INDIAN.] ARCHITECTURE 395 by pillars of tho natural rock left in their places, and surrounded by a number of small sleeping-places or cells. Fid. 28. Cave at Baugh, on the Xerbudda ; plan. The most wonderful excavations are those at Ellora, near Aurangabad. These are a series of hypogea or caves sunk in the solid rock, extending a distance of 3 or 4 miles. Canina has given plans and interior views of six of them. Those called Parasova Rama and Diajannata are simply halls supported on massive piers with level architraves. The piers are richly carved with figures and friezes, and have a sort of cushion capitals, and square abaci, and stand round, forming a kind of atrium. That called Indra has a court open to the sky, in which is a small shrine or temple. In the solid rock are two halls similar to those above described, a larger and smaller. The piers of the Tin Tal are quite plain. In the Viswakarma is a quadrangle, open to the sky and surrounded by pillars. This leads into an atrium with three aisles and an apse, and exactly like a basilican church, las. The most magnifi cent of the Ellora caves, and indeed of the native Hindu works, are the chambers and halls called the Kylas, or Kailasa. These arc sunk into the rock, and occupy a space of 270 feet deep and 150 feet wide. The roofs are solid rock, supported by pillars, or rest on the walls, or on the divisions of the as semblage of cham bers. There is a porch, on each side of which are two columns. This con ducts into a hall, supported on 1G such columns, and Fio. 29. Kylas, Ellora ; plan. colonnade or cloister encircling the whole. Great part is open to the sky, for the sake of light and air, but the work is entirely cut out of the solid rock. leading into a sort of adytum. Round this is passage space and five chambers. The whole forms a temple, with its usual appendages, just such a one as would be built on the ground, and round this a wide open space, with a FIG. 30. Kylas, Ellora. The date of the construction of the Kylas is about 1000 A.D. The earliest existing work of the Jains seems to be of the Jain work. 1 Oth century. They were a sect which arose in the endea vour to re-establish Brahminism, and which first seems to have acquired importance about 450 A.D. This sect rejects the doctrines of Buddhism, as also the practice of monasticisrn. The famous temple at Somnauth belongs to them. Mr Fergusson has given a description of that built by Vimala Sah, on Mount Abu, as a type of the ordinary Jain temple. In the centre is a cell in which is a cross- legged figure of one of the twenty-four saints worshipped by this sect ; in this case it is that of Parswanath. The cell is always terminated by a pyramidal roof. In front of this is a portico of 48 pillars, disposed much like a cruciform church with a dome at the intersection of the transepts. The whole is surrounded by a species of cloister formed by double rows of columns, and a series of small chambers like the cells of a vihara. But as the sect abjure monasticism, each cell is used not as a dwelling, but as a kind of small chapel, and contains one of their cross-legged deities. One of the peculiarities of this style is that richly-carved brackets spring from the pillars at about two- thirds of their height, and extend to the architraves, forming a sort of diagonal strut to strengthen and support them. The Jains probably adopted the dome at a very early period, and it is doubtful whether the Buddhists ever used this species of construction. " No tope," Mr Fergusson observes, " has the smallest trace of such a structure, though of domical shape outside, and the design of the rock-cut temples, with the upright supports, the raking struts, and the level architraves, has manifestly been deduced from timber construction." The Indian dome has no voussoirs radiating from the centre, as in European architecture. Tho courses are all horizontal; and the domes are therefore necessarily pointed in section, for they would not stand if circular. The Indian dome, however, has this merit, it requires no abutments, and has no lateral thrust. The pressure is entirely vertical ; and if the foundation be sound, and the pillars stout enough, there can be no failure. The leading idea of the plan of the Jain temple is that of a number of columns arranged in squares. Wherever it was intended to have a dome, pillars were omitted, so as to leave spaces in the form of octagons. By corbelling over the pendentives in level courses the dome was gradually
formed. Tho plan and view of the temple at Sadrec