On the Principles of Pump-JVork. 227 



large; and when T had finished it, I found it answer all my 

 expectation, and, indeed, beyond what the common theory 

 of pump-work required. 



For, as the theory has hitherto considered a pump consist- 

 ing only of a single barrel and piston upon the sucking- pipe, 

 it supposes that no water rises into the barrel but while the 

 piston is ascending; which is true in all common or sino-le 

 pumps ; but in a double pump, or that here proposed with 

 two baireis upon one pipe, the case is otherwise; for the 

 water not only rises while the piston rises, but continues to 

 do so even after the piston begins to descend ; and therefore 

 we soon found the pump delivered more water than was ex- 

 pected from calculation. 



To account for this hydraulic paradox, it must be consi- 

 dered, that as this pump has both its large pistons workino- 

 (alternately ascending and descending) at the same time, 

 there must be produced a constantly rising column of 

 water in the pipe, whose velocity throuch a bore of five 

 inches to supply the barrels of twelve inches diameter each, 

 nmst be so great, that it cannot be checked or stopped at 

 once, or upon the first descent of the piston, and therefore 

 a surplus of water must be produced, as we found there ac- 

 tually was, by the experiment. 



This is therefore an argument which alone proves the 

 superior perfection of this construction of a pun)p, as such 

 an effect can never be produced in any other. • 



How deservedly this pump may be considered as without 

 friction, will appear from its own nature, and from experi- 

 ments that have been made upon it. With respect to the 

 nature and construction of the piston, it is evident that no 

 friction can arise from thence; for, while any body is in a 

 fluid, it.is affected by a pressure every way equally, and is 

 therefore in the same condition as if it was not acted upon 

 by the fluid at all ; that is, it will be absolutely free to move. 

 Now this is the case of the leathern pistons in this pump ; 

 they touch not the surface of the barrels, but produce their 

 effect by a motion of the several parts in water only, and 

 therefore entirely preclude even thi; very possibility of Iric- 

 tion from that f[uarter. 



The only source of friction, then, must be in the axle, 

 arising from the pressure of the lever, pistons, &.c. But 

 how very small and inconsiderable this is will be evident 

 Irom two experiments made on a pump at lai-gc, viz. the 

 lieight sixteen feet, the dian)etcr oC the piston twelve iuchi-S, 

 ^nd the length of the stroke nine inclies. 



P 2 Experiment 



