302 Primitive Greece: Mycenian Art. are fitted with the utmost nicety ; horizontal beds, however, are out of the question. The stones are of varying size and form ; each seems to be cut of the required shape in order that its pro- jecting angles shall marry the re-entering ones of the contiguous units on the wall surface. The polygons thus constitute a net- like structure of indestructible firmness and solidity (Fig. 94). In this fashion is constructed the wall which circles in front of the graves on the acropolis discovered by Schliemann ; as also a species of tower to the southward (Fig. 90, l), and again towards the north-east extremity of the fortified enclosure. Wherever the wall has kept its two faces, its thickness varies from three to seven metres ; on two points, however, north and south, it has been ascertained, from the fragments still lying on the ground, that the primordial depth of the rampart was as much as fourteen metres. The only possible explanation of this exceptional bulging out is that at these points the walls were pervaded by inner galleries akin to those of Tiryns (Figs. 74-78). These passages may have been similar to our casemates ; the structure on this side, however, is so hopelessly ruined that nothing can be asserted. The case is different in the north wall. There (Fig. 90, i) Schliemann at first thought, with much show of probability, that the entrance to the gallery (Fig. 95) led to a series of chambers.^ Later excavations have shown that the real purpose of the structure was to provide a covered passage between the acropolis and the upper portion of the aqueduct conveying the water of the Perseia;^ so that the existence of the latter should remain secret and unknown to the enemy, and that the inhabitants of the castle should not be reduced to the uncertain supply from the three cisterns, which are partly rock-cut and partly built, on the acropolis ; these, though of good size enough, would prove very inadequate in a long siege,^ and not save ^ Schliemann, Mycena, 2 TSOUNDAS, Report in IT/oouTiica ri/c apX^^'oXoyaiyc haipiaCf 1889. ^ Steffen thinks that the aqueduct passed near the Lions Gate and entered the interior of the citadel. We believe with Belger that excavations carried on in this region of the acropolis have not made good the assumption. No pipes have been discovered front or back of the gate in which water under pressure could have risen to a level approaching that at the point of outlet. As may be seen from Steffen's map, the Turkish aqueduct, which he is inclined to think follows the line of the old one, turns the north-west corner of the circuit at a much lower level than the ground-sill of the Lions Gate and the esplanade close by. The water-supply was conveyed to the town.
