FORTIFIED WALLS. 357 than the main rampart, but it afforded a shelter to the catapults and other machines, and to the troops who served them. Finally, behind the third ditch, there was an outer defence of palisades, which served to at least prolong the siege and to put off for some days the moment when the main wall should be seriously attacked. Daux tells us that he found easily traceable remains of a triple enceinte like this both at Thapsus and Adrumetum. We give his restorations (Figs. 246 and 250). Thus there is perfect accord between the theories handed down to us by Philo and the evidence collected by examining the Punic ruins. Appian himself admits the distinction between the wall and the advanced-wall, if not in so many words, at least by implication. 1 The idea of three exactly similar walls must, therefore, be given up ; and the dimensions given by Diodorus and Appian must be taken as applying to only one of the three, the innermost one, which was the real bulwark of the city. When the historians of the siege spoke of the triple wall, it was merely to distinguish between the fortifications where they were complete, on the side towards the isthmus, and the mere skirt of masonry by which the town w r as embraced on the side towards the sea. So that we must not multiply by three the numbers given by Appian for elephants, horses, and foot-soldiers. We must be content with 300 elephants, 4,000 cavalry, and 20,000 infantry, all of whom could easily, according to Graux, have found accommodation in the casemates of a single wall, especially as it was not less than 7,000 yards long. The distance from the Lake of Tunis to the Lake of Soukhara, across the isthmus, is about 5,500 yards, and we must allow at least 1,500 for the windings of the rampart, for its salients and re-entering angles. A detailed discussion of the topography of Carthage would here be out of place, but it is important that her fortifications should be clearly understood. Even when shorn of the magnitude ascribed to them by some writers, they still remain perhaps the most 1 APPIAN, viii. 97. He speaks of the TrpoT^urp-a. at the end of this paragraph, and in an earlier passage we should no doubt read TrpoTeix'oyw* instead of the cTrtTei'xio-px of the manuscripts. Graux's correction to that effect seems beyond dispute ; the word riTeixr/ia has quite a different meaning. No other word but irporeixurfia could be rightly opposed to ra vi/^Xa retx 7 ?, " the elevated wall," which Censorinus wished to attack after having filled up the ditch and beaten down the rampart low enough.
