The floods of the Tigris and Euphrates combine with the ordinary action of their streams upon their banks to produce a constant variation in their courses, which in a long period of time might amount to something very considerable. It is impossible to say, with respect to any portion of the alluvial plain, that it may not at some former period have been the bed of one or the other river. Still it would seem that, on the whole, a law of compensation prevails, with the result that the general position of the streams in the valley is not very different now from what it was 4000 years ago. Certainly between the present condition of things and that in the time of Alexander, or even of Herodotus, no great difference can be pointed out, except in the region immediately adjoining on the gulf, where the alluvium has grown, and the streams, which were formerly separate, have united their waters. The Euphrates still flows by Hit and through Babylon; the Tigris passes near Opis, and at Baghdad runs at the foot of an embankment made to confine it by Nebuchadnezzar. The changes traceable are less in the main courses than in the branch streams, which perpetually vary, being sometimes left dry within a few years of the time that they have been navigable channels. The most important variations of this kind are on the side of Arabia. Here the desert is always ready to encroach; and the limits of Chaldaea itself depend upon the distance from the main river, to which some branch stream conveys the Euphrates water. In the most flourishing times of the country, a wide and deep channel, branching off near Hit, at the very commencement of the alluvium, has skirted the Arabian rock and gravel for a distance of several hundred miles, and has entered the Persian Gulf by a mouth of its own. In this way the extent of Chaldaea has been at times largely increased, a vast tract being rendered cultivable, which is otherwise either swamp or desert.