architectural effect is marred by the confused grouping of the later shrines clustered round its base. According to Brahmanical tradition, the temple was built by a king who reigned in the seventh century a.d. This would make it one of the oldest at Bhuvanēshvar, and, from its location in the centre of the sacred circle, one might expect this to be the case.
But archæologists assign to it a later date—the ninth or tenth century. Such discrepancies between Indian and European chronology are often accounted for by the fact that a great temple built as a votive offering by a ruling dynasty frequently enclosed a smaller shrine of venerable antiquity. We have seen that this was the case with the Great Stūpa at Sānchī; and as Bhuvanēshvar was a sacred city long before the ninth century, it is not unlikely that the king who built the Linga-rāj temple to the glory of his patron deity enclosed within the royal sikhara an ancient shrine where his ancestors had been accustomed to worship. The "linga" of Siva worshipped there may have been originally a Jain or Buddhist stūpa. It would be as easy to enclose the cubical dome-shaped shrine, such as is depicted in early Buddhist sculpture, within Vishnu's lofty steeple as it would be to cover a stūpa by a stūpa. Whether this method was adopted in particular cases could only be ascertained by careful examination of the holy of holies, into which the inquisitive archæologist is never allowed to enter.
It is, however, perfectly clear that this was one of the main principles of Indian temple design. A shrine in which a god had deigned to dwell for centuries might fall into decay from natural causes, but the temple architect who was priest as well as builder would never profane it by rebuilding it on a larger scale. Another site might be chosen upon which to raise a
