vineyards, crossing a valley nearer to Tivoli, and there is a considerable construction, perhaps connected with it, between the road and the Anio. On the other side of the Fosso degli Arci, we meet again with the remains of two of these aqueducts. They are soon lost, but about a mile further on, we find a massive construction upon arches crossing a little valley, and just beyond this, an apparent ramification, and a much more extensive range of arches, but more destroyed. I endeavoured to trace this branch, which seems to be that which Frontinus mentions as having been made and then deserted. It perforates the hill, crosses another valley in two parallel lines, which unite again, passes through another hill, and is found coasting the valley of the Fosso degli Arci, considerably above the place where I had quitted it. The foundations are carried to the edge of the brook, and the watercourse here could only be just high enough to enable it to pass the natural channel of the little stream without obstructing it: here we lose it. Where the channel itself can be examined, we find a deposit of tartar, but not of the thickness that it is down below. All the hills thus perforated are of a volcanic tufo, or peperino, for although the valley of the Anio be essentially among the limestone of the Apennines, yet for some miles above Tivoli, it abounds in abrupt eminences of volcanic deposits. Towards Vico Varo we again see traces of an aqueduct crossing the Anio, which is probably the Marcia, or a tributary stream called Augusta, said to be equal in purity to itself. Below Tivoli, we find again three of these aqueducts, on the slope of the hill above the villa of Hadrian. The Anio vetus is traced round an ancient sepulchre, at a very small distance from Tivoli, and the sharp curve it makes to avoid the tomb, has occasioned so large a deposit, as to prevent the free passage of the water; and a new channel was consequently formed for it. Their course here is in a direction nearly opposite to that of Rome, and I am always at a loss to know why their constructors should take so much pains to maintain an elevation so greatly beyond what appears necessary.