An Automatic Merc m- If ]\(()(U))i-j>wnj:>. ol7 



advantage is that the current is only on while the ta[) is being 

 turned, and there i.s thus no waste of electrical energy. 



As is well known, for an efficient action of the Topler 

 pump the air from the pum])-bull) must first be exi)ressed 

 into a fairly highly exhausted ''auxiliary chamber/^ D, and 

 from D finally, when a sufficient amount of air has been 

 accumulated there, into the not very high external yacuum 

 EE' which is all that can be produced by a water-pumj). 

 Without some arrangement of this kind the impossibility of 

 driving the very minute air-bubbles, trapped when a high 

 stage of exhaustion is reached, down a mercury column neces- 

 sarily several centimetres long soon puts a limit to the degree 

 of vacuum obtainable. Hence results the necessity of ar- 

 ranging for two classes of pump-stroke in an automatic in- 

 strument, one — which we may call a " partial '* stroke — 

 expelling the air from B into D only, and the other — a 

 '* complete " stroke and requiring a longer time — in which 

 the mercury goes right over into E, and expels into it the 

 air previously accumulated in D. The necessity for complete 

 strokes, which is continual at first, falls off very rapidly with 

 the exhaustion, and the loss of time produced by allowing 

 complete strokes to occur unnecessarily becomes considerable 

 in the long run. The theoretical requirement for the least 

 waste of time is clearly that the air in the auxiliary chamber 

 should be expelled when, and only when, it has reached a 

 certain limiting pressure — say half a centimetre of mercury ; 

 this condition_, wdiich it is difficult to satisfy in the usual form 

 of automatic pump, can be exactly satisfied in an electrically 

 controlled apparatus. It is accomplished in the instrument 

 bv the gauoe Gr attached to D, workino- in connexion with a 

 small contact-breaker B?\ 



The gauge consists of the two tubes p and ^joined by a short 

 tube r at the bottom ; in q two electrodes Y Y' are sealed 

 about ^ cm. above the entrance of r. The limb q contains 

 just so much air as will extend down as far as the entrance 

 of r, when it is under the pressure of the mercury column 

 jjr only, i. e. when there is a perfect vacuum in D. When 

 the pump is set up and used for the first time it is clear that, 

 after two or three complete strokes have been carried out, 

 the gauge will automatically empty itself until just this 

 quantity of air remains permanently trapped behind. The 

 action of the gauge is that, whenever the contact Y Y' is 

 made by the rise of the mercury in q, a current passes 

 through the coils of the contact-breaker B?', which is thus 

 caused to break a contact Q in a circuit through the two 

 electrodes XX' in D and the tap coils K'. The effect of 



