Chap. VII.] THE FIXE AKTS. 489 



exhibit a very slight advance beyond the rudest attempts 

 at excavation. 



On examining the cave temples of continental India, 

 they appear to exhibit three stages of progress, — first 

 mere unadorned cells, hke those formed by Dasartha, 

 the grandson of Asoca, in the granite rocks of Behar, 

 about B.C. 200 ; next oblong apartments with a veran- 

 dah in front, hke that of Ganesa, at Cuttack ; and lastly, 

 ample liaUs mtli colonnades separating the nave from the 

 aisles, and embelhshed externally with facades and agri- 

 cultural decorations, such as the caves of Karh, Ajunta, 

 and EUora.-^ But in Ceylon the earliest rock temples 

 were merely hollows beneath overhanging rocks, like 

 those still existing at Dambool, and the Aluwihara at 

 Matelle, in both of which advantage has been taken of 

 the accidental shelter of rounded boulders, and an en- 

 trance constructed by applyuig a fagade of masomy, de- 

 void of all pretensions to ornament. 



The utmost effort at excavation never appears to 

 have advanced beyond the second stage attained in 

 Bengal,— a smaU cell with a few columns to support a 

 verandah in front ; and even of this but very few exam- 

 ples now exist in Ceylon, the most favourable being 

 the Gal-wihara at PoUanarrua, which, according to the 

 Eaj avail, w^as executed by Prakrama I., in the 12th 

 century.'^ 



Taking into consideration the enthusiasm exhibited 

 by the kings of Ceylon, and the munificence displayed 

 by them in the exaltation and extension of Buddhism, 

 their failm^e to emulate the labours of its patrons in India, 

 must be accounted for by the intractable natm^e of the 

 rocks with wdiich they had to contend, the gneiss and 



^ See Fergttsson's Uhidrations of 

 the Rock-cut Temples of India, Loud. 



1845, and Handhooh of Architecture, 

 cli. ii. p. 23. 



^ Mahuicumo, cli. Ixxvii. 



