412 Prof. A. P. Chattock on the Velocity and 



since been found that they are subject to two sources of error? 

 both of which tend to lower V. Firstly, the ring was too 

 near the wall of the tube and so impeded the return current 

 of gas; and, secondly, the wire of the point B was uncovered 

 for more then a centimetre along its length. That this was 

 also responsible for some of the excess of pressure was 

 suggested by experiments with ( J0 2 in which the wire became 

 discoloured for some millimetres at its end, showing that dis- 

 charge had occurred from the sides of the wire as well as 

 from the point. Such sideway discharge would tend to creep 

 farther back from the point as z was increased, and this would 

 make the pressure-curve too steep. The following results of 

 two experiments made with the same point alternately bare 

 and covered with thin glass to within a millimetre of its end 

 illustrate this : — 



v+. 



V-. 





330 

 419 



406 

 496 



point bare, 

 point sheathed. 



The containing-tube, E, was of diameter 2'44 centim. in 

 this case. 



It is possible that there may have been a trace of this 

 effect present in the experiments with the " hole in plate " 

 apparatus ; but the point was in such a strong parallel field 

 that it could hardly have been marked. 



Even with the glass sheath there was opportunity for a 

 certain amount of side discharge as it was impossible to extend 

 the sheath right up to the point without increasing the 

 potential of discharge to an inconvenient extent. It was 

 therefore important to work with large values of z so that the 

 position of the discharge region on the point might be constant. 

 With the two E-tubes mentioned above this was not possible 

 as the effect of their walls on the discharge became apparent 

 when z was increased to a couple of centimetres or less. I there- 

 fore constructed a third apparatus, of which the inner diameter 

 of the glass tube was 6*26 centim., and the length 34 centim. 

 With this the readings for curves IV. were taken as well as 

 all others on which I rely. The ring (mean diam. = 2"5 centim. 

 and thickness 0*4 centim.) v\as of burnished brass, and in 

 most of the experiments there was a thin layer of vaseline on 

 its surface to catch any dust that might be deposited by the 

 current. All glass surfaces were carefully washed with soap 



