p. 113

             (a) Kantakasela = K£\£h£n£\£e£l£m£m£h£f£\

   In recent years explorations of great importance have been
conducted  on a Buddhist  site  in the  Palnad  taluk  of the
Guntur  district  of the Madras  Presidency, lastly under the
superintendence of Mr. A. H. Longhurst, of the Archaeological
Survey of India. The site in question which comprises several
ancient  mounds  is situated  in the midst of wooded hills on
the right bank of the river Kistna or Krishna, the Kannapenna
or Kannavanna  (Skt.  Krishnavarna)  of Pali literature, at a
distance  of some 15 miles from Macherla and on the border of
the Nizam's  dominions.  One of those mounds  is known by the
name of Nagarjunikonda.  Mr.  Longhurst  claims  it to be the
most important Buddhist site hitherto discovered  in Southern
India.

   The discovery  of several.  ruined stupas and monasteries,
of remarkable  pieces of sculpture in a late Amaravati style,
and  of  numerous  Brahmi  inscriptions  fully  confirms  Mr.
Longhurst's estimation. The inscriptions, more than thirty in
number and all composed in Prakrit, refer to the same Ikkhaku
(Skt.  Ikshvaku)  dynasty  which  is also  mentioned  in  the
Jaggayyapeta  inscriptions  discovered  by  Dr.   Burgess  in
February, 1882.  On  palaeographical  evidence  they  may  be
assigned to the third century of our era.  A paper containing
transcripts and translations  0f these interesting records of
Buddhism   will  shortly  be  published   in  the  Epigraphia
Indica.(1)

   In the present  note I only wish to draw attention  to one
point which relates to the ancient topography of Southern

  1.  A preliminary account of the discovery will be found in
the Annual  Report  on South-lndian  Epigraphy  for  the year
ending  31st March, 1926, Madras, 1926, pp.  4 and 92 f., and
for the year ending 31st March, 1927, Madras, 1928, pp, 71 f.
Some of the statements  made here regarding  the contents  of
the inscriptions  require  correction  in the  light  of more
minute  study.   Cf.   also  Annual  Bibliography  of  Indian
Archaeology for the year 1926, Leyden, 1928, pp. 14-16.

                           p. 114

India.  Among the Prakrit  inscriptions  found on the site of
Nagarjunikonda  there are two of considerable length, each of
which was incised  on the stone  floor of an apsidal  shrine.
One of these two inscriptions  is of peculiar  value  for the
ancient topography.  It records that a chetiyaghara-evidently
the apsidal  temple in question--with  a floor of stone slabs
and  with  a cetiya  had  been  founded  by an upasika, named
Bodhisiri, on the Siripav[v]ata (Skt.  Sriparvata),(1) on the
east side of Vijayapuri in the monastery at Culadham[m]agiri.
Besides,  the  inscription   enumerates  a  number  of  pious
foundations  which were due to the same donor.  Now among the
latter we find the following: Kamtakasele(2)  mahacetiyas[s]a
puv[v]adare  selamamdavo, "at Kantakasela  a stone shrime  at
the eastern gate of the Great Cetiya (Skt. Caitya)."

   There can be little doubt that the locality indicated here
by  the  name  of  Kantakasela  must  be identical  with  the
K£\£h£n£\£e£l£m£m£h£f£\ £`£g£ko£lo£h mentioned  by  Ptolemy
(vii, 15)(3) immediately  after  the  mouths  of the Maisolos
River. It follows that this river has been rightly identified
 with the Kistna.

   Several  of the  Nagarjunikonda  inscriptions  refer  to a
Mahacetiya, the ruins of which are represented by a mound now
called Ubagutta or " Owl Mound ".  In all probability this is
the  same  stupa  which  is  mentioned  in  connection   with
Kantakasela.

  As to the exact position of Kantakasela we shall have to
await the further results of Mr. Longhurst's excavations. If

  1. According to Tibetan tradition Nagarjuna spent the last
part of his life in a monastery called Sriparvata.
  2.  The vowel-sign  over  the s has the appearance  of an o
stroke.  But in these inscriptions the rendering of the vowel
marks is far from accurate. Moreover, if we compare the names
of  other  localities  which  occur  in  this  passage,  viz.
Culadham[m]agiri,  Mahadham[m]agiri,  Devagiri, Pu[p]phagiri,
and Puv[v]asela, there  can be little  doubt that the correct
form must be Kantakasela, and not sola.
  3.  Variant readings  are Cantacasila, ssilla, Canticosila,
and Cantacosyla.  Cf. Louis Renou, La geographie de Ptolemee.
L'Inde (vii, 1-4), Paris, 1925, p. 8.]

                             p. 115

we are right in identifying it with Ptolemy's K £\£h£n£\£e o-
£m£m£h£f£\ we may be sure that this place was situated on the
right bank of the Kistna  and at a considerable  distance  up
that  river.  Ptolemy  calls  it  an  E'urropiov, i.e.  "  an
authorised  sea-coast mart ".(1) We may assume that it was an
important  port in the second century  of our era when such a
vivid  trade  was carried  on between  the Roman  Empire  and
Southern India.  This sea-borne commerce, testified by hoards
of gold coins of the Roman Emperors, accounts  for a thriving
population of merchants at Kantakasela and indirectly for the
existence  of the great  monuments  which  once adorned  that
place.  For it was especially  among  the wealthy  commercial
classes that the Buddhist religion found many devotees.

                     (b) THE BINDUKA RIVER

   In Richard Schmidt's Nachtrage zum Sanskrit-Worterbuch  in
Kurzerer  Fassung  von  Otto  Bohtlingk, an  important  lexi-
cographical  publication  which  has lately  been brought  to
completion.  we find on page 133 the following entry: "Kandu-
Kabinduka f. N. pr. eines Flusses, Festgr. 16."

   I do not know which " Festgruss  " is the one referred  to
by the author. Anyhow, the name Kandukabinduka is, I believe,
due to an error  made  by Buhler  in editing  the  two Sarada
prasastis found in the temple of Siva-Vaidyanatha at Baijnath
(the modern  form of Skt.  Vaidyanatha)  or Kiragrama  in the
Kangra district of the Panjab.(2)

 The river-name was supposed by Buhler to occur in Prasasti
ii, verse 10;  which he transcribed and translated as follows
(3):-
sailasyankac calitva ruciranavayah Khelativa sahelam Kulya
Kanyeva yatra sphuradura(4)-lahari Kandukabindukakhya  Kira-

1. E. H. Warmington, The Commerce between the Roman Empire
and India.  Cambridge, 1928, p. 107.
2. Cf. Cunningham, Arch. Survey Report, vol. v, pp. 178 ff.;
plates xliii and xliv, and Fergusson, Hist. of Inian  and
Eastern Architecture, revised edition, London, 1910, vol. i,
pp. 297-301.
3. Ep. Ind., vol. i, pp. 97 ff.
4. Evidently a misprint for -uru-.

                             p. 116

gramo 'bhiramo gunagananilayo  vartate 'dhitrigartam  so 'yam
rajanakena prabalabhujayuja raksito Laksmanena. " There is in
Trigarta  the  pleasant  village  of Kiragrama, the  home  of
numerous  virtues, where  that river  called  Kandukabinduka,
leaping  from the lap of the mountain, with glittering  waves
sportively  plays, thus  resembling  a bright  maiden  in the
first  bloom of youth (who jumping  from the lap of the nurse
gracefully  sports).  That  (village)  is  protected  by  the
strong-armed Rajanaka Lakshmana."

   Buhler's  reading of the verse is unobjectionable, but the
word  binduka  must, in  my opinion, be  connected  with  the
preceding  and not with the following  compound.(l)  In other
words, we ought to read the second pada:
kulya kanyeva yatra sphuradurulaharikanduka Bindukakhya.  The
river  is compared  with a playful  maiden, and the waves  of
that  river  are likened  with  the playing-balls  which  she
tosses  up  and  down.   We  would,  therefore,  propose  the
following  rendering  of the passage in question: "Where that
river called  Binduka, leaping  from the lap of the mountain,
with sparkling  wide waves resembling  playing-balls  merrily
plays, like a bright maiden in the first bloom of youth."

   It was rightly recognized by Buhler that the river so well
described  by the poet  is the modern  Binnu, on the left  or
east bank of which the village of Baijnath is situated. It is
one of the feeders  of the  Bias  (ancient  Vipasa  or Vipas)
which, flowing  through deeply cut river-beds, have given the
hill- district of Kangra its ancient name of Trigarta.

   From  the  modern  form  "Binnu"  it is evident  that  the
ancient   name   of  the  river   was  "Bindu(ka)"   and  not
Kandukabinduka.  The forms Binoa and Binwa  used by Moorcroft
and  Cunningham  respectively  do not agree  with  the  local
pronunciation.  The Kangra District  Gazetteer  in its latest
edition (Lahore, 1926, p. 10) has "Binnun".
------------------------------
1. Cf. Annual Report Arch. Survey of India for the year 1905-6,
pp. 17 ff., plates v and vi. The correct date of the inscrip-
tions must be Saka 1126. corresponding to A.D. 1204.