The three walls are so near together, and follow so closely in the same direction, that if we prefer to consider the Arachtu to be a canal of the Euphrates, it here lies so close to the Euphrates that its existence is very problematic. The Euphrates wall of Nabonidus has here obviously replaced the Arachtu wall of Nabopolassar, which further argues for the identity of the Euphrates and the Arachtu (see p. 140). That the Nabonidus wall and the stone bridge are buildings on the Euphrates, no systematic investigator can doubt. Otherwise we must assume that besides the two buildings found by us there existed yet a second embankment wall of Nabonidus which lay on the Euphrates, and a second stone bridge that led over the Euphrates. Without wishing to anticipate further research, I am inclined to assume the Arachtu to be, not a canal nor an arm of the Euphrates, but a semicircular widening of the river (see Hommel, _op. cit._ p. 283, note 1, Arach) (moon, fem. Arachtu?), which possessed a special name, and for which the name Arachtu could be used as well as that of Euphrates, as in the case of the Binger Lock on the Rhine. Possibly it was the haven of Babylon. Nebuchadnezzar mentions his own wall among others in the Eharsagila cylinder (_K.B._ iii. 2, p. 41, l. 41): “I ... built the embankment walls of the Arachtu of asphalt and burnt brick, and strengthened it by means of the embankment walls that my father had made.” XXXV ESAGILA, THE TEMPLE OF MARDUK (A. THE PRINCIPAL BUILDING) The ancient celebrated temple, Esagila, according to Jastrow “the lofty house” (_Religion of Babylonia_, p. 639), the temple of Marduk, lies beneath the hill of Amran Ibn Ali (Fig. 123) buried to a depth of 21 metres below the upper level of the hill. We have already excavated some part of it, and by means of deep shafts and galleries we have established the ground-plan and the different divisions. There are two buildings adjacent to each other; the principal one on the east is very regularly and magnificently planned, of the Western Annex we have only recovered the outer circuit. We will first survey the principal building. The temple is almost square, with its northern front of 79.3 metres and its western front of 85.8 metres long. Inside it is a court 31.3 metres broad and 37.6 metres long. On the west of this court, as we learn from the mighty-towered façade, there was the principal cella, that of Marduk. The chambers are not yet excavated. On the south side towards the east there is a smaller cella, which can be recognised as such by the niche in the wall. The cella lies on the east side of a square, which on the west side has a door leading to a small chamber which may also be the remains of a cella. [Illustration: FIG. 123.—Plan of the mound Amran. ] [Illustration: FIG. 124.—Section through Esagila. ]