2/2 Essay on the Avian Order of Architecture, [Sept. 



XV. where the perspective view of these transverse stones is not shown, 

 the general effect looks rather heavy, which is not really the case ; for 

 excepting those of the pair of pillars, immediately in front, all these 

 transverse beams are seen resting upon the wall. Their moulded ends 

 cannot, therefore be mistaken for upper capitals. The greatest and 

 most characteristic distinction therefore, between the Arian and Classic 

 orders, lies in the disposition of the architrave. In the latter it lies 

 immediately over the line of pillars ; whilst in the former it is placed 

 over the transverse beams. There are consequently no metopes in the 

 Arian architecture. 



29. — About one-third of this entablature still exists, principally on 

 the north-eastern side of the quadrangle : but the mouldings have been 

 so much injured by the weather, that their character could only be con- 

 jectured, from the general outline relieved against the sky, to be much 

 the same as that of the transverse beams. The upper part of the roof 

 of the quadrangles has entirely disappeared, but with reference to the 

 pointed character of other Kashmirian roofs, its form might have been 

 restored conjecturally as triangular in section, the height being some- 

 what less than the base. Luckily the enclosing walls of the temple on 

 the Takht, and of the old Hindu temple now occupied by Zein-ul-ab- 

 ud-din's tomb are still perfect ; and although they are on a small scale, 

 and of a primitive style, without columns, yet the division of their walls 

 into arched recesses is precisely the same as that followed in the main 

 wall of Marttand. The roofs of the former are both triangular in sec- 

 tion ; and such no doubt was that of the Marttand quadrangle. Fur- 

 ther, as the Zein-ul-ab-ud-din example exhibits small breaks or mould- 

 ings on each face, so might it be presumed that the roof of the 

 Marttand peristyle was likewise broken into two portions by an orna- 

 mental band, exactly similar to that which I have assigned to the temple 

 itself. As, however, this would impose the observance of the same 

 treatment with the roofs of the central porches and corner buildings, I 

 have not adopted it in my restorations ; principally because I do not 

 think that the general appearance would thereby be improved, and 

 partly because the intervention of the ornamental band would make the 

 total height of the central porches somewhat more than twice their 

 own breadth, which was the proportion strictly adhered to during the 

 best days of Kashmirian architecture. 





