to ascertain the exact level of every single fragment of the aqueducts that is now visible, and in certain cases it is difficult, if not impossible, to be sure from mere inspection to which of the four a given piece belongs.

Of the Appia but little is to be found, while the course of the Virgo, still in use, is accurately known; the Tepula and the Iulia travel, like the Appia, at a low level, and have no difficult country to traverse; and the remains of the Alexandrina are well known and prominent enough. Of the Alsietina, on the other hand, little is known. Lanciani records his inability to find traces of it, and doubts the accuracy of Nibby's account; while, though the course of the Traiana is followed by the modern Acqua Paola, it would seem, from the description that we have of certain of its remains, that it deserves a more careful study than has hitherto been devoted to it.

Of the four aqueducts of the first group the earliest in point of date is the Anio Vetus, which was constructed forty years after the Appia in 272-269 B. c. It took its supply from the river Anio, as its name implies, at a point about ten miles above Tivoli, in the gorge of S. Cosimato. The Marcia was the next, though considerably later in point of date (144 B. C.) and took its origin from springs about thirty-eight miles from Rome, three miles, Frontinus tells us, to the right of the thirty-sixth mile of the Via Valeria, and two hundred paces to the left of the thirtyeighth mile of the Via Sublacensis. Augustus tapped additional springs eight hundred paces higher up the valley for

the Marcia by which he doubled its volume.

The

springs of the Claudia, which was begun in 38 A. D. and completed in 52 A. D., were only a hundred paces beyond the original source of the Marcia. Attempts have been made by topographers to identify the individual springs; but as Lanciani remarked twenty-five years ago in his monograph, the question though interesting to the topographer is of minor importance and not one in which certainty can be attained. The fall of nature has changed considerably since Roman times, the level of the floor of the valley having risen owing to the calcareous deposits of the Anio and of the springs themselves; in fact, the engineers of the modern aqueduct (the Marcia Pia, which uses the same springs) have found the ancient aqueducts at a depth of seven or eight metres, and have also discovered that the large pools in which the water rises are merely due to leaks in the ancient collecting channels, the springs themselves being situated under the rocks at the edge of the valley. Considerable modifications have also been caused by the necessity of enclosing the modern springs so as to preserve them from pollution; so that further study of the problem would be difficult, even if it were worth while.

The Anio Novus dates from the same period as the Claudia, but drew its waters originally from the river four miles further up; as, however, the water was apt to be turbid, Trajan carried out a project of Nerva's, according to which the three lakes constructed by Nero for the adornment of his villa above Subiaco were used as filtering tanks

for the aqueduct. This increased its length considerably, the distance from Rome to Subiaco being about forty-seven miles by road, whereas the original intake was at the fortysecond mile of the Via Sullacensis. '

It may be added that both the Anio Vetus and the Marcia shared in the general restoration of the aqueducts under Augustus, and were resurveyed by his engineers, so that both of them had a regular series of inscribed cippi bearing his name. These served to mark the ground belonging to the aqueduct and protect it from encroachment, and perhaps sometimes to denote the place where ventilating shafts existed. The Claudia and Anio Novus had no cippi as far as is known.

The question of the relation between the lengths of the aqueducts is indicated by the inscriptions-in the case of the Anio Vetus and the Marcia by the cippi erected by Augustus (which were never, as far as we know, renewed or substituted by others by any later emperor) and in the case of the other two by the inscriptions of Claudius on the great double arch now known as Porta Maggiore, close to the point where they reached their terminal cisterns, as given by Frontinus, and as traceable by the actual remains, is by no means an easy one.

Almost due east of the railway station of Castel Madama, two streams, the Fosso Maiuro and the Fosso Ariana, unite and soon fall into the Anio. They are crossed just below the point of junction by a long stretch of aqueduct, over 200 metres in length in concrete: it is constructed in two

horizontal sections, with small arches in the lower section and buttresses from top to bottom.

There is another bridge lower down the stream but at the same level (belonging, therefore, probably to the Claudia) constructed with brickwork with buttresses, huge masses of which have fallen, and some masonry in opus quadratum at the actual crossing of the stream.

The Claudia and the Anio Novus, after arriving in the Valle d'Empiglione, separate at right angles. The former runs straight on (i. e., south) and crosses the valley by a fine series of arches about 800 metres long. They were originally built in opus quadratum, but were later on strengthened with concrete, faced with brick and opus reticulatum. The highest portion at the crossing of the stream is twenty-five metres in height; here massive buttresses have been added to the original construction.

It would seem that before 88 B. C. the Aqua Claudia must have followed the course of the other three aqueducts; but the question of its length is a difficulty. It appears, as Lanciani points out, that the first channel was hastily executed; for the inscription of Vespasian on the Porta Maggiore, where the aqueduct entered Rome, states that he, in 71 A. D., restored the aqueduct which Claudius had constructed, and which for nine years had been out of use and ruined; and this shows that the original aqueduct can have been in use for ten years only. It is still more noteworthy that his son, Titus, ten years later (in 81 A. D.), describes Vespasian's work as already in ruins especially near the

springs, so that a new channel had to be constructed.

We

do not hear of any similar trouble with the Anio Novus, and the problem is a difficult one. It may be that the Claudia followed the Anio Novus very closely, round to Tivoli, and used the same arches before the tunnel was made in 88 A. D., which avoided this portion of its course.

The Porta Maggiore is the ornamental archway by which the aqueduct of the Aqua Claudia and the Anio Novus crossed the Via Labicana, and the Via Prænestina just after their divergence constructed by Claudius. Aurelian incorporated it in the walls which he began in 272 A. D.