In spite of all their precautions time and experience compelled the inhabitants of Mesopotamia to recognize the danger of crude brick as a building material; they endeavoured, therefore, to supplement its strength with huge buttresses. Wherever the ruins have still preserved some of their shape, we can trace, almost without exception, the presence of these supports, and, as a rule, they are better and more carefully built than the structures whose walls they sustain. Their existence has been affirmed by every traveller who has explored the ruins of Chaldaea,[185] and in Assyria they are also to be found, especially in front of the fine retaining wall that helps to support the platform on which the palace of Sargon was built.[186] The architect counted upon the weight of his building, and upon these ponderous buttresses, to give it a firm foundation and to maintain the equilibrium of its materials. As a rule there were no foundations, as we understand the word. At _Abou-Sharein_, in Chaldaea, the monument described by Taylor and the brick pavement that surrounds it are both placed upon the sand.[187] Botta noticed something of the same kind in connection with the palace walls at Khorsabad: "They rest," he says, "upon the very bricks of the mound without the intervention of any plinth or other kind of solid foundation, so that here and there they have sunk below the original level of the platform upon which they are placed."[188] This was not due to negligence, for in other respects these structures betray a painstaking desire to insure the stability of the work, and no little skill in the selection of means. Thus the Chaldaean architect pierced his crude brick masses with numerous narrow tunnels, or ventilating pipes, through which the warm and desiccating air of a Mesopotamian summer could be brought into contact with every part, and the slight remains of moisture still left in the bricks when fixed could be gradually carried off. These shafts have been found in the ruins of Babylon and of other Chaldaean cities.[189] Nothing of the kind has been discovered in Assyria, and for a very simple reason. It would have been impossible to preserve them in the soft paste, the kind of pise, we have described. Another thing that had to be carefully provided for was the discharge of the rain water which, unless it had proper channels of escape, would filter through the cracks and crevices of the brick and set up a rapid process of disintegration. In the Assyrian palaces we find, therefore, that the pavements of the flat roofs of the courtyards and open halls had a decided slope, and that the rain water was thus conducted to scuppers, through which it fell into runnels communicating with a main drain, from which it was finally discharged into the nearest river.