Automatic Mevcurij V(icinim-pa}}ip. iUU 



B, aiul that in E and D comes back to the reservoir by the 

 j)ath//" (sealino- columns beino- left in ee and dd'),u\\i\\ 

 the contact U V is made when the tap is again turned. 



In the subsequent strokes the action dei)ends on the state 

 of the gauoe G at the time when the mercury is passing 

 through </</'. So long as the pressure ot! the air in D is 

 under ^ cm. of mercury the air in q will reach below the 

 level of the electrodes YY', and break the circuit through 

 Br. The contact Q is consequently unaffected, and the circuit 

 P U X X' Q K' /' / will become closed, and the three-way taj) 

 turned, as soon as the mercury has filled up D to the level of 

 X X'. With this state of the gauge the air is thus expelled 

 only into D. As soon, however, as by the accumulation of 

 air in D the pressure there has risen above the limiting- 

 amount, the contact Y Y' is made by the rise of the mercury 

 in q, Q is broken, and the three-way tap is consequently not 

 turned until the mercury from the pump-bulb has expelled 

 the accumulated air in D into the external vacuum-chamber, 

 and has closed the circuit P U Z K' /' /. 



On stoppage of the water-pump the mercury simply rises 

 in the apparatus to some point of E depending on the pres- 

 sure of the air in EE^, and remains there permanently, sealing 

 off the vacuum-tube against any leakage of air through the 

 pump. With the form of stopcock M, of which a description 

 has already appeared in this Magazine^, leakage into EE' is 

 practically nonexistent, and a partially exhausted tube can 

 thus be left any length of time in connexion with the pump 

 without its vacuum being affected. 



Rate of Working. 



The rate of working of an instrument of this kind of course 

 depends chiefl}' on the efficiency of the water-pump. Apart 

 from this the best effect is olrviously obtained wdien the con- 

 nexion c to the A-acuum-tube is as high as possible above the 

 reservoir A, as the latter then requires exhausting to the least 

 extent. A limit to the height is, however, set by the fact 

 that the height from the mercury in A up to the top of d or e 

 (whichever is the higher) plus the vertical height of ee^ must 

 be under 76 cms. for pumping to go on at all stages of the 

 exhaustion. The addition to the apparatus of a Winchester 

 \y, acting as an air-reservoir to the water-pump, has also a 

 beneficial effect on the efficiency. On the downward stroke 

 the air in A now empties itself quickly into W, whence it is 

 carried off more slowly by the water-pump. The result, with 

 a reservoir of suitable size, is that the mercury in B falls very 

 * Phil. Mag. July 1903, p. 78. 



