[ 349 ] 



As it is ufelefs to attempt to make this plainer *, I proceed 

 to explain the meafures taken to diminifli the quantity of air 

 left under the pifton, in this pump, at each ftroke. 



Suppose the cyV to be open at the top, or the valve to be 

 taken ofF, which is the fame thing ; then, as the pifton fits the 

 bottom of the barrel, there is no fpace left for lodgment of air 

 under it when it is put down and the communication with the 

 dudls is fhut, except the two little cavities in the plate at the 

 bottom, being the holes of the fhell of the cock lying over 

 the key ; and thefe will be always full of air of the common 

 denfity, at the inftant when the pump communicates with the 

 rec", for the external air prefles into the top of the barrel, which 

 is fuppofed open, and therefore into the end of the circula- 

 ting pipe which opens there ; confequently as the pifton cannot 

 be put down, unlefs the air under it, which it protrudes in its 

 dcfcent, be difcharged through this opea pipe, fo the outward 

 air having accefs to it, and to the bottom of the barrel (with 

 which it muft communicate during the whole dcfcent of the 

 pifton) will fill it and any void under the pifton ; but when the 

 pifton is at the bottom, and the key is turned, the air in the 

 pipe is cut off from that in the bottom of the barrel, none al- 

 moft remaining there, but what is in the little cavities over the 

 key ; which air will pafs into the reC on opening its dudl into 

 the barrel ; — but if, by a balf turn of the key flopping iotb 



duds, 



*l have proved this by calculation, and it will be farther explained hereafter. 



