stone quarry, but of later character than the one first mentioned. In the earliest one, at the source, the stone has been split off the rock, not cut; in the other caves, the stone has been well cut, and of these many are to be seen, the openings being large square apertures like doorways, but side by side in the face of the scarped cliff, not following each other: a practice common in quarrying, where the circumstances require it, down to the present day. Between the chambers are massive square or oblong piers, left to support the vault and the earth above. At the bottom of one of these chambers is water running from another spring, and the shepherds state that there is an opening from it to the _specus_ of the aqueduct; but as the bottom is filled up to the depth of several feet with broken stone, it cannot be seen. Some of the chambers of this ancient quarry have been used as a burying-place, probably for the neighbouring town of Collatia. The graves are cut out in the side of the walls, like the _loculi_, and there are also chambers like the _cubicula_ of a catacomb. The tombs of this necropolis of Collatia are very ancient, perhaps anterior to the Kings of Rome, and certainly not posterior to the Republic. This stone has been cut out in large square or oblong blocks, such as are usual in work of the time of the Kings; it is probable that Appius Claudius made use of a quarry which he found, and that the ground was not excavated for the purpose of the aqueduct only. From this point the course was entirely underground, therefore no traces are visible without some difficulty. Incidentally, as we have seen, Frontinus notes, that at the head the water ran as much as fifty feet below the surface, and was thus protected from being fraudulently diverted. It could not have taken a very direct line, as the levels of the crossing valleys must have necessitated here and there a circuitous course; and moreover this is clearly implied by Frontinus, who says that the total length from the source to the “Salinæ” was more than eleven miles, while the distance in a direct line is less than nine miles. As it had no _piscina_[16], it flowed on in an uninterrupted course to the spot where it was distributed for the public service. One branch seems to have entered Rome under the line of the Claudian arcade, near the Porta Maggiore[17], and its course within the city is not difficult to follow, since for certain purposes access has been made in more than one place into the old _specus_, part of which is used to carry the lead pipes of the Aqua Felice to the Hospital of the Lateran and other places. From the surface where one of the shafts is situated to the bottom of the channel, is a depth of twenty-five feet. No water runs along it at present, except in the metal pipes from the Aqua Felice, belonging to the modern system.