46 



Messrs. Harvey and Hird on 



and each convolution separated from the next by a strand 

 of silk equal in diameter to the wire. 



The secondary was suspended by silk cords, so as not to be 

 in contact with any part of the primary, and the whole was 

 immersed in oil, in order to prevent the passage of brush- 

 discharges. 



With this apparatus vivid brush- discharges are obtained. 



It was observed that the brushes were most brilliant when 

 allowed to take place between a point and a plate, especially 

 if the latter had a fair capacity ; also that the plate was 

 always charged positively, although the discharge was of 

 course oscillatory. 



In order to put this fully to the test the experiment was 

 varied in many ways. The current in the primary of the 

 induction-coil was reversed, both mechanical and automatic 

 breaks were used, and the two ends of the secondary of the 

 transformer were in turn connected to the point (the unused 

 end remaining immersed in oil), but the result was always the 

 same, the plate being strongly charged positively. We may 

 add that the charge was steady and, as near as we could tell, 

 equal in amount with any of the above combinations. 



We now added to our apparatus in order to test these brush- 

 discharges in various gases. For this purpose a bell-jar 

 (fi.g. 2) was fitted with a rubber bung, through which passed a 



Fig. 2. 



glass tube with a platinum point sealed in the lower end ; 

 from this point a wire leads into an upper tube and terminates 

 in a small clip for making contact with either end of S. 



Both tubes were entirely filled with oil, this being the only 

 means we found of preventing brushes onto the glass of the 

 bell-jar, which produced local charges and obscured the result. 



The jar stood in a deep glass dish containing mercury, the 

 level of which could be adjusted by means of a syphon (not 



