326 INFRA-RED EMISSION SPECTRA. 



For this second effect the energy increases with the potential gradient, 

 since the radiation from the gas is caused by the colHsions between 

 molecules and electrons, which latter are moving with high speeds. As 

 already mentioned in the high temperature radiation, where the collision 

 is between molecules, the energy increases with the temperature. The 

 electrical temperature distribution will be different from the thermal, 

 and Kirchoff's law for the relation between emission and absorption 

 will not hold for radiation from a gas in a vacuum-tube. The spectral 

 distribution of intensity will be a function of the velocity distribution 

 of the electrons. The greater the number of electrons with high speed 

 in a volume element, the more will the intensity maximum be shifted 

 towards the short wave-lengths. 



For this reason the cathode glow is blue, since the cathode fall is 

 about 300 volts. On the other hand, in the position column, where the 

 fall is only about 30 volts, the light emitted is red (nitrogen). Accord- 

 ing to Stark's (loc. cit.) computations, in which the kinetic energy is 

 equated to the temperature, this would indicate an electrical tempera- 

 ture of some 6,000° for the cathode glow. 



The ionic energy^ (the minimum kinetic energy) necessary to disrupt 

 an atom is in the order, metals Hg (8 volts), H, N (27 volts), and O. 

 From this it would follow that if we mix H with N the fall of potential 

 and the electrical temperature of the positive column will be changed. 

 The temperature will be increased if the gas to be added has a higher 

 ionic energy, £?. ^., N to Hg., or N to H. (Heuse,' Herz). This will 

 be of interest in comparing the relative intensities of the 4.75 /x band of 

 COo when it occurs as an impurity in N (very intense) and in O or 

 NH3, where it is weak. 



These views will now be briefly considered in connection with the 

 results obtained in the present research. 



Prior to this investigation on vacuum-tube radiation only one type of 

 selective emission of gases in the infra-red had to be accounted for, viz, 

 emission bands of water vapor and CO,. They were thought to be due 

 to the thermal temperature of the gas. However, the data bearing upon 

 this subject are so scarce that writers, in referring to them, generally 

 expressed their opinions rather cautiously. Some have vaguely inti- 

 mated that it might be something similar to luminescence in the visible 

 spectrum — call it thermalescence. 



From the present research on the intensity of the infra-red emission 

 bands of N and CO,, for constant current and var}ing pressures, and 

 vice versa, it becomes evident that we have to deal with two distinct 



^Herz: Ann. der Physik (3), 54, p. 244, 1893, 

 ^Heuse ; Verb. d. d. phys. Ges. , i, 269, 1899. 



