36 J. E. Moore— Electrical Discharge from the point 



distribution of the flow lines will be very much modified. 

 Although the actual determination of the lines of flow would 

 be a somewhat complicated problem, jet it is quite plain that 

 the linos of flow in the region of the edges of the spherical 

 cup would almost wholly vanish, and that the lines from the 

 central portion would appear more crowded together. If be- 

 side the cylinder surrounding the spherical cup and heated 

 near the cup, there be a region of very low temperature op- 

 posite the concave side of the cup, the lines of flow from the 

 central portion of the cup will be still more condensed. It is 

 also clear that if the gaseous medium in the vicinity of the 

 spherical cup, move in streams from the walls of the surround- 

 ing cylindrical surface towards the spherical cup, with very 

 considerable velocity, this motion of translation of the medium 

 will also affect the form of the heat-flow lines from the spheri- 

 cal surface. When this velocity of the medium becomes equal 

 to the velocity of the molecules transmitting heat along the 

 flow lines, then those flow lines where this equality occurs will 

 vanish. 



The above considerations relating to the modification of 

 heat-flow lines for a hot body immersed in a cool gas, plainly 

 suggest causes for the remarkable diminution of size and 

 change of form of a discharge stream from the cathode in an 

 ordinary discharge tube, as the pressure of the gas is reduced. 



The considerations, and the experiments described in § 5 

 made it appear that a discharge stream from a cathode is 

 but one part of the cycle performed by the gaseous medium in 

 the operation of transmitting electrical discharge by discharge 

 streams. The energy transmitted by these streams, being de- 

 pendent on the number and velocity of the molecules forming 

 the streams, it is clear that, as the pressure or molecular den- 

 sity of the gas is reduced, the velocity of translation of the 

 molecules throughout their cyclical course, must increase, in 

 order to transmit the same energy. It thus appears that 

 the velocity of flow of the gaseous medium in the return 

 stream to the cathode, increases as the molecular density of 

 the gas decreases. It was also shown in § 5, that sec- 

 ondary discharge streams can spring from regions electri- 

 fied by discharge streams from the cathode ; and one ought to 

 expect these secondary streams to spring from surfaces electri- 

 fied in any other way. That the walls of a discharge tube 

 near the cathode are electrified there can be no doubt, for not 

 infrequently the glass walls of the tube break down at this 

 point under the electrical stresses, when the pressure of the 

 gas is very low. It ought thus to be found (due to the electri- 

 fied walls near the cathode acting as sources of secondary 

 discharge streams), that the stream from the cathode dimin- 



