1898.] M. M. Chakravarti —Language and literature of Orissa. 377 
one version of the royal geneology, while there are five or six versions 
in the Madala Patiji. The truth often lies in one or other version. 
Appendix IY. will give some idea of the difficulties in constructing a 
royal geneology from these different versions. 
At present the Madala Panji is kept in two sets, one by the Sebaka 
entitled Deulakarana, the other by the Sebaka named Tarha-u. The 
Deulakarana (lit. the writer of the temple) as his name implies, seems 
to have been the original keeper of the records, but a second was added 
apparently for check and for better preservation of the informations. 
It is not known when these records began to be kept. But from the 
fact that a large number of details dates from the time of Anaqga 
Bhima Deva, the system would appear to have been established by him. 
As a rule the later the accounts, the more reliable they are. 
N.B .—Since writing this article, a kind letter of Dr. Fleet has 
drawn my attention to his remarks on the Mada]a Panji in the Epigra- 
phia Indica. 1 Dr. Fleet has analysed the earlier list of kings carefully 
and comes to the conclusion that the annal is “absolutely worthless 
for any purposes of ancient history.” (p. 335). I came nearly to the 
same conclusion when T was studying the original manuscript of the 
Madala Panji. In an article read before this Society, an abstract of which 
was published in the Proceedings for July 1892, I noticed the general 
unreliability of the historical portion of the Madala Panji, and remarked 
that at best it can be used only as a corroborative document. Dr. Fleet’s 
analysis confirms me in that view. 
On some of the points discussed in the very interesting article of 
Dr. Fleet, I venture to differ. Firstly, he thinks that in respect of 
most of the Ke 9 arl kings it can only be said that “ not one of them 
has any ring of antiquity in the sound of it; they may possibly be 
real names of later rulers, misplaced in order to make out a consecutive 
chronological series.” (p. 336). I know at*least of one Orissan inscrip¬ 
tion not later than tenth century, which mentions K^ari kings. It 
is the Brahme 9 vara inscription of Bhuvane 9 vara. 2 This inscription 
mentions Udyotaka Ke 9 ari, and of his ancestors Janamejaya, Vicitravira, 
Candihara and Kolavati. The names of both Janamejaya and Kolavati 
are to be found in the Madala Panji. The K^ari kings need not 
therefore be considered as myths or later rulers, as suggested. 
Secondly, Dr. Fleet has “no substantial doubt that the Yavana 
invasions which were repulsed, as the annals say, by Vajradeva and his 
1 “ The Records of the SdmavamQi kings of Katak.” Ep. Ind., Vol. III., pp. 
334-340. 
2 Prinsep, Journ. As. Soc. Vol. VII., p. 558 et. seq ,; 1. c. Dr. R. L. Mittra s 
Antiquities of Orissa, Vol. II. pp. 88-9, 
J. I. 48 
