210 ON DIVERGING S'FREAMS 



beam by a string, and balanced by weights placed 

 in a scale E, attached to the opposite end of the 

 beam. The valve being thus placed on the seat 

 without any weight of its own to press downward, 

 the stream of compressed air was admitted into 

 the pipe A, when the valve D, rose from the flange 

 or seat C, l-32nd of an inch, and there remained 

 stationary. Thirteen ounces, avoirdupoise weight, 

 were now put into the scale E, which raised the 

 valve to 1-1 2th of an inch above the seat. 

 Twenty-six ounces raised it to l-8th of an inch, 

 and thirty- two ounces raised it to l-4th of an 

 inch, but any weight beyond this last caused the 

 valve to fly abruptly oft^. 



It thus appeared, that when the valve was 

 raised from its seat a quarter of an inch, there 

 was the greatest difference between the force of 

 the issuing current of air pressing against the 

 under side of the valve, and of atmospheric 

 pressure on the upper side of the valve. The 

 pressure of the atmosphere was greater than the 

 force of the issuing stream of previously com- 

 pressed air, a weight of thirty-two ounces being 

 requisite to establish an equilibrium. 



That we might ascertain what was the state of 

 the stream of air under the valve, in different 

 parts of it, four double syphon tubes were 



