116 JOURNEY FROM vent any communication between the marsh and the sea except in this place and the one which we are about to mention. In this place, particularly, such connexion seems formerly to have existed; for here a passage still remains, through the higher land forming the separation, by means of which the waters of the lake might have emptied themselves into the sea, and on one side of this channel is the structure in question which we have supposed to be a landing place. The small vessels of the ancients might have entered this passage, and have found sufficient shelter behind the high land* which formed it ; on the inner side of which they might also have been hauled up when the current through the channel was too strong for them to remain afloat f .

Nearer to Mesurata, a little to the southward of the Marabut of Sidy Abou Shaifa, are the remains of what seem to have been a small fort or station : its outer walls enclose a square of about an hundred feet and there are vestiges of smaller walls within, which appear to have divided it into several compartments. On the north-western side there are some small blocks ol stone, about two feet square, which seem to have been the abutments of arches formerly supporting the roof of the building ; and which are about eight feet distant from * No part of this land can properly be called high — but only by comparison with the lower level of the mai'sh.

•}• It was not, however, necessary, in places of this kind, that the vessels should be drawn upon shore ; for vtpo/xos- is the term here employed by feti’abo, and the oi v(poqfxoi, were somewhat similar to our own docks, and consisted of walls parallel with each other, between which vessels would be perfectly secure from wind and waves, as well as from the effects of strong current.