CHAP. III. DHARWAR. 119 but the central belt is more pronounced, and always apparently was on the west side of India. It will also be observed in this tower that every third course has on the angle a form which has been described as an amalaka in speaking of the crowning members of northern temples. Here it looks as if the two intermediate courses simulated roofs, or a roof in two storeys, and then this crowning member was introduced, and the same thing repeated over and over again till the requisite height was obtained. In the Para^urame^war there are three intermediate courses (Woodcut No. 312); in the great tower at Bhuvane^war, five ; and in the more modern temples they disappear from the angles, but are supplied by the miniature temple-forms applied to the sides. In the temple at Bodh-Gay& the same form occurs (Woodcut No. 19) on the angle of each storey ; but there it looks more like the capital of a pillar, which, in fact, I believe to be its real original. But from whatever form derived, this repetition on the angles is in the best possible taste; the eye is led upwards by it, and is prepared for the crowning member, which is thus no longer isolated and alone, but a part of a complete design. The frequency of the repetition of this ornament is, so far as is now known, no bad test of the age of a temple. If an example were found where every alternate course was an amalaka, it probably would be older than any temple we have yet known. It would then represent a series of roofs, five, seven, or nine storeys, built over one another. It had, how- ever, passed into conventionalities before we meet with it. To the north-west of Aihole is a .Saiva cave-temple, 1 and near it on the north-west is an old temple with a porch on four plain square pillars, the mandap built of massive stones, with a sloping roof, a pradakshina round the shrine, and, from the figure of Karttikeya on the roof of the entrance porch and of Garuda on the lintel of the shrine door, it was evidently dedicated to Vishnu (Plate XXIII.) It is known as the temple of Huchchhimalligudi, and appears to be of quite as early a date as any at Bhuvane^war or elsewhere. The Sikhara is relatively small, and if we compare this temple with that of Para.mrame.ywar (Woodcut No. 3 1 2), we observe that the latter is much more developed in style than the former. Unfortu- nately we have no direct record of its construction, the only indication of its date is an inscription on the north side of the west front, recording a grant for oil made in the thirteenth year of the Chalukya King Vijayaditya, that is in 718 A.D. ; 2 but the temple was clearly then established, we know not how 1 ' Archaeological Survey of Western I Kaladgi,' pp. 38-40. India,' vol. i.; 'Report on Belg&m and | ' 2 ' Indian Antiquary,' vol. viii. p. 284.
