which blocked the entrance to it was not so much a gulf as a lagoon of uncertain and shifting extent; the water flowed only in the middle, being stagnant near the shores; the whole expanse was irregularly dotted over with mud-banks, and its service was constantly altered by the alluvial soil brought down by the Tigris, the Euphrates, the Ulai, and the Uknu.  The navigation of this lagoon was dangerous, for the relative positions of the channels and shallows were constantly shifting, and vessels of deep draught often ran aground in passing from one end of it to the other.*

* The condition I describe here is very similar to what Alexander’s admirals found 350 years later.  Arrian has preserved for us the account of Nearchus’ navigation in these waters, and his description shows such a well-defined condition of the estuary that its main outline must have remained unchanged for a considerable time; the only subsequent alterations which had taken place must have been in the internal configuration, where the deposit of alluvium must have necessarily reduced the area of the lake since the time of Sennacherib.  The little map on the next page has no pretension to scientific exactitude; its only object is to show roughly what the estuary of the Euphrates was like, and to illustrate approximately the course of the Assyrian expedition.

[Illustration:  048.jpg MAP THE NAR-MARRATUM IN THE TIME OF SENNACHERIB]

Sennacherib decided to march his force to the mouth of the Euphrates, and, embarking it there, to bring it to bear suddenly on the portion of Elamite territory nearest to Nagitu:  if all went well, he would thus have time to crush the rising power of Merodach-baladan and regain his own port of departure before Khalludush could muster a sufficient army to render efficient succour to his vassal.

More than a year was consumed in preparations.  The united cities of Chaldaea being unable to furnish the transports required to convey such a large host across the Nar-Marratum, it was necessary to construct a fleet, and to do so in such a way that the enemy should have no suspicion of danger.  Sennacherib accordingly set up his dockyards at Tul-barsip on the Euphrates and at Nineveh on the Tigris, and Syrian shipwrights built him a fleet of vessels after two distinct types.  Some were galleys identical in build and equipment with those which the Mediterranean natives used for their traffic with distant lands.  The others followed the old Babylonian model, with stem and stern both raised, the bows being sometimes distinguished by the carving of a horse’s head, which justified the name of sea-horse given to a vessel of this kind.  They had no masts, but propelling power was provided by two banks of oars one above the other, as in the galleys.  The two divisions of the fleet were ready at the beginning of 694 B.C., and it was arranged that they should meet at Bit-Dakkuri, to the south of Babylon.