MERGE TO GYRENE. 529 edge of the precipice appears from the upper seats to be close at the foot of the ranges, although the whole of the arena intervenes, and it often made us giddy to look down from them. As the lower ranges of seats are not in their places, it is difficult to ascertain the diameter of the arena, but it seems to have been more than a hundred feet across ; and to have been, like that which we have mentioned at Ptolemeta, of a perfectly circular form. There is no appearance of any praecinctions, owing probably to the absence of interior communications, which are not to be found in this building ; and it seems to have been chiefly approached from the top, which is equal in height with the level summit of the hill, against which the seats are on this side built. The most natural approach would certainly have been that which leads from the fountain of Apollo, along the edge of the descent which we have mentioned : this will be evident from the plan of the city; but strong walls, which are undoubtedly of ancient construction, cross the road here so completely as to preclude the possibility of any approach from the city to the amphitheatre in this direction. If the walls which we allude to had not been standing at the present day many feet above the level of the road, we should have concluded that they must originally have contained gates which led to the arena ; but there is no appearance whatever of such communication, even supposing that the gates were approached by flights of steps, which would not have been an unreasonable conjecture.

The only road which remains (under the difficulties stated) must at the same time have been a circuitous one ; and as it communicated merely with the level summit of the hill, against which the seats