Pupin — Action of Vacuum Discharge Streamers. 265 



but the oscillations were always in the plane common to their 

 rectilinear paths. 



The pressure was still increased by turning the stopcock 

 once around and allowing some more air to get in. The dis- 

 charges became still less diffuse along their path through the 

 reservoir and when allowed to pass one at a time they were 

 quite steady and rectilinear. Both being allowed to pass 

 through the reservoir simultaneously they were appreciably 

 concave toward each other, but the concave arcs were in the 

 plane passing through their shortest path in the reservoir. 

 There was evidently a repulsive force acting between the dis- 

 charges. This repulsive force increased continually with the 

 increase of pressure in the vacuum, but with the increase of the 

 force the steadiness of the discharges along their paths through 

 the reservoir when they were both passing simultaneously 

 gradually diminished. Fig. 3 (Plate XII) is a photograph of 

 the discharges when taking place simultaneously in a vacuum in 

 which the repulsive force just described was moderate. Fig. 

 2 is a photograph of the same discharges, but when taking 

 place one at a time. Fig. 2 indicates that there is a repulsion 

 between the discharge and the nearest wall of the reservoir, 

 but this repulsion is very small in comparison to the repulsion 

 acting between the streamers, when they take place simulta- 

 neously, as fig. 3 indicates. The photographs were taken by 

 very short (i to \\ second) exposure, so strong is the luminosity 

 of these discharges. They were taken when the repulsion was 

 moderate, because when the repulsion is very strong the dis- 

 charges oscillate very rapidly, so that they could not be well 

 photographed with the apparatus employed. When the repul- 

 sion was so strong as to cause the streamers to curve way out 

 and from time to time strike the nearest wall of the reservoir, 

 the vibrations became very violent every time the streamers 

 struck against the walls. They rebounded against the walls 

 just as if they were luminous vibrating strings. 



At this pressure in the vacuum it was very difficult to start 

 the discharge and I had to strain the alternating machine and 

 the induction coil to their utmost to make the start.* But 

 when once started the discharge continues without any inter- 

 ruption even if the potential at the tinfoils is lowered 15 to 20 

 per cent. 



The rate at which the temperature of these discharges in- 

 creased with the increase of pressure in the vacuum seems to 

 be much more rapid than the increase of the pressure ; (the 

 impressed e. m.f'., of course has to be increased with the pres- 

 sure) and with the increase of the temperature the luminosity 



* A gentle touch, of the apparatus with a conductor will sometimes start the 

 discharge at once when all other means of making this start fail. 



