The Aqua Alexandrina is mentioned by Lampridius in his life of Alexander Severus, but it was probably a branch from the Anio Novus only, as the Nymphæum engraved on one of his coins[230] has been identified with the ruins near S. Maria Maggiore (2126, 2127). This stands on very high ground, and the only aqueduct that could reach it was that of the Anio Novus. This branch can be traced in the wall by the piers of the arcade, which have been built into the Wall of Aurelian near the Porta Maggiore (80 A, B). The _specus_ and the arches were destroyed by the engineers of the Aqua Felice; the piers only remain, and these cease just before we arrive at the point where the railway now enters Rome through the wall (99). They come to an end directly in a line with a large reservoir, now a gardener’s house, near the Minerva Medica, almost between that and the wall, but a little to the south of it; one pier, however, of the tall arcade of the third century remains, as a sort of buttress, against that side of the fine building of the same period called the Temple of Minerva Medica (537). There are great remains of thermæ and fountains in the large vineyard in which this building stands. Some of these are of an earlier period; but a considerable part of them are of the third century, and of the time of Alexander Severus. The Nymphæum before mentioned was at the north end, and is a fine picturesque ruin, with very evident remains of the aqueduct in it (61, 963, 964). The water was here divided into several branches, of which we see portions of the _specus_ going in different directions. One of the most important of these goes to the great reservoir of the Thermæ of Titus and Trajan, called the Sette Sale. This was made long before the time of Alexander Severus, who only rebuilt the Nymphæum; and there is an arcade of the first century leading to this Nymphæum from a reservoir near the Porta di S. Lorenzo. At a short distance to the south of this celebrated Nymphæum is another very curious reservoir, now in a very bad state, having been turned into a gardener’s house, but which must have been of considerable importance, and probably belonged to the thermæ of the third century (2322, 2323). There is a _cippus_ with an inscription, which indicates that the building was a _castellum aquæ_ (2324). XVI. ALGENTIANA, A.D. 300. This aqueduct was made to supply the great Thermæ of Diocletian on the Viminal Hill, but is believed to have been entirely subterranean, so that little is known about it. There was a reservoir for it on the eastern side of the Thermæ, under the present railway station. Drawings and a plan of this were preserved by Visconti[231]. XVII. THE AQUA CRABRA, THE MARRANA, A.D. 1124, AND THE RIVER ALMO.