186 



DESCRIPTIVE REMARKS ON 



35-h. Turning now to the niches on the south end wall the 

 first (or easternmost) niche contains the figure of a tall-capped 

 man with the cord over his left shoulder, holding with his left 

 hand the right arm of a lady who is followed by another. This 

 is said to be king Harisekhara (see Carr, p. 211), who is men- 

 tioned also in the Sthalapurana (Carr, p. 178). 



35-k. The remaining niche occupies the west part of the south 

 end of the main chamber, and contains a standing four-armed 

 figure with three heads, each crowned with a tall topi. The 

 figure wears the cord over left shoulder, and in this case the 

 right lower hand rests against the right thigh whilst the left 

 (lower) hand is raised. 



The superhuman hands bear emblems, one of which resembles 

 a disc or ring surrounded by knobs, attached to a longish handle. 

 The other might be a vase or a flower at the end of a tapering 

 handle. 



Over the niches a prominent cornice has been left carved out 

 of the native rock, of the usual convex curvature ornamented with 

 a couple of the upright horse-shoe facets, each containing a human 

 face in bas-relief. 



These and the style of the pillars, &c, are so like those of the 

 adj acent caves and monoliths (Nos. 15, 16, 24 and 25) for instance, 

 that they would all seem to be of nearly the same age. But not 

 so the additions. Without actual inspection it is difficult to 

 say for certain ; but it appears that the open front of the original 

 excavation was first bricked up by a wall and that the outer 

 structural Mandapa of stone was added subsequently, as has been 

 done also in the case of the Krishna Mandapa No. 15, the 

 pillared hall in front of the great Pastoral Bas-relief. 



The Tamil inscription in the niche cut in the face of the rock 

 on the north side of the original excavation, recording and describ- 

 ing grants of land in favour " of the Alvar in the temple of Sri 

 Paramesvara Mahavaraha Vishnu , in Mamallapuram or Jananatha- 

 pura," which has been found to belong to the year 1073 A.D., 

 proves that the original work is older than this date and also 

 that the additional wall by which a part of it has been covered 

 up is considerably more recent still, for it is not likely to have 



