TRIPOLY TO BENGAZI. 141 It is difficult to fix any precise date to these buildings, but we may perhaps conclude, with some appearance of reason, that the greater number of them were erected by the Eomans under the emperors, who possessed, at various times, the whole of the north coast of Africa, and kept open an extensive communication along the shores of the Mediterranean, as well as with some parts of the interior *. The quadrangular form of these structures is the same as that used by the Komans in their stations and encampments ; and the small number of troops which was allotted by the empire for the defence of Africa, made it peculiarly necessary that their garrisons should be well intrenched -f-. It has been calculated that a square of seven hundred yards was sufficient, according to the Roman method of encampment, for containing a body of twenty thousand men ; and a square of one hundred feet would, at that rate, suffice for the accommodation of nine hundred and fifty. The habitable parts of the forts above mentioned very rarely exceeded a square of that size, and this portion of the structure, in by far the greater number of them, seldom * The tower of Euphrantas is however stated to have been a boundary fort under the Ptolemies; and the fortress of Automala, at the bottom of the gulf, is mentioned by Diodorus to have been in existence before the occupation of Gyrene by the first of those princes. — See Strabo, lib. 17, and Diod., lib. 20.

f “ With regard to Egypt, Africa, and Spain, (says Gibbon, in describing the distribution of the Roman forces,) as they were far removed from any important scene of war, a single legion maintained the domestic tranquillity of these great provinces.”

“ We may compute (says the same writer) that the legion, which was itself a body of six thousand eight hundred and thirty-one Romans, might, with its attendant auxiliaries, amount to about twelve thousand five hundred men.”