308 On Electrolytic Convection in Geissler-tubes. 



in the diagram, pieces of platinum wire fused into it at the 

 points A, A', B, B' being used as electrodes. Hydrogen and 

 a little chlorine were allowed to. enter the tube, and the elec- 

 trodes A and B connected to the terminals of an induction-coil 

 giving a spark about 6 centim. long. 



With a mixture of the gases containing 7 per cent, of 

 chlorine at a pressure of 2 millim. mercury, the spectrum of 

 chlorine was obtained from a small portion of the tube sur- 

 rounding the anode, while in the rest of the tube no chlorine- 

 could be detected. The current was repeatedly reversed, 

 and the effect described by Prof. Thomson observed. 



The electrode B was then made the anode, and no chlorine 

 being detected on careful examination of the part AA' the 

 two parts AA' and BB' were fused off at the points C and D 

 whilst the current was still passing. 



Making use of the remaining electrodes at AJ and B', each of 

 the tubes thus obtained was connected to the induction-coil 

 and examined with a spectroscope. 



If Prof. Thomson's theory were correct no chlorine should 



