recesses on either side there runs what seems to be a road or path,
which sweeps away down the hill whereon the temple stands in a bold
curve, each path closely matching the other. The whole building is
perfectly symmetrical, except that the panelling is not quite uniform in
width nor arranged quite regularly. On the second stage, exactly in the
middle, there is evidently a doorway, and on either side of it a shallow
buttress or pilaster. In the centre of the third story, exactly over the
doorway of the second, is a squared niche. In front of the temple, but
not exactly opposite its centre, may be seen the _prophylaea,_
consisting of a squared doorway placed under a battlemented wall,
between two towers also battlemented. It is curious that the paths do
not lead to the propylaea, but seen to curve round the hill.

[Illustration: PLATE 53]

Remains of _ziggurats_ similar to this have been discovered at
Khorsabad, at Nimrud, and at Kileh-Sherghat. The conical mound at
Khorsabad explored by M. Place was found to contain a tower in seven
stages; that of Nimrud, which is so striking an object from the plain,
and which was carefully examined by Mr. Layard, presented no positive
proof of more than a single stage; but from its conical shape, and from
the general analogy of such towers, it is believed to have had several
stages. [PLATE LII., Fig. 2.] Mr. Layard makes their number five, and
crowns the fifth with a circular tower terminating in a heavy cornice;
but for this last there is no authority at all, and the actual number of
the stages is wholly uncertain. The base of this ziggurat was a square,
167 feet 6 inches each way, composed of a solid mass of sun-dried brick,
faced at bottom to the height of twenty feet with a wall of hewn stones,
more than eight feet and a half in thickness. The outer stones were
bevelled at the edges, and on the two most conspicuous sides the wall