As the Ganeśa Ratha is the most nearly finished of any, it may be as well to begin with it, though it would be rash to say it is in consequence, the earliest. It does seem probable however, that the masons would first select a suitable block among the many that exist, on the hill for an experiment, before attempting the much more serious undertaking of fashioning the southern ridge or group into the Rathas bearing the Pându names.

As will be seen from the annexed woodcut the Ganeśa Ratha is, though small, a singularly elegant little temple. In plan, its dimensions are 19 feet by 11 feet 3 inches, and its height 28 feet. It is in three storeys with ery elegant details, and of a form very common afterwards in Dravidian architecture for gopuras, or gateways, but seldom used for temples, properly so called, in the manner which we find employed in this instance.

The roof is a straight line; and was adorn at either end by a triśula ornament, and similar emblems adorned four at least of the dormer windows that cut into it. It is, however, no longer the triśula of the Buddhist, but an early form of the trident of Śiva, who is the god principally worshipped in this place. Between the tridents the ridge is ornamented with nine pinnacles in the form of vases which also continues to be the ornaments used in similar situations to the present day. The roof itself is pointed, both internally and externally, in a manner entirely suitable to the wooden construction from which it is copied. It is true that in most of the western caves the internal form of these roofs is of a circular section, but externally there always is and must have been a ridge, to throw off the rain water, so as to make the external form an ogee, and so it is always represented. In some instances, as the Son Bhandar cave at Râjgir (woodcut No.7) and at Sita Marhi (woodcut No. 11), the internal form was also pointed, and so I fancy it generally was in the wooden structures from which these Raths were copied.

Like all the many storeyed buildings of this class with which we are acquainted; this temple diminishes upwards in a pyramidal form, the offsets being marked by ranges of small simulated cells, such as no doubt existed in Buddhist viharas on a large scale, and were thus practically the cells in which the monks resided, or at least slept. In this instance they are more subdued than is usually the case, but throughout the whole range of Dravidian architecture, to the present day, they form the most universal and most characteristic feature of the style.

The pillars in the porch of this temple are of a singularly elegant form, but so very little removed from their wooden prototypes as to be very unsuited for the position they here occupy in monolithic architecture. Their capitals, though much more slender, are of the Elephanta type, and their bases are formed by yalis or lions, which are clearly derived from some wooden originals, and are singularly unlike any lithic form (woodcut No. 29). They are, however, the most characteristic features of the architecture of the place, being almost universal at Mahâvallipur, but not found anywhere else, that I know of.

On each side of the entrance there is a dwârpâla or porter, and on the back wall of the verandah is an inscription in a long florid character. dedicated to Śiva, and stating that the work was executed by a king Jayarana Stamba1, but his name occurs nowhere else, and we can only guess his age from the form of the alphabet in which it is written, which, as before stated, is certainly not far removed from the year 700.

The image in the small shrine inside is not cut in the rock, but of a separate stone, and has been brought and placed there, instead of a lingam, which in all probability, originally occupied the sanctuary.

  • 1. Trans. R. A. S., vol. ii. p. 266, Plate 14, Carr. pp. 57 and 201.