February i6, 1893] 



NATURE 



;69 



falls down on account of its own weight out of the 

 pump Q back into the ball H. "JT 



The reversing of the three-way cock, and there- 

 with the automatic action of the pump, is effected in 

 the following manner : — The ball H rests on a frame D, 

 revolving about the axis h, and the motion of which is 

 limited by the ledges c and c-^. A lever G is attached 

 to the frame not far from the axis, and by means of a 

 peg, when the balance D reverses its position, also turns 

 the cock. When the ball H is entirely filled with mercury 

 the balance D rests on the upper ledge c. If the pump is 

 set in motion the left side of the balance D becomes 

 lighter in proportion to the amount of mercury forced 

 out of the ball H into the pump, until at last the weight C 

 on the right-hand side becomes heavier, and the balance 

 thereby attains the position shown by the figure. The 

 three-way cock is also reversed by this motion. 



the adjusting of the height of the mercury can be easily 

 and accurately done up to a centimetre. 



It goes without saying that every mercurial air-pump 

 not provided with cocks can be worked by the apparatus 

 just described. But the improved construction of the 

 Toepler pump, drawn likewise in projection in the figure, 

 has proved to be especially practical. The following is a 

 description of its automatic working : — 



If the cock /i is connected with an hydrostatic 

 air-pump, the ball Q of the pump and the space R, 

 which is to be evacuated through the tube s, is 

 pumped out up to the tension of the vapour. The mer- 

 cury then rises in the tube R almost to the height of the 

 barometer above its level in the ball H. If the automatic 

 apparatus is then set in motion, the mercury enters the 

 ball Q and the tube S, thus cutting off the connection 

 with R, while any further rising of the mercury 



«r 



Thus, as already described, the water current is now 

 cut off, the water present in M flows out through Kg, and 

 the mercury goes back from the pump Q into the ball H. 



During the tipping over of the balance, however, the 

 sliding weight C has run down its inclined plane to a 

 ledge, E, so that it now exerts a pressure on the lever 

 arm. Its momentum is so calculated that the mercury in 

 the pump must have fallen to the point/, and flowed back 

 into the ball H before it again overweighted, and moves 

 back the balance. The weight c then slides back again 

 to the left until it rests against its left ledge, and the play 

 of the pump recommences. It will easily be seen that 

 the height to which the mercury rises in the pump, the 

 mass of the sliding weight being a constant quantity, 

 depends only on its final positions, and that, therefore, 



NO. 12 16. VOL. 47] 



in the tube S is prevented by a glass valve v, it passes 

 through the first V-tube rj filling the little vessel r^ 

 and rises through s-^ into the ball n, driving before it 

 the air which was before shut off in Q. At this 

 moment so much mercury has been forced out of the 

 ball H into the pump Q, that the balance is turned, the 

 mercury flows back out of Q into H, forming vacua in r j. 

 and Q, as the little mercury-threads remaining in the side- 

 tubes rj and .y, form shut-off valves. As soon as the 

 mercury has fallen below the entrance-point of s into E 

 the pressure in R and Q become equal, the denser air 

 flowing out through S into Q. The time during which 

 Q is connected with R may be determined at will by 

 changing the right ledge of the sliding weight. Then the 

 balance again changes its position, the mercury rises in 



