368 Prof, E.Wiedemann on the Behaviour of Gases 



in two portions connected by a ground joint, tho form of 

 which is that recommended by Qimmingham *. The lower 

 portion of the tube is provided with a funnel-shaped portion 

 which is filled with mercury; so that if the joint becomes a 

 little loose it is not necessary to stop working. 



The lower end of the tube b is connected at o with one end 

 of an india-rubber tube, the other end of which communicates 

 with a vessel of mercury 0, which can be raised or lowered 

 by a pulley not shown in the drawing. 



The action of the pump is extremely simple. The vessel 

 is first raised and mercury allowed to enter the bulb A, by 

 which the communication between A and B is interrupted. 

 The air is driven out of A, and escapes through the mercury 

 in L. When a certain quantity has escaped, the vessel is 

 lowered. The air is now drawn out of B and the parts of the 

 apparatus communicating w T ith it, into the bulb A. This takes 

 place in a somewhat violent manner, since large bubbles of 

 air enter the exhausted space above A and throw the mercury 

 about; so that too much air must not be driven out of A 

 at first f. When the exhaustion has advanced somewhat, 

 another precaution is necessary, since if the flask were raised 

 too high the mercury might be driven violently into A, and 

 might possibly cause fracture, a result less to be apprehended 

 if the tube B is of sufficient width. The pump has the great 

 advantage that it is constructed entirely without taps, it works 

 very easily and surely, and gives a very good exhaustion. 

 Tubes may easily be prepared by its means which show the 

 phenomena discovered by Hittorf \ and recently further 

 described by Crookes § . 



The tubes to be exhausted were connected with the pump 

 by means of a joint on the further side of the tap 8. In 

 order to make connexion of the different pieces of apparatus 

 with the pump, glass tubes of greater or less length must be 

 employed connected with each other by joints. In order 

 to make a secure joint, not liable to be affected by the 

 shaking which occurred in pumping, the two parts a and 

 b of the joint (Plate IX. fig. 6) were connected by an india- 

 rubber band e passed round side-tubes c and d. Any acci- 

 dental loosening was thus immediately remedied, and it was 

 unnecessary to cover the joint with wax||. 



* Gimmingham, Beibl. i. p. 175, 1877. 



t For this reason, the pump employed in the physical laboratory at 

 Berlin has been modified so that the points Q and 8 are connected by an- 

 other tube. The shocks are therefore much less powerful. 



\ Hittorf, Pogg. Ann. cxxxvi. p. 8, 1869. 



§ Crookes, Beibl. iii. p. 527, 1879. 



|| The whole of the glass apparatus employed was manufactured in an 

 admirable manner by Mr. Gbtze, glass-blower in Leipsic. 



