ON HYDRAUtlC MACHINES. 356 



wholly intermitted, is, however, by no means uniform in its velocity. 

 (Plate XXIII. Fig. 314. . 317) 



The pipes, through which water is raised, by pumps of any kind, ought to 

 be as short and as straight as possible ; thus, if we had to raise water to a 

 height of 20 feet, and to carry it to a horizontal distance of 100 by means 

 of a forcing pump, it would be more advantageous to raise it first vertically into a 

 cistern 20 feet above the reservoir, and then to let it run along horizontally, or 

 find its level in a bent pipe, than to connect the pump immediately with a 

 single pipe carried to the place of its destination. And for the same reason 

 a sucking pump should be placed as nearly over the well as possible, iu 

 order to avoid a loss of force in working it. If very small pipes are used, 

 they will much increase the resistance, by the friction which they occasion. 



Water has been sometimes raised by stuffed cushions, or by oval blocks of 

 wood, connected with an endless rope, and caused by means of two wheels 

 or drums, to rise in succession in the same barrel, carrying the water in a 

 continual stream before them ; but the magnitude of the friction of the 

 cushions appears to be an objection to this method. From the resemblance 

 of the apparatus to a string of beads, it has been called a bead pump, or a 

 paternoster work. When flat boards are united by chains, and employed 

 instead of these cushions, the machine may be denominated a cellular pump; 

 and in this case the barrel is usually square, and placed in an inclined posi- 

 tion, but there is a considerable loss from the facility with which the water 

 runs back. The chain pump generally used in the navy is a pump of this 

 kind, with an upright banel, through which leathers, strung on a chain, are 

 drawn in constant succession ; these pumps are only employed, when a 

 large quantity of water is to be raised, and they must be worked with con- 

 siderable velocity in order to produce any effect at all. Mr. Cole has im- 

 proved the construction of the chain pump, so as materially to increase the 

 quantity of water raised by it. (Plate XXIII. Fig. 318.) 



It is frequently necessary to procure alternate motion in pumps by means 

 of wheelwork, and for this purpose the application of a crank is the most 

 usual and perhaps the best method. Provided that the bar by which it acts 

 be sufiiciently long, very little will be lost by the obliquity of its situatioji, and 



■ - ■ ( 



