The most famous temple at Ashur was that of the god Ashur himself, but unfortunately it is badly preserved, and is consequently of less archæological importance than the Anu-Adad temple or the temple of Nebo. One point of interest about the ancient temple of Ashur, is that the rooms appear to have been broad rather than long. In the oldest part of the building, an alabaster block[65] bearing an inscription of twenty-four lines written in archaic characters was discovered. The characters somewhat resemble those found in Irishum’s inscriptions and are similar to the characters used in early Babylonian inscriptions, while like them, they read longitudinally and not laterally, but the lines run from left to right instead of from right to left, and in this they resemble a few inscriptions found at Tellô.[66] This alabaster block is possibly the oldest Assyrian inscription as yet brought to light. In the fore-court of this same temple, some fragments of a diorite sculpture with small figures similar to those of the Khammurabi period were found. The best-preserved ziggurat in Mesopotamia is that which was discovered at Khorsabad; four stages of this tower still remain, and the colours with which they were painted are yet visible. It is in close proximity to though not in immediate connection with the group of buildings formerly regarded as the harem of the palace, but recently shown by Koldewey[67] to be in reality a group of temples (cf. Fig. 24 B). The argument upon which the harem-theory was based was the fact that this block of buildings is separate from the palace, but this argument could be used with even greater force in support of the temple theory, while its proximity to the ziggurat, and the general correspondence in form and shape of the several buildings which it comprises, to the normal Assyrian temple as revealed by the excavations, makes Koldewey’s contention a practical certainty. Furthermore, though the ziggurat, as is the case at Borsippa, is not connected with the theoretical “temple-complex,” there seems to be no doubt they belong to each other as there is no room elsewhere in the neighbourhood for a temple proper, and the adjacent parts of the palace were certainly used for secular and not religious purposes. The block would appear to contain three temples the entrance to each of which was through a central court; the temples consisted in a broad-room or vestibule, a long-room or hall at the end of which was another room—presumably the sanctuary where the statue of the god was enshrined. The entrance to the sanctuary from the hall was through a broad opening and up some stairs.