446 Journal of the Asiatic Society of Bengal. [July, 1911. 
There is a caste in the Sambalpur tract called Sudh. This 
term is supposed to be a contraction either of the term Sidra 
or of the word Suddha (pure). There are two sections of the 
Sudh people, namely, the Butka Sudhs and the Bad Sudhs. 
The Butka Sudhs are treated still as an aboriginal tribe and 
are not touched by the high class Hindus. But the Bad (big) 
Sudhs are allowed to offer water to the Brahmans. The 
Dumils interdine with the Bad Sudhs, but the Sudhs and the 
Dumils do not intermarry. This shows that the Dumals and 
the Sudhs are akin to each other, while the Sudhs must be 
supposed to have once belonged to the tribe of the Butka 
‘adhs, who are considered to be of low origin. 
Even where the Dumils have their temples (called by the 
Telugu name gudi by all the Hindus of the Sambalpur tract) 
for their goddess, they fix in the earth two pieces of wood, one 
to represent Khamsiri or Khambesvari and the other to 
represent Parmasiri or Paramesvari. For the Paramesvari @ 
piece of rohint wood is obtained The word rohint is in the 
feminine gender, and it means red-coloured goddess. The 
wood rohini is Indian red wood which is known to the 
Botanists as soymida febrifuga. The Brahman priest wor- 
ships the Paramesvari for the Dumals, while the Dumals 
themselves worship their Khambe§vari. : 
It is difficult to say whether the KhambeSvari has come 
over to the Diimals from the home of the Kandhs. The 
Aryan form of the name points to a time of Hindu or Hin- 
duized influence both over the Dumals and the Kandhs, 
at least in the translation of the name of the goddess. e€ 
sacrificial post of the Kandhs is also known to be of black 
wood. Regarding this, however, I have not got yet very 
satisfactory information. 
I now relate another account of Khambesvari, though I 
cannot assert whether the KhambeSvari, I am going to describe, 
has any connection or not with the goddess of the Dumals. 
family goddess of some Rajas whose copperplate grants were 
published by Babu Manomohan Chakravarti in the ‘J ournal of 
is that given by the author, and that the plates were not aval 
able for comparison. Since Babu Manomohan Chakravartt 18 
not himself sure whether his reading is correct all throughout; 
Pies us one half foot and one full foot of the Indravajra verse 
aine If the text 
could be carefully read in the light of those meters, reconstruc- 
