Some idea of the appearance of the quays may be gathered from the right-hand corner of the restoration in Fig. 5.[55] It is true that the outer quay-wall appears to have been built to replace the inner one, while in the illustration both are shown. But since the height of the citadel and of its walls was continually being raised, the arrangement there suggested is by no means impossible. But in the later part of his reign Nebuchadnezzar changed the aspect of the river-front entirely. To the west of the quay-walls, in the bed of the river, he threw out a massive fortification with immensely thick walls, from twenty to twenty-five metres in breadth.[56] It was constructed entirely of burnt-brick and bitumen, and, from his reference to it in an inscription from Sippar, it would seem that his object in building it was to prevent the formation of sandbanks in the river, which in the past may have caused the flooding of the left bank above E-Sagila.[57] A narrow channel[58] was left between it and the old quay, along which the river water continued to flow through gratings. This no doubt acted as an overflow for the old northern moat of the citadel, since the latter fed the supply-canal, which passed round the palace and may still be traced along its south side.[59] It is possible that the subsequent change in the course of the Euphrates may be traced in part to this huge river-fortification. Its massive structure suggests that it had to withstand considerable water-pressure, and it may well have increased any tendency of the stream to break away eastward. However that may be, it is certain that for a considerable time during the Persian and Seleucid periods it flowed round to the eastward of the Ḳaṣr, close under three sides of the citadel and rejoined its former bed to the north of Marduk's temple and the Tower of Babylon. Its course east of the Ishtar Gate is marked by a late embankment sloping outwards, which supported the thicker of the crude-brick walls at the point where they suddenly break off.[60] Beyond this embankment only mud and river sediment were found. The water-course to the south of the citadel is probably the point where the river turned again towards the channel it had deserted. A trench that was dug here showed that the present soil is formed of silt deposited by water, and beyond the remains of the earlier canal no trace of any building was recovered. This temporary change in the river's course, which the excavations have definitely proved, explains another puzzle presented by the classical tradition--the striking discrepancy between the actual position of the principal ruins of Babylon in relation to the river and their recorded position in the Persian period. Herodotus,[61] for example, places the fortress with the palace of the kings (that is, the Ḳaṣr), on the opposite bank to the sacred precinct of Zeus Belus (that is, E-temen-anki, the Tower of Babylon). But we have now obtained proof