ON HYDRAULIC MACHINES.

333

is placed within the piston itself, so that the same barrel serves for the ascentof the water, which rises in one continued line, while the piston is raised,and rests on the fixed valve while it is depressed. The velocity of the strokeought never to be less than 4 inches in a second, nor greater than two orthree feet; the stroke should also be as long as possible, in order to avoidunnecessary loss of water during the descent of the valves. The diameter ofthe pipe, through which the water rises to the barrel,ought not to be less thantwo thirds of the diameter of the barrel itself. (Plate XXIII. Fig. 309.)

A bag of leather has also been employed for connecting the piston of apump with the barrel, and in this manner nearly avoiding all friction: but itis probable that the want of durability, would be a great objection to such amachine. (Plate XXIIT Fig. 310.)

Where the height, through which the water is to be raised, is considerable,some inconvenience might arise from the length of the barrel through whichthe piston rod of a sucking pump would have to descend, in order that thepiston might remain within the limits of atmospheric pressure. This may beavoided by placing the moveable valve below the fixed valve,and introducingthe piston at the bottom of the barrel. Such a machine is called a lifting pump:in common with other forcing pumps, it has the disadvantage of thrusting thepiston before the rod, and thus tending to bend the rod, and produce anunequal friction on the piston, while, in the sucking pump, the principal forcealways tends to straighten the rod. (Plate XXIII. Fig. 311.)

The rod of a sucking pump may also be made to work in a collar of leather,and the water may be forced through a valve into an ascending pipe. Byapplying an air vessel to this, or to any other forcing pump, its motion maybe equalised, and its performance improved ; for if the orifice of the airvessel be sufficiently large, the water may be forced into it, during the strokeof the pump, with any velocity that may be required, and with little resistancefrom friction, while the loss of force, from the frequent accelerations andretardations of tlffc whole body of water, in a long pipe, must always be con-siderable. The condensed air, reacting on the waiter, expels it more gra-dually, and in a continual stream, so that the air vessel has an effect analo-gous to that of a fly wheel in mechanics. (Plate XXIII. Fig. 312.)