538 MERGE TO GYRENE.

it : the stadium would probably afford little of interest, for the stones which were employed in its construction appear to have been carried away in later times to serve in other buildings ; and, indeed, little more could be expected from excavating the temples, than fragments of architecture too much decayed by time to render them particularly useful in furnishing details, or of statues which enthusiasm and bigotry have probably defaced, if they should even have been spared by the hand of time.

The city walls approach closely to the southern extremity of the stadium, and are in this part very decided. They begin from the verge of a deep ravine, as will appear by the plan, and continue in an unbroken line to the spacious reservoirs (at the south-eastern angle of the city) which are mentioned in the publication of Dr.

Della Celia. Here we lose traces of them, but they again make their appearance on the south side of the buildings just alluded to, and extend to the brink of the large ravine with which the aqueduct communicated. Beyond this (the aqueduct), a wall was unnecessary, for the mountain descends perpendicularly to the bed of the ravine, and renders all approach to the town in this direction impossible ; and as the wall of the aqueduct has not been built with arches, but carried along the mountain in a solid mass, it would have been fully sufficient for the purpose of defence, and was probably built solid with this intention.

Square towers were attached to the city wall in various parts, not apparently at regular intervals, but approaching each other more closely where the ground was low, and consequently more favourable