Dr. B. R. AMBEDKAR
)PEN UNIVERSIT
(J fj-
AUTHOR : Jclfc V6*U-
TITLE :
Date Due
Borrower's Name
Dr. B. R. AMBEDKAR
OPEN UNIVERSITY
LIBRARY
HYDERABAD-500033
Dr. B. R. AHBEDKAR OPEN UNIVERSITY
LI B RA R Y
Call No. -fa ( Accession No. W - <$ ^. I/
Author: ^fcU l/y~u. -
Title :
P^
This book should be returned on or before the date
last marked below.
N08631
THE
PALLAVAS
BY
G. JOUVBAU-DUBBEDIL
Doctor of the University of Paris,
Professor, College, Pondicherry.
TRANSLATED FROM FRENCH
BY
V. S. SWAMINADHA DI^SHITAR, B.A., L.T.,
Officier d'Acadtmie
Professor of English, Colonial College, Pondicherry.
PONDICHERRY :
SOLD BY THE AUTHOR,
6, DUMAS STREET.
1917
CHAPTER I:
THE ROMAN ORIGIN OF PALLAVA ART.
I do not know if any author has so far attempted to
ascertain what kind of art flourished at the time of the first
kings of the Pallava dynasty. That is what we shall try to
determine in this chapter.
There is a document which is particularly important for a
study of the early history of the Pallavas; it is the Viripara
plates discovered at Mayidavdlu, a village situated at a distance
of 12 miles from Narasar&opet in the Gunttir district which
lies on the southern bank of the Krishna.
From his capital, K&nchfpura, and in the 10th year of the
reign of his father whose name is not given, the heir-apparent
(Yuva-Mah&rftja) Sivaskandavarman, of the Pallava dynasty
and of Bhfiradvfija Gdtra, sent to the governors of Dhafifiakada
an order concerning the village of Viripara, situated in the
province of Andhrfipatha (Ep. Ind. 9 Vol. VI, p. 84.)
There is no doubt that this prince belongs to the dynasty
whose history we are studying, for it is expressly stated that
6ivaskandavarman lived at K&ncht, and was of the family of
the Pallavas and of Bh&radv&ja Gotra.
Again, as the plates were discovered in the Gunttir District
and the village for which they were engraved was in. Andhrft-
patha, and as the order was issued to the officers of Dhannakada,
that is to say, the town of Amarftvatf, it is certain that the
Pallava -princes of this period reigned not only over Tondai
mandalam, that is to say, the province of KMchf, but also over
the country up to the banks of the Krishna which was occupied
by the Andhras and in which was situated the town of Amar&vati
where, in the middle of the 2nd century, King Pujumfiyi II. built
the white marble Stfipa, the sculptures of which, almost entirely
Roman in workmanship, now adorn the Madras Museum.
The Pallavas have thus succeeded the Andhras on the banks
of the Kjishijft; but how long after and at what epoch?
We shall now show, that, in all probability, the father of
Sivaskandavarman of the Pallava dynasty reigned at Amarftvatf
shortly after Pujum&yi II. had built the famous Sttipa.
The gift of the village of Viripara bears only the date of
the year of the king's reign ; however, the alphabet in which
it is written would enable us to determine the date, if we have
any points of comparison.
Fortunately, in this case, we have such points of comparison :
The Mayidavolu plates are written in the same alphabet as
the plates found at Kondamftdi (a village situated in the
Tenali T&luk, Guntfir District). And in its language and its
phraseology the Kondamtidi document resembles so much the
K&rlS inscription of "Gautamiputra 6&takarni, and the N^sik
inscription of V&sishthiputra PulumS,yi, that there cannot be
any great difference of date between them :
The alphabet of this inscription shows that he must have
lived in the same period as the Pallava prince fivaskandavarman
who issued the Mayidavolu plates. Further, the language and
a phraseology of the inscription are so similar to the N&sik
inscriptions [a still closer resemblance exists between Jaya-
varman's plates and the K&rl inscription No. 19 (A. 8. W. I.
Vol. IV, p. 112)] of Gautamiputra Satakarni (Nos. 4 and 5)
[ Dr. Bhagwanlal Indraji's in Bombay Gazetteer Vol. XVI, and
Inscriptions from the cave temples of Western India Kdrld ] of
V&sishthiputra PulumSyi (No. 3), that Jayavarman's date cannot
a have been very distant from that of 'those two Andhra Kings. D
(Ep. Ind. Vol. VI, No. 31, p. 315).
If we bear in mind that it was V&6isthiputra Pujum&yi who
built the Stfipa at Amarftvati, we may conclude that the Pallava
king, father of 6ivaskandavarman, who engraved the Mayidavdlu
plates, reigned at Amar^vatl shortly after Pujumftyi II.
A strange coincidence indeed : the son of this Pujum&yi
reigned from 177 to 184 A.D., under the name of 6ivaskanda which
was also the name of the son of the Pallava king.
Is this coincidence something purely accidental ? We may
believe it is not. In fact it is possible that the Pallava king
had married the daughter of Sivaskanda 6&takarni and that the
" Yuvamahar&ja " of the Pallava dynasty received, according to
the custom of the Hindus, the name of the Andhra king who was
his grandfather.
Since we are now dealing with suppositions, we shall say a
few words here about the theory of the Persian origin of the
Pallavas.
The father of Pujumayi II. had fought with the Pahlavas, and
this same Pulumiiyi had to struggle against the satrap Rudra-
daman, who, about the year 150 A.D. had for minister a Pahlava
called Suvififtkha. (Ind. Ant., VII, 257 Junagadh Insc.) Relying
upon the analogy of names, certain authors think that these
Pahlavas, who fought with the Andhras, succeeded in establishing
a kingdom for themselves between the KrishnS, and the Plar and
founded the Pallava dynasty.
This theory is subject to variations in is details. For
example, we may suppose that the Pahlavas took possession of
one part of the Andhra empire not by conquest but by marriages
such as the one that is supposed to have taken place between
the father of Sivaskandavarman and the daughter of the Andhra
King of the same name.
Similar marriages between the Andhras and those of other
dynasties have surely taken place ; we know that Pu}umiiyi IL
married the daughter of the satrap Rudradaman, whose minister
was a Pahlava.
All these hypotheses are based on the similarity of names :
Yuvamah&r&ja Sivaskandavarman and Sivaskanda Satakarni,
Pahlavas and Pallavas ; but they are contestable.
All that we can consider as certaian is that the Pallava
King who reigned at Amarvati when the Mayidavdlu plates
were engraved lived shortly after Pujum&yi II. who built the
Sttipa at Amar^vati ; but it is not possible to say exactly how
long after.
The Andhra dynasty came to an end about 236 A.D, There is
nothing, however, to prove that the Pallavas did not succeed the
Andhjas many years earlier at AmarSvati, and that the prince
Sivaskandavarman was not the contemporary of the last Andhra
kings whose kingdom had now become much smaller. In that
case, the Mayidav61u plates may be dated about 200 A.D. If, how-
ever, the Pallavas succeeded the Andhras after 236 A.D. it must
have been immediately after, for the resemblance between the
Mayidav&lu plates and Jayavarman's plates discovered at Konda-
mfidi is so close that we cannot suppose that more than a century
would have elapsed between Sivaskandavarman and Pujumiiyi II.
We can therefore conclude that the Paliava kings reigned at
Amarftvatt in the first half of the third century after Christ, that
is, about 50 years after the famous Stftpa was built. These kings
reigned over a very extensive territory : their empire extended
from the banks of the P&l&r to those of the Kristin k They were
therefore powerful and glorious.
What about the art in this kingdom? There is no doubt
about the answer. Paliava art at the time of fiivaskandavarman
cannot be very different from that which flourished at the time
when Pujum&yi II built the Stflpa at Amarftvatt
A visit to the remains that are kept in the Madras Museum is
enough to convince one that this art had.attained great perfection.
The subjects are Buddhistic, the costumes and the ornaments
are Hindu, but their workmanship is European.
But, above all, it is in the representation of the human body
that the European influence manifests itself. The hair is curled
in the Greek manner, the face is symmetrical, the limbs are
sculptured according to the rules of Anatomy with conspicuous
muscles, and some of them are dressed in clothes that remind us
of the Roman toga.
This Indo-European art was not peculiar only to the banks of
the Krishna ; it was found throughout India and specially in
Gandh&ra.
At this epoch, the Roman Empire extended its influence over
almost the whole of the civilised world and was also connected
with India in various ways.
How long did this influence last ? There is no doubt that it
disappeared from South India in the VII century, whereas it was
probably not introduced there before the Christan era. It attained
its zenith probably in the II and the III centuries. There is no
doubt that in its origin the Paliava art was strongly influenced by
the principles of the Latin Art.
In many places on the banks of the Krishna, we find sculp-
tured marbles of which the subjects are Buddhistic and the
workmanship Roman. These are the bas-reliefs that once adorned
the Stftpas or the mutilated images of Buddha. Up to the present
they have always been attributed to the Andhras. This view,
I am sure, is not always correct.
The Graeco-Buddhistic art did not certainly disappear with
the Andhra dynasty. It is not likely that the technical methods
Plate
Roman head of Buddha
(Discovered by the Author near Bezwada.)
11
employed by the sculptors who built the Stflpa at Amar&vatl,
disappeared completely in the space of a few years. It is almost
certain that this art that flourished in the middle of the II century
lived on for more than one century and that it was only very
slowly that the sculptors abandoned the ancient models and .
forgot the noble methods that they had been taught by the artists
that had come from Rome. To be more precise, I believe that the
Indo-Roman art attained its zenith in the II century with the
sttipa of Amarftvati; that throughout the III century, the art
used on the banks of the Krishna was entirely Indo-Roman ;
that in the IV century there were still very evident traces of this
influence and it was only in the V century that all vestiges of
Latin influence disappeared completely.
Indeed, many of the marbles that have been discovered on the
banks of the Krishna do not date from the time of the Andhras,
but have been sculptured when the Pallava kings reigned over
this country*
On the 1st January 1917, I went to Bezwada with the object
of visiting the caves of Undavalli. When going about those regions,
I luckily discovered the ruins of a Buddhistic temple at Vijiader-
puram, a village half a mile to the west of Bezwada. Amongst
the remains of the brick walls, there were two heads of Buddha
and a trunk dressed in the Roman toga. One of these heads was
very beautiful. I bought it from the owner, a temple priest, for
two rupees and have it now in my possession.
Plate I represents it. Were it not for the elongated ears and
the sign of Buddha on the forehead, one would take it for the face of
a Roman of the earliest times. The hair is curly, the eyes have no
pupil, the general aspect is Roman, and this image is made of
white marble like the antique models of Europe. I think that
this work done in India by an Indian is strong proof of the great
influence that Latin Art exercised in the early centuries on the
banks of the Krishna.
I shall not certainly affirm that it was the work of a Pallava
sculptor, but we are not sure either that this statue dates from
the time of the Andhras. I believe that even if the sculptors
of the time of Sivaskandavarraan have not made this head, they
had at least the habit of making similar ones.
On my return to Madras, I informed the archaeological
Pepartment of the existence of the Buddhistic remains that I
-12-
had discovered at Bezwada. The other head and the trunk have
since been brought over by the Government to the Madras
Museum.
Was this Indo- Roman art confined to the banks of the
Krishna, or did it extend through the whole Pallava empire ?
The latter is very probable ; the Buddhists at K&iichtpuram built,
in their capital, Sttipas that were perhaps in the same style as
those in the north of the Empire. The only researches, so far
attempted, into the Buddhistic remains at K&fichi have been made
by M. R. Ry. T. A. Gopinatha Rao when on a short visit to the
town, and his rapid investigation has been very fruitful in as
much as he has discovered, in the last pr&k&ra of the K&ra&kshi-
d6vi temple, a Roman statue of Buddha which is shown in Fig. I
of the article " Bauddha vestiges in KMchipura " by T. A. Gopi-
natha Rao (Ind. AnL, Vol. XLIV, Part DLVII, June 1915). If we
compare the head of this statue with that of the one I have found
at Bezwada, and the trunk with the one since brought over to the
Madras Museum, we shall find that the resemblance between them
is complete.
We can therefore conclude by saying, that, at K&nchipuram
as well as at Amar^vati and Befcwada, the Pallava Art, inspired
by Roman models, attained great perfection at the time of those
early Pallava kings.
CHAPTER II.
THE VA.YALfiR INSCRIPTION.
The record found at Mayidav61u is written in Pr&krit.
There are also two others of the same kind; but the other
Pallava records are in Sanskrit ; the former are surely the oldest
and can be considered to belong to the III century.
One of these two records (Ep. Ind. Vol. I, p. 2), that of
Hirahadagalli, is dated in the 8th year of the reign of Sivaskanda-
varman, King of Kfi-iichi, who is of the Pallava dynasty and
Bharadv&ja g6tra and who by this document confirms a gift
made by his father Bappa-deva. We may suppose that this
6ivaskanda was the person who was " Yuvamah&r&ja " when
the Mayidavdlu gift was made, but we are not quite sure of it.
This record is specially important as it shows the extent
of the Pallava empire. These plates (discovered in the Bellary
District) mention the province of S&t&hani (S&thani-rattha) as
forming part of the Pallava kingdom. This province comprised
a portion of the Bellary District. Thus, we know that the
Pallavas reigned over an empire, which, having K&fichipuram
for its capital, extended not only along the Coromandel coast
up to the mouth of the Krishna, but also to the West, in the
Deccan, up to the banks of the Tungabhadra river.
Here also the Pallavas had succeeded the Andhras. The
inscription discovered at My&kadoni, in the Bellary district,
(G.O. ifo. 99, 29th August 1916. Report on Epigraphy for 1915-
1916) says that King Pulumfiyi II. reigned about 140 A.D. over
the province of SStavahani-hfira which is none other than Sfttft-
hani-rattha of the Hirahadagalli copper-plates.
We see then that, at first, the Pallava empire extended more
in the Deccan than in the Tamil country ; there is nothing to
prove that it comprised the banks of the KftvSri ; the Trichi-
nopoly region was probably occupied by the Chdjas; on the
contrary, we are sure that the Pallavas reigned over the country
extending from Bellary to Bezwada, that is to say, over au
important portion of the Andhra empire. These geographical
considerations lead to the probability of the theory that the
Pallavas were a Northern dynasty, who, having contracted
marriages with the princesses of the Andhra dynasty, inherited
a portion of the Southern part of the Andhra empire.
The other record was discovered in the District of Guiitflr
(Ep. Ind., Vol. VIII, p. 143). In the reign of Vijayaskanda-
varman, Ch&rud^vi, wife of " Yuvamah&r&ja M Vijaya-Buddha-
varman, a Pallava prince of the Bh&radv&ja g6tra, and mother
of BuddhyaAkura, made a gift to the temple of N&rftyana at
Dftlflra.
The alphabet of these plates resembles that of the Hira-
hadagalli plates. However, as the name of the king is not
exactly the same, we cannot, with certainty, identify Vijaya-
skanda with Sivaskanda.
We shall conclude by saying that the three Pr&krit records
(those of Mayidavolu and Hirahadagalli, and the grant made
by Ch&rudvi) found in the districts of Guntfir and Bellary, prove
that many princes of the Pallava dynasty of K&nchfpuram
(1) the father of Yuvamah&r&ja 6ivaskandavarman [called Bappa-
deva in the Hirahadagalli plates], (2) Yuvamahftrfija Siva-
skandavarman, (3) Mah^r&ja Sivaskandavarman, (4) Mah&r&ja
Vijayaskandavarman, (5) YuvamaMrSja Vijaya Buddhavarman,
(6) His son Buddhyankura (2 and 3 being perhaps one and
the same person) reigned towards the III century, not only over
Tondai Mandaiam, but also over the lands bordering the Krishna
and Tungabhadra, and so over a great part of the Telugu
country.
We know from the inscription on the pillar at All&hftbftd
that about 340 A.D. the great emperor Samudragupta vanquished
Vish^ugdpa, King of K&fichl.
Certain authors have supposed : (a) that Vishnugdpa was
a Pallava ; (b) that Samudragupta advanced as far as K&achl
(Conjeeveram) in the Tamil country.
I think that the last conclusion is not correct. I believe
that Samudragupta never entered the Tamil country but that
Vishnugdpa was, in fact, a Pallava king. We know that the
kings of KMchi reigned on the banks of the Krishnft; it is,
therefore, very likely that events happened as mentioned below t
Samudragupta came from the North of India and vanquished
the kings who reigned at Pithfipuram, Mahtodragiri, and Kothfoa.
When they saw him advance to the South of the Gddftvari,
Mantaraja who reigned near the Colair lake, his neighbour, the
King of Vengt, Vishiiugdpa, King of Kftnchl who reigned on the
right bank of the Krishn^ and his neighbour UgrasSna, King
of Pftlakka, formed a coalition to stop the invader. ' But he
affirms that he was the victor ; and until the contrary is proved
we have to believe in his words.
Be that as it may, it is probable that Samudragupta did not
care to advance more to the South in a country which must have
been difficult of access, and so returned to the North.
Let us now examine the Sanskrit records.
Five copper plates :
1 6mg6du No. 1 ( (?. 0. No. 99, 29 Aug. 1616, Part II, No. 3),
2 Uruvupalli (Ind. Ant. Vol. V, p. 50),
3 6mg6du No. 2 ( G. 0. No. 99, 29 Aug. 1916, Part II, No. 4),
4 Pikira (Ep. Ind., Vol. VIII, p. 159),
5 MangalAr (Ind. Ant., Vol. V, p. 154),
enable us to establish with certainty the following genealogy :
Kum&ravishnu
Skandavarman (I)
Viravarman
Skandavarman (II)
Yuvamah&r&ja Vishnug&pa
Simhavarman.
These six princes belonged to the Pallava dynasty and Bh&rad-
vftja g6tra. There is nothing, however, to prove that their capital
was K&nchipuram. It is also probable that, had K&nchf been
their capital, they would have dated their grants from there.
But Skandavarman II. was encamped at Tftmbrfipa when he made
the 6ibg6du No. 1 grant ; Yuvamah&r&ja Vishnugdpa was at Palak-
kada when he made the Uruvupalli grant; Simhavarman was
at Da&anapuram when he made the M&ngaitir grant and at Men-
raatura at the time of the Pikira grant. It is therefore probable
that these three sovereigns never reigned at Kanchtpuram. The
geographical position of these towns from which the several
grants have been made is not known ; but we know accurately a
region which formed one of the provinces of their kingdom. The
two. 6mgddu records mention " Karmm&-r&shtra " as forming
part of their empire, and this district is often mentioned in the
grants made by the Eastern Ch&lukyas. We know also that the
village of ChendaUir (Ep. Ind., Vol. VIII, p. 233) which is only
another name for Chandalftru, a village in the Ongole taluq of the
Gunttir district, was situated in this province. Besides, it is also
probable that the village of Omgddu was adjacent to Santarfivtlru
where the plates were discovered. Santarvflru is in the Bftpatla
taluq of the Guntfir district.
There is therefore no doubt that the environs of the present
town of Ongole watered by the river Gundlakamma was named
Karmm&-r&shtra at the time of the ancient Pallavas. There is no
other information about these kings except their genealogy.
The word " Yuvamah^r^ja " preceding the name of Vishnu-
g6pa shows that he never reigned. He has however made
a grant to the village of Uruvupalli in the llth year of the reign
of Sirtihavarman. So, Vishnugdpa was probably the brother of a
king named Simhavarman.
The grant of the village of Chftra ( G . 0. No. 920, 4th Aug.
1914, Part II, No. 1) enables us to suppose that the donor Vijaya-
Vishnug6pavarman was the son of Simhavarman and grandson of
Yuvamah&r&ja Vishnugdpa, though he is called in the Chfira
plates " Maharaja " and not " Yuvamah&raja ".
An isolated copper plate relating to a grant made at
DaSanapuram by the great-grandson of King ri-Vira-K6rcha-
varman, has been discovered at Darsi (Nellfir District) [Ep. Ind.,
Vol. I, p. 397], Be that as it may, we must bear in mind the
incontestable fact that a dynasty of six or seven Pallava kings
reigned over the Telugu country about the V century of the
Christian era.
But what connection had they with the kings of K&fichi-
puram ? we do not know it exactly.
We may suppose that the Ch6las occupied KMchf for the
time being and about the V century drove back the Pallavas to
the north of the kingdom.
The existence of a dynasty of Chdja princes in the Telugu
country seems to confirm this presumption (G. 0. No. 518, 18th
July 1905, Part II, No. 5).
18
Written in characters which are not very ancient ; there is tto
means of knowing if it is a forgery or a copy of a genuine
document.
The Jain work called " Lokavibhaga," discovered by M.E.Ry.
R. Narasi&hachar, is dated . 380 (458 A.D.), the 22nd year of
the reign of the Pallava king Siihhavarman ; this evidence is
rather suspicious, and besides it does not say which Siihhavarman
it was and to what branch of the Pallava family he belonged.
The Penugoiida plates (G.O. No. 920, 4th Aug. 1914. Part II.,
No. 4 and J. R. A. S. Oct. 1915) mention two Pallava kings
Simhavarman and Skandavarman, but the age of these plates is
not known.
The Pallavas themselves have attempted to answer the
important question who were the ancestors of the Pallavas of the
Simhavishnu dynasty. The Kasakfidi plates give a few of the
names of kings (S.LI., Vol. II, Part III. p. 356).
The VSlflrpalaiyam plates (&/./., Vol. II, Part V, p. 510) give
us more information ; but, unfortunately, they do not give us a
complete list. On the contrary, this author warns us that the
kings he mentions, are only a few among the numerous prede-
cessors of Simhavishnu. These isolated names do not enable us to
build up a genealogy.
Moreover this record which is dated in the IX century, not
being relatively ancient, we do not know how far we can rely on it.
The V&yalfir inscription presents much greater guarantee as
it is the most ancient genealogy of this kind. Whereas the
Kasaktidi and Vfilfirpajaiyam plates are dated during the time of
the princes of the dynasty of Nandivarman, the Vay&lftr inscrip-
tion is of the time of a prince of the dynasty of Siriihavishim.
Besides, this inscription gives the names of such a large
number of kings that it seems to be a complete list of them in the
order of their succession.
It is for these two reasons that the VSyaltir inscription
becomes extremely interesting.
The Vftyalflr inscription (No. 368 of 1908) is very much
damaged: The report on Epigraphy for 1908-09 (G.O. No. 536,
28th July 1909 Part II, No. 17, p. 77) speaks of it only in a very
summary manner. The order of succession of the predecessors of
Silfahavishnu is not given in the report, but it is precisely this
order that it would be interesting to know.
The report says :
The pillar in the Vyaghrapurfsvara temple is a very
a interesting one on account of the genealogical record which
y> is engraved on it. The latter begins with the usual mythical
a names Brahma, Angiras, Brihaspati, Samyu, Bh&radv&ja, Dr6na,
ASvatthaman, and Pallava. Then, the quasi-historical names
A66ka, Harifcupta, Aryavarman and others are mentioned.
After these, the names of nearly twenty-five kings are registered
whose relations to one another are not specified. The order in
which the known names are mentioned at the end suggests,
however, that these twenty-five are also to be taken in the order
of descent Among these occur the following which are already
familiar to us from copper-plate inscriptions : (1) MahSndra-
a varman (once), (2) Karanda (Kalindavarman ?) (once), (3) Vishnu-
J> gopa (thrice), (4) Kum^ravishnu (twice), (5) Buddhavarman
i> (twice), (6) Skandavarraan (five times), (?) Simhavarman (four
J> times), (8) Viravarman (once), and (9) Nandivarman (once). In
many places the record is damaged and the names are not
* legible. If, however, the whole of the inscription is made out, it
will perhaps establish a connection more definite than that
hitherto set forth by the copper-plates between the Pallavas of
the Pr&kjit records, those of the Sanskrit records, and those of
the Simhavishiiu line. The Amar&vati pillar epigraph (which is
evidently a copy of some older record) gives a list of early
Pallava Kings some of whom also occur in the V&yalfir inscrip-
tion. After mentioning the twenty-five names above referred to,
the latter introduces Simhavishnu for the first time. From him
were descended apparently in the order of father and son,
MahSndravarman I, Narasimhavarman I, Mah&idravarman II,
a and Param66varavarman I. His son was RSjasiihha ' the (very)
king* of lions on the high mountain (viz.) the prosperous Pallava
family 1 .
Seeing the importance of the Vyaliir inscription for a history
of the Pallavas, I went over there with the object of studying it
myself on the spot.
It is engraved on a cubical pillar of the Pallava style and
runs round it in the form of a helix. It begins with the well-
known series of names : Brahma, Angiras, Brihaspati, &amyu,
Bharadv&ja, Drdija, A6v$.th&raan, Pallava, A$6ka, Harigupta
-20-
Aryavarman, and then two or three names hardly legible, and
then Kajinda, Byftmalla, [E]kamalla.
After this last name begins a series of 36 names. The estam-
page of this part of the inscription is given in Plate II.
1. Vimala
2. Konkanika
3. KSjabhartri
4. Chfttapallava
5. Viraktircha
6. Chandravarman
7. Karaja
8. Vishnug6pa
9. Skandamtila
10. Kanag6pa
11. Viraktircha
.12. Skandavarman
13. Kum^ravishnu
14. Buddhavarman
15. Skandavarman
16. Kumaravishnu
17. Buddhavarman
18. Skandavarman
19. Vishnugdpa
20. Vishnud^sa
21. Skandavarman
22. Simhavarman
23. Viravarman
24. Skandavarman
25. Simhavarman
26. Skandavarman
27. Nandivarman (I)
28. Simhavarman
29. Simhavarman
30. Visnugdpa
31. Simhavarman
32. Simhavishnu
33. MahSndravarman (I)
34. Narasimhavarman (I)
35. Mahndravarman (II)
36. Paramsvaravarman (I).
A fact of very great importance is that in the legendary
series of ancestors of the kings belonging to the dynasty of
Plate. 11.
Simhavishnu there are personages called Aryavarman and Kon-
kanika, names that are surely of Western Ganga origin.
The Penugonda plates (G. 0. No. 920, 4th Aug. 1914, Part II,
No. 4, p. 83) which are the only authentic records that we have
about the ancient W. Gangas give the dynasty :
Konkanivarman
M^dhava
Aryavarman
I
Sirhhavarman.
The presence of names of Ganga origin in the legendary
portion of the list of Pallavas goes to confirm the fact that those
kings had political relations with the Pallavas as is shown by
the Penugonda plates : (Aryavarman and Siihhavarman were
crowned by Pallava kings).
Let us now compare the Vayalflr inscription with the K; sa-
kfidi plates.
The series of names is the same up to Pallava.
Between Pallava and Simhavishnu the list of names given in
the Kasaktidi plates is rather short. However, as at Vayalir,
we read : Asoka, Kanag6pa, Skandavarman, Vishnugopa, Vira-
kftrcha, Simhavarman. Kalindavarman of Kas&ktidi is probably
none other than Kalinda of Vayaltir.
If we compare the V^yaltir inscription with that of Amara-
vati (8. 1. 71, Vol. I, p. 25) which, we know, is not an ancient one,
we find little resemblance between them.
On the contrary, the Vayalfir inscription exactly coincides
with that of Veltirpajaiyam. The V^yaltir list is complete ; but in
the Vayalflr plates, we have many u et caetera" in place of names.
'After a few names which are all of them found in the Vaya-
Iftr list Asdka, KSijabhartri, Chfitapallava, the coincidence be-
comes complete :
Velurp&laiyam. \ Vayalur.
Vfraktircha (11) VirakHrcha
Skandasishya (12) Skandavarman
Kum&ravishnu (13) Kum&ravishnu
Buddhavarman (14) Buddhavarman
The Velurpajaiyam plates flay that (11) VtrakOrcha was the first
who " grasped the complete insignia of royalty ", that i$ to say,
that his predecessors were not kings and that Vlrakflrcha was
the first Pallava king.
In fact, with these four kings we enter into the domain of
history leaving the series of legendary names behind.
The existence of a king called Virakdrcha is proved by the
plate discovered at Dar& (Ep. Ind., Vol. I, p. 397). The existence
of a king of the name of Skanda­a is established by the
Tirukkalukkujiram inscription. [Ep. Ind., VoLJLII, p. 277].
This king is identified by the Vfiyalflr inscription with
(No. 12) Skandavarman. This identification admits of no doubt
since in both the inscriptions he is mentioned as the son of
Virakfiroha, and father of Kum&ravish^u and grand-father of
Buddhavarman.
It is to be noted that the name Skandavarman given in the
Vyaltir inscription to the second Pallava king{Vlrakfircha being
the first) is also the most ancient name known to history. In
fact, the donor of the Mayidavdlu plates is called " Yuvamah&-
raja " diva Skandavarman ; the donor of the Hirahadagalli plates
goes under the name of Siva-Skandavarman ; and the names
Skandavarman and Buddhavarman figure in the grant of Chftru-
devi that has been found in the Guntfir district (Ep. Ind.,
Vol. VIII, p. 143).
The list :
Skandavarman
1
Kumaravishnu
Buddhavarman
*
is the same as the one found in the genealogy given in the
Chendalftr plates (Ep. Ind., Vol. VIII, No. 23, p. 233) which we
have supposed to be a copy of an ancient record ; so we have to
note th&t the V616rp&laiyam and the V&yalftr records place them
among the earliest kings.
Let us now conclude by saying that with (11) Virakflrcha we
deal no more with the legendary series of descendants of Pallava
but with the historic line of kings* *
What makes the coincidence existing between the
and Vdltirpftjaiyara records interesting is that the latter gives an
important information :
<r (V. 6)... Vtraktircha, of celebrated name, who simultaneous-
ly with (the hand of) the daughter of the chief of serpents
j> grasped also the complete insignia of royalty and became
j> famous.*
We have already concluded from this passage that Vira-
kflrcha was the first Pallava king.
Here we have to make a remark : the VSlflrp^laiyam plates
together with the VSyalfir inscription lead us to think that, in the
Pallava family, there existed the following tradition which was
probably perpetuated from century to century. "The earliest
Pallavas were not kings, and they were alien to South India.
One of them married the daughter of one of the kings of that
country and thus became a king himself. Skandavarman was a
son born of this marriage."
Is it not strange that this tradition coincides exactly with the
theory that we have propounded in the preceding Chapter ?
A Pahlava married the daughter of the Andhra King iva-
Skanda and thus became the first king of the Pallava dynasty.
The son born of this union was jiva-Skandavarman.
We should not rely too much on the order of succession of
the kings given in the ViyaWr inscription after (11) Virakfircha.
It has to be noted, however, that among the early kings we find
(19) Vishnugdpa who must be identified with Vishnug6pa of
Kaiichi who was the adversary of Samudragupta about the year
339 A.D.
But, from which king does the order of succession given in
the VftyaWr inscription become trustworthy ? I believe it is from
(23)' Viravarman. In fact, if we compare the ViyaWr series with
the genealogies given in the undermentioned plates :
6rbg6du No. 1 [O., 1]
Uruvupalli [Ur.]
6mg6du No. 2 [O., 2]
Pikfra [P.]
Chftra [C.]
Udayendiram [TTd.]
J64
we obtain the following table in which the numbers indicate those
oftheVayaiarlist:
dates:
A.D.
(23) Vfravarman 422
[O., 1J [UrJ [0., 2] [PJ [MJ
(24) Skandavarman 450
[0., 1J [UdJ [Ur.J [O., 2J [PJ [M.J [CJ
(25) Simhavarman Yuvamah&r&ja Vishnugdpa... 478
[UdJ [UrJ [UrJ [O., 2J [P.J [M.J [CJ
(26) Skandavarman (29) Simhavarman 506
[UdJ [0., 2] [P.] [MJ [CJ
(27) Nandivarman. (30) Vishnugdpa 534
'[CJ
(28) Simhavarman (31) Simhavarman.. 562
(32) Simhavishnu 590
I think [that the Vayaltir record is extremely important on
account of the series :
(24) Skandavarman
(25) Simhavarman
(26) Skandavarman
(27) Nandivarman
which exactly coincides with the genealogy given in the Uda-
yendiram plates (Ep. Ind., Vol. Ill, p. 142). We ha^e already
said, that, the alphabet of these plates not being ancient, we
cannot say how far they can be trusted. The deciphering of the
V&yalftr inscription has shown that this genealogy is quite right
and that the Udayendiram plates are a true copy of an earlier
record.
1 ho succession list :
(29) Simhavarman
(30) Vishnugdpa
(31) Simhavarman
(32) Siiiihavishnu
unites the dynasty of Sithhavishnu with that of Vishnug&pa of
the Chfira plates, (G. 0. No. 920, 4th Aug. 1914, Part. II.J No. 1).
The Uruvupalli plates prove that (25) Simhavarnian was the
brother of Yuvamahar^ja Vishnug6pa and that both of them
were the sons of (24) Skandavarman and grandsons of (23)
Vtravarman.
It is probable that (24) Skandavarman did not reign at
Kanchi, since the 6mgodu No. 1 plates are dated from Tftmbrftpa
camp.
It is also probable that his son (25) Simhavarman resumed
possession of Kanchi and confided to his brother Yuvamah&rja
Vishnug6pa the government of the Northern provinces (Guntftr
and Nellore Districts) in which were situated Palakkada, Dasana-
pura and Mnmtura as well as the districts of Vengor&shtra,
Mundar^shtra and Karmarashtra. Vishnugdpa was never
crowned and always remained a subordinate of his brother ; that
is why he is called Yuvamahrja.
His son (29) Simhavarman became independent.
So there were two dynasties simultaneously : in fact, when
(30) Vishnugopa. reigned at Palakkada (Chfira plates), (27) Nandi-
varman reigned at Kanchi.
It is probable that (31) Simhavarman or (32) Simhavishnu
took possession of K&iichi.
Simhavishnu is then a descendant of the dynasty that reigned
at Palakkada. That is why the Vayaltir inscription gives first
the names of the kings of the dynasty of (27) Nandivarman of
Kanchi, and then of those of the dynasty of Palakkada.
Yuvamah&raja Vishnugopa is not mentioned in the Vayalfir
^list as he was never crowned king.
In the preceding table, we have given in the margin the
approximate dates of the predecessors of Sirahavishnu. Starting
from Simhavishnu who lived at the end of the VI century, about
590 A.D., we have given to each generation an average duration of
28 years.
This calculation seems to be correct, for we find, that, accord-
ing to it, (25) Simhavarman must have lived in 470 A.D. And if
we can depend on what is said in the " Lokavibh&ga " discovered
by M. R. Ry. R. Narasimhachar, there seems to have been a king
of that name in S, 380, ?>. 458 A.D.
-26 -
Dr. Fleet has assigned the date of about 500 A.D., to the Penu-
goiida plates, which is the date we give to (26) Skandavarman,
son of (25) Siihhavarman who crowned king M&dhava II alias
Si&havannan.
The name Simhavarman given to a king of the western GaAga
dynasty shows that Aryavarman who had been crowned by the
Pallava King (25) Simhavarman had married his daughter and his
son Madhava II. received the name of his grandfather, the Pallava
king Siihhavarman.
The Vayalfir inscription enables us to believe that the
Penugonda plates belong to about 500 A.D.
In the Vftyaltir list, the predecessor of Simhavishnu is called
Simhavarman, and this name is engraved very clearly. This is a
remarkable fact, since it confirms what is stated in verse 10 of the
VgltirpSjaiyam plates.
Then from the king named Simhavarman, who wiped off
the pride of (his) enemies, was born the victorious Simhavishnu
J> whose prowess was widely known on earth.
If we admit : 1 the identity of (19) Vishnugdpa with the
adversary of Samudragupta in A. D. 338 and, 2 the identity of
(21) Skandavarman with Skandavarman (I), of the 6mg6du No. 1
plates, we obtain the following chronology :
rVishnugdpa A.D. 338
Vayalflr -j Vishnud&sa (Kum&ravishnu) 366
^ Skandavarman (I) 394
6mg6du -I Viravarman 422
No - L ^Skandavarman (II) 450
Chflra | (Y * M-) Vish * u S 6 P a 478
j Simhavarman 506
,Vishnug6pa .*. 534
Simhavarman 562
V&yal6r -| Simhavishnu 590
Mah^ndravarman I 618
- NarasiAhavarman I 646
It is not necessary, in conclusion, to say what a flood of light
the deciphering of the V&yaltir inscription has thrown on the
history of the early Pallavas : the importance of the inscription is
naturally very great as it gives us the most ancient and complete
list of kings.
CHAPTER III.
TELUGTJ ORIGIN OF THE MAH^NDRAVARMAN STYLE.
The deciphering of the VSyalfir inscription has led us to think
that the genealogy of the predecessors of Simhavishnu stood thus :
(29) SiAihavarman
(30) Vishnug&pa
(31) Sirfihavarman
(32) Simhavishnu.
It is certain that the father of Simhavishnu was (31) Sitfiha-
vartnan who seems to have been so named after his grandfather,
(29) Simhavarman. Again we have admitted the identification of
(30) Vishnugdpa with the one who granted the Chftra plates, from
Palakkada. It is also highly probable that (29) Simhavarman, the
donor of the No. 2 6mg6du, Pikfra and M&nga|flr plates was not
king of Kanchi but reigned in the districts of Nellore and Gunttir.
From this we have to conclude that the direct ancestors of
Sirfihavishnu and Mahfindravarman lived perhaps in the Telugu
country. We shall now proceed to show the importance of this
detail.
When, last year, I wrote the first volume of " Pal lav a
Antiquities", I was struck by the fact that king Mahendra-
Vikrama whose inscriptions I copied at Pallftvaram (Pall. Ant.,
Vol. I, Plate XXI, A.) and at Trichinopoly (Pall. Ant., Vol. I,
Plate XXIII.) and who in all probability cut the caves containing
these inscriptions had many surnames; such as Chivibhundupdu,
Nilvil&nayyambu, Ventulavittu, Pasarambu, etc., which seem to
be all of Telugu origin. Those who have studied the inscriptions
at Trichinopoly (Archaeological Survey of India Annual Report
for 1903-4, p. 271) and at Pall&varam ((?. 0., No. 538, 28th July
1909 Part II, No. 14, p. 75), have mentioned this fact, but thejr
have not drawn any conclusion from it,
-28-
I believe that Mahndravarman I. had names of Telugu origin
because he or his father perhaps reigned over the Telugu country.
Now the following important question presents itself : What
was at the time of the Pallavas the line of demarcation between
the Tamil and the Telugu countries ?
Epigraphy has furnished an answer to this question :
It is worthy of note that the earlier inscriptions at Tondama-
n&d, KSJahasti, Gudimallam, Tirupati, Tiruchch&ntir and Y6gi-
J> Mallavaram in the North Arcbt district are in Tamil, though the
prevailing language is at present Telugu. The same was the
case at Nellore in the 12th and 13th centuries. We may, there-
fore, conclude that the tract comprising these villages and the
southern portion of the Nellore district was originally Tamil
country and the change of language into Telugu probably began
a during Vijayanagara times.)) (G. 0., Nos. 678, 679, 12th Aug.
1914, p. 7).
From the above, it follows, that at the time of the Pallavas
the region now forming the Nellore district served as the frontier
between the Tamil and the Telugu countries.
We have therefore to conclude that we find Telugu names in
the caves of Mahendra because either Simhavishnu or MahSndra
himself reigned over the country lying to the north of the modern
town of Nellore. In Vol. I. of ' Pallava Antiquities ' I have said
that in the Tamil country there is not a single antiquity which
could with certainty be attributed to the time anterior to that of
Mahendra; I have therefore been led to think that it was this
king who, by his own initiative, spread in the Tamil country a
taste for sculpture in general and rock-cut temples in particular.
But whence did Mahendravarman himself get this taste for
temples sculptured in rocks ?
The reply can be easily found if we take the two preceding
propositions together. As Mahendra reigned in the Telugu
country it was probably in the banks of the Kyishn& that he
acquired a taste for rock-cut temples ; and so the Pallava Art of
the time of Mahendravarman had its origin in the Telugu country.
It is but a hypothesis, but a hypothesis that can be verified by
studying the art that flourished in the VI century on the banks of
the K?ishn&. It is certain that this art existed : There are rock-
cut temples at Bezwada and at Mogulrazapuram on the northern
bank of the Krishn&; and on the southern bank there are the
temples of Sittanavasal and Undavalli. It must however be
admitted that this art has remained quite unknown up to the
present : only, the Undavalli temple has been described not very
distinctly but in such a manner that it is impossible to form a
correct idea of the style of those sculptures.
For these reasons I thought it necessary to go and see the
caves on the banks of the Krishna and particularly the temple of
AnantaSayana at Undavalli.
We know that this temple does not contain any very ancient
inscription and that the age of this monument can be ascertained
only from its architectural style.
And the authors who have handled this subject hold such
different opinions that it is very difficult to arrive at a decision on
this matter.
Sir Walter Elliot (Ind. Ant., Vol. V, p. 80) was struck by the
resemblance they bore to the sculptures at Mah&balipuram ; but
there he speaks only of the general impression ; this authordoes
not seem to have made a minute study of these monuments ; at
all events, he gives no arguments in support of his thesis.
Mr. Sewell (List of Antiquities, Vol. I, p. 77) is of opinion that
it is the work of the Chalukyas. And M. R. Ry. H. Krishna Sastri
who visited these caves on the 20th and the 28th December 1908
says very correctly that it does not seem to be the work of the
Chalukyas, because, not a single cave is known to have been dug
by them (G. 0., No. 538, 28th July 1909, Part II, No, 13, p. 74).
He remarks a resemblance with the Pallava sculptures:
<c The ornamental designs on the tops of four of the niches in this
hall resemble very much those on the " Rathas " at Mahabali-
J> puram.a
M. R. Ry. H. Krishna Sastri remarks, however, certain points
of resemblance with the caves of Orissa and thinks that the
temple at Undavalli might have come into existence in the
J> Andhra periods
Thus we see that the opinions of the various authors differ
much. Mr. H. Krishna Sastri attributes these sculptures to the
Andhra period, i.e., the I or II century of the Christian era ; Mr.
R. Sewell, on the contrary, dates them in the VII or the VIII
century. From this we may conclude that a general impression
will not do for fixing the age of these sculptures and that only a
-3d-
toost attentive study of the minutest details of ornamentation catt
lead to any adequate result.
In 1912 I visited the caves of Bftd&mi which are the works of
the ancient Ch&lukyas. I examined the style of these temples
very attentively and have given my impressions about them in
my book (Archfologie du Sud de I'lnde, Tome I, Architecture,
ps. 173, 174, 175 and PL CLVIII, B). The style of these caves
differs entirely from the Pallava style and there is no resemblance
between them. Again, the old theory that the Chalukyan artists
were the authors of the Pallaya temples cannot at all be admitted
and I do not believe that anybody will dream of maintaining it
to-day.
It was not without curiosity that I visited the Undavalli
caves. What was the style of these sculptures and what did they
rese nble ? The Andhra or the Chlukya or the Pallava style ?
What was my surprise when I stood before those sculptures
at Undavalli ! The Undavalli sculptures belong entirely and even in
the minulest details to the style of MaMndravarman.
We have given a description of this style in Chapter II, Vol. I
of " Pallava Antiquities " ; and the resemblance between the caves
of MahSndra and those at Undavalli is so complete that there is
no difference to be pointed out.
The plans are the same. The principal cave at Undavalli is
a four-storied one. The ground-floor and the top floor remain all
unfinished. The last story but one where the image of Ananta-
Sayana (Vishnu lying on a serpent) is placed, is also left un-
finished. There is thus but one story that has been completed.
Here the plan is very simple : they are three caves of Mahendra
placed beside one another.
The middle cave is almost in the same plan as the rock-cut
temple at Paliavaram (Pall. Ant., Vol. I, Plate XX). The right
and the left caves have the same plan as the Mandagapattu cave
(Pall. Ant., Vol. I, p. 56) and the left cave at M&mandtir.
Besides, there are many other caves in this very rock at
Undavalli; they are almost all of them in ruins but their plan is
easily ascertained. It is the very simple one used in all the
temples cut in the rock during the time of Mahendra.
The pillars belong to the well-known type represented in
Plates IX, XII, XIII, XV, XVI, XIX, XXI, XXVII, XXVIII,
XXIX of Vol. I, of " Pallava Antiauities ". Thev are souare in
Section and adornedr-with lotus flowers as at Atah6gd rflV Mi
Dalav&nfir. The Dv&rap&las resemble neither those of the
temples of Rajasiihha nor those of Mahftbalipuram, They are
like those found in the caves of Mah6ndra. Some of them have
their hand raised in sign of adoration as at Tirukkalukkunram
(Pall. Ant, Vol. I, Plate XXVI), at Singavaram, at M,raandfir and
at Dalavftnftr (Pall. Ant., Vol. I, Plate XVI, B). But most of them
rest it on a club with the same pose that we have noticed at
Trichinopoly (Pall. Ant., Vol. I., PL XXII), at Vallara (Pall. Ant.,
Vol L, PI. VIII) and at other places : Dalavinflr, Mandagapattu,
Tirukkalukkunram, Mamandtir, Siyamangalam, etc.
The niches are ornamented with a very special kind of frame-
work that we have noticed at Dalavftnfir (Pall. Ant., Vol. I.,
PI. XVI,) and at Siyamangalam (Pall. Ant., Vol.1., PI. XVIII)
and which we have called 4 double-arched tiruvatchi' (Torana).
The shrines are empty as elsewhere in most of the temples
of Mahindra but there exist on the right side of the temple at
Undavalli certain niches that contain lingams.
At Undavalli we find ktidus with the head of Gandharva
resembling those seen in the Pallava temples (vide the kfidus of
Dalavanfir represented in Pall. Ant. f Vol. I, PL XVI, A.)
The floors of these temples are ornamented outside with
pavilions resembling those at Mahabalipuram ; they are the
" Karnakftdu " and the " Sfilai " like those represented in Fig. 23
of our work on " Dravidian Architecture."
Besides, the small niches, mentioned above, which are found
to the right of the facade, are but small " Rathas ". They exactly
resemble the small shrine seen in the middle of the bas-relief, ,
" Bhagiratha's penance," at Mahabalipuram. The style of
architecture is identically the same.
All the authors that have spoken of the temple at Undavalli
have believed that it was dedicated to Vishnu ; it is not so. No
doubt the Vishnu cult occupies a predominant place there but
there are also many shrines dedicated to Siva. We have already
said that the small " Rathas " to the right of the facade contain
lingams.
The principal image in the temple is indeed that of Ananta-
&ayana ; but this image is found also in the temples of Siva.
At Mah&balipuram in particular, the bas-relief representing
Vishnu lying on the serpent is not in a temple of Vishnu. It is
-32 -
found in the cave near the light-house which was undoubtedly
dedicated to Siva ; god Anantasayana is represented in the same
form both at Undavalli and Mah&balipuram. Unfortunately at
Undavalli almost the whole of the body of the god has dis-
appeared ; it may be due either to the wear and tear of time or to
the iconoclastic work of the Mussulmans. The image we now
find there, is a rough figure made of cement. However, a close
examination of it shows that at one time the god had his arm
stretched out as at Mah&balipuram and Singavaram.
In short, the caves of Undavalli differ completely from those
of Bad ami and resemble jnuch those of the Pallavas.
However they are not exactly like those of Mah&balipuram :
Nowhere do we see the squatting lions supporting the pillars
[we have said that these squatting lions should have been an
invention of the artists of the epoch of Narasimhavarman L] ; the
Dvrap*Uas at Undavalli do not resemble those of Mah^balipuram.
On the contrary, the sculptures of Undavalli very much
resemble those of the caves of Mahndra.
(a) The plan of the caves is the same.
(b) The pillars have cubical parts ornamented with lotus
flowers.
(c) The doors and the niches have a kind of framework
which is different from the ' doubled-arched tiruvatchi '.
(d) The Dvaraplas have the same pose.
In short, the caves of Undavalli belong to the style of
Mahendra.
We have said above, that, during many centuries, many gene-
rations of Pallava kings had reigned over the country near the
banks of the Krishna in the districts of Guntfir and Nellore.
We have also observed that the caves of Trichinopoly and Pall&-
varam contain the surnames of Mahendra which are all of
Telugu origin and we have supposed that MahSndravarman I
reigned over the Telugu country and imported into the Tamil
country the art that existed on the banks of the Krishna.
Undavalli is in the Guntflr taluq of the Gunttir district. The
caves are situated on the southern bank of the Krishna. In the
absence of any inscription enabling us to know the origin of these
caves, we might suppose that they are the work of the Pallavas
who reigned over this country before it was conquered by the
Ch&lukyas.
Fig. 1.
A lion (simha) at
Undavalli.
Well, that is not my opinion.
I do not think that the caves of Undavalli were dug by the
Pallavas. On closely examining the sculptures, I have made a
discovery which I believe will prove to be of great importance in
ascertaining the origin of these temples.
On the cubical portion forming the
foot of two of the pillars I have found the
image of lions (Fig. 1).
At the foot of another pillar I have
observed the image of a vase.
It is true that these lions resemble
very much those of 6iyamangalam, an
image of which is given in Vol. I. of
Pallava Antiquities (PL XIX).
Lastly certain coins bearing the
images of a lion and a vase have been
attributed to the Pallavas (vide Vincent
A. Smith's Early History of India, Plate
facing p. 1).
It would therefore be possible to use this argument to affirm
that the caves of Undavalli are the work of the Pallavas.
My opinion, which is quite different, is that the Caves of
Undavalli are the work of the Vishnukundins.
There is no doubt that this dynasty reigned on the banks of
the Godavary and the Krishna before that country was conquered
by the Ch&lukyas. The village of Peruv&daka in the district of
Plaki-r&shtra, which belonged to the Vishnukundins at the time
when Indravarman granted the Ramatirtham plates (Ep. Ind.,
Vol. XII., No. 17, p. 133), passed into the hands of the Chaiukyan
king Vishnuvardhana I. while he was still a vassal; of Pulakesin II.
iTimmapuram plates Ep. Ind., Vol. IX., p. 317).
The Vishiiuku^dins were therefore the predecessors of the
Chftlukyas in the Vengi country. One of their capitals was
Lendulftru (Dendultiru in the Ellore taluq). The cradle of
this family was probably Vinukonda in the Krishna district.
They were fervent worshippers of the god at &riparvata(6rifiailam)
in the Kurnool district.
As it is certain that the Eastern Chftlukyas reigned in the
Guntftr district when Sarvaldk&grftya granted the village of
Chandalflr in 673 A.D. (Ep. Ind., Vol. VIII, No. 24), we may affirm
a
-34 -
that the Vishnukundins reigned before this epoch, probably
towards the end of the VI century.
From our point of view, the Chikkula plates (Ep. Ind.,
Vol. IV, No. 25, p. 195) are very interesting.
1 The se.al represents a lion exactly like the one we have
remarked at Undavalli. The Chikkula seal is represented in
Ep. fad.. Vol. IV., to face page 244. This seal is analogous to the
Ramatirtham plates which M. R. Ry. H. Krishna Sastri has
described as follows :
An advancing lion or tiger with its left fore-paw raised, its
neck erect, mouth wide open and the tail swung over the back so
as to end in a loop. (G. ., No. 538, 28th July 1909).
It must be noted that coins bearing a lion and a vase are
found on the banks of the Godavary and the Krishna ; and I am
convinced that they do not belong to the Pallavas but to the
Vishnukundins.
2 The village gifted away by these plates is Regonjram
which is situated to the south-east of the village of R&virva on
the bank of the Krishna : Regonram, which was south-east of
the village of Revireva on the bank of the Krishna benna, i.e., the
river Krishna. .
So it is beyond all doubt that the Vishnukundins reigned on
the banks of the Krishna.
3 The following genealogy is found in these plates :
Madhavavarman
i
VikramSndravarman I.
i
Indrabhattarakavarman
VikramSndravarman II.
We know also that the mother of Vikramdndravarman I. belonged
to the family of Vakatakas. Indeed, it has been said of Vikra-
mndravarman I. that his birth was embellished by the two
families of the Vishnukundins and Vakatakas. .
On the other hand, a donation of the Vishnukundins dis-
covered in 1914 (No. 7 of Appendix A., G. 0. No. 920, 4th Aug.
1914 Part II, No. 35) shows that the name Vikram^ndra is a
corrupted form of VikramahSndra.
If we remember, then, that in the 6iyamangalam cave there
is the image of a lion resembling the one found in the seal of the
Chikkula plates, that the Undavalli sculptures resemble those of
the Pallava caves, that Undavalli stands on the banks of the
Krishna where the Vishnukundins had reigned, and that the
inscriptions of Mahendravarman I. at Trichinopoly and at Pall&-
varam contain Telugu epithets, we shall be struck with these
coincidences.
But there is another point of coincidence. Among the Pallava
kings that have reigned before Simhavishnu, there is not one
named MahSndra. It would appear that this name was not
hereditary in the Pallava family and that Mahendravarman I.
was really the first king of that name.
The account given of the V^yalfir inscription in the report on
Epigraphy for 1908-09 mentions Mahendravarman (once) among
the 25 predecessors of Simhavishnu. Now that the VHyalilr
inscription is completely deciphered, it is manifest that that name
is not there, but we find the name (6) Chandravarman. Perhaps
it is the ending "ndravarman" that led to the reading [Mah]ndra-
varman.
In the Trichinopoly and Pall&varam inscriptions this king is
called Mahdndra-Vikrama. But then, there were Vishnukundin
kings who bore the name of VikramahSndra.
In my opinion, these coincidences can all be explained in a
very simple manner: The Pallavas at the end of the VI century
reigned in the districts of Nellore and Gunttir ; their neighbours
were the Vishnukundins who reigned on the banks of the
Krishna ; Simhavishnu married probably the daughter of a
Vishnukundin king named Vikramahendra and gave his son the
name of his grand-father, Mahendravikrama.
We know that VikramSndravarman 1. was the son of a
V&k&taka princess ; and the inscriptions of the V&k&taka kings
are found engraved in the caves of Ajanta. It is probable that it
was owing to their VSMtaka origin that the Vishnukundins had
the idea of digging caves on the banks of the Krishna caves that
we see even now at Bezwada, Mogulrazapuram, Undavalli and
Sittanagaram. The Pallava king Mahendravarman I, who was
the grand-son of a Vishnukuiidin king, having had many occasions
to admire those caves that had been dug by his relatives, had
similar ones cut on the rock around K&nchipuram,
CHAPTER IV.
THE DYNASTY OF SlMHA VISHNU.
I. The V61flrpfi]aiyam plates coupled with the Vftyaltir
inscription inform us that Sifhhavishnu was the son of (31) Simha-
varman.
The VglfirpSjaiyam plates say of Shhhavishnu :
He quickly seized the country of the Ch6Jas embellished by
the daughter of Kavira (i.e. the river K&v6rf), whose ornaments
are the forests of paddy (fields) and where (are found) brilliant
J> groves of areca (palms).
From this it would appear that the Ch6|a country did not
belong to the Pallavas before Simhavishnu and that it was he
who conquered it.
This military operation was perhaps difficult, for it seems
that all the southern kings opposed it : the Kas&kudi plates say,
indeed, that Simhavishnu vanquished The Malaya, Kajabhra,
)> Malava, Oh6|a and P&ndya (kings), the Simhala (king) who was
proud of the strength of his arms, and the
II. MahSndravarman I. is the first king about whom we
have precise information.
It is probable that MahSndra gained a victory at PuJJaltira
[according to the K&s&kudi plates]. It is thought that this town
can be identified with Pullalftr (Chingleput District Conjeeveram
taluk) which is at a distance of 15 miles north of K&fichf. No one
knows for certain who was his adversary ; it is supposed it was
Pulakgfiin II. -
The Aihole inscription (Ep. Ind., Vol. VI, No, 1, p. 11) praises
PulakSsin thus :
(V. 29) ......... He caused the splendour of the lord of the
Pallavas, who had opposed the rise of his power, to be obscured
by the dust of his army, and to vanish behind the walls of
K&fichtpura.
'67
(V. 30) When straightway he strove to conquer the Chdjas,
the Kftv6ri, who has the darting corps for her tremulous eyes, had
her current obstructed by the causeway formed by his elephants
> whose rutting-juice was dripping down, and avoided the contact
with the ocean.
(V. 31) There he caused great prosperity to the Chdjas,
a K3rajas and P&ndyas, he being the hot-rayed sun to the hoar-
frost the army of the Pallavas.
We have said in the preceding chapter that the country lying
between the towns of Ellore and Guntfir probably formed part of
the kingdom of MahSndra.
Again, it is certain, that, in the middle of the VII century,
this country belonged to the Chaiukyas.
It is, therefore, likely that Pulak6sin II conquered it about
610 A.D., i.e., at the beginning of the reign of Mah^ndravarman I.
The latter, thus dispossessed of the northern provinces of his
kingdom, lived in the Tamil country during the latter part of his
reign and it was then he encouraged the arts of that country as
we shall presently show.
In Pallava Antiquities (Vol. I. p. 40), we have admitted with
V. Venkayya (Ep. Ind., Vol. Ill, p. 277) that MahSndra who was
first a Jain, was converted to the Siva cult by saint Appar
(Sfikkilfir's Periapur&nam, Madras, 1870); we have also said,
though without any positive proof, yet, owing to our moral
conviction, that it was MahSndravarman I. who was the author
of the rock-cut temples that we have described in Chapter II. of
the aforesaid book.
In the course of this year (1916-17) we have learnt much more
about Mahendravarman I.
In " Pallava Antiquities " Vol. I., I have spoken about the
right side cave at Mamandfir (ps. 53, 54, 55,). Judging from the
style of Architecture, I have attributed this cave to Mahendra-
varman I. and have described the inscription found there in the
following terms : Mr. E. Hultzsch who mentions this inscription
(No. 38 (?. 0., No. 424, 20th April 1888), declares that it is
" illegible ". However, one important remark has to be made
here : the alphabet is identically the same as that of MahSndra-
vftdi. PI. XXVII, B. is a photograph of a small part of the
D inscription which will enable us to judge of the form of the
characters. D
38 -
So then, relying on :
1 the style of architecture,
2 the palaeography of the inscriptions,
I concluded that this cave must be attributed to MahSndra-
varman I.
When I examined the inscription, it seemed to me that,
though it was very much damaged, it would perhaps be possible to
make something out of it.
Mr. Hultzsch, when he copied it, did not understand it, but
that is not a reason why it should^be abandoned for ever.
I believed that a more attentive study of it might enable us to
obtain some interesting information.
So, in January 1917, 1 went to M&mandfir to copy the inscrip-
tion. ^
This labour was well rewarded :
In the midst of many incomprehensible phrases I read the
words : " Mattavilasdipadamprahasana " (see PL III, A). I did not
understand their meaning; but I was at once struck with the
name Mattavil&sa which is a name of Mahendravarman I. that
I have seen engraved in the caves of Trichinopoly and Pallvaram
(Pall. AnL, Vol. I, p. 39); and I was glad to have discovered a
new proof of what I had asserted in attributing the Mamandtir
cave to Mahendravarman I.
I sent a copy of the Mamandtir inscription to M.R.Ry. T. A.
Gopinatha Rao requesting him to tell me what he made out of it.
A few days after, I received from him an article which he had
published in February 1917 in the " Madras Christian College
Magazine " which mentioned the following important discovery :
"Pandit T. Ganapati j&stri," the curator of Sanskrit Manu-
scripts, Travancore, has very recently discovered a manuscript
called " Mattavil^sa-prahasana ".
What is very remarkable is that the author of this Sanskrit
poem is a king named Mahendravarman. It is specified that this
king belonged to the dynasty of the Pallavas of Kdnchi, that he
was the son of Simhavishnu, and that he had the surnames of:
AvanibhSjana, Mattavil&sa, Gunabhara, Satrumalla.
As soon as I received this communication I replied to
M. R. Ry. T. A. Gopinatha Rao by a letter dated 12th April
informing him that I was very pleased with the discovery he
_!_!_ 1
Plate III.
A.
Mamandftr
cave insc.
B
Kit ram
Siva temple
inscription.
C.
Kanchipuram
Airavatesvara
temple insc.
words that I had read at the end of the 6th line of the M&mandflr
inscription seemed to confirm the fact that the poet king
MahSndravarman I. was the author of the earliest rock-cut
temples. Besides, as the musical inscription at Kudumiyamalai
(Ep. Ind., Vol. XII, p. 226) is written in an alphabet resembling
that of Trichinopoly (Pall. Ant., Vol I, PI. XXIII), we may
suppose that Mahendravarman I. was also a musician.
Mr. Gopinatha Rao in his reply dated 19th April said :
No doubt the discovery of the burlesque Mattavillisa-
i> prahasana is important ; but what is more important is the
mention of this work very definitely in the mutilated inscription
J> of M&mandfir The musical composition discovered and
copied for the Epigraphist's office by me [at Kudumiyamalai],
D is not a composition of Mahendravarman though it belongs no
a doubt to the same period. It was composed by one Rudr&charya.
a No doubt in the Annual Report Venkayya says :
The inscription was apparently engraved at the instance of
an unnamed king, who was a disciple of a certain Rudr&charya
and who composed these " svaras " for the benefit of his pupils.
We may only suspect if the unnamed king be not MahSndra-
varman, but cannot be sure of it.
And in a letter dated 28th April, he added :
The [Mamandflr] inscription seems to refer to " svaras " and
a"varnas" of Music, is your surmise that the Kudimiy&malai
Musical record was engraved at the instance of MahSndra going
to be true ? talks of Kavis (poets), mentions Valmiki and the
Matavilasaprahasana. As suspected by you, the record perhaps
gives a panegyric on the literary and musical talents of the
Pallava king Mahendravarman.))
The village of MahSndramangalam in the subdivision of
M&vandfirppaml mentioned in the inscription No. 41 of 1890 on
the store-room (northern wall) of the Vardham^na Temple at
Tirupparuttikkunru (Ep. Ind., Vol. VII, No. 15, p. 115) has
certainly been named after Mahendravarman I.
The tank at Mamandtir^was named CitramSga, probably after
a " biruda " of MahSndra.
The discovery of the similarity of architectural style existing
between the caves of MahSndra and those at Undavalli and the
other discovery that in the M&mandfir cave there is mention of
Mattavilftsaprahasana, these two discoveries taken along with
-40-
others enable us to conclude that King Mahdndravarman I. is one
of the greatest figures in the history of Tamilian civilisation.
1 From a military point of view, he has checked at Pullalftr
the invasion of the Ch&lukyas.
2 As for religion, he has given a new impulse to Saivism.
3 As for the arts, being himself a royal artist, he has
glorified poetry and music.
4 As for the plastic art, he has transported the taste for
rock-cut temples from the banks of the Krishna to those of the
P&lar and the K&vri.
5 As for the administration, he built the tanks at MahSndra-
vftdi, M&mandtir, and probably also at Dalav&ntir.
Mahendravarman I. has opened a new era whose apotheosis
we shall see in the reign of his son Narasimhavarman the Great.
III. In 640, the Chinese pilgrim Hiuen Tsang stayed at
Kanchipuram and it is probable that this year falls within the
reign of Narasimhavarman I. The chronology of the Chfiiukyas
says indeed that it was about 642 that this king took possession
The Ktiram plates (/./., Vol I., p. 152) give us information
about this event. They say first that Narasirhhavarman I.
vanquished Pulakein in the battles of Pariya}a, Manimangala,
6uram&ra, etc. Probably Manimangala is nothing but Maniman-
galam (Saidapet taluk, Chingleput district) which is at a distance
of 20 miles from K&fichi. In that case it would appear that the
Tamil country was invaded a second time by the same Ch&lukyan
king. For the second time Pulak6in II. was repulsed by the
Pallavas. The glorious Ch&lukyan emperor who had vanquished
Harsha Vardhana and whose friendship had been sought by the
King of Persia was thoroughly routed and his capital Vftt&pi was
destroyed.
In regard to this last point, all documents (Ktiram, Kfts&kudi,
Udayendiram, etc.) agree. The Vfiltirp&laiyam plates add also a
detail : Narasimhavarman (I) topk t (from his enemies) the pillar
of victory standing in the centre of (the town of) Vfttapi (8.1. 1.,
Vol. II, part v., p. 511). This fact is confirmed by the inscription
at B&d&mi (2nd. Ant., Vol IX) the alphabet of which is the same
as that of the Pallavas and which mentions Narasimhavishnu,
alias Mah&malla as also a commemorative column (Jayastambha)
Speaking of Narasiifahavarman I., the Kftfiftkudi plates say :
Who surpassed the glory of the valour of Rftma by (his) con-
quest of LankH
This event is confirmed by the Chronicles of Ceylon (Trans-
lation of the Mahavamsa by Wijesinha) and it took place after
642 A.D., because, according to these chronicles, the Singhalese
prince Mftnavamma aided Narasiihha in repulsing Vallabha
(Pulakftsin) and the two expeditions to Ceylon took place only
after that.
It was perhaps during these expeditions that Narasimha
vanquished the Chdlas, Keralas, Kajabhras and Pandyas as men-
tioned in the Kftram plates.
It is probable that the navy took part in the conquest of
Ceylon, for the Mahftvamsa says that M&navamma crossed the
sea in ships.
We may suppose that the port of M&mallapuram served as the
naval station for the Pallava fleet. Even now Mahabalipuram
serves as a landmark for all vessels.
In the first volume of Pallava Antiquities 1 have proved that
the monuments at Mahabalipuram do not belong to the style of
Mahendra, that this town did not probably exist before the time of
Narasimhavarman I., and that it is this king that founded it and
and gave it the name of Mahamallapuram after his own name of
M^raalla and began the cutting of the " Rathas " and " Caves "
there.
I am convinced that in the year 650 A.D., the sculpturing of
the rocks of Mah&balipuram was being executed.
The second expedition to Ceylon was crowned with success ;
the Pallava army conquered Ceylon ; and M,navamma cut off the
head of King Hattha-datta II.
In what year was Ceylon conquered ? Since we know that
after the capture of Bd&mi there were two expeditions to this
island, this conquest probably took place several years after
642 A.D., and I would put it after 650 A.D.
The chronology of Mah&vamsa affirms that the death of
Hattha-data and the coronation of M&navamma took place in
691 A.D. What value are we to attach to this information ?
I think we may affirm that for the X century (900 A.D.
1000 A.D.) the chronology of Mahftvanisa is very correct but on
condition that the dates are all reduced by 24 years.
42
Does this rule apply to the VII century ? If we take away
24 years from 691, we get 667, i.e., 17 years more than 650.
We have said that Ceylon was conquered perhaps after 650,
but it appears that it was less than 17 years after that date.
I think that the following hypotheses may be admitted :
1 Ceylon was conquered about 660 A.D., at the end of the
reign of Narasimhavarman I.
2 The chronology of Mahavamsa is accurate enough pro-
vided we reduce the dates by nearly a quarter of a century.
We shall have occasion l^ter on to utilise this result.
IV. The reign of Mahendravarman II, was probably short
and uneventful. The Ktiram plates, dated in the time of his son,
only say that his reign was prosperous and that he was a legis-
lator <( who thoroughly enforced the sacred law of the castes and
the orders .
V. The Ktiram plates give a long description of the
military exploits of Paramesvaravarman I. and also mention the
name of his royal adversary: He made Vikram^ditya, whose
army consisted of several lakshas, take to flight, covered only
by a rag..
On the other hand the Udayendiram plates (8.1.1., Vol. II,
Part III, p. 371) give us the name of the battle : Paramesvara-
varman, who defeated the army of Vallabha in the battle of
D Peruvalanalltir .
An extremely important detail, which, in my opinion, has not
so far been noted well, is the date of the conflict between
Paramesvaravarman I. (called Ugradanda and Lokaditya in the
inscriptions of the Kailasanatha temple at K&nchipuram, S.l.L,
Vol. I.) and Vikram&ditya I. (Ranarasika).
From 1910, we have been in possession of a document which
gives this date with great certitude, I mean the Gadval plates.
This grant (Ep. Ind., Vol. X, No. 22, p. 101) has been made when
Vikram&ditya I. was encamped in the Ch6la kingdom on the
southern bank of the K&vSri.
Besides, they bear the date of the year of the reign and of the
&aka era. The exact date is incontestably 674 A.D.
We shall now proceed to determine a second point of vefcy
great importance.
- 43 -
Mr. Hultzsch who edited these plates has said :
When Vikram&ditia made this grant, his arfny had invaded
the Chdlikavishaya, i.e., the Ch6Ja province, and was encamped
in Uragapura on the southern bank of the K&v6ri river (1. 25 f.).
J> In sending me the impressions, Mr. Venkayya drew my atten-
tion to the fact that Uragapura is mentioned in Kaiid^sa's
Raghuvamsa (vi. 59) as the capital of the P^ndya king, and
J> proposed to identify it with the ancient Choja capital Uj-aiytir
near Trichinopoly ; I rather think that Uragapura, " the snake-
city " is a poetical equivalent of N^gapattanam (now Nega-
patam) .
But, I believe that Mr. Hultzsch is entirely mistaken.
In connection with this subject I have made a discovery
which appears to me to be decisive : I have spotted the village of
Peruvalanalltir where was fought the battle between Para-
mesvaravarman I. and Vikram,ditya. This village is in the
Trichinopoly Taluq of the district of the same name, on the left
bank of the K&v6ri, and at a distance of 10 miles north-west of
Trichinopoly.
I am of opinion that it is not possible to doubt that the town
of Uragapuram where Vikram^ditya I. was encamped is any
other place than Uraiyur (Trichinopoly).
In the inscription of Rjasimha found at the base of the
Vimana of Kail&san&tha temple at Kanchipuram (/./., Vol. I.
p. 13) Paramesvara is named Ugradanda the destroyer of the
city of Ranarasika . It is to be noted that in the Gadval plates
the surname of " Ranarasika " is given to Vikram^ditya. But
which is the town designated by the words the city of Rana-
rasika ?
I do not think it could be Bad&mi, for, in that case, Para-
mesvara would not have failed to assume, as his grand-father,
the title of " Vat&pikonda ".
I believe that " the town of Ranarasika " is Uragapuram
(Uraiyur).
The Gadval grant which was made when Vikramftditya was
encamped at Trichinopoly gives us a detailed description which
is very life-like of the third invasion of the CMlukyas into the
Tamil country.
The Pallavas are called there "The family of MSmalla "
(verse 5).
-44-
How did Param66varavarman I* succeed in repulsing the
Chftiukyas ?
The Kfindftr plates (Ep. /we?., Vol. IX, No. 29, p. 205) tell us
that the Ch&lukyas had to contend against the P&^dyas, the
Chd}as, the K&rajas, the Kajabhras and the kings of KavSra,
P&rastka, Sirhhaja. It is therefore probable that all the people of
the south combined together against the invaders. The aforesaid
plates themselves say that there was a confederation of three
kings. One of these was perhaps the king of K&nchi. Mana-
vamma, the king of Simhaja, was probably another. The Mahft-
vaihsa says that this king reigned 35 years. We have said that
he ascended the throne about 660 A.D. So he must have surely
reigned in 674 A..D. He had been Minister to Narasimha-
varman I. and he was perhaps attached to Paramesvaravarman I.
by ties of friendship. If our suppositions are correct, by 674 A.D.,
only 14 years would have passed since he ascended the throne
with the aid of the Pallavas. And at a time when the latter
were in danger it was his duty to act according to the dictates
of the simplest feelings of gratefulness.
The third king that entered the coalition against the
Chaiukyas was probably the king of the P&ndyas. We shall
refer to this question again when we examine the history of the
Pa^dyas later on.
It is certain that Paramesvaravarman lived some time after
the defeat of Vikramftditya, for the Kflram plates are posterior to
this event, but unfortunately they are not dated.
The object of the Kfiram record (8.8. /., Vol. I., p. 154) is a
gift made to the temple of Siva called Vidy&vinita-Pallava-
Paramsvara which had been built at Ktiram by Vidyftvinita-
Pallava.
There is a temple of Siva now at Kftram. The Epigraphical
Department have copied an inscription of Nandivarman Maharaja
found in this temple (No. 38 of 1900); the writing seems to
belong to the IX century, but they have not found any inscription
which could be attributed to the epoch of Paramfifivaravarman I.
When I visited this place, I discovered an inscription which
has not been mentioned by any one till now. I do not know how
it has escaped investigation so long, seeing that it is well
preserved and the letters are all cut fair and deep. This inscrip-
tion which is reproduced in Plate III, is the following :
Text:
U) ff f IT 9 69T
&ri-Pallava
The name of the king is not given, but the alphabet employed
leaves no doubt as to the antiquity of the inscription. The letters
' Sri ' and * va ', in particular, are so formed that we may say that
the inscription belongs to the VII century.
So, the Siva temple at Kfiram is certainly the VidySvinita-
Pallava-Paramesvara temple* mentioned in the plates, but alas !
in what a dilapidated condition ! No doubt the adjoining
mandapam has been rebuilt with the old stones. As for the
sanctuary itself, only the ground-work remains, which, however,
enables us to know two important facts :
1 The temple was apsidal in form as the Sahadevaratha at
Mah&balipuram.
2 The entrance into the sanctuary was set towards the
west which is a peculiarity frequently seen in Pallava temples ;
almost all the temples of Mahabalipuram and many of the Pallava
temples at K&nchipuram face the west.
The discovery of the temple of ParamSsvara at Kflram is
important from two points of view :
1 It is interesting to identify the temple that was the object
of the Ktiram grant.
2 This temple is in ruins; however, what remains of it
constitutes the most ancient monument in South India which is
known to have been built of stones placed one above another.
VI. In Chapter I of " Pallava Antiquities " we have spoken
about the son of ParamSSvaravarrnan L, Narasimhavarman II.
surnamed Rfijasimha. He is perhaps the only Pallava king who
had a long and peaceful reign. He does not seem to have done
any thing else during his reign except loading the Sivite priests
with favours and building the temples mentioned above, the
Kailftsanfttha temple at K&nchlpuram, the Shore temple j.tJMaha.
balipuram and the Panamalai temple. To this list we may add
the^XirftvBt66vara temple at KMchipuram. This temple with its
rearing lions, the image of Sdmftskanda and the prismatic lirigam
-46-
presents all the characteristics of the style of Rajasimha. Mr.
A. Bea, relying upon its general aspect, has attributed this
temple to the Pallavas. However, Epigraphy has as yet furnished
no proof that will allow us to classify this temple definitively
among the works of the reign of R&jasiihha.
This temple whose Pallava origin is doubted has never been
visited, however, by the officers of Epigraph] cal Department.
When, in January 1917, I went to MS,mandtir, I stopped at
KMchipuram and visited all the temples there hoping to make
some discovery.
The temple of Air&vatSsvara stands amidst a group of houses
found opposite to the entrance of Kakhesvara temple.
On examining this monument, I found all round the base of
the temple some inscriptions which are very much damaged.
One part is written in Tamil and the other in Grantha.
To the right of the entrance and at the base of the temple
I found a fragment of an inscription and copied it as its import-
ance cannot be questioned : the name T^fiT5 Narasimha is written
very clearly (vide PI. Ill, C.).
VII. ParmSsvaravarman II, son of R&jasimha, is probably
the author of Vaikuntha Perumal temple at K&iichipuram.
The temple of Virattanesvara at Tiruv&di contains an inscrip-
tion (No. 56 of 1903) belonging to the reign of Paramesvara-
P6ttaraiyar. In p. 72 of " Pallava Antiquities" Vol. I., I have
spoken about the resemblance between the Vim,na of the Tiru-
vftdi temple and that of the KailSsanatha temple at Kafichipuram.
Therefore, we must perhaps attribute the Siva temple at TiruvSdi
to King ParamSsvaravarman II. However, this temple seems to
have been repaired many times (we know, from inscription 35 of
1903, that it was done during the reign of Nripatunga).
The Tiruv&di inscription is dated in the 3rd year of the reign
of ParamSsvaravarman. It is probable that his reign was short.
CHAPTER V.
INTRODUCTION TO THE HISTORY OF THE DYNASTY
OF NANDIVARMAN.
1. The Ganga-Pallava theory.
In 1887, there lived in India a gentleman who has rendered
eminent service to the history of the Pallavas and whose name
must not therefore be forgotten : It was Mr. J. Delafon, who was
a Magistrate at Pondicherry. He discovered at Bslhtir (near
Pondicherry) the plates of Nripatunga and at Kftsaktidi the
plates of Nandivarman Pallavamalla.
These two important discoveries are not due to chance ; they
are the happy result of the active and disinterested researches of
a savant who studied the history and the languages of India for
a long time.
Mr. Vinson, Professor in the School of Oriental Languages,
Paris, has given the complete text, Sanskrit and Tamil, of the
Bahfir plates, in an article that forms part of the memoirs
published by the above School for the Congress of Orientalists
held in 1905. However, the works that are published in French
are generally ignored by the scholars of India.
So, the B&hftr grant came to be known only by a shorl
summary of it given by Mr. Hultzsch in his article No 23, page
180 in Vol. IV. of Epigraphia Indica.
The text was republished a few days ago, but withoul
translation, in Part V, Vol. II of 8.LL, p. 514.
This text is incomplete : it is the work of a Pandit who livec
at Pondicherry at the time of Mr. Delafon, and in many place*
the transcription is defective.
I thought it necessary to give the readers of this book th(
translation of the B&htir plates. So, I requested M. R. Ry
T.A. Gopinatha Rao to translate for me the Sanskrit portion o
the text as published in Vol. II of /./., and he has been goo<
enough to send me the following version of it :
48
THE BlHtfR INSCRIPTION.
Text.
(Please retain the text as it appears in Part V, Vol. II. of
South Indian Inscriptions, with Mr. Krishna Sftstri's footnotes
also. In addition to these, please add the following) :
1. For ftSB in 1. 1 of verse 1, read ft?.
'2. The words ^fa^-wain* in 1. 1 of verse 11, is an incorrect
reading of the passage by the original transcriber, the
Pandit. As it is, the passage is not clear.
3. For ^si**Sli*r in 1. 2 of the same verse, read vaift.
4. For b<AIII^L ^ft, I propose to read ^FWft, in the battle of
Kal&va or Kal&pa, a place which requires identification.
5. For fRWWpft: in 1. 2 of verse 16, read SKl^fa. The river
Arichit is the same as the Ari&il&gu. It is usual for the
final t of Sanskrit words being read in Tamil countries
as I. This custom is still retained in its entirety and
often in a ridiculous manner in the Malabar country.
For example, in almost all Tamil inscriptions it will be
found that the phrase chandrddityavat written as
chandrddityaval. In Malayajam, words like tasmdt is
pronounced as tasmdl ; k&chiti kinchit, as k&chil and
kifichil respectively. On this analogy we may argue
that the Tamil name Arisil or Arichil was Sailskritised
into Arichit. It must be remembered that almost all
the battles between the Pallavas and the allied armies
of the Pandyas and the Ch6las, were fought in the
vicinity of Kumbhak&^am, near which is also the
river Ari6tl&U.
6. For ^RL in 1. 1 of verse 20, read $H^. It is only then
the passage makes any sense.
7. For BTHWT in 1. 2 of verse 20, read
8. For ^rt^: in 1. 2 of verse 20, read
9. For UI3R in 1. 1 of verse 21, read MIH4|4.
10. For **l*<^l in 1. 2 of verse 27, read <*<<^H ; and
11. For qJJUfl in L 2 of verse 29, read
Translation.
Verse 1. May Madhustidana, whose lotus-feet are rubbed by the
kirttas of the d^vas, who is the cause of the destruction of
the Rftkshasas who are dreaded by all the worlds, who is
eternal and whose eyes resemble the petals of the lotus,
give you prosperity.
V. 2, ......... From his naval rose the lotus flower which is the
origin of every (created) thing ; from it came Brahmft.
V. 3. From the lord of the world, the four-faced (BrahmS), was
born Aflgiras. From him, Brihaspati, the minister of
Sakra (Indra), the destroyer of (the asura) Vala.
V. 4. From him Samyu ; from him was born Bharadvftja. From
him the great archer Dr6na, who was as powerful in
battle as Indra.
V. 5.-XThen) came into existence from Dr6na Asvatth&man of
great power who was well-versed in the use of all
weapons and who was a (partial) incarnation of Pin&kin
V. 6. From AsvatthSma was born the king named Pallava. He
protected every one, from the cultivators up to the kings,
in the Navakhanda (the nine divisions of the earth ?).
V. 7. In his lineage were born the host (of kings), Vimala,
Konkaiiika etc., before whom the wives of the enemies
bowed, whose commands were obeyed by other kings also,
who were most dear (to their subjects ?) and who were
ever attended by the sound * jaya* (be ye victorious.)
Vv. 8-9. After Vimala and others, having ruled the earth which
is girdled by the four oceans, by their prowess had gone
to heaven in celestial chariots (vimdnas), there lived the
king Dantivarman, who was equal to Indra (in his power),
who was an intense devotee of 'Muradvish (Mur&ri,
Vishnu), who was powerful and who was worshipped by
their crowns by (other) kings.
V. 10. This sovereign, on account of his ruling the earth even in
this Kaliyuga with justice and of his pouring gifts, shone
like Indra.
V. 11. (This verse is incorrect. It appears to convey the follow-
ing idea in it. Nandivarman cut his enemies' heads
which resembled parcels of food offered to their souls
-50-
which were preparing to pay a visit to the house of
Yama).
V. 12. From Dantivarman was born tjie powerful Nandivarraan,
who, singly and unaided, took the earth in battle.
V. 13. Just as Lakshmi was (the wife) of Muradvish (Vishnu),
the queen named 6ankh&, who was born in the R&shtra-
ktita family, was the wife of Nandivarman.
V. 14. 6ankh&, the queen of the king, who resembled the earth
in forbearance, who was kind to the people like a mother,
was resplendent as Lakshmi incarnate.
V. 15. Nripatungad^va who was esteemed for his high birth, who
was lord of the three worlds, who was resplendent as the
rising sun and who was victorious in the battle of
Kalava (?) was born of that queen who had intelligence,
beauty, learning etc.
V. 16. The army Xof the Pallavas) which on a former occasion
sustained defeat in the hands of the P&ndya, was, by the
grace of this king (i.e., by being led by him), able to burn
down the hosts of the enemies together with the pros-
perity of their kingdoms, on the bank of the river Arichit.
V. 17. This illustrious king named Nripatunga though he was
young (in age) was not only famous in this, but also in
the other worlds, as Rama was. 1
V. 18. Marttanda of Ves^li who was born in the race of Kuru
and who was solicitous in protecting his subjects, was
bound by obligation to this king (Nripatunga.)
Vv. 19-20. He was great in the world like the moon ; in majesty
etc., he resembled the ocean ; he was an asylum to the
world (the people) inasmuch as he afforded protection
(to it) like the sun. Therefore it is but right that there
should be similarity between his body and name. 2 Or, he
is certainly a king because it is quite patent at sight (?)
1. Just as Kama, even as a boy, had conquered Rakshasas for Visvamitra and
become famous in this world and was held as an object of praise even by gods, Nripa-
tunga became famous as a boy- warrior and king even in the world of gods.
2. He is named Marttanda (meaning the sun) ; he exerts his energy in protecting
the country ; in this act he resembles the sun which sustains life on the earth by its
warmth and energy. So, Marttanda of the Vdsali family and of the race of Kuru is
akin in name and action to the sun. Here the reading has been slightly altered ;
instead of dtvavat, I am inclined to read ddhavat which makes sense,
Vv. 21-23. He, who was increasing (the prosperity of) the Kurti
race, having petitioned NripatuAga and duly obtained
permission through the Secretary (ajnapti), granted to
the vidydsthdna the three villages, namely, ChStup&kkam,
. ViJ&rigadirSphantam (i.e., Vi^ngattfir) and Iraippunaich-
chSri situated in his province (VSs&lippadi).
Vv. 24-29 Just as Dfirjati (Siva) bore on one of his^as (the
river) Gang& who was descending with a large number
of waves, similarly the river of learning consisting of
fourteen gu'nas (or divisions of knowledge) was spreading
round the abode of the residents of the Bahu-village :
because it is the abode of learned men, it is called a
vidyas1h&na.
This king, having given them (the learned men of
Bahtir) by his ajnapti the villages to the extent marked
by the circumambulation of an elephant, freed from all
taxes and protected thereby, honored himself.
The minister of the king ri Tungavarman, who has
reverential awe for the commands (of the king), who is of
a charitable disposition, who is held in regard by the
lord of the three worlds (perhaps this refers to the king),
who is famous like Brihaspati (the minister of the lord of
the celestials Indra), exhorts the future kings to protect
this charity (made by him).
V. 30. Dasaya, the servant of the vidydsthdna of the resident of
B&hu and who is himself learned in the principles of the
sdstras, wrote this eulogistic document.
[TAMIL PORTION, LEFT UNTBANSLATED.]
V. 31. " If equal merit accrues to both the giver of a charity and
its protector ,then do you protect it." so saying, the king
Nripatungavarman, by bowing lowly his head which bears
on it the feet of Mukunda (Vishnu), exhorts future kings.
V. 32. The goldsmith Nripatunga, who was an ornament to the
family of Uditodita, who was well-versed in all sdstras
and who was an hereditary servant of the Pallavas wrote
this document.
The Bahflr plates were discovered at a time when the Epi-
graphy of South India was yet unborn. The dynasty : Danti-
ftandi-Nfipatunga, was quite new. No sooner had a few storie
inscriptions been copied than Mr. Hultzsch tried to reconcile the
new records with that of Bfthftr.
Owing to an unlucky chance, there was, among the dis-
coveries that were made first, the Kil-Muttugftr inscription which
Mr. Hultzsch has published with the following remark (Ep. Ind.,
Vol. IV, p. 177):
The inscription is dated in the eighteenth year of the reign
of the king, the Victorious Narasimhavarman. The same name
occurs among the Pallava kings of Kftfichf. But the two centre
figures of the bas-relief below the inscription make it impossible
to attribute this record to the Pallava Dynasty, whose crest was
D a bull and whose banner bore a club. The elephant appears at
Dthe top of three stone inscriptions of the Western Ganga
dynasty which have been published by Mr. Kikel, and the goose
*> (hamsa) is said to have been the device on the banner of the
mythical Ganga king Konkani.
In editing the two inscriptions of Nripatunga found at
Ambtir (Ep. Ind., Vol. IV, p. 180), Mr. Hultzsch has said :
Besides, the Bhtir plates mention among Nripatuiiga-
D varman's remote ancestors Konkanika. This name seems to be
a reminiscence of Konkani, who is believed to have been the
ancestor of the Western Gangas.
And again (Ep. Ind., Vol. IV, p. 182) :
The Kil-Muttugtir inscription of the latter (Narasiihha-
varman) bears, however, the emblems of the W. Ganga kings
and its alphabet is more archaic than that of the two Ambtir
inscriptions of Njipatunga. If it is kept in mind that the
Bahtir plates represent the latter [NripatungaJ as a descendant
not only of Pallava, but also of KoAkani, the ancestor of the
*> Western Ganga kings, we are driven to the conclusion that the
old dynasty of the Pallavas of K&fichi came to an end with
J> Nandivarman, the opponent of the Western Ch&lukya king
Vikramaditya II ; that Narasimha varman, a Pallava by name,
J> but Western Ganga by descent, succeeded them ; that two of
his successors, Dantivarman and Nandivarman, were the con-
temporaries of the R&shtrakftta kings G6vihda III. and Amdgha-
varsha I ; and that Nandivarman's son, NripatuAgavarman or
> Nripatunga - Vikramavarman, who ruled over Noi*th Arcot,
a Tanjore and Trichinopoly, discarded the emblems of the Western
> Gangas and adopted those of the Pallavas.*
The Ganga-Pallava theory that has been imagined by Mr.
Hultzsch is very simple ; it can be summarised in a few words :
Njipatunga is not a descendant of Nandivarman Pallava-
malla ; he is not a Pallava ; he has usurped this title ; the name
Konkanika, found in the genealogy given in the Bahflr plates,
proves that he is a descendant, not of Nandivarman Pallavamalla,
but of certain chiefs of Western Ganga origin.
The whole of the Ganga-Pallava theory is, therefore, based on
the word Konkanika. Was this name quite enough to justify the
creation of a new dynasty ?
Mr. Hultzsch has made a supposition, a hypothesis, based on
a very weak argument, which cannot be admitted unless con-
firmed by other discoveries ; and until this is done, it must remain
what it really is, viz., a simple conjecture.
The inscriptions dated during the reigns of kings like Danti-
varman (e. #., inscription No. 80 of 1898), Nandivarman (e. g.,
No. 72 of 1898), Nripatunga (e.g., No. 81 of 1898), ought to be
classed among the Pallava inscriptions up to the time when the
Ganga origin of Nripatunga and his ancestors is proved in an
irrefutable manner.
Mr. Hultzsch has not had the discretion to do so. In his
report on Epigraphy for 1897 (0.0., Nos. 1093-1096, 29th Aug.
1898) inscription No. 304 of 1897 of Vijaya-Nandivikramavarman
17th year is classed under the head of " Ganga-Pallava ".
So, even when there is not sufficient evidence to confirm this
imprudent theory, it has received official sanction.
The Ganga-Pallava dynasty was ranked with the great
Pallava, Ch6la and P^ndya dynasties.
For that, and that alone, Mr. Hultzsch deserved to be re-
proached.
He had every right to imagine that Nripatunga was of Ganga
origin and write in the "Epigraphia Indica" the sentences we
have quoted.
But he had no right to put during 6 years (from 1897 to 1904) a
great number of inscriptions in the Ganga-Pallava category, as if
the existence of this dynasty had been completely established,
First of all, it must be proved :
54
1 that Nnpatunga was actually descended from a king
named Konkanika ;
2 that the above Koiikanika was no other than the Gariga
king Konkani ;
3 that Njipatunga was not also a descendant of Nandivar-
man Pallavamalla.
M. Hultzsch seems to have thought such proofs useless :
similarity of names, vague suppositions, unfounded identifications
based on imaginary genealogies, all these have formed a respect-
able whole for enabling him to consider the Ganga origin of
Nripatunga as something conclusively proved.
It was not necessary, since then, to prove the Ganga- Pallava
theory ; the hypothesis of Mr. Hultzsch need not be confirmed by
other discoveries : it was true by supposition.
So, we have to lay much stress on the fact that " the Ganga-
Pallava theory has been admitted without any demonstration."
We have so far examined " The Hultzsch theory.'' We shall
now enter into the second stage and examine what we may call
44 The Venkayya theory."
In editing the inscription of Dantivarman found at Triplicane,
V. Venkayya has written (Ep. bid. Vol. VIII, No. 29, p. 291) :
The /inscription is dated during the reign of King
J> Dantivarma-Mah^raja, who was " the ornament of the Pallava
family " and belonged to the Bh^radvaja-gotra. There is thus no
doubt that he belonged to the Pallava dynasty.*)
In his report on Epigraphy for 1905-06 (O. 0. No. 492, 2nd
July, 1906) V. Venkayya mentions inscription No. 541 of 1905
dated during the reign of Dantivarman of the Pallavatilaka
family, which sprang from the Bh^radvaja-gotra : From other
inscriptions we know that the queen of a certain Nandippotta-
raiyan of the Pallavatilaka family lived as late as the time of
the Ganga-Pallava king Nripatunga and the reign of the Ch6la
king R&ja Kesarivarman (Annual Report for 1900-01, paragraph
10). Consequently, the original Pallavas of Conjeeveram seem
to have continued in some form or other long after the defeat of
Nandivarman Pallavamalla by the Western Chalukya Vikra-
m&ditya II., when they apparently ceased to be the dominant
power in Southern India. The political relationship of the
Pallavas to the Ganga-Pallavas, who gradually took their place,
D is not known.
Bo V. Venkayya lays down a principle : When, in an ins-
cription, a king is said to belong to the Pallava dynasty there
is no doubt that he belonged to the Pallava dynasty.
Applyii^g this principle, V. Venkayya affirms that kings like
Dantivarman of Triplicane and TiruveUarai and Nandippdtta-
raiyan " of Pallavatilaka family " are really Pallavas, and that,
therefore, the Pallava dynasty has continued to exist after the
death of Nandivarman Pallavamalla.
If V. Venkayya had been logical, he would have carried his
conclusions further. In the B&hftr plates, it is distinctly stated
that Danti, his son Nandi and his grandson Nripatunga are all
descended from the Pallavas and belong to the Bh&radvaja-g6tra.
If here V. Venkayya had applied the principle laid down by
him, he would have said of Nripatunga : There is thus no doubt
that he belonged to the Pallava dynasty.))
But V. Venkayya could not agree to it. The name Konkanika
found in the genealogy given in the Bahfir plates has been con-
sidered by Mr. Hultzsch as positive proof of the Ganga origin
of Nripatunga. It is not therefore possible to doubt it. The
descendants of Nandivarman Pallavamalla could not therefore be
identified with the ancestors of Nripatunga ; there were therefore
two different dynasties reigning at the same time.
Then, V. Venkayya created a new theory founded on the
following principles :
* 1 The Pallava dynasty existed after Nandivarman Pallava-
malla. Dantivarman of Triplicane and Nandippottaraiyan are
their representatives. They differ from the " Gariga-Pallavas " by
their epithets " P6ttaraiyan," " potavarman " [pota in Sanskrit
and pottu in Tamil mean ' the sprout (of a plant) ' and are thus
synonymous with pallava, "a sprout" /./., Vol. II, p. 341].
" Pallava-kula-tilaka " : "The title Pallava-Kulatilaka of the for-
mer (Dantivarman) which occurs in the Triplicane inscription
might have furnished the family name Pallavatilakakula of his
successors, which must have been invented in order to distinguish
them from the rising Ganga-Pallavas ((?. 0. No. 492 2nd July
1906)."
2 At the same time there existed a dynasty of chiefs of
Ganga origin, such as Narasimhavarman of Kil-Muttugfir, Danti,
Nandi and Nfipatunga of Bahfir. These chiefs have succeeded
58
by degrees in supplanting the true Pallavas. They are dis-
tinguished by the prefix "Vijaya" or the suffix "Vikrama-
varman ".
When V. Venkayya enunciated this theory, there was but
one person who strongly protested against it.
In an article published in April 1907 in the " Christian College
Magazine " under the heading " The Pallavas and the Ganga-
Pallavas" M. R. Ry. T. A. Gopinatha Rao, Superintendent of
Archaeology, Travancore State, Trevandram, has enunciated the
following propositions :
1 The Ganga origin of Nripatunga which has been treated
as a sort of axiomatic truth by later epigraphists (p. 1) J> is a
wrong supposition.
2 There existed but one dynasty, that of the Pallavas ; the
descendants of Nandivarman Pallavamalla must be identified
with the ancestors of Nripatunga.
3 Dantivarman, the grand-father of Nripatunga is the son
of Nandivarman Pallavamalla.
The last hypothesis is ?ery remarkable. M. R. Ry. T. A.
Gopinatha Rao was the first to affirm that Pallavamalla was the
father of Dantivarman. So, the genealogy of the Pallavas stands
thus :
Nandivarman Pallavamalla
Dantivarman
I
Nandivarman
I
Nfipatunga
However, no one cared for what was said by M. R. Ry.
Gdpinfttha Rao. V. Venkayya continued to maintain his own
opinion, and the public, relying on the authority of the Govern-
ment Epigraphist, continued to believe in Gangd-Pallavas.
This was in 1907. Ten years have rolled on since and new
discoveries have only confirmed in a striking manner the theory
of M.R.Ry. T.A. Gopinatha Rao.
1 Although we possess a large number of documents, we
have not as yet found any proof of Nripatunga being related to
Narasiihhavarman of Kil-Muttugfir or to the Gangas.
2 In none of the numerous known inscriptions has it been
possible to find the least trace of any internal struggle enabling
us to believe that the Ganga-Pallavas supplanted the descendants
of Nandivarman Pallavamalla. There has been no civil war, no
revolution, no dissension in the Pallava Kingdom.
3 The inscriptions of the Pallavas and the so-called Ganga-
Pallavas are found scattered all over the country, and it must be
admitted that these two different dynasties reigned over the same
country at the same time :
There would be an insuperable difficulty in locating these
J>two contemporary dynasties in proper geographical regions.
(The Pallavas and the Ganga-Pallavas p. 8) .
4 It has been proved that Nripaturiga had the name that
marks his Pallava origin : The ending "pottaraiyar" which is
applied to [Nripatungapp6ttaraiyar of the Valuvtir inscription
No. 68 of 1908] without the characteristic prefix K6-visaiya,
* makes it suspicious if we could include his name among the
Ganga-Pallavas (G.O. No 538, 28th July 1909).))
Moreover, the discovery of the Velfirp&laiyam plates has
shown (G.O. No. 832, 28th July 1911) that the descendants of
Nandivarman Pallavamalla had the prefix K6-vijaya and the
suffix Vikramavarman added to their names.
The same VglfirpSjaiyam plates have proved the following
genealogy in an incontestable manner :
Nandivarman-Pallavamalla
i
Dantivarman
I
K6-vijaya-Nandivikramavarman.
If we bear in mind, that, in 1907, i.e., 4 years before the dis-
covery of these plates, M. R. Ry. T. A. Gopinatha Rao had affirm-
ed that Dantivarman was the son of Nandivarman Pallavamalla,
we shall see the barrenness of the Ganga-Pallava theory and the
fecundity of the theory propounded by M. R. Ry. Gopinatha Rao.
And still, up to this time, it has been possible to doubt it, as
the conclusive argument in favour of Mr. T. A. Gopinatha Rao's
theory has not been found. What was, in fact, the origin and basis
of the Ganga-Pallava theory ? It was the name Konkanika in the
genealogy given in the Bahfir plates.
Mr. Hultzsch said, and V. Venkayya repeated with him,
Nripatunga is not a Pallava; he is a Ganga because he is
& descended from Konkanika . That was the only reason. M.R.Ry.
-58-
T. A. Gopinatha Rao said that that reason was good for nothing ;
he gathered proofs to try to demolish the Ganga-Pallava theory,
but its supporters retorted, Njipatunga is not a Pallava ; he is
descended from Konkanika.
In publishing the V61tirp&laiyam plates in Pqrt V of Vol. II
of 8.1. 1. that appeared only a few days ago (Madras, 1917),
M. R. Ry. H. Krishna Sastri says of the theory of M. R. Ry.
T. A. Gopinatha Rao that it connects the names in the Bfthtir
plates with those of the VMftrp&iaiyam plates, and suggests
that Vijaya Nripatungavarman of the former was apparently
the son of Nandivarman III of the latter *> (8. L /., Vol. II,
Part V, No. 98, p. 505).
Rao Sahib H. Krishna Sastri then says distinctively :
Against this, the only objection is the ancestry which, in the
one case includes the clear Western GaAga name (or surname)
D Konkanika, while in the other it does not.
To-day (June 1917) this objection does not exist any more :
the deciphering of the V&yaltir inscription has cleared all doubt.
We have said in Chapter II of this book that the Vftyaltir
inscription gives us a complete genealogy of the ancestors of
Narasimhavarman II (Rajasimha).
After Pailava, As6ka, Harigupta, Aryavarman and some
others, we have the following series : Kajinda, By,malla, [E]ka-
malla, Vimala, Konkanika, Kalabhartri, Chtitapallava, Virakfircha.
We have reproduced in PL II the estampage of this part of
the inscription.
We have said that the presence of names of Western Gariga
origin, such as, Aryavarman and Konkanika, in a genealogy of
the Pallavas engraved on stone in the VII century, shows the
political relations and perhaps also the bonds of affinity that
existed between the Western Gaiigas and the Pallavas in the
VI century of the Christian era. The Penugonda plates (G.O.,
No. 920, 4th Aug. 1914, Part II, No. 4, p. 86) confirm this hypo-
thesis, as they say that the grandson of Konkanivarman who was
called Aryavarman was installed on the throne by the Pallava
king Simhavarman and that perhaps he married the daughter of
this Pallava king since we see that his son was named Simhavar-
man. The son of Simhavarman, alias Madhava II, was Konkani-
varman II alias Avinita [SringSri plates Mysore Archoeological
Report for 1916, p. 33.]
Approximate Genealogy of the W. Gangas.
A. D.
420 Konkanivarman (I).
450 Madhava (I).
480 Aryavarman (installed on the throne by Simhavarman,
a Pallava king).
510 Simhavarman alias Madhava (II) (installed by Skanda-
varman, Pallava). (Penugonda plates).
540 Konkanivarman (II) alias Avinita
[Sringeri plates Mysore Report for 1916, p. 33.]
I [E.G. 10, Malfir 72.]
580 Druvinita
(Mujbagal plates Mysore Report for 1916, p. 44.)
[Gummareddipura plates Mysore Report for 1912
dated the 40th year.]
We have proved with the help of the Vayalur inscription that
the names Vimala and Konkanika existed in the Pallava genea-
logy from the VII century.
The presence of these names in the Bahfir genealogy goes
only to confirm the purely Pallava origin of Nripatunga.
Thus then, from the moment that the Vayalfir inscription
was deciphered, the Ganga-Pallavas ceased to exist ; full light
has been thrown on the dynasty of Nripatunga ; the truth is seen
to be clear and simple, and it can be summed up in the following
words : M. R. Ry. T. A. Gopinatha Rao was right in affirming
that
1 in the VIII and the IX centuries there existed but one
dynasty, that of the Pallavas ;
2 the Velfirpajaiyam plates coupled with those of Bahfir
give us the following genealogy :
Nandivarman (Pallavamalla).
I
Dantivarman
Nandivarman (donor of the Veltirpalaiyam plates).
Njipatunga (donor of the Bahfir plates).
-60-
2. Chronology.
Let us first determine the duration of each reign. This ques-
tion presents no difficulties in regard to Dantivarman and Nyipa-
The inscription (No. 262 of 1904) at Tiruchch&nftr is dated in
the 51st year of Dantivarman. That implies a long reign and we
shall admit that Dantivarman reigned little more than 51 years.
It is certain that Nripatunga reigned 26 years, for the in-
scriptions of this king, which are very numerous, do not go
beyond the 26th year. .
The question is not so simple in the case of the two other
kings. We know, as a matter of fact, that both of them bear the
name of Nandivarman.
How to distinguish the one Nandi from the other ? and how
long did each of therti reign ?
I hold a very definite opinion on this subject. I shall now
proceed to prove my proposition, which is quite different from
what has been imagined till now. I am the first to enunciate it
and I strongly affirm it as I am quite convinced of it.
Nandivarman Pallavamalla reigned 62 years. His grandson
reigiued for about 24 years.
My demonstration is based on the Tandantdttam plates.
In the 58th year of his reign, a king Vijaya-Nandivikrama-
varman has made this grant (/./., Vol. II, Part V, No. 99, p. 517).
These plates have been, till now, attributed to the king that
granted the VSltirpalaiyam plates, viz., to the son of Danti. My
opinion is quite the reverse.
The donor of the Tandantottam plates is Nandivarman Pallava-
malla. Here are my reasons for thinking so :
1 From the palceographic point of view :
(a) If the donor of the Tandantdttam plates were the same
as the donor of the Veltirpajaiyam plates, as the first is dated in
the 58th year and the second in the 6th year, the alphabet of the
VSltirpSjaiyam plates must be more archaic: but it is just the
reverse : The alphabet of the Tandantdttam plates is more archaic
than that of the Vglfirpdlaiyam plates.
' In the Sanskrit portion of the Tandantdttam plates, the
letters, ft, kha, ga, da, dha, na, ya, va, &a, sa, ha, nft, mu, td, to,
show signs of archaism, when compared with the corresponding
letters in the V61tirp&Jaiyam plates.
It is the same case with the Tamil letters : na, ta, na, ya, la,
tft, n&, yi, mfi, lai, ko, to.
It must therefore be admitted that the Nandi of the Tandan-
tdttam plates was a predecessor of the Nandi of the Vlfirpa]aiyam
plates.
(b) If, instead of examining the letters separately, we
observe the general aspect of the writing, we shall find that the
Taii^antSttam plates remind us of the K&s&kudi plates, whereas
the VSlflrp&laiyam plates resemble those of Hastimalla found at
Udayendiram. For example, vertical lines predominate in the
TandantSttam plates ; besides that, the secondary i is almost
circular ; on the contrary, the vertical lines are scarcely visible
in the VSlfirp&laiyam plates and the secondary i is semi-circular.
(c) If, instead of comparing the plates, we view them
individually, we are led to the following considerations :
Jf we suppose that the Tandant6ttam plates were dated in the
58th year of the reign of the son of Danti, they would evidently
belong to the end of the IX century.
We know, in fact, that Danti was vanquished about 804 A.D.
by Govinda III ((?. 0. No. 919, 29th July 1912, p. 59, Part II,
No. 7). Granting that this date falls at the end of Danti's reign
and that he died in 805, the 58th year of the reign of his son
would be 805 + 58 -863.
So, if we admit that the Tandantottam plates are dated in the
reign of Nandi, son of Danti, they would be dated 863 or some
years after, i.e., at the end of the IX century. There would then
be a difference of 50 years between them and the plates of B&n&
King Hastimalla (S.H., Vol. II, Part III, plate facing p. 385.)
Any one who has a look at the Tandantdttam plates (8.L /.,
Vol. II, Plate XVIII) will, I think, admit that the alphabet of the
Tandantdttam plates is much too archaic to be attributed to the second
half of the IX Century (about 870).
I think that these plates are one century older and belong to
the second half of the VIII century (about 775).
2 A reign of more than half a century is rare in history ;
and there is no room to doubt that the inscription (No. 10 of 1895)
at Paftchapandavamalai dated in the 50th year of the reign of
-62-
Nandippottarasar belongs to the same reign as the Tandant6ttatn
plates which are dated in the 58th year of the. reign of Nandi.
If then this king was the son of Danti, it must be admitted
that the Panchap&ndavamalai inscription belongs to the second
half of the IX Century. I think It is absolutely impossible to do
so. The alphabet of this inscription is archaic : for example, in
the latter half of the IX century, the letter va began with a
curve ; in the Paiichap&ndavamalai inscription there is no curve
(Vide Plate : Ep. Ind., Vol IV).
3 From the genealogical point of view, we know that there
was only one Pallava prince who was called Hiranyavarman.
The inscription (No. 37 of 1888) on the wall of the southern
ve'randah in the Vaikuntha-Perum&l temple at Raiichipuram leads
us to believe, and the K&sakudi plates say very clearly, that this
prince was the father of Nandivarman Pallavamalla.
And the Taiidant6ttam plates say distinctly that Nandi-
varman, the donor, is the son of Hiranyavarman.
There is therefore no room for any doubt. Nandivarman, son
of Hiranyavarman, who, in the 58th year of his reign, made the
Taiidantottam grant is no other than Nandivarman Pallavamalla.
4 When Nandivarman Pallavamalla granted, by the
kudi plates, the village of Kodukojli, it received the new name of
fikadhiramangalam ; from this we may conclude that Pallava-
malla was surnamed fikadhira.
And Nandi, the donor of the Tandantottam plates, is desig-
nated in verse 4 of these plates by the surname fikadhira.
M. B. Ry. H. Krishna Sastri says : " Dantivarman mentioned
in No. 88 of Appendix C from TiruveJla^ai is reported to have
belonged to the Brahma, kshatra family and to have been entitled
Pallava-Mahar&ja. Evidently, the double name Dantinandivar-
man has to be understood in the sense of Nandivarman, son of
Dantivarman ...... The title Pallava-Mah^r&ja borne by him, may
point to the fact that he was a Pallava king, and perhaps identical
with Nandippdttaraiyar of the Tiruchchennambtindi inscription,
(No. 283 of 1901), " who was victorious at TeU&ru."
This hypothesis is confirmed by the fact that the inscriptions
.of the victor of Tejia^ are written in an alphabet which is not
archaic enough to be attributed to the epoch of Nandivarman
Pallavamalla.
Besides, the Nandikkalambagam gives a description of the
victor of TeUa^u which cannot apply to Nandivarman Pallava-
malla.
And M. B. Ry. T. A. Gopinatha Rao, relying upon the fact
that a certain K&duvetti-Tarailapperaraiyan was a contemporary
of both the victor of TeU^ru and Nripatunga, concluded, as early
as 1907 (Madras Christian College Magazine April 1907, page 8),
that the former (Nandi) was the father of the latter (Nripatunga).
It is therefore probable that the Victor of Tejiaru was the son
of Dantivarman and the father of Nripatunga.
As, on the other hand, we do not know of any inscription of
this king dated after the 22nd year of his reign, we may
conclude :
1 Nandivarman Pallavamalla reigned 62 years.
2 Nandivarman Teliarrerinda reigned only a little more
than 22 years.
As I am convinced that the Gudimallam inscription (No. 229
of 1903) is dated in the 23rd year of the reign of this latter king,
[the alphabet of this inscription mostly resembles that of the
inscription No. 228 of 1903 dated during the reign of Nripatunga]
I think we can give him a reign of about 24 years.
The inscription of Tiruvallam (No. 76 of 1889) being dated in
the 62nd year of the reign of Nandi, we come to the conclusion
that Nandivarman Pallavamalla reigned 62 years.
We have now determined the duration of the reigns :
Nandi- J^kadhira-Pallavamalla : 62 years.
Dantivarman : 51 years.
Nandi-TeJiarrerinda : 24 years.
Nripatunga : 26 years.
It would be desirable to know approximately when these
kings reigned.
We can at once say for certain that they have reigned after
the VII and before the IX Century.
1 Pallavamalla began to reign after the VII century.
In the preceding chapter, we have established, in an irrefu-
table manner, the fact that ParamSsvaravarman I. repulsed the
Chaiukyas in A.D, 674. Here we have a correct date.
-64-
between this and the accession of Nandivarman Pallavamalla
there have passed away :
(a) the end of the reign of Param36varav&rman I ;
(6) the whole of the reign of Narasiihhavarman II. (Rftja-
simha), which was probably long enough ;
(c) the reign of Parameevaravarman II., which was probably
short.
I think these events must have taken 40 years, and so I
conclude that it is unlikely that Nandivarman Pallavamalla
began to reign before 716 or 717.
2 Nripatunga's reign came to an end before 900.
(a) Inscription No. 28 of 1903 is dated in the 24th year of
the reign of Nripatunga, during the reign of his feudatory, the
B&iia king V&navijj&dhara ; and the inscription No. 223 of 1903
is dated during the reign of the son of the Ban a king and in the
6aka year 820, i.e. 898 A.D. We must therefore conclude that the
24th year of Nripatunga's reign is anterior to 898 and that the
26th and last year is before 900.
(b) Inscription No. 735 of 1905 at Gram am assures us that
the Chdia king Parantaka I. began his reign in the first half of
907. (G.O. No. 492, 2nd July 1906, p. 68).
The year 906 is therefore the last year of the reign of his
father Aditya I. But the Tirukkalukkunyam (Ep. Ind., Vol. III.,
p. 277) and the Brahmadesam (No. 230 of 1915) inscriptions are
dated in the 27th year of a king called Rajakesarivarman, who
must unquestionably be Aditya I., for he is the only king named
R&jakdsari who was able to reign so long before R&ja-r&ja. It is
thus certain that Aditya I. ascended the throne about 880.
This king, who was surnamed To^dafim&najEtir-Tunjina-
Udaiy&r gave his donations, in the North Arcot District, in the
21st and the 22nd years of his reign (G.O. No. 503, 27th June 1907,
p. 71, part II, No. 29.) There is no doubt that this king was
Aditya I, for an error that had been made in the donation was
rectified during the reign of his son Par&ntaka I.
It is thus established that the Pallava empire was conquered
by Aditya I. before the 21st year of his reign, i.e., before 900 A.D.
So, it is between 717 and 900 that we must put the four
reigns: Nandi, 62 years; 'Danti, 51 years; Nandi, 24 years;
26 years.
The length of these four reigns put together g.ves a period of
62+51+244-26-163 years, and from 715 to 900 there are 185
years.
I therefore consider the following result as something certain :
The coronation of Nandivarman Pallavamalla
took place between 717 & 737.
da Dantivarman do. 779 & 799.
do. Nandi of TeUftru do. 830 & 85o!
do. Njipatunga do. 854 & 874.
These figures are confirmed by the following two facts which
we shall prove later on :
1 Nandivarman Pallavamalla was vanquished by the Chaiu-
kyas about 741.
2 Dantivarman was vanquished by the Rashtrakfltas about
803.
If we now take the average of the two dates in which these
kings might possibly have been crowned, we find that :
Nandivarman Pallavamalla was crowned in cir. A.D. 727.
Dantivarman do. do. 789.
Nandi of TeU,ru do. do. 840.
Nripatunga do. do. 864.
These figures may be considered to be approximately correct,
allowing for an error of 10 years, more or less. This result
must be considered sufficient.
I believe, however, that these 10 years can be taken away
from the figures I have given above. I give below the reasons
that have led me to this conclusion :
1 The Ambftr inscriptions (Nos. 7 and 8 of 1896 Vide also
Ep. Ind., Vol. IV, pp. 182 & 183) which are dated in the 26th and
last year of the reign of Nripatunga mention Pirudi-Gangaraiyar
as a contemporary of this king.
It is probable that this Pirudi-Gangaraiyar is no other than
Prithvipati I. whose last known date is 879. If we admit that the
26th year of Nripatunga almost coincides with this epoch, we
must admit that Nripatunga ascended the throne about 854, which
is 10 years before the approximate date we have given to it
2 Inscription No. 222 of 1911, dated in the second year of
RajakSsarivarman and found at Tirunag&svaram in the Tanjore
district, has been Attributed by M. E. fty. H. Krishna Saatri to
Aditya I.
Inscriptions Nos. 101, 104, 105, 127, 130 and 133 of 1914 have
been admitted by Diwan Bahadur L. D, Swamikannu Filial (G.O.
No. 1260, 25th Aug. 1915, p. 72, Appendix F.) to be " earlier th*n
907 A.D." and attributed by M. R. Ry. H. Krishna Sastri to
Aditya L (Part II, No. 20, p. 96, of the same Report).
These inscriptions are dated in the 5th, 6th, and 7th years of
his reign and are engraved in the temple at Tiruvejumbur in
the Trichinopoly district.
If, then, we admit that these inscriptions are dated during the
reign of Aditya I., we must also admit that this king was from
the very beginning of his reign, i.e., in 882 (2nd year of his roign)
and 885 (5th year of his reign) the supreme lord of the districts of
Tanjore and Trichinopoly.
But we must also bear in mind that Nfipaturiga reigned $t
Ka^diytir, near Tanjore, in the 21st year of his reign (insc. No. 17
of 1895) and that in the 22nd year he reigned at Lalgudi (insc.
No. 83 of 1892) and at Kdviladi (insc. Nos. 300 and 301 of 1901)
which are situated not far from Tiruvejumbur near Trichinopoly.
Thus then, the districts of Tanjore and Trichinopoly formed
the kingdom of Nripatunga at the end of his reign and of Aditya
at the beginning of the reign of the latter.
My impression is that Aditya I. began to reign after the death
of Nripatunga who ceased to reign about 880. He should there-
fore have ascended the throne in 854, i.e., 10 years before the
approximate date (864) we have fixed for it.
3 The Udayendiram and the Kaskudi plates which are
dated in the 21st and the 22nd year respectively of the reign of
Nandivarman Pallavamalla do not mention the Ch&lukyan inva-
sion that took place about 741 . We may therefore conclude that
those years of that king's reign are anterior to this event, and that
Nandivarman Pallavamalla began to reign before 719-720.
For these reasons, I think that the approximate dates we
have settled must be reduced by 10 years and I therefore fix the
following chronology :
Nandivarman-Pallavamalla : from 717 to 779 A.D.
Dantivarman : 779 to 830
Nandi of TeUa^u : 830 to 854
Nfipatunga: 854 to 880
$ 3. The History of the Pftn4yas,
We have said that, about 741, during the reign of Nandi-
varraan Pallavamalla, there was a Chaiukyan invasion and that
about 803, king Dantivarman became a tributary of the Rashtra-
kfltas.
Apart from these two events, I believe that most of the wars
that the Pallavas had to wage were against the Pa^idyas ; so, in
writing the history of the Pallavas we must speak of the
Pa^dyas.
We shall begin the history of the Pandyas by putting the
genealogy of VSlvikudi by the side of that of the two Sinnamandr
as shown below :
1) Kadungon
2) Majavarman
Avanisujamani
3) 6eliyan ndan-l) Jayantavarman \ Smaller
" | I ISinnamantir
4) Arik^sarin 2) ArikSsarin j [Report for
Ma^avarraan Ma^avarman J 1907]
5) K6chchadaiyan
6) Ter Maran 1) ArikSsarin
Rajasimha Pararikusan
7) Jatila 2) Jatila
Parantaka 1
3) Rajasimha
4) Varaguna
5) Mara, ^kavira
Srivallabha
Parachakrakoiahala .
The first king that seems to be a little known to fame is No. 4
Arikdsarin Majavarman.
The V&lvikudi plates say: He overcame the 'ocean-like
army of VilvSli at the battle of Nelvdli and conquered the army
of the king of KSraJa ; and the Tirutto^dattogai of Sundara-
mflrti Nayanar (verse 8) mentions a king named " Nedumayan,
who was victorious in the battle of NelvSli," who was first a Jaia,
grant ^
[Report for
1908]
Bigger
Sinnamantir
plates.
[Report for
1907]
-68-
was afterwards converted to Sivism by Tirufi&nasambandar and
became one of the 63 devotees. If therefore this saintly king is
identified with No. 4 M&yavarman, he will be the contemporary of
Sambandar.
The latter was a contemporary (Periyapuranam p. 318) of
Siyuttonda - N^yanar who conquered V&t&pi with Narasimha-
varrnan I. -King No. 4, M&javarman would therefore have lived
about 650.
His son, No. 5 Kdchchadaiyan would have lived about 675.
The VJvikudi plates, say of him : He destroyed at Marudfir
the ocean-like army and at the great city of Mangalapuram
the Mah&ratha was overcome and destroyed.* This victory at
Marudtir, this ocean of enemies, this " Mah&ratha ", what are all
these ?
We have saicj just now that Kdchchadaiyan lived probably
about 675. Strange coincidence ! It is precisely in 674 that
Vikram&ditya I. was encamped at Uragapuram, i.e., at Trichino-
poly. And the Kendtir plates (Ep. 2nd. Vol. IX, No 29, p. 205)
say that Vikram&ditya I. fought with the P^ndyas ; we have also
supposed that the P&ndya king was one of the three confederates
who gained the victory at Peruvalanalltir. The "Mah&ratha"
will then be the Ch^lukyan king.
The son of K6chchadaiyan was RHjasimha. Here we find a
new coincidence: we have supposed that Kochchadaiyan and
Paramesvaravarman I. were friends allied together against Vikra-
m&ditya. The son of Kochchadaiyan and the son of Paramsvara-
varman I. both bore the same name of R^jasirhha. This can be
explained by supposing that the Pallava prince R^jasimha (who
was a lion to his enemies, probably the Chaiukyas) fought under
the orders of his father, Paramesvaravarman I., on the side of
K6chchadaiyan ; K6chchadaiyan married the daughter of RSja-
simha and the son of the P^ndya king was given the name of his
grandfather, the Pallava king.
There again we shall find a new coincidence :
The Udayendiram plates (S.I.L Vol. II, part III, p. 376.) say
that Nandivarman, Pallavamalla faced a great danger. He was
attacked by the Pallava prince Chitram&ya and other princes
among whom was the P&ndya king who fought the battle of Mai}-
Xiaikudi. Nandivarman Pallavamalla repulsed his enemies in the
Since Pallavamalla seems to have been an usurper, it is
probable that ChitramHya was no other than the heir to Para-
mSsvaravarman IL, who had called his relative R&jasiihha
Pa^dya to help him.
As Kdchchadaiyan lived at the end of the VII century, his
son Rajasifhha lived in the first half of the VIII century at the
time when Nandivarman Pallavamalla usurped the throne.
Let us now examine the VSJvikudi plates; from the Tamil
portion, we learn that the son of Kochchadaiyan gained a victory
at Mannikurtchchi (perhaps Mannaikudi) over the Pallavas ; and
the Sanskrit portion says that Rajasimha defeated in battle King
Pallavamalla . Here, there is no room for any doubt, as the name
of the king is stated clearly.
Again, the "bigger" plates of !innamanur say that M,ra-
varman ArikSsarin Parankusa subdued the Pallavas at Samkara-
mangai, a village which has been identified by V. Venkayya with
Samkaragr&ma.
These events took place in the first half of the VIII century.
It is also natural to identify Jatila Pariintaka, son of Arikesarin
Parankusa Rajasimha, with Maraiijadaiyan Par^ntaka of the
inscriptions (Nos. 453 and 454 of 1906) found in the Anaimalai
cave which give 769-70 as the exact date of this king.
This identification is justified by the fact that the poet
who composed the Anaimalai verses was called M&rangari,
alias, Madhurakavi Mtivdndamangalapperaraiyan, exactly as the
Ajnapti who composed the Velvikudi record. Both of them lived
at Karavandapura : We may therefore conclude with V. Ven*
kayya that these two personages are but one and the same.
Jatila - Parantaka according to the Velvikudi plates con-
quered the K^dava in battle at Penn^gadam on the southern
bank of the Kav&ri ; the AyavSJ and the Kurumbas at the battle
J> of N&ttukkurumbu.
We know that the Pallavas were also called K^davas.
Jatila is probably the donor of the Madras Museum Plates,
the king mentioned in the Tirupparankunram and the Trivan-
dram Museum stone inscriptions (according to M. R. Ry. T. A,
Gopinatha Rao) and the donor of the smaller 6innamantir plates
whose name must have been mentioned in the missing plates.
70
In the bigger plates of &iQ&amagftr, we find no details either
about the son of Jatila called R&jasimha (who was named after
his grandfather) or about his grandson Varagu^a.
The son of Varaguna was Srimara Parachakrakol&hala. The
Sanskrit portion of Sinnamantir plates say that he vanquished
M6y&~P&ndya, the Keraja, the king of Simhala, the Pallava and
the Vallabha. The Tamil portion says that he was victorious
&t Kunntir, at Singajam, at Vilinam and that he repulsed the
Gangas, the Pallavas, the Chojas, the Kaiingas, the Magadhas
who attacked him at Kudamtikku.
V. Venkayya has written : At any rate the mention of a
aPiindya prince as an enemy of the reigning king implies the
ft existence of conflicting interests in the family. We have also
remarked that the Singhalese are mentioned among the enemies
of &rlmftra. The king of Ceylon joining M&y&-P&ndya enables us
to make an identification :
In chapter LI of Mah&wamsa (Translation of Wijesinha) it is
said (No. 27): Now, it came to pass that at that very time a
* prince of the royal family of Pandu was come hither, having
ft formed a design to overthrow that kingdom because he had been
ft ill-treated by his king.ft
Sena II., king of Ceylon, allied himself with the rebel P&jdya
prince and began to lay siege to the town of Madura.
The king of P^ndyas (No. 38) fled from the field of battle on
ft the back of an elephant, and gave up his life in the wrong place,
ft and his queen also died with him at the same time.ft
According to Mah&wamsa, the king Sena II reigned from
866 to 901.
Here, we have to speak again about the chronology of Maha-
wamsa.
The Chdlas had to wage wars frequently with the kings of
Ceylon and invaded that country several times.
And we know the exact dates of two of these invasions :
1 Parantaka I. invaded Ceylon a little before the 37th year
of his reign (Tiruppalaturai inscr.), about 943.
2 Rdjar&ja conquered the island about 1002.
These invasions are of course mentioned in the Mahftwadisa,
the first under the reign of Udaya IIL and the second under the
reign of Mahinda. V.
But to make the dates given* in the Mahftwaftsa agree with
71
the correct ones, we must reduce the numbers of the Singhalese
chronology by 24 years.
Besides, when we spoke of the invasion of Ceylon by Nara-
siihhavarman L, we saw that the dates coincided, when we took
away a little more than a quarter of a century from those given
in the MaMwartisa.
Subject to this condition, we may safely depend upon the
Mahawaihsa chronology and admit that Sena II. reigned from
841 to 876 and that the P&ndya king who was dethroned by his
relative was probably Srtm&ra Parachakrakoiahala.
We must here note a new coincidence. The Sinnamanftr
plates say that 6rimara encountered the Pallavas in the battle of
Kudamftkku.
There is no doubt that in the time of the ancient P&ndyas, in
the IX century, the town of Kumbhakdnam was called Kuda-
mdkku ; in the inscription of M&anjadaiyan-P9,ndya (No. 13 of
1908) in the N&gSsvara temple at Kumbhak6nam, this town is
called Tirukkudamfikku.
And we have also learnt from the B&hftr plates, that, in the
first eight years of his reign, i.e., from 854 to 863, if our 1 chrono-
logy be correct, Nripatunga gained a victory over the P&ndyas on
the banks of the river Arichit which M. R. Ry. T. A. Gopiiratha
Raa has identified with Ari&iaru that runs near Kumbhakdnara.
We have said that Srimara was a contemporary of Sena II
who reigned from 841 to 876 and that the first eight years of
Nripatunga's reign fall between 854 and 863.
We may therefore suppose that the Pallava Nripatunga
profited by the invasion of the P&ndya kingdom by the Singhalese,
x to march against &rtm&ra who was defeated at Kumbhakdnam.
The alliance of Nripatunga and Sena II. seems to be confirmed by
the B&hflr plates which say that Njipatunga's fame had spread
beyond the seas as that of R&ma.
The Singhalese chronicles say that the P&ndya king was
slain. It was perhaps about 860.
In the foregoing pages we have tried to show the importance
of a knowledge of Pft^dya history for the proper understanding
of that of the Pallavas.
>H
E-i
$
<
<
tm
9
i<
<0
1 1 s
>* P
{j 80 .,H
S ^ S
I 5
i
<g
c*
<oS
P
^4 *'"*
<
a
03
2 ^
^
S cr
^ ^-s
*o
>N f^
a 2
o5 Q
o o
PH
03 <
O <
co
fl "
M
o3
^ O
o
i W)
bO
tf
c8 ,g
^ .S
S CT
^ M
H? H < ^ s ^
c? 9
s~*
c3 ,j
:^7 ^
fl "^^ A N-*' ^J ^^
CO O*
1C
pbd o
J *
s P e g.| I
M ^ ^ ^ <Co CO
^ *S ^2 p
CO CO ^
o : ^
4 1 ft
C>
^
: *o
'd g A c3 ^
i ! i ii i
w 3 pq M
W bd ^ aj
" O*
.* .* ^
o c?
^ ^
1 ^
1
M
S
g
ii
d
3
S g
> <*
cd M
M &0
> 1
S 5
-<rt 'd -
g g o3
J A s
pq QQ p>
Skandava
is
**H
oo c
cfl r^l
^ NQ
JS
H
(-t
oO
X;-
1
Hi
i
00
P*
P ^
4 S
. 00
I
f4
3 33
x<
GQ
CO
P
i
rman
i
1
rman
03
Hiranyavarman
Nandivarman II
Dantivarman (c:
Nandivarman II
Nripatunga (cir,
Bhtmava
Buddhav
03
03
1
G6vinda\
i
S ^.2
.
o 6 M
M
*s
*
mhavarman (
ishnugdpavar
(king of Palal
rhhavarman (
s I
0*
CO
P
co ^"^
! 3 '
i
P
he IX centur
1-4 :"
CQ |>
iH
CQ
I
.!3
^
*S -ti
'S
<4-l
iO
M O
"o
^^,
O
Q
^ ^
I~H
M
M
M
-g
|
M
a g
2 P
M
03
O HH
I i'
B
03
1-4
g g
>
F-l
to co
O 10
iO
L ?->
OQ
'P
2 .3
"6
*M 'S
<0 M
1
03
P 1
03
a 's
< ^
<
8 'S
g
^
r^
pH
^j g
<V
|
03
c
E
o3
M J<
b
"H W
CQ J^
S
a*
02
CHAPTER VI.
ESSAY ON THE HISTORY OF THE DYNASTY
OF NANDIVARMAN.
1. Nandivarman Pallavamalla (about A. D. 717-779).
When ParamSsvaravarman II. died, the succession was pro-
bably disputed. There was perhaps an heir belonging to the
dynasty of Simhavishnu who is called " ChitramSya " in the
Uday&ndiram plates. There was, however, at this time, a Pallava
prince, called Hiranyavarman who claimed to be a descendant of
a brother of Siihhavishnu and had great power in the kingdom.
He was supported by some nobles: one of his friends was the
** Muttaraiyan ", a vassal of the Pallavas, who governed the prin-
cipality of Tanjore which served as a buffer state between the
Pallava and the P&ndya territories (vide : Some insc. of the
Mutt^raiyars by T. A. Gopinatha Rao Journal of the S. I.
Association July 1911.)
Another friend of Hiranyavarman was Udayachandra who
governed Vilvala (probably Villivalam) on the banks of the
Vfigavati, not far from Kaiichlpuram. (UdaySndiram plates).
Hiranyavarman succeeded in having his own son Nandi, then
a minor, crowned at Kafichi. (G.O., No. 492, 2nd July 1906, p. 61).
Chitram&ya, however, managed to obtain the help of many
kings among whom were R&jasiifaha (TSr M&ran), king of the
Pftiidyas who was related to him, and perhaps also the king of the
Majavas (father-in-law of RHjasimha) and the king of the Western
Gangas, a relation of R&jasimha Pandya.
The Pftndyas were overthrown at Sankaramangai and Man-
naikudi (Mannikujichchi) which is probably Manni near Tiruvi-
salur and Id aval (not far from Eumbhakdnam).
The other enemies were beaten back in all directions up to
the confines of the Eastern Chfilukyan kingdom in which Vishjiu-
rftja (Vishpuvardhana III) was reigning (A.D, 709 to 746).
About 741, Vikramftditya II. (733 to 746) led the fourth
Ohftlukyan' invasion to "Tundftka" (Toiidai-MaTidalam).
In the KtadAr plates (Ep. Ind. Vol. IX, No. 29, p. 205), the
Ch&lukyas say that Vikramftditya II. beat and put to flight, at
the opening of the campaign, the opposing Pallava king named
Nandipdtavarman, took possession of particular musical instru-
ments, called ' Kaiumu k hav&ditra and Samudraghdsha, the Khat-
i vdnga-dhvaja, many excellent and well-known intoxicated ele-
iphauts and a heap of rubies, which dispelled darkness by the
* brilliancy of the multitude of their rays, who entered, without
destroying it, the city of K&ficht, which was, as it were, a girdle
adorning yonder lady, the region of the south, who had rejoiced
D Brfthmanas and poor and helpless people by his uninterrupted
> liberality, who acquired high merit by restoring heaps of gold to
* the stone temples of Bdjasimhesvara and other gods, which had
been caused to be built by Narasimhapdtavarman, who distressed
* Pdndya, Chola f Kerala, Kalabhra and other kings.*
The last phrase seems to show that the Pandyas joined the
Pallavas.
The donations made to the Kailsanatha temple at KMchf-
puram by Vikramfcditya II. are confirmed by an inscription
caused to be cut in this temple by this king (Ep. Ind. Vol. III.,
No. 48, p. 359.)
That Nandivarman Pallavamalla reigned for a long time
is confirmed by five documents :
47th year : inscription No. 55 of 1887 at Viriiijipuram.
50th year : No. 10 of 1895 at PanchapSndavamalai.
52nd year : No. 27 of 1887 at 6adupperi.
58th year : Tandanttdttam copper plates.
62nd year : inscription N6. 76 of 1889 at Tiruvallam.
The last inscription mentions that the Bana king Mavali
Vanarayar was a vassal of the Pallavas.
2. Dantivarman (about A. D. 779-830).
The Vglfirpfilaiyam plates (&!./., Vol. II, Part 7, p. 511) tell
us that Nandivarman Pallavamalla had for wife R6v&, who, like
)> (the river) Rv&, had (her) birth from a great king (or, from a
high mountain) J> ; and it was of her that Dantivarman was bonu
The name Danti seems to be of R&shtrakfita origin ; and Rdvft
was probably the daughter of a R&shtrakfita king named Danti.
-76-
This will explain the political relations that subsisted at this
period between the Pallavas and the R&shtrakdtas.
In 780 probably at the beginning of the reign of Danti-
varman the king of K&nchi helped Gdvinda II. in his struggle
against Dhruva. The latter succeeded however * in dethroning
his brother. (Ep. Ind. Vol. III., No. 17 p. 104).
There is no doubt that at this epoch the Pallava kings reigned
not only in K&nchl, but also over the region on the banks of the
K&vrl up to the modern Pudukdtta State. In fact, in the 5th
year of his reign, donations have been made at Kunn&ndftrkoil
(Pudukotta) (No. 348 of-1914) and at Tiruvellarai (near Trichino-
poly) (No. 541 of 1905).
Inscriptions were engraved in the 6th year at Tondur, near
Gingee (No. 283 of 1916), and in the 7th (No. 80 of 1898), 9th
(No. 74 of 1898) and 10th years (No. 51 of 1898) of his reign at
Uttaramallur (Chingleput district.)
The inscriptions at Triplicane (Madras) [No. 234 of 1903 Ep.
Ind. Vol. VIIL, No. 29, p. 291], and at Kdram (No. 35 of 1900) near
K^ncht are dated in the 12th year of his reign.
In the 16th year, Dantivarman was still reigning in the
Muttaraiyar kingdom, for we see that VidSlvidugu Muttaraiyan
declared himself a vassal in two of the Malaiyadipatti inscriptions
(Nos. Ill and IV, p. 23 Some inscriptions of the Muttaraiyars
Journal of S. I. Association July 1911.)
Inscription No. 61 of 1892 at Uttaramalltir is dated in the
21st year of his reign, i.e., in 800 according to our chronology.
It is about this time that Govinda III, having conquered
Dantiga, who ruled over K,nchi, forced Dantivarman to pay him
tribute ( Ind. Ant. Vol. XL, p. 127).
It is probable that this event considerably weakened the
power of the Pallavas. It is, in fact, very remarkable that there is
no inscription dated between the 21st and the 49th years of the
reign of Danti 800-828.
We have two inscriptions dated at the end of Danti's reign.
In the 49th year of his reign, a grant was made to Gudimallam
(No. 226 of 1903) when the Bftna king Vijay^ditya was the vassal
of the Pallavas; and the Tiruchch&nfir inscription is dated in
the 51st year (No. 262 of 1904).
It is remarkable that no inscription dated at the end of
Dantr's reign has been found in the southern portion of the
Pallava empire and particularly in the Tanjore and the Trichino-
poly districts.
The name of Danti is not mentioned in any of the monuments
found on the banks of the K&vri, for a period of 35 years, i.e.,
from the 16th year (Malaiyadipatti) to the end of his reign.
This period extends from 795 to 830 and it is strange that not
a single donation was made during this time. We shall now
proceed to make a remark :
A dozen inscriptions have been found :
No. 17 of 1907
No. 51 of 1895
No. 414 of 1904
No. 413 of 1904
No. 364 of 1907
No. 358 of 1907
No. 13 of 1908
No. 10 of 1899
No. 84 of 1910
No. 105 of 1905
No. 690 of 1905
which seem to prove that the southern part of the Pallava
kingdom was occupied by the P,ndyas, from the 4th to the
16th year of the reign of a P^ndya king called Majaiijadaiyan ; it
appears to have been a military occupation.
It is impossible to suppose that this occupation took place
during the reign of the Pallava kings that succeeded Dantivarman,
for, from the beginning of the reign of Nandi of Teljaru to the
advent of the Chola king Parantaka, we have a series of inscrip-
tions that will not allow us to suppose that the P&ndya invasion
took place at this epoch.
There are, in fact, inscriptions in the Tanjore and the Trichi-
nopoly districts dated in the 6th (Tiruvellarai), 10th {Tillas-
thanam), 12th (Sendalai), 18th (K6viladi and Tiruchchennambtindi),
22nd (Tiruppal&tturai) years of the reign of Nandi, son of Danti ;
in the 7th (Narth^malai), 18th (Kdviladi), 21st (Kandiyfir), 22nd
"(Laigudi and Kdviladi) years of the reign of Nripatunga ; in the
2nd, 5th, 6th, 7th (Tiruverumbur) years of the Chola king, Aditya.
We cannot, therefore, possibly suppose that the Pftndyas
occupied the banks of the KSiveri after the reign of Dantivarman.
So, we think that this occupation took place at the end of
Danti's reign, about 825 ; and it is precisely this date that we
have given to Varaguna when we studi
P&ndyas. It would therefore appear t
whose inscriptions are found on the
Varaguna who united Ch6la and To
lists made from H&l&sya-Mahatmyai
nam].
>f the
an
Perhaps, it was the saate Varagwja that during the reign of
Sena I, i.e., from 821 to 841 (the Mahftvamsa dates being reduced
by 25 years) conquered Oeylon.
Be that as it may, the inscriptions enable us to give a history
of the invasion of the Pallava kingdom.
x The attack began with an expedition into Idavai (insc. No.
690 of 1905) in the Ch6Ja country which can be identified with
the town of the same name in the Manni-nftdu (5.1.7. Vol II.,
part V, Introduction, page 23), i.e., in the same n&du as Tiru-
vifialtlr where there is an inscription (No. 17 of 1907) of Varagufla-
Mah&r&ja. And the Trichinopoly and the Ambftsamudram inscrip-
tions tell us that Varagu^a destroyed Vetnbil which is probably
Tiruvifialur that was originally called Vfemban^r.
All the Ch&Ja country was conquered ; for, the inscriptions
found at Tillasthfinam (No. 51 of 1895) and at Trichinopoly (No.
414 of 1904) that are dated in the reign of M&jranjadaiyan mention
Varaguiia.
All these inscriptions being dated in the 4th year, it is
probable that the campaign was conducted at that time.
Other inscriptions :
4th 4- ? year at Trichinopoly (No. 413 of 1904),
6th year at Idutujai (No. 364 of 1907),
8th year at Adutujai (No. 358 of 1907),
8th year at Kumbhak6nam (No. 13 of 1908),
10th year at Sendalai (No. 10 of 1899),
13th year at TiruveUajai (No. 84 of 1910),
prove that the occupation of Pallava territory was durable.
The Sendalai inscription, methinks, is specially important for
fixing the date of Varaguna. We see that Vid&lvidugu Muttaj-
aiyas lived in the 16th year of the reign of Dantivarman
(Malaiyadippatti inscriptions) and that he made a donation at
Niyaman (Sendalai) in the tenth year of Mfi^anjadaiyan's reign.
That shows well that the invasion of Varaguiia Pfrjdya took
place in the reign of Dantivarman.
The Amb&samudram inscription which is dated in the 16th
year of the reign of Varaguna Mahftrftja (No 105 of 1905) is the
latest in date concerning the occupation of Pallava territory.
It says, that, at this period, Varaguna was encamped in the town
of Araisflr* which is situated in the Togdai-n&Ju, on the banks of
the Pennar.
At this time the Pfttjdyas had overrun half the Pallava
empire and probably were preparing to march on K&iichipuram.
Who was the Pallava king that repulsed the invader? We
shall now try to discover it.
3. Nandi of TeJl&m (about 830-854).
The mother of Nandi, son of Danti, was a Eadamba princess
named Aggalanimmati. The Vldrp|aiyam plates that contain
this information add that he had to fight for the throne of his
father: (V. 20). This (Nandivarman), puffed up with the
prowess of his arms, acquired the prosperity of the (Pallava)
J> kingdom, not easy for others to obtain, by killing (his) enemies
on the battle-field . There is here perhaps an allusion to the
battle of TeU&ju which was fought within the first ten years of
his reign, since the Tillasthanam inscription (No. 52 of 1895)
dated in the 10th year gives Nandi the title of Tejjarrerinda.
Who was the enemy with whom he fought this battle ? Very
probably, it was the P&$dyas. The poem Nandikkalambagam, the
hero of which is TeU^nerinda, tells us, in verse 64, that he was
victorious on the banks of the Vaigai (river in Madura) [Ind. Ant
Vol. XXXVII, page 172].
The town of TeU&ru is undoubtedly the village of Tell&r
which was once the capital of TeUarjrupparru (insc. No. 73 of 1908)
and now forms part of the Wandiwash t&luq of the North Arcot
district. The existence of this town during the Pallava period
hap been proved in Pallava Antiquities Vol. I. page 69 : I have,
in fact* found in this place numerous remains of temples built in
the Pallava style.
Is it rational to suppose that the Pandyas advanced as far as
Teller which is only at a distance of 30 miles south of K&nehi-
puram ?
It is, on the contrary, very probable, if we admit that, at the
end of Danti's reign, the southern portion of the Pallava empire
was invaded by the P^dyas. We know that at the end of this
occupation, Varaguna was encamped at Araisur which is
situated on the Pennar river in Toiidai-mapdalam (Ambft-
samudram in so., No. 105 of 1905).
We may suppose, that, at the accession of Nandi, Varagufla
tried to seize the rest of the Pallava empire and marched on
K&nchi. Nandi stopped the invader at Tell&r, inflicted many
defeats on him, notably at Palaiy&ru (Nandikkalambagam, verse
31) alias Palaiy&rai near Kumbhakdnam and pursued him up to
the banks of the Vaigai.
This glorious campaign in which Nandi earned the surname
of Tell&neginda enabled him to reign peacefully not only at
Kafichl, (insc. No. 12 of 1895) but also on the banks of the K&v6rl.
In the country round Tanjore and Trichinopoly, we find some
inscriptions of the victor of TeU&yu, in the 10th year of his reign,
at Tillasth&nam (No. 52 of 1895), in the 12th year, at Sendalai
(No. 11 of 1899), in the 18th year, at Koviladi (No. 283 of 1901), and
in the 22nd year, at Tiruppaiattu^ai (No. 180 of 1907).
The VSltirp&jaiyam plates tell us that he reigned in the N&yar
region (Saidapet taluk- Chingleput district).
More in the north, he reigned at Gudimallam where the
Bana king Vikramaditya was his vassal : (insc. No. 229 of 1903
and No. 1, a of 1890 S.I.L Vol III., No. 42, page 93).
His reign was also rich in poets. We have already referred
to Nandikkalambagam ; there were also other works :
It was probably during the reign of this Pallava king that
the poet Perund^vanar, the author of a Tamil poem entitled
ft B&ratavenba, flourished, because in the invocatory verses of the
> Uttiy6gaparvam of this poem, he refers to his patron as the
* Pallava king " who conquered his enemies at Tell&jru "
The invocatory verses prefixed to the Tamil works Pu j*a-
ft n&ntiru, Agan^ntiru, Narjinai, Ku^ondogai and Ainguyunfljru are
believed to have been composed by the same Perundvan&r
(G.O. Nos. 922, 923, 19th Aug. 1899, p. 6).
4. Nripatunga (about 854-880).
The B&htir plates give us a minute account of the birth of
this king. Nandi varman (III) had married SaAkha, a princess of
the Rashtraktita family, and their son was N^ipatuAga. This
name having been borne by the R&shtrak(Ha king Amdgha-
varsha I. who reigned from 815 to 879, there is little doubt that
was the daughter of Amdghavarsha I. (Ep. Ind., Vol. IV.,
-81-
The Bfthtir plates tell us that Nfipatunga was young when he
came to the throne and that the beginning of his reign was
glorious, for, he overcame the P^ndyas on the banks of the
Ariohit.
We have already spoken about this event. M. R. Ry. T. A.
Qopinatha Rao has identified this river with the Arisil which
is only another name for the Arasil (Vide 8.I.L Vol. II., p. 52);
and the Arasil or the Arasileiy&ru is but the well-known Arsalar
that flows eastwards near Kumbhakonam and falls into the sea at
Karikal.
We have already said that the adversary of Nripaturiga was
M&ravarman Parachakrakolahala who encountered the Pallavas
at Kudamukku (Kumbhakonam), on the banks of the Arichit
The B&htir plates compare the glory of Nripaturiga to that of
Rama which calls to our mind the island of Ceylon ; we have
already mentioned that Sena II. and MHya-P&ndya were allied
with Nripatunga and that it was at the end of this campaign
(about 861) that Madura was taken.
The Bfthtir plates belong to the 8th year of his reign, and, if
our chronology is correct, their date must be about 862 shortly
after the capture of Madura.
Nripatunga seems to have been a peaceful and glorious king,
as a great many inscriptions dated during his reign are found
scattered all over the Pallava kingdom from TiruvaiangH-du
(Nos. 460 and 461 of 1905) and Tiruvoniy^r (No. 162 of 1912) in
the north, to Kandiytir (No. 17 of 1895), Koviladi (Nos. 303, 301
and 300 of 1901) and L&lgudi (No. 84 of 1892) in the south. Besides
this, the Gudimallam (No. 228 of 1903) and the N&rtMmalai
(No. 365 of 1904) inscriptions show that the Bna king [Vidy&-
dhara] in the north and the Muttaraiyan king [6&ttam-Paliyili]
in the south were his vassals, so that Nripaturiga's domination
extended from the river Swarnamukhi in the north to the
Southern Vellaur in the south, i.e., from K&lahasti to Pudu-
kdtta.
Again, as theae inscriptions are dated at the beginning ad
well as at the end of his reign, we have to conclude, that, during
all the 26 years of his reign, the glory of Nripatudga was never
5. Aparftjita.
We have a dozen inscriptions in which Apar&jita is men-
tioned ; he was a Pallava king, since, in inscriptions NOB. 159
and 190 of 1912, we find the title " P6ttaraiyar " affixed to his
name.
These inscriptions are dated between the 3rd and the 19th
year of his reign.
They are :
3rd year No. 351 of 1908 at Mftng&du,
4th year No. 158 of 1912 at
4th year No. 161 of 1912 at
4th year No. 31 of 1912 at SatyavSdu,
5th year No. 32 of 1912 at SatyavSdu,
6th year No. 190 of 1912 at Tiruvoyyiyftr,
7th year No. 163 of 1912 at Tiruvoniyftr,
8th year No. 159 of 1912 at Tiruvojyiyfir,
12th year No. 180 of 1912 at Tiruvo^iyftr,
18th year No. 435 of 1905 at Tiruttani,
18th year No. 433 of 1905 at Tiruttani.
There is also an inscription dated in the reign of Tribhuvana-
vira-d6va found at Uttaramallur which refers to the 19th year of
the reign of AparSjita.
We may add here that the village of Ukkal (near M&mandftr)
bore the name of Aparftjita-Chaturv^dimangalam (/. /., Vol. Ill,
P. 2).
We know with certitude of two events that took place in the
reign of Apar&jita :
1 The battle of Sripu^arnbiya (Tiruppirambiyam).
The UdaySndiram plates tell us that the PS^dya king
Varaguna fought at Sripugambiya with king Apar&jita and his
ally, the western Ganga Prithvlpati I. (Q. 0. No. 492, 2nd July
1906 Part II, No. 9, p. 64). The latter was slain in that battle.
As the last inscription of the reign of Pj-ithvfpati I. is dated
879, it must be admitted that the battle of Sripuyambiya took
place after this date, i.e., after 879.
Varaguna has been identified with a P&ndya king of that
name who ascended the throne in 862-63 (insc. No. 705 of 1905).
The Chdjas were perhaps the allies of Varagu^a, for, inscrip-
tion No. 337 of 1912 seems to mention the P&^dyas and the
Chftjas as the enemies of the Western Ganga Priduvayya.
2 The downfall of the Pallavas and the conquest of Aditya.
The Tiruvftlarigadu plates say (Verse 49) that the Ch&la king
Aditya I. defeated the Pallava king Aparftjita in battle and took
possession of his kingdom (G. 0. No. 492, 2nd July 1906, p. 65).
Inscriptions Nos. 286 and 287 of 1911 seem to show that, in
this campaign, Aditya, had as his ally the Chera king Sth&nu-
Ravi (G.O. No. 919, 29th July 1912, Part II., No. 11, p. 61).
We have said above that Tondai-Mandalam was conquered
before the 21st year of the reign of Aditya, i.e., before 900. A.D.
(G.O. No. 503, 27th July 1907 Part II., No. 29, page 71).
It is therefore incontestable that Apar&jita lived at the end
of the IX century and that he was the last king of the Pallava
dynasty.
There is another interesting question : What place does
Aparftjita hold in the Pallava genealogy ?
There are two theories concerning this subject :
1 V. Venkayya (G.O. No. 492, 2nd July 1906, Part II.,
No. 9, page 64) has supposed that Aparajita was the son of
Nripatunga.
In that case we must admit that Nripatufiga was killed in the
battle of Sripurambiya and that his son who was present at the
battle succeeded to the throne and assumed at the same time the
title of Apar^jita (the unconquered.)
The provinces of Tanjore and Trichinopoly would have
remained in the possession of the Ch6Ja prince, who, at the battle
of Sripur&mbiya, was crowned king of Chdjas under the name of
Aditya I.
But Tondai-Mandalam continued to belong to Apar&jita who
reigned there for about 20 years from 880 to 900 A.D., and it was
only in the 20th year of his reign and that of Aditya that the
latter succeeded in taking it with the help of the ChSras.
In support of this theory, we may point out that the inscrip-
tions of Aparajita are found only in the northern parts of the
Tamil country, at Mangftdu, Tiruvoniyfir, Satyavgdu, Tirutta^i,
Uttaramallur, and none are known to exist in the Tanjore and
Trichinopoly regions.
-84-
On the contrary, we know from inscriptions No. 222 of 1911
(2nd year at Tirun&g66varam) and NOB. 101, 104, 105, 127, 130,
133 of 1914 (5th, 6th and 7th years at Tiruveyumbur) that Aditya
I. held sway on the banks of the Kverf in A.D. 882, 885, 886, 887.
On the other hand, It is not possible to explain why Tiru-
vojjriytir, in which so many Gariga-Pallava epigraphs have been
discovered, should not contain any record belonging to the reign
of R&jakgsarivarman (Aditya I). Perhaps, the northern corner of
Tondai-Mandalam in which Tiruvoniyfir is situated had not yet
fcbeen completely brought under subjection by him
D (G.O. No. 961, 2nd Aug. 1913; Part II. No. 18, p. 94).
It must, however, be noted that the Tiruvoniyfi* temple
contains 6 inscriptions dated in the 4th, 6th, 7th, 8th, and 12th
years of the reign of Apar&jita while there is none belonging to
Aditya.
2 M. R. Ry. T. A. Gopinatha Rao (the Pallavas and the
Ganga-Pallavas Madras Christian College Magazine April
1907) has put forth the theory that the name AparSjita was but a
pseudonym of Nripatunga.
In favour of this, we may say that in case Apar&jita was no
other than Nripatunga there would have been only one battle,
that of Sripurambiya.
If, on the contrary, AparSjita is supposed to be the son of
Nripatunga, there must have been two conflicts : 1 the battle of
6ripurambiya about the year 880 where AparSjita would have
defeated the P&ndya king Varaguna ; 2 the conquest of Tondai
by Aditya about the year 899 in which AparSjita would have
been himself defeated.
It is to be hoped that new discoveries would be made that
might go to confirm one or other of these two hypotheses.
CONCLUSION.
In the preceding pages, we have tried to increase our know-
>dge of the history of the Pallavas.
No doubt, there are still many points that are obscure, many
Btails that must be stated with precision, but I am convinced that
le main facts are, after all, known to us.
1 The Pallavas before Simhavishnu.
A large number of copper plates (Mayidavdlu, Chendaltir,
ihg&du, UdaySndiram, Chtira, etc.) have proved the ancientness
f the Pallaya family ; but these records mention isolated dynasties
r hich it was difficult to connect with one another ; and the rela-
onship existing between these kings and the dynasty of Siihha-
ishnu was completely ignored. The VSyaltir inscription has
iven us the key to most of these mysteries.
We have assumed that the first Pallava who became king
wed his position to his marrying the daughter of the AndhrS king
awards the end of the 2nd century of the Christian Era.
The VSyalftr inscription seems to indicate that kings Skanda-
irman - Kum&ravishnu - Buddhavarman, whose names have been
jvealed to us by the Chendaltir plates, must have reigned before
ishnugdpa, the adversary of Samudragupta, about 339 A.D.
The UdaySndiram plates give us the genealogy : Skanda*
imha Skanda Nandi ; but these plates have been looked upon
ith suspicion. It was not known if they were a copy of an
uthentic record or a forgery committed by an impostor.
The Vftyalfir inscription that has given us an absolutely iden*
cal pedigree has proved the authenticity of the Udayfindiram
enealogy.
Besides, the discovery, in the Vfiyaltir inscription, of thto
iries, Vira-Skanda-Siihha, which is identical with Uruvupalli
enealogy, proves that Simhavarman who reigned at the time of
tie Uruvupalli grant was but the grandfather of Nandi of the
Idayendiram plates.
Lastly, the V&yaldr inscription, in which we find the series
i&havarman - Vishiiugdpa - Simhavarman - Sirhhavistmu, is the
88
first record that gives a correct list of the names of the immediate
predecessors of Simhavishnu.
The deciphering of the V&yalftr inscription has given us
results which are of capital importance for the history of the
predecessors of Sirhhayishnu,
2 The dynasty of Simhavishnu.
The important discovery made in the State of Travancore of
the work called Mattavil&sa prahasana , and the reference to
this comedy in the M&maijdftr inscription have thrown much
light on this portion of history.
There is no longer any room to doubt that it was Mah6ndra-
varman I., son of Simhavishnu, that cut the caves of M&maQdftr,
Vallam, Dalav&nfir, Slyamangalam, Mahdndrav&di, Pallftvaram,
Trichinopoly.
So, we now possess very precise information about the saintly
king Mahendravarman I., poet, musician, architect and adminis-
trator, who introduced in South India the art that obtained on the
banks of the Krishna and gave a new impulse to religion and
literature in the Tamil country.
In Pallava Antiquities Vol. I, I have asserted, that, to
determine the age of the sculptured rocks of Mahftbalipuram, it
is absolutely necessary to study the whole of the Pallava art.
In making this study, I have shown that the style of the
rathas and caves of Mahftbalipuram belongs to the stage of
transition from the style of MahSndra to that of R&jasiifaha ; there
remained however a doubt : Was the MahSndra of Trichinopoly
and Pall&varam the same as Mah&ndravarman I., son of Simha-
vishiiu ?
Now that all doubts on this point have been cleared, we can
definitively attribute to STarasiifahavarman I, most of the* sculp-
tures in the rock at Mah&balipuram.
These works that were continued during the reign of Para,
mfifivaravarman I., were probably suspended at the time of the
Chftlukyan invasion which took place in 674, when Vikra-
m&ditya I., was encamped at Uragapuram (Trichinopoly).
3 The dynasty of Nandivarman.
There was no history of the Pallavas for the VIII and the
IX centuries. The one name " Koiikaiiika " found in the Bfthflr
plates had led the historians completely astray. M. R. By. T. A.
Qopinatha Rao was the only one that understood the question
clearly : he had affirmed that the Ganga-Pallava theory was quite
wrdng.
The defenders of that theory, however, could still argue that
the name Kotikanika found among tlie names ~of the ancestors
of Nripatufiga seemed to prove th^t he was not a descendant
of Nandivarman Pallavamalla Hii4 of the princes of the Western
Ganga family. ^ ^
- The d^dtpRering of the Vfiyalfir inscription has completely
elucidated the question.
This inscription, which is dated in the VII century and during
the reign of a king belonging to the dynasty of Simhavistmu,
mentions Vimala, Korikanika among the ancestors of the Pal-
lavas.
So, the Vfiyalftr inscription has removed all doubts and the
genealogy of the last Pallava kings now stands established in a
definite manner.
h June. 1017. G. JOUVEAU-DUBREUIL.