CHAP. II. KANARAK. 107 the form of the design of the roof of the Jagamohan, or porch the only part now remaining. Both in dimensions and detail, it is extremely like that of the great temple at Bhuvane^war, but it is here divided into three storeys instead of two, which is an immense improvement, and it rises at a more agreeable angle. The first and second storeys consist of six cornices each, the third of five only, as shown in the diagram Woodcut No. 184. The two lower ones are carved with infinite beauty and variety on all their twelve faces, and the antefixae at the angles and breaks are used with an elegance and judgment a true Yavana could hardly have surpassed. There is, so far as I know, no roof in India where the same play of light and shade is obtained with an equal amount of richness and constructive propriety as in this instance, nor one that sits so gracefully on the base that supports it. Internally, the chamber is singularly plain, but presents some constructive peculiarities worthy of attention. On the floor, it is about 40 ft. square, and the walls rise plain to about the same height. Here it begins to bracket inwards, till it contracts to about 20 ft., where it was ceiled with a flat stone roof, supported by wrought-iron beams Stirling says nine, nearly i ft. square by 12 ft. to 18 ft. long. 1 My measurements made the section less 8 in. to 9 in., but the length greater, 23 ft; and Babu Rajendralal points out that one, 21 ft. long, has a square section of 8 in. at the end, but a depth of 1 1 in. in the centre, 2 showing a knowledge of the properties and strength of the material that would be remarkable, were it not that they seem to be formed of blocks of short lengths, 3 or 4 in. square, built together, like bricks, and then covered with molten metal. The iron pillar at Delhi (Woodcut No. 373) is a more remark- able example than this, and no satisfactory explanation has yet been given as to the mode in which it was manufactured, though it may possibly have been by a similar method. Its object, however, is plain, while the employment of these beams here is a mystery. They were not wanted for strength, as the building is still firm after they have fallen, and so expensive a false ceiling was not wanted architecturally to roof so plain a chamber. 3 It seems to be only another instance of that pro- fusion of labour which the Hindus loved to lavish on the temples of their gods. 1 'Asiatic Researches,' vol. xv. p. 330. 2 These discrepancies arose from the fact that the beams lay on the floor buried under the ruins of the stone roof they once supported, and it was extremely difficult to get at them so as to obtain correct measurements. 3 See ante, p. 95. The present survey furnishes no information, nor seems to have made any architectural drawings of the structural arrangements and details of the interior before burying it from all future examination.
