satakarni, Gautamiputra Sri-Yajna
Satakarni.
Whilst Rapson's view that sometimes Satakarni was used gener-
ally is correct, the example given by him, i.e., the Girnar inscription
where Satakarni must mean Pulumavi is, as we shall see, rather
unfortunate. Better examples are Sivamaka Satakani called simply
Satakarni in the Girnar inscription, and perhaps the Siri-Satakani
of some coins closely allied to Siri-Yaiia's coins by type and fabric,
as all the successors and immediate predecessors of the latter bear
personal names. The Satakani of the Naneghat relievos would
seem to have borne a personal name ending in * siri.' Satakani was
sometimes abbreviated into Sata, Sati, 32 Sa$a (Sada?) 33 and
Sataka. 34 Sadakana of the Chitaldoorg coins is a Prakrt form of
SatakarrLa. 35
The meaning of the term is, however, not settled. Rapson did
not attempt to elucidate it. The Puranic forms Satakarni, Sata-
kar^i, Svatikarni, Svati, 36 Svativarna and Santikarni show how
little the Puranic writers understood the meaning of the word
Satakarni in Prakrt. Prof. Jean Przyluski's suggestion that kani
32. El, Vol. XVIII, p. 318, t. 1. 3. Rapson: op. cit., PI. Ill, G. P.
and PI. VHI, G.P.3.
33. Rapson, op, cit.
34. ASWI, Vol. V, No. 24, Kanheri Inscriptions. According to Rapson,
Sadakana and Sataka may be forms of Satakanam (op. cii., Ixxxii). As
Sata is an abbreviation of Satakarni or Satakani, as the Banavasi inscrip-
tion of Haritiputra Visnuka<Ja-Cutukulananda Satakarni (IA, 1885, p. 331)
and the Malavalli inscription of a Kadamba king (EC, Vol. VII, p. 252 and
PI.) make it clear that more than one prince in the Cutu line bore the name
Satakarni, Sataka as a form of Satakani is more probable.
35. 'Sadakana' occurs in a clay tablet from Candravalli which was exhi-
bited at the Eighth Oriental Conference at Mysore (1935) . The reading is
mine. The tablet bears the Trisula emblem in the centre.
36. The Puranic Svati is possibly a mis-Sanskritization of Sati, which,
like Sata, is an abbreviation of Satakarni. Krishna Sastri remarks : "... .the
name -ending svatikarna is more likely to have been the origin of the later
fiatakarni than the fanciful 6ata-karna (the hundred-eared)." (El, Vol.
pp. 317-18).
THE EARLY SATAVAHANAS 35
is derived from the Munda kon meaning * son ', and that Sata is the
Munda 'Sadarh' meaning horse, is ingenious, but not convincing.
Long ago Mr. Coomaraswamy proposed to identify the Ntirruvar
Kannar of the Tamil Epic, Silappadikdram^ with a Satavahana
Satakarni. Since then the Pandits 'have sought to derive Satakarni
from Satakarna (Satakarnasya putrah Satakarnih) . So far as we
know Simuka, the father of Satakarni I, did not bear either
the name Satakarni. In all the Sanskrit inscriptions
in which the term occurs 38 we have Satakarni and not
Satakarni. True the Puranas spell the word sometimes with S and
sometimes with S. Vatsyayana in his Kdmasutra gives the from
Satakarni. 39 But the evidence of the inscriptions which belong to
the third and fourth centuries A.D. is conclusive. The mean-
ing given by the Tamil annotator cannot therefore be readily
accepted. Sata cannot be connected with Sdttan for the reasons
stated above. Satakarni would be the name of a descendant of
Satakarna. Satakarna is as curious a name as Kumbhakarna,
Lambodara and Jatikarna. 40 If we read the name as Satakarna it
may mean one with ' a sharp ear '.
It is not true to say that Satakarni is only the dynastic name
of both Satavahana and Cutu families. 41 It was also borne by
ministers and ordinary persons. In a Kuda inscription 42 a minister
bears the name Hala Sata, a contraction for Satakani. In Nasik
No. 3, the preparation of the plates or the cloth or the palm
leaves is attributed to a '-takani/ and the lacuna could have con-
tained one letter only. So (Sa)takani is most probable. 43
37 . Pp . 540-41 . He figures as a close ally of Sehguftuvan ; he is here
represented as being prepared to secure for Senguttuvan, a stone from the
Himalayas, out of which was to be carved a figure of Pattini.
38. (a) Daksin&pathapates Satakarner dvir api " the Girnar
inscription of Rudradaman, El, Vol. VIII, p. 44. 1. 12.
(b) " presubhis-Sdtakarny-adibhis .... "the Talagun^a
inscription of Kakusthavarman, El, Vol. Vm, p. 33. 1. 14.
(c) "Vasisthlputrasya Satakarnisya," Kanheri inscription of the
daughter of Mahaksatrapa Rudra, ASWI, Vol. V, p. 78, PI. LI.
39. Vide supra.
40. Vedic Index, q.v.
41. Rapson: op. cit., Index, V, p. 264.
42. CH, No. 18, p. 15.
43. Pace Senart who says (El, Vol. Vm, p. 70): "It is most improbable
that we should have to read S&takanind, and it would indeed be extremely
puzzling if this royal name were borne by a simple engraver."
36 EARLY HISTORY OF THE ANDHRA COUNTRY
Sdtakani's Imperial Position
The long record at Naneghat incised during the minority of
Vedisiri by the regent Nayanika mentions a number of sacrifices
performed. Among those mentioned are the Asvamedha, Raja-
suya, Agnyadheya. Anvarambhaniya, Gavamayana, Bhagaladasa-
ratra, Aptoryama, Angirasamayana, Gargatriratra, Angirasatri-
ratra, Chandogapavamanatriratra, Trayodasaratra, Dasaratra, and
some others as the lacunae would show.
Buhler supposes that these sacrifices were all performed by
Nayanika though he admits that "according to the Sastras, women
are not allowed to offer Srauta sacrifices and that those who per-
form such sacrifices for them (strzydjaka) are severely blamed ;
yet that seems hardly probable for in the sentence which ends with
yanehi yitham, 'the sacrifices were offered/ we have the impersonal
passive construction and the genitive rayasa, * of the king ' . . . . "
Prof. D. R. Bhandarkar objects : 44 " It is inconceivable that Naga-
nika, even as queen-regent, celebrated it (the Asvamedha) of her
own accord and to indicate her paramount sovereignty as
Naganika's husband Satakarni has been styled apratihatacakra, it
is proper and natural to suppose that it was he who celebrated the
sacrifice twice. What appears to be the case is that Satakarni it
must be, who carried out the sacrifices referred to in the epigraph,
and as all sacrifices are performed by Yajamanas along with their
consorts, Naganika has been associated with him." The fact that
Naganika's husband is called vira, sura, Dakhind (patha) pa (ti) ,
and apratihatacakra, and the words ' rdyasa . . . (ya) nehi yitham '
support Bhandarkar's conclusions. After 'caritabrahmacdriydya
dikavratasumdaya yana huta ', 'vano' appears, and after
'vano' there is a stop. It is therefore probable that Nayanika's
part is only the description of the sacrifices performed by her hus-
band, and we know that the record was incised after the death of
Siri-Satakani I. The epithets dhamadasa y kdmadasa, varadasa,
putradasa, if they apply to Siri-Satakani, would be another piece
of argument in favour of our conclusion. 45 .
44. IA, Vol. XLVH, p. 72, n. 11.
45. Biihler's reading 'aWagavaradayiniya' is uncertain. Neither the
a nor the na is certain.
The epithets 'putradasa varadasa' etc., cannot apply to Vedisiri as
his name is in a compound with matuya. Nor can they apply to SaMri-
matasa, as they are too far removed from it.
THE EARLY SATAVAHANAS 87
It would then be that Siri-Satakani I was a powerful monarch
and that most of the sacrifices were performed by him to com-
memorate the expansion of his empire of which we have evidence
from the coins. The Naneghat record is then the funeral oration
of a disconsolate wife.
Satakani I Kharavela Synchronism?
According to the Hathigumpha inscription, Kharavela, in his
second year, sent an army to the west disregarding Satakamni. The
army reached the Kanhabena river and struck terror into the
Musika capital or city. 46 K. P. Jayaswal and R. D. Banerji
have shown that the Hathigumpha" inscription mentions a Yavana
king Dimita who was forced by Kharavela's victorious invasion of
Northern India to retreat to Mathura. As he could only be
Demetrius I, who on his coins wears the head-dress made of ele-
phant's skin, and who would have come to the throne about 190
B.C., king Kharavela's reign would on this synchronism fall in the
second and third decades of the second century B.C. Satakani I
would, according to the chronological scheme adopted here, have
reigned between 200 and 190 B.C. As Satakani II would have come
to throne in 172 B.C. the synchronism of Kharavela and Satakani
II is as probable as that of Kharavela and Satakani I.
Hakusiri
An inscription on one of the pillars of the Cetiya cave at Nasik
mentions Maha-Hakusiri and his grand-daughter Bhatapalika,
daughter of the royal officer (amaca) Arahalaya and wife of the
royal officer Agiyatanaka. The early type (i.e. in low relief) of
the decoration of the fagade. the simple lotus-shaped capitals of the
pillars, and the proximity of the cave to that excavated in the reign
of king Kanha, stamp it as a very early excavation in the series.
Senart has pointed out that " if this Mahahakusiri is the same as
Kumara Hakasiri at Nanaghat, two generations would not be too
much to explain the difference in the forms of letters which exists
between our epigraph and the Nanaghat inscription." 47 Buhler 48
assigns the inscription to a very early period and supposes that the
46. Dutiye ca vase acitayita Sdtdkaihnim pachima-disam haya-gaja-
nara-radha-bahulam dam$am pathapayati Kanha-bemnft gatdya ca
senaya vitdsitam M(u)sika-nagaram. El, Vol. XX, p, 79.
47. El, Vol. VIH, p. 92.
48. BG, Vol. XVI, p. 608.
38 EARLY HISTORY OF THE AWDHRA COUNTRY
change in the characters of its alphabet is due not only to time, but
to the development of the 'Malwa and Upper India style.* The
fact that the grand-daughter of Maha-Hakusiri is the daughter of
one royal officer and wife of another makes it highly probable that
the Hakusiri of this inscription belonged to the royal family and
was therefore the Kumdra Hakusiri of the Naneghat relievos. But
since he does not bear the title of Rdjan which Satavahana kings
invariably do, we cannot subscribe to the view of Rapson and
Biihler that the Hakusiri of our inscription ascended the throne.
Satisiri
Satisiri mentioned as a son of Nayanika in the sacrificial ins-
cription was probably represented in the relievos between Kumdra
Bhaya , and Mahdrathi Tranakayiro and therefore before
Hakusiri. Rapson and Biihler think that he may be the Sakti
Kumara of the Jain legends. 49 It has been shown above that he
cannot be identified with Kumdra Hakusiri. In the e Vdyu and
Matsya lists the successor of Satakani I is Piirnotsantu (Purnot-
sanga or Purnotsarga also in Mt). As Sati and Sata (abbrevia-
tions of Satakani) were incorrectly Sanskritized into Santikarna or
Satakarna, 'Santu' may likewise be an incorrect Sanskritization of
Sati, (San. Sakti) ; in that case Satisirimato of the Naneghat
inscription would be the Purnotsantu of e Vdyu.
Skandastambhi
This king, the sixth in the list, is mentioned only in some ver-
sions of the Matysa Purana. 50 As will be shown below two or
three kings have to be added to the Puranic list; and the number of
kings is nowhere mentioned to have been more than thirty. 51 Pro-
bably some of the names in the first half of the list have to be
deleted as imaginary names mentioned to bring up the total to
thirty. Skandastambhi's existence may therefore be reasonably
doubted.
Satakani II
If the Puranic chronology may be trusted Satakani II ruled
for 56 years, the longest reign in the annals of the dynasty. From
Western India come certain square coins (potin and copper) bear-
49. ASWI, Vol. V, p. 62, n. 1; Rapson: op. cit., xx.
50. Pargiter: op cit., p. 39.
51. Ibid., p. 36.
fiAftLY SAf AVArfANAS 3$
ing partially the legends Rano Satakamnisa;* 2 some bear the device
of the springing lion, others that of an elephant with upraised
trunk. According to Rapson they bear a close resemblance in size,
shape and types to the coins of Eran (East Malwa). He cites the
authority of Cunningham according to whom while the coins of
Ujjain are invariably round pieces, those of Besnagar (according to
him the capital of East Malwa) and Eran are nearly all square. 53
The double line border with the fish and swastika symbols are
strikingly similar to the device supposed to represent the river
Bina on the coins of Eran. The Satakani of these coins would
therefore seem to have ruled over East Malwa. West Malwa, as
the Siri-Sata coins show, had already passed into Satavahana
hands in the reign of Satakani I. 54
The Satakani of the coins would seem to have been earlier than
Apilaka, eighth in the Puranic lists. And so he must be No. 6
of the Puranic lists. The angular ta of the Satakani coins brings
them near the Naneghat inscriptions. But too much reliance can-
not be placed on the results arrived at by a comparison of coin
legends and stone inscriptions. In the Satakani coin the ka has no
nail-head, and has a longer horizontal member than the ka of the
Apilaka coin. East Malwa which, according to Kalidasa's
Mdlavikagnimitra, was ruled by Agnimitra from Vidisa would
therefore seem to have come under Satavahana rule during the
reign of Satakani II, some time between 180 and 130 B.C.
According to Rapson the coins may be those of Gautamiputra
or some earlier Satakarni, 55 and the Bhilsa inscription that of the
time of Vasithiputa Vilivayakura, predecessor of Gautamiputra
Satakarni. 56 But the early forms of ka, da and ra 57 in the Bhilsa
Tope inscription place it long before Gotamiputa Satakani 's time.
The Vilivayakuras do not belong to the Satavahana line (at least
the main line). 58 Moreover, Rapson's conjecture that Vasithiputa
applies to the king is not supported by the plate.
Apilaka
The authenticity of the Puranic lists which mention Apilaka
as the 8th king has been proved by a large copper coin of this
52. Rapson: op. cit., PI. I, Nos. 5 to 12.
53. CAI, p. 95.
54. Rapson: op. cit., xcii.
55. Op cit, xcvi.
56. Ibid., xxvii, n. 2.
57. Buhler's Tables m.
58. Vide supra.
40 EARLY HISTORY OF THE ANDHRA COUNTRY
king from the Central Provinces. 59 Rao Bahadur K. N. Dikshit
reads the legend as ' rdn& siva-siri-Apilakasa.' The legend, I think,
should be read as ' rdno siva sirisa-ApSafcasa/ 60 the i sign is repre-
sented by two short strokes one vertical and the other horizontal
attached to the right arm of pa. The coin bears the device of an
elephant goad. What is above the elephant may be nandipada.
Dikshit remarks " on numismatic grounds the place of thisi
59. The coin which belongs to the Maha-Kosal Society was exhibited at the
Annual Exhibition of the Asiatic Society of Bengal by Rao Bahadur K. N.
Dikshit, Director-General of Archaeology, who has been kind enough to
furnish me with an excellent photograph of the coin (PI. I, No. 2).
About coin No. 3 in his catalogue of coins in the British Museum PI. I
xciv Rapson says: "It is inscribed with Brahmi characters which are
apparently of the same period, (i.e., Satakani I's) but which are too
fragmentary to allow of any satisfactory reading. All that can be said is
that the name of the king seems to have begun with Aja or Aji . No
form occurring in the Puranic lists suggests any very probable identifica-
tion, though it is possible that the curious name Apitaka or Apilava which ap-
pears early in these lists . . . may be a corruption of the name of this king."
It is no longer possible to hold with Rapson that Apilaka is a Puranic
corruption of a name beginning with Aja or Aji. We have a silver coin
from Mathura with the legends Ajadeva, and bearing the same symbols
as our coin, i.e., the swastika with ma attached to each of the four arms,
man standing, and representation of a river with fish swimming in it.
On palaeographical evidence this coin belongs to the same period as No. 3
in Rapson's Catalogue. (Mathura is the fmdspot of many unga coins),
and resembles in type and symbols the silver coin of Sumitra, identified
with Sumitra of the Har^acarita, a Suhga prince in whose kingdom Malwa
might have been included. (JBORS, 1934, PI. facing p. 5, No. 2 and the
following) .
It might be remarked that both the Satavahana and the Sunga
coins from Mathura bear the same symbols. Rapson thinks that the Aja-
coin (lead) is clearly connected by type with the potin coins of Siri-Sata.
Only future research can show whether the Satavahanas were indebted to
the 6uhgas or vice versa, or whether both were indebted to a particular
locality for these symbols.
60. From the numerous forms of the name in the Puranas, Pargiter
long ago chose Apilaka of c Vfiyu op. cit., p. 39, n. 45. See Plate IX,
No. 3.
This is an instance of an honorific prefix having a case ending in
Satavahana inscriptions and coins; the only other instance is afforded by
the legends on a coin attributed conjecturally by Rapson to Pulumavi II
(G. P. 3, PI. op. cit.), which should be read in the following order:
'samisa s(i)r(i) ( ).' The combination of sa and d in s& is also
curious, the only parallels for this being givadatt-tibhlraputrwya and
Abhirasy-esvarasenasya. But these occur in a Sanskrit inscription while the
legends on our coin are in Prakrt.
*HE &U&Y SATAVAfcANAS 41
ruler is more with the later kings of this dynasty than with the
earlier ones as indicated in the Puranas." 61 But the blank reverse
of the coin certainly attests its early age. The early forms of sa and
ra which are only slightly developed forms of those of the Siri-
Sata coins, and the primitive i sign (a short curve) stamp the coin
as an early one in the series. No doubt the elephant is better
executed but this is not without a parallel ; the lion on Satakani
IFs coins is better executed than that on Sakasena's coins.
His reign: Extent of his kingdom
The coin, like the inscriptions, bears witness to a growing
empire. In his time the Satavahana power would seem to have
extended as far north-east as the modern Central Provinces. It
is hazardous to build too much on the provenance of a single coin.
It is even significant that Satakani I and II struck potin coins, and
potin coins are found "exclusively in the Chanda District of the
Central Provinces." 62 The Puranas are agreed that Apilaka 63 ruled
for 12 years (dosa dve or dvddasa) and that he was the son of
Lambodara.
Hdla
From Apilaka to Hala (8th and 17th in the Puranic lists res-
pectively), we have a period of absolute darkness, and the Puranic
Satavahana kings between them are to us mere names. But it is
probable that fresh evidence like the Apilaka coin may not only
confirm the order in the Puranic list but also open a vista into the
period.
Hdla: Meaning of the name
The king is mentioned by his name in the Puranas, the Sapta*
satakam, LUdvati, Abhidhdnacintdmani } and Desindmamdld. 64 In
the last two works mentioned, Hemacandra considers Hala as a
variation of Salahana and Satavahana. In the Gdthdsaptasatl the
61. YB of the AS of B, Vol. I, 1935, p. 28.
62. Kapson : op. cit., cfxxxiii.
63. The Sanskrit form as given in the Puranas is the same as the Prakrt
form on the coins.
Like sand Siri-Pulumavi, Apilaka does not bear the name Satakani.
Unlike otfcj^ jS&tavahana kings he has the prefix 'siva' in addition to 'siri'
(&ya=:auspicioiis) .
, M. Abhidh&nacintdm&ni, V. 712.
: HA. 6
42 EAHLY HISTOfcY OF THE ANDHRA COUNTRY
king under mention is called Hala, and Satavahana. Rajasekhara
also calls him Hala and Satavahana. 65 Hemacandra gives Kuntala
and Cauricinda as synonyms of Hala. 66 But we know from the
Puranas that Kuntala and Hala are the names of two different
kings. On coins and inscriptions Satakani is abbreviated into Sata,
Sati, Sada and Sataka. No instance of the abbreviation of Sata-
vahana is to be found. Hala cannot then be a variation of Sata-
vahana. It can therefore be considered as a variation of Sata-
kani only. Sata and Sati are contractions of Satakani. The form
Hatakani occurs on coins; and la is sometimes used for ta.
Satakani as a personal name is borne by many a king in the line.
The reign of Hala introduces us to an epoch of literary acti-
vity. From inscriptions we know that the official language under
the Satavahanas was Prakrt. The works attributed to or to the
time of Hala show that the Satavahanas encouraged the use of
Prakrt in literature. In this respect they played a part opposite
to the part played by the Ksatrapas. Only in the inscriptions of
Uavadata at Nasik and Karla do we have a mixture of Sanskrit.
The Satavahana son-in-law got some Sanskrit from his father-in-
law, and his wife uses Sanskrit in her inscription at Kanheri. But
the official records of Gotamiputa and his son Pulumavi II at
Nasik and Karla are in pure Prakrt. 67
Gdthasaptasati (700 verses in seven chapters), an anthology
of erotic verses in Arya metre and in Maharastri Prakrt is said to
have been compiled by Hala. Hala evidently worked on the basis
of an earlier anthology by a certain Kavivatsala and unified and em-
bellished it considerably, retaining the names of the original compo-
sers in some cases, and adding other verses of his own. The work
must have undergone several changes at other hands in subsequent
times as its numerous recensions testify. But there is no doubt that
its kernel dates from the first or second century A.D. and that it
shows the previous existence of a considerable body of lyrical litera-
ture in Maharastri Prakrt. 68 In the mangala or introductory verse
adoration is paid to Pasupati or Siva. Although the verses are said
to have been composed by Hala alone, the commentator's notes men-
65. Detinamam&la, 8, 66.
66. Ibid., ii, 36, iii, 7.
67. It is however to be noted that the Naneghat record is not in pure
Prakrt as is often imagined. It is in mixed dialect, e.g., prajapatino (1.1).,
apratttiatacakrasa (1.2), bhdriyd, (1.4), and caritabrahmae&riy&ya.
68. Verses 3, 698 and 709. ed. Weber.
THE EARLY SATAVAHANAS 43
tion the following poets as contributors to the work: Bodissa,
Culluha, Amaraja, Kumarila, Makarandasena and Sriraja. 69
Verses from this work are quoted in Dhanika's commentary on
the Dasarupaka, in the Sarasvati Kanthdbharana and in the Kavya-
prakdsa. Bana evidently refers to this work when he says "Sata-
vahana made an immortal refined treasure (kosa) of song
adorned with fine expressions of character like jewels/' 70 Meru-
tunga in his Prabandhacintamani tells us of Satavahana of
Pratisthana who devoted himself to collecting the compositions of
all great poets and wise men; he bought four gdthas for forty mil-
lion gold pieces and had a book made which was a kosa of the
gathds that he had collected. 71
Events of his reign
Lilavati throws some light on the events of Hala's reign. A
theme in the work is the military exploits of Hala's Commander-in-
chief Vijayanaihda in Ceylon on behalf of his master. The king of the
Sringala dvlpa by name Silamegha had a daughter by name Lila*
vatl by his gdndharva wife Sarasri. She lived near Sapta Goda-
vari Bhimam which is identified with modern Draksarama. After
his military exploits Vijayanaihda camped with his troops at Sapta
Godavari Bhimam, and came to learn all about Lilavati. After
his return to the capital, he narrates the whole story to his king.
Hala then proceeds to the place, kills the demon Bhisanana and
marries Lilavati. After visiting the residence of her father the
count returned to Pratisthana in Svabhukti visaya.
69. Prfi. Spra. Sec. 13; Winternitz, Geichichte der Indischen Litteratur,
HI, pp. 97-103; Keith, History of Sanskrit Literature, pp. 223-5.
70. Harsacarita, Trans, by Cowell and Thomas, p. 2.
71. EHD, p. 241; JRAS, 1916, p. 819.
CHAPTER IV
THE LATER SATAVAHANAS
(a) Cotam/iputa Siri-Satakani
(i) Metronymics : Gotamiputa Siri-Satakani is the first
known Satavahana king to bear a metronymic. 1 If Satakani of the
Sane! inscription is a king later than the sixth in the Puranic lists
the early Satavahanas would not seem to have borne metronymics.
Nearly all the successors of Gotamiputa Siri-Satakani, known to
us through lithic records, bear metronymics. 2 It may be noted at
the outset that the Satavahana metronymics, like those in many
primitive communities, were apparently an institution for regulation
of marriages and not for descent of property, for the kingdom was,
as is shown elsewhere, transmitted in the male line. 3
The Satavahana metronymics are derived from Vedic gotras.
Gotamiputa means the son of Gotami or of a lady belonging to the
Gotama gotra. 4 Vasithiputa means the son of a Vasithi. Madhari-
puta means the son of a Madhari. 5 It has not been pointedly
emphasized by scholars, that the Satavahanas and their successors
in eastern Deccan, the Iksvakus, bear metronymics derived from
only the three Vedic gotras mentioned above. The Iksvaku records
offer an explanation for this curious feature. The institution of
cross-cousin marriages especially with the father's sister's daughter
was the cause. Occasionally a wife might be taken from a new
1. After the materials for a discussion of the views of Biihler and
Cunningham had been collected and presented by me, D. R. Bhandarkar's
criticism of the old theory appeared in Epigraphia Jndica, Vol. XXII, pp. 30 ff.
I have made, use of a few of his ideas.
2. The exceptions are Rajan Sivamaka Sada of an Amaravati inscrip-
tion and Pulumvi of the Myakadoni inscription.
3. Frazer, Totemism and Exogamy, Vol. HI, p. 321.
4. El, Vol. VIE, Nasik, No. 2.
5. The Satavahanas bear a personal name, or a surname, or both, along
with the metronymics : Gotamiputa Siri-Satakani, Vasithiputa Catfa Satakani,
Gotamiputa Siri-Yana, etc. We cannot accept Prof. Bhandarkar's remark
that he knows of no instance where the metronymic alone without the per*
sonal name is mentioned. In some Andher Stupa inscriptions (Liiders,
List, Nos. 680, 681, 682 and 683) Vachiputa and Grotiputa occur unaccom-
panied by a personal name.
THE LATER SATAVAHANAS 45
family, e.g., Vasisthlputra Sri-Satakarni married the daughter of
a Maharaja of Ujjain, evidently a Western Katrapa.
It is interesting to note that the feudatories and the successors
of the Satavahanas in eastern Deccan also bear metronymics. An
inscription at Karla belonging to the first century B.C., mentions a
Mahdrathi Gotiputa. 6 A Mahdbhoja of the first century A.D. bears
the metronymic Kociputa. 7 A Maharathi of the second century
A.D. bears the metronymic Vasithlputa while his father bears the
metronymic Kosikiputa. 8 Abhira Isvarasena of a Nasik inscrip-
tion bears the metronymic Madhariputa. 9 The Iksvakus bear the
Satavahana metronymics.
Even as early as Vedic times people bear metronymics like
Kausikiputra, Kautsiputra, Alambiputra, and Vaiyagrahapadi-
putra. 10 Pali canonical literature calls Ajatasattu a Vedehiputta. 11
But it is from Malwa that we get a good crop of metronymics. Two
inscriptions from Safici Stupa III and an inscription from Sat-
dhara Stupa II mention a saint Sanputa. 12 In another Safici ins-
cription an artisan under a Satakani is called Vasithlputa. 13 In two
inscriptions from the same place a Mogaliputa (Maudgalyiputra)
and a Kosikiputa, both of whom are Buddhist saints, are mentioned.
In a Barhut inscription of the second century B.C. a Vasithlputa
Velamita is mentioned. In another inscription, which begins with
'Suganam raje ' a king is called Gotiputa. 14 But his father
and son bear the Vedic gotra metronymics Gagiputa and Vatsi-
puta. A Besnagar inscription reveals a Kaslputa (Kasikiputra or
Kasiputra) Bhagabhadra, perhaps a ruler of Ujjain in the time of
Antalkidas. It is also worthy of note that some Pitalkhora cave
inscriptions of the third century B.C. mention a royal physician by
6. El, Vol. VII, Inscriptions from Karle, No. 2.
7. CTI, Vol. X, p. 17, No. 23.
8. El, Vol. VII, Inscriptions from Karle, No. 14.
9. El, Vol. VHI, p. 88 No. 15.
10. Vedic Index, Keith and Macdonell, q. v.
11. Barhut Inscriptions, Barua and Sinha, p. 2. Vaidehi means 'one
who belonged to Videha.'
12. Luders, List, Nos. 665 and 667.
It. Ibid., No. 346.
14. As has been suggested by Buhler Gotiputa is Sanskrit Gauptfputra,
'son of a lady of the Gupta race or clan.' Dr. Bhandarkar has suggested
that Kotiputa of a Sonari Stupa II inscription (Luders, No. 158) may be
derived from 'Kota' the name of a ruling family whose coins have been found
round about Delhi and in the eastern Punjab (El, Vol. XXH, p. 35) .
46 EARLY HISTORY OF THE ANDHRA COUNTRY
name Vachiputa (Vatslputra) Magila. 15 In the case of Malwa
metronymics we are thus able to trace three classes of metronymics
derived from (a) locality (b) race or clan (c) Vedic gotras; and
these seem to be used as indifferently as Satavahana and non-
Satavahana metronymics are used among Mahdrathis of west-
ern Deccan.
How did the institution of metronymics enter into the Sata-
vahana family ? As metronymics are borne by the later Sata-
vahanas and not by the early Satavahanas, surely it is not the
result of the change of father-kin into mother-kin. 16 To go further
we have only circumstantial evidence to depend upon. Like the
early Satavahans the early Mahdrathis bear no metronymics. 17 Later
Mahdrathis and later Satavahanas bear them, and the Mahdrathis
and the Satavahanas were matrimonially connected. The Maha-
talavaras who are feudatory nobles under the Iksvakus and are
matrimonially connected with the latter, bear Iksvaku metronymics.
The Iksvakus, originally servants of the Satavahanas, certainly got
their metronymics from the Satavahanas, for do they not bear the
Satavahana metronymics and are not metronymics quite foreign to
the Andhradesa? 18 The Abhira servants of the Satavahanas also bear
metronymics. From these facts two conclusions emerge. The feuda-
tory nobles under the Satavahanas bear Satavahana metronymics and
the Satavahanas and the Iksvakus are matrimonially connected with
their feudatories and with one another. In the case of the Mahd-
rathis it is not clear whether they gave it to the royal family, or
got it from them ; for, the first Mahdrathi to bear a metronymic
would on palaeographical evidence have to be assigned to the first
century B.C. 19 But, in other cases it is clear that the Satavahanas
transmitted their metronymics through the channel of marriage as
the Iksvaku records unequivocally show. 20 The question then arises,
15. CTI, Nos. 3, 4, 5, 6 and 7 ; PI. xliv.
16. "Whereas a system of father-kin once established is perfectly stable
never exchanged for mother-kin, the system of mother-kin is on the other
hand unstable being constantly liable to be exchanged for father -kin."
Frazer, Totemism and Exogamy, Vol. IV, p. 31.
17. Naneghat Inscriptions.
18. The dynasties that succeeded the Iksvakus in the Andhradesa do
not know metronymics.
19. Karla, rimhastambha inscription.
20. It has not been noted that while in many cases metronymics and
the father's name are mentioned (personal name or gotra name), e.g., in
the Barhut inscription referred to above, in all the later Satavahana records,
LATER SAtAVAHANAS 4t
could they not have got metronymics through the same
channel? It has been shown, that the inscriptions of an early period
in the Paithan region and in East and West Malwa, mention metro-
nymics. West Malwa came under Satavahana sway in the first
decades of the second century B.C., and East Malwa some time
between 175 and 125 B.C. It may be that marriages between the
Satavahana and Katriya noble families, more probably in Malwa,
gave thelse metronymics to the Satavahanas.
Btihler held that "the usage of calling sons after their mothers
was caused not by polyandria as some Sanskritists have suggested
but by the prevalence of polygamy, and it survives among the
Rajputs to the present day" and that the surnames f the Sata-
vahana queens which are derived from Vedic gotras and which
form the metronymics borne by their children were originally
the gotras of the Purohitas of the royal or noble families, from
which the queens were descended and kings were affiliated to
them for religious reasons as the Srautasutras indicate.
The title 'ekabamhanasa' applied to Gotamiputa Siri-Satakani 21
and the adoption of metronymics derived from Vedic gotras need
not mean that the Satavahanas were Brahmins. D. R. Bhandarkar
objects to Senart's translation of 'ekabamhanasa,' as 'the unique
Brahmana/ and adopts that of Sir R. G. Bhandarkar, i.e., 'ekabrah-
manyasya', 'the only protector of Brahmans' and the epithet 'eka-
brdhmanyasya* applied to Viravarman on the Pikira grant and to
Madhavavarman in his Polamuru grant makes Bhandarkar' s inter-
pretation more acceptable. There is, however, no difficulty in look-
ing upon even 'ekabrdhmanyasya' as an eulogistic expression.
That the Satavahanas were Ksatriyas is shown by the fact that
Gotami Balasiri styles herself as one who fully worked out the
ideal of Rajarsi's wife. 22 According to the Visnu Purana, 23 there are
three kinds of ?*$is, Devarsis, 'sages who are demigods also', Brafo-
the father's name is not mentioned at all. In Nasik No. 2 Vasithiputa
Pulumavi refers to his father in the expression pitupatiyo (t.1.11) but does
not mention his name. In the early Satavahana records, the father is men-
tioned, e.g., Naneghat inscription and the Nasik inscription of the grand-
daughter of Maha-Hakusiri. In the Nagarjunikon-tfa inscriptions, ladies
usually mention their mothers, brothers and nephews. Only twice is the
father's name mentioned (Inscriptions, H & L.)
21. El, Vol. YIII, Nasik No. 2.
22. Ibid* P. 60 t. 1. 10.
23. IE, 6, 21; Muir: Sanskrit Texts, Vol. I, p. 400n.
EARLY HISTORY OF THE ANDHRA COUNTRY
'sages who are sons of Brahmans' and Rajar$is, 'Ksatriya
princes who have adopted a life of devotion.' If the Satavahanas
were Brahmans it would be difficult to explain the absence of
gotra name in their early records especially in the Naneghat record
which gives an account of the sacrifices performed by a Satavahana
king and queen. 24
It may be pointed out that this theory is contradicted by the ex-
pression 'khatiya dapa madanasa* applied to Gotamiputa Siri-Sata-
kani in Nasik No. 2 itself. According to Prof. D. R. Bhandarkar 25
'Khatiya' (Ksatriya) refers not to the Ksatriya caste but to a tribe,
the Xathroi of Arrian, placed on the confluence of the Chenab and
the Indus, and the Ksatriyas of Kautilya, Manu Smrti and Ptolemy.
In his Girnar inscription Rudradaman refers to the reinstatement of
deposed kings and the defeats he inflicted on the contemporary Sata-
karni. If these were the descendants of feudatories of Nahapana
dethroned by Gautamiputra, might not 'khatiya' of the Nasik record
refer to the Ksatriya princes deposed by Gautamiputra? 26
The results of the foregoing discussion may be summarised as
follows : The Satavahanas were Ksatriyas and bore the gotras of
their mothers. They got this institution of tracing descent by
mothers through intermarriages with Ksatriya families in certain
localities. The system was one for the regulation of marriages and
not for descent of property. The system of cross-cousin marriages
explains the occurrence of only a few gotra names along with the
Satavahana names.
24. In Jain legends Satavahana is made the son of a Brahman virgin
girl and gesa. JBBRAS, Vol. X, p. 132.
25. El, Vol. XXII, p. 33.
26. In some recensions of the Visnu Purana, Sudra heads the list of
Andhra kings, instead of Simuka; this is obviously a mistake, for the
numerous sacrifices performed by Satakani preclude us from looking upon
Satavahanas as 6udras, for according to Manu, the Veda is never to be read
in the presence of a 6udra (iv, 99) and for him no sacrifice is to be per-
formed (iii, 78) .
Brahmans also bear metronymics. The Buddhist teacher Vasubandhu,
the son of a Brahman of the Kausika family, was named Bi-lu-ci Vatsa.
Bi-lu-ci was -his mother's name and 'vatsa' signifies 'son* (IA, Vol. IV,
p. 143). Two inscriptions from Malavalli (Liiders, List Nos. 1195 and
1196) mention Kosikiputa Siri-Nagadatta of the Kontfamana family and of
the Kauntfinya gotra and Haritiputa Kontfamana of the Kauntfinya gotra.
In a Nasik record (Liiders, List, No. 1131), a Brahman is called a Varahi-
putra.
THE LATER SATAVAHANAS 49
(ii) Gotamlputa Siri-Satakani and the Ksaharatas 27
In Nasik No, 2 Gotamlputa is spoken of as the destroyer of the
Sakas, Yavanas and Pahlavas and as the exterminator of the
'Khakharata-vasa'. The Jogalthembi hoard has brought to light
numerous coins of Ksaharata Nahapana, % of which have been
restruck by Gotamlputa Siri-Satakani. Nasik No. 4 and Karla No. 19
show him as engaged in conquests. 28 The Satavahana records at
Nasik and Karla show that the Nasik and Poona Districts, Akara,
Avanti, Kukura, Suratha and Anupa countries which, on epigra-
phical and numismatic evidence, would seem to have been included
in Nahapana's kingdom, were conquered by Gotamlputa Siri-Sata-
kani.
Ksaharata inroads into Satavahana dominions
According to tradition preserved in the Kalakacarya Katha-
naka, the Saka invasion of Western India and Ujjain took place
some years before the Vikrama era. 29 After some time the Sakas
are said to have been driven from Ujjain by Vikramaditya, only to
return in 78 A.D. The identification of the Sakas of the Kalaka
legends with the Ksaharatas is rendered difficult by the fact that
while the son-in-law of Nahapana is called a Saka, Nahapana and
Bhumaka are nowhere so called. Says Rapson, 30 "It is possible
. . . .that the Ksaharatas may have been Pahlavas and the family of
Castana Sakas." But that they were "of foreign, i.e. non-Indian
nationality is certain." 31 In later Indian tradition they might have
figured as Sakas, much like the Kusarias. But it is highly improba-
ble that the Ksaharatas are included in the 18 Sakas figuring in the
Puranas as the successors of the Andhra Satavahanas; these
eighteen Saka rulers are doubtless the kings of the Cabana line
who ruled up to the time when the earliest Matsya account was
closed according to Pargiter, C. 255 A.D. 32
27. Chaharada, Chaharata, Khaharata, Khakharata are various Prakrt
forms of Ksaharata.
28. The Nasik and Karla inscriptions of Usavadata, son-in-law of
Nahapana, the Jogalthembi hoard and Bhumaka's coins. Also Junnar inscrip-
tion of Ayama, minister of Mahdksatrapa Nahapana, Luders, List, No. 1145.
29. CH, Vol. II, xxvi, xxvii. Says Sten Konow, "I cannot see the
slightest reason for discrediting this account as is usually done,"
80. Op. cit., civ.
31. Ibid.
32/ Vide infra.'
H.A.-7 *
50 EARLY HISTORY OF THE ANDHRA COUNTRY
Bhumaka
The earliest known member of the dynasty of Nahapana is
K?atrapa Bhumaka, known to us from coins only, According to
Rapson, considerations of type and fabric of coins and of the nature
of the coin legends leave no room for doubting that Bhumaka pre-
ceded Nahapana. There is, however, no evidence to show the re-
lationship between them. But the forms of Brahml and KarosthI
letters on their coins make a long interval impossible. According
to Bhagavanlal Indraji, the fact that Nahapana' s coins are found in
'the coasting regions of Gujerat, Kathiawad and sometimes in
Malwa' proves the K?aharata conquest of Malwa from the Satava-
hanas and the establishment of Kaharata power in western India.
One coin of Bhumaka comes from Pukar near Ajmer. 33
Nahapana
On coins he bears the title r&jan, and in inscriptions those of
Katrapa and Mahdksatrapa. The inscriptions of his time, his coins
and his titles prove the extension of Kgaharata power over fresh
territory 34 including north Maharastra, the heart of the Satavahana
empire. The Satavahana power must have been confined to the
territory around and to the east of Paithan. We do not know how
far Nahapana succeeded in the south, though it seems probable that
portions of southern Maharastra passed under his sway and had to be
reconquered by Gautamiputra. The Periplus has preserved some-
thing of the Kaharata-Satavahana struggle in the statements, that
the Kingdom of Mambanes 35 (Nambanus Nahapana) began with
Ariake and that the Greek ships coming into the Satavahana port of
Kalyan were diverted to Barygaza. 36
33. Rapson: op. cit., p. 64; CAI, p. 6. PI. 1.4.
34. Vide supra.
35. JRAS, 1916, pp. 836-37. Kennedy says, "the MS is so illegible that
it is impossible to restore his (the ruler of Ariake's) name with any con-
fidence. It has been read as Manbaros, Mambaros, and Mambanos. Fabri-
cius says that only the final letters (Barou) are certain. Boyer proposed
to read Nambanos, and in an essay full of learning and acuteness identified
him with Nahapana " (JA, July-Aug. 1897, pp. 120-51) Kennedy
thinks that like Pandion and Kerebotros the name of the ruler of Ariake
may be a general designation. But Sandanes (Sundara) and Saraganes
(Satakarni) are personal names. Dr. Fleet has shown how Nahapana could
have been misread into Mambanes (JRAS, 1907, p. 1043 n, 2) .
36. Schoff : The Periplus of the Erythraean Sea, pp. 43 and $2,
LATER SATAVAHANAfe 51
(iii) Comparative Chronology of Gotamiputa Sdtakani
and Nahwpana
Prof. Rapson refers the dates in Nahapana's inscriptions (years
41, 42, 45 and 46) to the Saka era (78 A.D.). According to him the
evidence of Nahapana's coins restruck by Gotamiputa Siri-Satakani,
Nasik No. 2 and Nasik No. 4 issued from a victorious camp, and the
ascertained date (S. 72) of Rudradaman show that Gotamiputa was
the conqueror of Nahapana. "... it would seem improbable that
Nahapana's reign could have extended much beyond the last record-
ed year 46=124 A.D. Gautamlputra's conquest of Nahapana seems
undoubtedly to have taken place in the 18th year of his reign.
We therefore have the equation: Gautamiputra's year 18=124
A.D., or 124 A.D.+x. On this synchronism, on the recorded regnal
dates in the inscription of other Andhra sovereigns, and on the
known date 72=150 A.D. of Rudradaman as Mahak?atrapa rests at
present the whole foundation of the later Andhra chronology/' 37
Cunningham proposed to refer the dates in Usavadata's and
Ayama's inscriptions to the Vikrama era; Rapson objects to it and
quotes, "the empirical remark" of Kielhorn that "in the majority of
the Saka dates the term year is rendered by vara" and that
in "the inscription of the Western K$atrapas the word for
'year' everywhere is vara . . . . ", 38
It may be noted here that the unit figure in Karla No. 19 read
as 8, by Rapson has been read as 4 by Buhler. Senart thinks that
it may be any number between 4 and O. 39 A study of the inscription
from the stone itself and of an impression of it taken by me, shows
that 7 is more probable than any other figure. No doubt it would
then be a later form of the symbol for seven. But it is certainly not
8 as assumed by Rapson or 4 as read by Buhler since it is quite un-
like the symbols for 8 and 4. 40 This would show that Gautami-
putra's conquests were accomplished at least in part in or before the
year 17.
Since Rapson wrote, the Andhau inscriptions of Cabana and
Rudradaman's time (year 52) have been studied and edited by
37. Op. cit., xxvii.
38. IA, Vol. XXVI, p. 153.
39: El, Vol. VH, p. 65.
40. The symbol for 4 appears in the same inscription*
52 EARLY HISTORY OF THE ANDHRA COUNTRY
scholars like D. R. Bhandarkar and R. D. Banerji. 41 The Andhau
inscriptions shatter the theory of the Gotamiputa-Nahapana
synchronism. Scholars are not agreed as to whether the inscrip-
tions should be referred to the joint reign of Ca$tana and Rudra-
daman, or the reign of Rudradaman. In all the four inscriptions
we have: 'Kafio Castanasa Ysamotika-putrasa, rano Rudradamasa,
Jayaddmasa putrasa vasa 52 . . . ' Prof. D. R. BhMdarkar originally
thought that pautrasya (which the construction would not allow
us to insert) had been omitted. But later 42 he states that "Mr. R. C.
Majumdar of the Calcutta University has kindly offered the sugges-
tion that the date had better be referred to the conjoint reign of
Castana and Rudradaman," R. D. Banerji objects: "Apart from the
possibility of such an event in India, nobody having ever thought or
tried to prove conjoint reigns of two monarchs except Messrs Bhan-
darkar, there is sufficient evidence in the Andhau inscriptions them-
selves to prove that the author of the record was quite ignorant as
to the exact relationship between Chashtana and Rudradaman ....
the Andhau inscriptions are the only records known which mention
Chashtana or Rudradaman as Rajas and not as Maha-Kshatrapas,
The only possible explanation of this is that in a remote place like
Andhau on the Rann of Cutch the people were not aware of the
new titles of the new dynasty of rulers, titles on which Rudra-
daman set great store The cause of the absence of any word
or phrase indicating the relationship between Chashtana and Rudra-
daman now becomes clear." 43 There is many a weak link in this
argument. In the genealogical portion in all other inscriptions of
the Western Ksatrapas of the Castana line, the great-grandfather
is mentioned first, then the grandfather, then the father and then
the son. In the Andhau inscription the names of Cabana
and Rudradaman precede those of their fathers. One
cannot believe that the people of Cutch who knew the relation-
ship between Ysamotika who does not seem to have been even a
K$atrapa and Ca?tana did not know the relationship between the
latter and Jayadaman. On his coins Nahapana is known as 'rdjan*
only. Does it mean that the people were ignorant of his titles of
K$atrapa and Mahdksatrapa which he bears in the\ inscriptions
of his son-in-law and minister ? The objection to the joint rule of
Catana and Rudradaman does not seem to be well taken; for as
Rapson has observed: "Among the later Western Ksatrapas we find
41. JBBRAS, Vol. XXXHI; El, Vol. XVI, pp. 19ff.; A.R. for 1915.
42. IA, Vol. XLVH, p. 154 n. 26.
43. El, Vol. XVI, pp. 22-23,
LATER SATAVAHANAS 5ft
the father and son ruling concurrently as Mahaksatrapa and
Ksatrapa." This would explain why Jayadaman bears only the
title of K?atrapa. 44
If then Castana was a Mahaksatrapa in the year 52 (130 A.D.)
with Rudradaman as Ksatrapa, he must have been a Mahaksatrapa
during the reign of his son Jayadaman as Ksatrapa. As on the
testimony of the coins Castana was a Ksatrapa for some time, if
Nahapana's dates are referred to the Saka era, a three years' inter-
val between Nahapana and Castana is the utmost that can be postu-
lated. Even taking for granted that Nahapana was defeated in year
46 itself (124-25 A.D.) , we are led to the paradoxical conclusion that
a year after the rooting out of Khakharata race and the destruction
of the Sakas, Pahlavas and Yavanas, Castana was on the scene
avenging Saka defeat. 45 The Nasik record of Balasirl which men-
tions in such glorious terms Gotamlputa's conquests of Aniipa,
Akara, Avanti, Suratha, Kukura, Asaka and Mulaka would become
a record of a fleeting conquest. Was then the 'Sdtavdhanakulaya-
sapratisthapana' referred to in an inscription incised 25 years after
the event such a shortlived glory? 46 Ptolemy's (139 A.D.) state-
ment that Ozene was the capital of Tiastanes (Castana), and the
Andhau inscriptions which show that Cutch was in possession of
Castana and Rudradaman in 130 A.D., are clear proofs of the re-
establishment of Saka power in the lands between Malwa and Cutch
at least. 47 It has been pointed out by Prof. K. A. Nilakanta Sastri
44. Rapson's view that between the reigns of Cas$ana and Rudradaman
there was an interval during which there was no MaMksatrapa and that
this may have been the result of a defeat, is no doubt partly based upon the
fact that Jayadaman bears only the title of Ksatrapa.
45. The theory by Biihler in JRAS 1890 that Nahapana and Castana were
contemporaries was abandoned by him later. ASWI, Vols. IV and V.
46. Says Rapson, op cit., xxxvii: " Rudradaman's conquest took place
c. 150 A.D., and before the 19th year of Pulumavi. The inscription of
Balasri seems to be a record of glory which has only recently passed away."
47. R. D. Banerji (JRAS, 1917, pp. 286-87) not only holds with Prof. D.
R. Bhandarkar (JBBRAS, Vol. XXIII, pp. 68-9) that Rudradaman had
conquered all the dominions mentioned in the Girnar prasasti before . 52,
but also says that it would not have been possible for Rudradaman to con-
quer Aparanta (N. Konkan) without conquering north Maharastra (Nasik
and Poona districts). We cannot subscribe to these views. The former
lacks conclusive proof. The identification of Mulaka and Asmaka with
N. Maharastra, and the absence of their mention in the Girnar prasasti are
conclusive arguments against the latter.
54 HlARLY HISTORY 6F THE ANDHRA COUNTRY
that there is nothing in the inscriptions of Px4umavi's time to show
that his dominions had shrunk so much in their area as the
Andhau and Girnar inscriptions would show. The silver coin of
Vasithiputa [Hi]ru Hatakani (or ni), (Siri-Sataka$i) which
like similar coins of Siri-Yaiia would seem to have been current in
Aparanta, and the Kanheri inscription of Rudradaman's daughter,
prove that Aparanta was held by the son-in-law of Rudradaman, a
successor of Pulumavi.
The chronological scheme of Rapson requires that Vasithiputa
sami Siri-Pulumavi should be the son-in-law of Rudradaman and the
Satakarni of the Girnar inscription twice defeated in fair fight by
him. I was fortunate enough to trace in the Prince of Wales Museum,
Bombay, a silver coin of Vasithiputa Siri-Satakani; this closely
imitates as regards type, size and weight the silver coinage of
the Western Ksatrapas. And we have the Kanheri inscription of
Vasisthiputra Satakarni's queen, the daughter of Mahaksatrapa
Rudra (daman) . These make it difficult for us to identify a king
who, on coins and in inscriptions, is called Satakarnd, with Pulu-
mavi who does not bear the surname Satakani either in inscriptions
or on coins. He must therefore be a successor of Pulumavi. In
the Puranic lists Sivasri, the Vasithiputa siva Siri-Satakani of the
coins, figures as the successor of Pulumavi (perhaps his brother).
As 'siva 1 and 'siri* are honorific prefixes, no insuperable difficulty is
involved in the identification of Vasithiputa Siri-Satakani with
Vasithiputa siva Siri-Satakani. Prof. K. A. Nilakanta Sastri has
rightly pointed out that it is improbable that Pujumavi, who accord-
ing to Ptolemy was a contemporary of Castana, married the latter's
great grand-daughter.
The Satakarni of the Girndr inscription
As Rudradaman is said to have defeated Satakarni not distantly
related to him sambamdhdvidurayd (ratayd) , the Girnar Satakarni
is to be sought for in one of the successors of siva Siri-
Satakani, not far removed from the latter in point of time. In the
Puranic lists Sivaskanda Satakarni (the Sivamaka Sada of an Ama-
ravati inscription) and Siri-Yaiia figure as his successors. The diffi-
culty in identifying the Girnar Satakanji with Siri-Yana is that
Pujunifivi would then have to be placed between 86 and 110 A.D.,
while the contemporaneity of Castana and Pulumavi mentioned
by Ptolemy would bring Pujumavi to 130 A.D. at least. So the
Girnir Satakarni must be Sivamaka (Sivaskanda) Satakarni,
probably a brother or nephew of siva SrI-Satakan>i. Working
THE LATER SATAVAHANAS 55
backwards with the ascertained regnal periods of these Satavahana
kings we get the following scheme of chronology :
Gotamlputa Satakani ,. 82106 A.D.
Vasithlputa Pulumavi ,. 107131 A.D.
6iva Sri-Satakarni .. 132 145 A.D.
Sivamaka Sada .. 146153 A.D.
It now becomes still more impossible to place Nahapana
between 119 and 124 A.D.
Then how are the dates in the inscriptions of Nahapana's son-
in-law and minister to be interpreted? In 1908 R.D. Banerji revived
a point 48 made by Prof. Bhandarkar years before that U?ava-
data's inscriptions are palaeographically earlier than that of Sodasa
of the year 72, and added, *it is extremely probable that as Naha-
pana is prior to Sodasa the dates in his inscriptions refer to the era
in which the dates in the inscriptions of the Northern satraps are
dated.' Subsequently 49 he refers these dates to the regnal years of
Nahapana. Prof. K. A. Nilakanta Sastri has lately revived Cun-
ningham's theory 50 that Nahapana 's dates must be referred to the
Vikrama era. His arguments are as follows:
The letters of Uisavadata's inscriptions resemble those of the
inscription of the Northern Satrap Sodasa and Biihler admits that
in the former southern peculiarities are wanting. The discus, arrow
and thunderbolt on the coins of Nahapana and Bhumaka remind us
(as V.A. Smith has pointed out) of the coins of Hagana and Haga-
masa, the Northern Satraps. In EL Vol. XIV, Dr. Sten Konow has
referred the year in Sotfasa's record to the Vikrama era. Rev. H. R.
Scott has observed that the letters on the coins of Nahapana belong
to the near middle of the period from 350 B.C. to 350 A.D. ". . .the
state of Kharosthi on Nahapana's coins. . . .seems now to secure for
him a place distinctly earlier than Castana." The Usavadata bha,
va, and sa, a and ka with longer verticals than those of Sodasa a and
ka, the more angular ja, da with better curved back and the ornate
i sign, show that Usavadata alphabet was later than Sodasa alphabet.
Conclusions based on similarity of alphabets and the northern affi-
nities of Nahapana's family cannot be final; as Sten Konow has
observed, " We do not know for certain in which era the So^asa
inscription of Sam 72 is dated. I do not think it can be the same as
48. IA, Vol. XXXVH, p. 43.
49. JRAS, 1917, p. 285.
50. Op. cit., pp. 650# .
56 EARLY HISTORY OF THE ANDHRA COUNTRY
in the Taxila plate of Sam 78. I think that Patika, who issued that
record, is identical with the Mahakshatrapa Padika of the Mathura
lion capital which mentions Sudasa, i.e., So$asa as Kshatrapa." A
comparison of coin legends with stone inscriptions cannot yield safe
conclusions. The palaeography of the Andhau inscriptions renders an
interval of more than 160 years between the Girnar inscription and
the inscription of Usavadata, as Prof. Nilakanta Sastri would postu-
late, too long. Noteworthy are the form of a ka ra and pa which ap-
proach those of Usavadata alphabet. The state of Kharosthi on
Castana's coins does not point to a long interval between them, for,
as on Nahapana's coins, and on the silver coins struck by Castana as
Ksatrapa, Kharosthi is used to transliterate the Brahmi legends in
full. It is only on coins struck by Castana as Mdhdksatrapa that
Kharosthi shows decline it is used only for the genitive of the king's
name. This fact points to a rapid decline of Kharosthi rather than
to a long interval. 51
The development of the Brahmi alphabet in the Nasik and
Poona districts in the 1st and 2nd centuries A.D. makes a long in-
terval between Nahapana and Gautamiputra highly improbable.
According to Biihler there is a striking similarity between the al-
phabet of the Nasik inscriptions of Gautamiputra Satakarni and
Usavadata. 52 Prof. Nilakanta Sastri rejoins 'it is a similarity which
is only to be expected if the area of their location had been ruled
by the Khakharatas for some time before Gautamiputra recovered
it for the Satavahanas.' 53 The ftasik alphabet of Pulumavi's time
especially ca, ja, da, na, and no,, (with a slightly curved base) , ta,
bha, yo>, ha (with a notch at the left), which is more developed
"inan the alphabet of Gotamiputa Satakani's inscriptions, deprives
this argument of its force. The Nasik inscriptions of Gotamiputa
Siri-Yana Satakani only three generations later than those of
Gotamiputa Satakani show letters of the ornate, type. The Karla
inscription attributed to Gotamiputa Satakani is not only engraved
immediately below that of Usavadata, but also shows only slightly
developed forms over those of Usavadata's inscription (e.g., ta, pa,
bhi, va, ra and ma) . In the inscriptions of the time of Pulumavi
51. According to Rapson on the copper coins of Nahap&na, only the name
Nahapana in BrShml legend can be deciphered. It is uncertain whether
or not this was accompanied by an inscription in Kharostfil characters,"
Op. cit., cix.
52. IA., xxxiii, App. Ind. Pal. pp. 42-43,
53. JHAS, 1926, p. 652.
tttis LATER sAf AVAHANAS &
the ornamental type has appeared. 54 The quick and distinct deve-
lopment from Gotamiputa to Pulumavi of the alphabets at Karla
and Nasik does not allow us to place a long interval between Naha-
pana and Gotamiputa.
The Jogalthembi coins of Nahapana, more than two-thirds of
which are restruck by Gotamiputa, point to the same conclusion.
If Gotamiputa defeated a remote successor of Nahapana, we would
have found in the hoard, coins of Nahapana's successors restruck
by Gotamiputa. The coins of Bhumaka show, that among the
Ksa'haratas other princes than Nahapana, if they existed, would
have struck coins ; and there is more point in restriking the coins
of the vanquished ruler rather than those of a remote predecessor
of his.
The village of Karajaka which is granted by Gotamiputa to the
monks of the Karla caves is surely the Karajaka granted to them
previously by Usavadata. Nasik No. 4 records the grant of a field
in western Kakhadi, a field which was held by Usavadata. Buhler
and Bhagwanlal Indraji look upon ajakalakiyam in the phrase e ya
khetam ajakalakiyam Usabhadatena bhulam' as a Prakrt form of
ajakala (Sans, adyakala) and translate 'the field which has been
possessed by Rsabhadatta up to the present time.' Senart looks
upon it as the name of the field. 55
Liiders' No. 795 where Ajakalaka is the name of Yaksa makes
Senart's suggestion extremely probable. Prof. Nilakanta Sastri
denies that the Usavadata of No. 4, the possessor of a single field,
was the Usavadata, son-in-law of Nahapana. But it is not
impossible that Usavadata held all the fields in the village and
that Gotamiputa gave one field to the monks retaining the rest for
himself . **
54. El, Vol. VII, PI. Ill, No. 20.
55. El, Vol. VIII, p. 72.
56. It is, however, risky to look upon Usavadata as a rare name, and
identify all the Usavadatas of epigraphic records with the son-in-law of
Nahapana. 'Datta* as a name ending is very common and 'Rsabha* often
enters into the composition of names (El, vol. XVI, p. 24; Liiders, List,
Nos. 56, 69a). A 6ailrwa4i inscription (Sailarwfitfi is near Karla) men-
tions Usabhanaka native of Dhenukakata.
An inscription in a pillar of the Karla Cetiya cave records the dona-
tion of a pillar by Mitadevanaka, son of Usavadata from Dhenukakata.
According to Senart, the 'mita* in the name of the donor and that of IJsabha-
datta's wife (i.e., Daksamitra) supplies "a link which may perhaps connecf
H.A.-S
58 EARLY HISTORY OF THE ANDHRA COUNTED
The guiding points in the determination of Ksaharata chronology
are the short-lived reigns of Nos. 18, 19, 20 and 21 in the Puranic
lists of the Andhras (44-54 A.D.) and reference to Mambanes in
the Periplus ; the kingdom of Mambanes (corrected by Schoff into
Nambanus and identified with Nahapana) is said to have begun
with Ariake, which according to Lassen would represent the terri-
tory on either side of the gulf of Cambay. Scholars like K. P.
Jayaswal have challenged this identification. 57 But as Nahapana
is known to have ruled over Surastra and as the seaboard of the
kingdom of Mambanes was, according to the Periplus, Surastrene,
and as the capital of Nambanus, viz., 'Minnagara' also shows Scy-
thian or Saka associations (Min^ Scythian) this identification is
highly probable, and Schoff has shown that a consideration of the
authorities Roman, Parthian and Arabian fully supports 60 A.D.
for the Periplus, and Nahapana would seem to have been ruling
22 years before the accession of Gautamlputra Sri-Satakarni.
The question whether Nahapana's dates should be referred to
an era or to his regnal years, is not then so important as it might
be if we had no independent grounds to fix his date. The old view
that it is dangerous to assume the existence of eras other than the
Vikrama and Saka eras is not accepted by Dr. Sten Konow. 58 The
years may then refer to an era which started somewhere in the
closing years of the first century B.C., or in the beginning of the
first century A.D. Prof. K. A. Nilakanta Sastri has admirably
met Rapson's objection to referring the 'vosa dates' to any
other than the Saka era. He says, " it must be noted that
there are very few dates in these records, and even among these
few, we are unable to trace any consistency in the manner of
dating. An important inscription at Nasik is undated; only the
year and month is mentioned in another, though the day of the
month is also added elsewhere in the same inscription. The use of var-
sha for year does not seem to have been so rare in early dates that
are not in the Saka era." 59 R. D. Banerji's view that these dates
must be referred to the regnal years of Nahapana is supported by
them." The absence of the usual epithets applied to Nahapana's son-in-law
makes the Usavadata of this inscription a different person. The pillar
inscription and the Cetiya cave would seem to belong to the first century
B.C.
57. JBORS, 1932, p. 9.
58. CH, Vol. n, Introduction.
59. Op. cit., p. 656.
THE LATER SATAVAHANAS 59
tHe numerous coins of Nahapana and by the tradition preserved in
the Patpdvali Gdthas and Jinasena's Harivamsa which assign a
period 40 and 42 years respectively to Naravahana (a corruption
of Nahapana). 60 Rapson makes Nahapana a feudatory of the
Kusanas, on the strength of the mention, in a Nasik inscription of
Usavadata, of suvarnas which according to him must refer to the
gold currency of the Kusanas " which we must suppose to have
been current or prevalent in Nahapana 's kingdom." 61 But suvarna
as a coin was prevalent in India as early as the Vedic times; 62 the
suvarna under reference need not necessarily refer to the gold
currency of the Kusanas.
Prof. Bhandarkar 63 has advanced another argument for con-
necting Nahapana with the Kusanas. To him the Kusana of Nasik
No. 12, "appears to have been (the name) given to the silver coinage
of Nahapana, because he issued it for his overlord who must have
been known as Kusana i.e., Kushana." He continues " Was there
any Kushana king who was also known by the mere name
Kushana? Certainly this must be the Kushana sovereign referred
to in the Taxila scroll inscription of the year 136 I have
elsewhere shown that he can be no other than Kujula Kadphises, or
Kadphises I as he is also known." Apart from the objections to
kusana being a form of Kusana or Gusana, the sense of the passage
does not admit of Bhandarkar's interpretation. In line 2 civarika
and kusanamula are mentioned together as the purpose for which
the money was invested in a guild. In line 3 we have 'civarika-
sahasrdni be 2000 ye padike sate eto mama lene vasavuthdna bhi-
khunam vlsdya ekikasa civarika bdrasaka, yd sahasra prayutam
pdyunapadike sate ato kusanamula. 9 But though the evidence
cited for Kusana overlordship over Nahapana cannot be relied upon,
the chronological scheme adopted here makes Kusana overlords-hip
not improbable (Kujula Kadphises 50-75 A.D.).
60. Rapson remarks (Op. cit., ex) : "Pandit Bhagvanlal Indraji sup-
posed that the portraits of the Nahapana on the silver coins indicated a very
long reign; but now that a vast number of specimens are available for com-
parison, it is clear that no such conclusion can be safely drawn from these
representations of the king's head They cannot possibly have been
portraits, in the true sense of the word, of any single individual "
61. CIC, Andhras and Western Ksatrapas, clxxv.
62. Vedic Index, Keith and Macdonell, Suvarna Yajna q. v.
~$3. IA, Vol. XLVH, p. 76,
60 EARLY HISTORY OF THE ANDHRA COUNTRY
(iv) Gotamiputa Sdtakani cannot be identified with Gotamiputa
Vilivayakura of the coins
A number of lead and copper coins flrom Kolhiapur have
revealed to us the names of 3 princes: on the evidence of restruck
coins, the order would be as follows :
Vasithiputa Vilivayakura
Madhariputa Sivalakura
Gotamiputa Vilivayakura.
The current theories about these princes are far from being
satisfactory . Says Rapson, 64 Vilivayakura and Sivalakura " are
probably local titles in the dialect of the district of Kolhapur ....
The question remains whether these peculiar titles are to be re-
garded as the designations of members of the imperial Andhra
Dynasty or of viceroys governing the district of Kolhapur. Although
the question cannot be decided with certainty, the former view is
perhaps the more probable, since there is some evidence to show that
Andhra monarchs were known by different titles in the different di-
visions of their empire (cf. sup. 48, 50) ." Long ago Sir R. G. Bhan-
darkar said that Vilivayakura and Sivalakura are names of the
feudatories whilst the metronymics are those of their suzerains. 65
The metronymics attached to the names of Mahdrathis are not those
of their suzerains the Satavahanas. 66 The father of a Mahdrathi
who dates his inscription in the regnal years of Vasrhiputa sami
Siri-Pulumavi, bears a metronymic not borne by the Satavahanas
viz., Kosikiputa. *Kura' as a name appears in Bhatfiprolu. Sivala
as the name of a queen occurs in one of Barhut inscriptions, 67
and at Amaravati we come across an updsikd Sivala. La is a com-
mon ending in names, e.g., Bhadila, Bhayila, Dronala, Sivakhadila,
Buddhila, Sarpila, and Rudrila ; 68 Vilavanaka is the name of a
village. 69 Regarding Ptolemy's mention of Pulumavi of Paithan
and Baleokuros of Hippokura (identified by R. G. Bhandarkar with
Vilivayakura), Rapson remarks that his Siri-PuJumavi and Viji-
vayakura might be one and the same person. "A foreigner might
be excused for not knowing, that in our own country, the Prince
64L Ixxxvii-lxxxviii.
65. EHD, 1927, p. 30.
66. Karla No. 14, El, vol. VIII.
67. Cunningham, Stupa of Bharhut p. 131; Liiders, List, No. 1268.
68. Liiders, List, Nos. 125a, 149a, 1124, 1247, 1292 and 1054.
69. El, Vol. XV, pp. 41 and 43,
THE LATER SATAVAHANAS 61
of Wales, the Earl of Chester and the Duke of Cornwall were the
same person." But Ptolemy mentions the capital 'Hippokura* (a
'kura* ending in the name of the capital too !), which has been
identified by Bhagwanlal Indraji with the modern Go$abandar. To
accuse a contemporary of ignorance we require strong proofs. In
seeking to identify these princes with the Satavahana rulers Rapson
places too much reliance on metronymics, yet we know that the
Satavahana and Iksvaku metronymics were borne by their feuda-
tories, the Makdrathis and the Mahdtalavaras. Rapson's identifica-
tion upsets the order of Gotamiputa Satakani and Vasithiputa Pulu-
mavi among the Vilivayakuras; and Ma$hariputa Sivalakura has
no place between Gotamiputa Satakani and Pulumavi in the Pura-
nic lists. 70 Prof. Rapson himself points out that the Ujjain symbol
which is employed by all the later Satavahanas on all their coins
is not found on the Kolhapur coins 71 and that the ' bow and arrow '
on these coins is nowhere found on the Satavahana coins. To a
certain extent the type (the tree within railing) and size of these
coins (lead) resemble those of the Cutu and Makdrathi coins.
The pieces of evidence which make it probable that they were
feudatories of the Satavahanas are that one of the Vilivayakuras
was a contemporary of Pulumavi, that Gotamiputa's dominions ex-
tended as far south as Vaijayanti and that they used Satavahana
metronymics. The title 'raj an' does not prevent them from having
been feudatories, for Nahapana, who was undoubtedly a feudatory,
bears the title 'rajan 1 .
(v) Gotamiputa 9 s empire
To return to Gotamiputa, he would seem to have wrested from
the Ksaharatas not only the ancestral dominions, but something
more. The epithet 'Sdtavdhana-kula-yasa-patithdpana-kartf applied
to him is no idle boast, for before his reign, the Satavahana power
would seem to have suffered considerable loss of territories. Accord-
ing to Rapson, the countries which are mentioned in Nasik No. 2 as
having been under his sway, "in no way represent the extent of
his empire." "The names themselves are those of the kingdoms
which had submitted to Gautamlputra ". 72 Rapson is evidently
thinking of the Andhrade&a. Not all the conquered kingdoms
would seem to have been mentioned, e.g., S. Mahara?tra as far south
70. Op. cit., sd,
71. Op. cit., clxvi-clxviii.
72. Op. cit., xxxv, xxx vi.
62 EARLY HISTORY OF THE ANDHRA COUNTRY
as Vaijayanti (Nasik NQ. 4 speaks of the Vejayanti army as 'sen&ye
Vejayamtiye') . Even so the extent of his kingdom is indicated
by the mountains of which he is said to have been the lord the
western and eastern portions of the Vindhya range (Vijiha and
Paricata) , the Satpura hills extending through the middle of Berar
nearly into west Bengal (Achavata), the northern and southern
portions of the Western Ghal^s (Sahya and Malaya respectively),
There is then no epigraphic evidence to show that Gautarmputra's
sway extended over the Andhradesa. 73 The reference to his
chargers having drunk the waters of three oceans need not neces-
sarily mean that his kingdom extended from the Arabian Sea to the
Bay of Bengal. 74 The most important countries mentioned are
Asaka, Mulaka 75 (Northern Mahara?tra) Asika, Kukura (Eastern
73. A life-size statue of a person, in the Madras Museum Amaravati
collections, holding a lotus in the left hand, bears the inscription Gotami
nama (o*) . From the dress we can infer that it is not a statue of the
Buddha. Even supposing that the inscription records an adoration to Gotami,
the mother of Gotamiputa Siri-Satakani, it is possible that during the reign
of Vasithiputa sami Siri-Pulumavi (when the Amaravati Stupa underwent
alterations and perhaps enlargement) a statue of Gotamiputa Siri-Satakani
was set up; and from Nasik No. 2 (in which both grandson and grandmother
have eulogised Gotamiputa Siri-Satakani) we know that Gotami Balasiri
was living in the 19th year of her grandson's reign.
74. In Bana's Harsacarita (Trans. Cowell and Thomas) the Satavahana
contemporary of Bhikku Nagarjuna is styled "lord of the three oceans"
(p. 252).
75. In the Sona-Nanda Jataka (Fausboll, Vol. V, pp. 317, 24; and 318, 6),
Assaka and Avanti are mentioned together (Assakdvantim; Assakavantim ti
Assaka rattham va Avantiratfham va etc . ) . For the precise identification of
Asaka (Sans. Asmaka not Asvaka. Panini mentions Asmaka iv, i, 173) and
Mulaka a clue comes from an unexpected quarter. Verse 1011 of the Sutta
Nipata (Alakassa Patitthanam purimam) mentions Patifthana of Alaka and
verse 977 mentions Alaka as a country in the neighbourhood of Assaka (Asaka) ,
a country on the banks of the Godavari. (So Assakassa visaye Alakassa
samasane van Godavankule uncena ca phalena ca). But in three manu-
scripts (a MS in the Phayre collection in the India Office Library, a Burmese
MS in the Royal Asiatic Society, London, and a MS in the Mandalay Collec-
tion of the India Office Library) it reads as Mulaka or Mujaka. Mulaka is
made more probable by the following facts. Asaka and Mulaka are men-
tioned together in the Nasik No. 2. According to the Puranas Mujaka was
the son of Asmaka of the Iksvaku line. (Wilson, Vi$nu Purdna, p. 382).
Patifthana is modern Paithan on the banks of the Godavari; as Assaka is said
to have included Godavari there is no doubt that N. Maharastra, at least the
Nasik district and the territory around Paithan, is to be identified with
Asaka and Mx4aka; see also IA, Vol. XLVH, p. 150 f.n.
tHE LATER SATAVAHANAS &
Rajputana) Akara (East Malwa), AvantI (West Malwa), Sura^ha
(Suratra), Anupa (a district in the upper Narmada), Vidarbha
("the western part of modern Berar and the valley country west
of that") 76 and Aparanta (N. Konkan) . 77
(b) Pulumavi II. 78
Gotamiputa was succeeded by his son Pujumavi who according
to the inscriptions would seem to have ruled for 24 years at least.
The Puranas assign him a period of 28 years. Since they assign
only 21 years to Gotamiputa Siri-Satakani who from inscriptions is
known to have ruled for 24 years, it is probable that the Puranic
total 49 has to be divided between them as 25 and 24 or 24 and 25.
Messers R.G. and D.R. Bhandarkars' theory of the conjoint rule
of Gotamiputa Siri-Sdtakani and Pulumdvi
The theory of the conjoint rule of Pulumavi and his father,
which was proposed by Sir R. G. Bhandarkar long ago, did not
find a single supporter. As it has been restated by Prof. Bhan-
darkar as late as 1914 and 1918, a discussion is necessary. 79 Accord-
ing to Sir R. G. Bhandarkar the Puranic discrepancy in the matter
of the duration of the Satavahana dynasty (Mt. 460 years, Va. 411
years and Visnu 300 years) is to be explained "by supposing that
the longer period is made up by putting together the reigns of all
the princes belonging to the several branches of the Andhrabhrtya
dynasty ". 80 According to V. A. Smith this discrepancy arises on
account of some of the Puranas omitting the Sunga and Kanva
years (112+45) or the latter from the Satavahana total. 81 The
other arguments of Sir R. G. Bhandarkar repeated by Prof. Bhandar-
kar are:
76. Pargiter: Markanfaya Purdna, p. 335.
77. For a detailed discussion of these names BG, Vol. XVI; El, Vol. VIII;
Rapson: op. cit., and Bhandarkar 's EHD.
78. His name is spelt as Pulumavi on coins and in Nasik Nos. 3 and 20
and Karla No. 20; as Pulumayi in Nasik Nos. 1 and 2 and Karla No. 20;
and as Pulumai in Nasik No. 25. Pujumavi would seem to be the right
form to start from. Rapson thinks that Pu}umayi like Vilivayakura is a
name whose meaning is not clear.
79. JBBRAS, Vol. XXXIII, Epigraphic Notes and Questions', IA, 1918,
Dekhan of the Satavahana period.
80. EHD, p. 26.
81. ZDMG, 1902, p. 6,
& EARLY HISTORY OF THE ANDHRA COUNTRY
(a) In Nasik No. 2 dated in the 19th year of Pujumavi II,
GotamI Balasiri is called Mahdrdjamdtd and Mahdrdjapitdmahi.
This statement would be pointless if she were not both at one and
the same time.
(b) Gotamiputa is called 'Dhanakatasami' (Dhanyakataka-
svami) in No. 3 and Pulumavi ruled at Paithan. It has been shown
elsewhere 82 that Dhanakatasamanehi refers to Dhanaka^a monks or
samanas and that Dhanakata cannot be a form of Dhamnakataka.
In No. 4 Gotamiputa is called 'Benakatakasami.' Every queen is
the mother of a king and grandmother of a king. In the Naneghat
record Satakani I is eulogised and yet he was dead when the record
was incised. In the inscription Pulumavi makes over the merit
of the gift to his father in the expression 'pitupatiyo. 9 Such an
application of merit can be made only in favour of a deceased
person. 83 As Prof. Nilakanta Sastri has asked, 84 why should not
the queen who refers to her son as one living (jwasutdya) in No. 5
dated in the 24th year of Gotamiputa refer to him so in No. 2 ?
According to M. Dubreuil the inscription is the funeral oration of
a disconsolate mother.
(c) If it was a fact that Gautamiputra was dead when the cave
temple was dedicated and Pulumavi alone was reigning, we should
expect to find the exploits of the latter also celebrated in the ins-
cription, but there is not a word in praise of him. It is improbable
that a king who had been dead for nineteen years should be high-
ly extolled in the inscription and the reigning king altogether pass-
ed over in silence. It will be shown below that the cave with all
its cells was planned even during the reign of Gotamiputa, but
executed only in part. Balasiri completed the cave later and made
it equal to the cave of Usavadata, 85 son-in-law of Nahapana, whose
dynasty Gotamiputa had extirpated; then the train of thought sug-
gested to a mother explains this puzzling fact.
The arguments adduced by Prof. D. R. Bhandarkar are:
1. Since Pulumavi is a contemporary of Cas^ana, who died
before year 52 (130 A,D.) and since the dates in the inscriptions
of Usavadata and Ayama are to be referred to the Saka era, we
should postulate a theory of conjoint rule if we are not to run into
82. Chap. H.
83. El, Vol. Vm, p. 65.
84. JRAS, 1926,Th Later Sdtavtihana* and the aka*.
85. Nasik No. 2.
THE LATER SATAVAHANAS 65
chronological absurdities or impossibilities. Bhandarkar himself
has admitted that Cabana was living when the Andhau inscriptions
were incised; it has been shown above that the inscriptions of
Nahapana's reign must belong to a time before 100 A.D.
2. If cave No. 3 was granted in the 19th year of Pulumavi's
reign, and if in the year 24 his father Gautamiputra speaks of it as
his own gift, is not the conclusion irresistible that Gautamiputra
was living when the cave in question was made over to the Buddhist
monks, i.e., in Pulumavi's 19th regnal year, and that the year 24 of
the other inscription although it records a donation of Gautamiputra
must be referred not to his, but to Pulumavi's reign ? No. 2 dated
in the 19th year of Pulumavi records a non-official or private grant
by the queen, while No. 5 (dated in the 24th year) like No. 4 is an
official grant. What stamp No. 4 and No. 5 as official records are
the order of the king to the officer in charge of the district where the
object of the grant lay, and the mention of the formalities connected
with the grant, i.e., oral order, drafting, preparation of the charter,
preservation in the archives of the State and delivery. Nasik No. 2,
where neither the order of the king nor the formalities connected
with grants are mentioned, is then a non-official record. The grant
of the cave and lands by Gotamlputa mentioned in Nos. 4 and 5 must
be different from the grant of the same cave by Balasiri. Prof-
Nilakanta Sastri has admirably pointed out that all official grants
open with the order of the reigning king and end by giving the date,
while in all private records (at Nasik and Karla) the date is mention-
ed at the beginning. He concludes that Nos. 4 and 5 must therefore
be referred to Gotamiputa's reign. 86
But how can one cave be granted by two persons at different
times ? The statement that Balasiri made the cave equal to the
mansion in Kailasa and the plan of the cave give us the clue.
Gotamlputa who emulates Usavadata in his grant of lands to the
Nasik and Karla Buddhist monks would certainly have planned a
cave as beautiful, perhaps more beautiful than Usavadata's cave
(and in general appearance and arrangement the Queen's cave re-
sembles that of Usavadata). According to Prof. Nilakanta Sastri,
the raised verandah with a bench at the left end and two cells, one
at the right and the other at the left end, show that the verandah
with its two cells and the bench was a self-sufficient unit and was
86. Op. cit., p. 650. It may be pointed out that in private records of the
Ik$vaku period the date comes at the end.
H.A. 9
68 EARLY HISTORY OF THE ANDHRA COUNTRY
the Sivasri of the Puranas, we have to take it that by some process
which is not now traceable Sivamaka was transformed by the Pu-
ranic writers into Sivasri, and Rapson is inclined to identify him
with the SivaSri of the Puranas. Sivamaka as the name of Sata-
vahana is quite probable. In Kuda No. 6, Sivama as a personal
name occurs, 96 and in the Satavahana official records Siva often
enters into the composition of the names of officials under them.
Ka as a suffix to personal names is also common in the records of
the Satavahana period, e.g., Sivaka, Saghaka. 97
(e) Mddhariputa Sakasena gs Satakani
But for the find of coins bearing the legends Saka Sada (Saka-
sena Satakani, (vide supra), the inclusion of his name in the list
96. Dhanama in Nasik No. 25 (El, Vol. VIII), is another instance of a
name with a ma suffix. For the Kuda inscription see ASWI, Vol. IV.
97. Liiders, List, Nos. 1177 and 1189.
98. This king bears the title 'sami.' Though Gotamlputa Siri-Satakani
bears the title Benakatakasami his son Pulumavi is the first among the later
Satavahanas to bear the honorific prefix 'sami' in addition to the usual 'siri.'
Pulumavi's successors Madhariputa Sakasena, Siri-Yaiia Satakani and Catfa
Sati bear it. In the Chinna Ganjam inscription Siri-Yaiia bears the titles
of 'araka* and 'siri'. Therefore 'araka* would seem to be a prefix identical
in meaning with 'sami' (lord). Since Nahapana and his successors, princes
of the Castana line, regularly bear the title ' svamin ', it may be asked whether
the later Satavahanas did not borrow the prefix 'sami' from the Ksatrapas.
Since in a Deotek inscription a 'sami' (lord, king) addresses his official
(amaca) at Cikambari, we have to cry halt to such a conjecture. The eye
copy of the inscription compared with the estampage of the inscription re-
cently prepared by Prof. Mirashi, and exhibited at the Oriental Conference
at Mysore (1935) furnishes the following reading :
1. Sami amnapayati Cikambari-sa sa
2. hanamto bamdhamto va tasdradam kururadheva
3. Amacdna la-namnaga
4. Dato lego (kho) he pa I di 4 badho
(The eye copy is in Cunningham's Cn, vol. I, old series, p. 102
. and PI. XV)
In the paper read before the Conference (Proceedings pp. 613-22) the
Professor called it an Asokan inscription. But the later forms of ta ca, and da
(in dato), ya, and the angular pa, make it difficult for us to subscribe to the
Professor's views. Cunningham was nearer the truth when he opined that
the inscription was not earlier than the first century B.C. The inscription
would seem to be a Satavahana inscription for the following reasons :
In the first century B.C. and even earlier the Satavahanas were in
possession of East and West Malwa. The inscription is dated in the Sfita-
vahana fashion by the seasons and fortnights. It also begins like the later
THE LATER SATAVAHANAS 69
of our kings would have been rendered difficult, for do not his
metronymic and personal name draw him nearer to Ma^hariputra
ISvarsena, the Abhlra ? Even so he cannot be identified with any
of the kings of the Puranic lists; we have to depend on the palaeo-
graphy of the two Kanheri inscriptions of his time. Their alpha-
bet so closely approaches the alphabet of Siri-Yana at Kanheri that
it is probable that one closely succeeded the other. But we are
not in a position to determine who preceded whom. Whilst the
looped ta (11. 10, 12 and 13), and the rounded va (as opposed to
the triangular va in No. 14) , and the more cursive ha in No. 15 stamp
it as later than No. 14, the other letters stamp Sakasena's inscrip-
tion (No. 14) as earlier than Siri-Yana's (No. 15) . No. 14 which
was incised on the 10th day of the 5th fortnight of the rainy season
on the 8th year of the king, records the excavation of a cave by a
merchant and householder, the son of Venhunamdi an inhabitant of
Kalyana along with his father, brother and mother (Bodhisama).
The other inscription (No. 19) records the excavation of a cave by
Halanika, wife of the donor in the previous inscription.
Like Siri-Yana, Madhariputa Sakasena Satakani ruled over
both western and eastern Deccan. This fact also places him before
Canda Satakarni who would seem to have ruled over only the east-
ern Deccan. His coins bearing the lion device have been picked
up in the Krsna-Godavari districts. The 'sena' ending in his name
makes it probable that he was a son of siva Siri-Satakani, the son-
in-law of Rudradaman.
(f) Gotamiputa Siri-Yana Satakani
His relationship to siva Siri-Satakani and Sivamaka Sada can-
not be ascertained. Formerly Bhagwanlal interpreted the reverse
legends on his silver coins in such a way as to make him the son
of Catarapana." Prof. D. R. Bhandarkar interprets it so as to make
him the father of Catarapana. 100 " But there can be no doubt that
Satavahana inscriptions. The formalities connected with the grant are also
Satavahana (dato, lego and badho). To add to these there is the title 'sami.'
It is noteworthy that M&^hariputa Sakasena has only the prefix 'sami/ the
usual prefix 'siri' being absent. For this reason Bhagwanlal would read 's
for 'saka' in the inscriptions. But the second letter is only ka.
like i sign over the first letter in Burgess' impression may be only
dental stroke. The omission of 'siri' may be likened to the
'sami' in some of the inscriptions of Pulumavi II.
99. JBBRAS, Vol. XV, p. 306.
100. JBBKAS, Vol. XXm, p. 66.
70 EARLY HISTORY OF THE ANDHRA COUNTRY
this reading and interpretation are incorrect. Ite rest of the rev,
legend agrees word for word with obv. legend ; but all that can be
said about the doubtful word, which was read as ' Catur&panasa ',
is that it was one of five or six syllables, the last two only of which
are legible with certainty, and that it was probably the equiva-
lent to the first word of the obv. legend Raiio." 101
His inscriptions at Kanheri, Nasik, Chinna Ganjam, his silver
coins of Sopara fabric, his numerous coins from Kr?na and Goda-
vari districts, his potin coins from the Chanda district (Cen-
tral Provinces), clearly show that he not only maintained the
eastern possession but also wrested from the Western K?atrapas
Aparanta and North Maharaja. 102
The Chinna Ganjam inscription dated in the 27th year of his
reign shows that the Purar/as are probably correct in assigning
him a reign of 29 years.
In the Harsacarita Bana refers to a mendicant by name
Nagarjuna who was brought to Hell by the nagas ; he begged for
(a wreath of pearls) from the snake King as a gift and received
it. When he went out of Hell he gave it to a king, his friend, i.e.,
Satavahana, "the lord of the three oceans." Cowell and Thomas
say that the latter therefore ruled over Jambudvipa, Plaksadvlpa
and dmaldwpa. W2 The Nasik record of Balasiri makes it clear
that the three oceans or seas are the seas to the west, east and
south of the South Indian Peninsula. Hiuen Tsang refers to
Nagarjuna P'usa, a contemporary of the king styled Sha-to-p'o-ha
or Leading Light (Yin-Leng) . The latter quarried for him a
monastery on the mountain Po-lo-mo-lo-ki-li, 300 li to the south-
west of the capital of South Kosala. His date is variously given
as 700, 500 and 400 years after the death of the Buddha. 104 Says
Watters, "the names of the kings Kanishka and Kilika, of Vasu-
mitra, Asvaghosha, Katyayaniputra, Dharmagupta, and Rahula-
bhadra occur in the writings ascribed to Nagarjuna, and we may
with some probability assign him to the third century A.D." (we
may say even to the end of the second century A.D.) . Since the
probable date of Nagarjuna coincides with that of Siri-Yafia and
' ^JdjJL. Rapson op. cit, xd.
lOSfc-. If Madhariputa Sakasena came between Siri-Yafia and Slvamaka
Sad$ the credit for recovering the Aparanta would go to him.
103. p. 252 /.n. 1.
J04. Watters, On Yuan Chwang, ii, p. 204 f
tttfi LATOft SAtAVAHANAg ft
since Siri-Yafia would seem to have been the last great Sata-
vahana king to rule over eastern and western Deccan, he may
be the Satavahana contemporary of the P'usa Nagarjuna.
Ma^harlputa Sakasena has equally good claims.
(g) Vdsithtputa Cada Sdti
Prof. Rapson identifies Vasithiputa Siri-Catfa Sati of some coins
from the Godavari and Krsna districts with the Skandasvati
of the Puranas on the score that he is closely connected with
Pulumavi II by the type of his coins and metronymics ; 105
according to the same scholar Ca$a Sati of some lead coins is
probably the Candasri who occupies the last place but one in the
dynastic lists in the Puranas. It has been shown that Sivaskanda
Satakarrii of the Puranas is the Sivamaka Sada of the Amaravati
inscription. 'Cada' is a variant of 'Catfa' much like 'Ruda' and
1 Ru$a '. 106 The Kodavolu inscription dated in the regnal years
of Vasithiputa Cada Sati makes two Cada Satis improbable. 107
If the Puranic account can be relied upon, the Abhiras would
seem to have risen to power in North Maharastra 15 years before
the rise of the Ik?vakus, i.e., about 193 A.D. This makes it
probable that the western dominions of the Satavahanas were lost
during his reign. His Kodavolu inscription shows that during
his reign Kalihga or a part of it came under Satavahana sway.
(h) Pulumavi III
An inscription from Myakadoni (in the Bellary district)
recording the construction of a tank by a certain householder,
resident in the village of Vepuraka 108 under Gdmika Kumaradatta,
in the S[a]tavahani-hara under Mahdsenapati Khamdanaga, is
105. Op. cit., xl.
106. Rapson: op. cit., p. 46.
107. The inscription records a donation by an officer (amaca) and is
dated in the second year of Catfa Sati, the Cantfasri of the Puranas. What
is read as 'amacosa bhumivea' by Sten Konow should be read as 'amacasa
bhumikhasa'. Bhumika is perhaps the name of the amaca.*
108. Vepuraka may be tentatively identified with Virapuram in the Adoni
Taluq. Veparla in the Hatfagalli taluq and Virapuram in the Hacjagalli and
Rayadrug taluqs have good claims to be identified with Vepuraka inasmuch
as Hatfagalli would also seem to have been included in the Satavahani-hara
or Satahani raftha. The Chilla(le)rekakodumka of the ffira-Hatfagalli in-
scription of 6iva-Skandavarman is perhaps Chillakaladona in the Adoni
taluq.
W EAftLY HISTORY OF TftE ANDHRA COUNTftY
dated in the 8th year of a Siri-Pu|umavi of the Satavahana family
(rano Sdtavahananam s[i]ri-Pulum[a]visa) . Mr. V. S. Sukthan-
kar who has edited the inscription 109 has identified Siri-Pulumavi
with Vasi^hiputa sami Siri-Pulumavi, son of Gotamiputa Siri-
Satakani. The absence of the metronymic and the honorific prefix
'sami' which are always borne by the son of Gotamiputa Siri-
Satakani, and the alphabet of the Myakadoni inscription, which
approaches that of the Nagarjunikonda and Jagayyapefca inscrip-
tions, make it highly probable that the Pulumavi of this
inscription is the last of the Satavahanas in the Puranic lists. 110
True, the Puranas assign him a period of 7 years. But this fact
presents no insuperable difficulty. Gotamiputa Satakani ruled
for 21 years according to the Puranas. Yet we know from
epigraphic evidence that he ruled for at least 24 years ! Vasithi-
puta Catarapana Satakar/i ruled for at least 13 years ; yet the
Puranas assign him a period of 7 years only !
Since all Puranic lists stop with Pulumavi it is highly probable
that he is the last representative of the great dynasty.
109. El, Vol. XIV, pp.
110. Following V. A. Smith, Sukthankar makes out four Pulumavis in the
Puranic dynastic lists (Nos. 15, 24, 26 and 30 in the Matsya List). In fact
the Puranas mention only three Pulumavis . The line * Siva&rir vai Puloma
tu saptaiva bhavitd nrpah ' must be translated as : " after Pujumavi Sivasri will
be king seven years" and not as "Sivasri Pulumavi will be king 7 years" if
we accept the reading 'Pwlomdt tu' (Pargiter, Purflna Text of the Dynasties
of the Kali Age, p. 71, f.n. 18).
CHAPTER V
ADMINISTRATIVE ARRANGEMENTS IN THE SATAVAHANA
EMPIRE
Royal Succession
The Satavahana polity conformed to the rule laid down in the
Sdstras ; its keystone was the sovereign. Monarchy was hereditary.
Though the Satavahanas bore metronymics, though they do not
mention their fathers in their inscriptions, succession was always
reckoned in the male line. 1 The -expression 'kulapurisa-parapara-
gata ' in which the term ' purisa 9 implies, according to Senart, * des-
cent by males ', is corroborative evidence in the same direction. 2
Generally the eldest sons inherited the crown. It is remarkable
that though polygamy seems to have been the rule, we have no evi-
dence, either in the lithic records or in the Puranas, of disputed
successions. 3 During the minority of the Crown-Prince 4 , especially
in times of stress, succession passed on to the brother of the late
king. Kanha Satavahana, brother of Simuka, 5 the founder of the
dynasty, would seem to have come to the throne during the
minority of the latter's son. Sometimes the Queen-mother assisted
by her father acted as the regent and according to Buhler per-
formed some sacrifices 6 like kings.
The King,
The King was the commander in war and led his armies per-
sonally to the battle-field. The detailed instructions issued to
amacas (amatyas, governors of provinces) 7 bear testimony to the
1. The Purana texts and Nasik No. 3 make this clear; also among the
Mahdrafhis bearing metronymics, titles and office pass from father to son
(Luders, List, No. 1100).
2. El, Vol. VIII, p. 63.
3. Vide supra.
4. All princes are called Kumdras. The Pallava practice of calling the
Crown-Prince Yuvamahdrdja and of associating him in the administration
of the country is unknown to this period.
5. Kf^no bhrdtd, yaviydmstu as^ddasa bhavi?yati,
6. But it has been shown above that the sacrifices mentioned in the
Naneghat inscription were performed by Siri-Satakani.
7. Karla No. 19, El, Vol. VH.
H.A.-10
74 EARLY HISTORY OF THE ANDHIRA COUNTRY
effective control exercised by the king over officers in every part
of the empire; and kings were not merely content with
issuing orders. They took the necessary steps for realising their
intentions. It is no wonder that in a simple administrative
machinery as the Satavahana one, the king was powerful both in
theory and practice.
But the king was no capricious Sultan. As the guardian of the
social and religious order, his " fearless hand " was to be " wet by
the water poured out to impart fearlessness." He was to prevent
" the contamination of the four castes." The true father of his
people, he should " sympathise with the weal and woes of his citi-
zens " and " never employ taxes except in conformity with justice".
He was to be the " furtherer of the homesteads of the low as well
as of the twice-born ". 8 He should properly " devise time and place
for the triple object of human activity." A king educated in these
precepts among a moralising people would have been more than
human if he had escaped the obsession of this conception of his
duties. Moreover he was aided by ministers some of whom were
confidential ministers (visvasya amdtya) .
It has been shown that Dr. Bhandarkar's theory of dual
monarchy is unproven and improbable. 9 But sometimes kings asso-
ciated their mothers with them in the administration of their coun-
try. Nasik No. 5 10 (dated in the 24th year) is a joint order of
Gotami Balasiri and her son Gotamiputa Siri-Satakani,
to the officer in charge of the Govadhanahara.
Prof. Rapson attributes this arrangement to the failing
health of the King. It might have been so. It might also be, that
the absence of the Queen-mother's name in Nasik No. 4 (dated in
the 18th year) is to be attributed to the fact that it was issued from
a military camp in Govadhanahara, 11 whilst the other was issued
from the capital. If so, tne arrangement might have been due to
reasons other than ill-health. A Kanheri inscription speaks of a
confidential minister who executed certain works, and of the queen
8. El, Vol. VIII, pp. 61-62.
9. Vide, supra.
10. EI, Vol. vm.
11. According to Senart, the genitive Govadhanasa is better construed
with skandhdv&rdt than with Benakataka. 'The sequence of words would
then appear somewhat less regular; but the presence of another genitive,
senaye Vejayamtiye, may have caused Govadhanasa to be placed after
khamdhav&ra.'
ADMINISTRA-nVU ARRANGEMENTS 75
of Vasith!putra Sri-Satakar^i . Unfortunately the inscription is
mutilated and the missing words may show us the part she played
in the government of the empire. 12
Feudatories
(a) Petty Princes. The administration of the empire was
carried on by the agency of ordinary officials, and feudatories, i.e.,
petty princes, the Mahdrathis and the Mahabhojas. Kolhapur and
the district around it would seem to have been governed by a line
of princes with Dravidian associations (2nd cen. A.D.) . That they
were feudatories of the Satavahanas is made clear by the follow-
ing facts : Ptolemy places one of these princes in the time of
Vasithiputa sami Siri-Pulumavi ; their coins are found at
Kolhapur. Gotamiputa Siri-Satakani's empire included lands
as far south as Vaijayanti, and Kolhapur is north of Vai-
jayanti ; these princes bear Satavahana metronymics . Like the
Ksatrapas of the Khakharata and Cas^ana line they bear the title
of rajan. Rajan Cutukadanamda and Rajan Mu^anamda of the
coins from Karwar in North Kanara according to Rapson
belong to the same period as the Mahdrathi of the Chitaldoorg coins
(Ixxxvi), a period long before that of the Cutu kings of inscrip-
tions (Ixxxv) . From the fact that Rapson has included these coins
in the catalogue, it may be inferred that he considers them as
feudatories of the Satavahanas.
(b) Mahdrathis and Mahabhojas. Rapson and Senart consi-
der the derivation of the term Mahdrathi uncertain. 13 The analog-
ous titles, Mahdsdntanta, Mahdsendpati 9 Mahddandanayaka, leave
no doubt that the prefix mahd denotes an officer of higher rank.
The word 14 ' rathi ' connects it with the Ratthika of Asoka's and
12. No. 11, ASWI, Vol. V.
13. JRAS, 1903, p. 297 El, Vol. VII, p. 49.
. 14. The orthography of the inscriptions Naneghat, Kanheri, Be^sa,
Karla No. 2 (El, Vol. VII) and Chitaldoorg coins (except Bhaja No. 2)
(CTI) and Karla No. 14 (El, Vol. VII) is 'rathi' and not 'rathi.* E. Senart
thinks that thi is probable in Karla No. 2 and th probable in Karla No. 14,
1. 1; but an examination of the stones makes the thi certain in No. 2 and
very probable in No. 14, 1. 1. Since in the numerous inscriptions there
is not one instance of a mistake of tha for tha and Vathi' occurs more often
than 'rathi', the former is the proper form to start from.
The terms 'Rathi' (Rastrin) and Rathika (Rastrika) have the same
meaning. Also the elision of ka in 'Rathi' may be compared with the elision
of it in Bhoja of Asoka's edicts (the Bhojaka of the Hathigumpha inscription
of Kharavela).
76 EARLY HISTORY OF THE ANDHRA COUNTRY
Kharav-ela's inscriptions; what is more striking, the Mdh&rafJiis
are as much associated with the Mahdbhojas as the Rattfiikas with
the Bhojas (Bhojakas) , 15
The Mahdrathi and Mohabhoja inscriptions are in western India;
and the Ratthikas and Bhojas of Asoka's inscriptions are to be
sought for in western India. 16
Asoka's inscriptions mention the Ratthikas and the Bhojas in
the plural along with border peoples like the Andhras, the Pulin-
das, Kambojas and the Yavanas. These terms would therefore
seem to refer to tribes or peoples. Ratthika (Rastrika) means
" ruler " or " governor of a province." In the Anguttara Nikaya,
Ratthika implies a hereditary office. 17 The Ratthikas 18 and Bhoja-
kas of the Hathigumpha inscription would seem to be local chiefs,
since, when referring to Kharav.ela's conquest of them, it mentions
the smashing of their coronets, helmets, umbrellas, etc. insignia
of a ruler. 19 According to the Aitareya Brahmana and the Maha-
15. The Hathigumpha inscription of Kharavela (El, Vol. XX, p. 79 1. 6);
in Asoka's edicts the Ratthikas are not mentioned with the Bhojas, but the
Petenikas are mentioned with the Ratthikas in R.E.V. and with the Bhojas
in R. E. XIII; see also Betfsa No. 2, CTI; and Kanheri Nos. 15, 24, 29.
ASWI, Vol. V. The proximity of the Thana and Kolaba Districts, which
would seem to have been held by the Mahabhojas, to Poona and the surround-
ing districts, -held by the Mahdrathis, should also be noted.
16. Unlike the Andhras and the Pulindas, the Ratthikas, Bhojas and
Petenikas are called "the western borderers" (R.E.V. Girnar, t. 1. 5; Sah.
t. 1. 12).
, 17. The Pancakaihmap&ta in the Nivarana-Vagga mentions as Mahanama
Kulaputtas, annointed Kings, Pettanika-Ratthikas, Sendpatis, Gdmagdmani-
kas and Pugagdmanikas . D. R. Bhandarkar is of opinion that 'Pettanika,'
which is explained in the commentary (Indices, Vol. VI) as 'Pitard dattam
sapateyam bunjati/ is an adjective qualifying Ratthika. Pettanika-Rat-
jhika therefore means ^hereditary Ratfhika.' Yadiva which separates Senfi-
pati from G&magdmanika and the latter from Pugagamanika does not sepa-
rate Ratthika from Pettanika. This is the only argument in favour of his
view though he does not mention it. He is also of opinion that Ratffiifca-
Pitinikesu and Bhoja-Pitinike?u of Asoka's inscriptions also mean 'among
hereditary Ratfhikas* and 'among hereditary Bhojas.' The parallel cases of
Amdha-Pulidesu and Yona-Gamdh&ra-Kambojesu make such an explana-
tion improbable ; and in &h V., we have Rastikanam Pitinikanam.
18. These coupled with the fact that in Asoka's inscriptions Pitinika is
the regular form makes the identity of the Pitinika of Asoka's inscriptions
with the Pettanika of Anguttara Nikaya problematical.
19. Liiders translates ' sava Rathika-Bhojake ' as 'of the Provincial and
local chiefs'.
ADMINISTRATIVE ARRANGEMENTS
the term Bhoja denotes a prince : 2 In a Mab
tion a Kanabhoa (Kanabhoja) is called a Kum&ra, a titli
princes (CTI, No. I). If 'bhojaka' of the Hira-Ha<
can be taken to mean "free holder" it can by a stre
meaning, imply a local ruler or chief. 21 In the Mahdvam,
Village headman.' It is probable that these titles have
cal or ethnical meaning. For all that we know, it may be
ethnical meaning started from the official title.
The important place assigned to Mahdrathi Tranakayiro father
of Nayanika, wife of Satakani I, 22 (2nd century B.C.) , a place just
below that of the Crown-Prince and above that of the two younger
princes may give the clue to the origin of the title. In the days of
their service under Mauryan suzerains, the predecessors of Sata-
kani I must have been in a close alliance with the powerful Rathikas
of the west, a source of strength for them ; they would have en-
listed Rathika help in their coup d'etat. The coup d'etat being
successful, the Rathikas would have been given a higher title and
status, but they had to exchange Mauryan suzerainty for the Sata-
vahana. The silken bond of marriage which made and unmade
empires in Mediaeval Europe must have been forged to strengthen
Satavahana imperialistic position.
20. Ait. Brah. VIII, 12, 14, 17.
Santi Parvan, chapter LXVIII,
Raja bhojo virat samrdt
ksatriyo bhupatirnrpdh ]
ya ebhih stuyate &abdaih
kastam ndrccitumarhati \ \
21. In the Ait. Brah. one who is installed on the thron^LWj\he gakg
of enjoyment (bhojydyaiva) alcne is called a Bhoja.
22. Biihler's (ASWI, Vol. IV, p. 60, n. 3) restoration of
Mahdrathi (in the long inscription) as '[b#]Zfi3/a* is open to qu
ral usage requires that [bd]ldya' which again should be
mistake for balikaya, should follow Mahdrathino. According to Prof. Rapson
(JRAS 1903, p. 238; and op. cit., xx) the Chitaldroog Mahdrathi coins make
the restoration of the word as '(Kala)laya' probable. Then the
long record does not show the Mahdrathi' s relations with the royal family.
Even so, the Mahdrathi of the long record is identical with Maharathi Trana-
kayiro of the relievos, which show him in the company of two kings, one
queen and three princes, apparently in the order of precedence after one oi
them and before the other three. He could not have been a Brother ol
Satakani I, for he is not called a Kum&ra. He could not have been a
minister only, for ministers have no place in the royal family. The lauda-
tory epithets used by the Queen towards the Maharathi following closelj
those used towards her husband, could only represent an outburst of filia
78 EARLY HISTORY OF THE ANDHRA COUNTRY
The titles of Mahabhoja and Mahabhofi might have had a
similar origin. But since they occur in the Ku or Be^sa inscrip-
tions which yield no date or point of contact with any known dy-
nasty, it is not easy to determine the period at which they came into
existence. That they existed under the Cutus is certain. 23 As the
ornamental alphabet of the Mahabhoja inscriptions at Kuda 24 is
found also in the approximately datable inscriptions of V&sithiputa
Siri-Pulumavi or his time, 25 and of the minister of the Queen of his
successor Vasithiputa Siri-Satakani, it may be ascribed to the second
century A.D. True, local influences may have played their own
part. Two Mdhabhoja inscriptions from Kuda 26 do not exhibit
the ornamental variety and are earlier than Kuda Nos. 1 and 9.
The primitive form of the dental da (open to the left), and the
ornamental treatment of medial i and u signs, of the lower end of
the verticals of ka and ra and the upper end of the verticals of ha
and la and finally the rounded bottom of ma and la, stamp
No. 19 as very early in the series. No. 17 with its
somewhat angular ma and cursive da, which occurs in later inscrip-
tions represents a transition to the ornamental alphabet. An inter-
val of two generations between the alphabet of No. 19 and the
ornamental alphabet may, therefore, be safely postulated. The
office and title of Mahabhoja, then, came into existence not later
than the 1st half of the first century A.D.
Nature of the titles: Mahabhoja
The Bhdgavata Purana gives the meaning * great prince ' to
Mahabhoja. 21 Since no Mahabhoja inscription is dated in the
fashion in which kings' inscriptions are generally dated, it is certain
that they were not independent rulers; 28 and it is very difficult to
separate the title from the feudatory titles Maharathi and Mahfi-
sdmanta. Like all feudatory titles, the title Mahabhoja is also a
23. Luders, List, Nos. 1021 and 1186.
24; CTI, Nos. 1 and 9.
25. El, Vol. VH, Nos. 20 and 22.
26. CTI, Nos. 17 and 19.
27. Petersburg Diet. (q. v) .
28. A title originally applied to princes can become a feudatory title.
The title Mahdrfijn, which in the Gupta inscriptions is associated with the
feudatory titles, Mah&admttnta, Mdhapratihara, Mdh&senapati and MaK4-
dandan&yakd, is an instanee to the point (CII, Vol. HI, pp. 252, 289, 290,
and 236, n.) .
ADMINISTRATIVE ARRANGEMENTS 79
hereditary one 29 and became fixed in a few families or clans. 30 That
the Mahdbhojas were feudatories of the Satavahanas is proved by
the fact that there could not have been a rival power in the western
Deccan in the first century A.D. (Kuda where most of the Maha-
bhoja inscriptions are found is only 150 miles from Karla which was
certainly included in the Satavahana empire), which could have
claimed their allegiance; they are moreover related to the Maha-
rathis (feudatories of the Satavahanas) by family ties and were
the feudatories of the Cufus who succeeded the Satavahanas in
northern Mysore and parts of the western Deccan. 31
29. Kutfa, Nos. 1 and 9, CTI.
30. Marhdava and Sadakara (kera) . What is read as .Sadakara (in
Ku4a No. 19) may well be read as Sadakera. When we remember that in
the same inscription as well as in Kuda No. 9, CTI, the a and e signs are
very short strokes (e.g. Vijayanikaya and lena) and that sometimes the
sign for a is the o sign mdlakarasa Kuda, No. 16 (CTI), we may treat what
appears as a nail head over ka in Sadakara as the sign for medial e. Con-
sidering the da in Sadakera and Sadagera (the mas. form of Sa<JagerI in
Kuda Nos. 1 and 9, CTI) and the use of ga for ka (the Soparaka of Nos. 988
and 1095 is called Soparaga in No. 995 Luders, List, one may equate Satfa-
kera with Sadagera. A lady of the Satfagera family bears the name Vijaya.
The daughter of Sadakara Sudamsana bears the name Vijayanika, a variant
of Vijaya. Do not these names also suggest the idea that they belonged to
the same family circle? Pandit Bhagavanlal Indraji remarks (CTI p. 15):
"Vijayanika is apparently the same as the Vijaya of Nos. 1 and 9 : the
epithets Mahabhoya and Sadakara applied to her father here (No. 19) cor-
responding with the feminine forms Mahabhoji and Sadageri applied to
Vijaya". But it has been pointed out already that Ku<jla No. 19 is sepa-
rated from Kuda Nos. 1 and 9 by at least two generations.
According to the same scholar, Mamdava may designate either the gotra
name Mandavya or the title Mantfapa 'lord of a town called
Manx&apa.' Says he: "This latter seems to be the preferable explanation
as Mantfapa is the name for towns all over India, and three
small villages called Madad or Manda^h i.e., probably Mandapagatfh
lie close to Kutfa". (p. 4). But the use of the dental d in Madavdnam and
Mamdava renders either explanation problematical. Moreover ' Mamdava"
s&misa* and not 'Mamdavasa can mean 'lord of the town of Mamdava'. If
Mamdava should designate the country, the cognate inscription would lead
us to expect Mamdavakasa. In only one inscription have we Kalianasa
for Kalianakasa (Luders, List No. 1179), but it is easily understood as a
scribal error. The occurrence of 'Madavana' (Mamdavandm) in Kua No. 14
CTI, coupled with the fact that in several instances (Junnar Nos. 5 and 6,
the Jangli Gwr\4u inscription of Pulumavi and the Pallava and Kadamba in-
scriptions) the proper name of a person is preceded by the name of the
title or family to which he belongs in the genitive plural, makes it certain
that 'Mamdava' is a family name.
81. Luders, List, Nos. 1021 and 1186.
80 EARLY HISTORY OF THE ANDHRA COUNTRY
It has been generally supposed, that the title Mahdbhoja is
exactly co-ordinate with that of Mahdrafhi. But the Mahdbhojas
seem to have enjoyed more independence than the Mahdrafhis, for
unlike the Karla Maharathi inscription, no Mahdbhoja inscription
is dated in the regnal years of a Satavahana king; a Ku$a inscrip-
tion 32 comes very near to being dated in the years of a Mahdbhoja
(Mahdbhoje Mamdave Kochipute Velidate). Whereas Senart has
said that title Maharathi cannot imply a title of nobility superior to
that of Mahdbhoja, 33 we can on the strength of the evidence cited
here go further and say that the title Mahdbhoja was superior to
that of Maharathi.
Mahdbhoji
Like the title Maharathi, Mahdbhoja became a title applied
even to women. The title Mahdbhoji is borne only by the wives of
Mahdbhojas and not by the daughters also as Bhagwanlal thought. 34
In Kuda Nos. 1 and 9 (CTI) Mahdbhoji Sadageri Vijaya is men-
tioned along with her son Mahdbhoja Khandapalita to the exclu-
sion of her husband's name ; this may go to show that, like some
of the Satavahana queens, the Mahdbhojis sometimes shared poli-
tical power with their sons. There is so far no evidence to show
that a Mahdrathini ever enjoyed such a position or influence.
Mahdrathis
What stamp the Mahdrathis as feudatories are the fact that
they were hereditary governors of provinces, 35 and the rank and
32. CTI, No. 23.
33. Senart says (EJ, Vol. VII, p. 50, n. 4): "in this instance (Betfsa
No. 2 CTI) Mamdavi precedes Mahdrathini. Seeing that Mahdbhoja always
precedes either attribute when connected with it, this position does not seem,
to indicate that Maharathi could imply a title of superior nobility, and con-
sequently still less that it could designate a very high dignity." But it is a
correct view based on wrong premises. Not much can be based upon
Mahdbhoja preceding Maharathi as it is Mahdbhoja-Bdlikd that precedes
Mahdrathini, and in a Banavasi inscription, (Luders, List No. 1186) Mafwz-
bhuvia (Mahdbhoji) precedes Maharaja. As for Mamdava the donor in
Be^sa No. 2 (daughter of a Mahdbhoja and a Mahdrathini) might have
combined the Mahdbhoja practice of mentioning the family name after the
feudatory title with the Mahdrathi practice of mentioning it before the
feudatory title (Karla No. 14, El, Vol. VII).
34. The daughters of Mahdbhojas are, however, referred to as MahdbKoja-
bdlikds. Mah&bhuvl in a Banavasi inscription is either a mistake or a variant
of Mahdbhoji.
35. " .. whatever the derivation of the term may have been," says Prof.
Rapson (JEAS, 1903, p. 300) "such an expression as Okhalakiydnam Mahd-
(Karld No. 14) shows conclusively that it denoted the governor over
ARSANGEMENtS 81
power enjoyed by them, a rank and power far superior to those
of ordinary governors or amacas in charge of districts. Whilst
amacas make grants of lands and villages to religious bodies under
the explicit commands of the sovereign (the detailed instructions
issued to them by kings would seem to have left no room for their
discretion) , the Mahdrathi, like a feudal vassal, grants villages with
the fiscal immunities attached to them, in his own name. 36 That
they were feudatories of the Satavahanas is shown, as was point-
ed out by Biihler, by Karla No. 14, which is dated in the regnal
years of Vasithiputa Siri-Pulumavi. It is not, however, known
whether they had the right of waging war with one another. But
their semi-independence is shown by the absence of any reference
to their suzerains in their inscriptions (except Karla No. 14) . The
Chitaldoorg Mahdrathi coins make it very probable, that during
a part of the kingdom." Senart has shown (El, Vol. VH, p. 50) reasons
for abandoning this construction and making Okhalakiydnam depend upon
Somadevena; it would then denote not the people over which the Maha-
rathi ruled, but the tribe or family to which he belonged.
Senart further saysr "the occurrence of the feminine Mahdrajhini in
Be$sa No. 2 also indicates rather that the term does not imply the actual office
of governor of a district or province, but an honorific or nobiliary title." But
in a Nasik inscription of Siri-Yana Satakani's reign (No. 24, El, Vol. VIII)
the wife of a Mahasendpati is known by her husband's title. From the
Jahgli Gundu inscription of Pulumavi Ill's reign, we learn that a Muhasena-
pati ruled over an dhdra much in the same way as an amaca. Modern in-
stances of finding appellations for women in the official titles of their hus-
bands are afforded by Viceroy, Putlin and Gandasami. That Mahdrafais
governed is shown by Karla No. 14, where a Maharafhi grants on his ac-
count a village with its taxes and by the Chitaldoorg coins bearing the
legends Sadakana Kalaldya Mahdrathisa. If Mahdrathi is not an official
title, we are led to the paradoxical conclusion that the feudatories are known
in their coins and official grants by their nobiliary titles, whilst the official
titles (given for even minor officers, Nasik Nos. 4 and 5) are the only ones
we miss here. Etymologically too, the term, which Senart himself admits
presupposes a Sanskrit form Mdhdrdstrin implies an office (vide supra).
36. Karla No. 14. We owe to Senart a proper explanation of the terms
sakarukaro and sadeyameyo which Biihler and Bagwanlal translate as
'this gift is in order to keep the Valuraka caves in repair*. Senart splits
sakarukaro into 'kara' and 'ukara,' the exact equivalent of which appears
at the head of customary formulas which begin generally with sddranga
soparikara. According to him while /caret is known in the sense of dues
payable to government, the meaning of uparikara is as unsettled as that of
sddranga. But upari means 'above' and uparikara may be taken to mean
taxes over and above the ordinary ones. The adjective Sadeyameyo is
etymologicaUy translated as 'what is taken (in money) and what has to be
measured (meyya to be measured, ddeya = to be taken) .
H.A.-11
& EAKLY HISTORY OF THE! ANDHRA COUNTRY
the second century A.D. the Mahdrathis were contributing their
share to the dismemberment of the Satavahana empire. 37
Like many feudatory titles, that of Mahdrafhi had purely
a local significance. The Mahdrathi and Mahdrafhinl inscrip-
tions are found in northern Mysore and the fhEUJ a and Kolaba
districts of the Bombay presidency. The expressions Okhala
kiydnam Mahdrathi, Sadakana Kalaldya Mahdrathi, and Amgiya
kulavadhana Mahdrathi suggest the idea that the title was restric-
ted to a few families or tribes. 38 Senart says: "It may be noted
that to Mahdrathi Agimitranaka corresponds a Mahdrathi
Mitadeva in No. 14; that this Mitadeva is a Kausikiputra, like Vish-
nudatta at Bhaja (No. 2) ; and lastly that the Mahdrathini Sama-
dinika at Bedsa (No. 2) was married to an Apadevanaka. Do
not these different names look as if they were connected with each
other in such a way as to suggest the idea that they may have belong-
ed to the same circle of families or relations ?" Much cannot be
built on similarity in names, especially when they are very common.
Names like Mitabhuti and Mitadeva occur very often in the western
cave inscriptions. Surely the bhayata Mitabhuti of the Kanheri
inscription (Lliders, List, No. 1012) has nothing to do with the
Mitadeva of No. 1187 and both have nothing to do with the
Mitadevanaka of No. 1097.
The Najjeghat and Kanheri inscriptions show that the Mahd-
rathis had marriage relations with the ruling family much in the
same way as the Mahdtalavaras of the Iksvaku period. Professor
Rapson remarks: 39 " That they were. . . .closely connected with the
Andhra kings by family or by caste seems to be shown, as Pandit
Bhagwanlal observed, by the use of metronymics which they have
in common with them." But metronymics are not peculiar to a
caste or family. They are borne by Brahmans, 40 artisans, 41 and
even Buddhists, monks and laymen. 42 Unlike the Satavahanas,
Mahdrathis sometimes bear metronymics not derived from Vedic
gotra names, and give their father's name also.
It must however be noted that unlike the inscriptions of feuda-
tories of later times, the Mahdrathi and Mahdbhoja inscriptions are
37. El, Vol. VIII, PI. m.
38. Karla, No. 14; Chitaldroog Maharathi coins; Naneghat inscription
of queen Nayanika.
39. JRAS, 1903, p. 299.
40. Liiders List, Nos. 1195 and 1196.
41. Ibid., No. 346.
42. Ibid., Nos. 457; 661; 662; 663; 664; 665; 667; and 1271.
ADMINISTRATIVE ARRANGEMENTS 83
as short as the votive inscriptions of ordinary persons. We miss in
them the laudatory epithets applied to feudatories of a later time,
Mahdsendpatis
Another dignitary in the empire was the Mahdsendpati, next in
rank above the Sendpati. The first known epigraphic record to
mention a Mahasenapati is the Nasik inscription of the 22nd year
of Vasithiputa sami Siri-Pulumavi . In the records of his succes-
sors the title is mentioned twice and in Iksvaku recbrds often. It
may, therefore, be presumed, that the title is not as eld
as that of Mahdrathi. 43 Its origin is perhaps to be sought
in the rapid expansion of the empire from sea to sea in the second
century A.D.
The Mahasenapatis of the Satavahana period have non-mili-
tary duties; but this confusion of functions though it may seem
curious to moderns was a common feature in Indian polity. In
Nasik No. 3, the drafting of the royal order is attributed to a Mahd-
sendpati. (Mahdsendpatind Medhunena Nokhaddsdtara(ne)na
chatho) . While editing the inscriptions Senart remarks 44 : " As
to the Mahdsendpati, the proper name alone seems obliterated or
doubtful; but the lacuna may have contained something else than
his name. Other inscriptions do not attribute to the Sendpati the
menial work of drafting, but perpetuate his name as that of a high
officer entrusted with this charge at the end of the grant; see e.g.
Dr. Fleet's Gupta Inscr. Nos. 55 and 56. In a still higher
degree the title of Mahdsendpati, which comes very near to that
of Maharaja seems to place the person who is honoured with
it above any such mean task. This is why I suspect that the obli-
terated letters, if exactly known, would let his part appear in a dif-
ferent light." No doubt in cognate inscriptions (Nasik Nos. 4 and
5) the drafting of the royal order is attributed to minor officials.
But here the third case ending in Mahdsendpatind precludes any
chance of his lekhaka's name having been on the stone. A close
examination of the stone renders 'Nokhadasdtara (ne) na 9 probable;
and in western inscriptions 'dasa' often enters into the composition
of names. The Hira-Hadagalli plates mention a Rahasddhikata,
43. The office of Senapati (Commander of forces) would seem to have
been coeval with the beginnings of Indian polity itself. We hear of it in
the Vedas (Vedic Index Sen&m) the Jatakas, the ArthaSastra, and
the Puranas; and an inscription from Ayodhya (El, Vol. XX, p. 57) shows
that Pusyamitra was a Sen&pati under the last of the Mfcuryas,
44. El, Vol. VTO, p. 70,
&4 EARLY HISTORY OF THE ANDHRA COUNTRY
the Khoh copper plate of Maharaja Hastin a
and most of the Valabhi grants of the sixth century A.D. a
Sdmdhivigrahika, as writers of charters. 46 The title Samdhivi-
grahika is sometimes used in connection with that of Mahddanda-
ndyaka which is associated with the great feudatory titles of Maha-
sendpati, Maharaja, Mahdprafthdra and Mahdsdmanta. 47 Mahd
sdmdhivigrahika would seem to be an officer equal in rank, if not
superior to Mahasendpati. In the case of such high officials the
mean task of drafting would have been done by clerks under them.
What would be a conjecture is raised to a certainty by the expres-
sion say am chato in the Kondamudi plates. 48 The task of reducing
royal writs to writing was a responsible one; the dangers atten-
dant upon a careless drafting and the large number of orders to be
drafted might have necessitated a lekha department under a res-
ponsible officer. 49
The Mahasendpati of the Jangli Gundu inscription of the
time of the last king of the Satavahana line, is, like the amaca,
in charge of only an dhdra. As late as the reign of Ca$a
Satakani, the eastern provinces, divided into dhdras, would
seem to have been under amocas. 50 It is, therefore, probable that
in the days of the Saka attacks and the dismemberment of the
empire, the outlying or vulnerable parts were put under Mahd-
sendpatis who would naturally have seized the opportunity to gain
feudatory rank and power. 51 Jangli Gundu is midway between the
45. CH, Vol. HI, p. 105; t.l. 28.
46. IA, Vols. IV, etc.
47. CII, Vol. HI.
48. Vide infra.
49. The Artha&dstra says that only persons possessed of ministerial quali-
fications, acquainted with one kind of customs, smart in composition, good in
legible writing and sharp in reading should be appointed as r&jalipikaras
(chap. IX; Bk. I).
Sometimes Dutakas carried the orders to local officers whose duty it
was then to have the charters drawn up and delivered (Nasik No. 5, op.
cit, CH, Vol. HI, p. 100, n.)
50. El, Vol. XVI, pp. 316-19.
51. Prof. D. R. Bhandarkar holds that Mahdsenapati of Nasik No. 21,
is a feudatory on the ground that the Anguttara Nikaya mentions Sen&pati
along with kings, hereditary Ratfhikas and heads of villages. But the office
of Mahasenapati was a generic one, and the three Mahdsenapatis of the
Satavahana inscriptions appear in three different capacities. So an inference
based on such argument cannot be conclusive.
V. S. Sukthankar remarks (El, Vol. XIV, p. 155); "The relation in
which the mah&sen&pati and the gumika stand to the janapada and the gama,
ADMINISTRATIVE ARRANGEMENTS 85
Myakadoni and Chinnakagiaburu, villages in the Bellary District
which along with parts of northern Mysore and Kanara
would have represented the southern limits of the Satavahana
empire in the second century A.D. It is also probable that Sata-
vahanihara was exposed to Cutu and Mahdrathi attacks. That,
under the Iksvakus, the title had become a feudatory one is certain.
These feudatory titles seem to have not only survived the
Satavahana rule in the Deccan but spread as far south as Mysore.
We hear of a Mahdrathi in the Chitaldoorg District, who struck
coins in his name and of a Mdhdrathirfi in Kanheri and Banavasi
inscriptions of the line of Haritiputa Cutukulananda Satakani. The
same inscriptions show the Cutus intimately connected with the
Mahabhojas and Makdrathis. It may be that the Cutus (who suc-
ceeded the Satavahanas in the south and in parts of the western
dominions), themselves started as Maha,rafhis or Mahdbhojas. But
unlike the title of Mahdsenapati, these titles do not seem to have
spread to the Andhra dominions of the Satavahanas; much less did
they become Mahdsdmanta. 52
is not explicitly mentioned . But, considering the position of these
persons, one might hazard the guess that, these were feudal lords of
the lands, holding them in the form of jagirs." For reasons given below,
what is read as Gumika is to be read as Gdmika; and this considerably
weakens the force of his arguments.
D. C. Sircar (Successors of the Satavahanas in the Eastern Deccan,
p. 15) says that under the Satavahanas, the Mahdsen&patis were feudatory
chieftains in charge of rastras. We do not know upon what evidence this
statement rests.
52. It is tempting to connect the Bhojaka and Mahabhoja of our inscrip-
tions with the bhogika and the Mahdbhogika of the later records. In
the Gupta inscriptions the son of a Bhogika is in charge of the drafting of
the order or charter. (Fleet GI, pp. 100, 105, 109, 120, etc.). In the in-
scriptions of the Gurjara Buddharaja (Kalacuri Samvat 361 El, Vol. VI,
p. 298), Dadda II, Prasantaraga (Kalacuri Samvat 380 and 385 respectively,
IA, Vol. XIII, pp. 82-88) and Jayabhatta III (IA, Vol. V; p. 110);
the Bhogikas are mentioned after Kdjas, Samantas and before
Visoi/apatis and Ra&rddhikdrikas . What is interesting, both Bhogika
and Bhoja literally mean 'one who enjoys*. Both refer to rulers of dis-
tricts also. (Bhogika may also be connected to Bhogapati, i.e., a governor
or officer in charge of revenue). According to the lexicographer Hema-
candra, both Bhogika and Bhojaka mean Village headman.* But the fact that
the Mahdbhojos are not referred to in any inscription after the second
century A.D., the long interval that separates the Mahabhogikas from the
Mahabhojas and the technical nature of the titles, make any connection
between them problematical.
86 EARLY HISTORY OF THE ANDHRA COUNTRY
Officials and administrative divisions
Barring the districts enjoyed by the feudatories, the empire
was divided into akdras (including the Andhra province), e.g.
Soparahara, 53 Govadhanahara, 54 Mamalahara 55 and Satavahani-
hara. 56 An dhdra would represent the same territorial division as
the rdstra of Pallava records, the visaya of the records of kings of
the Salankayana gotra, and the modern district. 57 Can we draw
from the compound arathasamvindyikam, which is translated by
Senart 58 as 'exempt from the magistrate of the district or of the
Rastrin' the inference that some divisions of the Satavahana empire
were called rdstrasl Such technical expressions are often, con-
ventional and in not one of the inscriptions of the Satavahana
period is a territorial division called rdstra. Ahdras were under
governors called amacas (Sanskrit, amdtya) who were non-heredi-
tary officers. 59 Quinquennial transfer was in force. Each dhdra
presumably received its name from the headquarters of the gover-
nor (nagara); Kudura (the Koddura of Ptolemy) is known to us
from an Amaravati inscription of the second century A.D. 60 and
yet we hear of Kudurahara only in a copper-plate grant of the
third century A.D. 61
In a Kanheri inscription, which, on palaeographical grounds,
can be ascribed to the time of Siri-Yana Satakani, 62 we have the
expressions Patithdne and Rdjataldka-Paithdnapathe. Could we
translate the latter as "the village of Rajatadaka (King's Tank) in
the Paithana division " ? If Rajatalaka is a mistake for 'Rajatalake'
it would mean "in Rajatadaka in the Paithana division." 63 The
analagous expression Amdhapata (Andhrapatha) in the Mayi-
davolu plates, which is synonymous with Andhrarattha, makes it
53. Kanheri No. 5, Vol. IV.
54. Nasik No. 3, El, Vol. VIH.
55. Karla No. 19, El, Vol. VII.
56. Jangli Gundu inscription of Pulumavi III, El, Vol. XIV, p. 155.
57. Vide supra.
58. Nasik Nos. 3, 4, and 5, El; Vol. VIII. Also VII; p. 68.
59. The amaca of the Ko^avolu inscription of Cada Sati and the r&ja-
maca of Ku$a No. 18 (CTI), might have been governors of fthdras. Some-
times, however, treasurers and officers in charge of the drafting of charters
bear the same title (Nasik Nos. 4 and 19).
60. Liiders, List, No. 1295.
61. The Kondamudi plates of Jayavarman, El, Vol. VI, p. 315jff.
62. No. 5 ASWI, Vol. V; Compare No. 4.
63. Liiders leaves the expression untranslated.
ADMINISTRATIVE ARRANGEMENTS 67
highly probable that 'patha* literally 'path or road* is synonymous
with dhdra (district) ; 64 may be the district of Paitfaana which con-
tained the seat of the king and was perhaps under the direct con-
trol of the king, was distinguished from the other divisions in this
way. 65
The next division below that of dhdra is gdma (grdma) . From
the Saptasatakam of Hala we learn that the officer in charge of a
village was called Gdmika (Grdmika). In the Jahgli Guri^u ins-
cription of Pulumavi we come across a Gdmika. 66
The other functionaries known to us are the Mahatarakas,
Mahd-dryakas 61 Bhdn^dgdn/cas, 68 Heranikas 69 Mahdmdtas
64. It has been shown that the Pallava rattha is synonymous with dhdra.
65. In the Alina copper plates of Slladitya VII (year 447) we have the
expression Sri Khetakhare VppalaMta pathake Mahila(?ld)bali n(d)ma-
gramah. Pathaka which Dr. Fleet (CII, Vol. Ill, p. 173, n.) connects with
paLhin or patha represents here a territorial division beween aliara and
grama.
66. V. S. Sukthankar would read it as 'Gumikasa' (Gaulmikasyaoi the
Captain) which, according to him, would agree with the Matutsena-
patisa oi the preceding line. True in the Hira-Ha$agalli plates, coining from
the same district, Gumikas are mentioned. But since, the officer over a
gdma (grama) is mentioned and as the u sign is not visible on the plates, it
is safer to read it as 'G(d)wifca.'
67. 68 & 69. Mahataraka means the Great Chamberlain. Hemacandra m
the Deslkosa (i. 16), gives aira in the sense of an. official, e.g., the lord of a
village. Etymology would therefore seem to be unsafe guide in the inter-
pretation of official terms. As regards the Maha-aryaka mentioned in
Nasik No. 3, Senart says (El, Vol. VIII; p. 68: " the part the monks are
playing in the first sentence seems to point to the name being that of a religi-
ous personage . Even admitting that the title araka given to Yanasiri-Satakani
(Siri-Yana Satakani) by an inscription (Ep. Ind. Vol. I, p. 96) be really =
aryaka, that would in no way prevent this epithet, which is commonly^ used
with reference to Buddhist monks, being applied to some religious functionary.
I am the more inclined to think so, because I find the similar title Chula-arya
conferred on the Arya Buddharakshita ; who is styled Arhat (Burgess' Bud-
dhist Stupas of Amaravati, Plate lix. No. 39, p. 104)." He also compares
this title to that of Mahdsdmiya in Nasik No. 4. But Maha-drya is used in
the singular, while the title of Mahasamiya is used in the plural. The latter is
perhaps an instance of pluralis majestatis, in which case it would not point
to a college of religious functionaries. Mahat-araka (Sans. Mahat-dryaka) ,
an official title, is mentioned in the Chinna Ganjam inscription of Siri-Yana
Satakani, and the way in which our Mahd-dryaka is connected with the
village of Samalipada ('This village of Samalipada. . . . by the Mahd-dryaka,
you (amaca), must deliver to be owned by the BhiJc$u, of the school of
the Bhadayaniyas), seems to point to a secular official, one in charge of a
d8 EARLY HISTORY OF THE ANDHRA COUNTRY
(Mdham&tras) in charge of Buddhist monks, 70 the NibwihdhaMras
or officers in charge of the registration of documents (the Aksapa-
talikas of later times), 71 the Pratih&ras, the Dutakas who carried
royal orders, and the Amacas in charge of drafting royal orders.
From a Nasik inscription of Uavadata we learn that every town
had a records office. 72
APPENDIX A
The Alluru Inscription
One of the notable discoveries of the Epigraphy Department in
the year 1924 was that of a Brahmi inscription on a fragment of a
marble pillar at Alluru, a village in the Nandigama taluq of the
Krna district. The pillar under reference must originally have
belonged to the Buddhist Stupa which stands at about two furlongs
to the west of the village. The Epigraphy Report for the year
ending with March 31, 1924 contains a photograph of a facsimile
of the inscription and a short note on it. It was subsequently edi-
ted by Dr. R. Shamasastry in the Calcutta Review for the year 1925.
His reading misses the truth in many a place.
gama perhaps. Lastly in the Amaravatl evidence cited by Senart, Cula-
Arya is a name and not a title conferred on Arya-Buddharaksita. (G8)
Nasik No. 19, El. Vol. VIII. Bhdnddgdrika which may mean both
store-keeper and treasurer is better construed here as store-keeper as in
cognate records. Heranika, is the term for treasurer. (69)
The various forms are Heranika, Heranaka, Heranika, Hiranakdra ;
this term which occurs in the inscriptions at Kanheri, Nasik, AmaravatI, and
Bhattiprolu is better construed as treasurer, as in them suvanakdra is the
term for goldsmith (Luders, List, Nos. 986 and 1117) . If the treasurer Dhaih-
manaka of No. 993, is identical with Dhammanaka, son of the treasurer
Rohanimita of Nos. 996 and 1033, the office would seem to have been, at
least to some extent, hereditary. At times we come across a treasurer who
was the son of a merchant or a gahapati (Luders, List, Nos . 1239 and 1249) .
70. Senart (El, Vol. VIII, p. 93) however considers the translation of
Ndsikakena, samanena Mahdmatena* as 'by the officer in charge of the
monks of Nasik,' as hypothetical.
71. Senart (ibid,, p. 74) takes them to be Dutakas carrying the orders
for registration. Senart himself has happily explained the meaning of
nibandh by a reference to Yajnavalkya I, 317. Says he: "nibandh was
perhaps a kind of the royal decision in the archives of the state." In such a
case the meaning given to Nibandhakdra here is more apt than the one pro-
posed by Senart.
72. 'Nigamasabhdya nibadha ca phalakavdre', ibid., No. 12.
ADMINISTRATIVE ARRANGEMENTS &
My reading of the inscription is as follows: Hie beginning
line or lines of the inscription are lost.
1. (ai) lasa Ma$a (vi) sa ca
2. sa ramo viharo deyadhama parica. . . .
3. nigala-simaya Vetaraku^o Na (ga) . . .
4. ti Khetasarasa Papikala simaya. . .
5. nivatanani raja datini. Carathe Macha . .
6. (pa) da-simaya batisa nivatanani Ra . .
7. (c)erapura-simaya ca (tu) visa-nivatanani. .
8. <Jalasa gavina pacasatani (co) yathl baliva(da)
9. sakadani pesa-rupani dasi-dasasa cata 1 (isa) . .
10. kubhi kadahasa catari lohiyo be kad (a) hani (kasa) . .
11. (sa) bhayanam catari vadalabhikaro karodiyo (yo)
12. (na) ka-dlvikayo ca Ataragiriya pica-pake talaka(ni)
13. kahapanana ca puranam sahasam akhayani v(i)
14. esa mahatalavarasa deya-dhama paricako
15. ata Pedatarapase bapana-nivatanani
16. eta sabhariyasa saputakasa sanatukasa
17. Ayirana Puvaseliyana nigayasa ....
The rest of the inscription is lost. It mentions the gifts made
by many, including a king, who perhaps out of modesty omits his
name, a unique feature in our records. The first two lines speak
of an Alia (Aira, or Arya) Madavi. Then comes the gift of some-
thing within the limits of Vetaraku<Ja. Next is mentioned the gift
of a beautiful vihdra, perhaps by the side of the Stupa,,
of some nivatanas of land within the limits of Papikala (for the
identification of Papikala, see the chapter on the Iksvakus). Then
come the gifts of 32 nivatanas of land within the limits of Macha- -
pada in the ratha of Ca, and 24 nivatanas of land within the limits
of the town of Ra- -cerpura, 500 cows, 64 bullock carts, 40 servants
some cauldrons, especially two brass cauldrons, 4 bronze vessels,
some hand lamps of the vadala fish shape, some Yonaka lamps,
a tank in the vicinity of Ataragiri, and one thousand purdna kaha-
panas as a permanent endowment ; 52 nivatanas of land were the
gifts of a Mahdtalavara along with his wife, son and grandson.
All these gifts were for the (acceptance of) the school of the
Pubbaseliyas....
The use of a peculiar form of ai which comes close to the Vaf-
teluttu ai is noteworthy. I am indebted for this reading to Mr.
K" N. Diksit. It is clear that in this record the Mahdtalavara is
playing a more important role than even the king.
HA. 12
90 EARLY HISTORY OF THE ANDHRA COUNTRY
As D. C. Sircar has also pointed out what was read by Dr.
Shama Sastry jayadhama is only deyadhama (I, 2). What has
been read as caradhama is undoubtedly Carathe Ma- -pada (I. 5).
What is read as Sanasa kata (made by Sana) King of the Ayis
(Ayiranam is interpreted as King of the Ayis), is only sanatu
kasa (with his grandson) (1. 16) and Ayiranam refers to the school
of Pubbaseliyas mentioned in the same inscription.
CHAPTER VI
SOCIAL, ECONOMIC AND RELIGIOUS CONDITIONS
I. SOCIAL CONDITIONS
The Bhattiprolu, Amaravati and western cave inscriptions give
us an insight into the social organisation of the Deccan from the
second century B.C. to third century A.D., and from the Amaravati
sculptures we get a vivid picture of life in the Andhradesa. The
fourfold division of society the caste system prevailed. We hear
of Usava data's charities and gifts to numerous Brahmans. Gotami-
puta Siri-Satakani boasts of having prevented the contamination of
the four castes and of having been the true supporter of Brahmans.
Yet one of his descendants took his wife from the Saka satrapal
family and the Sakas were looked upon as degraded Ksatriyas. It
is doubtful whether Buddhism toned down the caste system even
to the extent to which these foreign elements in society did. At
Kuda an upasaka Ayitilu calls himself Bammhana. 1 His wife
is called Bamraanl. Mahadevanaka of a Karla inscription who
would seem to be Ayitilu's son bears the title of gahata (Sans.
qrhasta). In the sarman ending in the names of Buddhists, monks
and laymen, we have probably one of the Brahmanical vestiges
in the Buddhist communities. 2 Ksatriyas would sometimes seem
to have followed the profession of the Vaisya caste. In a Kanheri
inscription Gajasena and Gajami(ta), the Khatiya brothers, follow
the profession of vdnijakas." Unfortunately the lacunae before
khatiyasa, makes the meaning of the word not quite certain. The
sub-castes met with in inscriptions are those of the hdlaka or
halika* (ploughman), murdhaka (according to the Sabdaratna
1. ASWI, Vol. IV, No. 13.
2. ASWI, Vol. IV, Junnar No. 19, p. 96 ; Liiders, List, Nos. 1101 and
1102. However sarman ending in names does not always indicate Brahmani-
cal origin. In an Amaravati inscription a vdniya (belonging to the Vaisya
caste probably) bears the name Bodhisamman (TSW 1873, p. 261, No. 8).
See also Fleet CII, Vol. Ill, p. 11, n.
3. ASWI, Vol. V, Kanheri Inscriptions No. 4.
4. Liiders is in doubt as to whether hdlika in No. 1084, is only a per-
sonal name, or a variation of hdlaka. An Amaravat! inscription (El, Vol. XV;
Some Unpublished Amaravati Inscriptions, No. 56), makes it certain that
hdlika is not a personal name. Wherefore it must be identical with hdlaka .
92 EARLY HISTORY OF THE ANDHRA COUNTRY
Petersburg Diet, s.v.) a branch of the Ksatriyas, and golika
(herdsman) . 5
Foreign Elements in Hindu and Buddhtet Society
(i) Yavanas. Yona or Yavana is an Indian form of the word
Ionian. In literature Yavana refers to all kinds of foreigners or
Mlecchas. At the present day the term Yona is applied by the
Sinhalese to the 'Moormen' or Arabs some of whose families have
been settled in Ceylon for centuries. It is however more than pro-
bable that Yavana of our inscriptions denotes the Greeks. For, the
Yonas of Asoka's inscriptions placed with the Kamboja's and the
Gamdharas in the north-west are certainly the Greek element that
Alexander's invasion and Seleucus' empire left in the north-western
India; as our inscriptions mention akas and Yavanas, 6 a confusion
between Sakas and Yavanas is ruled out; as the term Yavana
occurs in,the inscriptions of the foreigners also, it is improbable that
they did not clearly state their racial affinities.
We do not know how and when these Yavanas entered wes-
tern Deccan. According to the Mahavamsa, some 250 years after
the Nirvana of the Buddha, the Yona priest Dharhmarakhita was
gent to Aparanta as a missionary, 7 while the priest Maharakhita
was sent to the Yona country. This shows that there was already
in western Deccan a large element of foreign Yavana population.
Yavana Tusaspha was governor of Surastra under Asoka. Accord-
ing to Strabo, Menander, the Greek prince, penetrated into
'Isamus* (Jumna) and subjugated Patalene (the Indus delta) and
Saraostes (Surastra). This statement is corroborated by the curi-
ous observation of the author of the Periplus that the coins of
Apollodotus and Menander were current in his time at Barygaza. 8
Of a crowd of Yavanas in western India thoroughly Indianised we
hear in the inscriptions at Karla. 9 A Sarici inscription mentions a
Yavana. 10
5. ASWI, Vol. IV, Junnar No. 2 pp. 92 ff.
6. Nasik, No. 2, El, Vol. VIII, speaks of gakas, Palhavas and Yavanas.
7. MaMvamsa, Geiger p. 82 and Intro, xxxi; Ivii.
8. Schoff, The Periplus, pp. 41; 42, and Sec. 47.
9 & 10. Sten Konow is of opinion, that the 'Yavana' of the Karla inscrip-
tions wherever it is followed by a name in the genitive plural, is a personal
name. Liiders looks upon Yona in No. 547 also as a personal name. There
are weighty considerations to be brought forward against this view. Firstly
it is improbable that many persons bore the same name Yavana, The ethni-
con Yavana denoted to the Indian a foreigner whom he looked upon as a de-
SOCIAL, ECONOMIC AND RELKH0U& CONDmONS 93
As these inscriptions are incised on the Cetiya cave pillars,
they are as old as the cave itself. The palaeography of the oldest
Karla inscriptions would support a first century B.C. date for the
Cetiya cave. It is possible then, that the Yavanas entered the
graded Ksatriya ; it is therefore improbable that Indians bore 'Yavana' as a
personal name. Much less could a Yavana have done so.
While Sten Konow looks upon the names in the plural found along with
the word 'Yavana' as a family or corporation name, Senart looks upon them
as personal names in the genitive plural. Senarl looks upon the genitive
plural in Simhadhayanam in Karla No. 7 (El, Vol. VII), as a personal name
in the plural (pluralis majestatis) , and following him M. Swarup Vats has
treated the other names in the genitive plural in the other Yavana inscriptions
as personal names. In all the Karla epigraphs and in the Yavana epigraphs
at Junnar (CTI Nos. 5 and 33), the personal name is in the singular while
the family name is in the plural ' Okhalakiydnam Mahdrathisa Kosikiputasa
Mitadevasa' (Karla No. 14 El, Vol. VII), 'Yavanasa Irilasa Gatdnam' and
'Yavanasa Citasa Gatdnam (Junnar, Nos. 5 and 8; ASWI; Vol. IV, pp. 93 and
94) . It is therefore, not proper to consider Camddnam (Junnar) Culayakhanam,
Dhamadhaydnam, Vitasamghatdnam, Simhadhayanam and Yasavadhandnam
(Karla) as personal names. The Junnar inscription under reference is as-
signed by Dr. Burgess to the first century B.C. on palaeographical grounds,
and it is the period of the Karla Caitya cave. In the Sand Yavana inscrip-
tion we miss the personal name ' Setapathiyasa Yonasa ddnam . ' In Liiders,
Nos. 82 and 1035 the personal names of the donors are omitted though their
gotras are mentioned. While editing the Sand inscriptions, Dr. Buhler
remarks (El, Vol. II, p. 94): 'Peculiar and noteworthy are the names of
monks and nuns, which like Kaboja, Pratithana, Chirati, and perhaps also
Odi, consist of adjectives derived from the names of countries, towns and
races. In these cases it would seem that the real name of the donors has
been left out."
Senart translates the compound 'Dhamma- Yavanasa' in Karla No. 10 in
the same way as Buhler did viz., 'of Dharhma, a Yavana.' He adds : "
the simple name of Dhamma applied to a Buddhist surprises me I feel
tempted to take Dhamma in a specifically Buddhist sense, and to under-
stand by dhammanigama 'a member of the guild of Buddhist merchants';
compare nigamasabhd at Nasik (No. 12, 1. 4). On this analogy Dhamma-
Yavana would be 'the community of the Buddhist Yavanas' or rather a Bud-
dhist Yavana who has modestly omitted his personal name" (El,
Vol. VH, p. 56). Liiders considers Dhamma- Yavana as the name.
But in a Nagarjunikon<Ja inscription Dhamma occurs as a personal
name (El, Vol. XX, Ins. J). Dharhmila, a name of very common occurrence,
is only Dhamma with the la suffix. Names Cetiya and Sagha are of the
Dhamma class. Nor is the compound a source of difficulty. In the Nasik
inscription of the time of Abhira-Isvarasena we have the compound Sivadatt-
Abhvraputrasya ; the analogy is not, however, very close, since a compound
is necessary in the latter case and since the one inscription is in Prakrt and
the other in Sanskrit. In an Amaravati inscription (El, Vol. XV, No. 11)
we have Sa[m]ghalasamana3a; Safmjghala cannot be anything else than a
94 EARLY HISTORY OF THE ANDHRA COUNTRY
Satavahana dominions in the wake of Saka conquest. The Yavanas
who are credited with donations at Karla are: one of the Sirhha-
dhaya family, one of the Yasavadhana family, one of the Dhamadha-
ya family, and one by name Dhama.
As all the Karla Yavanas except one, profess to be natives of
Dhenukakata, this place would seem to have contained a Yavana
settlement. As most of the donors in the Karla inscriptions, come
from Dhenukakata, 11 and as the place name occurs frequently in
Karla epigraphs and once in an inscription at Sailarwadi, a place
very near Karla, it has to be sought for in the vicinity of Karla.
It would therefore seem to have been included in Mamialiahara. 12
A point that deserves mention is that these Yavanas besides
embracing Buddhism adopted thoroughly Hindu personal and
family names. 13 They use Prakrt in their inscriptions and
it is not unreasonable to infer that they adopted Hindu man-
ners and customs. This is no wonder since even a casual visitor
to Ujjain from the kingdom of Antalkidas became a Bhagavata. 14
So completely did the Yavanas merge into Hindu society that Indian
Buddhists had no scruples whatsoever in joining with these foreig-
ners in making donations. The Karla Cetiya cave was a result of
such a joint effort.
(ii) Sakas. "Like the Yavanas, the Sakas too merged into
Hindu society. The Saka son-in-law of Nahapana bears the Indian
name Usavadata (Sans. Rsabhadatta) , while his father bears the
un--Indian name Dmika. Another 6aka bears the name Agnivar-
personal name. Liiders, List, No. 1283. The analogy between Dhath-
manigamasa and Dhamma-Yavanasa suggested by Senart breaks
down at every step. Yavana is not like nigama, a collective noun.
Another objection is that in all the other Karla epigraphs either
the family name or the personal name or both appear. [Dr. Tarn has
suggested that Dharhma- Yavana was a naturalized citizen of an Indo -Greek
polis. See, however, JRAS 1939 pp. 217 ff and 1940 pp. 179 ff Ed]
(10) El, Vol. II, p. 395, No. 364.
11. El, Vol. VII, Nos. 4, 6 and 10, El, Vol. XVIH, Nos. 3, 4, 9, 11
and 12.
12. It is mentioned once at Kanheri. Liiders, List, No. 1020.
13. In Nasik No. 18, El, Vol. VIII, we hear of a Yonaka from Dattamitri.
According to Biihler, it is the same as Demetrias, a town in Arachosia, men-
tioned by Isidore of Kharax. He bears the name Indragnidatta, his father
is called Dhammadeva, and his son Dharhmarakhita.
14. It is stated in the Milinda Pafiho that the Yavana king Milinda
(generally identified with Menander) was converted to Buddhism by the
teacher Nagasena. According to a legend mentioned by Plutarch no less than
seven cities fought for his ashes. (IA, Vol. VIII, p. 337),
SOCIAL, ECONOMIC AND RELIGIOUS CONDITIONS 95
man, and his daughter that of Visnudatta. A fourth bears the
name Vudhika (Sans. Vrddhika). 15 If Nahapana was a Pahlava,
even Pahlavas would seem to have followed the example of Yava-
nas and Sakas, for Nahapana's daughter bears the Indian name
Daksamitra.
Unlike our Yavanas all of whom are Buddhists, Sakas embraced
both Brahmanism and Buddhism. Ku$a inscriptions mention a
Brahman updsaka, named Ayitilu, and according to Senart, it is a
foreign name corrupted and curiously reminding us of Azilizes. 16
Saka Usavadata's charities to Brahmans and Brahman institutions
stamp him as a staunch adherent of the Brahmanical religion. We
are told that he gave money and tirtha on the river Barnasa and
also 300,000 cows and 16 villages to Brahmans. He bathed at the
Poksara tanks and gave the Brahmans 3000 cows and a village.
He also gave eight wives to Brahmans and fed thousands of them
all the year round. The Carakas also received something at his
hands. In a Nasik inscription, however Usavadata says that "ins-
pired by true religion, in the Trirasmi hills at Govardhana" he
caused a cave 17 to be made. Says Senart: "I dare not decide if
this phrase (dharmdtmand) implies an express conversion to Bud-
dhism, or only puts a first gift in favour of Buddhism in contrast
with the previous grants which were inspired by Brahmanical feel-
ings. I do not think the wording allows us to settle this shade of
meaning. On the strength of this explanation I propose in N. 18
to take dhammatmana in a similar way. I believe the reading
'tmana, not 'imano, is certain, and the manner in which the cons-
truction is interrupted after the preceding genetives confirms the
impression that dharmdtmana is intentionally put forward, in ord^r
to dwell on the fact of a change having taken place in the religious
belief or inclination of the donor Indragnidatta," (a Saka). If
reliance could be placed upon names, Indragnidatta's father and son
would both seem to have been Buddhists. As three years after the
foundation of his Nasik cave Usavadata makes donations to
Brahmans, his change of faith is extremely improbable. 18
15. In No. 26, El, Vol. VIII we have akasa D&macikasa lekhakasa Vudhi-
kasa Vifn/udata-putasa Dasapuravathavasa . As the donor's race and father's
name are mentioned Damacika is the name of his native town. Senart, how-
ever, does not agree with Bhagwanlal who looks upon it as a corruption of
Damascus.
16. El, Vol. VH, p. 52.
17. Govardhane Trirasmiqu parvate$u dharmatmanS, idam lenam kdritam,
El, Vol. Vm, p. 78, No. 10.
18. The cave was consecrated in the year 42.
& EABLY HISTORY OF THE ANDH&A COUNTRY
Vi^nudatta, the daughter of Saka Indr&gnidatta, was an upa-
sifca. As Dr. Bhandarkar observes: "These Saka kings (Western
Ksatrapas) had thus become so thoroughly Hinduised that another
Hindu royal dynasty (the Satavahana) had no scruples whatever,
social or religious, in entering into matrimonial relationship with
them." 19
We do not hear of the Sakas and Yavanas in the Satavahana
dominions in the western Deccan after the second century A.D.
The epithet Saka-Yavana-Palhava-nisudanasa applied to Gotaml-
puta SirinSatakani seems to be no mere boast; evidently he drove
out these foreigners from his newly rebuilt empire the Sakas
were only to return in the wake of Rudradaman's conquest for a
short time.
Yavanas and Sakas in Eastern Deccan
Of Yavanas in the eastern Deccan we hear nothing; it is how-
ever certain that Graeco-Roman influences played a great part in
the fashioning of the Amaravati tope, and as will be shown below
the inscription from Alluru 20 is another piece of evidence for Greek
influence. Of the Sakas we hear something, An Amaravati ins-
cription of the second century A.D., mentions a Saka-giri (not
(S)akagiri as read by Chanda, or Pi (Si?) giri as read by F. W.
Thomas). 21 Another mentions a * ratika Nekhavana,' and
Nekhavana curiously reminds us of the Persian name Nahapana. 22
More Sakas would seem to have entered eastern Deccan in the
wake of the marriage of Virapurisadata with the daughter of a
Western K^atrapa. A Nagarjunikonda epigraph 23 mentions a Saka
Moda, 24 and his Buddhist sister Budhi. Among the sculptures ex-
cavated by Mr. Longhurst at Nagarjunikonda there are two show-
ing a warrior in Scythian dress.
Family
At Amaravati not only father and mother, but also sons and
daughters, sons-in-law and daughters-in-law, and brothers and
19. IA, Vol. XL, p. 15.
20. Vide infra.
21. Some Unpublished Amaravati Inscriptions, No. 55, El, Vol. XV, and
PI. facing p. 273.
22. Ibid., No. 54.
23. El, Vol. XX, p. 37.
24. Sanskrit moda means 'joy/ This is an instance of an Indian name
borne by a &aka.
SOCIAL, ECONOMIC AND RELIGIOUS CONDITIONS 97
sisters, are associated with the donor. The wife has a place above
the brothers and sisters, the son enjoys precedence over the
daughter and the daughter over the daughter-in-law. At Kanheri
and Nasik the donor is mentioned with all his blood relations, and
so high was the social sense in the Buddhist world that the donor
shared the merit of his donations with all his fellow beings. 25
Could we infer from what we have stated above that the joint-
family system was in vogue 7 s6 Could Amaravati No. 38, which
speaks of Khada and ' his daughter-in-law in her house ' show
that it was at the time going to pieces or had done so ?
Women
Women occupied a prominent position in society. The idea of
woman being the chattel of her lord with no rights and privileges
which make life worth living, was quite alien to the period. In
the western cave and Amaravati inscriptions we come across a
bevy of ladies making sometimes very costly donations. A great
number of the exquisitely sculptured rail pillars, toranas and stupa
slabs at Amaravati were donated by ladies. Of the nearly 145
epigraphs from Amaravati 72, out of the 30 at Kutfa 13, out of the
29 from Nasik 16, either record gifts by ladies or gifts in which the
ladies are associated. The Cetiyagharas at Nasik and Ku4a were
founded by ladies. Women joined hands with men in the construc-
tion of the Caitya cave at Karla, 'the most excellent (?) mansion in
Jambudvipa.' The base to the right of the central door carved
with rail pattern, and a similar piece on the left were the gifts of
two nuns. A belt of rail pattern on the inner face of the gallery
was also a bhikkhunl's gift. The remaining pillar on the open
screen in front of the verandah was the gift of a housewife. These
instances unmistakably show that ladies were allowed to possess
property of their own. At Nasik, a Saka lady (Visnudatta) gives
to the Samgha of Nasik more than 3500 kdr?apanas. Ladies even
25. ASWI, Vol. V, Kanheri Inscriptions, No. 9.
26. The word saparivarasa in the Nasik Kanheri and Junnar epigraphs
is translated by Senart as 'with his (or her) next/ He remarks (El, Vol.
VIII, p. 77): "It is, I think, too precise to translate saparivara by 'with
his family ' Parivara may, together with the family or even
excluding it, apply to companions of the donor, fellow- workers or caste -
partners." In Junnar No. 7 (ASWI, Vol. IV) the donor associates with
him his son in the merit of his donations and in No. 9 his parivara. As it is
probable that in both cases he has associated with him the same kind of
persons, parivara would refer to members of the family only. In Kanheri
No. 18, (ASWI, Vol. IV), we have bitiyikaya ca saM porivara and in
No. 27 sarvaseva kulasya.
H.A.--13
98 EARLY HISTORY OF THE ANDHRA COUNTRY
got the titles of their husbands e.g., Mahdbhojl, Mahdrathinl, Bfeoji-
kl f Kufumbini, Gahini, Vaniyim etc. In the Amravati sculp-
tures we often come, across ladies, worshipping Buddhist emblems,
taking part in assemblies, playing on instruments, enjoying music
and dance and entertaining guests along with their husbands. In
one of the panels of an outer rail pillar, 27 we find depicted a disputa-
tion between a chief and another, and the audience consists mostly
of women who are represented as taking keen interest in what is
going on. In some panels they are represented as watching proces-
sions. Widows were to shun ornaments and to be bent on self-
control and restraint and penance. 28
On dress and ornaments, the Amaravati stones, and the figures
cut in the western caves, furnish ample information. Except in
some minor details, the dress and ornaments in vogue on both sides
of the Deccan are the same. The most striking item of the dress
of ladies and men is the head-dress as in the Indus
valley. The former have their hair divided in front and
running down to a knot at the back. Hung on the
knot is a cord of twisted cloth or hair drawn in two or four
rows. Sometimes we come across two strings in four rows ending
in tassels. Some ladies have their hair done in a pointed knot
sideways. 29 In some the knot is done near the forehead with a
string of beads. In western Deccan ladies sometimes cover their
heads with a piece of cloth. 30 Sometimes a thick cloth runs round
their head. At Kuda a lady wears a long cap of conical shape.
Perhaps it is the coiffure done to that shape. Generally a string or
strings of beads adorn the forehead and the knots. Men wore high
head-dress. The general custom was to have hair knotted in front
and covered to a great extent by twisted cloth running down. The
knot was adorned in front by a horse-shoe-shaped or caitya-arch-
shaped ornament. Some Amaravati and Nagarjunikonda men wear
knots unadorned by ornaments. Lay disciples and even servants
have hair done in knots. In one of the Amaravati sculptures a
groom has let the hair run down and secured it by bands at three
places. One of the male figures in the fagade of the Caitya cave at
Kanheri has a very low turban fully ornamented, the ornaments
even hiding the knot of hair on the left.
27. ASSI, Vol. I, PI. XI, Fig. 2.
28. Kuda Sculptures CTI, PI. facing page 10. Nasik inscription, No. 2,
El, Vol. VIII.
29. ASSI, Vol. I, PI. XV, No. 3.
30. The Karla Caitya cave figures.
SOCIAL, ECONOMIC AND RELIGIOUS CONDITIONS 99
Women are as scantily dressed as men, and sometimes even
more so. Twisted cloth running in two or three rows below the
waist and knotted at the right, the ends, however, hanging from
the knots, and sometimes also four or five strings of beads held to-
gether by a clasp, constituted the main part of their dress. Men
wear an undercloth. There is only one instance among our sculp-
tures of a woman covering her breasts. Laymen and monks and
perhaps others also had also a loin cloth, part of which was thrown
over their shoulders. The cloth worn by Brahmans covers them
down to their knees. Some men have twisted cloth thrown over
their shoulders. At Amaravati and Nagarjunikonda we also come
across men in breeches and long tunic perhaps akas. 31
Men and women alike wore ornaments. Heavy rings, some-
times two in each ear, sometimes rows of beads joined together,
constituted their ear ornament. Even kings wore ear ornaments.
The representations of Vasithiputa Siri-Satakani and Siri-Yafia Sata-
kani on their silver coins show us well-punched ears. 32 Both men
and women wore bracelets and bangles with this difference, that
sometimes women wore bracelets covering the whole of the upper
arm, and bangles running up to the elbow. Men did not wear
anklets while all women had them. 33 Sometimes the anklets are
heavy rings, two for each leg, while in other cases each is a spiral
of many columns. Both men and women, even servants, wore
necklaces strings of beads and of medallions. The noses of women
were unadorned as it seems to have been at the Indus Valley. In
this connection it is interesting to note a description of
some of the Bhattiprolu remains given by Rea in his
South Indian Buddhist Antiquities. They are coral beads, beryl-
drops, yellow crystal beads, amethyst beads, double hollow beads,
garnet, trinacrias, pierced pearls, coiled gold rings and gold flowers
of varying sizes.
Luxuries
Jugs, jars, and vessels of attractive shapes, chairs, tables, stools
and cots seem to have been used by many. Whilst kings, great
31. ASSI, Vol. I, PI. xlviii No. 2; ABIA 1927, PI. vi.
32. In this connection it is interesting to note the head-dress worn by
Satavahana kings as represented on their coins. Siri-Yana's head-dress
consists of a strap on the forehead and from the temple locks of combined
hair fall over the strap. Behind the head hangs a string knotted at the
end, probabl> a braided lock of hair. Vasithiputa Siri-Satakani is repre-
sented with short curly hair.
33. For the solitary exception see TSW, 1868, PI. No. LXH.
102 EARLY HISTORY OF THE ANDHRA COUNTRY
padaka on the south-west side of mount Tiranhu (Trira6mi) , 47 and
Sudisana on the southern road in the Govardhana district. 43
The other places mentioned in Nasik inscriptions ar.e Chakalepa, 49
Pimditakavada, Suvarnamukha, Ramatirtha near Soparaga,
Cecimna, Sakha, Anugami and Dasapura. As regards Dasapura
mentioned in the inscriptions of Usavadata and Saka Vudhika,
Senart says: "I see no means of choosing between the Dasapura in
Rajputana (Buhler) , that in Malwa (Bhagwanlal) , or others which
might be added, as Mandasor, etc. In No. 26 we see that some
Sakas dwelt in that place; this is at least a hint that it ought
to be searched for towards the north." 50 D. R. Bhandarkar prefers
Mandasor since Usavadata's inscriptions mention places in the
47. Nasik, Nos. 12 and 20, El, Vol. VIII.
48. As regards the two villages Senart remarks (El, Vol. VIII, p. 65) :
"Buhler seems to entertain no doubt as to the identity of the
village named here with that mentioned at tho beginning of
the following inscription. It is certain that the date of the
donation mentioned there is exactly the same as in the present
epigraph, and that this donation is made in favour of the same sect of
Bhadayaniyas . It is above all evident from the place it occupies,
and from the fact that the following text has been compressed in order
that it might be inscribed here, that that place has been chosen intentionally.
It must, however, be stated that the village called here Pisaji-
padaka, i.e., I suppose Pisachipadraka, gets in the following epigraph the
name of Sudisana, and that the description is not identical in both texts, Pisaji-
padaka being located at the S. W. of Tiranhupavata, and Sudisana at the
south of the Govadhanahara . The two may after all be the same; but
the difference in the name and description deserves to be noted, especially
because a perfect agreement would naturally be expected. Further N. 3
brings in the Sramanas from Dhanakata, who are not mentioned in connec-
tion with the donation which N. 2 records. It is therefore impossible to
affirm that the beginning of N. 3 refers to the present donation; nor is
it absolutely impossible that the king should have consented on the same
day to a double donation, although it would, in that case, be difficult to
understand why he should not have combined the mention of both." It has
already been shown that there can be no connection between No. 3 and
No. 2 as one records an official grant while the other records a non-official
grant. Even the grant of the village of Sudisana was an official grant as
the recording of it in the archives of the state is implied in the words
' Sud na gdmasa ca Sudasan&na vimbadhakarehi anata' (1.14). The
non-mention of the previous donation in a separate inscription is of course
inexplicable under the circumstances.
49. "Chhakalepa or Chhagalepa, a village or a town, a region or clan; has
not yet been identified...." (El, Vol. VHI, p. 90). Liiders, List, Nos. 214,
477, 547, 626 and 937, make a place name more probable. On the analogy
of Parvatiya, Chakalepaka is more probable.
50. El, Vol. VIII, p. 79,
SOCIAL, ECONOMIC AND RELIGIOUS CONDITIONS 103
pradaJcsiria order Bharukacha, Daapura, Govardhana, and Sopa-
raga. As Ujjain which was for some time the Saka capital was in-
cluded in Nahapana's kingdom, the choice is between Mandasor
near Ujjain and Dasapura in Malwa.
The places mentioned in the Junnar inscriptions which from the
context of the inscriptions would seem to have been near the caves
are Puvanadagama, Konacika Va^alika, the village of Mahaveja, the
village of Seuraka. Kapicita would seem to be the name of the
locality in which the cave containing the Junnar inscription No. 15 is
situated. 51 Manamuka^a is the Manmcxji hill.
As regards Mamala and Mamalahara 52 mentioned in Karla
inscription Burgess remarks: "The name Mamala is evidently the
ancient form of the modern Maval (Maul) : the change of medial
ma to va is common in Marathi ; Maval being still the
name of the tract along the Sahyadri or Ghat range,
fully corresponds with the position of the ancient
Mamala. We have thus another proof that the lapse of two thou-
sand years has not changed much the geographical names of Western
India and its territorial divisions. " M Valuraka mentioned as vil-
lage in Mamalahara designates, according to Burgess and Senart,
the modern village of Karla a few miles from the caves. It is pro-
bable that Viharagaon which the caves overlook designates Valuraka.
The places mentioned in the -eastern inscriptions are however
difficult of identification. Velagiri mentioned in the Jagayyapeta
inscriptions would designate the modern village of Jagayyapeta,
probably deriving its name from the hill on which the ruins of Slupa
complex stand. Todatura of the same inscriptions was situated in
the Kammakaratha. Mahakariidurura would seem to have been
outside Kammakaratha as unlike Todatura it is not stated to be in
the Kammakaratha. 54 The places mentioned at Amaravati are
Hiralura, Kevurura, Kudura, Turulura, 55 Devaparavana, Mahava (i) -
nasa(e)la, Mahemkhanaja, Narasa (se) la, Mamdara and Rajagiri.
The last mentioned would seem to be identical with Rayasela; Vira-
pura of the same inscriptions may be the Viripara of the Mayidavolu
51. ASWI, Vol. IV, pp. 30tf.
52. El, Vol. VII, No. 19.
53. ASWI, Vol. IV, p. 113, n. 4.
54. The situation Kammakaratha has been discussed in the chapter on
the Ik^vakus.
55. ASSI, Vol. I, PI. LVII, No. 17 and El, Vol. XV, p. 274, No. 56.
What is read as Turughura in this case is Turulura.
104 BlABLV HISTORY OF THE ANDHfcA COUNTRY
plates. As Chadaka is mentioned twice at Amaravati, and as
Chadapavata is mentioned once in a Nagarjunikon^a inscription,
they must not be far away from Amaravati and Nagarjunikontja. A
feature to be noted is the ura and sela ending in names. Amongst
divisions may be mentioned Tompuki (?) district, Caratha and
Aya-SakasathI 56 in which Suvarnamukha was.
Military Arrangements
An Amaravati inscription of the third century B.C., speaks of a
Sendgopa Mudukutala. In later times Mahasendpatis would some-
times seem to have been in charge of armies. The traditional four-
fold division of the army mentioned in the Hathigumpha inscription
of Kharavela, obtained in the Satavahana dominions. An Amaravati
rail pillar 57 shows vividly the art of war and military organization of
these times. Cities were well defended with high walls, ramparts
and gates. Walls and gateways wer.e often built of brick and mortar
and the gateways were surmounted by tor anas as at Saiici. In
battle the foot soldiers armed with round shields, and short swords,
with a band round their abdomen intended to protect them from the
enemy's spears, led the army. The infantry was flanked by the
cavalry and elephants and the rear was brought up by bowmen.
Soldiers used sometimes long spears. The battle-axe is little differ-
ent from that used in mediaeval and modern times. Mallets also
were used. While the cavalrymen and elephant drivers have tur-
bans, the foot-soldiers have no turban. Does this indicate a higher
status enjoyed by the cavalrymen and elephantmen?
II. ECONOMIC CONDITIONS
Foreign Trade: Seaports: Eastern Ports
The eastern and western ports in the Satavahana empire were
throbbing with trade, though the former were not either as many
or as important as the latter. To take the eastern Deccan first,
Ptolemy notes that Kantakossyla (Kantakasela of epigraphic re-
cords) , 58 Koddura (modern Gu^ur in the Bandar taluq of the Krsna
district), arid Allosygne were ports in the Maisolia region which,
according to the author of the Periplus, * stretched a great way along
56. Liiders, List, No. 1339. .
Liiders looks upon it as the name of a committee.
57. ASSI, Vol. I, PI. XI, fig. 1.
58. Liiders, List, No. 1303, and El, Vol. XX, Nagarjunikoii^a Inscrip-
tions, F.
SOCIAL, ECONOMIC AND RELIGIOUS CONDITIONS 105
the coast before the inland country'. 59 North of AUosygne was
Apheterion, the starting-point of ships bound for Golden Chryse, i.e.,
Farther India, the Malay Peninsula and the Archipelago. A stupa
pillar fragment bearing an inscription of the time of Siri-Yana Sata-
kani was found at Chinna Ganjam (Repalle taluq, Guntur district)
near the sea-shore. A fragment of a stupa pillar was also found at
Kollitippa a few miles to the north of Chinna Ganjam. In the
vicinity of Chinna Ganjam, Rea found the remains of three stupas.
As Jouveau-Dubreuil has pointed out, an inscription dated Saka 1166
mentions Motupalle as a port, 60 and Motupalle is only three miles
to the north-east of Chinna Ganjam. It is therefore highly probable
that Motupalle or its vicinity contained a port or ports 61 through
which a part of the trade of the Maisolia region flowed. And the
Maisolia region was also located between two great navigable rivers.
Lower down the Maisolia region also, there would seem to have been
ports though their names have not been preserved. Coins bearing
the device of ship with masts and the legends, ' sdmi Pu (lumd) visa '
and coins with simha device bearings the legends, 'sdmisa s(i)r(i)'
have been picked up on the Coromandel Coast between Madras and
Cuddalore. 62 Rapson remarks: "The maritime traffic to which the
type 'a Ship,' whether on Andhra, Pallava or Kucumbar
coins, bears witness, is also attested by the large num-
bers of Roman coins which are found on the Coroman-
del Coast". Regarding the migration of the Hindu colonists
to the Far East in the first century of the Christian
era and later, Prof. Nilakanta Sastri remarks " it seems
probable that the movement towards the East was the work of the
entire coastal tract on the eastern coast of the modern Presidency of
Madras, and that the Andhra Country in general, and the Kingdom
of Verigl with it, had a good share in this movement/' 63 Krom re-
59. The Periplus of the Erythraean Sea, ed. Schoff, Sec. 62.
60. 600 of AER 1909. Bulletin of the School of Oriental Studies, Vol. VI,
Part 2, pp. 389 # .
61. The inscription mentions the remission of taxes on articles of export
and import at the harbour of Motupalle alias Bisyuyyakontfapat.tana. Nos.
601 and 602 mention a concession given to merchants at sea.
62. Regarding the coin G. P. 3, PI. CIC, Andhras and Western Katra-
pas etc., Rapson reads ' ( ) Pu ( ) '. Says he (p. 24): "The
restoration lu seems possible, but by no means certain." The two obliterated
letters before pu are sa and ma. After lu there are traces of ma. We
have then 'sami-Pulu (ma) '. Regarding G. P. 2, PL V. Rapson
reads *ara[ ] (X) Samisa, The more plausible way of reading the legends
is jSamwa Siri( ).
63. Telugu Academy Silver Jubilee Volume (English Sec.) p. 11.
HA 14
lOd EARLY HISTORY OF THE ANDHRA COUNTRV
marks: "We understand from Vogel's study, that the history of the
Pallava princes first attains prominence at a time which must be con-
siderably later than the first setting out of the Hindus towards the
east; direct connection between the data here and those from the
other side is not thus to be expected." Again: ''What we know ol
the history the Pallavas concerns the later fortunes of a land which
must have served as the starting-point for the trade voyages of the
Hindus towards the East; they do not relate to that period when such
expeditions had their beginning. We are thus at a time much earlier
than Sivaskandavarman, and it is wholly uncertain if the dynasty
or the kingdom of the Pallavas in general can be taken to a time
earlier than the fourth century; and none can guarantee that the
Pallava script was begun by the Pallavas or during their rule and
had not received earlier the characteristic traits which distinguish it
from other scripts We must leave open the possibility that the
'Pallava 1 script was brought to foreign lands not from the Pallava
kingdom itself but from a kingdom which preceded it in the same
locality." 64
Western Ports
According to the author of the Periplus Barygaza, the Bharu-
kacha of a Junnar inscription 65 and modern Broach, and at his time
a Saka port, was the northern-most port in the Dachinabades. 60 Its
imports and exports so graphically described by him are: Italian,
Laodecian and Arabian wine, copper, tin, lead, coral, topaz, fine and
rough cloth, storax, sweet clover, flint, glass, realgar, antimony;
gold and silver coin, "on which there is a profit when exchang-
ed for the money of the country," and small quantities of oint-
ments. 67 For the use of the Saka Satraps were brought very costly
vessels of silver, fine wines, beautiful maidens of the harem, thin
clothing of the finest weaves, and the choicest ointments. The ex-
ports were spikenard, costus, bdellium, ivory, agate, cornelian, ly-
cium, silk cloth, mallow cloth, long pepper; and "such other things
as are brought from the various market-towns." He adds: "Those
bound for this market-town from Egypt make the voyage favoura-
bly about the month of July, that is Epiphi." The inland commer-
64. Ibid, pp. 10-11.
65. No. 19, ASWI, Vol. IV. It records the foundation of a two-celled
cave by the brothers Budhamita and Buddharakhita (Buddhista) of the
Lamkuoliyas, and sons of Asasama ( Asvasarman) , the Bharukachas.
66. Op. cit., sec. 51.
67. Op. cit, sec. 49.
SOCIAL, ECONOMIC AND RELIGIOUS CONDITIONS 107
cial entrepots from which streams of trade flowed to Barygaza are
Ozene, Paethan and Tagara. To quote the author of the
Periplus : " There are brought down to Barygaza from
these places by waggons and through great tracts with-
out roads, from Paethana carnelian in great quantity and,
from Tagara much common cloth, all kinds of muslins and
mallow cloth, and other merchandise brought there locally from
the regions along the sea-coast." 68 Tagara has been identified by
Dr. Fleet with modern Ter or Thair. 69 The question arises, why
68. Op. cit., sec. 51.
69. JRAS, 1901, pp. 537 ff.
Bhagwanlal Indraji identified Tagara with modern Junnar, IA, Vol.
XIII, p. 366. Fleet has identified it with Ter on the following grounds:
Ter or Thair represents a variation of Tagara which is mentioned undei
that name in two Silahara records (El., Vol. Ill, pp. 267 and 273, 11.
43-44, and CTI, pp. 102-103, 11. 26-27) ; the author of the Periplus says thai
Tagara is nine days' journey from Paithan. As Paithan is twenty days 1
journey from Broach, and Ter is half as distant from Paithan as Paithan is
from Broach, there can be no objection to the identification; modern Tei
is as big a town as Paithan.
Fleet continues (p. 548): "A study of the maps has shown me the
former existence of an early trading route, of which well-marked traces
still remain, from the east coast through Golconda or Haidarabad, Ter and Pai-
than, to Broach There were two starting-points. One was Masulipa-
tam, on the coast, in the northern part of the Kistna district; and the road
from this place took, not only the local traffic from the coast districts, on
the north of the Krisna but also the sea-borne traffic from the far east. The
other starting point was probably Vinukonda, inland, in the southern pan
of the same district, which would serve admirably as a collecting centre
for the local products of the sea-side country on the south of Krisna.
The roads from these two places joined each other at a point about twenty-
six miles towards the east-by-south from Haidarabad, or perhaps at a point
about twenty-three miles further in the same direction. And from thai
point the single road ran in the most natural manner, through easy country
via Haidarabad, Kalyani, Ter, Paithan and Daulatabad, to 'Chandore* and
Markintfa in the west of the Nasik district. And only there, in the Western
Ghauts commenced the real difficulties of the journey "
According to the author of the Periplus most of the seaport towns,
especially Barygaza, were connected with inland market-towns like Paethana,
and Tagara by great tracts without roads; and waggons drawn by bullocks
could do only twenty miles a day. This might have been the case. But
the difficult nature of the country described has to be taken into account.
Nasik official records speak of roads. We have the benefactions of the
natives of Nasik at Barhut of Vaijayanti at Karla, of Bharukacha and Kal-
yan at Junnar, of Sopara at Naneghat and Karla. Usavadata's constant and
wide pilgrimages in the north-western parts are well-known. It is therefore
108 EARLY HISTORY OF THE ANDHRA COUNTRY
should goods from Tagara and Paethan be sent to a distant port like
B'arygaza, in preference to the ports of Kalyan and Sopara which'
were less distant. Was it because as the author of the Periplus
himself says the Sakas had closed the port (Satavahana port) of
Kalyan? According to the author of the Periplus again from Ozene
. were brought to Barygaza all things needed for foreign trade, agate,
cornelian, Indian muslins, mallow cloth, and much ordinary cloth.
Below Barygaza stood the Satavahana port of Sopara, the
Soupara of Ptolemy, the Supara of the Periplus, the Soparaka or
Soparaga of epigraphic records and modern Sopara, a few miles to
the north of Bombay. Of the two Kanheri records which mention
the place, one mentions a jeweller and the other a merchant from
Sopara. The Sagarapaloganas (?) of a Kanheri inscription 70 are
probably the sea-faring traders at Kalyan or Sopara.
Sopara would seem to have been a port of great antiquity. It is
mentioned as Suparapattana in the Mahdvarhsa. 71 The early Bud-
dhist story of Purnamaitra speaks of Soparaka as a great seaport
and the residence of a king Puma, a very prosperous merchant of
the city, who had made several successful voyages in the great
ocean. The Buddha is said to have visited the town and preached
his law to two Naga kings there. Purnamaitra Yanlputra built a
vihara to the Buddha. 72 There are at Sopara even to-day the
remains of Buddhist Stupa 73 Jaina literature and the Rama-
yana also mention Sopara. 74
The greatest port in Satavahana western Deccan was Kalyan,
on the .eastern shore of the Bombay harbour, the Kaliana of inscrip-
tions and the Calliene of the Periplus 75 Of the two Junnar inscrip-
tions mentioning Kalyan, one records gifts by a trader and the other
gifts by a goldsmith, from Kalyan. Of the seven inscriptions at
probable that the less mountainous country was covered with roads whilst
in the mountainous regions only great tracts without roads were found
possible.
70. ASWI, Vol. V, No. 23.
71. VI, 1. 46. The Mahdvaihsa is not older than the fifth century A.D.
But it embodies earlier tradition.
72. IA, Vol. XI, pp. 293-94.
73. Antiquities of Sop&ra, by Bhagwanlal Indraji, pp. 4 ft .
74. Ibid.
75. The various forms in inscriptions are Kaliana, Junnar Nos. 11 and
13, ASWI, Vol. IV; Kalyana, Kanheri No. 15; Kaliana, Kanheri No. 5;
Kaliana, Kanheri No. 18; Kallyina, Kanheri No. 25; ASWI, Vol. V. These
are Prakrtic variations of Kolyflna meaning *blest\
SOCIAL, ECONOMIC AND RELIGIOUS CONDITIONS 109
Kanheri mentioning the port, four record gifts by merchants or their
sons, and two record gifts by artisans (goldsmith and blacksmith).
It is clear that Kalyan was a prosperous trading and industrial
centre. This is also known from the fact that the Sakas captured it
with a view to destroy the resources of their enemies, the Satavaha-
nas. Of the three dated inscriptions at Kanheri two are dated in
regnal years of Madharlputa Sakasena and one in the regnal years
of Siri-Yana Satakani (after 160 A.D.). Could we infer that as the
port was closed by the Sakas in the closing years of the first cen-
tury A.D. it was not functioning as a port when Ptolemy wrote, and
that it regained its old position in the reigns of Ma<Jhariputa Saka-
sena and Siri-Yana? As late as the sixth century A.D. Cosmas
Indicopleustes found it one of the five chief marts of western India
and the capital of the powerful Calukya kings with a trade in brass,
blackwood logs, and articles of clothing.
The other ports mentioned by the author of the Periplus and
Ptolemy and identified with places in the Satavahana western Deccan
are: Mandagora (Mandaragiri) identified with Bankot at the mouth
of the Savitri river, and now a fishing village of no importance;
Palaepatmae (Ptolemy's Balipatna) probably modern Dabhol;
Melizagara (Malayagiri) placed by McCrindle and Miiller at the
modern Jaigarh, formerly a port of importance but now little more
that a fishing village, and by Schoff at the modern Rajpur near
which the Kuda caves are situated; Simylla (the Symulla of Pto-
lemy, the Chimulo of Yuan Chwang and the Cemula of two Kanheri
inscriptions), 76 certainly modern Chaul about twenty-five miles
south of the Bombay harbour; Hippokura, 77 which Campbell would
identify with Ghoregaon in Kolaba; and Byzantion identified with
Vijayadurga the south entrance of the Vaghotan river. 78
Market-towns in the interior: Western Deccan
The market-towns in the interior besides Paithan and Tagara
were Junnar, 79 Karahakata, 80 Nasik, 81 Govardhana, 82 and
Vejayantl. 83
76. Liiders, List, Nos. 996 and 1033.
77. Hippokura is not mentioned in the Periplus. It is however different
from the capital of Baleokuros mentioned by Ptolemy. The latter is an
inland town.
78. IA, Vol. Xm, p. 327.
79. 80, 81, 82 & 83. The inscriptions at Junnar refer to a town near the
caves, and Junnar is situated in a pass in the Western Ghats, which is on
110 EARLY HISTORY OF THE ANDHRA COUNTRY
Eastern Deccan
The market towns in the eastern Deccan would seem to have
been not as many or as important as those in the western Deccan.
Even as early as the third or the second century B.C., Dhannakataka
was a market-town. In some AmaravatI inscriptions Kevurura,
Vijayapura and Narasala (sela?) are mentioned as the residence of
merchants. Liiders, No. 1261 mentions a sethipamukha and a mem-
ber of a guild of merchants from Cadaka, 84
Merchants
Merchants are generally all called vanija or negama. 85 A mem-
ber of a guild of merchants is known as nigama; the alderman of the
guild is called sethin (Sans, sresthin). Sathavdha (Sans. Sartha-
vdha) means 'caravan trader'. The wife of a vaniya is called vani-
yini, and the wife of a sathavaha, a stitakavahim. 86
the line of communication through the Naneghat pass to the ports in the
west. Moreover in the Junnar caves a number of senis or guilds are
mentioned. The traveller's rest house at Naneghat shows undoubtedly that
much trade flowed through it. A Soparaga executed a cistern at Naneghat
hill (Liiders, List, No. 1119) called the Satagara mountain in another inscrip-
tion. ((JBBRAS, Vol. XV, pp. 813 ff). (80) Modern Karadh 70 miles from
Mahad. It is mentioned as a nigama in a Barhut inscription (Cunningham,
Stupa of Barhut, p. 131, No. 16); it is mentioned also in a Ku$a inscription
as *he residence of an ironmonger (ASWI, Vol. IV, p. 87, No. 18). (81) A.
Be$sa inscription speaks of a sethi from Nasik (ASWI, Vol. IV, p. 89 No. 1).
(82) An inscription of Usavadata mentions a number of guilds, and the guild-
hall (nigamasabha) at Govadhana (Nasik No. 12, El, Vol. VIII). (83) Vide
supra.
84. The inscription (No. 1261) runs as follows: "Cadakicasa seMpa-
mukhasa (bha)-daniga(ma)8a suci dfina." Liiders translates it as follows:
"the gift of a rail bar by the pious town Chadakica." On the analogy of
Junnar No. 4 (ASWI, Vol. IV) where Virasenaka is a personal name and
dhammanigama has therefore to be interpreted as 'the member of Buddhist
guild of merchants' as pointed out by Senart, bhadanigama has to be in-
terpreted as 'members of a Buddhist guild merchant.* Chadakicasa would
then mean 'native of Chadaka' and sefhipamukhasa prominent among the
sethis'. Instances of the omission of the personal name in votive inscrip-
tions are not wanting.
85. In Liiders, List, No. 987, both vanijaka and negama are used in such
a way as to show that they had identical meaning,
86. Ibid., No. 30 f
SOCIAL, ECONOMIC AND &ELIGJOUS CONDITIONS in
In our epirgraphs merchants bear the title of gahapati, gahata
(Sans, grhasta) being a title applied to Brahmans who had pas-
sed the pupil or the Brahmacarya stage. 87 According to Senart
the use of gahapati (Sans, grhapati) in Nasik No. 6 "favours the
opinion I have formerly stated, and which I must maintain against
the doubts that have been raised by a learned opponent (Fick,
Sociale Gliederung zu Buddha's Zeit, p. 164) , viz. that grihapati is,
in the Buddhist language, specially restricted to people of
various castes, who are included in the large class of Vaisyas."
In the Saptasatakam we have references to the philandering of the
gahapati with a girl of the hdlika class; 88 a SailarwEwjli inscription
mentions a hdlakiya (cultivator) kudubika U$abanaka whose son
is called gahapati.** In Kanheri No. 15 and Nasik No. 6, the
wives of negamas are called kutumbinis. Therefore kuturiibika
and gahapati are identical titles applied to persons of the culti-
vator class also. In Nasik Nos. 11 and 13, the wife of Rsabhadat-
ta styles herself kntumbirii? Does this show that the word also
meant wife, or does it show that the Ksatriyas also styled them-
selves as kutubin and kutubini?^ 1 Later on kutumbin came to mean
* cultivator.' A point against Senart 's conclusion is that in the Ama-
ravati inscriptions many a gahapati and merchant is mentioned, yet
we have only one instance of the father of a vaniya bearing the title
of gahapati, but even here the latter is not styled a vanija. la the
western cave inscriptions too, not all the merchants bear the title
of gahapati.
Coins and Trade
If the state of trade, internal and external, can be judged by
the coinage i.e., by the variety and number of coins used or issued,
it would appear that western Deccan throbbed most with trade and
industry during the period of the early Satavahanas (third and
second centuries B.C.) , and during the first period of Katrapa occu-
pation (first century). The Naneghat inscription of queen Naya-
nika describes the daksinas given on the occasion of the various
sacrifices performed by the queen and her husband Siri-Satakani I.
They are 1700 cows and 10 elephants, 11000 cows, 1000 horses, 17
87. Ibid., 1091.
88. II, 7, vi, 100.
89. CTI, No. 1, p. 38.
90. ASWI, Vol. V, Kanheri Inscriptions, El, Vol. VIII, Nasik Cave
Inscriptions.
91. Does- it indicate that she was a laic?
ii2 fcA$LY HISTORY OF THE ANDHRA COUNTRY
silver pots and 14000 kdr$dpana$, one horse chariot, 30002 cows,
silver ornaments and dresses, 1100 cows on three occasions, 1000
cows on four occasions, 40001 fcar$dpatias on three occasions, 12
golden ?, and 14000 (?) karsapayas etc. 92 As for the first
period of Ksatrapa occupation, do we not possess a hoard of Naha-
pana's silver coins? Does not Usavadata endow large sums of
money, viz., 70,000 kdrsapanas and 3000 kdrsapanas? The vast trade
of the period is also indicated by the other charities of Usavadata. 93
Some may be inclined to look upon them as the results of oppressive
taxation or successful wars.
We have not a single Satavahana coin of the period from the
first century B.C. to the reign of Gotamiputa Siri-Satakani (end of
the first century A.D.) , Even the latter only restruck the coins of
Nahapana unless it be that some copper coins attributed by Rapson
doubtfully to him, were struck by him. Even among the coins of
the later Satavahanas from western India, we have more coins from
the Chanda district (Central Provinces) , than from Maharas^ra, and
Aparanta.
Eastern Deccan where lead coinage predominates over copper
and where not even a single silver coin has been picked up, 94
would seem to have entered upon an epoch of great commercial and
industrial activity during the reign or Pulumavi II, an activity
which reached its culminating point in the reign of Siri-Yaiia.
Pulumavi issued y 16 ; V 4 ; %; 2; %; %; 7 /s; 1, kdrsapanas. 95 The
reign of Siri-Yaiia not only saw the issue of a large number of coins
which an empire from sea to sea demanded, but a correspondingly
larger issue in the Andhradesa than in the western Deccan. In
addition to coins of the denominations already mentioned 1%, and
92. Rapson has pointed out that the kdr$dpanas of this inscription, as well
as those of Kanheri No. 15 (ASWI, Vol. V) and Nasik inscriptions of
Usavadata, are silver kar?apanas. CIC, Andhras and Western Katrapas,
etc., clxxxiii, n. 1. The ratio was 35 kdrsapanas to one Suvarna.
93. Nasik Nos. 10, 12 and 14a, El, Vol. VIII.
94. The Alluru inscription which is a composite record of gifts, mentions
an investment of 1000 karsapanas as permanent endowment. As this is men-
tioned along with a gift of 53, 32, and 24 nivartanas of land and 500 cows, 64
bullock carts, some Greek lamps and a tank, it is all but certain that only
silver kdrsapanas are meant. It must be noted that this is the only
inscription where a kdr?apana is called purdna or old coin.
95. 1|16 is more probable than 1|12, as the other denominations follow
the division into 16, and as an inscription at Nagarjunikontfa mentions dindri-
Md?aka is the weight five ratis and 16 md?aka$~l kdrsa.
SOCIAL, ECONOMIC A&D RfeLIGIOtfS CONDITIONS US
1%, kdrsdpanas were issued, an indication of increased trade. 96
Matfharlputa Sakasena's and Ca<Ja Sati's reigns did not witness such
varied and large issues. The large lead and potin coins from
Kolhapur, also attest a busy trade in the south-western part of
the empire which was left under feudatories.
Industrial and Commercial Organization
The inscriptions introduce us to various classes of workers
dhamnikas (corn-dealers), mdldkdras (florists), kolikas (weavers),
tilapi$akas (oilpressers) , odayantrikas (fabricators of hydraulic
engines), kdsdkdras (braziers), tesakdras (polishers), kamdras
(iron- workers) , lohavdniyas (iron-mongers) , kularikas (potters?), 97
avesanis (artisans) , and lekhakas (writers) some of whom were in
the service of kings and Mahdbhojas, Cammakdras (leather-
workers), gadhikas (perfumers), suvanakdras (goldsmiths), mani
kdras (jewellers), mithikas (stonepolishers) , selavadhakins (stone-
masons) , and vadhakis (carpenters) . Workers connected with build-
ings are ndyakamisas, ka$hicakas? 8 and mahdkatakas. Among
these workers only gadhikas, kammakdras, avesanis and lekhakas,
are mentioned in epigraphs from the eastern Deccan.
Most of these craftsmen were as well-to-do as the craftsmen of
the Middle Ages. Their artistic taste was something unsurpassed."
It is to their munificence that we owe some of the Buddhist monu-
ments of our period. 100
&6. It is here necessary to bear in mind the remark of Rapson that : "
it is important to insist on the fact that any supposed uniformity in the weight-
standards of the ancient coins of India appears on examination to be
quite illusory. It is impossible to read the various passages quoted from
Sanskrit authors in the Vdcaspatya Dictionary, s.v. 'karsa,' without realising
that the diversity of weights may have been very considerable. This diver-
sity seems certainly also to be proved by the actual specimens, many of
which cannot, in the present state of our knowledge, be satisfactorily as-
signed to any particular denomination." (op. cit. t clxxxii) .
97. "For kularika at least I see nothing better than BUhler's conjecture,
taking it to be, kutala. = 'a potter'." Senart, EL, Vol. VIII, p. 89.
98. According to Buhler, fcodhicaka may be the Gujerati ka^hiyo=:brick-
layer.
99. A gadhika has lavished all his art on an Amaravatl pillar (TSW,
1868 PI. LXXXIX).
100. Mugudasa, a fisherman (ddsa), excavates a cave at Nasifc (Nasik
Nos. 8 and 9, El, Vol. VIII). As the donation of one cave is mentioned in
two epigraphs Senart thinks that the word lena in No. 9, points no more to
the verandah but "to the cell which the same donor Mugftdasa, must have
HA. 15
ii4 EARLY HISTORY OF THE ANDHRA COUNTRY
Some and perhaps most of the crafts and trades mentioned
above were organized into guilds. We hear of a Dhamnikaseni, a
Kasakaraseni, and a Tesakdraseni in Junnar inscriptions; Kolikani-
kdyasenis at Govardhana; of a Kularikaseni, a Tilapisakaseni, and an
Odayantrikaseni we hear in an inscription from Nasik. Each guild
had an alderman called sethin (sresthin) , 101 Guilds had their office
in the town-hall or nigamasabhd. Usavadata's investment in guilds
were read (srdvita) , and registered in the nigamasabhd, 1 ^ 2 Perma-
nent endowments, 103 especially in favour of religious institutions
were sometimes invested in fields, and sometimes in these guilds.
Since Usavadata invests some of his religious charities in guilds,
does it mean that the guilds were looked upon as very stable orga-
nizations, as stable, if not more so than the government?
In India the rate of interest is stated monthly. 104 Some of our
western cave inscriptions show conformity to this method. A
Junnar inscription mentions investment of money in two guilds at
the rate of paonaduka mdsa in one guild and at the rate of pod
mdsa in the other. Usavadata invested two thousand kdrsdpanas in
one weaver's guild at Govardhana, at the rate of pratika per cent
(12% per annum), and 1000 kdrsapanas in another weaver's guild
at the same place at % pratika per cent (9%) . But the normal rate
of interest would seem to have been 12 per cent per annum. 105
Guilds were like the goldsmiths of the Middle ages in Europe
bankers receiving deposits and lending out money.
Land
In the Satavahana empire, as in India at all times, agriculture
was the main industry. Western and eastern inscriptions record
added to his cave. This interpretation seems the more tempting as the second
donation has for its object, to supply with clothes the pavajita, i.e., the
monk residing in the cell." But in our epigraphs there is not a single in-
stance of the confusion between lena and ovaraka; and the money for clothes
is to be applied to the monk or monks living in the cave. It is possible that
while making another donation for monks, the previous donation was re-
capitulated.
101. Liiders translates sethin as 'banker'.
102. A Bhattiprolu inscription mentions a negama (guild), and a number
of persons, the members of the guilds. (El, Vol. II, No. VIII d.)
103. Akhayanivi. We do not hear of permanent endowments in the Ama-
r&vatl, Jagayyapeta and Nagarjunikontfa inscriptions. The Alluru inscrip-
tion however mentions it.
104. Manu, VHI, v. 141.
405. Nasik, No. 17, El, Vol. VIII, Kanheri No. 15, ASWI, Vol. V.
SOCIAL, ECONOMIC AND RELIGIOUS CONDITIONS 115
gifts of villages and fields (for cultivation and plantation of trees) to
monastic institutions. The king had what in the Middle Ages was
called the royal domain. 106 The king did not expropriate the lands
of the subjects but bought them, even when whole villages were
granted to monastic institutions. 107
Sometimes the field was possessed by more than one person;
the share of each person in the field was specified in fractions of a
pana. A Kanheri inscription mentions the owner of an adhapana-
kheta.
Land was subject to more than one tax. 109 Salt was a royal
monopoly. 110
III. RELIGION
Buddhism in Eastern Deccan
The stupas at Alluru, Gummadidurru, Ghantasala, Bhattiprolu,
Gu$iva<jla, and Goli, and the Amaravati sculptures and epigraphs
give us in their own way the history of Buddhism in Andhradesa
from the second century B.C. to the third century A.D. Much cre-
dence may not be placed in the story that the Buddha miraculously
visited the Andhradesa. 111 As Andhradesa is not mentioned among
the countries to which monks were sent by Tissa after
the Third Council and as the Bhattiprolu inscriptions show that
even before 200 B.C. Andhradesa had become a stronghold of Bud-
dhism, it is very probable that it embraced Buddhism long before
the time of the Third Council and the reign of Asoka. It is also
probable that Buddhism spread more quickly among the non-Aryan
Andhra tribes than in Aryan societies. A feature of Bhattiprolu
Buddhism is the worship of the relics of the Buddha (sarira) placed
in crystal caskets which were in turn placed in stone caskets. The
faithful in each village organised themselves into groups e.g.
Sihagothi, Aya-Sakasathigothl etc. 112
The objects of worship at Amaravati are the stupas, small and
big, the sacred tree with the empty throne, the footprints (paduka)
of the great teacher on a stool in front of the throne, 113 the trisula
106. Nasik, No. 5, Vol. VIII: etha nagarasime rajakam kheta.
107. Nasik, No. 12.
108. ASWI, Vol. V, p. 79, No. 15.
109. El, Vol. VH, Karla No. 14.
110. El, Vol. Vm, Nasik NOB. 3, 4, and 5.
111. Waiters, On Ywan-Chwang, Vol. n, p. 209.
112. Bodhagothi, Luders, List, No. 234 and 351.
J13. The footmarks of the Buddha have been objects of,
116 EARLY HISTORY OF THE ANDHRA COUNTRY
emblem, 114 the Dhammacakka on a pillar with an empty throne be-
fore it, relics of the Buddha and great teachers, Sothikapafas (sans.
Svastikapattas) , 115 the Buddha or Naga RajS with the snake hoods
above his head, life size statues of the Buddha, and the Buddha
preaching. These objects of worship are often found in the open
and sometimes in the midst of a hall with caitya arches and rail
pattern bands. 116 To these objects of worship offerings of cloth,
flowers and liquids, brought in vessels of attractive shapes were
made by the rich and the poor, young and old, men and women.
Pilgrimages to Buddhist centres would seem to have been the order
of the day. Men and women are represented in sculptures as being
in ecstatic devotion. The common way of worshipping is kneeling
before the object with joined hands. Sometimes it is falling flat on
the ground.
The Naga cult in Buddhism is also noteworthy. Serpents are
represented as entwining stupas. Nagas and Naga Rajas and their
wives are represented as worshipping the stupa and hearing the
sermons of the master. Both at Amaravati and Nagarjunikonda,
the Buddha is represented as seated on Nagas with their hoods just
above his head. As there is no halo behind these figures, and as
Nagas are sometimes represented on stupa slabs, in the place of the
Buddha preaching, as the figures under mention have turbans and
his followers from very early times. "And special representations of it sup-
posed to have been left by the Buddha himself as that on Adam's Peak in
Ceylon were objects of pilgrimage. And the legends that enumerate the
thirty-two marks of personal beauty or superiority ascribed to the Buddha,
specially mention two beautiful brilliant wheels (cakra) with a thousand
rays on the soles of his feet.'* But though the essential feature in the repre-
sentation of this footmark called carana-nyasa or &kya-carana is the cakra
on the middle of the sole, there were almost always others also and in the
eastern peninsula they have been multiplied largely. The Sapta&atakaik
speaks of the worship of the feet of the Buddha (g. 308) .
114. Regarding the tri&ula emblem and the pillar supporting it Burgess
says (ASSI, Vol. I, p. 30): "The sides of the pillar supporting the trisula are
always represented as in flames, and, as Mr. Fergusson has remarked, this
seems to be the counterpart of the Agni-linga of &va." An examination of the
sculptures shows clearly that what is taken to represent flames, is only a
representation of twisted cloth, much like that worn by men round their
waist and hanging from the top of the pillar. In one of the sculptures
(TSW, 1865 PI. LVin, No. 2) offerings of cloth to the tree is represented.
115. In the Bhattiprolu Stupa 24 silver coins arranged in the Svastika
shape were found along with other relics. * (Rea, South Indian Buddhist
Antiquities, PI. IV, p. 12).
lie. AS&J, vol. i, PI. xxi, 2,
SOCIAL, ECONOMIC AND RELIGIOUS CONDITIONS 117
no upper cloth, it is probable that they are Naga Rajas. One is
tempted to identify them with Nagarjuna, but the snake cult was so
widely prevalent in these parts that such an interpretation must re-
main doubtful until fresh evidence turns up. As we shall see, any
supposed connection between Nagarjuna P'usa and the Andhrade&a
of the second century A.D., is open to doubt.
Western Deccan
The earliest Buddhist remains from western Deccan come from
Pitalkhora and Kolhapur. The former has yielded us a number of
caves and epigraphs which palaeographically belong to the third
century B.C., and which record the gifts of Buddhists from Paithan
and other places. (The most prominent among the donors is a
royal physician). The Kolhapur Stupa has yielded a stone box
and relic casket on the square lid of which is cut in pure Maurya
characters (earlier than the Pitalkhora inscription) , the inscription:
Bamhasa ddnam
Dhamayute-na kdritam.
In one of his former births the Buddha is said to have gone to
Suppara, in Aparanta or north Konkan, at the request of Purna, the
son of a slave girl, who had risen to be one of the chief merchants of
Suppara; then a Brahman and some widows got relics over which
they built a Stupa.
Bhagwanlal Indraji has excavated the remains of a Stupa at
Sopara. According to the Mahdvamsa and a Nagarjunikonda ins-
cription, Aparanta was converted by a missionary Dharhmarakhita,
a Yavana sent by Tissa after the Third Council. Pitalkhora and
Kolhapur show that Buddhism had made rapid strides in western
Deccan even before the Third Council. Next in age to Pitalkhora
come Nasik, Bhaja, and Bedsa. The Karla Caitya cave belongs to
the first century B.C. It is however the first and second centuries
A.D. that constitute the glorious epoch of Buddhism in the western
Deccan. Kings patronised various sects. Rival powers vied with
each other in scooping vihdras or caves at Nasik, and in making
grants of villages, lands, and money to monks spending the vassa in
such caves. MaMbhojas, Maharathis, ministers, and minor officials,
merchants, craftsmen and ladies of all rank and denominations, vied
with one another in making donations to the order. Monks and
nuns vied with laymen in donating caves, cisterns, caityas, stupa-
marbles and permanent endowments. 117 They readily joined with
117. Kanheri Nbs. 17 and 21, ASWI, Vol. V,
118 EARLY HISTORY OF THE ANDHRA COUNTRY
laymen in such enterprises. No less than six monks and nuns join-
ed hands with Yavanas and laymen from VejayantI, NSsik, etc., in
constructing the Karla Cetiyaghara. Were these expenses met from
the savings effected out of the cloth money and the kusanamula
given to them? Did enterprising monks who were enjoined to beg
for alms only, go about collecting money for such pious works?
Whatever might have been the case, the possession of money by
monks and nuns was evidence of relaxation of the rigid rules of the
Order. The Buddhist monuments at Kuda, Mahad, Kol, Bhaja,
Redsa, Karla, Junnar, Nasik, and Kanheri are such donations. A
Kanheri inscription mentions the construction of a Cetiyaghara,
upathdnasdld (hall of reception) , three cells in the Abdlikdvihdra at
Kalyan, a Cetiyaghara and thirteen cells at some vihdra at Pati-
thana, a kuti (temple), and a kodhi (hall) at Rdjataldka in
Paithanapatha, and a saghdrdma at the vihdra at Sadasevaju(?) 118
Sects: Buddhist Sects in western Deccan
Our epigraphs introduce us to a number of sects. The Bhadaya-
niyas were the most favoured at Nasik and at Kanheri. Dhammot-
tariyas flourished at Sopara and in the town near the Junnar caves.
The Mahasamghikas had their stronghold at Karla and its vici-
nity. Both Bhadayaniya and Dhammottariya are subdivisions of
the Theravada school. In tv/o Junnar inscriptions we have Sidha-
ganesu Apardjitesu and Apajitesu gane. 119 Apajitesu is evidently a
mistake of the scribe or the engraver for Apardjitesu. Siddhagana
denotes a holy assembly. Another Junnar inscription 120 mentions a
gandcdrya. As in our inscriptions teachers are generally mentioned
with reference to sects, 121 gana of the inscriptions under reference
also refers to a school or sect. The Aparajita sect is not mentioned
in the books.
It does not however mean that only one sect flourished at a place
or in a group of caves. Caitikas flourished at Nasik, and Usava-
data's inscriptions show that monks of different sects kept the vassa
in the same cave. We know that Mahisasakas, Bahusutiyas, Apara-
mahavinaseliyas and Ceylonese monks and nuns lived in the Nagar-
junikonda valley. A preacher of the Dhammottariya sect donates
118. Liiders, List, No. 988.
119. ASWI, Vol. IV, Nos. 25 and 20.
120. No. 17.
121. ASSI, Vol. I, No. 49, p. 105; ASWI, Vol. V, No. 4, pp. 75jff. ASR
31906-07, pp. 96 f; Luders, List, Nos. 1158 and 1171.
SOCIAL, ECONOMIC AND RELIGIOUS CONDITIONS lid
a pillar of the Karla Cetiyaghara which would seem to have been the
property of the Mahasamghikas.
Unlike Gotamiputa Siri-Satakani and Pujumavi II, Usavadata
patronised all the sects. At Karla and Nasik, the benefits of his
donations are assigned to ascetics of every denomination or resi-
dence, who could be brought to take their abode in the caves at
Karla and Nasik, along with their resident hosts, during the
vassa. 122 Another foreigner donates a navagabha man^apa at
Karla to the Samgha as the special property of the Mahasamghikas.
Senart remarks that this grant seems "to stipulate that gifts attribut-
ed to particular sects should be meant for monks of every origin and
of every denomination without distinction; compare No. 13, 1.4" 123 .
On the other hand Gotamiputa Siri-Satakani's donation at Karla is
* for the support of the sect of the Mahasamghikas of the mendicant
friars dwelling here in (these) caves of Valuraka . . . . ' Balasiri's
cave at Nasik was for the acceptance of the Bhadayaniyas. The
official grants of land or village recorded in Nasik Nos. 3, 4, and 5,
were in favour of the same sect.
Eastern Deccan
The Amaravati epigraphs mention some sects that flourished
during the Satavahana period. The earliest among them would
seem to be that of Caityavamda (Caityavada) , or Cetika, or Ceti-
kiya, which is mentioned in four epigraphs. This is the only sect
mentioned both in eastern and western inscriptions. Since an Ama-
ravati epigraph speaks of Cetikas at Rajagiri, and as the commentary
on the Kathdvatthu mentions Rajagirika as one of the Andhaka sects,
it is probable that this sect was an offshoot of the Cetika nikaya.
Whilst the Pubbasela (mentioned in the Alluru inscription), and
Avarasela schools, (Andhaka schools), are known to the commen-
tator on the Kathavatthu, the Mahavinasela (not to speak of its later
offshoot Aparamahavinasela) and Ayira-Utayipabhaha nikdyos 124
122. Eto mama lene vasatdnam cdtudlsasa bhikhusaghasa mukhdhdro
bhavisati, Nasik No. 10. Samghasa cdtudisasa ye imasmim lene vasamtdnam
bhavisati civarika, No. 12, El, Vol. VIII.
123. Mdhdsamghiydnam parigaho saghe cdtudise dina, Karla No. 20, El,
Vol. VII, p. 71.
124. Utayipabhaha is perhaps Sans. U tiara Prabli&sa. Prabhasa is the
name of a place of pilgrimage in western Deccan. (Nasik, No. 10, El, Vol. VIII)
Most of the schools that took their rise in the Andhradesa, derive their names
from places, presumably places where the sects in question took their origin.
'Ayira' or 'aira' (Sans. Arya) is at Alluru, Amaravati, and Nagaifunikon^a
120 fcAftLY HISTORY OF fHE ANDHRA COUNT&tf
are not so known. The latter group was therefore later than the
commentary on the Kathavatthu. But the epigraphs which men-
tion them cannot, however, be ascribed to a period later than the
second half of the second century A.D. The Mahdcaitya at Amara-
vati was dedicated to the Caitikas. As another but smaller Stupa in
the same place was dedicated to the Utayipabhahis they were per-
haps an offshoot of the Caitikas. Rajagiri would also seem to have
been a stronghold of the Caitikas. Each sect had its Mahdnava-
kammas and Navakammas, monks some of whom were sthaviras,
mahdsihaviras and bhadantas. 125
Monks and Nuns
Monks are called bhikhus, pavajitas, samanas, and penujia-
pdtikas. Nuns are called samanikas, pavajitikds, and bhikkhunls.
It is no wonder that the flourishing Buddhist communities in western
and eastern Deccan abounded in great teachers. In western Deccan,
mahasthaviras, sthaviras, bhdnakas, and tevijas (sans. Traividyas
those who know the Tripitaka; also adopted by the Buddhists as an
epithet of arhats) 126 trod the land, enlightening the faithful on the law
of the master. In eastern Deccan, monks, nuns and laymen flocked
to teachers versed in the Vinaya and Dharhma (Dhammakathikas)
and had bhdna under them. Even nuns were teachers (upajhiya-
a title prefixed to the Samgha in general. While in the western cave inscrip-
tion 'aya', 'ayya', and 'ayira' (arya) is a title prefixed to the names of Buddhist
arhats, teachers, monks and nuns.
125. Navakamma is a religious building dedicated by some lay member to
the Samgha. The superintendent of such work is Navakammika. The CwZ-
lavagga VI, 5Vinayapi^akam supplies us with the following information.
" If the buildings were for the Bhikkhus, then a Bhikkhu, if for the Bhikkhunls
then a Bhikkhuni, was appointed to superintend the works in order to
ensure the buildings being in accordance with the rules of the Order as to
size, form, and object of the various apartments." (ed. Oldenberg, Trans.
SBE, Vol. XX, pp. 189 ft n.). But a Nagarjunikon^a epigraph states that a
Cetiyaghara was erected for the them of Ceylon, in the Nagarjunikontfa valley,
and yet the superintendents were not theris, but three monks, theras Camda-
mukha, Dhammanamdi and Naga. The Navakammikas for the Caitya Cave at
Kanheri were the theras, bhadantas, Acala, Gahala, Vijayamta, Bo(dhika),
and Dhammapala. Along with these Navakammikas is mentioned as Upara-
khita, a monk, whose function, however, is not clear. As sometimes ka is
used for kha the Uparaka of an AmaravatI inscription (Vol. XV, No. 33,
p. 269) may be identical with Uparakhita. The Kanheri inscription cited
above mentions along with Navakammikas a Samdpita a layman and the
son of a merchant.
126. Rhys Davids, Buddhist Suttas, p. 162.
SOCIAL, ECONOMIC AND RELIGIOUS CONDITIONS m
yini), and had scores of female pupils (atevasini) under them.
Some monks and nuns were persons who had led the life of
grhaathas. Monks and nuns were recruited from the lowest classes
also. 127
The monks spent the rainy season (kept their vassa) in the
caves scooped out on prominent rocks or in monasteries built by
the faithful. 128 The remaining part of the year was spent in
religious tours. That is why most of the Buddhist monuments
were erected in trade centres like Dhamnakataka, Kalyan, Paithan
and Nasik, and at Karla, and Junnar which are situated in the
passes leading from Konkan to the Ghats. The caves at Kanheri,
which is near the sea and the sea-port of Kalyan, and Kuda, Mahad,
and Chiplun situated on creeks, show that monks and nuns travel-
led by sea also.
Monks and nuns lived in caves called lenas (Sans, layana) or
viharas. Viharas cut out of rock, open with a verandah; inside
is a hall surrounded on three sides by rows of cells, each with a
stone bench for the monks to sleep on. Structural viharas were also
built on the same plan. Attached to the vihdra or as parts of it
are bhojanamatapa or bhojanacatusdld (refectory) , upathdnasdld
(hall of reception), saghdrdma, 123 pdniyapodhis, and sdnapodhzs.
The bhojanamatapa at Junnar 130 is an open hall 19 feet wide by 14^
feet deep and 8 feet high, with a bench round the three inner walls;
the upathdnasdld at Karla is a hall 21 feet wide by 11 feet deep
and seven feet high, and is the upper story of a vihdra 24 feet wide
by 22 feet deep and 7% feet high, with ten cells four in the back
and three on each side, but half of them in an unfinished state. The
vihdra hall is sometimes called kodhi. Thdnaka would also seem
to be another name for vihdra, as the cave in which the inscription
containing the former word is incised bears a close resemblance
to the small vihdra at Bhaja. The Buddhist temples attached to
viharas are called Cetiyagharas or Selagharas, or Cetiyakodhis.
Cetiyagharas are of two main types, both ddgoba shrines. One is
the flat-roofed vihdra like cave with a ddgoba shrine at the back
and with cells on the sides or in front of the hall. The other is
127. ASSI, Vol. I, No. 36, p. 91.
128. In Kanheri No. 18, ASWI, Vol. V the vassa is said to have been
kept in summer.
129. Luders translates sagh&rdma No. 988 as 'monastery'; but in the inscrip-
tion (ASWI, Vol. V, No. 6, p. 77) it is spoken as part of a vihftra. Ar&ma
means 'garden' and saghdrama means 'the garden around the monastery*.
130. ASWI, Vol. IV, p. 30.
H.A, 16
122 EARLY HISTORY OF THE ANDHRA COUNTRY
vault-roofed, has horse-shoe shaped windows over the entrance
and has an interior consisting of a nave and side aisles with a small
circular Stupa at the inner circular end, the object or centre of cult.
To the former class belong those at Mahad, Kuda, and Junnar. In
a Cetiyakodhi at Mahad the ddgoba is in half relief. At Ku<Ja the
abacus of the capital of the Cetiya just touches the roof and the cell
has a stone bench or bed. At Junnar the Cetiyaghara consists of a
verandah with a flat roof forming the porch to the cave. Behind the
arched nave of a Cetiya stands a dagoba of the same style as at
Bedsa. The next step was to pierce the rock over the verandah
with an arched window. The ddgoba shrines at Bhaja, Bedsa,
Karla, Nasik, Korwjane, and Kanheri, some of them belonging to
a period considerably anterior to the Ctmstian era, are of the latter
type. One on the Manmodi hill (Junnar) is of this type. But
those at Nasik, Kondane, Bhaja, and Bedsa, have no screen in
front, though at Bedsa the returns of the rock at each side of the
facjade of the Caitya cave favour the idea that something in the front
was intended. Burgess remarks: "Judging from the examples at
Bedsa Karle, and Kaqheri, it seems as if the great windows in the
original wooden structures from which these Caitya caves were
copied were always covered by a screen in front, which partially
hid them while it protected them from the weather. 1 " 131
Brahmanical Religion
Brahmanism was also in a flourishing condition. Most of the
Satavahana kings were followers of the Brahmanical religion. The
third king of the line performed a number of Vedic sacrifices and
named one of his sons Vedisiri. 132 In the Saptasatdkam an antho-
logy of erotic verses attributed to Hala, adoration is paid to Siva. 133
Later Satavahanas were also followers of the Brahmanical religion.
Gotamiputa Siri-Satakani was not the only supporter of the Brah-
mans. He was not only learned in the traditional lore, but emu-
lated epic heroes like Rama, Kesava, Arjuna, Bhimasena, and
Puranic figures like Nabhaga, Nahusa, Janamejaya, Sagara, Yayati,
and Ambarisa. Since Gotaml speaks of Kailasa, were she and her
son devotees of Siva? Another king bears the name Yana.
131. ASWI, Vol. IV, p. 39.
132. The minister to whom the daughter of prince Hakusiri was given
would seem to have been a Buddhist, as he bears Hie name Arahalaya, and
as his daughter excavates the only Caityagrha at Nasik, (Nasik, No, 19, El,
Vol. VHI).
133. First and last verses.
SOCIAL, ECONOMIC AND RELIGIOUS CONDITIONS 128
U?avadata's inscription 134 speaks of Carakas, a special category of
ascetics at Ramatirtha in Soparaga, Suvarnamukha, Govardhana and
Plmditakavada. Usavadata's inscriptions however, show that
Brahmanism was more flourishing outside Satavahana dominions,
viz., in Gujerat, Kathiawad, Rajaputana, and Ujjain; all his Brahma-
nical austerities are located in them. 135
The Naneghat record begins with adoration to Dharma, Sam-
karsana, Vasudeva, Indra, the Sun and the Moon, the guardians of
the four quarters of the world viz., Vasava, Kubera, Varuna and
Yama. The Saptasatakam mentions wooden images of Indra which
were worshipped. 136 Worship of Krsna is indicated by the names
like Govardhana, Krsna, and Gopala. In the Saptasatakam we
find the Krsna legends fully developed. Here Krsna is called
Madhumathana 137 and Damodara. Gopis and Yasoda are also
mentioned. 138 We also hear of the jealousy of shepherdesses
against Radha. 139
Names like Sivapalita, Sivakhadila, Sivadatta, Kumara etc.,
point to a worship of Siva, and Skanda. The Saptasatakam fur-
nishes us interesting data in this direction. In the opening and
closing verses Pasupati and Gauri are adored. Temples of Gauri
are mentioned in gdthd 172. Siva is also called Paramatadhipa
(Prakrt. Paramahahivam) in gdthd 440. Kdpdlinis or ash covered
and skull-bearing women ascetics are also mentioned. 140 Ganesa
is mentioned as Ganadhipati. 141
134. Nasik, No. 12, El, Vol. VIH.
135. According to the Mah&bharata, a forest near Sopara was in times
of yore, the scene of austerities and sacrifices performed by kings. It also
contained the holy shrines of Vasu, of the Marutganas, of Asvini, Vaivasvat,
Aditya, Kubera, Indra, Vi^nu, etc. (Vanaparvan, Chapter CXVIII).
136. Saccam cia katthamao
Suranaho, jena haliadhute \
Hatthehi kamaladalako
Malehi cikko na pallavio \ \
Weber, Dos Saptacatakam des Hala, p. 470, g. 864.
137. Op. cit, p. 323, g. 657.
138. Ajja vi valo Ddmo-
aro tti ia jampe jasoae \
Kanhamuhapesiaccham
nihuam hasiam Vaavdhuhim \ \ g. 112.
139. Weber, op. cit., p. 31, g. 89.
140. Gathc , 408.
141. Gdthas 403, 372,
124 EARLY HISTORY OF THE ANDHRA COUNTRY
Names like Vinhupalita, Venhu, and Lachinika point in the
same way to the worship of Vinu. In the Sapta&atakam, Hari
or Trivikrama is said to be superior to other gods. Birth of
Lakshxm from the ocean of milk is also mentioned. 142
In conclusion it may be noted that one of the interesting reli-
gious data supplied by Saptasatakarh is the vrata of fire and
water. 143
142. G&thda 411 and 388 respectively.
143. Gdth4, 135.
CHAPTER VII
THE IKSVAKUS THIRD CENTURY A.D.
The Purdnic label for the Ik$vdku kings
The Iksvakus of the Nagar junikontjla and Jaggayyapeja records
are none other than the Puranic Sriparvatiyas 1 i.e., the dynasty
whose capital or home or kingdom lay in the Srlparvata region, 2 also
called Andhras 3 of the lineage of the servants (bhrtydh) of the
' Andhra Satavahanas.' 4 This identification which Mr. K. P. Jaya-
swal 5 was the first to suggest is supported by the following pieces
of evidence 6 : The Nagarjunikonola epigraphs make it clear that
during the period under review ' Sriparvata ' signified not any parti-
cular hill on the Nagar junikoruja site or ' Srlsailam ' 7 but the whole
range of Nallamalai hills of which the hills surrounding the Nagar-
junikon^a plateau and the ' Srisailam ' peak are offshoots. 8
1. a, c Mt. The rest '6riparvatiya,' Pargiter, Purana Text of the
Dynasties of the Kali Age, p. 46, n. 30.
2. ' Chakalepakiyasa ' 'of an inhabitant of Chakalepa (Nasik No. 17, El,
Vol. VIII, p. 90 ; ' Padukulikiya ' (Liiders, Nos. 571 and 576) ' of an inhabitant
of Paclukulika.'
3. Bd. and Vs. 'Andhrabhrtya.' Pargiter, op. cit., p. 72, n. 5.
4. In the early Ml, Va, Bo!, and Vs. accounts.
5. JBORS, 1933, Parts I and II, p. 171.
6. These were not worked out by Mr. K. P. Jayaswal.
7. ASK, 1926-27, p. 186 and ASSI, Vol. I, p. 7.
8. In the Nagarjunikon^a ins. giri is the word for 'peaks' and 'hillocks'
(Cula-Dhaihmagiri is certainly the hillock now called NaharjUJabotfu on
which the apsidal shrine built by Bodhisiri stands. Ins. F.). So 'Siri-
pavata' of the same inscription cannot refer to a peak or hillock but to a
whole range. The later riparvata in the Kurnool District makes it improba-
ble that the hills surrounding the Nagar junikonda site monopolised that
name.
Scholars like Burgess and Dr. J. Ph. Vogel take seriously the
Tibetan tradition preserved by Taranatha that Nagarjuna, the expounder
of the Madhyamika philosophy (second century A.D.), lived at Sri-
parvata. Whilst they are agreed in identifying it with Hiuen-Tsang's
Po-lo-mo-lo-ki-li where a Satavahana is said to have quarried a monastery
for Nagarjuna, Burgess identifies them both with 'grlsailam' in the Kurnool
District, and Dr. J. Ph. Vogel and the Epigraphy Department with Naha-
raHabotfu or Nagarjunikonfla, a lofty hill overlooking the Krsna at the north-
ern end of the plateau. The Chinese pilgrim places Po-lo-mo-lo-ki-li
in Daksina Kosala in a place 300 li (50 miles) to the south-west of its
H.A.-17
126 EARLY HISTORY OF THE ANDHRA COUNTRY
No dynasty other than the Ikvaku could have ruled over the
Krna-Guntur region immediately after the Satavahanas. The
ornate alphabet, with long verticals, of the Iksvaku records shows
only slightly developed forms over those of the Chinna Ganjam
inscription of Siri-Yaiia Satakani and the Jangli Gurwju inscription of
Pulumavi III (probably the last of the Satavahanas). That the
Iksvakus were once ' bhrtyas ' of the Satavahanas is shown by the
Satavahana metronymics and prefixes to the names which they bear
capital. While speaking of T'o-na-kie-tse-kia (Dhanyakataka where the
Nagarjunikonda plateau would have lain) he speaks of neither Nagarjuna nor
his monastery. To identify Po-lo-mo-lo-ki-li of the Daksina-Kosala, which is
placed by Hiuen Tsang 1200 li to the north of T'o-no-kie-tse-kia, with ' Sri-
parvata' in the Guntur and Kurnool Districts is to go too wide of the mark.
General Cunningham has identified Daksina-Kosala with the province of
Vidarbha, modern Berar, and its capital with modern Nagpur AGI, p. 595.
This agrees with the Tibetan tradition that Nagarjuna was a native of
Vidarbha (Wassiljeu, Appendix to Taranatha, pp. 301, 303). The Tibetan tradi-
tion that Nagarjuna surrounded the Stupa at Dhanyakataka (the Amaravatl
Stupa) with a railing is supported by none of the extant Amaravatl rail
inscriptions of the second and third centuries A.D. N. Dutt has pointed out
(1HQ, Vol. VII, p. 639) that the Gandavyuha, a work of about the third
century A.D., speaks of Dhanyakara as a great city of Daksinapatha and
a seat of Manjusri, who lived in an extensive forest and converted a large
number of Nagas and inhabitants of the place, but refers neither to Nagar-
juna nor to Sriparvata.
A Jaggayyapeta inscription which, on palaeographical grounds, should be
ascribed to the seventh century A.D., mentions Candraprabha, his teacher Jaya-
prabha and the latter's teacher Nagarjuna (Nagarjunacarya, ASSI, Vol. I, p. 112.
PI. LXIII) . The Sadhanamald mentions a tantric Nagarjuna, one of the 84
Mahasiddhas, who came after Sarha. B. Bhattacarya places the former in the
seventh century A.D. (SddfianamaZd, Vol. II, Intro, xliv-xlv, Gaekwad's Orien-
tal Series, XLI). The tradition preserved in the Pag-bsam-ljon-bzan (p. 86;
is that, according to the account of the 84 Mahdsiddhas, one Nagarjuna was
born at Kahora, a part of Kanci, and educated at Nalanda, where he practised
the Mdhis and visualised the goddess Tara. He came to Sriparvata,
(IHQ, Vol. VII, p. 637). Taranatha also mentions him (ibid., 638, n. 1).
While there is thus evidence, literary and epigraphic, for connecting the
second Nagarjuna with Sriparvata, there is at present no evidence which
ellows us to associate the first Nagarjuna with Nagarjunikonda.
"Hie Matsya Purdna speaks of a family of ri-Parvatiya Andhras,
which may refer to a petty dynasty either at $ri-$ailam itself or across the
river at Chandaguptapatnam in the vicinity." ASSI, Vol. I, p. 7.
According to Prof. Rapson, "they were probably the Cutus, who
rose to power in the western and southern districts after the reign of
grl-Yajfia" CIC, Andhras and Western K?atrapas, Intro. Ixix. But the
identification suggested here makes Rapson's conjecture wrong.
THE IKVAKUS THIRD CENTURY* AJX 127
(e.g., siri and sami) 9 ; and it will be shown below that they were
Mahdtalavaras under the Satavahanas. Names into the composi-
tion of which ' Skanda ' enters and the * anaka ' ending in names
also point to Satavahana influence. 10 The Ik^vaku capital Vijaya-
puri is, in inscription F, stated to have been situated to the west
of the Lesser Dharhmagiri (Naharallabddu mound.) 11
The find of nearly 148 lead coins of the Satavahana period at
Nagarjunikonda 12 they are now in the Calcutta Museum and are
said to be much corroded indisputably shows that the kingdom of
the Iksvakus or at least a part of it was included in the Satavahana
empire. Whilst then, the south-western parts of the Satavahana
empire fell to the Cutus, and the western parts to the Sakas, Abhi-
ras and Gardabhillas, the eastern parts passed into the hands of
the Iksvakus.
Duration of the dynasty
According to the Puranas, there were seven kings in the dy-
nasty. But epigraphy has disclosed the names of only three. As
regards the duration of the dynasty, the Puranic account is far
from being clear. The Matsya which gives us the oldest version
has ' dvi pancaSatam' which, according to Pargiter, may mean 52
or 100. The Vdyu and Brahmdnda accounts, which according to
him are corrupted, give ' dve ca satam.' However, c Vayu which
next to the Matsya gives us the oldest version has ' dve arddha
9. Ins. L. El, Vol. XXI.
10. Kamdasiri, A2, C3, etc., Kharhdacalikiremmanaka, B4.
11. "Siripavate Vijayapuriya puva-disd-bh&ge vih&re Cula-Dham-
magiriyam cetiya-gharaih . . .at Siripavata (the Nagarjunikonda site with
the hills which form parts of the range which went by that name) , a cetiya on
the Cula-Dhammagiri standing to the east of Vijayapuri." In the central part
of the valley which is now marked by cultivation and which is certainly
to the north-west of the Naharallabodu, Mr. Longhurst discovered a palace
site (ASR, 1928-29, p. 104) . From the elaborate ornamentation and the
curious semi -classical objects portrayed on some of the pillars, it would
seem that they once supported the roof of some royal palace. The pillar
set up in memory of Siri-Camtamula lay buried in the north-western part
of the valley, not far from the river. Moreover, several ruined mandapas
or pavilions mark the site of the ancient city. Finally, the plateau shut
in by hflls on which there are remains of fortifications offered an ideal
site for a capital. The Vijayapura of the Amaravati inscription (ASSI,
Vol. I, p. 85, No. 30, PI. LVm) is perhaps identical with the Vijayapuri
of our inscription.
12. ASR; 1928-29; p. 103.
128 EARLY HISTORY OF THE ANDHRA COUNTRY
fata' which is undoubtedly 52. This conclusion is arrived at in
another way. The earliest Matsya account which mentions the
seven griparvatiya Andhras ends with the mention of Kilakila
kings; even the Vktaka Vindhyasakti is not mentioned in this
recension. The synchronism between Vakataka Rudrasena II and
Devagupta (Candragupta II 380-419) 13 would place Vindyasakti
between 260 and 285 A.D. 14 The earliest Matsya account was,
therefore, closed before 260 A.D. 15 The Satavahanas could not
have disappeared from the political stage before 205 A.D. The
Iksvakus, their successors, could, therefore, have ruled for only
52 and not 100 years. We know from inscriptions that Siri-Vlra-
purisadata and Ehuvula Camtamula divided between themselves
at least 31 years. A reign of 15 years may be assigned to Siri-
Camtamula, 'the Unobstructed/ who is credited with many sacri-
fices and danas and who seems to have lived to middle age. 16 The
short reigns of the last four kings would be evidence of the
troublous times. 17
Home of the Iksvakus
Dr. Sten Konow seeks their home in the western Deccan. 18
The sources of his suggestion are the * anaka ' suffix to personal
names in the Nagarjunikonda and western cave incriptions and
the possibility of explaining some terms in the former through
Kanarese. According to him, * Khanda ' is Kanarese c Kanda '
meaning ' child.' ( Caliki-remmanaka * is probably Kanarese
r Calikiranaka > ' moon.' " It also strikes me that Kanarese
karrambu means * envy.' " " The K for s also points to Kanarese."
We may point out, however, that ' Khanda ' is a Prakrt form of
1 Skanda.' ' Karumbudhina ' is a contraction for ' Karumbudhi-
naka' 19 and l karum y in Tamil means 'black* (adj.), and 'anaka'
13. Allan, Gupta Dynasties, pp. xxxiv, ff.
14. Vindhyasakti, Pravarasena I, Gautamiputra Rudrasena I, Prthivi-
sena I and Rudrasena II, CII, Vol. Ill, p. 235.
15. Pargiter arrives at the result in another way. Op. cit., Intro, xxv,
Sec. 44.
16. Vide infra, pp. 132-4.
17. Dr. J. Ph. Vogel says (El, Vol. XX, p. 6): "Dr. BUhler's assump-
tion, based on palaeographical evidence that Siri-Vlrapurisadata flourished
in the third century of our era, may be accepted as probably correct." The
identification suggested and the chronology worked out above definitely
place the Iksvakus in the first half of the third century A.D.
18. El, Vol. XX, pp. 25-26.
19. Virarhna in Inscription F. is a contraction for Vlramnaka,
THE K$VAKUS-THIRD CENTURY AD. 129
is a name ending ; and names like Ehuvula, A^avi-Camtisiri and
Damila-Kanha betray Tamil influence. Since in an AmarSvatl
inscription of the time of Vasithiputa sami Siri-PuJumavi, an ordi-
nary person bears a name with ' anaka ' suffix, 20 the western influ-
ences shown by the Nagarjunikonda inscriptions are best consi-
dered as the result of Satavahana rule over the eastern Deccan
and have no significance for the question of Ik?vaku origins.
According to Biihler and Prof. Rapson 21 the southern Iksvakus
were Rajputs of northern descent a dynastic drift of which we
have instances like the Mauryas of Konkan, the Guttas of Guttal
and the Colas of Renandu. According to the Vayu Purana, 22
Iksvaku, the progenitor of the Solar race, was the eldest among
the nine sons of Manu; he ruled from Ayodhya and had a hundred
sons of whom Vikuksi was the eldest and successor; of the other
sons fifty were entrusted with small states in the north and 48 in
the south. According to the Visnu Purdna, Kosala (southern) was
founded by Kusa who ruled from Kosasthali. 23 The foundation of
Asmaka and Mulaka on the upper Godavari is ascribed to princes
of Iksvaku descent. The Iksvaku drift into the AndhradeSa must
have taken place very early for them to have merged in the Andhra
tribe, for according to the Puranas the Iksvakus were Andhras.
A Kanarese work entitled Dharmdmrta affords evidence of an early
drift. In the time of the 12th Tzrihankara Vasupujya (third or
second century B.C.), Iksvaku Yasodhara of Anga carved a king-
dom for himself in the Vengi country, to use the later designation of
the heart of the Andhradesa, and founded the town of Pratipala-
pura identified by Mr. M. S. Sarma with Bhattiprolu. 24
20. Vide supra, p. 66; Liiders' List No. 1248.
21. IA, Vol. XI, pp. 256 ff.; CIC, Andhras and Western Ksatrapas; xliv.
22. Chap. 88, 8ff.
23. Visnu Purdna, Hall's edition, ii, 172, n.
24. In a paper on Jainism in South India read before the Archaeological
Society of South India.
Scholars like Burgess (ASSI, Vol. I, p. Ill) and Caldwell
(Comparative Grammar of Dravidian Languages^ p. 115) look upon the
Iksvaku descent claimed by the southern princes as an idle boast. Writes
Caldwell : "The Aryan immigrants to the South appear to have been generally
Brahmanical priests and instructors rather than Kshatriya soldiers; and the
kings of the Pandyas, Cholas, Kalingas and other Dravidians appear to have
been simply Dravidian chieftains whom their Brahmanical preceptors and spiri-
tual directors taught to imitate and emulate the grandeur and
cultivated tastes of the Solar, Lunar and Agnikula races of kings.*' What,
however, invests the claim of the Iksvakus with authenticity is that while
130 EARLY HISTORY OF THE ANDHRA COUNTRY
Rise of the lk$vakus
The Alluru Brahmi inscription (Alluru is a village in the Nandi-
gfima taluq of the Krsna District) discovered a decade ago throws
welcome light on the rise of the Iksvakus to power. 25 The ayaka-
pillar inscriptions from Jaggayyapeta 26 prove beyond doubt that the
Nandigama taluq or part of it was included in the Iksvaku king-
dom. The Alluru inscription which, on palaeographical grounds,
is slightly earlier than the Amaravati inscription of Vasithiputa sami
Siri-Pulumavi, mentions a Mahatalavara and a king. 27 As the
the Gangas (EC, Vol. VII, Sh. Nos. 4 and 64) and the Colas (El, Vol. XVIII ;
p. 26 and Kalihgattupparani) trace their descent from Iksvaku, they did
not assume Iksvaku as their dynastic name.
25. ARE, 1923-24.
26. ASSI, Vol. I, pp. 110-11. Pis. LXII and LXIII.
27. The epigraph does not give us either the name of the king or that
of the Mahatalavara. In the Calcutta Review for July 1925 Dr. Shamasastry
edited this inscription. According to him, 11. 16-17 refer to Sana king
of the Ayis; 11. 16-17 of the inscription however read :
" kahapanana (m) ca pura(ne)ka sahasam akhayani(vi) esa Maha-
talavarasa deya-dhama paric&ko ata utarapase bapana-nivatandni
eta sa-bhdriyasa sa-putakasa sanatukasa ayirana(rh) Puvaseliyana
nigdyasa "
The third letter in sanatukasa read as sa is clearly tu; a mention
of the grandsons of the Mahatalavara (and not of the name of a king),
after the mention of the wife and sons, is what is to be expected.
In the ARE, 1923-24, it was stated that "palaeographically it (the
inscription) may be assigned to about the second century A.D. Most of the
characters resemble those of the inscription of Siri-Yafia Satakani,
while others are like those of Satakani I and Usavadata."
The Chinna Ganjam inscription of the time of Siri-Yana written
in the ornate alphabet of the Nagarjunikon^a epigraphs, is admittedly later
than the Alluru inscriptions. A comparison with the Amaravati inscrip-
tions would have served the purpose far better ; the early square
characters of the Alluru type are to be found in some of the Amaravati
inscriptions which, belonging as they do to the various periods between the
second century B.C. and third century A.D., (the Nagarjunikonda alphabet is
to be found in Nos. 36 and 42, ASSI, Vol. I, pp. 91 and 104, Pis. LVHI, LIX ;
and El, Vol. XV, Nos. 27 and 54), enable us to trace clearly the evolution of
the Brahmi alphabet in the Krsna valley. The Alluru characters resemble
clearly those of No. 16 (ASSI, Vol. I, page 63 and PI. XLIV, ta, ya, sa,
ja, ma, ha, a, ka, da and lo). True, the Alluru na, ta, a, da and ka some-
what resemble those in the inscription of the time of Pulumavi II, but
the i and u signs in the former inscription as in Nos. 16 and 18, and the
rounded form of pa are certainly earlier than those of PulumaVi's time.
It might be argued that No. 16 is on a coping stone and must, therefore,
belong to the reign of Vasithiputa sami Siri-Pu|umavi when the railing was
THE IKSVAKtJS THIRl> CENTtfRY A.D. i3i
title and office of Mdh&talavara combined with those of Mdhasenft-
pati and Mahddan$an&yaka was a feature of the Iksvaku period,
and as the Iksvakus like the Maharathis were matrimonially con-
nected with the Satavahanas, they were, like the Maharathis, feuda-
tories under them. The feudatory title often met with in the
Andhradesa is Mahatalavara. We would be justified in concluding <
that the ancestors of the Ik?vakus were Mahdtalavaras under the
Satavahanas. After their fall Siri^Camtamula 28 founded his dynas-
ty much in the same way as the Mahdrafhi Cufcus in the south-
western parts and the Abhiras in the western parts.
The founder of the line Siri-Cdmtamula ' The Unobstructed ^ 9
Whilst Vasithiputa Siri-Camtamula is extolled by his sisters,
his father is not even mentioned in their inscriptions. 30 Camta-
enlarged and new stupa slabs set up. No. 52 is on a coping stone of the
outer railing, but in early square characters. An inscription of
the reign of Sivamaka Sada (karri) which is palaeographically later than
that of Vasithiputa sami Siri-Pulumavi, is also on an outer rail coping stone.
It would, therefore, seem that alterations in or additions to the railings
of the Stupa (Mahacetiya) were made from time to time.
28. Camtamula in A2, A3, A4, Cl, C2, C3, D4, and X; Cdla is clear in
E, G, and H.
Camta is clear in C4, G2, G3, L, M, etc. Dr. J. Ph. Vogel thinks
that probably Camta is the correct form. The dictum of Senart
that when the anusvara is found in some cases and absent in others, it is
certain that the scribe or engraver omitted it by mistake makes it certain
that Camtamula is the correct form.
Prakrt Camtamula has been Sanskritized by D. C. Sircar and K.
P. Jayaswal as Santamula and by Dr. Vogel as Ksdntamula. Cula which
in the Nagarjunikonda inscriptions is opposed to Mahd is certainly Sanskrit
Ksudra small. On this analogy Camta would be Ksanta only. In the
inscription under reference &i becomes sa and not ca. (See Childers Pali-
Eng. Die., cullo) .
29. Apatihatasamkapa.
30. On the other hand he is mentioned by his daughter Atfavi-Camtisiri ,
Among Iksvakus there is a preference for names beginning with
camta. A sister of Siri-Camtamula is called Camtisiri. His grandson is
called Ehuvu}a Siri-Camtamula.
Cula-Carhtisiri of the Kulahaka family would seem to have been
a descendant of an Iksvaku princess married into the Kulahaka family ;
as the office of the Mahatalavara would seem to have been hereditary in
the Kulahaka family as in the Dhanaka and Pukiya families, as Cula Camti-
sirinika is herself married to a Mahatalavara and as the Iksvakus are matri-
monially connected with the Mahdtalavaras such a conjecture has strong
support.
1S2 EARLY HISTORY OF THE ANDHRA COUNTRY
mula would therefore seem to have been the founder of the line. 31
That he possessed sovereign powers is indicated by the title ' Mafea-
raja' 32 attached to his name, and fittingly enough his sisters,
mother and consorts* erected a pillar and perhaps a stupa also in his
honour. 33 But we have no lithic record of his reign. All that we
know about him is furnished by the memorial pillar and by the
inscriptions of the reigns of his son and grandson.
His reign
In a passage which occurs in most of the inscriptions, Siri-
Camtamula is credited with the performance of Agnistoma, Agni-
hotra, Asvamedha, and Vdjapeya sacrifices. Whilst Jyotir-Agni-
stoma is the simplest of Soma liturgies and Agnihotra a modest
Havir-Yajna, Vdjapeya was a complex rite at the end of which the
performer sat upon the throne and was hailed ' Samrat ' * emperor'.
The fact that only three south Indian princes of the early period
are said to have performed it (Siri-Satakani I, Pallava Siva-Skanda-
varman and Siri-Caihtamula) shows how powerful Siri-Camtamula
must have been. In the Satapatha Brdhmana?* and Katyayana's
Srauta-sutra? 5 it is said that by offering Rdjasuya one becomes a
31. The Memorial Pillar Inscriptions (El, Vol. XXI, L.) shows
that Siri-Camtamula's father indulged in a plurality of wives (mdtdhi)
unless indeed it be that the term is an honorific plural employed
by the daughters of his only wife. But even a plurality of wives does
not indicate the father's kingly position, for even nobles and high dignitaries
of state would have followed the example set up by kings, as
their metronymics would show.
32. Caihtamula is referred to as Rdjan in the Memorial Pillar Inscrip-
tion. K. P. Jayaswal's contention that the title of Mahdrdja applied to
Siri-Camtamula indicates his feudatory position whilst the title of Rdjan
applied to Siri-VTrapurisadata shows that the royal position was assumed by
the latter cannot therefore be upheld. In the Iksvaku records the titles
Rfijem and Mahdrdja are indifferently used. Siri-Virapurisadata bears the
title of Rdjan in most of the inscriptions and Maharaja in inscriptions G and
H. (El, Vol. XX). Ehuvula Siri-Caihtamula is styled Mahdraja in G
and Rdjan in G2 and G3.
33. Though the inscription records the setting up of the pillar only,
the dome with the railing, having cetiya-arches over the gates, in the first
panel might be a representation of a stupa erected in his honour. However
Stupa No. 9 near which the pillar lay buried, contained only the bones
of an ox, deer and hare along with a broken doll's head made of red pottery.
ASR, 1929-30, p. 149.
34. V, 1.1.13.
35. XV, 1.1. 2.
THE IKSVAKUS THIRt) CENTURY A.D. 138
Rajan, and by offering the Vdjapeya a Samrdt. According to them
tfce office of * Rajan ' is the lower and that of ' Samrdt ' the higher.
The passage in the inscription also credits Siri-Camtamula with gifts
of lumps of gold, ploughs of land 36 and cows and oxen. Dr. J. Ph.
Vogel looks upon the passage as conventional. 37 But in the
fifth panel of the memorial pillar Siri-Camtamula is represented as
standing in plain attire, bareheaded, wearing sandals and hold-
ing a staff in his hands. But for the parasol over his head he
would look an ordinary person. By his side is an attendant hold-
ing a vessel containing libation water. Before him are five Brah-
mans; 38 one of them who is very young is stretching his right hand
to receive a gift from the king. On the ground is seen a heap of
round pieces of uncoined metal. 39 Here is undoubtedly a repre-
sentation of the gift of crores of ' hiramna ' (hiramnapindas or un-
coined gold pieces of a definite weight) with which Siri-Camtamula
is credited. 40
His personality, sisters and queens
Of the representation of the Satavahanas or kings of their
bhrtya lines on stone, we have only two clear instances. One is
that of king Simuka, his son Siri-Satakani I and the latter's family,
the other is that of Siri-Carhtamiila on the memorial pillar. In all
the four panels he is represented as a corpulent person. In the
second and third panels he wears a low cap. The fourth panel
shows him riding the state elephant fully caparisoned, with the
attendant seated behind him holding a parasol over his head. He
is followed by five or six marching attendants, one of whom is a
dwarf.
36. 'go-satasahasa-hala-satasahasa paddyisa.' Regarding the various
kinds of measure of land called 'plough' see Kulluka on Manu VII, 119.
37. El, Vol. XX, p. 6.
38. Dr. Vogel calls them monks. (El, Vol. XXI, p. 63). According to
Mr. Hirananda Sastri they are royal ladies and officials; the young person
is prince Virapurisadata . ASR, 1929-30, pp. 165, 166.
39. Unlike coins they are thick globules.
40. D. C. Sircar would see in the compound ' aneka-hirana-kofi-go-
satasahasa' a reference to some of the mdhddfLnas. Probably the gift of
cows or oxen and gold pieces was made on the occasion of sacrifices. It
is noteworthy that in the last panel Siri-Camtamula with his hair cut and
carrying a staff is like a performer of Agnistoma. What appears to be
thrown over his shoulders is perhaps the skin of an antelope. Barnett,
Antiquities o/ India, p. 162.
H.A. 18
134 EARLY HISTORY OF THE ANDHRA COUNTRY
Unlike his father and son who indulged in many wives, Siri-
Camtamula had only two queens. In the panels he is represented
as sitting with two queens. In the inscription where his ' maka-
devis' are mentioned we have ' subhatarikdhi ca Sarasikdya, Kusu-
mdlatdya.' ' Subhataraikdhi ' is a mistake for ' Sublwtarikdhi * (Sans.
Svabhattdrikdbhih) ' by his own ladies ' or wives. Thus in the ins-
cription also only two queens are mentioned.
Harhmasirinika and Camtisiri are the sisters of Siri-Camta-
mula. The former would seem to have died before the twentieth
year of Siri-Virapurisadata and the latter between his eighteenth
and twentieth regnal years. Unlike Siri-Caihtamula, a staunch fol-
lower of the Brahmanical religion, the two sisters were ardent
Buddhists (lay disciples) , and it is to the latter's munificence that
we owe some of the most important monuments in the Nagarjuni-
kon^a plateau. 41
Conclusion
Since the portrait representation of Camtamula shows us a
middle-aged person, he would seem to have died at middle age ; this
is made very probable by the fact that his mother and step-mothers
(mdtdhi) lived up to the twentieth year of his son's reign. No
Buddhist monument in the valley can be definitely attributed to
his reign. He was, like some of the Western Calukya kings, a
protege of Mahasena, * the Virupakhapati ' ( ( lord of Virupaksa
hosts') .^ Until fresh evidence turns up, his attitude towards Bud-
dhism and the Buddhist activities of the royal ladies must remain
unknown. Since daughter, sister, grand-daughter and daughter-
in-law are all anxious to state their relationship to him, Siri-Camta-
mula was evidently looked upon as the most famous in the line. 43
41. Camtisiri who was born of a Vasithi, was married to a Mahasenapati,
Mahatalavara Vasithiputa Kharhdasiri of the Pukiya clan. Such an endo-
gamous marriage seems peculiar. As the Mah&talavaras borrowed metro-
nymics from their kings it is no wonder they did not look upon the former
as an institution for regulating marriages.
42. El, vol. XX, p. 6.
"From the expression Virupakhapati-Mahasena-parigahitasa, which
is applied to Chariitamula, it may perhaps be concluded that he was a votary
of the god Mahasena or Skanda, 'the lord of the Virupakhas'. The
term Virupakha (= Skt. Virupaksha) seems to be used here to indicate the
hosts of which Skanda is the lord and leader."
"The word Virupakha (Skt. Virupaksha) , indicating a class of
snakes, occurs in an ancient snake-charm. Vinaya Ptyakam . . . ." ibid.
43. B2, C3, E, C2, H and G. El, vol. XX.
THE IK9VAKUSTHIRD CENTURY A.D. 135
VSrapurisadata 44
Ma^hariputa Siri-Virapurisadata was not as great as his father
Siri-Carhtamula. Even so his reign marks a glorious epoch in the
history of Buddhism in the Krsna valley. Probably one or two
matrimonial alliances of far-reaching importance were contracted
during his reign.
Alliance with the powerful house of Castana
Like their masters the Satavahanas, the Iksvakus also contract -,
ed matrimonial alliance with the Saka dynasty of Ujjain. An dyaka
pillar epigraph dated in the sixth year of Siri-Virapurisadata
records the donation of a pillar and 170 dinari-masakas 45 by *Maha-
devi Rudradhara-Bhattarika, an ' Ujenika Mahardjabalikd.' Now
that we know the names of the two queens of Siri-Camtamula it is
44. Buhler, (IA, vol. XI, p. 257) and following him Burgess (ASSI,
Vol. I, p. Ill) take both 'siri 5 and Vlra' as honorific prefixes. Buhler's
argument is that a name like 'Virapurisadata' would compel us to assume
the existence of a deity called ' Virapurusa ' which hitherto is not known,
(datta given by, Purusa=Vi$rm) . Names like Virarhnika and Viramna
(inscription F.) make Vlra' part of a personal name. When two or more
honorific prefixes adorn a name, they always precede ' siri ' ; i.e., what imme-
diately follows 'siri' is the personal name, e.g., Siva Siri-Apilaka, Siva Siri-
Satakani. Liiders, op. tit,, No. 1127.
45, D. C. Sircar would attribute the presence of dinari-m&sakas to 6aka
matrimonial alliance. Dr. Vogel would attribute it to the vast seaborne
trade between the Krsna-Godavari region and the West through
the emporiums of Kantakossyla, Palura, Koddura, etc. Dr. J. Ph. Vogel's
theory is supported by the following facts : The find of Roman coins of
the period from 68-217 A.D. (JRAS, 1904, pp. 599 ff.), at Vinukontfa in
the Guntur district and in the Nellore and Cuddapah districts; the mention
of a 'vaddlabhikaro (ra) yondka divikayo' (a Greek lamp resembling
the 'vaddla' fish in shape) in the Alluru inscription (A lamp of the fish
shape has been found at P'ong Tuk. It might have gone there direct from
Europe or Asia-Minor or the eastern coast of South India, ABIA, 1927,
PI. 8) ; and the Graeco-Roman influences discernible in the AmaravatI
sculptures of the middle of the second century A.D. As Gotamiputa Siri-
Yana Satakani and Vasithlputa Siri-Satakani struck silver in imitation of
Ksatrapa coinage with the Head of the King (the Ksatrapas derived the
obverse of their coins, i.e., Head of King from the denarii brought into
India by way of commerce, Rapson, op. cit., cviii and cix), it is not impro-
bable that the denarius was introduced in the wake of Satavahana conquest
of the eastern Deccan. The relic casket from Stupa No. 6, has yielded two
coin-like medrllions of thin gold % inch in diameter, each embossed with
a head which makes the impression of being meant for a portrait (ASR, 1929-
30, PI. 37),
138 EARLY HISTORY OF THE ANDHRA COUNTRY
not possible to consider her as his queen; she would therefore
appear to have been a queen of Virapurisadata. 48 As she is said
to have donated money for the building of the Mahdcetiya while
the work was going on, the matrimonial alliance must have been
contracted before the sixth year of Siri-Virapurisadata's reign, 47
possibly even during the reign of his father. As Ozene is men-
tioned by Ptolemy as the capital of Tiastanes 48 (Castana) and as
' Rudra ' often enters into the personal names of Western Ksatrapas
of the Castana line who style themselves Rajan* 3 Dr. Vogel's
conjecture that Rudradhara-Bhattarika belonged to the house of
Castana is very sound. 50 As Siri-Virapurisadata's reign would fall
in the second and third decades of the third century A.D. she might
have been a daughter of any one of the following Western Ksatra-
pas: Rudrasena I (S. 122-44), Rudrasiihha I (son of Rudradaman),
Prthivisena (son of Rudrasena I), Sanghadaman and Damasena
(. 144-58) , sons of Rudrasiihha I. The alliance would have gained
Ksatrapa recognition for the new dynasty. 51
Other queens
Other queens of Siri-Virapurisadata were Chathisirl and Bapi-
sirinika, daughters of Harhmasiri, and Bhattideva, 52 the daughter of
Carhtisirinika. 52a The marriage between the king and Bhattideva
would have taken place between the sixth and fifteenth
46. Unlike the other queens of Virapurisadata, Rudradharabhattarika
does not state her relationship to him.
47. The year in which the Mahacetiya was consecrated.
48. McCrindle, Ptolemy, Book VII, Chap. I, sec. 63.
49. Rudrasena III is styled Mahar&ja on some coins of Mahaksatrapa
Simhasena. Rapson, op. cit., p. 190.
50. Dr. Vogel reads 'Ujanika Maharabalikd' but the e sign over ja is
partially visible (the Prakrt form of Ujjain is Ujeni or Ujjem) . The learned
doctor's correction of 'Mah&rdbalika,' into 'Mahar&jabdlika' is certainly
warranted by the sense of the passage and the numerous mistakes of the
scribe or the engraver to be found in the Nagarjunikonda inscriptions.
51. This alliance accounts for the sculpture of a Saka warrior on one of
the Nagarjunikonda pillars (ABIA, 1927, PI. VI), and for the donations
by a Saka girl (El, vol. XX, p. 37) .
52. a is clear in G3. El, vol. XXI, PI.
52a. Bhattideva does not however expressly call herself a daughter of
C&mtisirinika . It can only be inferred from the facts that both of them
were Vasithis. and that Camtisiri calls Siri-Virapurisadata, 'her own son-in-
law' (apano j&matuka} in some inscriptions (E, Ml, etc.).
THE IK$VAKUS THIRD CENTURY A,D. 137
years 58 of his reign. As Bhattideva's son ascended the throne not
long after the twentieth year of Siri-Virapurisadata's reign, the
marriage must have taken place soon after the sixth year. 54
Alliance with ike Cutus
Another power of importance in south-western India of the
third century A.D. was the Cutu whose kingdom extended as far
north as Kanheri and as far east as^Anantapur. The political sense
of the Ikvakus that dictated an alliance with the Western Ksatra-
pas also dictated an alliance with the Cutus. An inscription dated
in the eleventh year of Ehuvula Siri-Camtamula records the bene-
factions of Mahddevi Kodabalisiri, 55 daughter of Siri-VTrapurisa-
data, half sister (bhacjirii, not sodara hliacfirii) of the king and wife
of a Vanavasaka-Maharaja? Q Scholars are agreed that Banavasi
53. Dr. J. Ph. Vogel is of opinion that the marriage took place between
the 6th and 18th year. But inscriptions Ml to M5 (El, Vol. XXI) dated in
the 15th year mention Siri-Virapurisadata as the son-in-law of Carhtisiri; the
latter calls herself his aunt (pituca) in an inscription of the sixth year of his
reign.
54. The significance of the Iksvaku system of marriages has been dis-
cussed along with metronymics (vide supra) .
55. Ins. H, El, Vol. XX.
56. Both forms Vanavasa or Vanavasi occur in the Nagarjunikonda
inscriptions (H and F). Vanavasakas or Vanavasins are mentioned in the
Mahdbhdrata (6,366) as a people dwelling in South India. The Mah&vamsa
mentions the conversion of Vanavasi by Rakkhita (Chap. XII, p. 84, evidently
the kingdom) .
The ka suffix indicates the place to which the king belonged, i.e., his
capital or his kingdom e.g. Kantakasolaka^:' inhabitant of Kantakasola'
(ASSI, Vol. I, p. 106) ; and Nasikaka^ 'inhabitant of Nasik,' (El, Vol. VIE ;
Nos, 20 and 22). The compound 'Ujanika-Mahara(ja)balika' makes it more
probable that the capital is referred to here. We may also note the
forms 'Vaingeyaka Hastivarma' and 'Kanceyaka Visnugopa' which occur in
the Allahabad pillar inscription of Samudragupta (CII, Vol. Ill, p. 3 t. 1. 4).
Dr. Vogel places modern Banavasi, which represents the remains of the old
town, in the Shimoga Dt. of the Mysore State (El, Vol. XX, p. 8) . It is,
however, in the Sirsi taluq of the North Kanara Dt. (Bombay Presidency)
(Survey map 48 F|14; Imperial Gazeteer of India, New Series, Vol. VI).
Dr. Buhler has quoted St. Petersburg Dictionary to the effect that Vaijayanti
occurs both in Brahmanical and Jain books as the name of a town in the
coast of the Konkan, and has suggested that it is the seaport Byzantion of the
Greeks. (CTI, p. 28, n.). The identity of Vaijayanti with Banavasi is how-
ever established by the following points: Jayanti as the name of Banavasi
occurs in manv records notably in an inscription at Banavasi itself at the
temple of Madhukesvara which records that the stone cot of Madhukebuara
188 EARLY HISTORY OF THE ANDHRA COUNTRY
Vanavasi, or Vanavasa is another name for Vaijayanti (Pr&kyt,
Vejayanti) , 57 The Malavalli pillar inscription of a Cutu Satakani
mentions Vaijayanti as his capital. Since Kanheri, which, as late
as the reign of Siri-Yana, was in the Satavahana empire, came into
the possession of the Cutus, 58 and since neither the Banavasi nor
the Malavalli inscriptions 59 can be ascribed to a period later than
the third century A.D., it is certain that the Cutus rose to power in
the third century on the ruins of Satavahana power. The Cutus
bear the title of 'Maharaja.' 60
The Buddhist monuments of his reign
The central royal Buddhist figure m the reign is Carhtisiri. To
this donatrix, the ' mahaddnapatini' , the ' veldmika ddna patibhaga
vochirhna dhdra padayini ' 61 goes the credit of having given an impe-
tus to the beautification of the Nagarjunikonda valley, 62 which bids
fair to prove of more interest than Amaravati. She would even
seem to have been responsible for the Buddhist leanings of Rudra-
dhara-Bhattarika, Chathisiri Cula-Camlisirinika and Bapisirinika. 63
The most important foundation of Camtisiri was the ' Mahdcetiya 3
enshrining the 'dhdtu' of the Great Teacher, 64 consecrated in the
was presented at the town of Jayanti (IA, Vol. IV, p. 207, No. 8), and this
god who was the family deity of the Kadambas of Hanagal is always called
in their records Madhukesvara of Jayanti.
Senart remarks (El, Vol. VII, p. 49) : " In addition to the instances
quoted by Dr. Burgess, where Jayanti seoms to represent Banavasi, one
might perhaps ask if in the Banavasi inscription the letter which has been
read sa or sam before jayamtakasa (1.2) might not be a ve, in which case
the sculptor Damoraka would be designated as a native of Vaijayanti."
The letter is clearly sa and nothing else.
57. EC, Vol. VII, p. 251, No. 263 t.1.1.
58. Luders' List, No. 1021.
59. IA, Vol. XIV, PI. EC, Vol. VII, pi.
60. IA, Vol. XIV, p. 331; ASWI, Vol. V, p. 86.'
61. C3, El, Vol. XX, p. 16, t. 11. 8 and 9.
62. The reason for her choice of this site must have been its vicinity to
the capital.
63. They want to attain Nirvana. B4, B5, C2, and C4.
64. Whilst Dr. Vogel thinks that Camtisiri built the Mahdcetiya, Mr.
Hirananda Sastri maintains (ASR, 1928-29) that she only rebuilt or enlarged
an older stupa. The latter view rests upon the following arguments, some of
them not expressly stated: The Amaravati, Ghantasala and Jaggayyapeja
Stupas, which on epigraphical evidence belong to a period much earlier than
the second century A.D., were enlarged, and dyaka platforms were added
to them during the second century A.D. If the Stupa was built by Carhtisiri,
the inscriptions would have told us how the relics of the Teacher which the
THE IK$VAKUS~THIRD CENTUBY A.D. 139
sixth year of Siri-Virapurisadata. The Stupa which is fifty feet in
diameter, is in the shape of a wheel, with spokes, hub, tyre and all
complete. 65 Camtisiri was aided in the undertaking by other Bud-
Stupa is said to enshrine were obtained; whilst the terms ' patithapita' and
'thdpitd' are used, with reference to the erection of 'dyaka-khartibhas,' 'cetiya-
gharas' and 'sela-mandavas,' 'samuthdpiya' and 'nfyhdpitd' are used with
reference to the Mahdcetiya (B5 and Cl) . Dr. Vogel cites the authority of
the Mahdvamsa (El, Vol. XX, p. 30) to show that 'nithdpita' means completed.
In inscription Cl, the Malidcetiya is called ' navakammam ' ; the ' navakam-
mika' is said to have been the Reverend Ananda, who knew the Majjhima
and Digha Nikdyas by heart. According to the Vinaya Pitaka (SHE, XX,
pp . 189 ff) a ' navakammam ' is ' a religious edifice ' erected by a lay member
(upasikd or updsaka) for the Samgha.
Whilst the fact that the outer drum and the interior of the Mahdce I iya
are built of bricks of the same size negatives the theory of enlargement, the
fact that it is built of bricks of the same size as those used for the apsidal
temples built during the Iksvaku period (20"xlO"x3") and other Buddhist
monuments in the valley, and the fact that the relic caskets in the Nagarjuni-
kon^a stupas are all nearly alike, prove Dr. Vogel's theory. If the Mahdcetiya
is older than the Iksvaku period, we would have found older epigraphs and
sculptures of which no traces remain; no doubt if the pot containing the silver
relic casket found in one of the northern chambers of the Stupa had been in-
tact, it might have given us an inscription and proved beyond doubt the age
of the Stupa. (The relic consists of a fragment of bone of the size of a
pea found inside a tiny round gold box %" in diameter. This with a few
gold flowers, pearls and garnets was placed in the silver casket shaped like
a stupa. The latter was, however, found corroded and broken to pieces).
Vogel translates "namo Bhagavato samma sambudhasa dhdtu-
vara parigahitasa Mahdcetiye" into "adoration to the Blessed one the supreme
Buddha absorbed by the best of elements at the Mahacetiya "
Dr. Hirananda Sastri links dhdtuvaraparigahita with Mahdcetiye and
thinks that the Mahdcetiya was protected by the corporeal remains of the
Buddha. I am wholly unable to accept this as we cannot link the genitive
parigahitasa with the locative Mahdcetiye (El, Vol. XX, p. 29 note 1) . The
interpretation proposed by Vogel on the authority of M. L. de la Vall6
Poussin for the difficult phrase dhdtuvaraparigahita is quite acceptable.
Mahdcetiye commences a fresh sentence and must be linked up with ayaka-
khabho thdpitd.
While stupas of less importance at Nagarjunikon^a (Stupa No. 6 and
Stupa No. 9 especially) were decorated with carved marble slabs and coping
stones, the Mahdcetiya would seem to have been executed in simple style
like the stupas of Ceylon. As will be shown below, there was active commu-
nication between Ceylon and the Nagarjunikontfa valley.
The remains of the Mahdcetiya are the drum fifty feet in diameter,
and twenty feet high at the centre, the dyaka platforms, dyaka-khambhas,
some of them in fragments, the foundations of the enclosure wall and the
gateways.
65. According to Mr. Longhurst, all the Andhradesa stupas are built in
140 EARLY HISTORY OF THE ANDH&A COUNTRY
dhist and non-Buddhist royal ladies and private individuals. 66
Mahadevi Rudradhara-Bhat^arika donated money and an ayaka
pillar. 67 Adavi-Camtisiri, Chathisiri, Bapisirinika and Cula-Camti-
sirinika of the Kulahaka family, each contributed an 'dyafca-JcTiam-
bha.' A slab and a coping stone were donated by a Chadakapavatica
and his wife Padumavan! together with their sons and daughters.
However, nine dyaka-pillars or more were set up by Camtisiri.
The practice of erecting vihdras and cetiyagharas or apsidal tem-
ples 68 by the side of stupas was followed in the Nagarjunikon^a
valley also. Close to the Mahacetiya on the eastern side is an apsi-
dal temple; 69 an inscription incised on the marble floor of the shrine
in two lines records its foundation by Camtisiri in the eighteenth
year of Virapurisa data's reign 'for the sake of his victory and
longevity of life/ 70 To the east of the apsidal temple are fragments
of thirty-six pillars, some of them just rising out of the ground,
pillars which must have supported the roof of a mandapa. Frag-
this style (IA, Vol. 61, p. 188). But the Jaggayyapeta Stupa was formed
of earth in layers about two feet thick over each of which was laid a close
flooring of very large bricks closely fitted together (ASSI, Vol. I, p. 108).
66. El, Vol. XX, p. 25, Inscriptions I and J.
67. The meaning of e ayaka' is not settled. Some would connect it with
'ayako', 'ayaka' (d^yaka) meaning Venerable or worshipful'. In our epi-
graphs we have 'ayaka' not 'ayaka'; 'dyaka' is used as a noun; and the fact
that ayaka pillars bear inscriptions (El, Vol. XX, H; Vol. XXI, G2, G3)
little favours the theory that they were objects of worship. Ayaka is
rendered by Luders and Burgess as 'entrance'. Dr. Vogel objects to this
rendering on the ground that in the Nagarjunikon<Ja and Amaravati inscrip-
tions, the word for entrance or gate is 'dara' (F. El, Vol. XX; ASSI, Vol. I,
No. 44). An inscription on a coping-stone (ASSI, Vol. I, Pi. XLVIH; fig. 4)
of a Cetiya platform has ' utarayake pato dana' (ibid, p. 93) . Another inscrip-
tion (No. 47, p. 86) on an ayaka-pillar (PI. XLV, No. 1), records the gift
of a 'Cetiya- khab ha' (ayaka-pillar) at the ' dakhindyaka' (dakhindyake not
dakhindyaka as Burgess has read it) . But hi No. 15 PI. Ivi we have 'utar&yake
unisa ddnam.' Since the inscription is on an outer rail coping, 'ayaka' here
means 'gateway.' Pali 'ayo 1 means 'entrance'. Thus in the Amaravati inscrip-
tion 'dyaka is used to denote 'something at the entrance/ as well as projections
facing entrances. It would, therefore, seem that the projections received
that name from the fact that they faced entrances or 'gateways.'
68. 'Cetiya-ghara' . It seems that 'ghara' was restricted to halls used
for worship, e.g., 'sela-ghara' Karla No. 1, (El, Vol. VII); Cetiya-ghara
Kutfa Nos. 15 and 23 (CTI), Nasik Nos. 18 and 19 (El, Vol. VIH).
69. Plate VI, Nos. 2 and 3.
70. apano jam(d)tukasa rano M (d) thariputasa-lkh (a) kunam 5iri-Vird-
purisadatasa dyu-vadhanike vejayike, (El, Vol. XX, E.).
THE IK$VAKUS THIRD CENTURY A.D. l4l
ments of inscriptions on those pillars put together, record the
foundation of a pillared-hall surrounded by a cloister (cdtusdla-
parigahitam) 11 and its consecration in the eighth fortnight of the
rainy season of the fifteenth year of the reign.
A private donatrix who emulated Camtisiri was the lay disci-
ple Bodhisiri, daughter of the householder Revata and Budhaihnika,
belonging to Govagama 72 and niece of the treasurer (Kothagdrika)
Bhada (Sanskrit Bhadra) . The foundations attributed to her are
two 'Cetiya-gharas' (one on the Lesser Dhammagiri by the side of
a vihara 13 as the special property of the theris (nuns) of Ceylon, 74
and another at Kulaha-vihara, 15 a shrine for the Bodhi-tree (i.e.,
71. El, Vol. XXI, p. 65. But no traces of the quadrangular building
remain.
72. As Bodhisiri dedicates her apsidal- temple to the theris of Ceylon,
she was probably a native of Ceylon. N. Dutt identifies Govagama with
Gonagamaka, mentioned as a port in Ceylon in the Mahdvamsa (IRQ. Vol.
VII, p. 653, n. 2).
73. The vihara referred to still stands on the Naharaljaboijlu mound.
74. Dr. Vogel's translation of ' (bha)damta (rd) jdcariydnam Kasmira-
Gamdhdra - Cina-Cildta-Tosali-Avaramta - Vamga- Vanavdsi-Yavana-Da (mila)
(Pa) lura-Tambapamni-dipa pas (a) dakdnam theriydnam Tariibapa ( m) nakd-
nam-suparigahe Cetiyaghararh kdritam' (F, t. 1. 1) as
"Caitya-shrine erected for the acceptance of the fraternities of Ceylon who
have converted Kasmira etc./' is not satisfactory. Not even the Mahdvamsa
does credit Ceylonese monks with the conversion of various countries. Keeping
in mind the analogy of the Karla and Nasik inscription (Karia Nos. 19 and
20, Nasik Nos. 2, 3 and 4; also El Vol. I, p. 240), we might translate thus;
" Caitya shrine erected for the venerable teachers who converted Kasmira
etc., for the special acceptance of (as the special pro-
perty) of the theris of Ceylon." This agrees with the account of the Ceylonese
chronicles (the Dipavamsa and Mahdvamsa) that Ceylon, Kasmira, Gamdhara,
Mahisamandala, Vanavasa, Aparantaka Maharat{ha Himalaya and Suvanna-
bhumi, were converted by monks sent from India by Moggaliputta Tissa.
Ceylon is said to have been converted by Asoka's son Mahinda (Mahdvamsa
Chap . XIII, pp . 88 ff ) . The inscription "Moggaliputtasa" on relic caskets
from Andher and Safici stupas makes Moggaliputta Tissa a historical per-
sonality (Luders' Nos. 664, and 682).
In a learned paper in the Indian Historical Quarterly (Vol. VII, pp.
651 ff) N. Dutt has objected to Vogel's translation of pasddakdnam as 'who
converted.' According to Dutt as ' posada ' in the Mahdvamsa means * serene
joy,' pasddakdnam ' of those who brought serene joy.' But the account of the
conversion of countries given in the Mahdvamsa makes Vogel's translation
acceptable (Childers, pasdda).
75. Vide supra.
H.A.- 19
142 EARLY HISTORY OF THE ANDHRA COUNTRY
a railing around it) at the Sihala-vikdra, one cell at the Great
Dhariimagiri, 77 a mawjiava pillar at the Mahavihara, 18 a hall for
religious practice at Devagiri, a tank, verandah and manijiava at
Puvasela, 79 a stone raarw^aua at the eastern gate of the Mahdcetiya
at Kan^akasela, 80 three cells at Hirurirthuva, seven cells at Papila 81
a stone manfava at Puphagiri, and a stone mandava at the
yihdra.
Ehuvula Cdmtamula
The son of Matfhariputa Siri-Virapurisadata by Vasithi Bhatti-
deva was Ehuvula Camtamula, the last known king of the dynasty. 82
T^at he ruled for at least eleven years is shown by the inscription
of his half-sister Kodabalisiri (H) .
76. On the analogy of Kulaha-vihdra, Sihala-vihara would mean the
vihdra built by the Ceylonese monks or nuns or laymen. Perhaps a branch
of the Bodhi-tree at Ceylon was planted at Nagarjunikon^a.
77. Could it be the mound to the north-west of, and not far away from
the Cula-Dhammagiri? On this mound are the remains of a stupa and sela
man&ava ; the latter consists of a series of four cells on each of the north-
east and west sides and a pillared hall 30' 6" x 30' in the south with an open
court situated between the pillared hall and the stupa. The roof of the hall
rested on sixteen pillars; all of them with the exception of two at the south-
east and south-west ends have fallen down. (Plate V Nos. 1 and 4 ASR
1926-27, pp. 158-59) .
78. Could this be the Mahd-vihdra (lit. Great monastery) to which the
' Mahdcetiya ' is said to have been attached (B5) .
79. According to Hiuen-Tsang it stood to the east of the capital of
Tona-kie-tse-kia (Dhanyaka^aka) . It gave its name to a Buddhist school.
80. The name occurs in an Amaravati inscription (ASSI, Vol. I, PI. Ixi;
No. 54) ; Burgess and Lliders have read it as Katakasola. But the anusvara is
clear on the plate. An inscription from Peddavegl mentions Kan^akosala
(MER, 1926-27 No. 219). It is certainly the port Kontakossyla which is men-
tioned by Ptolemy (Bk. VII, Chap. 1, Sec. 14) and placed by him just north
of the mouth of the Krsna and which transliterates Kanjakasula (the spear of
thorns) the Sanskrit form of Kantakasola. The name is preserved in the
modern Ghantasala, a village thirteen miles to the west of Masulipatam and
the sea. There is a Cetiya in the village (Rea, South Ind. Bud. Antiquities,
pp 4 ff ) , and on the southern boundary of the village there is a mound named
Polimeradibba on which loose bricks appear. Surely these are traces of Bud-
dhist buildings (Ibid., p. 42). But no trace of the sela-mantfcwa at the
eastern gate of the Mahdcetiya is available.
81. Papila is perhaps identical with the Papikala of the Alluru inscrip-
tion.
82. H. t. 11. 3 and 10, Ehuvala ; G2 t. 1. 8, Ehuvula ; G3 Ehuvula. Only
in G3 is e turned sideways. K. P. Jayaswal agrees with Hirananda Sastri
THE KVAKUS-THIRD CENTURY A,D. 143
Buddhist monuments of his reign
The second year of his reign witnessed the completion of a
monastery called ' Devi-vihara, 93 provided with everything, 84 by
queen (Devi or MahudevT) Bhattideva for the grace and acceptance
of the masters of the Bahusutiya sect. 85 This monastery, with a
roofles man<Japa, stands at the north-east foot of Nagarjunikon^a
on the Itikarallaboiju where stand also a stupa (No. 5) and two
apsidal temples. 86 Inscriptions G-2, G-3, are borne by the ayaka-
khambhas belonging to Stupa No. 5. Perhaps Bhattideva built the
stupa and apsidal temples, in addition to the Devi-vihara, thereby
emulating her mother. In the eleventh year Kodabalisiri conse-
crated to the masters of the Mahisasaka school a monastery and
a cetiya. 1 The cetiya is probably Stupa No. 6 on the top of the hill
above the monastery. The latter, which is on the Kottampalagu
mound a few furlongs to the north of Nagarjunikon<Ja, and com-
mands a fine view of the Krsna, has a pillared hall or pavilion in
the centre with a row of twenty cells all round. The hall is sixty-
one feet square and provided with a flat wooden roof supported
by thirty-six lofty marble pillars. Stupa No. 6 has yielded a num-
ber of sculptured beams, 88 two small medallions, 89 a silver relic
in reading 'Bahuvala'. Says he: "In the plate G, the letter b is misformed,
but the full form is seen in H, where it occurs twice and is clearly the four-
cornered b" (JBORS, 1933, p. 173, n. 1) . The so-called ba has no resemblance
whatsoever with the four-cornered ba occurring in these inscriptions. It
certainly resembles the Jaggayyapeta c (ASSI, Vol. I, PL LII; t. 1. 5).
Like Pulumavi it is a Dravidian word and both are difficult of inter-
pretation. 'Ehu' of the name may be Tamil Ehu 'steel', a very old Tamil word.
83. This is another instance of a religious foundation named after its
founder. Here the monastery receives the latter part of the queen's name.
84. savajataniyuto.
85. G, G2, G3.
86. Structural apsidal temples of the very early centuries A.D., are
very rare in India. Barring those at Nagarjunikonfla, one has been dis-
covered at Sanci, two at Taxila, and one at Sarnath.
87. Dr. Vogel reads imam khaniyam vih&ro ca and explains khaniya by
a resort to 'khanu pillar' (Childers, Pali-Bug. Diet, khdnu). What is read
as ni is certainly ti, for unlike the loop in na the loop in ta ends in a down-
ward curve. The word is clearly cetiyam. At Nagarjunikonda cetiyas and
vib&ras are found together.
88. Pis. VI, 4; VII, 1, 2, 3 and 4, VJH, 1 and 2.
89. ASR, 1929-30, PL 37 d and c, one is embossed with the head of a
Greek male figure and the other with the head of an Indian lady. They
are not king and queen as they do not wear crowns.
144 EARLY HISTORY OF THE ANDHRA COUNTRY
casket much like the one found in the Mahdcetiya and bone
relics. 90
The monuments which would seem to belong to the Iksvaku
period, but which cannot be assigned to any reign definitely, are
two cetiyas (Nos. 3 and 4) and a vihara 91 on the hill to the north-
west of and near the Naharallabodu, the vihara on the Naharalla-
bo$u and Stupas Nos. 5, 7, 8 and 9.
Buddhism of the period
Like the Amaravati inscriptions the Nagarjunikonda inscrip-
tions introduce us to a number of Buddhist schools; the Mahisa-
sakas, an offshoot of the original Sthaviravddas; the Bahusutiyas,
a sub-division of the Gokulikas who belonged to the Mahdsamghi-
kas, the original schismatics; 92 the Puvaseliyas (offshoot of the
Mahdsamghikas) who find place in Buddhaghosa's commentary on
the Kathavatthu, along with Avaraseliyas, Siddhaihikas and Raja-
yiriyas, are called Andhakas, i.e. schools that took their rise in the
Andhradesa?* and the Avaramahdvinaseliyas, who were patronised
by Carhtisiri. According to Dr. Vogel ' Avarasela ' is perhaps the
abbreviated form of 'Aparamahavinasela' of our inscriptions. 94 Since
a 'Mahdvinasela? school is mentioned in an Amaravati inscription, 95
since we meet with 'Puvasela* and not 'Puvamahavinasela' in a
Nagarjunikonda (Inscription F.) and the Alluru inscriptions, Dr.
Vogel's suggestion cannot be accepted. A fragmentary inscription
from Amaravati has 'liyanam' and before it space for four or five
letters (Mahavinase?) , The teacher belonging to this school is
referred to as "Mahavinayamdhara" 9B (versed in the Great Vinaya) .
Can Mahdvinasela be a form of Mahdvinayasela, the elision of ya
being not uncommon ? Like Puvaseliya and Aparaseliya, Puva-
90. Dr. Vogel thinks that the eastern side of the stupa is the most
important one. But the relic caskets from the Mahacetiya and Stupa No. 6
were found in chambers on the north-eastern side.
91. Vide supra.
92. Walleser, Die Sekten des alien Buddhismus. pp. 6 and 21.
Hiuen-Tsang found Mahasanghikas in Tona-kie-tse-kia. Watters 'On
Yuan-Chwang' O.T.F. Series, Vol. II, p. 214.
93. Edited by Mrs. Rhys Davids, p. 104. Avarasasela is also mentioned
by Hieun-Tsang as a hill which stood near the capital Dhanyakataka
(Ibid.).
94. El, Vol. XXI, M2, 1. 5, M3, 1. 6 ; Vol. XX, E. 1. 5.
95. ASSI, Vol. I, p. 105, PI. LX, No. 49: Mahava (vi) nasela is men-
tioned as a place-name in PI. Iviii No. 35, (Ibid.).
96. JbW., p. 102.
THE IKSVAKUS- THIRD CENTURY A.D. 145
mahavinaseliya and Aparamahavinaseliya schools might also have
arisen. It is well to remember the remark of Rhys Davids. 97 :
" As the so called sects were tendencies of opinion, the number of
them was constantly changing." Dr. Vogel takes Ayira-Hamgha
(Cl, C2) to be the name of a school. In his paper, entitled 'Notes
on the Ndgdrjunikonda Inscriptions,' N. Dutt contends that ' Ayira-
Hamghdnam' is another form of 'Mahasdmghikanam'. 97 * But c Aira-
Utayipdbhdhinam ' in an Amaravati inscription, 98 'Aryamaha-
sdmghikdndm Lokottaravddindm Madhyadesikdndm pathena Vin-
ayapitakasya, Mahdvastuyeddi^ e Ayirdnam Puvaseliydnam ' in the
Alluru inscription, prove that Ayira (Arya) is not used even with
reference to sects in the sense of 'mafia? but only in the sense of
Venerable.' ' 'Ayira-Hamgha^ \ like 'Catudisa Sagha' of the Nasik and
Karla inscriptions, would, therefore, mean the venerable Samgha. 100
Camtisiri's wish that the families to which she belonged and the
whole world might attain happiness in both the worlds, reminds us of
the Mahayanist who places the attainment of Bodhi knowledge and
liberation from worldly miseries of all creatures, before his own. 101
According to Hiuen-Tsang monks studying the 'Great Vehicle' lived
in the Andhradesa. Nagarjunikonda sculptures also show tbe Mafia-
ydnist tendencies at work (worship of large Buddha
figures, one of which was found in the large square chamber at the
north-west end of the monastery on the mound to the north-west of
the Naharallabodu; three others were found at site No. 4) , 102 Where,
among all the Hinayana schools mentioned, does Mahayanism come
in ? The answer is given by Mrs. Rhys Davids : " The exten-
sion of the Mahayanist school was and is of a very vague and fluid
kind. Those to whom it applied formed no close corporation." 103
Mr. Rhys Davids compares the relation of the Mahayana to Hina-
yana schools with that of the various Roman and Greek Catholic
schools to those of the early Christians. 104
97. ERE, q. v. Hinayana.
97a IHQ, Vol. VII, pp. 647 ff.
98. Luders, No. 1276; ASSI, Vol. I, p. 87, No. 45, PL LX.
99. Mahdvastu, ed. Senart, p. 2.
100. M. 15, El, Vol. XXI, mentions (Mahabhi) khu Sam (gha).
101. This ideal finds expression in K&randavyuha where Avalokitesvara
Bodhisattva is represented as refusing to accept Nirvana, until all creatures
were in possession of the Bodhi knowledge and were freed from worldly
miseries (Sarras vami's ed. p. 121).
102. ASR, 1926-27.
103. Points of controversy, Preface, XLVI.
104. EKE, Hinayana.
146 EARLY HISTORY OF THE ANDHRA COUNTRY
Ceylon and Ndgdrjunifcon^a
Ceylon and Ceylonese Buddhism were in touch with Nagar-
junikonda. Aryadeva, a disciple of Nagarjuna (3rd century A.D.) ,
was a native of Ceylon, but spent the greater part of his life in
India. The pot containing his relics has been found in the Guntur
District. 105 Theris of Ceylon would seem to have lived in the
Nagarjunikonda valley, for the apsidal temple on the Naharallabo<Ju
is dedicated to them. A vihdra built by a Ceylonese Buddhist is
also mentioned (Si hala- vihdra, inscription F.) . Probably the trade
routes from the eastern ports and the Mahdcetiya enshrining the
'dhatu' of the Great Teacher attracted these pilgrims to eastern
Deccan.
Buddhist Canonical Books Mentioned
A point of interest is the mention of the Digha and Majhima
sections of the Sutta Pitaka and of the five Mdtukas. 1 ^ The Mdtukas
are the condensed contents especially of the philosophical parts of
the Canonical books in the Abhidhamma. On the authority of
Burnouf s translation of the Saddharma Pundarika, Childers says
that it means also the list of Vinaya precepts omitting all the ex-
planations and other details.
Administration : Administrative divisions
The biggest administrative division was the rdstra, a division
identical with the Satavahana dhdra. But the rdstra division was
known to the pre-Iksvaku period in the Andhradesa. The Alluru and
Amaravati inscriptions mention ratfoas. 107 The division below the
rdstra was gdma. 108
105. Vide infra.
106. ' Digha-Majhima-pamca-Mfitu'ka-osaka (desaka) v&cak&naih acarlyti.-
nam Ayira-Hamghdnam a(m)tevasikena DtgTia-Majhima-NiJcdva-dfiarena
bhaja(da)mt Anadena' (Cl, C2). Pali-English Diet. q. v. mdtikd.
107. ARE, 1923-24, p. 97 and PI. 11 4-5: Carathe Maca(pa)da; ASSI,
Vol. I, No. 17, 'Towpukirathe(?) adhiihane'.
According to Burgess adhithana may be the name of a town or may
mean capital. After adhithane we have a lacuna with traces of four letters
and after it vathavasa. The lacuna would then seem to have contained the
name of a town.
108. The villages mentioned in the Iksvaku records are Pamnagama at
which masters of the venerable Samgha are said to have resided, Govagama,
Na^atura in Kammakaratha, Mahakamdurura and Velagirl,
THE IK9VAKUS-THIRO CENTURY A.D. Uf
Officiate
The official titles known to us are those of Mahdsenapati, Mahd-
talavara, Mahddanfandyaka and Ko^jhogdrika. 109 A feature of the
Iksvaku period is the bearing of two or more titles by the same per-
son. 110 Vasitfilputa Karhdasiri, Vasithlputa Maha-Karhdasiri and
Vi^husiri of the Pukiya family 111 and Vasithlputa Khamdacali-
kiremmarjiaka of the Hiramfiaka family, bore the titles of Maha-
sendpati and Mahdtalavara', the son-in-law of Siri-Camtamula bears
the titles of Mahasenapati, Mahdtalavara, and Mahddan<jundyaka;
perhaps this indicates a higher position than that of the other
Mahdtalavaras, which he enjoyed in virtue of his being the son-in-
law of the king.
Official Titles: Mahasenapati
Vogel looks upon Mahasenapati as a nobiliary title on the score
that the Satavahana Mahasenapati was in charge of rastras. It is
highly improbable that high dignitaries are known in their inscrip-
tions by their nobiliary titles only. In the Jangli Gur><jlu inscription
a Mahasenapati in charge of an ahdra is mentioned along with a
Gdmika, an official. A Mahasenapati with non-military duties was
possible in an age of confusion of titles and duties. In all the re-
cords of the Iksvaku period the title Mahasenapati precedes that
of Mahdtalavara and except in one instance the wives of Mahatala-
vara-Mahdsendpati-Mahddandandyakas bear the title of Mahatala-
vari only. 112 Could this indicate that the title of Mahasenapati was
superior to that of Mahdtalavara ?
109. Since Ko$thd,gara means 'storehouse'; Prakrt Kolhagarika is best
construed as 'keeper of royal stores.' See also Luders' No. 937, Kothagala.
110. In the Cinna Ganjam inscription of the time Siri-Yana Satakani we
have Mahatarakasa Maha-e ' The latter is perhaps the mutilated form
of Mah&sendpatisa; in that case the combination of two or more titles in
the same person is not peculiar to the Ifesvaku period.
111. According to Vogel Maha -Kamdasiri of C5 is identical with Kamdasiri
the husband of Camtisirinika C3. He however considers the lady mentioned
in C5 as a co-wife of Kamdasiri, for while Camtisirinika mentions Khamda-
sagaramnaka as her son, the other lady mentions Mahdsendpoti-Mafidtala-
vara-Vinhusiri as her son, and both C3 and C5 were incised on the same
day. In our epigraphs 'Maha' and 'Cula' are used to distinguish
a younger from an elder person, that is when they both bear the same
name, e.g., Camtisirinika and Cula-Carhtisirinka (B4, C5); Damila-Kanha
and Cula-Kanha (Luders' No, 1243). On the analogy of these names Maha-
Kamdasiri is either an elder brother, or some senior member of the Puklya
family.
112. B2, El, Vol. XX.
148 EARLY HISTORY OF THE ANDHRA couN*R5?
Mahdtalavara
In his note on the Nagarjunikon^a inscriptions Vogel says 112a
" The curious term Mahdtalavara which is also met with in other
inscriptions of Southern India must likewise denote a high dignitary,
wh6se exact function, however, is not clear. The second mem-
ber of the compound is not a Sanskrit word, but seems to be
a term borrowed from some Dravidian language." As
' maha ' is a prefix denoting a higher title, ' talavara ' is the term
to be explained. Says Vogel: "We must leave this question to the
decision of students of South Indian Languages. Can the word have
any connection with Tamil talavay (~a general), Tamil talaiyarl
(=a village watchman) or Canarese talavara, talavara (a watch-
man, a beadle)?" Since Canarese talavara (a watchman) very
nearly corresponds to our Talavara, since Tamil talaiyari also means
watchman, talavara and talaiyari are the words with which Tala-
vara should be connected. 113 The title of Mahdsendpati borne by
Mahatalavaras may also exclude Tamil talavay (commander).
Talaiyari as Tamil form of Talavara or Talavara as a Prakrt varia-
tion of talaiyari is phonetically possible. 114
Since Talavaras are mentioned in the Kalpasutra with eighteen
ganardjas, Vogel looks upon Talavara as military title. But the
Subodhikd, a commentary on the Kaiposutra, shows that it was an
official title. 115 The Matoalavaras would seem to have been
viceroys. As has been suggested above, the office of Mahdtalavara
was a Satavahana one derived from the local office of Talavara
much in the same way as the offices of Mahdbhoja and Mahdrathi
were from those of Bhoja and Rathika. 116
112a. El, Vol. XX, p. 6.
113. Tamil talayarikkam (watchman's dues, SII, Vol. II, p. 119, n. 4)
is certainly the Canarese tdlavdrike (El, Vol. I, p. 402, n); talapdfaka as a
revenue term occurs in the Kalimpur plates of Dharmapala (IA, Vol. XI,
text, 1. 51). See Wilson, Glossary talapada. D. C. Sircar connects talavara
with Tamil talaivan (op. cit., p. 16) .
114. Pischel, op. cit., Sec. 254.
115. Talavarah tusta-bhupdla-pradatta-pafta bamdha vibhu$ita Rdjasthd-
nlydh. This passage has been quoted by Vogel himself.
116. On the evidence of the Alluru inscription we may say that during
the Satavahana period the title of Mahdtalavara is not found in combination
with those of Mahasendpati and Mahdtalavara. On this score even the
Ramaretfdipalle inscription which mentions a Mahdtalavara would also
belong to the Satavahana period.
IKVAKUS THIRD CENTURY AJX 149
" .... It penetrated also into Northern India, for there can be
little doubt that it is identical with the mysterious word taravara
which coupled with mahdpratlhdrd (=" a great chamberlain ") is
found in the legend of one of the clay sealings excavated by the
late Dr. Bloch at Basarh, the site of ancient Vaisali. This document
belongs to the Gupta period. It was suggested by Dr. Bloch that the
word tarika, which occurs in the lists of officials in mediaeval
copper-plate charters, may quite well be a corrupted form of tara-
vara." 11Ga These instances show that the office of Talavara survived
the Iksvaku period. The Kondamu^i plates mention a Mahatala-
vara-Mahddandandyaka. Besides the instances cited by Vogel
we have the Deo-Baranark inscription of Jivitagupta II of
Magadha, which mentions a Taldvdtaka-, 111 the Kudopali plates of
Mahabhavagupta, assigned by Keilhorn to the first half of the
twelfth century A.D., 118 mention a Talavargin ; and the Katak
plates of Mahasivagupta mention a Tala7ii(?)ta, 119 and in an ins-
cription of the Kalacuri king, Rayamurari Sovideva (A.D. 1173),
the son of the governor of Ehur is called Talavara Camdeya-
nayaka. 120
Mahddandandyaka
The title Mahddandandyaka is unknown to the Sata-
vahana period. As danda means ' rod ' as well as ' army ' (danda-
mti~ administration of justice), the title can be explained as judi-
cial or military. Since Mahasendpati was in origin a military title
Mahddandandyaka would be a judicial one. All these titles were
hereditary and sometimes more than one son inherited the father's
titles.
Other Conclusions
The few glimpses that the Iksvaku records and sculptures
afford into the social life of the period have been discussed in the
chapter entitled ' Social, Economic and Religious Conditions ', as it
represents a continuity with the conditions in the Satavahana period
and in many aspects presents no break with the past. According
to Vogel seaborne trade was "no doubt also largely responsible for
116a. El, Vol. XX, p. 7.
117. CH, Vol. El, No. 46.
118. El, Vol. IV, p. 258, 7i,
119. El, Vol. HI, p. 352.
120. El, Vol. XII, p. 335.
H.A. 20
150 EARLY HISTORY OF THE ANDHRA COUNTRY
the flourishing state of Buddhism in this part of India. The devotees
of the Good Law were largely recruited from the commercial classes
and it was their wealth which enabled not only the merchants
themselves, but also their royal masters, to raise monuments of such
magnificence as the great stupa of Amaravati." The remains of an
ancient quay discovered by Longhurst on the right bank of the
Krsna near the Nagarjunikoiida plateau, 121 and the emporiums of
Kantakossyla and Allosygne would seem to be evidence of a vast
seaborne trade with Ceylon, Farther India, and the West. This trade
was perhaps responsible for the gifts of crores of gold with which
Siri-Caihtamula is credited. But the Buddhist monuments of the
Iksvaku period were almost all of them constructed neither by
merchants nor by their royal masters. All of them were, except
the foundations attributed to Bodhisiri and a slab donated by Cada-
kapavatica, set up by royal ladies. It is then seen that Nagar-
junikonda cannot by itself prove that the flourishing Buddhism was
a result of flourishing trade. It must also be noted that Nagar-
junikonda does not introduce us to such a glorious epoch of Bud-
dhism as Amaravati, Giunmadidurru, Alluru and Kantakasela, for
we do not find at Nagarjunikon^a such a crowd of Buddhist de-
votees as at Amaravati. Evidently the rise of the Brahmanical
dynasties was silently undermining Buddhism from the beginning.
Of the Iksvaku currency we know nothing except that the
denarii were current. Not a single Iksvaku coin has been picked
up. As the Satavahanas made large issues of lead and copper coins
which are discovered in heaps in the Kr$na, Godavari and
Guntur districts even to-day, especially at Nagarjunikonda, it is
possible that the Iksvakus did not find it necessary to issue new
coins.
121. IA, 1932, p. 188.
CHAPTER VIII
KINGS OF THE BRHATPHALAYANA GOTRA
The material for a study of the dynasties that succeeded to the
political heritage of the Satavahanas in the Andhradea and in the
areas south of the Krsna as far as Palar is scanty. Of the kings
of the Brhatphalayana 1 gotra, we have but a single copper-plate
grant (the Kondamudi plates of Maharaja Jayavarman) .
Chronology
The chronology of the period is far from being satisfactorily
settled. While editing the grant just mentioned Dr. Hultzsch
says 2 : " The alphabet of his ( Jayavarman's) inscription shows
that he must have lived in the same period as the Pallava king
Siva-Skandavarman who issued the Mayidavolu plates." This view
has been followed generally by other writers on South Indian
History. Prof. Jouveau-Dubreuil not only follows Dr.
Hultzsch but even says that the unknown predecessor of
iva-Skandavarman Pallava ruled between 225-250 A.D. 3
A fuller knowledge of the Iksvaku dynasty than was
possible before the discovery of the Nagarjunikonda
inscriptions has led Prof. Dubreuil to change his views. He
now places the Iksvakus in the third century A.D. and shifts the
early Pallava s to the fourth; but even here he maintains that
Maharaja Jayavarman and Yuvamahdrdja Siva-Skandavarman were
1. Brhatphalayana as a gotra name is not to be found in other records.
But the phrase Brhatphalayana-sa-gotto occurring in the Kondamudi plates
of Jayavarman leaves no doubt on the point. In these plates as well as
in the records of the kings who are said to have belonged to the 6alankayana
and Ananda gotras we miss the dynastic names. This recalls to our mind
some of the Satavahana inscriptions and coins where we have metronymics
derived from Vedic gotra names, but miss the dynastic name. In the total
absence of their dynastic names scholars have labelled the former group
of kings as 'the Brhatphalayanas, the Salankayanas and the Anandas.' It
is like calling the Pallavas and the Kadambas as the Bharadvajas and the
Manavyas respectively. We would avoid the confusion between the dynastic
and gotra names if we call these kings Ttings of Brhatphalayana gotra,'
'kings of the Salankayana gotra' and so on.
2. El, Vol. VI, p. 316.
3. Ancient History of the Deccan, p. 54.
152 EARLY HISTORY OF THE ANDHRA COUNTRY
contemporaries. 4 Dr. K. R. Subramaniam'Has also followed Dr.
Hultzsch. He would make the Iksvakus, Jayavarman and the Pal-
lavas contemporary powers and give them a third century date. 5
A comparative study of the alphabet of the Kondamudi and
the Mayidavolu plates throws some doubt over the contemporaneity
postulated by Dr. Hultzsch. No doubt the alphabets of the two
grants have some common characteristics. Both exhibit a cursive
writing. Both have the peculiar e which according to Dr. Hultzsch
resembles the archaic Tamil sa* the ma with a loop at the bottom,
and the semicircle or triangle open at the top, replaced by a rudi-
mentary vertical to which is attached on the left a curved stroke.
Sometimes the curved stroke does not touch the vertical as in
'Yuvamahardja,' 'barhhadeyam* and 'vitarama' (Mayidavolu 11. 1, 12
and 13 respectively) ; and in 'Mahesvara', 'Jayavammo' and 'amTie'
(Kondamudi 11. 3, 5 and 7 respectively) . The sa consists of two
curves one below the other but not connected still. Besides these
common features mentioned by Dr. Hultzsch there are others like
the peculiar forms of Jen, ke, ha, la, and na. 7 These common peculi-
arities are evidently to be explained by the fact that the records
come from one and the same area, from adjoining taluqs in the
Guntur District. 8
4. JAHRS, Vol. V, p. 91.
'The Mayidavolu plates are written in the same alphabet as the plates
of Jayavarman' Ibid.
5. "At the time of his (6iva-Skandavarman's) rule, about the middle of
the third century A.D., the Brhatphalayanas ruled what was later known
as the kingdom of Vehgi (Veiigl) and the Iksvakus were in possession of
the Andhra country stretching from about 6ri 6ailam northward and extend-
ing indefinitely into Dakshina Kosala and along the coast north of the Goda-
vari." Buddhist Remains in Andhra and Andhra History, p. 78.
While writing this chapter I got D. C. Sircar's monograph entitled
The Successors oj the Satavdhanas in the Eastern Deccan. He has also
followed the old view.
6. El, Vol. VI, p. 86.
7. It is interesting to note that the cursive ha which agrees with the
northern Gupta form (Biihler, Tables IV, 39) occurs in the Jaggayyapeta
inscriptions of the time of the Iksvaku king Virapurisadata, but does not
occur in the Nagarjunikonda inscriptions of his reign. It occurs in a Karla
inscription EL Vol. XXIV, p. 282 and PI. XVI 1. 3). The peculiar ha on
some of the coins of Gotamiputa Siri-Yana Satakani was probably developed
out of this ha.
8. Mayidavolu is a village near Narasaraopet, the headquarters of the
taluq of that name, and Konclamu^i is a village in the Tenali taluq. It Is
only to be expected that alphabets vary not according to dynasties but
according to localities, and in establishing any comparative system of palaeo-
KINGS OF THE B$HATPHALAYANA GOTRA 153
By the side of these common characteristics stand out certain
differences in the Mayidavolu grant which seem to indicate a
further stage of development from the Korwjamudi alphabet. Says
Dr. Hultzsch : " The group jd (Kon^amu<Ji plates 11. 5, 11 and 34)
has a different shape, the vowel-mark being attached on the right,
and not at the top of the letter as in the Mayidavolu plates (11. 18
and 24). The n (or n) 9 is identical in shape with the lingual d
but the dental d is represented by a separate character while in
the Mayidavolu plates no distinction is made between all the four
letters/' 10 The other differences which are more important for
settling the relative chronology of the two dynasties and which
have not been pointed out by Dr. Hultzsch are as follows:
(a) Whilst the Mayidavolu sa consists of two equal curves,
the upper curve of the Kondamu<Ji sa is bigger than the lower one
and resembles the upper curve of sa in all other inscriptions. This
fact and the peculiarity common to the Mayidavolu sa and the
Kondamudi sa which has been noted above make it probable that
the latter represents a transition to the former. Whilst the two
curves of the Mayidavolu sa, which are more developed than those
of the Kondamudi sa, end in strong hooks on the left, the curves of
the latter have no such finish.
(b) The broad-backed na of the two grants differs from the
na of the Hira-Hadagalli and the Gunapadeya grants, which has
graphy we must select territorial rather than dynastic names. The highly
cursive writing of the Hira-Hadagalli grant shows in its ductus a certain
relationship to the Jangli Gundu (in the Bellary District) inscription of Siri-
Pulumavi, the last of the Satavahanas. The similarity between Gotamiputa
Siri-Satakani's and Usavadata's Nasik and Karla inscriptions is explained
by the area of their location. (JRA.S, 1926, p. 625). The differences bet-
ween the Mayidavolu and the Hira-Hadagalli grants of giva-Skandavarman
Pallava must be explained by the same principle. The peculiar ma, so,, ha,
la, na, and e are not to be found in the Hira-Hadagalli grant. The Udayagiri
cave inscription of Candragupta II of the year 82 (G. Era) is the box-
headed variety of the Central Indian alphabet. The Sanci inscription of
the same king is in the character of the southern alphabets (CII, Vol. Ill,
p. 4). Instances of this kind can be multiplied.
9. na does not occur in the Kondamudi plates.
10. It may be argued that this greater absence of differentiation in the
Mayidavolu plates may indicate an earlier period. But differentiation does
not always represent a later stage of development. The looped ta and na
are later form? of the na with the horizontal or curved base and the ta with
the semicircle at the bottom, and yet there is less difference (sometimes no
difference) between the former than between the latter.
154 EARLY HISTORY OF THE ANDHRA COUNTRY
a strongly curved base line, the beginnings of which we see in the
Girnar prasasti of Rudradaman 11 and in some Kusana inscriptions. 12
The Mayidavolu na is more broad-backed than that of the Konda-
mudi plates and this fact certainly points to a later period.
(c) Whilst the e of both the grants has a form not met with
elsewhere the Mayidavolu e is more cursive and ends in stronger
hooks than the Kon^amudi e.
(d) Whilst the verticals of both the Kondamu^i and the
Mayidavolu la are bent to the left, they have longer tails and smal-
ler bodies than those of some of the Amaravati inscriptions of the
first and second centuries A.D. 13 The body is smaller in the Mayi-
davolu than in the Kondamu<Ji grant. This development can be
traced further. In the western script of the sixth and seventh
centuries the body becomes smaller and the tail correspondingly
longer. True in the la of the grant of Carudevi 14 (later than the
Mayidavolu grant) the body is more prominent than in the Mayi-
davolu la, but even so the former registers a development over the
latter in the enormous tail which is a feature of the la of the
Eastern Calukya and later Pallava charters.
Again the Kondamudi la has not the angular or slightly curved
base of the Mayidavolu la. 15 However the la in ' alonakhadakam ' 16
resembles the Mayidavolu la (angular base) but even here the
upper vertical is not bent to the left as in Mayidavolu and the
medial o sign over it is an earlier form of that found in the Mayida-
volu plates. 17 The vertical of the Kondamudi la does not continue
the curve of the body but starts from the middle of its right arm
so that even careful epigraphists like Dr. Hultzsch cannot distin-
guish between la and gi. In line 42 Dr. Hultzsch reads 'tagivarena'
for 'talavarena,' and Dr. Vogel 18 thinks that the former is a mistake
11. Btihler, Tables III (vi).
12. Ibid., V.
13. ASSI, I. Nos. 8, 16, 32 and 44.
14. The Carudevi grant was mistakenly called a grant of Nandivarman
of the 6alankayana gotra by Dr. Fleet in IA, Vol. V, p. 176. However, he
corrected himself in Vol. IX. Since writing these lines I find that D. C. Sircar
has also noticed the .mistake, vide infra.
15. la with the curved base occurs in 11. 13, 15 and 22 while la with
the angular base is found in 11. 2 and 6.
16. Kontfamudi line 32.
17. Vide infra.
18. El, Vol. XX, p. 7, n,
KINGS OF THE BRHATPHALAYANA GOTRA 155
of the scribe for the latter. This error arises from the rudimentary
form of the Kondamudi la.
(e) The curved horizontal member of the Mayidavolu ka
like that of the Carudevi grant is surely later than the straight line
of the Kondamudi ka.
(f) The ba of the Kondamudi grant with the notch in the
left vertical which does not, except in a few cases, show a serif at
the upper end (11. 4, 10, 37 and 38) and is slightly open on the left
at the top is less cursive than the closed ba of the Mayidavolu and
the Carudevi grants. The Kondamudi pa in its narrow and curved
base is an obviously earlier form.
(g) The letter ya shows practically the same features of
development from the Kondamudi to Mayidavolu as the letter pa.
(h) The vertical member of the Mayidavolu a like that of
the Carudevi grant is longer than that of the Kondamudi a. The
lower end of the vertical of the former shows sometimes a bend
and sometimes a reascent to the left. The reascent is more pro-
nounced in the Carudevi 19 than in the Mayidavolu grant; in the later
Pallava charters there is a reascent to about half the length of the
vertical. 20 The curves at the base of the verticals of the Kongia-
muoli a are rudimentary, as those in Usavadata's inscriptions from
Karla. The upper and lower limbs of the two Pallava grants under
reference are connected to the middle of the vertical by a straight
line; in the Kon^amu^i grant they are represented by a wavy line
connected to the top of the vertical by a slanting stroke. While
the vertical and left upper limb have nail heads which in the
Carudevi grant are turned into small curves attached to the top oi
the vertical (11. 8 and 10), the Kontjamu^i one has no such nail-
head or curve. The general appearance of the Kon^amu^i a is
altogether more primitive.
(i) The medial a, i and o signs of the Mayidavolu plates show
a more developed form than those of the Kon<Jamu4i plates. While
in the Kon(Jamu4i plates the sign expressing the length of the
vowel in d is a short stroke attached to the middle of the vertical
(11. 6 and 9), in the Mayidavolu plates it is a curve (line 4.) which
becomes stronger in the Carudevi plates. (11. 6, 8, 10 and 16) . Even
where the strokes are attached to the top of the letter those in the
19. 11. 6, 8 and 9.
20, Biihler, Tables VII, i, xx and xxii.
156 EARLY HISTORY OF THE ANDHRA COUNTED
Mayidavolu plates end in better curves and have longer tails (as
in the CarudevI grant) than those of the Kon^amufo plates. In
the latter the medial i sign is a semi-circle; in the other two it
approaches the closed circle. The medial o sign in the Mayidavolu
lo (line 13) is a wavy line while in Kon^amudi (line 32) it is a hori-
zontal stroke. Even where it is a horizontal line over the letter,
the tail on the right is longer in the Mayidavolu than in the Kon^a-
mu<Ji grant. In the CarudevI grant it is longer still.
Scholars are agreed that the CarudevI grant is later than the
Mayidavolu grant. The line of development is therefore from the
Kon<Jamudi to the Mayidavolu grant, and then on to the CarudevI
grant. 21 Even where the Mayidavolu alphabet does not approach
the CarudevI alphabet, it shows more developed forms than those
of the Kontfamu^i plates.
The Mayidavolu plates were issued by Yuvamahardja Siva-
Skandavarman during the reign of his father (Bappa) . The palaeo-
graphy of the plates makes it therefore highly probable that Jaya-
varman reigned at least a generation before the predecessor of Siva-
Skandavarman.
The general facts of history point to the same conclusion. It
may be presumed that the find place of the Kondamu^i grant
(Tenali taluq) is not far from the object of the grant. 22 Then
Jayavarman's sway would have extended over lands south of the
Krsna. The alphabetical peculiarities common to the Mayidavolu 23
21. The orthography of the grants furnishes corroborative evidence.
Whilst the writer of the Kon^amutfi plates follows the practice of the Sata-
vahana inscriptions where every double consonant is expressed by a single
letter (the exceptions are ' bammhadeyam ' and * Jayavammo ') , the writer
of the Mayidavolu plates adopts the etymological spelling in 'pattifca',
' datta ', ' sa-gotto ' and ' kardpejja '. In the HIra-Ha$agalli grant the etymo-
logical spelling of the Pandit is more pronounced than in the other two. The
orthography of the CarudevI grant is in accordance with that of literary
Prakrt. Dr. Hultzsch himself remarks (El, Vol. VIII, p. 144. n. 5) that in
this respect the two grants of iva-Skandavarman occupy an intermediate
position between the Satavahana inscriptions and the CarudevI grant.
22. supra, p. 152.
23. Prof. Jouveau-Dubreuil sees in the Kudura and the Kudurahara,
according to him roughly coresponding to the modern Bandar taluq of the
Krsna District, the capital and the kingdom respectively of MaMr&ja Jaya-
varman. Hie Prof, has narrowed down too much the kingdom of Jayavar-
man. Scholars like Dr. Hultzsch and Kielhorn (El, Vol. VI, p. 316 ; Vol. IV,
p. 34; Vol. V, p. 123) are agreed that the Kudurahara of the Konflamufli
plates is the same as the Kudrahara viaya of some of the Salankayana
itlNGS OP THE B^HATPHALAVANA GOTRA 157
and the Kon^amu^i grants are corroborative evidence in the same
direction. 24
Whilst the Mayidavolu grant of Yuvamahdraja Siva-Skanda-
varman proves that the Guntur District or part of it was included
in the Pallava division of Arhdhapata, 25 the Carudevi grant shows
that the Guntur region continued to be part of the Pallava Domi-
nions during the reigns of Dharmmamahdrajadhiraja Siva-Skanda-
varman and his immediate successors Skandavarman and Yuvama-
hdraja Buddhavarman. 26
inscriptions and the Gudrahara, Gudravara and Gudrara visaya of the Eastern
Calukya grants. In a Kakatiya inscription on the right door-pillar of the
Bhlmesvara temple at Gudivaola (593 of 1893), GuoUvada is said to have be-
longed to the district of Gudrara. Dr. Hultzsch has identified Kudura with
Gu^ivatfa, the headquarters of the taluq of the same name in
the Krsna District. A grant of Maharaja Nandivarman 6alankayana
from the Kollair Lake in the Kaikalur taluk of the Krsna
District makes it probable that the latter was also included in
the Kudurahara or Kudrahara visaya of the grant. Thus the Kudurahara of
Jayavarman's time included besides the Bandar taluq the territory as far west
as Gudivada, as far north as the Kollair Lake and as far south as the northern
part of the Guntur District. Besides, in the Satavahana and Salarikayana
records ahara (or hdra) and visaya denote a division of the kingdom, not the
whole kingdom. Kudura mentioned in the Kondamudi grant is spoken of as
the headquarters of the governor of the district, and as a ' vijayakhamd&vdra '
i.e., royal headquarters in camp. According to Hemacandra (Desiko&a q.v.) it
may also signify a capital. In Nasik No, 4 (inscription of Gotamiputa Siri-
Satakani) which the inscription under reference resembles in phraseology, a
' vijayakhamdavara ' in Govadhanahara is mentioned. The capital of Gotami-
puta Siri-Satakani was Paithan, far away from Govadhana or Govadhanahara.
24. Pamtura, the village mentioned in the grant as ' bamhadeya ,' can
be identified with Potturu in the Guntur taluq. Prof. Jouveau-Dubreuil
identifies it with Panduru in the Bandar taluq of the Krsna District (Ancient
History of the Deccan, p. 85).
25. ' Viripara' the village situated in the AmdMpatQi) a and the object of
6iva-Skandavarman's grant is certainly the Virparu mentioned in the
Kopparam plates of Pulakesin II (El, Vol. XVIII, p. 258) and perhaps the
modern Vipparla in the Narasaraopet taluq of the Guntur District.
26. Siva siri-Apilaka, siva Siri-Satakam, siva Skanda Satakarni; the
prince called Kharhda-naga in a Kanheri inscription (ASWI, Vol. V, p.
called siva-Khamda-naga-siri in a Banavasi inscription (IA, Vol.
These instances of the use of 'siva' hi the inscriptions and on
the second century B.C. and second century A.D., make it highly
'siva' in the expression Siva-Skandavarman is an honorific
therefore 6iva-Skandavarman and Skandavarman are iden
'siva' enters into the composition of names, e.g., Sivamaka.
HJL 21
158 EARLY HISTORY OF THE ANDHRA COUNTRY
It is thus clear that the reign of Jayavarman in the same region
must be placed before that of Yuvamaharaja Siva-Skandavarman.
The rise of the dynasty
The rise of the dynasty of Maharaja Jayavarman is shrouded
in mystery. Even so, the Iksvaku records from Nagarjunikonda
and Jaggayyapeta make a tentative suggestion possible. The cursive
writing of the Kontfamupli grant obviously places it after the Iksvaku
inscriptions. It has been shown above that the Iksvakus ruled not
only north and south of the Krsna, but as far east as the delta of
the Krsna; then their dominions must have included at least a part
of what was later on the kingdom of Jayavarman. The continu-
ance of the Iksvaku offices of Mahdtalavara and Mahadandandyaka
under Jayavarman is another link in the chain of evidence that
suggests that Jayavarman or his predecessors had a large share in
weakening the power of the Iksvakus. 27
As in the case of the Pallavas we are in the dark as to the
founder of the dynasty. The Kondamudi plates do not mention
the father of Jayavarman even in the Pallava or Salankayana
fashion, i.e., under the form Bappa, 28 But could Jayavarman
have carved out a kingdom for himself out of the debris of the
Iksvaku kingdom, built up an administrative machinery, earned
the title of Maharaja 22 and entered upon a career of further con-
quests 30 within the short span of ten years? 31 Obviously Jaya-
varman's dynasty rose to power before Jayavarman came on the
scene .
The Capital of Jayavarman
It has been shown that Prof. Jouveau-Dubreuil's and
Dr. Hultzsch's view that Kudura was the capital is untenable.
27. Sircar boldly attempts to carry Jayavarman's dynasty to the second
century B.C. He says: "If we accept the reading Pithwfa in a
passage in the Hathigumpha inscription (1. 11) of Kharavela and the interpre-
tation that King Kharavela of Kalinga besieged the city of Pithuo!a, it is not
impossible to think that the Brhatphalayanas were ruling at Pithu<Ja=Pitundra
as early as the time of Kharavela (second or first century B.C.)", op. cit. p. 38.
28. Like Pallava kings Jayavarman assumes Brahmanical gotra, has a
name ending in ' varman ' and does not bear a metronymic.
29. He is called a MaJidrdja on the seal and a Raja in the plates.
30. The source of this assertion is the term 'Vijayakhamddv&ra'
(Kontfamutfi 1. 1).
31. The Kontfamutji plates were issued in the tenth year of his reign.
KINGS OF THE BFtHATPHALAYANA GOTRA 159
D. C. Sircar would locate it in Pityndra, 32 mentioned by Ptolemy
as the metropolis of the Maisolia region. 33 But Ptolemy places it
in the interior of the Maisolia region and there is no evidence to
show that Jayavarman's dominions extended beyond the modern
Gutjlivada taluq in the west. Ptolemy wrote in the middle of the
second century A.D., and the Iksvakus of the third century A.D.
had their capital in Vijayapuri. Nothing compels us to look upon
Pityndra as the established capital of every dynasty that ruled
over the Andhradesa. Under the Pallavas who would seem to
have succeeded Jayavarman in the Guntur region Dhaihfia-
kaa(ka) is the headquarters of the Andhra province and the
town is as old as Pityndra if not older. 34 Dhamnakataka has
equally good claims to be considered as Jayavarman's capital. 35
Administrative Organisation
The kingdom was mapped out into districts called dhdras as
under the Satavahanas, each under an executive officer called
Vapatam. The Sanskrit word corresponding to Vapatam is
' Vydprta. ' Prta is the past participle of pr ; with the prefix vya
it means * busied with or engaged.' Viyapata in the sense of
* engaged ' occurs in the Edicts of Asoka. 36 The Kdsikd, a com-
mentary (probably seventh century A.D.) on Panini by Vamana
and Jayaditya, equates Vydprta with the Ayukta of Panini (II, 3,
40) . The latter term (Pali ayutto) means ' superintendent or
agent.' It occurs in the Carudevi grant and with the ka affix in
later inscriptions. 37 In the Damodarpur inscription of Budha-
gupta (fifth century A.D.), it is said that Kotfvarsa visaya was
administered by the Ayuktaka Sandaka. Vydprta and Ayukta
were therefore officers in charge of districts much the same as the
amacas of the Satavahana records and the Visayapatis of later in-
scriptions. The office of Vapatam is met with in the Kondamudi,
the Mayidavolu and the Carudevi grants only. 38
32. The Pithumela of the Hathigumpha inscription of king Kharavela.
33. Op. cit., p. 38.
34. It is mentioned in two Amaravati inscriptions of the second century
B.C. (El, Vol. XV, "Some Unpublished Amaravati Inscriptions).
35. McCrindle (IA, XIII, p. 370) would go to the length of identi-
fying Pityndra with Dhamnakataka. But the fact that the former is placed
by Ptolemy north of the Maisolos (the Krsna) is against such an identifica-
tion.
36. Shah. V, Kal. V, Man. V, etc.
37. El, Vol. XI, p. 175, t. 1. 17; XII, p. 154; t. 1. 60.
38. Dr. Hultzsch corrects 'viya' of the Carudevi grant to 'viya' and
160 EARLY HISTORY OF THE ANDHRA COUNTRY
Another, perhaps higher, dignitary in the kingdom bore the
titles of Mahatalavara and Mahddawjiandyaka. These titles stand
out in the Iksvaku period as prominently as those of Mahdbhoja
and Mahdrathi in the Satavahana period and Jayavarman's
dynasty is obviously indebted to the Iksvakus for these titles.
It may be puzzling that a Mahatalavara Mahddandandyaka? 9
probably a feudatory like the Iksvaku Mahatalavara and Mahd-
dandandyaka and the Satavahana Mahdrathi is entrusted with the
task of preparing the plates, a task ordinarily entrusted to minor
officers in the Satavahana stone records, where also engraving does
not mean the preparation of the stone, but that of the copper-plates
or palm-leaves. 40 Like the Mahdsendpati of Nasik No. 3 and the
Rahasddhigata of the Hira-Hadagalli plates who are said to have been
entrusted with the drafting of the charter and who, as has been
shown above, 41 would have only supervised the drafting, the
Mahatalavara under mention was perhaps in charge of the depart-
ment for the preparation and custody of the charters In the
same grant the king is said to have drawn up the protocol (sayam
chato) which can only mean that a lekhaka drafted it under the
immediate supervision of the king, for the king is also said to have
issued the order by word of mouth (aviyena dnatam) f*
restores ' viya (patam) .' The occurrence of 'viyapata* and 'vapata* in Asokan
edicts (vapata: Shah. V: and viyaputa: Man V makes 'viyapatam' another
Prakrt form of 'vapatam'. In the edicts 'viyapata' occurs more often than
'vapata' or 'vaputa',
39. Dr. Hultzsch who edited the Kondamudi plates long before the Nagar-
junikonola, Alluru, and Ramareddipalle inscriptions were discovered read
' Mahatagivarena ' and con jectur ally translated it as ' the best of the Mahatagi
family.' The peculiar form of la in ' Brhatphalayana' (1. 4) i.e., the vertical
starting from the right arm of the curve leaves no doubt that the letter read
i>s gi is to be read as la. In his edition of the Nagarjunikon^a inscriptions
Dr. Vogel and following him other writers have merely remarked that ' Maha-
taqivarena ' is a mistake of the scribe or engraver for ' Mahdtalav arena '.
40. The Kondamudi plates 11 36 and 37 ; Nasik Nos. 3, 4 and 5. El,
Vol. vm.
41. Vide supra, pp. 83-4.
42. Following Senart who derives the word from 'ksan* (to cut)
Dr. Hultzsch translates chato by 'signed'. He says (El, Vol. VI, p. 319):
44 The king's signature may have been affixed to the original document, which
was deposited in the royal secretariat, and from which the copper-plates
were copied." This leads us to the paradoxical conclusion that the operation
i.e., the drafting of the charter which is not wanting in the cognate inscrip-
tions (especially the later Satavahana inscriptions which the inscription
KINGS OF THE BRHATPHALAYANA GOTRA 161
Whilst grants of villages to religious bodies or Brahmans is
a feature of every reign, the similarities of Jayavarman's grant to
the later Satavahana grants, which show the extent to which
Satavahana administrative traditions were carried on to the period
of their successors in the Andhradesa, invest it with special inte-
rest. The immunities expressly attached to the bamhadeya in
Jayavarman's grant are the same as those conferred on the
bhikhuhala in the Satavahana charter. 43 The operations or form-
alities connected with the grants are in both cases verbal order,
drafting, preservation of the record in the archives of the state,
engraving, and delivery. The parihdras mentioned in and the
operations connected with the Mayidavolu grant are different.
More striking is the similarity in the wording. 44 Dr. Hultzsch
remarks 45 " the language and phraseology of the inscrip-
tion (Kon(Jamu<Ji) are so similar to the Nasik inscriptions of
Gautamiputra Satakarni (Nos. 4 and 5) and Vasishthiputra
Pulurnayi (No. 3) that Jayavarman's date cannot have been very
distant from that of those two Andhra kings." This statement
which implies Jayavarman's indebtedness to the Satavahanas for
his political lessons was made long before the Nagarjunikonda
remains were brought to light. It has already been shown that
some of the political institutions of Jayavarman's dynasty were
inherited from the Iksvakus. Since the Iksvakus took up the
thread where the Satavahanas left it, it is not improbable that Jaya-
under reference closely resembles in phraseology) is the only stage of which
there is no trace in this inscription.
43. Karla and Nasik inscriptions of Gotamiputa Siri-satakani and Vasithl-
puta sami Siri-Pulumavi .
44. The Kondamudi grant opens in the same manner as Nasik No. 4 (El,
Vol . VIII) . The parihdras are expressed by the terms apdpesam, anomasam,
alonakhddakam, arathasamvinayikam and savajdtapdrihdrikam. The
instructions to officials are also couched in the same language :
compare the Kondamudi grant " etamsi tarn gdma Pdtura bammmhadeyam
kdtuna oyapdpehi ", " etasa casa g&masa Pdmturasa bammhadeyam kdtuna
parihdre vitar&ma " and " etehi nam parihdrehi parihardhi etam casim
gdma[m] Pdfuratfi ba (m) mhadeyaih katuna etha nibhamdhdpehi" with
Karla No. 19 " etesa (tu) gdma Karajake bhikhuhala deya (ot/a) pdpehi"
" etasa casa gdmasa Karajakdna bhikhuhala-parihara vitardma " and " etehi
na parihdrehi pariharah et. casa gdma Karajake bhikhuhala-parihdre ca
etha nibadhdpehi ". The formalities connected with the grants are expressed
in the same terms. This is all the more interesting since the formulae of
immunities were variable.
45. Op. cit.
162 EARLY HISTORY OF THE ANDHRA COUNTRY
varman is indebted immediately to the Iksvakus. No copper-
plate grant of the Iksvakus and the later Satavahanas has been
found in the Andhradesa (the western cave inscriptions of Gotami-
puta Satakani and Vasithiputa Pulumavi are merely copies
of inscriptions engraved on pattika, i.e., copper plates or palm
leaves or cloth) . It is, however, highly improbable that the former
who outshone the latter in their zeal for the Brahmanical religion,
and who were as tolerant of the Buddhist religion as the later
Satavahanas, did not make grants of lands to Brahmans
and Buddhists ; in such a case the phraseology of the Satavahana
grants would have been kept up by them and transmitted to
their political successors. 46
46. The disappearance of the charters or paftifcd, of which the cave ins-
criptions of Gotamiputa Siri-Satakani and Vasithiputa Pulumavi's time are
copies is to be attributed to the fact that they were written not on copper-
plates but on perishable materials cloth and palm-leaves,
CHAPTER IX
THE VAINGEYAKAS
Sdlankdyana A Gotra and Not A Dynastic Name
While editing the Ellore plates of Devavarman 1 Dr. Hultzsch
remarked that Devavarman and his successors might be desig-
nated the Salahkayana Maharajas of Venglpura. He has been
followed by other scholars. Says K. V. Lakshmana Rao: "The
earliest of the dynasties of kings that Epigraphy has disclosed to
us as having ruled at Vengi in the Krishna district 2 is that of the
Salahkayanas ". 3 Says D. C. Sircar: "It is therefore not quite
impossible that the Bull banner of the Salankayana kings was
connected with the name of their family." 4
6alankayana is a gotra and not a dynastic name . The Pallava
Vakataka and Kadamba charters which give the kula and gotra
names make the distinction between them clear. The Vaingeyaka
grants also make a distinction between kula and gotra names. 5 In
all the records ' Salankayana ' occurs in the singular (Salankaya-
nasya, Salankayanah) . In the Satavahana, Pallava, 6 and Vaka-
taka records the dynastic name is always in the plural, while the
gotra name is in the singular ; 7 and the alphabet and phraseology
of the early Pallava Sanskrit charters bear a striking resemblance
to those of the charters of the kings of Salankayana gotra. In the
1. El, Vol. IX, p. 58.
2. Now West Godavari District.
3. JAHRS, Vol. V, Pt. i, p. 21.
4. Jl. of the Dept. of Letters, Calcutta, Vol. XXVI, p. 70: In some Sanskrit
Lexicons Nandi is called Salankayana vide infra.
5. ' asmat kula gotra dharma ya&obhi-vrddhyartham''
6. The Mayidavolu, Hira-Haolagalli, Carudevi, Pikira, Mahgajur Uruvu-
palli, Oihgodu and Cendalur grants.
7. In the Kadamba charters the dynastic and gotra names of kings are
in the plural.
Sometimes 'vam&a', 'jdti' and 'kula' names occur in the singular e.g.
'ksahar&tasa Nahapdnasa' (Nasik No. 10 El, Vol. VIIL, 'Khakharata' is referred
to as a ' vam&a ' name in Nasik No. 2) ; Abhirasya Isvarasenasya, ' (Abhlra is a
tribal name) ; ' Ikhdkusa ' (Nagar junikonola Inscriptions) . But our records do
not resemble these in any respect. Sometimes 'Ikhaku(ku)lasa' or
'Ikhakunam' occurs.
164 EARLY HISTORY OF THE ANDHRA COUNTRY
records of Yuvamaharaja Visnugopavarman, Maharaja Simhavar-
man and Maharaja Kumaravisnu, the family name immediately
precedes the personal name whilst the gotra name is separated
from the latter by laudatory epithets. In the Ellore grant of
Devavarman Salankayana, Salankayana is separated from his
name by the epithet assamedhaydjino. It may be argued that the
'sa-gotra' which is added to the golra names of the donors in the
Vaingeyaka grants themselves is not added to Salankayana. 8 But
the remarkable similarity in the phraseology of these grants and
the fact that sometimes gotra names occur without the suffix
sa-gotra knocks the bottom out of this argument. 9 It would seem
that in spite of the reference to their kula, the kings of the Salah-
kayana golra, like Jayavarman of the Brhatphalayana gotra bore
no dynastic name ; they were probably upstarts without any
renowned ancestors, real or eponymic. 10
While editing the Kollair plates Dr. Fleet remarked that the
Salarikayanas were descendants of Visvamitra and of lunar extrac-
tion ; he added " Perhaps these are the ' Solankis ' of Col. Tod, who
are included in the catalogue of the thirty-six royal races, and who
for a long time ruled over 'Anhilvaolpattana' in Gujarat." Else-
where, 11 he speaks of Salankayana as a gotra name. However, he did
not refer to the Pravarakdndas, nor did he emphasise the difference
between Salankayana, son of Visvamitra and Salankayana. 12 Four
gotra rsis bear the name Salankayana. 13
Salankayana not the no/we of a tribe
The attempts of some scholars to see in Salankayana the name
of a people, becoming subsequently the name of a dynasty also, is
wasted effort . Dr . Rayachaudhuri has identified the ' Salakenoi '
8. ' Maudgalya-sagotrasya ' The Kanteru plates of Nandivarman, I
9. ' Bhdraddayassa ' (The Carudevi grant); ' Bhdradddyo ' (The Hlra-
Ha^agalli grant) ; 'Bharadvajah' (The Pikira and Cendalur plates) ; 'Kasyapaya
Vilasa sarmmane (El, Vol. VIII, p. 162).
10. In the Komarti plates of Can<Javarman and the Chicacole plates of
Nandaprabhanjanavarman (El, Vol. IV, pp. 142 ff, and IA, Vol. XIII, pp. 48 ff)
we miss the dynastic and the gotra names. We miss the dynastic name, in
all but three Satavahana records.
11. IA, Vol. V, p. 175; Vol. IX, p. 101.
12. Ibid., p. 102. There is a Salankayana gotra which has the pravaras
Visvamitra, Katya, and Atkila.
13. Sircar, op. cit., p. 70.
*HE VAINGEYAKAS 1*5
of Ptolemy with the ' Salankayanas ' of Vengl. 14 D. C. Sircar
accepts this identification 15 and adds : " It has been noticed 16 that
the terms Salankayana and Salankayanaka (country of the
Salankayanas) are mentioned in the Ganapatha of Panini. It is
certain that the Salankayanas (Greek Salakenoi) ruled
over the Vengl region as early as the time of Ptolemy (c, 140
A.D.)." He would further consider Benagouron as a mistake
for Bengaouron 17 which would represent Vengipura. Having regard
to the fact that in Ptolemy's Book the n sound is not suppressed
e.g. Gangaridai (Book VII, Chapter 1, Section 81) and Peririg-
karei (Section 89), Salakenoi can be rendered Salakana and not
Salahkayana. The Salakenoi are placed north of the river Mana-
das which is almost certainly the Mahanadi, the great river
of Orissa, far north of the Andhradesa of literature. Kings of the
Salarikayana gotra ruled over the heart of the Andhradesa and
the suggestion of some scholars that they ruled over Kalihga and
Magadha (!) lacks proof; 18 and Ptolemy's description of the
eastern part of the peninsula is not as much vitiated by errors as
that of the western and southern parts. The Sdlankdyanaka of
Panini does not mean ' the Country of the Salankayanas '; it is the
adjectival form of Salatikayana which belongs to the Rdjanyddi
class. Names like Athenogouron make it highly improbable that
Benagouron is a mistake for Bengaouron. Benagouron would cor-
respond to Benanagara ; and as several Benas are known, and the
Benagouron of Ptolemy is on the banks of a river, a Benanagara is
not impossible. 19 The Benagouron of Ptolemy is not called a
metropolis, while Vengipura was certainly the capital of the kings
of the Salankayana gotra.
dlankdyana of the inscriptions has nothing to do with the bull
banner of the Vaingeyakas
According to the Trikdndasesa and the Medinlkosa, Salan-
kayana also means Nandin, the vehicle of Siva. 20 It is interesting
14. Political History of Ancient India, (4th ed.), p. 419, n. 1.
15. Op. cit., p. 70.
16. JAHRS, Vol. V, Pt. i. p. 23.
17. The letters might have been transposed in copying.
18. Vide infra.
19. A Benakataka is known from Nasik No. 4, El, Vol. VHI.
20. Mahakfllo mahabhimo, mahak&yo vrsanakah dvah sthastu naudt-
6alankayanastandavatalikAh (Trikandatesa si 49).
HA.-22
166 EARLY HISTORY OF THE ANDHRA COUNTRV
to note that the crest of the kings of the Salankayana gotra is the
bull. 21 Since Salankayana of our inscriptions is a gotra name, the
heraldic device cannot be connected with either the gotra or the
dynastic name but must be explained on other grounds. Neither
the bull banner of Pallavas nor the boar banner of the Calukyas
had anything to do with their dynastic names. Jayavarman of
Brhatphalayana gotra was a worshipper of Mahasena, and on the
seal of his plates we have a representation of the trident of Siva. 22
Some coins of Wima Kadphises bear the representation of Siva with
the combined trident and battle-axe, and the legends ' Maha-
rdjasa-Rajadirajasa sarvaloga Isvarasa-Mahlsvarasa Wima-Kathphi-
sasa'. On some others we have a representation of Siva and his
vehicle. 23 In the Mandasor pillar inscription of Yasodharman there
is a reference to Nandi as an an emblem on Siva's banner. 24 On
the seals of the grants of paramamdhesvaras like Dharasena II 25
the Maukhari Sarvavarman, the Ganga Indravarman 26 and the
Ganga Devendravarman 27 we have a representation of the bull.
The bull badge of the Vairigeyakas is therefore almost certainly
connected with their sectarian leanings. Once adopted by parama-
mahesvaras like Devavarman it was continued by even parama-
bhdgavatas like the Nandivarmans. 28
The Dynastic name adopted here is ' Vaingeyaka '
Since, as has been shown, the only name occurring in the
records of these kings is a gotra name, we would avoid a confusion
21. Of the five grants of these kings the seals of the Kollair and Pedda-
vegi plates are much defaced. On the seal of the Kanteru grants of Nandi-
varman I (No. 2 of 1924-25) and of Skandavarman a couchant bull facing left
with its hump is clearly visible. (However the horns and in the latter the
legs are not visible). The plates are now in the Government Museum
Madras. The device on the seal of the Ellore plates was thought by Mr. Ven-
kayya to be that of 'some quadruped, perhaps a tiger.' An examination of
the seal shows the body of a bull, facing left, much the same as that on the
seal of the Kanteru grant of 1 Skandavarman. The hump is partially visible.
22. El, Vol. VI, p. 315.
23. Catalogue of Coins in the Indian Muaeum, Calcutta, pp. 68 ff .
24. CH, Vol. HI, p. 146.
25. Ibid., p. 164.
26. IA, Vol. XIII, pp. 119-20.
27. Ibid., pp. 273 ff.
28. We may liken this to the Saivite names borne by the paramabh&ga-
vatas, Kumiragupta and Skandagupta and the Nandivarmans referred to
above.
THE VAINGEYAKAS 167
between gotra and dynastic names, if we cease to call them Salan-
fcayanas and tentatively gave them another dynastic name. For-
tunately, the Allahabad pra&asti of Samudragupta calls Hasti-
varman of this line a Vaingeyaka. As kings of the Salankayana
gotra are the earliest known ones who ruled from Vengipura, their
dynasty may be styled Vaingeyaka, 29
Origins of the Dynasty
A thick veil is drawn over the origins of the dynasty. It has,
however, been shown that D. C. Sircar's theory of the existence of
the dynasty as early as the time of Ptolemy, and perhaps of Panini,
lacks proof. The passing away of the great Satavahana power
would have given a tremendous fillip to the disintegrating forces
already at work during its decline. Whilst the southern part of the
empire came under the Pallavas, the south-western parts of the
empire under the Cutus and after them the Kadambas, the Andhra-
desa fell under less powerful and more short-lived dynasties. In less
than four and half centuries it saw the Iksvakus, the kings of the
Brhatphalayana gotra, the Vairigeyakas, the Kandaras and the
Visnukundins, rise and fall in quick succession. Since the later
kings of the Salankayana gotra were in possession of Kudrahara
which is identical with the Kudurahara of Jayavarman's plates, it is
certain that the former rose to power at the expense of the rulers
of the Brhatphalayana gotra. The general opinion of scholars is
that the Vaingeyaka did not rise at the expense of the Pallava. It
is argued that the starting-point of Vaingeyaka power was Vengi
identified with Peddavegi 30 near Ellore in the West Godavari dis-
trict, and therefore north of the river Krsna. There is no evidence
to show that the Pallavas ever crossed the Krsna. 31 Even so the
Mayidavolu plates show that Dhamnakataka (modern Dharani-
kot) and the Guntur and Narasaraopet taluqs were reached by the
Pallava arm ; as this part of the Guntur district came under the
Vaingeyakas later on, it is highly probable that the latter rose to
prominence at the expense of the former also.
29. We do not know whether Vaingeyaka refers to the kingdom of
Verigi or the city of Vengi. In the same inscription some kings are known by
the name of their country e.g. Daivarastraka Kubera and others by the name
of their capital e.g. K&nceyaka Visnugopa.
30. El, Vol. IX, p. 58.
31. All the Prakrt and Sanskrit grants of the early Pallavas have been
found south of the Krsnfi, in the Guntur, Nellore and Bellary Districts.
168 EARLY HISTORY OF THE ANDHRA COUNTRY
Devavarman : He was not the first king of the line
The earliest known member of the dynasty is Devavarman. 32
But the expression bappa-bhattdraka-pddabhakta in his Ellore
grant 33 makes it certain that his father (bappa) was an independent
ruler, for bhattaraka of the Vaingeyaka and Pallava grants, like
paramabhattdraka of the Gupta and Valabhi records, is a title
applied to independent kings. 34 The old view that Devavarman is
the first king must needs be given up and the origins of the dynasty
traced to the first quarter of the fourth century A.D.
32. Like the other Vairigeyakas Devavarman has the honorific prefixes
'Sri' and Vijaya'. Dr. Fleet (IA, Vol. V, p. 175) and Dr. Hultzsch (El, Vol. IV,
p. 143) considered 'vijaya' as an integral part of personal names. Later on
Dr. Hultzsch somewhat modified his view by putting a hyphen between
Vijaya* and the name proper. (El, Vol. IX, p. 58). Even here, he spells
Vijaya' before 'Vengipura' and 'sarhvacchara* with small v and that before
personal names with capital V. In the expressions ' vijaya samvacchara,'
' vijaya rdjya sathvatsara', 'sri vijaya Vehgipurdt' which occur in the Vain-
geyaka grants Vijaya' is certainly, an honorific prefix like 'sri' 'srimad' 'siva/
'jaya' (Bhandarkar List No. 1528) and 'deva' (Allan, Catalogue of Indian
Coins, Gupta Dynasties, Index) . The Kancipura of the Mayidavolu and Hira-
Hadagalli grants is called Vijaya Kancipura' in later records. In the Kadamba
grants we have 'sri vijaya Paldika' and 'vijaya Vaijayantipura.' True these
prefixes sometimes enter into the composition of names e.g. Jayavarman and
Devavarman. What makes it certain that in the Vaingeyaka charters Vijaya'
like 'err is an honorific prefix, is the fact that Nandivarman II calls himself
simply *6ri-Nandivarman' in the Peddavegi plates and 6ri vrjaya-Nandivar-
man in the Kollair plates.
33. This expression occurs in all the Vaingeyaka grants.
34. 'Bhattaraka' is a title applied to gods and priests (CII, Vol. Ill,
Nos. 28 and 46 and Luders' List, Nos. 43 and 1076). In
a Nasik inscription of Usavadata (El, Vol. VTEI, No. 10) 'Rdjan
Ksatrapa Nahapana' is styled 'bhataraka' (bhattaraka} . That it
was a title applied to Pallava Maharajas is shown by the Uruvupalli grant
where it is coupled with the title of Maharaja (bappa-bhattdraka-Mahdrdja-
pdda-bhaktah) . Dr. Fleet has pointed out that in the Gupta and Valabhi re-
cords 'bhattaraka' and 'paramabhattaraka' are titles applied to paramount
sovereigns (op. cit, p. 17 and n. 1); that 'bhattaraka' in the expression under
reference is a title applied to kings is evident from the fact that in Pallava
grants of the same period (which resemble the Vaingeyaga grants in many
respects) it is applied to Maharajas.
No term is so difficult of explanation and yet occurs over all parts of
India as the term 'bappa' in the expressions 'bappa-bhattdraka-pdda-bhaktah'
(in the Pallava, Vaingeyaka and some early Kalinga grants, El, Vol. IV,
pp. 142ff., Vol. XH, pp. 4ff., Vol. XXI, pp. 24ff.), bappa-pdddnu dhyatah
and f bappa-pddaparigrhita ' (in inscriptions from Nepal and in the copper
THE VAINGBYAKAS 169
But was the predecessor of Hastivarman
Before the Peddavegi plates were discovered Mr. K. V. Laksh-
mana Rao considered Devavarman as the successor of Hastivarman
of the posthumous pillar inscription of Samudragupta, whom he
rightly conjectured to have been a king of the Salankayana gotra.
Even with the Peddavegi plates (No. 3 of 1924-25) before him,
Mr. M. S. Sarma 35 considered Devavarman as the son of Hastivar-
man and as the elder brother of Nandivarman I. The mistaken
notion that a Prakrt grant of Nandivarman I existed in Sir Walter
Elliot's facsimiles was partly responsible for this view. It was
plate grants of the Calukya dynasty JBBRAS, Vol. XVI, pp. 3, 4 and
5) , and 'parama-bhattdraka-MaMrdjddhirdja-parame&vara-sri-bappa-pdddnu-
dhydtah' (in the Valabhi grants). Dr. Fleet's theory (CII, Vol. HI, p. 186 n.)
is that since in the Valabhi grants (we may add in the Uruvupalli grant too)
'bappa' is connected with the paramount titles of Mahdrdja, Mahdrdjddhirdja,
Paramabhattdraka and Paramesvara, since ajjaka, (Pali ayyaka, see Pischel,
Gram. Der Pra. Spra. Sec. 252) in the expression 'ajjaka-pdddnudhydtah'
applied to Dharasena IV (JBBRAS, Vol. X, p. 79 and IA, Vol. I, p. 16), is
the old Prakrt form of the modern Canarese 'ajja' and Marathi 'aja,' 'bappa',
suggests itself at once as the old Prakrt form of the modern 'bap, 'father,'
(We may add of the Canarese 'bappa' 'father').
But while 'ajjaka 1 is certainly a Prakrt word, 'bappa' meaning father
cannot be traced. Pali 'bappa' (Sans, baspa) means 'tear', since the founder
of the Mewar dynasty is called ( bappa' in one inscription (Prakrt and Sanskrit
Inscriptions from Kathiawar, pp. 88-89) and 'baspa' in another (ibid., pp. 75,
78), 'bappa' of our inscriptions also would seem to be the Prakrt form of
'baspa'. What then is the meaning of 'bappa'? A way out of the difficulty is
suggested by the personal names Bapisiri (Bappisiri) and Bapaka (Bappaka,
see Pischel, op. cit., sec. 305) which occur in inscriptions of the second cen-
tury A.D. (Liiders' List, No. 1213 and El, Vol. XVI, p. 235). In Nagarjuni-
konda inscriptions a royal lady bears the name Bapisirinika. In the Hira-
Hadagalli grant a predecessor, probably the father of Pallava Siva-Skanda-
varman, is called Bappasami. This is a personal name for the following
reasons: in later inscriptions Bappasvamin is a personal name, (El, Vol.
XIX, pp. 247 and 249, Vol. XI, p. 19). Bappisiri is the feminine form of
Bappasiri and 'siri' like 'samin' is a suffix. In the Hlra-Haolagalli inscription
itself Bappa is the name of a donee. It is then reasonable to suppose that an
early member of the dynasty of the Pallavas or the Vaingeyakas, perhaps a
founder, bore the name Bappasvamin or Bappa. His son or descendants may
have referred to him in the expression 'bappa- bhattdraka-pddabhaktah.'
Subsequently the practice of referring to father and the unwillingness to
change an old expression might have resulted in the same expression being
kept with 'banpa', however, referring to or meaning 'father*. This conjecture
is made probable by the fact that in Mara^hi 'boy means 'father',
35. ARE, 1924-25, H 2,
170 EARLY HISTORY OF THE ANDHRA COUNTRY
argued "that the assignment of the kings Vijaya Devavarman, Vijaya
Nandivarman and Yuvardja Buddhavarman of the Prakrt grants,
to a period subsequent to Samudragupta whose inscription is in
classical Sanskrit is not tenable." Here is indeed a weak argu-
ment for the correct view. In the Peddavegi plates which proba-
bly trace the descent in one line from Hastivarman to Nandivarman
II, Devavarman's name is not found and therefore Devavarman
must be placed either before Hastivarman or after Nandivarman II.
Devavarman's inscription is in literary Prakrt and as Dr. Hultzsch
has noticed, 36 in one respect the language is more archaic than
that of literary Prakrt, i.e., single consonants between vowels re-
main unchanged. The inscriptions of Nandivarman II (the Kol-
lair and Peddavegi plates) are in Sanskrit ; since it is an accepted
view that in official grants Sanskrit replaced Prakrt, Devavarman
cannot be placed after Nandivarman II. The palaeography of the
plates under reference furnishes corroborative evidence. The
general appearance of the Ellore plates is more primitive than that
of the plates of Nandivarman II. The Ellore subscript va (t. 11. 8,
9, 14) which resembles the Kondamudi and Mayidavolu va, the tha
with the dot in the centre which, in later inscriptions, is replaced
by a short stroke in the centre or a curve attached to the left, and
the sa are prominent instances.
Date of Devavarman
Since Devavarman's inscription is dated in his regnal year
(13th year) it is not easy to fix his date. The lower limit is the
date of Samudragupta's southern campaign; in the Allahabad ins-
cription Hastivarman Vaingeyaka, who came after Devavarman,
figures as one of the kings of the Daksinapatha 'captured and liberat-
ed' by that Indian Napoleon during his 'digvijaya? and all that we
can say at present is that Samudragupta's digvijaya took place
before 380 A.D. 37 The upper limit is arrived at thus: The Sata-
vahanas passed away about 200 A.D. According to the Puranas,
the Iksvakus who succeeded them in the Krsna-Guntur region,
ruled for fifty-two years. A period of twenty-five years may be
allowed to Jayavarman's dynasty that succeeded to the political
heritage of the Iksvakus in parts of the Krsna-Guntur region. It
has been already shown that at least a generation separates Jaya-
varman from Bappasami, a predecessor (probably the father) of
36. El, Vol. IX, p. 57.
37. Allan, Catalogue of Indian Coins, Gupta Dynasties, xx, xxxi, xxxii,
THE VAINGEYAKAS 171
Pallava Siva-Skandavarman. It has also been noted that the Siva-
Skandavarman of the Mayidavolu grant is in all probability the
Vrjaya-Skandavarman of the CarudevI grant. Calculating on the
basis of the usually accepted rate of twenty-five years for a Hindu
generation, we arrive at 340 as the approximate date of the Caru-
devi grant. 38 The language, orthography and the Sanskrit verses
of the CarudevI and Ellore grants would favour the same period for
both. 39 But the phraseology of the Ellore grant which came to be
widely accepted later on (especially in the Pallava grants) favours
a slightly later period for it. 40
Events of his reign
Sometime before the thirteenth year of his reign Devavarman
performed the Asvamedha ; for he styles himself ' AssamedhaydjV
Mr. K. V. Lakshmana Rao disposes of this epithet as a vain boast
that minor dynasties indulged in to emulate the Guptas;
examples being the Cedis, the Vakatakas, the Kadambas,
the Salarikayanas, and others. The Naneghat sacrificial
inscription (2nd century B.C.) enumerates the numerous
sacrifices performed by an early Andhra ruler and goes
into details about the daksinas offered. It states that a second
horse-sacrifice was performed. 41 The Vi$nukunoUn king Madhavar-
38. The date of the grant is lost.
39. Both are in literary Prakrt. In both the etymological spelling of the
Pandit is adopted. In both grants every side of the plates is numbered.
Palaeographically the two grants are apart. There is on the other hand
much resemblance between the archaic Telugu-Canarese script of the
Vaihgeyaka grants and the grantha script of the Pallava charters of the
4th or 5th centuries. (The few points of difference have been
noted by Biihler in his Ind. Pal. pp. 70-71, Sec. 31) . The Ellore grant is the
only Prakrt inscription where the letter n occurs (Vengi and Salarikayana) .
In other Prakrt inscriptions including the CarudevI grant it is represented
by the anusvara (the letter occurs in early, mixed dialect inscriptions
Liiders' List, Nos. 64a, 129, 130, 131, 133, etc.); and in Prakrt h sound is some-
times retained. (See Pischel op. cit., Sec. 381, 386, etc.). This fact also
favours a later date for the Ellore grant.
40. The 13th year of Devavarman would then be somewhere between
360-370 A.D. On palaeographical grounds, Dr. Burnell refers Nandivarman II
to about the fourth century A.D. (S. Ind. Pal., p. 14, n. 2) and Dr. Fleet
quotes him approvingly (IA, Vol. V, p. 176). Biihler leaves the question un-
settled (Ind. Pal,, p. 65, Sec. 29A) . Since Devavarman is removed from
Nandivarman II by at least four generations, the chronological arrangement
proposed by Dr. Burnell cannot fit into known facts.
41. 'Asamedho bitiyo (yi)tho'U B, t. L 1,
172 EARLY HISTORY OF THE ANDHRA COUNTRY
man I is said to have performed eleven Asvamedhas. The Vakataka
king Pravarasena is credited with the performance of four As-ua-
medhas. 42 The number of sacrifices mentioned makes the theory
of boast untenable. The Iksvaku Icing Caihtamula is credited with
the performance of Agnihotra, Agnistoma, Vdjapeya and Asva-
medha. Pallava Siva-Skandavarman is also styled an Assamedha-
ydji. The Asvamedha was performed either before a king set out
on a campaign of conquest (digvijaya) or in celebration of con-
quests. It involved an assertion of power and a display of politi-
cal authority. There is no reason why the king of a small kingdom
should not have celebrated his conquests and asserted his authority
over the various parts of his kingdom in this manner. Economi-
cally too, the Asvamedha was not impossible for a petty prince. If
'Assamedhaydji' were a mere boast, every king could have styled
himself an ' Assamedhayajl.' Among the Iksvakus only Caihtamula,
among the early Pallavas only Siva-Skandavarman, and among the
Vaingeyakas only Devavarman, bear this title.
Much less can this title be traced to a Gupta source. It is
borrowed from the Iksvakus and the Pallavas. The titles applied
to Samudragupta are 'Asvamedhaparakramah' on some coins attri-
buted to him, and ' Asvamedhdhartd ' in the inscriptions of his suc-
cessors. 43
The only other event of his reign of which we have record is
that in the thirteenth year of his reign, he granted 20 nivartanas of
land to Ganasarman of the Babhura(?) 44 gotra and house sites
for the Brahman, his tenants and doorkeepers.
42. CII, Vol. Ill, No. 55.
43. Mr. Divekar is of opinion (ABI, Vol. VII, pp. 164-65) that, since
the Allahabad inscription of Samudragupta does not make mention of the
Asvamedha performed by him, he performed it after the Allahabad inscrip-
tion was engraved.
In many cases, the Guptas seem to have been the borrowers. The
earliest known inscription in which the expressions ' paramamahe&vara '
(paramabhagavata is a similar expression) and 'vijayasamvatsara? occur, is
that of Devavarman.
44. Dr. Hultzsch read this word as 'Ba(bhura).' Prof. Keilhorn suggest-
ed that it might be meant for 'Babhru'. Considering the facts that the inscrip-
tion presents archaic and modern forms of some letters (compare va in 11. 1, 6
and 7 with va in 11. 8 and 14; and bhu in 1. 9 with bhu in 1. 10 and bha in
1. 18), and that the u sign in ku in the word 'Kuravaka' (Kollair plates) is
not represented by a short curve attached to the right of the vertical but by
a modification of the reascent of the vertical to the left, 'Babhura.' (Sanskrit
THE VAINGEYAKAS 173
Extent of his kingdom
Whilst it is certain. that Maharaja Devavarman ruled over a
small territory around Vengi, we do not know whether Kudrahara
of the later Vairigeyaka grants was included in his kingdom
or not. If Prof. Kielhorn's identification of Kurala (Allahabad
pillar inscription) with the Kollair Lake 45 can be accepted, the
Kollair region (excluding modern Ellore) with probably Kuraja
as its capital 46 was ruled over by a separate line of kings, one of
whom was Man^araja, a contemporary of Hastivarman. 47
Importance of Devavarman's Grant
Devavarman's grant is the most interesting of the Vaingeyaka
grants. It is the earliest known inscription in which occur the ex-
pressions 'paramamdhesvara' and 'bappa-bhattdraka-pdda-bhaktah,'
which are of frequent occurrence in later records all over India.
It is the first record in this side of India to abandon the
Satavahana method of dating (by the fortnight of one of the three
seasons of the year, i.e., gimhdnapakha, vasanapakha and heman-
tdnapakha) for that of dating by the month and tithi** This
method of dating is followed in the early Kalinga Sanskrit char-
ters, in the inscriptions of the Pallavas, the Guptas, the Vakatakas
and the Kadambas. 49
Babhru, i.e., Babhrava gotra) is the proper reading. (El, Vol. IX, p. 59,
n. 7).
45. According to Prof. Kielhorn (El, Vol. VI, p. 3, n. 3) 'jalam Kaun&lam'
of the Aihole inscriptions can only mean the 'Kollair Lake* as the description
of the water given in the poem would be applicable to it even at the present
day and as Kolanu of the later inscriptions is a corruption of Kunaja. Kunala
and Kurala of the Allahabad pillar inscription are identical, because the
former is mentioned just before and the latter after Pistapuram. Could it be
that Kurala changed into Kunala in less than three centuries much in the
same way as the latter changed into Kolanu in inscriptions of the eleventh
century?
46. In the Chellur plates of the reign of the Eastern Calukya Kulottunga
Coda n (6. 1056) we are told that hi the midst of a great lake in the Vengi-
man^ala (the Kollair Lake) there is a town named Saraslpurl.
47. Allahabad pillar inscription of Samudragupta.
48. The Western Ksatrapa records are dated by the month and tithi of the
year, but the expression Vijayasaihvatsara* does not occur in them.
49. The exceptions are the Devagiri plates of Kadamba Mrgesavarman
(IA, Vol. VIT, p. 37), the Dudia plates of Vakafcka Pravarasena (El, Vol. m,
p. 260), the Halsi plates of the reign of Kadamba Ravivarma (IA, Vol. VI,
p. 28), the Omgotfu plates of Skandavarman n (El, Vol. XV, p. 249), the
if 4 EARLY HISTORY OF THE ANDHRA COUNTRY
Hastivarman
The chronological arrangement proposed above not only makes
the identity of Hastivarman of the Peddavegi plates with the Hasti-
varman of the Allahabad pillar inscription of Samudragupta cer-
tain, 50 but also makes it probable that he was the immediate suc-
cessor of Devavarman. Much cannot be made of the epithet
'samardvaptavijayino' applied to him, as the credit for having won
many battles is taken by every prince and as similar epithets are
indifferently applied even in the case of sovereigns not far removed
from one another. 51 But it is certain that the whirlwind campaign
of Samudragupta, perhaps because of its transient nature in the
south, did not interrupt the fortunes of the Vaihgeyakas, and Hasti-
varman must have resumed his normal course, when the brief
disturbance of Samudragupta 's inroad passed away.
Nandivarman I: A misconception about him
The son and successor of Hastivarman was Nandivarman I; in
the Peddavegi plates he is credited with many gifts. An oversight
has vitiated all theories about Vaingeyaka genealogy. While edit-
ing the Kollair plates of Nandivarman II Dr. Fleet remarked 52 :
"In Sir Walter Elliot's facsimiles I have another copper plate ins-
cription of Vijayanandivarma and his Yuvamaharaja, whose name
seems to be Vijayatungavarma or Vijayabuddhavarma ;
the characters are, in fact, so rude and indistinct, that I doubt
whether a transcription of it can be made. The language,
seems to be Prakrt or Pali, as the first line commences
' (Sva)sti'Srivijayanandivarmma~malwrajassa, ' " Subse-
quently 53 Dr. Fleet gave a transcript of the inscription and observed
Hlra-Hadagalli and Mayidavolu plates of diva-Skandavarman I and all the
Visnukundin inscriptions except the Ramatirtham plates. In these, the Sata-
vahana method of dating is followed.
50. It is curious that in SewelTs Historical Inscriptions of South, India
(1932), Hastivarman of Verigl is called a Pallava Viceroy.
51. The epithets applied to Skandavarman in the Uruvapalli grant are
applied to his son, VIravannan in the Pikira grant. Epithets like 'prat&po-
panata rdjamandalasya' and 'abhyarchita 6akti siddhi xampannasya' are
applied to Yuvamaharaja Visnugopavarman's father in the former grant and
to his grandfather in the latter. Instances can be multiplied.
52. IA, Vol. V, pp. 175-176.
53. Jbid., Vol. IX, p. 101.
THE VAINGBYAKAS 175
'Vijayabuddhavarma is said to be a Pallava and of the Bharattayana
or Bharadvfija gotra. There is, therefore, no genealogical connection
between the Vijayabuddhavarma of this grant, and Vijayanandi-
varma of the Vehgi grant at Vol. V, p. 175, who was of the Salanka-
yana gotra? We might add, the father of Buddhavarma is Skanda-
varma. This grant, now called the British Museum plates of Caru-
devi, has been edited by Dr. Hultzsch in the pages of Epigraphia
IndicaJ 54 This correction has been overlooked by almost all scholars.
Prof. Jouveau-Dubreuil identified Nandivarman of 'Elliot's unpub-
lished Prakrt grant' with Nandivarman of the Kollair plates! K. V.
Lakshmana Rao, while repeating the mistake, placed 'Vijayanandi-
vannan' and 'Vijayabuddhavarman' after Devavarman and before
Carwjavarman. With the Peddavegi plates before them the Epigra-
phy Department identified 'Vijayanandivarman' of 'Elliot's unpub-
lished Prakrt grant' with Nandivarman I of the Peddavegi plates
and made Buddhavarman the elder brother of Candavarman. If
only Dr. Fleet's correction had been noted, all these mistakes could
have been avoided. 55
Nandivarman of the Kanteru grant is probably Nandivarman I
While editing the Kanteru grants, 56 K. V. Lakshmana Rao
identified Nandivarman of one of them 57 with Nandivarman II
(known to us from the Kollair and Peddavegi plates) . He has been
followed by D. C. Sircar and others. This view cannot, however,
be upheld. In both the grants Nandivarman II calls himself the
eldest son of Candavarman. 58 Nandivarman of the Kanteru grant
is not so called. True in two grants, 59 Kadamba Mrgesavarma is
called the eldest son of Santivarma, and in one, as simply
the son of Santivarma. 60 Even so, Nandivarman of the Kanteru
grant is not even called the son of Candavarman; and when we
consider the remarkable similarity in the phraseology of the Vain-
geyaka grants, this omission is all the more significant. In both
54. Vol. VIII, pp. 143ff.
55. D. C. Sircar also has pointed out this error. (Succetsors of the
vahanas in the Eastern Deccan, pp. 57-58) .
56. JAHRS, Vol. V, Part i, pp. 22 ff.
57. No. 2 of 1924-25.
58. 'Mah&raja Can^avarmanas-sunurjyetfali' Kollair plates.
jasya Candavurmanah putro jyeftah* Peddavegi plates.
59. IA, Vol. VI, pp. 24-25, and 27-29.
60. IA, Vol. VI, pp. 25-27,
176 EARLY HISTORY OF THE ANDHRA COUNTRY
the grants of Nandivarman II, 61 a 'Mulakura bhojaka' is mentioned
as the ajnapti. The Kanteru plates do not mention any djnapti. 62
The palaeography of the plates also supports our view. In the
Kanteru grant of Nandivarman the left half of the horizontal mem-
ber of ka is looped. This feature is not to be found in any other
Vaiiigeyaka grant. Whilst in the Peddavegi and Kollair plates the
reascent to the left of the verticals a, ka and ra and the u sign in
lu, yu, nu and mu, is up to half the length of the vertical and some-
times more, in the Kanteru plates, the reascent is much less. 63 This
makes their attribution to Nandivarman Ts time possible. 64
Extent of his kingdom
If the conclusion tentatively proposed here can stand, Kudra-
hara visaya was included in Nandivarman I's kingdom. 65 . The
extent of this province is not, however, easy of determination. In
the ARE of 1924-25, Kuravata (there read as Kuravata) was
identified with Kurada in the Gudivada taluq and Lakumari (there
read as Lekumari) with Lokamudi in the Kaikalur taluq. This
would make the Krjpna the southern boundary of the kingdom.
Since Kanteru, six miles north of Guntur, is the findspot of two
Vaingeyaka grants, since Cittapura can be identified with Cintala-
pudi in the Tenali taluq, it is highly probable that a part of the
modern Guntur District was included in Kudrahara, as in the time
of Jayavarman. What would be a conjecture is raised to a cer-
tainty by the term 'Vehgordstra* of the Mafigalur grant of
Yuvamaharaja Visnugopavarman issued from Dasanapura (tenta-
tively identified with modern Darsi by Mr. Venkayy*) . ' Ven~
61. Kollair and Peddavegi grants.
62. We might be in a position to discuss the question more satisfactorily
if the plate containing the injunctions to officers and the date had not been
missing.
63. Especially in U. 1 and 2.
64. An examination of the seal reveals five letters by the side of and
above the bull (not below the bull as K. V. Lakshmana Rao thought
JAHRS, Vol. V, Pt. i, p. 22) . They appear to be ma, ha and ra, na (or no)
and 6ri. Between the third and fourth letter there is space for two or three
letters. However, the first three letters are bigger than the other two.
According to K. V. Lakshmana Rao, there are only two letters "one of
which appears to be ra and another nu."
65. The inscription records the grant of 12 nivartanas of land in Kura-
vata village in the Kudrahara visaya to Svamicandra of the Maudgalya
gotra.
THE VAINGEYAKAS 177
gora?tra' is probably a scribal error for e Vengird&ra, 9 for in all
records we have Vengidesa or Vengimandalam. Mangalur may be
identified with Mahgalagiri in the Guntur taluq. Only a part of
the Vairigeyaka kingdom conquered by the Pallavas could have
been so named. No Pallava inscription speaks of the conquest of
Vengi and the grants of Simhavarman, his father Visnugopavarman,
and his grandfather Skandavarman (fourth and fifth centuries),
have been found south of the Krsna. That part of the Vaingeyaka
kingdom which was conquered by the Pallavas must have lain
south of the Krsna.
Candavarman
Cantfavarman was the son and successor of Nandrvarmau I. 68
While editing the Komarti 67 plates of Candavarman of Kalinga
(Kalingadhipati) , Dr, Hultzsch remarked that considering the
similarity in names, the expression c bappa-bhattdraka-pdda-
bhaktah ' and the close resemblance between the alphabets of the
two grants "the father of Vijayanandivarman may have been
identical with the Maharaja Charujavarman who issued the
Komarti plates. At any rate, the two Chandavarmans must have
belonged to the same period. An examination of the seal, which,
according to Sir W. Elliot, is defaced, would probably show if it reads
Pitribhaktah and if, consequently, the plates of Vijayanandivar-
man may be assigned with certainty to the same dynasty as the
Komarti and Chicacole plates." 68 Adopting this suggestion (i.e.,
that the Vaingeyakas ruled over Kalinga) Kielhorn classed the
Kollair grant as a North Indian inscription. 69 D. R. Bhandarkar
who has recently revised and enlarged Kielhorn's list has not
availed himself of the fresh evidence available. 70
In fairness to Dr. Hultzsch it must be stated that he was aware
that the phraseology of the Komarti plates resembles that of the
copper-plates of the Gangas of Kalinga and much more closely 71
that of the Chicacole plates of Nandaprabhanjanavarman. Moreover
66. The Peddavegl plates. 'Canda' means 'wrathful, violent, fearful';
probably here is a reference to the fearful form of Bhairava (Siva).
67. Ganjam District.
68. El, Vol. IV, p, 143.
69. Inscriptions of Northern India, No. 686.
70. El, Vols. XIX-XXI, No. 1908.
71. IA, Vol. XHI, pp. 49 ff. and PI.
178 EARLY HISTORY OF THE ANDHRA COUNTRY
when he wrote, only one Vaingeyaka grant was known. He could
have however withdrawn his statement when he edited the Ellore
grant of Devavarman, but did not do so. The phraseology of the
Komarti grant is totally different from that of the Vaingeyaka
grants which exhibit a remarkable similarity among themselves.
Candava>rman of the Komarti plates is styled 'Kalihgadhipati' and
issued the charter from Simhapura identified with the modern
Singupuram between modern Chicacole and Narasannapeta; all the
Vaingeyaka grants were issued from Vengipura, and in
none of the six Vaingeyaka kings known to us is the ruler
called ' Kalingadhipati'. It is not stated that Can<Javarman, the
'Kalingadhipati', belonged to the Salankayana gotra and was a
devotee of 'Citrarthasvami'! No conclusion can be based upon
the expression ' bappa-bhaftdraka-pada-bhaktali ' as it was used by
various dynasties. 72 Similarity in names (in the case of the
dynasties of the eastern Deccan between the third and sixth cen-
turies) proves nothing. Candavarman might have been as much a
name common to the Vaingeyaka and Kalinga dynasties, 73 as Hasti-
varman was to Ganga Vaingeyaka and Kandara dynasties and as
Skandavarman and Nandivarman were to the Pallava and Vain-
geyaka dynasties. 74
The epithet ' pratdpopanata sdmantah' applied to Candavar-
man in the Peddavegi plates would seem to be conventional.
Nandivarman II
Nandivarman II is referred to as the eldest son of Candavar-
man ; 75 the latter, then, would seem to have had two or more sons.
Of the events of Nandivarman's reign nothing is known except that
he granted the village of Vide (de) turapallika 76 in Kudrahara to
72. Pallava and Kalinga dynasties. Also vide supra.
73. Canda often enters into the composition of personal names. Canfla-
mahasena (El, Vol. XXI, p. 178), Candaketu, Candavikrama, Candasimha;
Canflaprabha, etc.
74. Dr. Hultzsch has not noted the difference between the alphabets of
the Kollair and Komarti grants. The thick dots or nail-heads at the top and
bottom of letters in the latter grant are not so prominent in the former. The
Komarti va lacks the notch of the Vaingeyaka va.
75. The practice of naming the grandson after the grandfather was not
only prevalent among the Ikhakus, the Guptas, the Vakatakas, the Pallavas
and the Visnukundins but was sanctioned in books. (El, Vol. XX, p. 6, n. 2).
76. Dr. Fleet read it as Videnurapallika. The reasons for adopting the
reading given here are stated below.
fHE VAINGEYAKAS 179
157 Brahmans of various gotras and caranas living in the excellent
Kuravaka agrahara in the 7th year, and 10 nivartanas of land in
Arutora, 10 in Mun^uru, 6 in Cenceruva and 6 in Karurhburan-
ceruva as 'devahalam' to Visnugrhasvdmin (God in the Visnu
temple), lord of the three worlds, in the tenth year.
Skandavarman
It is no easy to settle Skandavarman's place in Vaihgeyaka
genealogy ; those who identify Nandivarman of the Kanteru grant
(No. 2 of 1924-25) with Nandivarman II consider Skandavarman
of the Kanteru grant (No. 1 of 1924-25) as his brother. The
sources of this assertion are the epithet 'sunurjyesthah* applied to
Nandivarman II and the common findspot of the two grants. But
it has been shown above that Nandivarman of No. 2 of 1924-25, is
probably Nandivarman I. Since Skandavarman and Nandivarman
grant lands to persons belonging to the same gotra, probably father
and son, and since the two grants come from the same place,
Skandavarman might have been a brother and successor of Nandi-
varman I, omitted in the genealogical list much in the same way as
Devavarman in the Ramatirtham plates and Vikramendravarman I
in the Ipur plates set II. Palaeographically no objection can be
taken to this view. 77
Contemporary Powers
The theory that the Vaingeyakas were somehow related to the
Pallavas of the early Sanskrit charters (i.e., Skandavarman, I
Viravarman, Skandavarman II, Yuvamahdraja Visnugopavarman,
Simhavarman etc.) and hence were left unmolested by the latter
who were more powerful, lacks adequate proof. Similarity in the
names or the bull banner (adopted by both the dynasties) cannot
prove matrimonial or family ties. c Skanda ' enters into the com-
position of not only Vaingeyaka and Pallava, but also Satavahana
and Cutu names. Nandivarman, a name of frequent occurrence in
Vaingeyaka genealogy, occurs in Pallava genealogy only during and
after the sixth century A.D. (i.e., after the Vaingeyakas had passed
away). The bull crest of both dynasties has a sectarian, not
77. Since a part of the Vaihgeyaka kingdom, probably the whole of it
south of the Krna, would seem to come under Pallava sway during or before
the time of Simhavarman (middle of the fifth century), it is probable that
Skandavarman preceded Nandivarman II who has to be placed in the
last quarter of the 5th century AD.
180 EARLY HISTORY OF THE ANDHRA COUNTRY
dynastic, significance. 78 Since we do not know how Samudra-
gupta's invasion affected the Pallava, we cannot build much upon
the fact of the Vaingeyakas not being molested by the Pallavas.
Even supposing that the latter were as strong as ever, political and
geographical reasons might have prevented them from crossing the
Krsna to attack Vehgi. It has also been shown that during the
decline of Vaingeyaka power, the Pallava sliced off a part of the
kingdom i.e., the territory south of the Kr?na. This thrust must
have been one of the causes of its downfall.
Kalihga of the same period would seem to have been ruled by
powerful kings. The three kings who, on palaeographical grounds,
have been assigned to this period are Candavarman of the Komarti
plates, Nandaprabhafijanavarman of the Chicacole plates and
Maharaja Umavarman of the Brhatprostha grant. 79 One more
Kalingddhipati of the same period, who, however, would seem to
have belonged to another line, is Vasisthiputra Maharaja Sri-
Saktivarman. 80
Section II.
GOVERNMENT, RELIGION AND SOCIAL LIFE
Administrative Units
The land-grants incidentally throw some light on the admini-
strative system. The kingdom was divided into vi$ayas. Whilst
it is certain that the territory around the capital was not included
in the Kudrdhdra-vi$aya, we do not know how it was called. 81 It
has been shown above that ahdra, rattha and vi$aya denote the
same territorial division, not more in extent than a modern district.
Below the vi$aya was grama and below that pallikd or pallika-
grdma (a hamlet) . No division between vi$aya and grama is men-
tioned.
78. Vide supra.
79. El, Vol. XII, pp. 4 ff.
80. Ibid., pp. Iff.
81. The Ellore and Peddavegi grants in which the inhabitants of Eltir
(modern Ellore) and Pralura (Paloura, Ptolemy) respectively are addressed,
do not mention Kudrahara-vi$aya. The Kollair plates prove that the Kollair
region was included In the Kudrahara-vi?aya.
THE VAINGEYAKAS 181
Officers
Each 'viaya' was under an officer called Vi$ayapati. B2 The
Desadhipati of the Peddavegi and Kollair plates would seem to be
but the same officer under another name. True, under the Eastern
Calukyas, 'desa' denoted a kingdom of which 'vi$ayas' were divi-
sions. 83 However, in the Uruvupalli grant, Munclarastra is also
called a 'visayaS 8 * Like 'visaya' and 'rostra, 9 'desa' sometimes
denotes a kingdom and sometimes a province. 85 Cannot, therefore,
e vi$aya ' and ' desa ' in the same inscription denote the same terri-
torial division ? Since the Vaingeyaka kingdom would seem to
have comprised only two visayas, a higher division than a f viaya '
is impossible. Other officers (but subordinate to the Visayapati)
whose functions cannot be adequately defined are the Ayuktakas,
Niyuktas, Niyogas, Rajapurusas, and Vallabhos. Rajapurusas (lit.
royal agents) are probably the same as the Sasanasamcarins
of thfe early Pallava charters. 86 According to Hemacandra, 'Val-
labha' means 'Adhyaksa.' As, according to the lexicographer
Jatadhara, 'Vallabha' is a synonym of ' Asvaraksa srj and as 'Val-
labhas ' are distinguished from ' Go-vallab7ias ', ' Vallabha ' of our
records may mean either ' keepers of horses ' or * keepers of cows '.
Offices Borrowed
The writing of the Peddavegi plates as that of the Hira-Hada-
galli grant is attributed to a Rahasyadhikrta (confidential minister).
This title would seem to have been borrowed from the Pallavas as
also Vallabha. The practice of making bhojakas (freeholders)
executors of grants is also a Pallava practice.
82. C. P. No. 1 of 1924-25.
83. Their kingdom was called Vengidesa; and some of their grants (IA,
Vol. VIII, p. 76; Vol. XIII, pp. 213 ff.) record grants made in the Penna-
tava<Ji, Gudravara and Pagunavara visayas.
84. IA, Vol. V. pp. 50 ff., t. 11. 17 and 28.
85. CII, Vol. Ill, p. 31, t. 1. 4.
86. The Pulisas of Asoka's inscriptions (Pillar Edicts) are royal agents
whom all officers are asked to obey. Sometimes the word denotes ordinary
persons (subjects).
The Vallabhas are mentioned along with the Rajapurusas in our
inscription and with the Sdsanasamcarins in the Pikira, Marigalur and
Uruvupalli grants.
87. We owe this suggestion to D. C. Sircar, op. cit., p. 79, n.l. Dr.
Hultzsch translates Vallabha as 'favourite'.
RA. 24
182 EARLY HISTORY OF THE ANDHRA COUNTRY
' Mutuda' would seem to be the headman of a village. The
reading and meaning of the word are not settled. Burnell read
it as Munyada** Dr. Fleet was of opinion that for Munyada as a
common name no meaning could be found and that Munyada as a
proper name was out of place. 89 Reading it as Mutyada he cor-
rected it to Amatyadi. While editing the Ellore Prakrt grant of
Devavarman, Dr. Hultzsch read it as Muluda and remarked:
" The plates of Vijaya-Nandivarman seem to read, Munuda ;
but the apparent nu in the middle of this word may be in reality
the obliterated lu." But la has always a short curve attached to
right end of its back (the Nagarjunikonda inscriptions and the
Mangalur and Chikkulla plates) . 90 What is read as la has no such
curve. K. V. Lakshmana Rao reads it as Munuda? 1
The word which occurs in 4 out of the 5 grants can have only
one reading. In the Peddavegi, Kollair and Kanteru plates, the
reascent in the u sign in ' nu ' is to the left of the vertical, whilst in
the letter read as nu, the reascent is to the right ss in bhu or bhu.
The letter can only be tu. 92 That the last letter of the word is da
and not da is shown by No. 2 of 1924-25, where the back of da has
a notch at the right and whilst that of da lacks it. Mutuda being
the correct reading Dr. Hultzsch's reading in the Ellore grant must
be abandoned 7h favour of Mududa, which would be the Prakrt
form of Mutucte.
Though the Prakrt and Sanskrit forms are known, they cannot
be traced in dictionaries. Even so, the meaning of the word can
be ascertained. Munda and Mutalik in Hindi mean 'headman '. In
Telugu, Muttha denotes subdivisions of districts. Mutalpatta in
Malayalam denotes the office of the headman of the low caste term-
ed Chagon. In the Ederu plates of Vijayaditya II, (799-843 A.D.
according to Dr. Fleet's calculations) we have the expressions
'Kanderuv(a)di visaye va(nd)rupite(y)u-ndma gramasya Kutaka-
pramukhan Kutumbinas-sarvvan ittham djnapayati.' 93 These are
similar to 'Elure Mududa -pamukho gamo (gameyakd) bhdnitavvo*'
88. S. hid. Pal., p. 14.
89. IA, Vol. V, p. 176, n.
90. Ibid., pp. 159 ff., and PL; El, Vol. IV, pp. 193 ff.
91. JAHRS, Vol. V, pt. i, p. 31.
92. In the Kollair plates the in the syllable tu has no loop. But both
ta with the loop and ta without the loop are used (t. 11. 5, 10,).
93. El, Vol. V, pp. 118 ft.
THE VAINGEYAKAS 183
Dr. Fleet has shown 94 that l Kuta' has the meaning among others
* highest, the most excellent, first ' derived no doubt from its mean-
ings of any prominence, a peak or a summit of a mountain. In the
Bhdgavata Purana, * Kuta ' is used in the sense of chief. 95 Kutdka
is a variant of Kuta? G It is highly probable that the Mududas
(Sanskrit Mutuda) addressed in the same manner arc also heads
of villages.
The Citrarathasvami Cult
The tutelary deity of the Vaingeyakas was Citrarathasvami
('bhagavat Citraraihasvami padanudhyatah') , 97 Sanskrit Lexicons
give Citraratha as the name of the sun, the vahana of Agni and some
princes. 97a K. V. Lakshmana Rao thinks that Citrarathasvamin is
the Sun-God. While editing the Ellore plates of Devavarman, Dr.
Hultzsch referred to the existence of a mound " which, on a visit
to Pedda-Vegi in 1902, was shown to me by the villagers as the site
of the ancient temple of Citrarathasvamin, the family deity of the
Salankayana Maharajas." In the Khoh copper plate inscription
of Maharaja Sarvanalha, 98 a shrine of Aditya is spoken of. Ancient
temples of the sun exist at Asmarka, Gwalior, Deo-Baranark and
Tndore. In the Archaeological Survey of India" Cunningham has
noted an image of the sun at Shahpur, two feet and ten inches
high, holding a lotus in each hand, with, on each ^ide. a small
standing figure, that on the right being armed with a club. That
Sun-worship could have existed along with the worship of Siva or
Visnu is shown by Nirmand copper-plate grant of Maliawmantn
Maharaja Samudrasena. Here the divine Tripitrantaka is called
Mihiresvara. According to Dr. Fleet, the occurrence of the word
' Mihira ' (the Sun) as the first component of the God's nam^
seems to indicate that in this particular case, some form or other of
solar worship was combined with Saiva rites. It is, therefore, pos-
sible that tbp Va^nr^valTns some of whom had Saivn and others
94. El, Vol. VII, p. 221
95. 2, 9, 19.
96. IA, Vol. XL p. 112 and Vol. XVI. t>. 24.
In many Eastern Calukva grants we have 'visayp $an->ari(>va Rdstra-
hfitapramukhan Kutumbina ittham diiapayfitl' nnd 'rattraTtuta* is a technical
title meaning chief of a district.
97. That of the Eastern Gangas was Golcdrnasrftmin i.e., Siva.
97a. Peters, Diet. q. v.
98. CII, Vol. in. No. 28.
99. Vol. XV, p. 2, PI. xi. No. 1,
184 EARLY HISTORY OF THE ANDHRA COUNTRY
Valpnava leanings were at the same time worshippers of the Sun-
God also.
But on the analogy of the compounds Gokarnasvami, (i.e., the
God worshipped in Gokarna), Sriparvatasvami (i.e., the god wor-
shipped in Sriparvata in the Visnukurwjin inscriptions), cannot
Citmrathasvdmi mean the god worshipped in Citraratha ? We
do not know of any place called Citraratha. 100 Nor do the Pallava
records throw any light on this question.
Religion
The sectarian leanings of different kings are different ; some
were paramamahesvaras, others paramdbhdgavatas. It has been
shown that during the Satavahana rule Saivism was in a flourishing
condition. Some of the Iksvaku kings and their feudatories had
Saiva leanings. Jayavarman of the Brhatphalayana gotra was a
worshipper of Mahesvara. The earliest known king of the Vain-
geyaka line is a paramamdhesvara. Like the early Pallava kings
the later Vaihgeyakas are styled paramabhagavatas, i.e., worship-
pers of Visnu. Perhaps Pallava influence is to be seen in this
change. 101
100. Citraratha (M.BH. 6, 341; V. P. 184) is the name of a river.
101. Mr. K. V. Lakshmana Rao is of opinion that paramabhagavata need
not necessarily mean ' worshipper of Visnu'. He quotes Patanjali to show
that followers of iva were also called 'Siva bhagavatas'. We may add in some
inscriptions 'bhagavat' designates 6ambhu (GaoThwa stone inscription of
Candragupta II, CII, Vol. III). Paramabhagavata of the contemporary Pallava
inscriptions cannot mean 'worshipper of Bhagavat Siva' for at least one among
them would in that case have styled himself paramamahe&vara. What proves
conclusively that paramabhagavata of our inscriptions means 'a worshipper of
Visnu' is No. 38 in CII, Vol. Ill, in which Dharasena I and his younger brother
Dronasirhha are styled paramamahesvaras, while their younger brother
Dhruva II is called a paramabhagavata. The Peddavegi plates of Nandivar-
man II (a paramabhagavata) mention grant of lands to a Visnugrhasvamin,
i.e., the god inhabiting Visnugrha (Visnu).
CHAPTER X
THE KANDARAS
Di/nastic Appellation
Kings of this dynasty of whom three are known are said to have
belonged to the 'Ananda gotra.' 1 According to Dr. Hultzsch they
may be designated as 'kings of the family of the Ananda.' 2 Other
scholars have labelled them simply as "the Anandas'' or "the
Ananda kings of Guntur." We may once more emphasise that
confusion between gotra and dynastic names must be avoided. The
expression 'Kandara-nrpatikula-samudbhuta.,' 3 applied to Attivar-
man, a distant successor of Kandara, shows that the latter was the
first king of the dynasty who gave his name to it. 4 Hence these
kings had better be styled 'the Kandaras.'
Kandara : Meaning
Dr. Fleet considered 'Kandara' a variation of Krsna. 5 The
Prakrt forms of Krsna are Kanha (M; AMg; JM; S) and Kasana
(M and S) and Kasina (AMg; J.M) . The Dravidian form is Karma.
The variations of Krsna found in the Ratta and Yadava records are
Kanhara, Kanhara, Kandhara and Kandhara. 6 An argument
against Dr. Fleet's view is that Kandara nowhere occurs as a varia-
tion of Krsna. In a Rastrakuta record from Kyasanur in the Han-
gal Taluq of the Dharwar District, Lionel Barnett has read
'Kandara-vallabha' and equated it with 'Kannara-vallabha' of the
other Rastrakuta records, some of which are published in the same
volume. 7 In these records there is very little difference between
1. (a) Ananda-'kardnanda~maharsi-ma(hci)gotra^viyadamala'sakala--tu (hi)
nakiranasya .... Kandarajasya . . . . , ' 155 of 1899.
(b) A(na)nda-sa-gotrasya .... Damodaravarmano' El, Vol. XVII,
p. 329, t. 11. 2 and 3.
(c) 'Ananda-maharsivamSa-samudbhutena .... rdjnd AttivarmmanA',
1A, Vol. IX, pp. 102 and 103, t. 11. 1 and 5.
2. El, Vol. XVII, p. 328.
3. The Gorantla Plates, t. 11. 2 and 3.
4. Iksvaku, Pallava, and Gupta are some of the kings who gave their
names to their dynasties.
5. IA, Vol. IX, p. 103, n. 30 and BG, Vol. I, Part II, p. 334.
6. BG., ibid., p. 526.
7. El, Vol. XVI, p. 283,
186 EARLY HISTORY OF THE ANDHRA COUNTRY
nna and nda. Considering the fact that all the records including
the Kyasanur ones give 'Kannara-vallabha' only, what is read as
'Kandara-vallabha' is probably 'Kannara-vallabha.'
'Kanda' is a Prakrt and 'Kandan' a Dravidian form of Skanda. 8
'Kanda' might have become 'Kandara' much in the same way as
'Kanha' became 'Kanhara.' A fine Siva temple at Khajuraho
erected during the tenth century is called Kandariya Mahadeo.
Kings Kandara and Attivarman were worshippers of Siva. An
objection to this view is that in all the Satavahana, sarly Pallava
and Kandara records, the Prakrt form is Khrrhda.
In Sanskrit and Prakrt 'Kandara* means 'cave/ 'elephant goad.'
As some kings style themselves 'elephant goad to their enemies/
Sanskrit or Prakrt 'Kandara' as a name is not improbable.
King Kandara: His Dute
The founder of the dynasty was Kandara. Since the Allahabad
posthumous pillar inscription which speaks of Samudragupta's en-
counter with the Vairigeyaka Hastivarman and Kanceyaka Visnu-
gop?, does not mention king Kandara, whose kingdom lay between
the Vaingeyaka and Pallava kingdoms, it may be inferred that the
Kandaras rose to power after his southern campaign. Palaeo-
grrphy lends support to this view.
For the determination of Kandara's date, the palaeography of
the Chozarla inscription of his grandson supplies positive evidence.
On the back of the slab bearing this inscription there is record of
Mahendravikrama or Mahendravarman I. 9 To all seeming the
writing of the two inscriptions is almost contemporaneous. The
letters exhibit close agreement even in some details; but a careful
examination reveals certain developed forms in Mahendravikrama's
inscription. Only one pa and a few ha-s of the record of Kandara's
grandson reveal notches at the bottom while all the pa-s and 7ia-s
of the other have notches. The more or less angular la with the
notch at the bottom in Maharaja Mahendravikrama's inscription is
to be found only in a few places in the other; and in Chezarla we
8. Tarn. Lexicon, S. V. Kandan.
9. In the MattavtlAsa PraHasana, p. 3, the son of Sirhhavisnu (i.e., Mahen-
dravarman) is referred to as Mahendravikrama. Mahendravarman II men-
tioned only by the Kuram plates is not so known. The confusion between
the Pallava Mahendravikrama of the Chezarla inscription, and the Visnu-
kundin Vikramendravarman of the Ramatirtham plates which some scholars
have made must be avoided : the Mahendravikarama of the Chezarla inscrip-
tion belongs to the Bharadvaja gotra,
THE KANDARAS 187
find also the triangular va of the Nagarjunikon$a inscriptions and
the British Museum plates of Carudevi. We may then postulate
an interval of thirty to thirty-five years between the two inscrip-
tions. Since Mahendravikrama's reign would fall in the first de-
cades of the seventh century, the Chezarla inscription of king Kan-
dara's grandson may be assigned to the third quarter and king
Kandara to the first quarter of the sixth century. 10
In the Avantisundarikatha, of which we have, thanks to the
efforts of Ramakrsna Kavi, a fuller text than before, it is said that
the wife of Kandaresa (lord of Kandara) who was wounded in
battle (with the Pallavas evidently) sent a number of war ele-
phants to a Pallava King, there called a Magadha king.
Rise of the Dynasty
It is now possible to picture to ourselves the circumstances
favouring the rise of the dynasty. The Pallavas who were in posses-
sion of parts of the Guntur District in the reign of Sirhhavarman
in the early part of the fifth century were later on hard pressed by
the Kadambas in the south-west and Colas in the south. The bitter
rivalry between the Pallavas and the Kadambas under Mayura-
sarman continued long. Mrgesavarman, a successor of Kakusthavar-
man, was 'a destroying fire to the Pallavas,' 11 and Ravivarman up-
rooted Candadanda, lord of Kanci. The Anaji inscription 12 states
that Krsnavarman I's armies were totally defeated in a battle with
those of a Pallava king. From the fact that the early Pallava char-
ters are issued from Palakkada and Dasanapura, and from the fact
that Kanci is mentioned only in the Cendalur inscription of
Kumaravisnu, scholars like Venkayya have concluded a Cola occu-
pation of Tondaimandalam. 13 In the country just north of the
Krsna, the Vaingeyaka power had yielded or was yielding place to
the dynasty of the Visnukundins. Kandara took advantage of this
political turmoil to found a new dynasty. The Visnukuncjin and
the Pallava perhaps saw in the new kingdom a buffer state and so
connived at the coup. But the Kandara was too near the Pallava
to avoid conflicts later on. 14 One wonders whether the monkey
10. It is not possible to agree with the view expressed in the ARE,
1919-20, p. 95, that "by its early type of Pallava-Grantha character alone the
inscription must be referred to about the third century A.D." The Report
assigns Damodaravarman to the fourth century A.D.
11. IA, Vol. VI, p. 25.
12. EC, Vol. XI, Dg Ifil.
13. IA, Vol. XXXVII, p. 284 n.
14. Vide supra.
188 EARLY HISTORY OF THE ANDHRA COUNTRY
device on the banners of Kandara and of the Kadambas may be
taken to prove an alliance between them. 15
Kandara set about strengthening his dynasty by means of a
Pallava matrimonial alliance. The Chezarla inscription of Kan-
dara's grandson who bears the titles of 'Satsabhdmalla' and ' (Rana)
mahdmalla', 1 * introduces us to his father who bears the title of
'Prthiviyuvardja 111 and his mother, a daughter of king Kandara of
the Ananda gotra.
In the Chezarla inscription of his grandson, Kandara is said to
have been lord of two provinces or kingdoms and of the best city
Kandarapura. He is said to have engaged in sharp battles with the
Andhras near the banks of the river Krsna and brought about the
widowhood of many an Andhra woman. 18 His enemy was perhaps
the Visnukun^in. Like the Visnukundin Madhavavarman II, he
bears the title 'lord of the Trikutaparvata' Dr. Hultzsch's view that
Trikuta (Trirasmi of the Nasik inscription) in Bombay and Malaya
(in the western Ghats) , of which Madhavavarman claims to be lord,
were at a safe distance from his dominions, must, it would appear,
be abandoned. Any three-peaked hill might have been called
15. Rice, Mysore and Coorg frcmt Inscriptions, p. 26.
16. Appendix, 11. 30, 33-4.
17. I owe the suggestion that Yuvardjah in this inscription is the genitive
singular of Yuvardt to Professor K. A. Nilakanta Sastri. While editing the
Kopparam plates of Pulakesin II, Dr. Hultzsch wrongly considers Prthividuva-
rajah as in the nominative singular. As the sense of the passage requires a
genetive singular, he corrects it into Prthividuvardjasya.
The reasons for calling this a Pallava inscription are the following:
The birudas ' Satsabhdmalla ' and ' (Rana) mahdmalla ' are essentially
Pallava. Till the advent of the Eastern Calukyas, the office of Yuvardja or
Yuvamahdrdja was essentially a Pallava institution on this side of South India.
In a Dharmarajaratha inscription we have similar birudas e.g., Bhuvanabhd-
janah and Prthivlsara (El, Vol. X, pp. 5, 6; Nos. 1, 3, 7 etc.) . The practice of
mentioning the birudas to the exclusion of personal names would also seem to
be a Pallava practice. True Prthiviyuvardja and Prthivivallabha- Yuvardja
are titles borne by Visnuvardhana in the Kopparam and Satara grants
respectively (IA, Vol. XIX, p. 309). Prthividuvardja is according to
Dr. Hultzsch a form of Prthiviyuvardja (El, Vol. XVIII, p. 258). And
birudas into which 'Rana' enters are also borne by Kirttivarman I, and
Mangalesa (BG, Vol. I, ii, pp. 345-7). But the Chezarla inscription cannot
be brought to the beginning of the seventh century A.D.
18. Could this mean that king Kandara did not belong to the Andhra
tribe ? 'Prathitdn-Prthiviyuvardjah' applied to Kandara's son-in-law might
be corrected into ' Prathitdndhra-Prthiviyuvardjah '.
19. El, Vol. XVU, p. 388, t.1.5.
THE KANDARAS 189
Trikutaparvata. 20 But the association of Malaya with it forces us
to look for both in the west. A Vakataka inscription 200 gives us the
clue. Verse 18 in it records that Kuntala, Avanti, Kalinga, Kosala,
Triku^a, Lafa &&& Andhra were subjugated by one of the later
Vakatakas. At that time the Visnukiirwjins were an Andhra power.
Such defeats would have been repaid and these victories would
have provided the Visnukun^ins the occasion for taking over the
Vakafaka titles and glories.
Extent of his kingdom
The extent of Kandara's kingdom is not altogether easy of
determination. Dhanyakataka was in possession of his son-in-law,
a Pallava. Chezarla, where the inscription of Kandara's grandson
(by his daughter) and of Mahendravarman I are found, is in the
Narasaraopet taluq of the Guntur District. Damodaravarman's
grant comes from the Ongole taluk 21 and Attivarman's from the
Guntur taluq. The kingdom of Kandara would, therefore, seem to
have comprised at the most the Guntur, Tenali and Ongole taluqs
of the Guntur District.
Damodaravarman : His place in the Kandara genealogy
Damodaravarman came after king Kandara, for the Mattepa<J
grant of the former is issued from 'Kandarapura' and the Chezarla
inscription of Kandara's grandson, which refers to Kandara as 'the
lord of the best of cities, Kandarapura' makes it certain that Kan-
dara gave his name to the capital. 22 As Damodravarman's father
is credited with many Gosahasra and Hiranyagarbha mahadanas,
and as Kandara is not credited with them either in the inscription
of his grandson, where he is extolled in a lengthy passage, or in
the inscription of his distant successor, Attivarman, Damodara-
varman could not have been a son of king Kandara.
His date
Since the grant of Damodaravarman gives us neither dates nor
names and facts tending to establish definite synchronisms with
20. CII, Vol. HI, p. 9 n, (Introduction).
20a. ASWI, Vol. IV, pp. 124 ff.
21. Kamgura mentioned in the grant, which Dr. Hultzsch was unable to
identify, may be identified with Kanduluru, six miles to the east of Maftepatf,
the findspot of the grant.
22. Other instances of kings giving their names to their capitals are,
'Pravarapura' from which Pravarasena ITs edict is addressed, Kandharapura,
the city of Kr?na-Kandhara, etc. (CII, Vol. in, p. 236 and BG, I Pt. ii,
p. 556).
H.A.-25
lac EARLY HISTORY OF THE ANDHRA COUNTRY
kings whose dates are known, palaeography is our only guide in
the determination of his date. Dr. Hultzsch has only remarked
that the alphabet is of an early southern type. 23 We can say that
it resembles the alphabets of the later Vaingeyaka grants and of
the Pikira and Marigajur grants of Siihhavarman. Since Pikira is
said to have been included in the Mun^arastra which has to be
sought for in the Nellore District 24 and since Mattepad, the findspot
of Damodaravarman's grant, is in the Ongole taluq, once part of
the Nellore District, a comparison between the alphabet of the
Pikira and Mattepad grants is more apt. Striking is the resembl-
ance between the la, va, ya, ba, na, da, and tu of the two grants.
The close resemblance between the ja, bha, dbha and li of the
Mattepad and Mangalur grants is also noteworthy. The thick dots
or nail heads at the top of letters, especially over ga, a feature
common to the two Pallava grants under reference, are to be found
in the first few lines of the Mattepad grant. The Upadhmdniya
occurs in all the grants. 25 It may be argued that the Mattepad sa and
sa differ from the Pikira ones as they have bars instead of curves.
But the former are not quite unknown to the Pallava grants of the
same period. 26 The Mattepa$ numerical symbols for 1 and 2 are
more primitive than those in the two Pallava grants, but as two
symbols for one are used in the Mattepatf grant, and as the Manga-
lur symbol for four is slightly different from the Pikira one, much
reliance cannot be placed on numerical symbols; nor can archaic
forms be the main factor in the determination of dates. Saka 380 of
the colophon of the Jaina work Lokavibhaga would seem to corres-
pond to the twenty-second regnal year of Siihhavarman, king of
Kaiici, mentioned in the text. According to Krishna Sastri the
astronomical details given in the Omgotfu grant of Simhavarman
II, 27 do not fit into this date, and Siihhavarman of the Jaina work
must be Simhavarman I; and the Pikira grant belongs to Simha-
varman II. On these pieces of evidence, Damodaravarman would
23. El, Vol. XVII, p. 327.
24. Munolarastra is identical with Mundainadu of some Nellore inscrip-
tions (N. 19, 31 and 121) and Kandukura which, according to the Uruvupalli
grant, was included in the Mundarastra is perhaps identical with Kandukur
(IA, Vol. XXXVII, p. 283, n) .
25. Dr. Hultzsch has not noted the fact that in the Maftepaql grant as in
the Pikira and Mangalur grants, final m is once represented by small m (El,
Vol. XVII, PI. v, t. 1. 15).
26. Mangalur grant, IA, Vol. V, pp. 154 ff, PI. v-b, 1. 3; vi-a; 11. 1 and 3.
27. El, Vol. XV, pp. 246 ff.
THE KANDABAS 191
have to be assigned to the first quarter of the sixth century A.D.
But as he was not the son and immediate successor of Kandara, he
cannot be placed earlier than 550 A.D.
Dr. Hultzsch assigns Damodaravarman to a period earlier than
that of Attivarman partly on the strength of the Prakrt portion of
the former's grant. 28 D. C. Sircar rejoins, Attivarman is a Dravi-
dian form of Hastivarman through the Prakrt form Hatthivar-
man. 29 These arguments would compel us to place Damodara-
varman's inscription sometime before the Pallava Sanskrit charters
of the fifth and sixth centuries A.D., which, for reasons stated
above, would be impossible. Instances of Prakrt forms lingering
in the records of the fifth and sixth centuries A.D., can be cited. 30
The Buddhist leanings of the king can also explain this mixture of
Sanskrit and Prakrt.
Attivarman
According to D. C. Sircar, Attivarman 31 is the father of Damo-
daravarman. 32 The only source of this assertion is the expres-
sion 'Aneka-go-sahasra-aneka-hiranyagarbhodbhavodbhavasya **&
applied to the latter in his grant. He has rightly pointed out that
28. "It may be inferred from the Prakrt forms used that it could not be
far away in point of time from the Prakrt plates of the early Pallava kings
of the third or fourth centuries of the Christian era", ARE, 1919-20, p. 75.
El, Vol. XVII, p. 328.
29. Journal of the Department of Letters, Calcutta, Vol. XXVI, p. 48.
30. According to D. C. Sircar, Dr. Hultzsch's statement that the inscrip-
tions is partly in Prakrt is a 'misrepresentation.' But Dr. Hultzsch correctly
states the position when he says that the personal and most of the gotra names
of donees are in Prakrt and that the Prakrt ' samvaccharam f occurs in the
Sanskrit portion while the Sanskrit amko occurs in the Prakrt portion. For
Prakrt forms lingering in Sanskrit records see Visnukun^in records generally.
31. Dr. Hultzsch looks upon Attivarman as a Prakrt or Dravidian form of
Hastivarman. Since in Prakrt, the initial ha is nowhere softened into a, Atti-
varman cannot be another form of Hatthivarman. Neither in Telugu nor in
Canarese can 'Atti' meaning 'elephant' be found. That 'Atti' is sometimes a
Dravidian form of 'Hasti' is made certain by the Sholinghar inscription of
Parantaka (El, Vol. IV, pp. 221 ff.) where the Western Ganga Prthivipati II is
called Hastimalla (the name of the elephant of Indra) in the Sanskrit portion,
and Attimallan in the Tamil portion. Even now 'Atti' in Tamil means 'ele-
phant*. For names and surnames into the composition of which 'Atti' enters,
see also BG. I, ii, p. 507; SB, Vol. HI, No. 60, p. 121; Vol. I No. 74, p. 105;
El, Vol. XVI, p. 17; El, Vol. VH, pp. 195, 196.
32. Op. cit., p. 50.
33. T. 11. 2, 3.
192 EARLY HISTORY OP THE ANDHRA COUNTRY
Dr. Hultzsch's translation of this as well as the expressions 'Apra-
meya-Hiranyagarbhaprasavena'M t 'Hiranyagarbhaprasuta' 35 , and
'Hiranyagarbhasambhuta' 3 * as 'producer of many Hiranyagarbhas*
is defective, as the past participles prasuta and sambhuta can enter
into Pancami Tatpurusa and not Sasthl Tatpurusa compounds. He
has correctly translated them as 'born of the Hiranyagarbha. 9 The
theory underlying the performance of the Hiranyagarbha maha-
ddna is that the performer is born of the golden womb, etc perhaps
in the same way as Brahma. 37
But where he brings this translation to bear upon the problem
of Kandara genealogy one cannot agree with him. According to
him it is not 'quite impossible' that Attivaraman who is credited
with the performance of the Hiranyagarbha in the Gorantla plates
is the father of Damodravarman who is said to have been 'born of
one born of the Hiranyagarbha and a performer of the Gosahasra.'
He dismisses the palaeographical difficulty cheaply when he makes
the gratuitous assumption that the difference in time between the
Mattepad and Gorantla plates is so short, and that it is difficult to
determine which of them is the earlier.
(a) The grantha alphabet of the Gorantla plates is decidedly
later than that of the Mattepad grant, because in thorn the ra, ka
34. The Gorantla inscription IA, Vol. IX, p. 102, t. 1. 3.
35. Ipur and Polamaru grants El, Vol. XVII, PP. 334 ff, and JAHRS,
Vol. VI, Part i, pp. 17 ff.
36. The Mahakuta pillar inscription IA, Vol. XIX, pp. 7 ff.
37. The Hiranyagarbha jar is of the shape of a lotus ; it is provided with
a pedestal, golden stalk of lotus and a golden thread round its navel.
During the performance of the mahddana the performer enters the pre-
cints of the altar where the Hiranyagarbha is placed, holding the images of
Dharmaraja and Brahma in both hands, and takes five deep breaths after
placing his head between his ankles. The Brahmans versed in the Vedas per-
form the Garbhadana, Pumsavana and Simanta ceremonies of the Hiranya-
garbha and later on the rites following the birth of a child. Finally the per-
former addresses Vi$nu thus:
"O, the best among gods, before I came out of my mother's womb,
Now that I am born of your womb, mine is a divyadeha."
Moreover, udbhava (adj.) which occurs in the expression ' Aneka-
Hiranyagarbhodbhavodbhavasya* is mainly used in the sense of 'birth' or 'com-
ing from' e.g. Rastrakutavam6odbhava, Sailodbhava, etc. As has been point-
ed out by Sircar, 'Hiranyagarbhodbhavodbhavasya' in the sense of 'performer
of Hiranyagarbha ' is awkward in a prose composition and the Buddhist Damo-
daravarman would not have performed Brahmanical rites like '
and 'Hiranyagarbha,'
THE KANDARAS 19S
and the subscribed u consist of two vertical lines of nearly equal
length; 38 in the PIkira, Mangalur and Uruvapalli grants and even
in the Chezarla inscription of Mahendravarman I, it is considera-
bly shorter. In the Kailasanatha inscriptions the reascent is to the
full length of the vertical.
(b) Ja with the upper arm converted into a loop is on the line
of development to the Kuram ja which exhibits the head of the
modern Tamil ja. The Chezarla inscription has the archaic ja.
(c) The cursive ha and pa with the notch at the bottom occur
in the Dharmarajaratha inscription. The Chezarla ha and pa with
modest notches are earlier.
(d) The Gorantla a with a hook at the end of the lower arm
is like the Kailasanatha a and is more developed than that of
Mahendravarman's inscriptions. But the treatment of the vertical
in Kailasanatha a stamps it as later than the Gorantla . 39
(e) The hook at the right end of the body curve of the
Gorantla da is absent in Dharmarajaratha, but present in the
KaiUsandtha da. The latter is, however, more developed than the
Gorantla one.
(f) bha with two equal curves resembles the Dharmarajaratha
bha.
(g) The Kailasanatha na is more cursive and developed than
the Gorantla na.
(h) The Gorantla na is more cursive than the Dha^mardja-
ratha na. The development of na is from Siihhavarman's inscrip-
tion where it has a long stem to Dharmarajaratha and Chezarla
forms and then on to that of the Gorantla inscription where it
is very short.
According to Biihler, 40 the Gorantla alphabet went out of use
before the date of the Kuram plates of Narasirhhavarma's grand-
38. The reascent in r 1. 3 (paricarena) and in 11 in the syllable ju 1. 6, is
to the full length of the vertical.
Dr. Fleet compared the alphabet of the grant with that of the frag-
mentary inscription at Badami incised, according to his researches, by Pallava
Narasimhavarman I during his expedition against the Calukya Pulakesin II.
However, only the notched pa, ha, ma, la and va exhibit any resemblance to
the letters of the Gorantla plates.
39. Buhler's Tables, VII, Vol. XXIII, and IA, Vol. IX, pp. 102 ff., Plate 1;
t. 11. 1, 2, etc.
40. Ind. Pal IA, Vol. XXXHI, App. p. 70,
194 EARLY HISTORY OF THE ANDHRA COUNTRY
son ParamesVaravarman, which exhibit letters of a much different
type. It is thus seen that in many respects, the Gorantla alphabet
stands between the Chezarla and DHarmarajaratha writing on the
one hand and the Kaildsanaiha writing on the other. Dr. Hultzsch
has pointed out that the majority of the epigraphs on the monolith
now styled Dharmardjaratha resemble those of Mahendravadi and
Siyamangalam, assigned by him to Mahendravarman I 41 (first
quarter of the seventh century) . The earliest date for Attivarman
would, therefore, be the second quarter of the seventh century.
Sircar's assumption that only one prince in the line could have
performed the Hiranyagarbha mahddc.na enjoined by the Sacred
Books on kings and wealthy men of the realm, is a gratuitous one.
In the Srisailam plates of Virupaksa, Harihara, son of Bukka, is
said to have performed the 16 mahadanas , 42 In the Udayam-
bakam grant of Krsnadeva Raya 43 (. 1450) both Krsnadeva Raya
and his father are credited with the sodasa mdhdddnas. In the
Vellarigudi plates of Venkatapati Deva Maharaya I (S. 1520) ,**
both Vira-Bhupati and his son Krsna-Bhupati are credited with the
same. Nor can the numbering of both sides of the Mattepad plates
attest the posteriority of Damodaravarman to Attivarman. The
plates of Devavarman and Nandivarman are numbered in the
Mattepad fashion, whilst those of Nandivarman II and Skandavar-
man are numbered by plates. Yet the latter came after the former.
It is thus seen that Attivarman's reign falls in the second and
third quarters of the seventh century. Damodaravarman who on
palaeographical grounds has to be ascribed to the middle of the
sixth century could not have been a successor of Attivarman.
His Grant
Attivarman's Gorantla plates, which bear no date, record the
rjrant of the village of Antukkura (according to Dr. Hultzsch pro-
bably Gani Atukuru to the west of Bezwada) and the grant of land
called * Astasatapatti' 45 in the village of Tanthikontha or Tanri-
konra (which according to Dr. Hultzsch is the modern Tadikornja,
ten miles to the north of Gunttir and to the south of the Krsna as
41. El, Vol. IV, p. 152, and Vol. VI. p. 320.
42. El, Vol. XV, p. 24.
43. El, Vol. XIV, pp. 171-72.
44. El, Vol. XVI. p. 298 ff.
45. ' T&nthikontha (or Tanrikonra) gr&me catttrddiiam-astaiatctpatff
ksetran'; IA, Vol. IX, p. 103. It may mean 108 paffia.
THE KANDARAS i&
stated in the inscription itself) 46 to a Brahman named Ko^tisarman
who was versed in Rg, Yajus and Sdma Vedas and the Apastamba
Sutra.
Most of the epithets applied to Attivarman in his grant would
seem to be conventional. Such are, 'anuparata-dharmma-kriyd-
para', 'sura-guru sadrsa-buddhih ', 'sundara-sujdta-pesala-jana
paricdra. 9 But epithets like ' Hiranyagarbbtvprasava, 9 ' pratdpo-
panata-sakala-sdmanta-mandala ', and ' samyak-prajd-palanopdrj-
jitakirtti ' would seem to have facts behind them.
Attivarman's status
Whilst Damodaravarman and the Pallava kings of the period
bear the tittle of Maharaja, Attivarman is styled Rdjan. Could this
fact indicate a feudatory position ? We know that a part of the
Guntur district was under Pallava rule in the time of Kandara and
Attivarman. But the expression pratdpopanata-sakala'Sdmanta-
mandala applied to Attivarman, and the fact that Mahendravar-
man I is styled rdjan in some inscriptions 47 and also the absence of
any evidence of Pallava rule over the Guntur taluq, 48 make it highly
probable that Attivarman was an independent monarch. 49
Religion and Social Life
The available grants of the Kandaras do not give us any
direct insight into the administrative organisation. It may be sur-
46. Dr. Fleet re^td it as Tanthikontha and Dr. Hultzsch's reading is not
as certain as it has been assumed. In the Ipur and Chikkulla plates ra has
two cross strokes. In the Peddavegi plates of the Eastern Calukya Jaya-
siihha, I, (El, Vol. XIX, p. 261) and the Kon^anaguru plates of Indravarman,
(El, Vol. XVIII, pp. 1. ff.) letters similar to that under reference occur. In
the former it differs from the next letter xa in so far as it has not the latter's
notches. The word has been read as 'Kantheruvati' which has been identified
with the Kamjeru of the Eastern Calukya inscriptions. In the latter grant,
the subscript ret has assumed its modern form, and is different from the letter
above, though Dr. Hultzsch read them as rra- No doubt in all these grants the
subscript tha is open to the right and tha has a dot in the middle. Even so,
in a grant of the Eastern Calukya Visnuvardhana IV or V (IA, Vol. XIII,
p. 186) and perhaps in the Peddavegi plates of Jayasimha I, tha with the short
stroke occurs.
47. El, Vol. VI, p. 320, t. 11. 1 and 2.
48. Chezarla where the inscription of Mahendravarman exists is in the
Narasaraopet taluq of the Guntur District.
49. Attivarman claims to be as powerful as Mahendra (Mahendrasama-
vikramena). Is there a veiled reference to Mahendravikrama or Mahendra-
varman I (El, Vol. IV, p. 153)?
196 EARLY HISTORY OF THE ANDHRA COUNTED
mised that the traditions of the past were continued. Pallava
influence might have played its part. However, we know some-
thing about the religion of these kings and their subjects. Unlike
the Pallavas with whom they were matrimonially connected and
who were 'paramabhagavatas,' the Kandara kings, with the excep-
tion of Damodaravarman, were proteges of Sambhu Siva. 50 Their
family deity was thejBod in the Vamkesvara temple. 51 Thus the
Kandaras continued the Saiva traditions of the Iksvakus, the kings
of the Brhatphalayana gotra and some of the Vaingeyakas. If the
names of donees can be relied upon as an index to the state of pre-
vailing religions, Saivism would seem to have had a strong hold
over the Kandara kingdom. 52
But Buddhism was too firmly rooted in the land of the Bhatti-
prolu, Amaravati and Nagarjunikon^a and Ghantasala stupas for
monasteries to be completely eradicated even by four centuries of
Brahmanical rule. Damodaravarman was a Buddhist (Samyak-
sambuddhasya-padaniidhydtah) , and the Prakrt of his grant might
be an evidence of his Hinayanist leanings. In his day the tolera-
tion which the Brahmanical Satavahana and Iksvaku kings had
shown towards Buddhism was returned by Buddhism to Brahma-
nism. Like Brahmanical kings, Damodaravarman makes grants
of lands to Brahmans.
Scattered epigraphic and literary evidence attests the lingering
of Buddhism in the Andhradesa for a long time after it ceased to be
the dominant religion. Amaravati has given us not oii]y Buddhist
inscriptions of the third and fourth centuries A.D. 53 but also an
inscription in early Telugu characters from which we learn that a
Siihhavarman, son of Nandivarman, on his return from an expedi-
tion, came to Dhanyakataka and became a lay worshipper of the
Buddha and made donations. 54 An inscription on the base of a
pillar bearing the sculpture of a standing figure of the Buddha, in
50. ambhos~carana-kamala-rajah-pavitrlkrte, IA, Vol. IX, the Gorantla
plates.
51. The anusvdra over va is clear. Dr. Fleet was not sure of it (IA, Vol.
IX, p. 103, n.) . Only, as in other places, in the same inscription (11. 4, 6, etc.)
it is not exactly over the letter, but slightly to the right of it.
52. Names like 'Ruddajja', 'Nandijja', 'Khandajja', 'Kumarajja' and
' Devajja. '
53. ASSI, Vol. I, p. 106, No. 51 and PI. LXI, No. 4; for an unpublished
Amaravati inscription of the third century A.D., see PI. IIL
54. SII, Vol. I. No. 32,
THE KANDARAS 197
m the Nandigama taluq, Krsna District, mentions
the setting up of an image of the Buddha by the sramanaka Rahuja,
the disciple of Acdrya (Ma)deva who was again the disciple of
Acdrya Maudgalyayana . 55 According to Hiuen-Tsang, though
Buddhism was on the wane, there were in the Vengidesa twenty
and odd monasteries with three thousand brethren. 56 In the king-
dom of T'e-na-ka-che-ka (Dhanyakataka) there was a crowd of
Buddhist monasteries but most of them deserted, about twenty
being in use with one thousand brethren mostly adherents of the
Mahasanghika system. 57 The play entitled * Mattavildsa Praha-
sana' shows that Buddhist bhikkhus and their vihdras and caityas
existed in Kaiici in the seventh century, whatever might have been
the demoralisation that had set in.
About the social life of the period we do not know anything
except that the Brahmanical civilisation had once more gained
ground. A point deserving mention here is the 'arya' ending in
the names of the donees of the Mattepa$ grant. Such name end-
ngs are found only in grants coming from the territory immediately
south of the Krsna (the Kondamudi, the Mayidavolu, the Hira-
HatfagaJJi, the Kanteru (Nandivarman I) and the Mattepa^ grants.
' Arya ' (venerable) as an honorific prefix to the names of Bud-
dhist or Jain teachers and saints occurs hi inscriptions all over
India. It is used as an honorific title in the Hathlgumpha inscrip-
tion of Kharavela. 58 ' Arya ' as initial part of personal names
occurs in a Junnar inscription 59 (Ayama) and in a Nagarjunikon^a
inscription 60 (Ayako^usiri and Ayasiri, names of royal ladies) . The
celebrated disciple of Nagarjuna, who spent a great part of his life
in the Andhradesa, is Aryadeva. 61 But the earliest inscription to
exhibit names with 'arya' ending is the Kono!amudi grant of Jaya-
varman, where all donees have names ending in ' aja '. The same
55. 218 of 1926-27.
56. Walters 'On Yuan Ckwang\ ii. p. 210.
57. Ibid., pp. 214-215.
58. ' Aira maharaja Kharavela '; an Aya-SakasathI is mentioned in a
Bhat^iprolu inscription as the name of a group (gosjhi) .
59. ASWI, Vol. IV, p. 103, No. 11.
60. Ins. L. El, Vol. XXI.
61. The epigraph on a relic pot from the Guntur District, which has been
ably read and interpreted by Prof. K. A. Nilakanta Sastri (JOR, IX-13, 96, ff.)
states that it contains the ashes of Ayamani identified by the Professor with
Aryadeva.
HA. 26
198 EARLY HISTORY OF THE ANDHRA COUNTRY
is the case with the Mayidavolu and Maftepa^ grants. 62 'Ajja' is
another form of Prakrt 'ayya' (Sanskrit arya). 63 'Arya' started
as an honorific prefix and became a name ending much in the same
way as 'siri.' 64
APPENDIX B
155 of 1899
1. Sri siddhlsvaraya
2. Kalissara saravlraketoh sva sakti satita
3. satrava kari vara ghata samkata Dhanyakata rana
4. saptakoteh prathitan(dhra*)prthiviyuvarajah a-
5. nanda karananda maharsi mahagotra viyadamala-
6. sakala tu (hi) na kiranasya samavaga^handhrasundari
7. candananjanalakta lokamala parimosana pa-
8. ricaya krtaparadha vai(pu)lya sitetarabe (rtna)
9. nathasya Trikutaparvvata pate (r*) golangula vija-
10. yaketanasya hallisaka patu pataha rava s(a)la (?)
11. ma (tya or sya) nadyogasya Kandara-puravara-janapada-
dvitaya-
12. dhipateh Kandararajasya priyasutayamavanitala-
13. ntava(t)yam mahadevyam sanjatas-satguros-tasyaspa
(n)da-
14. madhi (sthe) ya gunaniadhitisthannalamkurva(nah)
sujatah
15. sujanamanah sravanasukhacarita-bhajano janodita
16. sakalayasovisesah rosasthira vrtta-janu-karah
17. Karibhala visasta satrava sarira sakalamalaman4a-
18. na bhavyamana pareta kantaguno gunonnata virodhi
19. parajan (au) gha nivarano varanottamabalo balonnatasya-
20 (ma) kusumake(tu)rabalajana manah pramathana-
patura-
21. . .gu. . . .rabhilasita phalapradanakrtamati-rahimagi-
22. ririva. . vilanghaniya mahima-gurus-salila nidhira
62. Ajja of the HIra-Hatfagalli and Mattepa^ grants is the etymological
spelling of the Pandit; 'aja' is the cave inscription form of 'ajja.'.
63. Pischel. op. cit, Secc. 236 and 252.
64. It is used as a prefix in ' Raya Simuka Satavahana sirimato ' (Nane-
ghat relievo figure inscription) and as a part of names in 'Vedisiri' 'Hakusiri'
'Balasir!' (Nasik No. 2) 'Bodhisiri/ 'Bapisiri' etc. (Amaravati and Nagarjuni-
inscriptions).
THE KANDARAS 199
23. gadha gambhrrya dhairya&laghi saganida-muraripu-sana-
thasasanah
24. paramasamrddha pallijanapadadhipati-raribala-jayamatih
25. sama (ra*) samaya samunnata grddhradhyasita-ketanah
pratinr-
26. pa vanitasrutipatha-Brotralbharaiia gauravavai phalya. .
27. . .pasamu. .inu. . .ta (ra) varinasaghoanassamarapatu-
28. patahara(va*) ravana dviguna maghaguna guna-
sagaraska-
29 (vijaya) sajayanuraktanuraga gururabhistuta
30. gunavasassakalakala visarada sa (t) sabhamaUah
31. (svaku)la salila nidhi samabhivarddhana tuhinakaranara
tilaka
32. ... dustanigrahogravisaghatah sistaparipalana
33. . . .prasanna nrpavara gunatulya kantih sriman sri (ra-
34. na) maha (mallah) sakta. . .tva (to) nvayasya prthujaya-
bala-
35. yurarogyabhilasitartha siddhaye punaratratirtha-
36. . . .rvvannata samamahimanasya . . .bhila
37. . . .bahuni bahu devatayatanani sthira
38. ... ardhatsarvaparihara sampa . . . grama
39. . . .maya. .ra. . .vrsabhagana
40 saklasadhananam godha
41 saha
42 parvatlpara. . . .ru
43 manekavi. . . .nu
44 rajitarajata (kapalapali)
45 ka . . . ghantakamsatala chatra
46 nise. . .hatabha. . .devopakara
47 sampradaya manvantaramabhiramabhi
48 bhagavan paramesvarastu . . .
49 tayapratyadhika (ra
50. . . . prasannasobitam
CHAPTER XI
THE VISNUKUNDINS
The Visnukundin is the last of the many minor dynasties that
ruled over the destinies of the Andhra Country since the disap-
pearance of the Satavahana. Whilst writers have emphasised the
fact that it was contemporary with the Pallavas in the south, the
Kalinga dynasties of the fifth and the sixth centuries A.D., the
Vakatakas, the Maukharis and the Guptas, they do not say that
immediately south of the Visnukundin kingdom lay the Kandara
kingdom. The chronological scheme adopted here makes the
Kandaras the contemporaries of the Visnukundins, and the Visnu-
kuridm-Vakataka matrimonial alliance was perhaps a political one
intended as an offset to Kandara-Pallava matrimonial alliance. The
fortunes of the Visnukundins were bound up with the fortunes of
the dynasties of the north and their sphere of action lay in the north,
especially, in Kalinga. It is with the Maukharis, the Vakatakas and
kin.rrs of Kalinga, that the Visnukundins waged a series of wars that
finally left them exhausted before the flood of Calukyrm invasion
came upon them.
Of this dynasty we have five records. That this dynasty rose
on the ruins of the Vaingeyaka dynasty is made clear by the fact
that the Chikkulla plates were issued from Lendulura, modern
Dendaluru. Peddavegi, the site of the capital of the Vaingeyakas,
is only five miles to the north of Lendulura. And in both
places there are ruins of a large of Siva temples and extensive
mounds. True some records were issued from other places, Ipur II
from (Ama)rapura, Ipur I from vijayaskandMvdra Kudava<Ja.
But as both Kudavada and Amarapura are called vdsaka, the for-
mer also was a temporary residence.
The Ramatirtham plates were issued from Puranisangama, a
vdsaka. Some scholars think that the Visnukundins were a
northern dynasty. They connect Trivara from which Madhava-
varman I took his wife with Tewar in the Central provinces and
Trikuta with Trirasmi and Malaya with the Western Ghats. They
also point to the northern wars of the Visnukundins as furnish-
ing evidence in the same direction. Amarapura cannot be identi-
fied with distant Amroati in the Central provinces. Nor can it be
identified with -Amaravati as even as late as the sixth century the
THE V1SNUKUNDINS 201
latter is known as Dhanyakataka Chezarla Inscription. No
one has pitched upon Pallava wars with the Western
Calukyas and Iksvaku matrimonial connections with the
house of Ujjain as evidence of their northern origin.
Kielhorn looks upon Vinukonda as the survival of the
dynastic name 1 and this looks quite probable . Visnukundm >
Visnukondin>Vinukondin> Vinukonda. Kundi really means 'one
who comes from the fire-pit.' But the fact that their family deity
was the Sriparvatasvami, may go some way to prove their southern
or Andhra origin. But even this piece of evidence should not be
pressed too far. Visnukundin used in the plural is, like Vakataka
end Iksvaku, a family name. Some writers look upon it as a gotra
name. They say that the Vakatakas with whom the VisTmkundins
had matrimonial connections belonged to the Visnuvrddha gotra,
Visrm also is a recognised gotra. But they admit thai the Visnu-
kundin is not to be traced in the gotra and pravara kdndas. More-
over we have the plural and not the singular.
It is, however, possible to build some history on the name. The
kin^s of the line though devoted to the God at Srlparvata (Malli-
ksarjuna) would, as their family name indicates, have been wor-
shippers of Visnu, and the names borne by these kings
also lends support to this view. Not one king in
the line among those known till now bears a Saivite
name. One wonders if the Sriparvatasvami of the Visnukundins
was a Vaisnavaite deity, otherwise as yet unknown, of the rtpar-
vata of the Nag&rjunikonda inscriptions. 2 Current local tradition,
it may be noted, associates the hill with the Rama cycle of stories.
Madhavavarman and Govindavarman, names which occur in the
lists, the latter twice, are indications of the Visnu worship in the
family, bef 
