Another branch of this wall extended all along the western bank of the torrent which the bridge spanned, and doubtless connected the latter with the south-western corner of the Acropolis. Of both branches of this wall very numerous traces remain, though with difficulty perceptible. Thus a part of the lower town, but scarcely one-third of it, was enclosed by a circuit wall. This was very insignificant, because its thickness on the ridge is only 6 feet, and it is still less on the bank of the torrent; so that it cannot have been high, and it was probably intended only to impart greater strength to the great Cyclopean walls of the Acropolis, and to prevent the Lions' Gate leading directly into the open country. After carefully examining the remnants of this city wall in numerous places, I see, in consideration of its weakness, no reasonable ground to object to regarding it as of later date than the walls of the citadel. The remaining part of the town has been, as the remnants of the house-walls show, a vast and well-built suburb, whence, when attacked by the enemy against whom their own means of defence were insufficient, the inhabitants could retire into the fortified part of the city and into the citadel. Some of the buildings of this suburb are very large, and show a most splendid Cyclopean masonry. I call particular attention to the vast building on the very bank of the deep glen in a westerly direction from the Lions' Gate, of which all the four walls are still visible. It is 93 feet long and 60 feet broad, and may have been a temple. I call attention also to the foundations of a large Cyclopean building, perhaps a temple, on the crest of a hill S.S.W. of the Acropolis and north of the village of Charvati. This hill appears to have been at the extremity of the suburb in this direction, for the Mycenean potsherds cease beyond it. I found there two well-polished axes of diorite. In two glens in the immediate vicinity of this hill are the only two wells of Mycenæ. The ruins of Cyclopean buildings close to them, and the Mycenean potsherds which extend beyond them, can leave no doubt that both wells were within the suburb. Strange to say, Professor E. Curtius has thought the ancient quarry of Charvati to be ruins of the city wall, and he has therefore put this village on his map still within the site of Mycenæ; but this is a great mistake; the city never extended so far. But not all the Cyclopean walls in the suburb are house-walls, for many of them are only intended for the support of the terraces. [Sidenote: TREASURIES IN THE SUBURB.]