The width of the tunnel excavation in rock is the same as that of open excavation in rock; and the manner of building the masonry to form the channel-way is the same, with the exception that the rock roof of the tunnel serves as the roof of the channel-way, where it is sound, but in cases where the rock is soft and liable to fall, a brick arch is built over the channel-way, and the space between its extrados, or outer surface, and the rock roof is filled with earth closely rammed in. In some instances where the tunnel perforated rock which was at first quite hard, the roofing has by exposure to the air, become soft and insecure, so as to render it necessary to turn an arch for its support. This is attended with inconvenience and some difficulty after the channel-way has been completed and closed through the tunnel. Plate IV. is a section of the Aqueduct in tunnel cutting in earth. When the earth is dry and compact, the excavation for the bottom and sides is made of a proper form to receive the masonry, which is built closely against it: the top is excavated sufficiently high to give room to turn the arch, and the space above is afterwards filled with earth closely rammed in. Where the earth is wet and there is difficulty in making it stand, the excavation is made larger, and props of timber and plank are used to support the top and sides until the masonry be completed; and the whole space exterior to the masonry is then compactly filled with earth. Plate V. is a section of the Aqueduct showing the manner of constructing it across valleys, or where the natural surface of the ground falls below the plane of grade. In such cases the Aqueduct is supported upon a foundation wall of stone laid dry, and formed by using large stones laid in positions to give proper bond, and to allow small broken stone to be closely packed in, filling up all the interstices so as to form a compact and uniform mass. The wall is generally allowed to stand some months after it is completed, before the masonry of the Aqueduct is commenced upon it, lest by this weight being placed upon it before it has found its bearing, it should settle and cause cracks in the masonry. That such settlement should in some instances occur, even after the Aqueduct is completed, is not surprising, for passing over so many different elevations, and encountering such numerous transitions from a hard soil, or from rock, to valleys of alluvial deposit, it would be beyond human powers of foresight and vigilance to prevent it. [Illustration: V _F. B. Tower._ _Gimber._ ]