280 Essay on the Avian Order of Architecture. [Sept. 



must have been the loftiest edifice, not only in Kashmir, but in India. 

 The width is 82-§- feet : its height therefore would have been about 165 

 feet, or perhaps a few feet less, being considerably more than twice 

 that of Marttand. 



3. — On each side of the temple there was a flight of steps with a front 

 of 28J feet, supported by flank walls 1 1\ feet in length. These walls 

 still remain, and I believe that the steps yet exist uninjured, beneath 

 what Moorcroft* justly calls a " confused mass of ruins." According 

 to him the edifice must have been " a square temple with four doors 

 approached by broad and spacious porches." This description corres- 

 ponds exactly with that which I have already given as the most proba- 

 ble style of superstructure of the other temple, which is the same as 

 that of the temples at Pathan. 



4. — Of the surrounding quadrangle nothing but the foundations can 

 now be traced, excepting to the westward, where parts of the gateway 

 walls, and of the sides of the recesses are still standing. The gateway 

 itself was similar in plan to that of Marttand, and much about the same 

 size ; but its width did not bear the same proportion to that of the 

 temple. In the Marttand example the width of the gateway was made 

 equal to that of the temple itself, or rather to that of the arddha-man- 

 dapa, or outer-chamber, whereas in both of the Avantipura examples 

 the width of the gateway bears a very different proportion. In the 

 smaller temple it is made two -thirds of the width, or exactly equal to 

 that of the projecting porches ; whilst in the larger one it is only one- 

 third of the width, or just equal to the front breadth of the flight of 

 steps leading up to the entrance of the temple. 



5. — In Plate XVII. I have restored the plan of the quadrangle of this 

 temple, from the few stories which still remain in their original posi- 

 tions, guided by the plans of the Marttand and Avantiswami examples. 

 The foundations of many of the pillars still remain ; and as the exist- 

 ing stones prove that there were both pilasters and recesses, the ground- 

 plan of this peristyle must have been almost the same as that of the 

 others. This plan shows a quadrangle 216 feet long and 190 feet 

 broad, containing 86 recesses, from which two must be deducted for 

 the side doors, leaving the favorite number of 84 for the reception of 

 as many Linga or emblems of Siva. For this number, although dedi- 

 * Travels, v. 2— p. 243. 



