394 



Prof. Thomson, On some experiments on the [May 10, 



used. With these vapours there was always at low pressures a 

 glow stretching across the space between the electrodes, and 

 though the disc near the negative electrode was distinctly brighter 

 than the rest, I was never able as in air to get the discharge 

 practically confined to the disc. The brightness of the glow was 

 always comparable with that of the disc. 



The stratifications of the discharge followed the same law as 

 in coal-gas; in the parts of the discharge where the field was 

 uniform no tendency to stratification could be detected, but any 

 secondary discharge started by some accidental inequality was 

 always distinctly striated. 



With the arrangement described above, the pressure could not 



Fig. 5. 



„ G H 



be reduced below that due to about ^ of a millimetre of mercury, 

 and it was with great difficulty that it was reduced as low as this, 

 so in order to investigate the phenomena at higher exhaustions 

 the parallel plates were put in the apparatus represented in 

 section in fig. 5. 



ABGDEF is a brass bed plate with a groove of the shape 

 CDEFG cut in it ; into this groove the glass vessel S which 

 terminates in the tube T fits, and is fastened against the side BG 

 of the groove by marine glue ; mercury is poured into the groove, 

 and the space outside the glass vessel between the mercury and 

 the brass is kept exhausted by a Sprengel pump, which is con- 

 nected with this space by means of the tube PQ. Thus since the 

 pressure in this space can easily be kept by the Sprengel lower 

 than the pressure due to 5 millimetres of mercury, if the depth 

 of the mercury in the groove be greater than 5 millimetres, no 

 air can pass into the vessel, however low the pressure inside may 



