(Continued from page 136 ) AIR VESSELS IN PUMPSmdash;Some experiments have been made by Messrs. Kirchweger and Prusman, engineers, of Hanover, on the positive effect produced upon the action of pumps by the application of air vessels on the suction pipes. Air vessels have been applied for many years on delivery pipes, but it is only lately that their value has been properly estimated, although it is obvious that it is of as much importance that the pump should befill-ed with water, as that the delivery should be constant. The apparatus employed by the German engineers is represented in section in fig. 1. A is a reservoir, which represents the source whence the pump draws its water; B is the suction pipe, and C is a valve-chest, containI ing a ball-valve, surmounted by a cock discharging at the side. The plug of the cock is stationary, whilst the shell is moved by the handle, E. D is the air vessel. Fig. 2 shows the details of the valve on a larger scale. It is obvious that, by causir.gthe cockto revolve by means of the handle, E, a certain volume of water will escape each time the passage is opened, the height of water column in the pipe, E, answering to the pressure of the atmosphere in causing the water to fill the pump. The result ot the trials was, that; when the air vessel was removed, and the opening stopped, an increased velocity of rotation of the cock gave less water ; but with the air vessel the increase of velocity gave more water. The trials were made with different speedsand different pressures of water, with the results shown in the following table :mdash; Gallons of water delivered per No. ofturnsper min.undera mean pressure of minute. 17 ft. 12J ft. 8Jft. 2 ft With air vessel. 80 129 12-78 8'79 283 100 156 1543 11-25 4'82 120 17-15 1663 12-23 S'44 140 18-28 16-75 1298 554 Without air vessel. 80 9'45 862 6902 2.36 100 803 808 6-05 T98 120 6-55 6 54 5'42 1-88. 140 5-42 6 29 517 151 The capacity of the air vessel is 68 cubic inches.The weight of the ball valve 2315 lbs. The area of the valve seat=11.5 inches. The smallest diameter of the feed pipe is l-48 inches. The quantities delivered at 80 to 100 turns are the mean of tour trials ; those of 120 and 140 turns are the mean of 3 only. If these trials are to be taken as the exact result which may be expected under similar circumstances with a pump, it is evident that a large increase of duty may be expected, by adding an air vessel on the suction side of a pump, working at a high speed. For, it will be observed that, whilst at 80 turns the increase is only 20 per cent, at 100 turns it is 133 per cent., at 120 turns 189 per cent., and at 140 turns 266 per cent.