That many of these structures, if not all, were aqueducts admits of no doubt, for some have been found to contain earthenware water-pipes, while others of them still carry into the city water brought by underground conduits from the hills on the west of the fortifications; and that they were dams seems the only explanation of the buttresses built against their lower side, as though to resist the pressure of water descending from a higher level. [Illustration: Aqueduct Across the Moat of the Theodosian Walls.] [Illustration: Coin of the Emperor Theodosius II. (From Du Cange.)] Certainly Dr. Paspates’ view has very much in its favour. It is, however, not altogether free from difficulties. To begin with, the idea that the Moat was flooded only during a siege does not agree with the representations of Manuel Chrysolaras and Bondelmontius on that point. The former writer, in his famous description of Constantinople, speaks as if the Moat was always full of water. According to him, it contained so much water that the city seemed to stand upon the sea-shore, even when viewed from the side of the land.[206] The Italian traveller describes the Moat as a “vallum aquarum surgentium.”[207] Are these statements mere rhetorical flourishes? If not, then water must have been introduced into the Moat by some other means than by the aqueducts which traverse it, for these, as Dr. Paspates himself admits, ordinarily took water into the city. Unfortunately, it is impossible, under present circumstances, to examine the Moat thoroughly, or to explore the territory without the city to discover underground conduits, and thus settle the question at issue. One can only ask, as a matter for future investigation, whether, on the view that the Moat was always flooded, the water required for the purpose was not brought by underground conduits that emptied themselves a little above the bed of the Moat. The mouth of what appears to be such a conduit is seen in the counterscarp of the Moat immediately below the fifth aqueduct to the south of Top Kapoussi. If water was brought thus to the elevation of Top Kapoussi and Edirnè Kapoussi, sufficient pressure to flood the rest of the Moat would be obtained. But, in the next place, it must be added that objections can be urged against the opinion that the Moat was flooded even in time of war. The necessary quantity of water could ill be spared by a city which required all available water for the wants of its inhabitants, especially at the season of the year when sieges were conducted. Then, there is the fact that in the accounts we have of the sieges of the city, all contemporary historians are silent as to the presence of water in the Moat, notwithstanding frequent allusions to that part of the fortifications.