Where the præhistoric temple has thus become the cathedral church, it is not out of keeping that the wall of the præhistoric _arx_ should become the wall of the cathedral close. This is the wall which we see from afar, a wall which seems so straight and regular, so clearly furnished with a modern finish at top, that it is not till we can distinguish the mighty blocks of which it is formed that it has the air of a wall even of Roman, even of mediæval, antiquity. Shall we say it? As we looked up at no very amazing distance, the wall of the _arx_ of Alatri had a good deal of the air of the wall of a modern prison. We could not yet see the construction, and the outline seemed more regular and rectangular than it proves to be. Nowhere do we better see than at Alatri the nature of these primitive walls. They are seldom walls in the same sense as the later walls of Rome or of other places, walls built on the ground and standing up clear on both sides. Their business commonly is, as is perhaps more clear at Alatri than anywhere else, to strengthen by masonry the scarped side of a hill. Hence they have little or no height within, and the gateways are necessarily reached from within by a steep descent. The open space at Alatri allows this arrangement to be studied with unusual ease. The wall is eminently a wall against a hill, and its arrangements are made with no small art. The weak corner has its double defence; the way up from the town at this point is carefully sheltered. And what stones they are with which the hill of Alatri is strengthened; above all, what stones they are which are piled together to form its main gateway. Nowhere indeed in the walls of Alatri, whether of temple, _arx_, or city, do we find anything quite so rude as the rudest part of the wall of Cori. All the stones, of whatever shape--and they are of many shapes--have clearly been cut; they are all laid according to some kind of system, though the system according to which they are laid is not the same in every part of the wall. In some parts they seem almost to take the shape of constructive arches, at least of attempts at arches, such as may be seen in gateways and roofs at Segni and elsewhere. The true arch, it is hardly needful to say, is nowhere found in the original work; nor do we find even any of the attempts at the arch in that position where we should have most naturally looked for them, in the gateways. The great gateway of the _arx_ at Alatri is indeed a wonderful work. Its builders either knew the arch and despised it, or else the thought of the arch had not come into their heads. It is as pure an example of the lintel-construction as any gateway at Athens or Mykênê. We suppose that the lintel-stone of the great treasury is yet vaster than the huge lintel-stone at Alatri; but the Anakim of Alatri were at least rivals whom those of Mykênê could not have despised. But, except in vastness of construction, we must not compare the gateway at Alatri, perfectly