Fiar. 2. 



Electric Intensity and Conductivity in Rarefied Gases. 507 



At W a tube 1 cm. in diameter and 5 cms. long was joined 

 on at right angles to the discharge-tube and two platinum 

 wires W, W were sealed in at the end of it. These wires 

 extended to the centre of the discharge-tube, and were coated 

 with glass except for 2 millims. at their extremities, which 

 were one above the other and 1*5 millim. apart. The wires 

 are shown on a larger scale in fig. 2. The difference of 

 potential between W and W was measured with a well 

 insulated quadrant electrometer. 



A telephone, galvanometer, and ad- 

 justable resistance were included in the 

 discharge circuit, and during a series of 

 observations the current through the 

 tube was kept at a constant value. 



The tube T was sealed on to a Topler 

 pump, McLeod gauge, and an arrange- 

 ment for introducing any gas into the 

 apparatus, &c. 



It was found that when a discharge 

 was passing the wires W, W took 



up their final potentials immediately, and that the electro- 

 meter-reading was not appreciably affected even by con- 

 necting either or both of them to earth through a very 

 high resistance such as a piece of damp thread, which of 

 course immediately discharged them when no discharge was 

 passing. The insulation between the wires W, W was very 

 good, and they were able either separately or together to 

 hold a static charge for a long time. 



A millimetre-scale backed by a mirror was attached to the 

 discharge-tube, and the position of the electrodes E, E' with 

 reference to the wires W, W read off on it. 



The method of obtaining a series of observations was as 

 follows : — The tube was first filled with the gas at the required 

 pressure and the discharge started, if necessary, by means of 

 an induction-coil. The current was then adjusted to the 

 required value and its steadiness tested with the telephone. 

 When a steady current of the required value had been ob- 

 tained in a sufficiently pure gas at the required pressure the 

 potential-difference between the wires W, W was observed 

 with the electrodes E, E' in a series of positions from that 

 with the electrode E in contact with the wire W to that with 

 the electrode E' in contact with the wire W. The current 

 was kept constant during the series of observations, and unless 

 the gas pressure remained practically unchanged whilst the 

 series was being taken the observations were rejected. 



It was found very difficult to obtain entirely satisfactorv 



2 N 2 



