high, cut out in the limestone-rock. At about 10 mètres from the entrance is a vertical hole, I m. in diameter, artificially cut in the superincumbent rock; it served, no doubt, to admit both air and light. At a distance of 18 mètres from the entrance, the large passage divides into two very narrow passages and a broad one, the former being only wide enough for one man to enter; the third passage, which turns to the north, is nearly as broad as the principal passage; of the two others, one turns to the east, the other to the south-east.
My architects having in 1882 thoroughly excavated the space before the cavern, as well as the cavern itself and the three narrow passages, we found that the passages which turn to the east and south-east are about as long as the large passage, namely, 18 m., but that the passage which turns to the north is somewhat shorter; and that at the end of each of them there is a spring, from which the water flowed out into the large gallery, and ran off by an earthen pipe of the Roman time; but this pipe being broken in many places, my collaborators and I in 1879 had not even observed it, and we thought the water flowed off, by an open channel in the floor, into an earthen pipe which we found outside. But now, on clearing the large passage most carefully of all the earth and dirt still contained in it, we discovered beneath the earthen pipe on the natural soil a water-conduit of a very primitive sort, composed of unwrought calcareous stones laid without any binding material, and covered with similar plates, which extends all along the large passage and its northern arm. It has the very closest resemblance to the Cyclopean water-conduits found by me at Tiryns and Mycenae (see my Mycenae, pp. 9, 80, 141). The channel was filled up with clay and dirt, and this may probably have been the case from a remote antiquity.
This conduit, which certainly belongs to a remote antiquity, appears not to have been noticed at all by the
