but walls are to be built to conduct the water through the earthy or sandy soils. Wells also, or air-holes are to be cut from the top of the water-channel to the surface, for the purpose of allowing the air .which might accumulate in the aqueduct to escape. These wells are directed to be placed at a distance of 120 feet. If the water be conducted by leaden pipes, a castellum or reservoir is first built at the spring head, then the diameter and strength of the pipes being suited to the quantity of water, they are to be carried from the castellum to that which is in the city. The pipes are not to be less than ten Roman feet in length, and were named from the breadth of the lead before it was rounded into a pipe. The manner of conducting water by pipes is thus regulated ; if the spring head had a sufficient current to the city, and no higher hills intervened, the interval is, by walling, raised to a proper level, as mentioned in the description of channels of ma- sonry, or else a circuit round may be taken if not very long ; but if there be frequent valleys, the courses are to be directed down the declivities, and when arrived at the bot- tom, a sub-structure is to be built, but not high, that the libramentum, (or level, or coun- terpoise,) may be as long as possible — this will be the venter. "When arrived at the op- posite declivity, as on account of the length of the venter, the water swells gently, it is pressed upward to the top of the ascent ; whereas if the venter should not be made in the valley, nor a sub-structure built level, but should be bent, the joints would be destroyed. In the venter, also, columnarics. are to be raised, through which the force of the vapors may be dissipated. These columnariae are supposed to have been always open at the top, and to reach above the level of the aqueduct. It was not unusual, when the level from the spring head to the city was obtain- ed, to erect a castellum at every 200 actus* distance, that if damage should happen at any place, the whole work needed not to be taken down, and that the defective part might * The actus is 120 feet, according to Columella and Pliny, as quoted in notes to Vetruvius, p. 170.— [Eo.] PRELIMINARY ESSAY. 11 be the more readily found. But those castellums are to be built neither in the decur- sions, nor in the plane of the venter, nor in the pressures on the part of the aqueduct where the water is raised by the weight or pressure of the descending water, nor in any of the valleys, but always in the even plane. But when it was required to con- duct water at less expense, tubes of earthen-ware were made, having a thickness of not less than two inches, and these tubes were so formed that one end being tongued, the one entered the other — then the joints were cemented with quick lime, tempered with oil. In the descents, level with the venter, a stone of the red kind is to be placed at the angles, so perforated that the last tube of the decursions and first on the plane of the venter may be joined to the stone ; so likewise at the opposite acclivity, the last in the plane of the venter, and the first of the expressure are to be in the same manner united to the red stone. Thus the tubes on the even plane, as well as those in the decursions, will not be split, for such violent vapors are apt to rise in conduits of water as would even burst through stone, unless the water was at first gently and spa- ringly admitted from the spring, and the bendings secured with ligatures or weights of ballast ; in all other respects they are built in the same manner as leaden pipes. When first the water is admitted, ashes are sent before it, that if any of the joints should not be sufficiently cemented, they may be stopped by the ashes. Aqueducts of tubes have this advantage — if any damage happen, any person may rectify it, and water from earthen tubes is far more wholesome than that from pipes, as the use of lead is found to be pernicious. We should not, therefore, conduct water in pipes of lead, if we would have it wholesome. The taste also of that from the tubes is better, as is proved by our daily meals ; for all persons, although they have tables furnished with silver vases, use fictile ware on