Discharge of Electricity through Gases. 



325 



end. Its internal diameter was 15 centitns., its height 38 centims. 

 The open end was closed by means of a brass disk into which a 

 groove bad been cut. The rim of the open end of the cylinder was 

 cemented into this groove. A circular hole 5'5 centims. in diameter was 

 cut into the centre of the brass disk, and another glass cylinder was 

 fitted into this hole. The negative electrode was formed by a piece of 

 aluminium foil 8 centims. long, which was wrapped round the inner 

 cylinder. The electrode had a surface therefore of over 120 square 

 centims. A long aluminium wire running parallel to the axis of the 

 aluminium cylinder at a distance of about an inch formed the positive 

 electrode. The aluminium cylinder was connected metallically to the 

 brass disk, so that the positive electrode had to be insulated from it. 

 This was done by passing it through a glass tube which was cemented 

 into the brass plate. This tribe also formed the opening through 

 which the exhaustion could be carried on. The aluminium wire had, 

 of course, to be connected with a platinum wire sealed into the tube. 

 The accompanying figure shows a section of the vessel. 



Fig. 1. 



As an air-pump I have used Hagen's form of the Toepler mercury 

 pump. As the vessel to be exhausted was rather large, the exhaustion 

 was a little laborious, and had to be carried on with great care. I 

 need hardly add that the effects to be described are not the result of a 

 single experiment, but that the vessel was often refilled and exhausted 

 in the coui-se of them. I shall now describe what is observed on 

 exhaustion. 



At first the discharge, of course, only passes as a spark. It runs 

 irregularly along the wire and passes to the nearest point of the 

 aluminium cylinder. Gradually the glow settles down on the cylinder, 

 if that is negative, as I shall assume it to be. The glow is quite 

 irregular and not always at the same place, but a tendency to avoid 

 the place opposite the wire is already apparent. In fact the effect 



