2. Herculea. This is mentioned by Frontinus[203] as receiving a part of the Aqua Marcia after its entrance into the city[204]. Another stream of the same name, thirty-eight miles from the city, was added to the Anio Novus[205]; but it seems most probable that the stream here intended is the Almo, which enters Rome under the Porta Metronia, and conveys the water of the Marrana, or Aqua Crabra, coming from these two mountain-streams united, as already described. This is always a very strong stream. It follows the line of the aqueducts for miles towards Rome, and received the surplus water in many places, where there were piscinæ; it passes very near to the point where the aqueducts enter Rome at the Porta Maggiore, and may very well have received the surplus water there also of the Marcia, as stated by Frontinus.
3. Cerulea. This is the name of one of the branches of the Claudia, among its sources, and is mentioned in the inscription at the Porta Maggiore; but again, this does not account for one of the nineteen Aquæ.
4. Augustea. A stream was added to the Aqua Marcia by Augustus, and called after him[206]; but that would not account for another stream in Rome here indicated as one of the nineteen Aquæ. It is more probable that the water here intended is the Augustan branch of the Aqua Appia, which was united at the Gemelli within the outer wall of Rome. Frontinus (c. 4), indeed, mentions that the Alseatina was also called Augustea, but that would not account for another of the nineteen streams[207].
5. Ciminia, another name for the Sabbatina [X.], according to Onuphrius Panvinius; but the sources of the Sabbatina are very distant from the Monte Cimino. Ciminia is perhaps written for Curtia, another source of the Claudia; but again, this does not account for another of the nineteen streams.
6. Damnata. This may be the stream that comes from the Quirinal and the Palatine, and which formed the lake of Curtius, whence it was conveyed into the Cloaca Maxima, as it still is. It was therefore condemned to serve for washing out a drain only.
There is reason to believe that some of these nineteen streams are the natural watercourses, and others merely branches from the great aqueducts, as has been said.
There were 1,452 lacus, that is wells, or cisterns of water, supplied by the aqueducts in Rome, according to the Horum Breviarium of the Notitia, or Catalogue of the fourth century, and they must in many instances have been very near together. A reservoir for the distribution of the water was almost a necessary termination of an aqueduct, and at a junction it was equally necessary.
Rutilius describes the aqueducts about A.D. 417, in such terms as shew that they were not destroyed in the first siege of the Goths[208]. Cassiodorus represents them as perfect, about twenty years before the siege carried on by Vitiges[209]. At the time of the siege by that king, Procopius (i. 19), writing in the sixth century, records only the fourteen aqueducts already mentioned[210].