The bad effect of the petrifying stream was noticed by the engineers of the Aqua Felice, and, as that stream was carefully excluded by them, it now runs to waste through the meadows. This discovery was probably made before the third century, and the _specus_ being then found to be choked up with stalactite, was restored to use on a higher level by Alexander Severus. Although the _castella_, or reservoirs, are all of the time of Trajan and Hadrian, the _specus_ and the arcade to carry it is of two periods, the later portion of it being of the third century. The principal intention of this aqueduct has been originally to supply the great villa of Hadrian at Cento-Celle, but a branch of it does appear to have been brought into Rome. Although it is very difficult to trace it for the last two miles, this may arise from the general use of all the aqueducts near Rome as quarries, by the engineers of the Aqua Felice. In the excavations made in 1871, in the high ground near the Minerva Medica, between the Porta Maggiore and the Porta di San Lorenzo, by the side of that portion of the Marcian arcade that has been mentioned as found there, was also part of another arcade of an aqueduct of the third century, which may have belonged to this, and it was in that neighbourhood that the inscription belonging to the aqueduct of Severus was found. About a mile nearer to Rome, at the Torre Pignattara[176], or Mausoleum of S. Helena, there is a branch from the Marrana, passing under the Aqua Marcia, apparently to convey water to an imperial villa. The construction of the arcade of this branch to the south of the road is not of the same period as that of the other arcade on the other side to the north. It is usual for modern topographers, following Fabretti, to consider this branch a continuation of the Aqua Alexandrina; but as the water from the Marrana, which is at a considerably lower level than the Aqua Marcia, runs _down_ the gentle incline of this arcade towards the road and the Mausoleum of S. Helena called Torre Pignattara, this is impossible. The construction is chiefly of the fourth century. The Mausoleum of S. Helena is on lower ground than this part of the Marrana. Constantine is said to have built a villa for his mother near her tomb; but there are no remains of an imperial villa nearer than those called the Cento Celle, a mile further along the road, or the one called “Torre de’ Schiavi[177],” i.e. Tower of the Slaves, about a mile across country, on another road, but on the same level. A milestone for the third mile, found opposite this mausoleum, with an inscription of the time of Maxentius, has been published by Ciampini. This indicates that some works were going on there at the end of the third century, and the arcade may be of that period.